The present disclosure relates generally to the field of drilling operations, particularly to monitoring and optimizing the same using surface measurements and the estimation techniques taught herein. More particularly, the present disclosure relates to methods to estimate the effective vibration attributes of the bottom of the drill tool assembly, such as at or near a drill bit, based on evaluation of selected surface operating parameters.
This section introduces various aspects of art that may be associated with some embodiments of the present invention to facilitate a better framework for understanding some of the various techniques and applications of the claimed subject matter. Accordingly, it should be understood that these Background section statements are to be read in this light and not necessarily as admissions of prior art.
Drill tool assembly vibrations are known to potentially have a significant effect on Rate of Penetration (ROP) and represent a significant challenge to interpret and mitigate in pursuit of reducing the time and cost of drilling subterranean wells. Drill tool assemblies vibrate during drilling for various reasons related to one or more drilling parameters. For example, the rotary speed (often expressed in revolutions per minute, or RPM), weight on bit (WOB), mud viscosity, etc. each may affect the vibrational tendency of a given drill tool assembly during a drilling operation. Measured depth (MD), rock properties, hole conditions, and configuration of the drill tool assembly may also influence drilling vibrations. As used herein, drilling parameters include characteristics and/or features of both the drilling hardware (e.g., drill tool assembly) and the drilling operations.
As used herein, drill tool assembly refers to assemblies of components used in drilling operations. Exemplary components that may collectively or individually be considered a drill tool assembly include rock cutting devices, bits, mills, reamers, bottom hole assemblies, drill collars, drill strings, couplings, subs, stabilizers, MWD tools, motors, etc. Exemplary rig systems may include the top drive, rig control systems, etc., and may form certain boundary conditions. Deployment of vibrationally poor drill tool assembly designs and conducting drilling operations at conditions of high downhole vibrations can result in loss of rate of penetration, shortened drill tool assembly life, increased number of trips, increased failure rate of downhole tools, and increased non-productive time. It is desirable to provide the drilling engineer and/or rig operating personnel with a useful but not overly complex tool utilizing readily available data and quickly estimating the vibrational tendencies of the drill tool assembly.
A fixed cutter bit often requires more torque than a corresponding roller cone bit drilling similar formations at comparable conditions, although both bits can experience torque issues. Increased bit torque can lead to an increase in the phenomenon known as “stick-slip,” an unsteady rotary speed at the bit, even when surface rotary speed remains substantially constant. Excessive stick-slip can be severely damaging to drill string assemblies. Roller cone bits may sometimes be more prone to axial vibration issues than corresponding fixed cutter bits. Although axial vibrations may be reduced by substituting fixed cutter bits for roller cone bits, some drilling operations with either type of bit may continue to experience axial vibration problems. Fixed cutter bits can be severely damaged by axial vibrations as the PDC wafer can be knocked off its substrate if the axial vibrations are severe. Axial vibrations are known to be problematic for rotary tricone bits, as the classic trilobed bottomhole pattern generates axial motion at the bit. There are known complex mathematical and operational methods for measuring and analyzing downhole vibrations. However, these typically require a substantial amount of data, strong computational power, and special skill to use and interpret.
Typically, severe axial vibration dysfunction can be manifested as “bit bounce,” which can result in a momentary lessening or even a momentary complete loss of contact between the rock formation and the drill bit cutting surface through part of the vibration cycle. Such axial vibrations can cause dislocation of PDC cutters and tricone bits may be damaged by high shock impact with the formation. Dysfunctional axial vibration can occur at other locations in the drill tool assembly. Other cutting elements in the drill tool assembly could also experience a similar effect. Small oscillations in weight on bit (WOB) can result in drilling inefficiencies, leading to decreased ROP. For example, the depth of cut (DOC) of the bit typically varies with varying WOB, giving rise to fluctuations in the bit torque, thereby inducing torsional vibrations. The resulting coupled torsional-axial vibrations may be among the most damaging vibration patterns as this extreme motion may then lead to the generation of lateral vibrations.
Recently developed practices around optimizing the Bottom-Hole Assembly (BHA) design (WO 2008/097303) and drilling parameters for robust vibrational performance, and using real-time Mechanical Specific Energy (MSE) monitoring for surveillance of drilling efficiency (US 2009/0250264) have significantly improved drilling performance. MSE is particularly useful in identifying drilling inefficiencies arising from, for example, dull bits, poor weight transfer to the bit, and whirl. These dysfunctions tend to reduce ROP and increase expended mechanical power due to the parasitic torques generated, thereby increasing MSE. The availability of real-time MSE monitoring for surveillance allows the driller to take corrective action. One of the big advantages of MSE analysis is that it does not require real-time downhole tools that directly measure vibration severity, which are expensive and prone to malfunction in challenging drilling environments. Unfortunately, MSE analysis may not provide reliable information about the severity of torsional or axial oscillations. Field data shows intervals for which MSE does detect such patterns and other instances for which there is no vibration signature in the MSE data. Therefore, it is desirable to have additional indicators complementary to MSE that can provide torsional and/or axial severity from surface data, thereby avoiding the costly step of deploying downhole tools just for this purpose.
Multiple efforts have been made to study and/or model these more complex torsional and axial vibrations, some of which are discussed here to help illustrate the advances made by the technologies of the present disclosure. DEA Project 29 was a multi-partner joint industry program initiated to develop modeling tools for analyzing drill tool assembly vibrations. The program focused on the development of an impedance-based, frequency-dependent, mass-spring-dashpot model using a transfer function methodology for modeling axial and torsional vibrations. These transfer functions describe the ratio of the surface state to the input condition at the bit. The boundary conditions for axial vibrations consisted of a spring, a damper at the top of the drill tool assembly (to represent the rig) and a “simple” axial excitation at the bit (either a force or displacement). For torsional vibrations, the bit was modeled as a free end (no stiffness between the bit and the rock) with damping. This work also indicated that downhole phenomena such as bit bounce and stick-slip are observable from the surface. While the DEA Project 29 recognized that the downhole phenomena were observable from the surface, they did not specifically attempt to quantify this. Results of this effort were published as “Coupled Axial, Bending and Torsional Vibration of Rotating Drill Strings”, DEA Project 29, Phase III Report, J. K. Vandiver, Massachusetts Institute of Technology and “The Effect of Surface and Downhole Boundary Conditions on the Vibration of Drill strings,” F. Clayer et al, SPE 20447, 1990.
Additionally, U.S. Pat. No. 5,852,235 ('235 patent) and U.S. Pat. No. 6,363,780 ('780 patent) describe methods and systems for computing the behavior of a drill bit fastened to the end of a drill string. In '235, a method was proposed for estimating the instantaneous rotational speed of the bit at the well bottom in real-time, taking into account the measurements performed at the top of the drill string and a reduced model. In '780, a method was proposed for computing “Rf, a function of a principal oscillation frequency of a weight on hook WOH divided by an average instantaneous rotating speed at the surface of the drillstring, Rwob being a function of a standard deviation of a signal representing a weight on bit WOB estimated by the reduced physical model of the drill string from the measurement of the signal representing the weight on hook WOH, divided by an average weight on bit WOB0 defined from a weight of the drill string and an average of the weight on hook WOH, and any dangerous longitudinal behavior of the drill bit determined from the values of Rf and Rwob” in real-time.
These methods require being able to run in real-time and a “reduced” model that can accept a subset of measurements as input and generate outputs that closely match the remaining measurements. For example, in '235 the reduced model may accept the surface rotary speed signal as an input and compute the downhole rotary speed and surface torque as outputs. However, the estimates for quantities of interest, such as downhole rotary speed, cannot be trusted except for those occurrences that obtain a close match between the computed and measured surface torque. This typically requires continuously tuning model parameters, since the torque measured at the surface may change not only due to torsional vibrations but also due to changes in rock formations, bit characteristics, borehole patterns, etc., which are not captured by the reduced model. Since the reduced model attempts to match the dynamics associated with relevant vibrational modes as well as the overall trend of the measured signal due to such additional effects, the tuned parameters of the model may drift away from values actually representing the vibrational state of the drilling assembly. This drift can result in inaccurate estimates of desired parameters.
Another disadvantage of such methods is the requirement for specialized software, trained personnel, and computational capabilities available at each drilling operation to usefully utilize and understand such systems.
A recent patent application publication entitled “Method and Apparatus for Estimating the Instantaneous Rotational Speed of a Bottom Hole Assembly,” (WO 2010/064031) continues prior work in this area as an extension of IADC/SPE Publication 18049, “Torque Feedback Used to Cure Slip-Stick Motion,” and previous related work. One primary motivation for these efforts is to provide a control signal to the drilling apparatus to adjust the power to the rotary drive system to reduce torsional drill string vibrations. A simple drill string compliance function is disclosed providing a stiffness element between the rotary drive system at the surface and the bottom hole assembly. Inertia, friction, damping, and several wellbore parameters are excluded from the drill string model. Also, the '031 reference fails to propose means to evaluate the quality of the torsional vibration estimate by comparison with downhole data, offers only simple means to calculate the downhole torsional vibrations using a basic torsional spring model, provides few means to evaluate the surface measurements, does not discuss monitoring surface measurements for bit axial vibration detection, and does not use the monitoring results to make a comprehensive assessment of the amount or severity of stick-slip observed for a selected drilling interval. This reference merely teaches a basic estimate of the downhole instantaneous rotational speed of the bit for the purpose of providing an input to a surface drive control system. Such methods fail to enable real-time diagnostic evaluation and indication of downhole dysfunction.
Other related material may be found in “Development of a Surface Drillstring Vibration Measurement System”, A. A. Besaisow, et al., SPE 14327, 1985; “Surface Detection of Vibrations and Drilling Optimization: Field Experience”, H. Henneuse, SPE 23888, 1992; and, “Application of High Sampling Rate Downhole Measurements for Analysis and Cure of Stick-Slip in Drilling,” D. R. Pavone and J. P. Desplans, 1994, SPE 28324. Additionally, patent application WO 2009/155062 A1, “Methods and Systems for Mitigating Drilling Vibrations,” provides further details on the methods presented herein. Numerous theoretical and analytical methods have been taught and disclosed in the art, but few have also provided methods for applying such technology. The art remains in need of a more reliable method for predicting downhole vibrational effects utilizing information that can be relatively easily obtained from surface measurements and data. The art particularly also remains in need of such methods that can be usefully employed at remote locations such as at a drill site, without the need for exceptional engineering and computational skills and equipment.
The present disclosure relates to improved methods to estimate the effective vibration attributes of the bottom of the drill tool assembly, such as at or near a drill bit, based on evaluation of selected surface operating parameters. The estimates may then be utilized, such as in advance of, during, or after drilling activities to enhance present or future drilling operations. These methods and systems may be used to increase overall drilling performance by adopting corrective measures to mitigate excessive inefficiencies and operational dysfunctions associated with vibrational energies within the drilling assembly. Vibrations may include but are not limited to torsional, axial, and coupled torsional/axial vibrations. Estimation of downhole vibrations from surface data can provide critical information to assess changes in operating parameters and bit selection. Since stick-slip can vary during a drilling operation due to both formation changes and operating parameter variations, maintaining an estimation of the amount of stick-slip severity for the entire drilling interval can provide important information for a drilling operation. It is desirable to implement a usefully accurate, reliable, and dependable remote surveillance program based on surface data that is broadly applicable, easy to teach, and easy to implement, using various selected aspects of a wide variety of rig data logging equipment that is readily available to the individual drill teams.
In one aspect, the claimed subject matter includes a method to estimate severity of downhole vibration for a wellbore drill tool assembly, comprising the steps: a. Identifying a dataset comprising selected drill tool assembly parameters; b. Selecting a reference downhole vibration index for the drill tool assembly; c. Identifying a surface parameter and calculating a reference surface vibration attribute for the selected reference downhole vibration index; d. Determining a surface vibration attribute derived from at least one surface measurement or observation obtained in a drilling operation, the determined surface vibration attribute corresponding to the identified surface parameter (step c); and e. Estimating a downhole vibration index by evaluating the determined surface vibration attribute (step d) with respect to the identified reference surface vibration attribute (step c). As used herein, the term drilling operation is defined broadly to include boring, milling, reaming or otherwise excavating material to enlarge, open, and/or create a wellbore, whether original drilling operation, planning a drilling operation, work-over operation, remedial operation, mining operation, or post-drilling analysis. Downhole vibration index for the drill tool assembly may include but is not limited to bit disengagement index, ROP limit state index, bit bounce compliance index, bit chatter index, relative bit chatter index, stick-slip tendency index, bit torsional aggressiveness index, forced torsional vibration index, relative forced torsional vibration index, axial strain energy index, torsional strain energy index, and combinations thereof.
In another aspect, the claimed technology includes a. identifying a dataset comprising (i) parameters for a selected drill tool assembly comprising a drill bit, (ii) selected wellbore dimensions, and (iii) selected measured depth (MD); b. Selecting a reference downhole vibration index for at least one of downhole torque, downhole weight on bit, downhole bit rotary speed, and downhole axial acceleration; c. Identifying a corresponding selected surface parameter including at least one of surface torque, a surface hook-load, surface drill string rotation rate, and surface axial acceleration, and calculating a corresponding reference surface vibration attribute for the selected reference downhole vibration index; d. Determining a surface vibration attribute obtained in a drilling operation, the determined surface vibration attribute corresponding to the identified selected surface parameter (step c); and step e. Estimating a downhole vibration index by evaluating the determined surface vibration attribute (step d) with respect to the identified reference surface vibration attribute (step c).
In other embodiments, the claimed improvements include a method to estimate severity of downhole vibration for a drill tool assembly, comprising the steps: a. Identifying a dataset comprising selected drill tool assembly parameters; b. Selecting a reference downhole vibration index for the drill tool assembly; and c. Identifying one or more ratios of: the selected reference downhole vibration index for the drill tool assembly (step b) to a calculated reference surface vibration attribute; d. Determining a surface vibration attribute derived from at least one surface measurement or observation obtained in a drilling operation, the determined surface vibration attribute corresponding to the identified surface parameter (step c); and e. Estimating the downhole vibration index by evaluating the determined surface vibration attribute (step d) with respect to one or more of the identified ratios (step c).
Additionally or alternatively, the methods above may include a step to estimate the quality of the downhole vibration index determined from surface data by comparison with downhole measured data, either during or after the drilling process.
In other embodiments, the methods above may include a step to evaluate the downhole vibration indices from at least two drilling intervals for the purpose of a drilling performance assessment to recommend selection of a drilling parameter for a subsequent interval, which may include selection of one or more bit features or characteristics, or a change in the specified WOB or rotary speed, or both.
In other alternative embodiments, the methods above may include the use of downhole vibration indices from surface data to evaluate drilling performance for an interval to adjust at least one drilling parameter to maintain a downhole vibration index at a desired value or below a maximum value not to be exceeded during the operation.
In the following Detailed Description, specific aspects and features of the claimed subject matter are described in connection with several exemplary methods and embodiments. However, to the extent that the following description is specific to a particular embodiment or a particular use of the present techniques, it is intended to be illustrative only and merely provides a concise description of exemplary embodiments. Moreover, in the event that a particular aspect or feature is described in connection with a particular embodiment, such aspect or feature may be found and/or implemented with other embodiments of the present invention where appropriate. Accordingly, the claimed invention is not limited to the specific embodiments described below, but rather, the invention includes all alternatives, modifications, and equivalents falling within the scope of the appended numbered paragraphs and claimed subject matter.
While the present systems and methods may be used in connection with any rotary drilling, milling, under-reaming, or boring operation, they are expected to be used primarily in wellbore drilling operations related to the recovery of hydrocarbons, such as for oil and gas wells. References herein to drilling operations are to be understood expansively. Operators are able to remove rock, other formation, casing components, cement, and/or related materials using a variety of apparatus and methods, some of which are different from conventional forward drilling into virgin formation. Accordingly, the discussion herein referring to drilling parameters, drilling performance measurements, drilling vibrations, drilling vibration severity, drilling vibration amplitude, etc., refers to parameters, measurements, performance, vibrations, and severity during any of the variety of operations that are associated with a wellbore rotary drilling process. As is well known in the hydrocarbon wellbore drilling industry, a number of factors affect the efficiency of the drilling operations, including factors within the operators' control and factors such as rock properties that are beyond operators' control. For purposes of this application, the term drilling conditions will be used to refer generally to the conditions in the wellbore during the drilling operation. The drilling conditions are comprised of a variety of drilling parameters, some of which relate to the environment of the wellbore and/or formation and others that relate to the drilling activity itself. For example, drilling parameters may include but are not limited to, any of rotary speed (RPM), weight on bit (WOB), measured depth (MD), hole angle, hole diameter, characteristics of the drill bit and drill string, mud weight, mud flow rate, mud viscosity, rock properties, lithology of the formation, pore pressure of the formation, torque, pressure, temperature, rate of penetration, mechanical specific energy, etc., and/or combinations thereof. Various parameters may be directly measured or must be indirectly measured, calculated, estimated, or otherwise inferred from available data. Typically, critical downhole determinations are more difficult or complicated to obtain than readily accessible surface parameters. It may be appreciated that these parameters typically may be measured and described quantitatively, and these measurements have certain attributes that characterize the data. Common attributes include mean value, standard deviation, root-mean-square, and other statistical values. Additional attributes of the surface parameters may include dominant period, dominant frequency, time rate of change, peak time rate of change (“slew rate”), peak-to-peak amplitude, moving average, spectral periodogram from Fourier analysis, and the like. Downhole vibration index for the drill tool assembly may include but are not limited to bit disengagement index, ROP limit state index, bit bounce compliance index, bit chatter index, relative bit chatter index, stick-slip tendency index, bit torsional aggressiveness index, forced torsional vibration index, relative forced torsional vibration index, axial strain energy index, torsional strain energy index, and combinations thereof.
The present inventions and claimed subject matter provide methods for reliably and conveniently estimating various downhole vibration indices from relatively available surface data, such estimations being useful to timely reduce unacceptable vibrations and improve drilling operations. The measurements and data acquisitions performed at the top of the drill tool assembly can be obtained by means of sensors or an instrumented sub situated close to the top of the drill tool assembly, or may be obtained at or near the drilling rig.
As drilling operations progress, the drill bit 108 advances through the formation 110 at a rate known as the rate of penetration (ROP, 108), which is commonly calculated as the measured depth (MD) drilled over time. As formation conditions are location dependent, drilling conditions necessarily change over time as the wellbore penetrates varying formations. Moreover, the drilling conditions may change in manners that dramatically reduce the efficiencies of the drilling operation and/or that create less desirable operating conditions. The presently claimed subject matter demonstrates improved methods of predicting, estimating, and detecting changes in drilling conditions and the response of different bits and cutting tools to these formations. Bit selection is a key parameter that affects drilling efficiency and the art of bit design continues to advance with new bit features that may be difficult to evaluate for a specific drilling application without using the bit to drill at least a portion of a formation of commercial interest. Means to evaluate the performance of such a drill test may include the propensity of the bit to generate drilling vibrations, including torsional stick-slip vibrations. Beneficially, the claimed subject matter provides means to efficiently quantify with a reasonable degree of accuracy the downhole vibration severity, relying only on calculable parameters and surface data measurements, thus avoiding the delays, costs, and complexity of providing actual downhole measurements.
This invention discloses a method to estimate the severity of one or more of rotary speed and WOB fluctuations at the bottom of the drill tool assembly in real time during drilling operations, or optionally before or after drilling to aid in drilling assembly planning or analysis. This severity estimate is computed based on a mechanical description of the drilling assembly and real-time operating parameters (including torque, rotary speed, WOH, WOB) and measured depth (MD) readings taken from one or more of a surface drilling rig recording system and an instrumented surface sub. Additional information such as the wellbore trajectory, drilling fluid density and plastic viscosity, and friction factors can refine this estimate but is not required. In some applications, the estimated severity level may be displayed to the driller or an engineer, in a manner similar to rig-determined and displayed MSE data, to assist in drilling surveillance and operational decisions. In one method, to analyze vibrational performance the driller may be provided (directly or indirectly) portions of the information in the form of predetermined tables or plots (e.g., for direct read and/or interpolation) that in conjunction with the rig-measured data can allow estimation of stick-slip severity, torque fluctuations, and axial vibration severity by monitoring the surface torque, rotary speed, and hookload on the driller's screen or rig parameters.
According to the present invention, the severity of a given type of vibrational dysfunction can be described by a dimensionless ratio that compares the amplitude of dynamic fluctuations in a drilling parameter to its average value. For example, stick-slip severity may be related to the ratio of dynamic rotary speed variations at the drill bit to the average rotary speed at the bit. Since there is rarely permanent torsional deformation of the drill string, the average rotary speed of the bit (downhole) is substantially equal to the average rotary speed of the drill string at the surface. When a vibrational dysfunction is present, a single dominant (“active”) vibrational mode at a specific frequency may account for a dominant portion of the dynamic variation in the observed drilling parameter. Thus, if the particular active mode can be reliably identified, it is possible to infer the amplitude of such vibrational modes from observations anywhere along the drill string, particularly at the surface where such measurements are already made.
In one aspect, the claimed subject matter includes a method to estimate severity of downhole vibration for a wellbore drill tool assembly, comprising the steps: a. Identifying a dataset comprising selected drill tool assembly parameters; b. Selecting a reference downhole vibration index for the drill tool assembly; c. Identifying a surface parameter and calculating a reference surface vibration attribute for the selected reference downhole vibration index; d. Determining a surface vibration attribute derived from at least one surface measurement or observation obtained in a drilling operation, the determined surface vibration attribute corresponding to the identified surface parameter (step c); and e. Estimating a downhole vibration index by evaluating the determined surface vibration attribute (step d) with respect to the identified reference surface vibration attribute (step c). As used herein, the term drilling operation is defined broadly to include boring, milling, reaming or otherwise excavating material to enlarge, open, and/or create a wellbore, whether original drilling operation, planning a drilling operation, work-over operation, remedial operation, mining operation, or post-drilling analysis.
As used herein, vibration relates to vibration of one or more components of the drill tool assembly and comprises one or more of torsional vibration, axial vibration, lateral vibration, coupled torsional and axial vibrations, and combinations thereof.
The step of “identifying a dataset” may comprise selecting, for example, one or more drill tool assembly design parameters, wellbore dimensions, measured depth (MD), projected drilling parameters, wellbore survey data, and wellbore fluid properties.
The “reference downhole vibration index” may be selected as, for example, a function of one or more of downhole drill tool assembly rotary speed, downhole axial velocity, downhole axial acceleration, downhole axial load, downhole torsional moment, and combinations thereof. In some embodiments, selecting a reference downhole vibration index may comprise, for example, selecting a downhole condition for the drill tool assembly for which the rotary speed is momentarily zero. Momentarily zero means that for at least some discernable increment of time the downhole rotary speed comes to a halt or is not greater than five percent of the average value. In some other embodiments, selecting a reference downhole vibration index may include, for example, selecting a downhole condition where a weight on bit (WOB) parameter is momentarily zero. In yet another embodiment, selecting a reference downhole vibration index may comprise selecting an undesirable downhole condition, such as for example full stick-slip of the bit, bit axial disengagement from the formation, or momentarily exceeding some design or operating limit anywhere along the drill tool assembly, such as the make-up or twist-off torque of a connection, a bucking limit, tensile or torsional strength of a component.
A corresponding surface parameter may be identified that is physically connected to the selected downhole vibration index. Using a mathematical model of this physical coupling, as described herein, a reference surface vibration attribute may be calculated for the corresponding reference downhole vibration.
Determining a surface vibration attribute may refer to calculating, estimating, or otherwise obtaining a quantity related to one or more measured values of a surface parameter. The term “surface parameter” as used herein is defined broadly to refer to physical properties, manifestations of vibrational energy, and operating conditions observed or measured at the surface. Typical vibration attributes of interest include but are not limited to the period of vibration of surface torque, peak-to-peak amplitude of surface torque, root-mean-square value of surface hookload, etc. Additional examples of surface vibration attributes are provided herein.
A downhole vibration index may be estimated from the determined surface vibration attribute obtained from the measured data, in consideration of the calculated reference level of the surface parameter corresponding to the selected reference downhole vibration.
A related but previously developed attempt to model downhole vibrational effects, WO 2009/155062, filed on Jun. 17, 2008, describes certain methodologies based upon a frequency domain model to design a drill tool assembly for use in a drilling operation, based on drilling parameters and drill tool assembly data, utilizing torsional and axial vibration indices that characterize an excitation response. The models described therein and presented as one embodiment below may optionally be used in conjunction with the present invention to compute the frequency response of the drill tool assembly and specifically to compute the dominant periods of vibration, as well as ratios of vibration amplitudes of one or more surface and downhole parameters for such periods.
In methods according to the present invention, the vibration attributes may provide information on the characteristic dynamic oscillations in one or more operating parameters such as torque, hookload, rotary speed, WOB and acceleration over a specified period or periods of vibration. Specifically, the vibration amplitude may be obtained from the Fourier component of the drilling operating parameters obtained at a specific frequency, or, if a single vibrational mode is dominant (active), from the maximum and minimum values that are observed during an interval longer than but comparable to the period of oscillation. A period of oscillation refers to the time required for completion of one cycle of dynamic variation. This period corresponds to the normal modes of vibration associated with the drill tool assembly.
Vibration amplitude may be determined by various methods that may be considered essentially equivalent for signals of interest with respect to accurately determining amplitude. In the time domain, the vibration amplitude is simply the coefficient A(t) in the expression x(t)=A(t)sin(ωt). The field of random vibrations teaches several ways to estimate A(t), which may in general vary in time, from a set of measurements. After means to remove a slowly-varying, steady, or “DC,” component, the residual signal typically has zero mean. The crossings of the signal with the time axis, in either the up or down direction, has significance because these time values help to determine the period. For one such cycle, the extreme values can be determined, and these values can be used to determine one estimate of the amplitude A(t). Alternatively, a sine wave could be fit to the data for one such period with the coefficient A(t) determined by a minimum error approach. Also, the standard deviation of the signal can be determined for some moving time window or interval, and using mathematical relationships one may estimate the amplitude A(t) from these values. As mentioned above, Fourier analysis is yet another way to calculate the amplitude of a sinusoidal signal. Therefore, the phrase “vibration amplitude” is used to refer to the strength A(t) of a time-varying signal that may be determined by these and other means that are known to those skilled in the art, including processes that use FIR and IIR filters, state observers, Kalman filters, derivatives, integrals, and the like.
The significance of vibration amplitude fluctuation about a nominal value of a signal is related to the strength of the signal overall. That is, severity of downhole vibrations (“vibration severity”) can be considered to be related to the ratio of the vibration amplitude to the mean signal strength. One convenient means to measure vibration severity of a signal x(t) is to define
In some references, the factor of 2 is absent. However, it is convenient to consider 100% stick-slip, or “full stick-slip,” to correspond to the condition wherein the sinusoidal oscillation of the bit about its mean rate of rotation is such that it momentarily has zero rotary speed, for which the amplitude of the vibration is equal to the mean rotary speed. Then Max(x)=2A, Min(x)=0, Mean(x)=A, and S(x)=100%. It is recognized that other more severe stick-slip conditions may occur, and the pattern may not be purely sinusoidal. This example is provided as a reference condition and is not limiting. Additional definitions of vibration severity are within the scope of the claimed subject matter.
Although the observed values of vibration amplitudes are affected by factors that can change continuously during drilling, the ratios of such amplitudes at different positions along the drill tool assembly for a given vibrational mode can be robustly estimated simply from the eigenfunction of the mode (also referred to as the “mode shape”), even under varying drilling conditions. Thus, with the knowledge of the active mode and its mode shape, it is possible to reliably estimate the vibrational amplitude of a parameter associated with downhole behavior from an observation or determination of a related parameter at another location, such as at the surface. Furthermore, it is not necessary to model either the instantaneous values or the long-term trends of the drilling parameters, both of which depend on many other uncontrolled factors.
The main benefit of the method outlined and claimed subject matter in this disclosure is that it allows real-time computation of the torsional and axial severity along with suitable alarm levels that diagnose downhole conditions without access to downhole vibration data. By diagnosing the axial and torsional behavior of the drill string, this invention complements the operator's ROP management process that uses the Mechanical Specific Energy (MSE) as a diagnostic surface measurement of downhole behavior. The downhole vibration indices presented herein are complementary to the MSE data. Estimates of downhole vibrations from surface data may be compared with downhole data measurements for use in an evaluation of the quality of the downhole vibration index. The accuracy of the physical model and proper selection of drilling parameter data will both contribute to increasing quality of the downhole vibration indices. Furthermore, downhole vibration indices for complete drilling intervals may be used in drilling performance assessment to aid in bit selection and drilling parameter selection for use in drilling a subsequent interval. It is therefore important to assess the quality of the downhole vibration index, using downhole data measurements, so as to understand the accuracy of the dynamic model and to conduct any necessary calibrations of the model. After a model has been calibrated and the quality of the estimate is known, it can be used with greater confidence for making operational and design decisions.
For example, downhole vibration indices may be obtained for a specific bit drilling a specific interval under certain drilling conditions. If the downhole vibration index indicates that the bit is not operating close to stick-slip, then one could reasonably choose a more aggressive bit or one or more other more aggressive operational parameters for a subsequent run, or a combination thereof. However, if the data shows that the bit is routinely in full stick-slip, a reduction in bit tooth or cutter depth-of-cut may be warranted, or alternatively less aggressive operating parameters would be advised. Such results are likely to be formation specific and thus one could contemplate the need to conduct such surveillance on a nearly continuous basis. Since it is most desirable to drill as long an interval as possible with a single bit, one important value of the diagnostics is to provide information for choosing a bit and operating parameters that have optimized performance over the interval taken as a whole.
Instead of investigating the total dynamic motion of the drill string, the inventive subject matter claimed herein separately investigates each of the zero and first order terms in a perturbation expansion. The fluctuation amplitudes of drilling operating parameters such as torque, WOH, WOB, and rotary speed are derived as the first order components of a perturbation expansion of the equations of motion of the drilling assembly. The zero-order terms determine the baseline solution. Second and higher order terms are not necessary for the claimed methods but could be calculated if desired. Using the fluctuation amplitudes provides a practical approach to calculation of the torsional and axial behavior at the bottom of the drill string. This is because the dynamic perturbation models do not require a complete understanding of the factors that affect the average steady-state amplitudes of these parameters, and there is a reduced requirement to tune the model to account for differences between estimated and actual average amplitudes of these parameters. This approach exploits the fact that stick-slip and bit bounce are dependent on the dynamic variations and not on the average values of these signatures. Also, with this approach, it is possible to provide additional information on fluctuations in rotary speed, torque, WOB, and WOH that is useful during drilling operations and in post-drill re-design. Furthermore, the methods and systems described herein differ from the approaches specified in other applications in the following ways: our calculations do not compute a real-time value of the rotational speed of the bit; our model calculations are not required to be carried out in real-time; our methods can make use of spectral analysis, and details from specific frequency(ies) may then be used for further computation; and we have no need to over-sample the data if the period of the active mode is known.
There are several techniques and devices that can acquire measurement data at the surface. These include measurement devices placed at the top of the drill string, which determine certain drilling mechanics properties including accelerations and drilling operating conditions such as torque, WOH, WOB, motor current or voltage fluctuations, and rotary speeds. Other devices exist that measure drilling mechanics data downhole and along the entire string. The advent of wired drillpipe offers additional possibilities for along-the-string measurements that can be used during a drilling operation, and, similarly, data from along-the-string memory devices may be used in a post-drill analysis. Typically, the driller who controls the surface rig operations can modify and control the WOB, torque, rotary speed, and the ROP. These operating parameters can be managed by one or more of: (a) real-time feeds of surface drilling mechanics data, (b) delayed feeds of downhole data using a mud-logging system or other suitable surface monitoring service, and (c) built-in automatic control devices.
One method provided herein includes a step of selecting a reference downhole amplitude or vibration severity for a torsional or axial state to be diagnosed. Examples of reference downhole conditions include: (1) the state of “full stick-slip” in which the torsional rotation of the bit momentarily comes to a full stop and then accelerates to a peak rate of rotation of approximately twice the average rotary speed; (2) the state of “bit bounce” for which the applied axial force of the bit on the bottom of the borehole is momentarily zero, after which it may increase to a value considerably in excess of its average value; (3) an axial vibration state in which the bit is lifted off the bottom of the borehole a sufficient distance such that the cutting element clears the present bottomhole cutting pattern; (4) extreme values of stick-slip such that the instantaneous torque value is negative and rises to a sufficient level to backoff drill string connections, which will depend on the specific hole size and drill string connections in use. The reference downhole condition may be expressed as a vibration amplitude or as a vibration amplitude ratio. For example, one may specify the rotary speed range or, alternatively, full stick-slip for which the ratio of the vibration amplitude (A(t) above) to the average rotary speed is 1, or 100% stick-slip. It follows that other natural reference downhole vibration conditions may be selected, but these are ones of present interest.
For such a reference downhole vibration, the amplitudes and severity of the corresponding reference levels of surface parameters are calculated using the drilling parameters and the physical model, which includes as much descriptive physics as may be necessary for an accurate modeling estimate. The reference surface condition may be simply a reference vibration amplitude of a single surface parameter (such as torque), or it may be a complex relation between multiple surface parameters (such as torque and rotary speed) for more complicated surface boundary conditions.
In the simplest case, it is possible to evaluate the downhole vibration severity by first selecting the downhole vibration index and its reference level, identifying a surface parameter and calculating its vibration amplitude for the corresponding downhole vibration reference amplitude (this is the “reference surface vibration attribute”). Then the vibration amplitude of the surface parameter is determined from measured data at the surface from a drilling operation, using one or more of the several methods indicated above. The “vibration amplitude ratio” is calculated as the measured surface vibration attribute, divided by the reference surface vibration attribute calculated from the model and the drilling parameters for the reference downhole vibration index. This vibration amplitude ratio is an estimate of the downhole vibration index. This method can be generalized to include more than one reference level and additional surface vibration attributes such as primary period and other measures of the effective vibration amplitude.
Consider a simple embodiment of the torsional stick-slip problem. The reference downhole condition is full stick-slip, for which the vibration amplitude of the rotary speed is equal to the mean value. The surface torque vibration amplitude may be calculated from the physical model for this downhole vibration reference condition. The vibration amplitude of the surface torque is determined from the measured surface data. In this simple embodiment, the ratio of the measured surface torque vibration amplitude to the calculated reference level is the torsional severity estimate (TSE).
In another embodiment, a post-drill analysis may be performed on a well for which downhole measurements were made while drilling. These measurements can be compared to the reference downhole vibration amplitude to obtain a measured downhole vibration index. Then any of a number of algorithms from the field of pattern recognition (also known as machine learning, statistical learning, data mining, and artificial intelligence) may be employed to train a computer program to automatically classify the severity of the downhole vibrations given only the corresponding topside measured data. Such algorithms include, but are not limited to, linear and logistic regression, discriminant analysis, and classification and regression trees. Once this post-drill analysis is complete for one or more wells, the trained algorithms may be employed to autonomously estimate downhole vibration severity in real-time while drilling new wells. Though such learning algorithms need only employ the drilling measurements, their classification performance is greatly improved by also using the results of the physical models described herein as a baseline during training.
Depending on the environment in which the present systems and methods are utilized, the adjustment of the at least one drilling parameter may be based on this one or more vibration amplitude ratio(s) and/or on the determined or identified drilling parameter change. For example, in a field operation, the identified change may be displayed for an operator with or without the underlying vibration amplitude ratio or severity level used to determine the change. Regardless of whether the vibration amplitude ratio or severity level is displayed to the operator in the field, the determined change may also be presented and the operator may act to adjust drilling conditions based solely on the displayed change. Additionally or alternatively, an operator or other person in the field may consider both the vibration amplitude ratios and the identified drilling parameter change. Additionally or alternatively, such as when the identified drilling parameter change is merely a change in operating conditions, the computer system may be adapted to change the drilling parameter without user intervention, such as by adjusting WOB, WOH, rotary speed, pump rate, etc. Again, depending on the manner or environment in which the present systems and methods are used, the manner of adjusting the drilling parameter may change. The present methods and systems may be implemented in a manner to adjust one or more drilling parameters during a drilling operation, but not necessarily in real-time. Furthermore, the data may be evaluated in a post-drill performance evaluation review, with subsequent recommendations on drilling parameter change, including selection of a drill bit or bit characteristics and features, for use in the drilling of a subsequent interval. A recent important innovation is the use of depth-of-cut (DOC) control features on PDC bits, which limit the amount of cutter penetration at higher bit weights. The DOC feature thus limits the bit torque at high bit weight. Evaluation of bit performance and optimizing the selection of DOC features has thus become more complex, and additional tools such as the present invention are necessary to maximize drilling performance.
The inventive technology may also include a software program that graphically characterizes the vibrational performance of the drill tool assembly. In some implementations, the software program will graphically characterize the vibrational performance or tendency of a single configuration design for one or more vibrational modes. The methodologies implemented to graphically characterize the torsional and axial vibration performance incorporate a common framework with some differences.
As will be described in greater detail below, the software program input consists of entering ranges for various drilling parameters, such as WOB, rotary speed, drilling fluid density and viscosity, and bit depth, as well as various drill tool assembly design parameters, such as pipe and component dimensions, mechanical properties, and the locations of drill tool assembly components, such as drill collars, stabilizers and drill pipe. It has been observed that the proper modeling of drill pipe tool joints affects certain modes of vibration, and model accuracy depends on including these periodic elements of greater wall thickness, weight, and stiffness in the drill string model. In some implementations, the program may allow for developing and maintaining multiple drill tool assembly design configurations as a storage record of the vibration amplitude ratios obtained for alternative drill tool assembly design configurations.
An exemplary method is provided below, along with the details of a model of the drill tool assembly response to torsional and axial excitations as described in WO 2009/155062. Useful information about the vibration characteristics of a drill tool assembly design under particular operating conditions can be obtained through frequency-domain modeling of the drill tool assembly response to excitations. The modeling is considered more robust because it is adapted to more thoroughly or explicitly incorporate factors previously ignored or represented by mere constants while maintaining tractability and computational efficiency. Exemplary factors that may be incorporated into the present frequency-domain models include drill string component inertial effects, the effect of tool joints on inertial and stiffness properties of the drill string, velocity-dependent damping relationships, drill tool assembly friction, drill bit friction, and complex borehole trajectory effects. Additionally, a number of complex factors influence the aggressiveness (rate of torque generation) and efficiency (energy consumed for penetrating rock in relation to rock strength) of the drill bit. These bit parameters depend heavily on details of the bit geometry, bit condition (new vs. dull), depth-of-cut (DOC) features, bottom-hole hydraulics, rock properties, etc. The model does not attempt to predict these parameters, which are measurable or known to a large degree during drilling operations, but uses them as inputs to analyze the response of the drill tool assembly to excitations caused by the bit action. The model is sufficiently complete that advanced modeling features may be examined, such as coupling between axial and torsional vibrations at the bit, as well as complex surface impedance characteristics, for which both torque and rotary speed may have dynamic variations at the surface, for example. It may also be noted that the effects of some of these parameters increase with increasing string length, and therefore greater model accuracy is required to maintain the downhole vibration index quality for increasing drill string length.
The data regarding drilling operations may include specific data regarding drilling operating conditions and/or may include drilling parameters, which are ranges of available conditions for one or more drilling operational variables, such as WOB, WOH, rotary speed, fluid density and viscosity, etc. An operational variable is an operational element over which an operator has some control. The methods and systems of the present disclosure may obtain input data, such as for use in the frequency-domain models, from a drilling plan. As used herein, drilling plan refers to the collection of data regarding the equipment and methods to be used in a drilling operation or in a particular stage of a drilling operation.
In the representative illustration, the data to be input into the systems and methods are stored in data storage device 412. The system computer 410 may retrieve the appropriate data from the data storage device 412 to perform the operations and analyses described herein according to program instructions that correspond to the methods described herein. The program instructions may be written in any suitable computer programming language or combination of languages, such as C++, Java, MATLAB™, and the like, and may be adapted to be run in combination with other software applications, such as commercial formation modeling or drilling modeling software. The program instructions may be stored in a computer-readable memory, such as program data storage device 414. The memory medium storing the program instructions may be of any conventional type used for the storage of computer programs, including hard disk drives, floppy disks, CD-ROMs and other optical media, magnetic tape, and the like.
While the program instructions and the input data can be stored on and processed by the system computer 410, the results of the analyses and methods described herein are exported for use in mitigating vibrations. For example, the obtained drill tool assembly data and drilling parameters may exist in data form on the system computer. The system computer, utilizing the program instructions may utilize frequency-domain models to generate one or more vibration amplitude ratios, or one or more vibration indices as described herein. The vibration amplitude ratios may be stored on any one or more data storage devices and/or may be exported or otherwise used to mitigate vibrations. As described above, the vibration amplitude ratios may be used by an operator in determining design options, drill plan options, and/or drilling operations changes. The vibration amplitude ratios may be utilized by the computer system, such as to identify combinations of drilling parameters that best mitigate vibrations under given circumstances.
According to the representative implementation of
In one aspect, the inventive methodology involves use of a “base model” to develop and/or calculate the baseline solution, the frequency eigenmodes, and the dynamic linear response functions for a given set of input parameters. An exemplary model of this nature is provided below, and more details may be found in WO2009/155062. The base model is a first order dynamic perturbation model of the equations of motion for the drill tool assembly under given input drilling parameters and conditions. Although both the zeroth and first order terms in the dynamic variables are calculated, the dynamic model comprises simply the first order terms in the dynamic variables. Higher order terms in the perturbation theory could be calculated but are not provided here. The tractability and computational simplicity of the present methods are preserved through the use of a robust base model used to determine a baseline solution, or a baseline condition of the drill tool assembly in which no vibration is present. Linear response functions are also developed based on the base model. The linearization of the motion around the baseline solution allows independent linear harmonic analysis of the eigenstates at each vibration frequency and the use of superposition to analyze the overall dynamic motion. In some implementations, the vibration-related factors may be incorporated into the frequency-domain models by way of one or more linear response functions, which in some implementations may be incorporated as a piece-wise wave propagator for which individual pieces of the solution correspond to sections of the drill string that have constant properties, such as inner or outer diameter.
Drill tool assemblies can be considered as slender, one-dimensional objects, and their properties can be effectively described as a function of arc length, 1, and time, t. Incorporating in its entirety the methods described in greater detail in WO2009/155062, the configuration of the drill tool assembly can be uniquely defined in terms of a total axial elongation, or stretch, h(l,t), and total torsion angle, or twist, α(l,t). It may be assumed that the borehole exerts the necessary forces to keep the drill tool assembly in lateral equilibrium along its entire length. When the drill tool assembly is in the borehole, it is constrained by the forces imparted to it by the borehole walls, such that its shape closely follows the trajectory of the borehole, which can be tortuous in complex borehole trajectories. The dynamics of the drill tool assembly are represented by partial differential equations along with suitable constitutive relations, external forces and torques, and appropriate boundary conditions at the ends of the drill tool assembly. In some cases, the reference levels of downhole and surface vibration parameters identified above may be applied to the boundary conditions.
An exemplary flow chart 200 is presented in
Vibration Amplitude Ratio (VAR)
In many embodiments, the inventive methods may determine a vibration amplitude ratio in estimating vibration severity. The vibration amplitude ratio is defined as the ratio of one or more vibration amplitudes at one or more locations. In one aspect, this could for example be a ratio of downhole fluctuations in rotary speed to the average value of the surface rotary speed. Alternatively, this could be represented as a ratio of fluctuations in surface torque to a reference value of torque vibrations estimated from a model. This estimate determines the severity level associated with torsional oscillations, or, simply, the torsional severity estimate (TSE). Other vibration amplitude ratios can be developed including those for axial vibrations, such as an axial severity estimate (ASE).
Estimation and Characterization of Downhole Torsional Vibration Severity
The developments leading to the mathematical relation (63) in WO2009/155062 may be summarized in a generic form, whereby it may be realized that other implementations are feasible within the spirit of this invention. As WO2009/155062 contains a complete description of the governing drill string physics, this reference may be considered to be available for use with some embodiments of the methods disclosed herein. The model disclosed therein makes the so-called “soft-string” approximation, i.e. it assumes that the drill string has negligible bending stiffness. The use of a “stiff-string” model that includes drill string bending stiffness may also be used within the scope of the invention described herein.
The state vector [αP(l),τP(l)]T represents a harmonic torsional wave along the drill tool assembly. Here, αP(l) and τP(l) are the (complex) twist and torque amplitudes of the wave mode of period P at a distance l from the bit end, respectively. For this mode, the actual harmonic twist angle (in radians) and torque are given as a function of position l and time t by:
Here, Re represents the real part and j is the imaginary number. A 2×2 transfer matrix SP(l,l′) relates the state vectors at two different positions along the drilling assembly:
In one embodiment disclosed herein, Eq. (87) and (96) below are representative SP matrices. Of particular interest is the transfer matrix that relates the state at the bit end to the state at the surface (rig) end: S(MD, 0)=S−1(0, MD). For harmonic motion with period P, the corresponding states at the bit and surface end are given by:
The baseline solution, frequency eigenstates, and linear response functions provided by the base model may be used with the techniques taught and claimed herein to evaluate bit bounce and stick-slip tendencies of drill tool assembly designs, which may be by means of “vibration indices” derived from these results. Specifically, the effective torsional compliance of the drill tool assembly at the bit position is defined as:
The torsional compliance relates the angular displacement amplitude to the torque amplitude. The compliance is a complex function of frequency, ω, and has information on both the relative magnitude and phase of the oscillations. Detrimental behavior associated with torsional vibrations can potentially occur at resonant frequencies of the drill tool assembly, where “inertial” and “elastic” forces exactly cancel each other out. When this occurs, the real part of the compliance vanishes:
Re[CTPnbit]=0; n=1,2, . . . (5)
The resonant frequencies of the drill tool assembly have an associated period of oscillation, Pn (seconds). For instance, the first fundamental mode has a period of oscillation, P1 (seconds).
The cross-compliance is defined for a particular harmonic mode with period P (seconds) as the ratio of the vibration amplitude at the bit (for instance, RPM) to the vibration amplitude (for instance, torque) at the surface (here 60/P represents the number of periods per minute):
In order to make an accurate estimate of the downhole rotary speed fluctuations, it is useful to identify the dominant harmonic mode P. This will depend on the type of torsional oscillations that are present. In particular, there are two specific types of torsional behavior of interest: (i) unstable torsional vibrations associated with the resonant modes, often the primary or fundamental period P1 (stick-slip), and (ii) forced torsional vibrations associated with the periodic excitation of the drill tool assembly at a particular frequency.
As a simple illustrative example of this general method, as well as in order to introduce an alternate embodiment, consider a very simple drill tool assembly configuration that consists of a uniform drill string of length L and torsional stiffness GJ (where G is the shear modulus of the drill string material and J is its torsional moment of inertia) attached to a Bottom Hole Assembly (BHA) that is much stiffer, much shorter and with a much larger torsional moment of inertia. For the first resonant mode, the twist and torque have the following form:
The rig-to-bit transfer matrix has the simple form
For this simple case the rotary speed fluctuations at the bit can actually be deduced from the time derivative of the surface torque signal, along with the known information about the drill string (G, J and L):
On the other hand, if the BHA has negligible torsional moment of inertia, the twist and torque have the following form:
Thus, a similar relationship can be established:
Note that the two results are very similar, the key difference being a multiplicative factor of π/2.
In other more general situations with a more complex drill tool assembly model, more complex boundary conditions, or other vibration modes of interest, the general method outlined here and described in more detail in WO 2009/155062, one exemplary embodiment, can be used to compute a more accurate proportionality factor CP that relates the time derivative of the surface torque to the rotary speed fluctuations at the bit:
wherein L and GJ are now the total length and the torsional stiffness of the uppermost drill string section of the drill tool assembly, respectively. Depending on the application and utility of the downhole vibration indices, the accuracy of the results may be more or less critical.
One practical benefit of this method is that it automatically detrends the average or slowly varying portions of both signals, i.e., it is not sensitive to the slowly varying baseline torque and rotary speed. It is also not necessary to separately keep track of the period P1. However, in some instances reliability may be somewhat compromised from noisy measurements, so the sampling rate has to be sufficiently frequent to allow a good determination of the time derivative; alternatively, the use of more sophisticated methods may be applied to obtain a smoother estimate of the derivative. Also, it may be necessary to increase the surface data acquisition recording rate to facilitate the torque derivative method described above.
Using a combination of several downhole vibration severity estimation methods can potentially improve the robustness of the overall method. For example, alternate means of processing surface parameter data may lead to different values for the torsional severity estimate. Average values and other means of combining the results of multiple measurements may be used to seek the best estimate. These different TSE estimates, from both individual and combined parameters, may be evaluated using quality factor calculations in wells for which downhole measurements are available. This calibration process will help to determine the optimal means for processing surface measurement data to assure that the torsional severity estimates have the highest quality factors possible.
Exemplary flow charts are presented in
Although all computations could in principle be carried out in the surface monitoring system, pre-computation of P1, X and NF allows specialized software to be utilized, possibly at a central location by qualified users, for these computations. This not only significantly reduces the real-time processing power needed in the surface monitoring system, but also circumvents the problem associated with compatibility and inter-usability amongst various systems that might be deployed at various drilling locations. The advent of modern web-based applications based on streaming data from the drilling rig may also enable alternate implementations of these methods.
Unstable Torsional Behavior:
Unstable torsional vibration is reflective of downhole torque fluctuations from various origins and is typically associated with a dynamic instability or near-instability of the downhole drilling assembly. “Unstable torsional oscillations,” commonly referred to as “stick-slip,” have a characteristic period P that is determined primarily by the drilling assembly design parameters such as material properties (steel), dimensions (length, OD, ID, relative position along the assembly), and the measured bit depth (overall length of the drilling assembly). An exemplary calculation of this period can be obtained with a torsional harmonic wave mode in a drill tool assembly system with a “fixed” dynamic boundary condition at the rig end (corresponding to a constant rotary speed imposed by the rig control system) and a “free” dynamic boundary condition at the bit end (corresponding to a constant torque at the bit).
Primary Period:
For the aforementioned boundary conditions, we are interested in states in which αrig=τbit=0. Note that αrig and τbit refers to the dynamic twist and torque amplitudes, i.e. they are differences between the current values of those variables and their average, steady-state values. A solution to the transfer matrix equation with these constraints exists only for specific values of the harmonic period P. There exists a sequence of such modes of decreasing periods, whereby each successive mode shape in the sequence has one more “node” (position along the drilling assembly with no harmonic motion, i.e., α=0). These are referred to herein as “resonance modes” of the drilling assembly. Of particular interest is the mode with the longest period (P1), which has its only node at the surface (rig) end. During unstable stick-slip, the primary contributions to the torque oscillations observed at the rig end arise from this mode. A number of search algorithms are known that can be employed to identify this period P1. This period increases as a function of measured depth (MD) and is commonly in the range from approximately two to eight seconds for typical drilling assemblies and MD's.
The relevant dynamic boundary conditions at the surface (rig) end may be different under special circumstances; notably if different types of rotary speed controllers such as Soft-Torque™ and Soft-Speed™ are used. In that case, the appropriate boundary condition at the surface, along with τbit=0, must be applied to solve for P1. In general, if the boundary condition at the surface is not known, it is possible to determine the effective boundary condition by measuring both torque and rotary speed and constructing the effective rig compliance from the measurements, using one of several state variable observer methods.
Unstable Torsional Severity (TSEu):
When the period P is known as a function of MD, the cross-compliance at the primary period can be pre-computed for the section to be drilled. During drilling, the surface monitoring system may use the real-time MD and model results to compute TSEu as described above. Typically, the unstable torsional severity is associated with the primary resonant mode with period P1 and is equal to the torsional severity TSE1 evaluated at period P1. TSEu is also referred to herein in by the often commonly used vernacular of “unstable stick-slip” (USS), but the term TSEu is preferable as it reminds that the value is only an estimate. However, the terms are interchangeable.
Forced Torsional Behavior:
A second potential source of severe torsional oscillations is associated with the periodic excitation of the drilling assembly at a particular frequency. In most cases, the excitation occurs at or near the bit, at a multiple of the rotary speed (RPM). If this excitation period is close to one of the resonant mode periods of the drilling assembly (see above) large fluctuations may result, leading to stick-slip. Often, the primary excitation at period P=60/RPM is the dominant excitation so if the primary period is not observed in the torque signal and the actual periodicity is not observable, this value can be assumed in order to provide a conservative estimate of forced stick-slip. In this case, the cross-compliance is computed for a range of periods corresponding to the expected rotary speed ranges and depths. These are then converted to normalization factors using the relationship:
An exemplary calculation for torsional severity estimation during drilling may be made using the streaming surface torque signal in the following way. The torsional vibration amplitude is computed as the “peak-to-peak torque,” delta-Torque, or dTorque, and consequently may be used to estimate the torsional severity TSE1:
Alternatively, TSE1 can be obtained using the concept of a reference dTorque. The reference dTorque as calculated from the model cross-compliance and the surface rotary speed is a reference surface condition associated with full stick-slip at the bit. This quantity represents the torque fluctuation level corresponding to a condition where the bit oscillates between 0 and two times the average RPM. The reference dTorque, dTo, can be obtained for a range of rotary speeds and is obtained as:
Consequently, the estimated torsional severity is then obtained as:
Additionally and alternatively, TSE1 can be obtained by identifying the reference time-derivative of the surface torque signal for the particular mode of interest:
Consequently, the estimated torsional severity is obtained as:
There may be alternate ways to represent torsional severity that are known to those skilled in the art, and TSE1 can be converted to any one of these alternate representations if desired. Here, a value of TSE1 less than 1 represents rotary speed fluctuations at the bit that do not involve actual stopping or reversal of bit rotation, whereas a value larger than 1 corresponds to actual “sticking” or stopping of the bit during the cycle and should be a cause for concern.
This computation will provide a value for TSE1 continuously, regardless of whether the dominant torsional behavior associated with the fundamental mode is present or not. If the value reported to the driller is a cause for concern, the driller can verify that unstable stick-slip is present by inspecting the torque indicator and noting that the torque fluctuations have a characteristic period close to or slightly longer than P1. This period is dependent on the MD and increases with increasing MD. For typical drilling operations, this period is in the range of 2-8 seconds and is easily observable. If confirmed, the driller can take corrective action as desired.
If the torque oscillations have a significantly smaller period or no period is easily discernible due to sampling rate limitations, it is likely that “forced vibrations” are dominant. In this case, the surface monitoring system can be configured to display the forced torsional severity FTS instead. This is estimated by multiplying TSE1 with the appropriate “normalization factors” NF:
Alternatively, the surface monitoring system can display both forced and resonant vibration amplitude ratios and the driller can consider the appropriate severity level depending on the period of the dynamic torque signal. Both the nth resonant torsional severity estimate, TSEn, and the forced torsional estimate, TSEf (sometimes called Forced Stick Slip (FSS), but which is also an estimated value), can be combined into one or more torsional vibration amplitude ratios or torsional severity estimates (TSE). Other methods can be adopted to identify when these estimates can be used. For instance, if the surface monitoring system is capable of real-time spectral analysis, the torque signal can be analyzed for the prevalent period to automatically decide the type of stick-slip that is present, and the appropriate severity level can then be displayed. It is beneficial to the driller to know the type of torsional oscillations as well as the severity, since mitigation measures may be different for each type.
In one exemplary embodiment, a reference surface dTorque (dTo) can be obtained by calculating the cross-compliance at the stick-slip period using the drill tool assembly description and the rotary speed. This calculation is obtained based on a spectral analysis method wherein at any given time a specific frequency associated with the stick-slip period is used to extract the cross-compliance. A plot of the reference surface dTorque as a function of measured depth and rotary speed is outlined in
Another aspect to note here is that as the rotary speed increases, the reference surface dTorque also increases. In other words, there is a greater capacity to allow dTorque without encountering actual stopping of the bit, i.e., there is an increased “dTorque margin.” Other alternate representations of the reference surface dTorque include descriptions in tabular form and a fitted equation that describes the reference surface dTorque per unit RPM as a function of measured depth. Yet another alternate representation is to directly incorporate the cross-compliance instead of the concept of reference surface dTorque.
As noted above, the reference surface dTorque is obtained based on the fundamental period P1 at each Measured Depth. Thereby, the reference surface dTorque associated with forced torsional severity is incorporated to obtain more conservative reference surface dTorque values.
In one embodiment, the period associated with reference torsional oscillations can be described in terms of the measured depth. An exemplary chart is provided in
A chart of this form can be obtained either during real-time operations or pre-calculated beforehand. The benefit of such a chart in real-time operations is that the precise period of interest can be obtained along with information on stoppage time and the severity of the torsional oscillations. Alternatively, the chart can be described in a tabular form.
Characterization and Estimation of Axial Vibration Severity
The calculation of axial vibration amplitude ratio and axial severity may be accomplished using similar methods to that described above for the torsional mode. There may be a variety of embodiments of axial drill string vibration models that may be used to calculate the compliance factor considered below. The exemplary embodiment is the physics model described in WO 2009/155062. In this reference, the discussion leading up to equation (45) describes the modeling of axial vibrations that includes consideration of all the forces and moments acting on a drill string, assuming what is known as a “soft-string” model, i.e. with no bending stiffness of the string. The use of a “stiff-string” model that includes drill string bending stiffness may also be used within the scope of the invention described herein.
In a similar manner to the torsional vibration mode, the state vector [hP(l), TP(l)]T represents a harmonic axial wave along the drill tool assembly. Here, hP(l) and TP(l) are the (complex) stretch and tension amplitudes of the wave mode of period P at a distance/from the bit end, respectively. For this mode, the actual harmonic twist angle (in radians) and torque are given as a function of position l and time t by:
Here, Re represents the real part and j is the imaginary number. A 2×2 transfer matrix Sp(l,l′) relates the state vectors at two different positions along the drilling assembly:
Of particular interest is the transfer matrix that relates the state at the bit end to the state at the surface (rig) end: S(MD, 0)=S−1(0, MD). For harmonic motion with period P, the corresponding states at the bit and surface end are given by:
In one embodiment disclosed herein, Eq. (69) and (93) below are representative Sp matrices. It is generally understood (see, for example, Clayer et al. SPE 20447) that unlike torsional excitations, axial excitations typically manifest themselves as “displacement sources” and the typical dysfunction of “bit bounce” manifests itself as a dynamic fluctuation of WOB whose amplitude exceeds the average WOB. Thus, an analysis analogous to the torsional case can be done for axial vibrations. Of particular concern are harmonic axial modes in which small displacements at the bit may cause large WOB fluctuations, which can be identified through the effective axial impedance of the drill tool assembly at the bit position:
Naturally, this impedance will depend on the axial boundary conditions at the rig end, which can be identified for a particular rig and specific rig configuration. Factors that affect the axial compliance include the block height, mass of the traveling equipment, and number of drilling lines in use. At axial resonant frequencies of the drill tool assembly, the real part of the impedance vanishes:
Re└ZAPnbit┘=0; n=1,2, . . . (16)
In this case, the measurement that is readily available at most rig systems is the weight-on-hook (WOH), so the response function of interest is the amplification factor that relates WOH fluctuations at the surface to WOB fluctuations at the bit:
If an accelerometer measurement is available at the rig end, a person skilled in the art can alternatively utilize a cross-compliance that relates accelerations at the surface to WOB fluctuations at the bit instead, based on the teachings of this disclosure.
An exemplary calculation for axial severity during drilling can be made using the streaming surface signal to compute the hookload vibration amplitude as the “peak-to-peak hookload,” delta-Hookload, dHookload, or dWOH, and consequently estimate the axial severity estimate ASE1:
where ASE1 is estimated using the amplification factor AP1 evaluated at the fundamental period P1. Alternatively, ASE1 can be obtained using the concept of a reference dWOH that is associated with bit bounce. The reference dWOH represents the hookload fluctuation level corresponding to a condition where the bit oscillates between 0 and two times the intended surface WOB. The reference dWOH can be obtained for a range of WOB conditions and is obtained as:
Consequently, the estimated axial severity is then obtained as:
If the hookload oscillations have a significantly smaller period, or if no period is easily discernible due to sampling rate limitations, it is likely that “forced vibrations” may be the dominant characteristic. In this case, the surface monitoring system can be configured to display the forced axial severity ASEf instead. This is estimated by multiplying ASE1 with the appropriate “normalization factors” NF:
Vibration Amplitude and Time Period Estimation from Surface Signals
Various methods to measure vibration amplitudes and periods associated with a real-time signal stream are known in the art. In an ideal situation where all of the peak-to-peak surface signal fluctuations can be attributed to a single harmonic mode, the peak-to-peak vibration amplitude corresponds to the difference between the maximum and minimum amplitudes of the surface operating parameter. In reality, a surface signal such as illustrated in
According to some embodiments, a desirable window size may be between 2 to 10 times the expected primary period P1 at that time, to facilitate obtaining an accurate estimate of an average value as well as a peak-to-peak envelope for the surface signal. In the exemplary case of
dX=2*(Xmaximum−Xaverage) or dX=(Xmaximum−Xminimum).
Another approach is to calculate the standard deviation of a time series in a rolling data buffer, wherein the new values displace the oldest values and the data window is continually refreshed. The constant or steady-state component is eliminated from the standard deviation calculation, and if the oscillating part is represented as a sinusoidal wave X(t)=A sin(2πt/P) then the standard deviation may be found as σX−A/√{square root over (2)}. Using the notation above, the “delta-X” for this parameter is then found to be
dX=σX2√{square root over (2)}. (22)
Still other methods are known for computing vibration amplitude, both offline and online. One offline method, which may involve only a slight time delay in the availability of the calculation results, is a phase-compensated moving average filter that can be used to compute the envelope of the surface operating parameter signal. Other methods may include computing a peak-to-peak value from a real-time data stream, including methods to reduce the effect of noise, including filtering. All such methods to obtain the peak-to-peak surface operating parameter fluctuations are within the scope of this invention. In certain instances, if downhole operating parameter fluctuations are available, these can then be used to obtain improved accuracy.
The period of oscillation may also be estimated from the surface signals such as surface torque, hookload, and rotary speed. An exemplary illustration of how this can be accomplished is provided in
The estimates of significant period(s) can be used to obtain more information about the downhole scenario. In one embodiment, knowledge of the reference peak-to-peak fluctuations in surface parameters and the reference period(s) associated with the dominant harmonic modes can be combined with information about the identified periods over the specified time intervals to obtain precise information on the extent of the “stopped” time. In an alternate embodiment, if this period is observed to be greater than the estimated fundamental period or other significant periods, a measure of the stopped time (the time that the bit stops rotating during any given cycle) can be obtained by direct comparison of the estimated and measured periods.
Fourier analysis can provide an estimate of the period of a signal, as well as provide the amplitude of the oscillation for use in the calculation of the vibration amplitude dX as discussed above. This added benefit provides further motivation to use Fourier methods, provided suitable data input streams with appropriate sampling rates may be obtained, and also provided that a surface data acquisition system can be properly configured to take advantage of the methods herein described. An exemplary illustration of how this can be accomplished is provided in
Determining Quality of the Estimated Vibration Amplitude Ratio
To determine the quality of the dynamic severity estimate and to calibrate the methods, comparison with actual downhole vibration severity information and/or vibration data is one exemplary means for evaluation. Downhole data could be obtained from one or more of a downhole instrumented sub with accelerometers, force and torque sensors, and downhole measurement-while-drilling (MWD) equipment that record rotary speed, acceleration, WOB, and other drilling parameters. The quantities that determine axial and/or torsional severity are then designated as VARmeasured to signify a measured vibration amplitude ratio. The surface estimated vibration amplitude ratios can be one or more of the torsional/axial modal vibration severity and torsional/axial forced vibration severity indices. These vibration amplitude ratios are designated as VARestimated. The reference value for the exemplary case is considered to be 1. If VAR=1, it is assumed that in the torsional case, we are at full stick-slip.
The quality factor may be defined in terms of conditional relations that depend on the values of the vibration amplitude ratios as follows:
Though complicated, this method gives full credit for estimates of full-stick slip that are detected, with no penalty for the amount of difference if there is actually full stick-slip at the bit.
For all values of VAR, another quality factor may be written as
or still alternatively,
While the quality factor QF describes the quality of estimation, both false negatives and false positives are lumped together. An alternative is to count the quality factor associated with false positives and false negatives separately and focus on false occurrences. Another alternative quality factor measurement is the goodness, which excludes false negatives/positives and counts the cases where both the measured and the estimated values are in agreement of the absence/existence of a vibration dysfunction. Cumulative statistics may be obtained and plotted in terms of histograms or other common statistical display measures. It is desirable to have a quality factor greater than 0.8 (80%) between surface estimates and downhole measurements to validate the methods described herein.
Combined Analysis of Torsional Severity and Drilling Operating Parameters
In one embodiment, the driller or engineer may consider the torsional vibration type and severity under different types of boundary conditions. In typical torsional vibration scenarios, observed under stringent rotary speed control where the rig end rotates substantially at the set rotary speed, the drill tool assembly can be considered as having a torsionally clamped boundary condition at the surface and a free condition at the bit. An alternate scenario is to have a free boundary condition at both the bit and the surface, corresponding to torque limit control. When more sophisticated top-drive controllers such as Soft-Torque™ and Soft-Speed™ are used, the boundary condition at the surface is effectively somewhere in between these extreme criteria, and both Torque and RPM fluctuations may be present at the rig end. In such situations, it is possible to solve the torsional model with various ratios of surface dTorque to dRPM and construct a hybrid reference condition that considers all such possibilities. An exemplary graphical form of a reference condition is illustrated in
Combined Analysis of Metrics
In another embodiment of the methods according to the present invention, the driller or engineer can consider the torsional vibration type and severity along with real-time MSE information to obtain a more comprehensive picture of downhole conditions. This may be facilitated by a display that combines all of the pertinent information advantageously. An example is illustrated in
Another embodiment of the inventive subject matter is to describe the variation in TSE and MSE in terms of a performance metric. This performance metric can be one or more of ROP, footage drilled, tool life, non-productive time associated with drilling, and formation, or some combination thereof. An example of how these performance metrics can be displayed is illustrated in
An exemplary method is to use the changes in performance metric, combined with changes in the severity estimate. For instance, objective functions of the following forms may be used to evaluate controllable parameters in conjunction with the concept of the “dTorque Margin”:
These functional forms may be augmented with comparable terms in MSE for completeness, without departing from the spirit of the invention.
The objective function here is to maximize ROP while minimizing TSE. For instance, maximizing ROP can be accomplished by increasing WOB. When the WOB is increased, the dTorque typically goes up and the TSE goes up. An objective is to ensure that there is sufficient WOB to drill efficiently without going into an undesirable operating zone. In other words, the operating conditions are maintained such that the measured dTorque is less than a specific percentage of the reference dTorque. The “dTorque Margin” represents the available excess dTorque with which drilling can be carried out without concern for severe torsional dysfunctions or stick-slip. The maximum value of the dTorque Margin is obtained by subtracting the surface dTorque from the reference surface dTorque, assuming that dTorque is less than the reference dTorque. The use of objective functions provides a formal approach for estimation of the “available” dTorque Margin in relation to the maximum dTorque Margin. It is also important to point out that the methodology and algorithms presented in this invention are not limited to these three types of objective functions. They are applicable to and cover any form of objective function adapted to describe a relationship between drilling parameters and drilling performance measurements.
Embodiment of a Base Model for Torsional and Axial Vibrations
One embodiment of a base model of torsional and axial vibrations of a drill string follows directly from the patent application WO 2009/155062. The zero-order and first-order terms of the perturbation expansion of the drill string equations of motion for axial and torsional vibrations are disclosed. This reference includes modeling elements that include the physical effects of wellbore profile (determined by the inclination and azimuth angles of the borehole as a function of measured depth), drill string description including the effects of tool joints, inertia, friction and viscous damping, and other details necessary to provide high quality model results necessary for the present invention. This is a “soft-string” model with no bending stiffness of the string. The use of a “stiff-string” model that includes drill string bending stiffness may also be used within the scope of the invention described herein. The present model will be disclosed in summary form, and patent application WO 2009/155062 should be referred to for additional details.
The present systems and methods utilize an exemplary “base model.” The present methods and systems can be adapted to apply to different equations of motion and/or different base models than those presented herein. Accordingly, for the purposes of facilitating explanation of the present systems and methods, one suitable formulation of a base model is described herein and others are within the scope of the present disclosure.
A borehole with a particular trajectory is created by the action of a drill bit at the bottom of a drill tool assembly, consisting of drill pipe, drill collars and other elements. Drilling is achieved by applying a WOB, which results in a torque, τbit, at the bit when the drill tool assembly is rotated at rotary speed,
The mechanical rotary power, ΩRPMτbit, is supplied to the bit and is consumed during the rock cutting action. The torque is provided by a drilling rig, and the WOB is typically provided by gravitational loading of the drill tool assembly elements. The application of WOB forces a portion of the drill tool assembly near the drill bit into compression.
The borehole centerline traverses a curve in 3-D, starting from the surface and extending out to the bottom of the hole being drilled. The borehole trajectory at arc length l from the drill bit in terms of the inclination θ and azimuth φ as a function of measured depth (MD), global (x, y, z) and local (t, n, b) coordinates and the local borehole curvature Kb can be written as:
Here, the unit normal vector n is in the plane of local bending and perpendicular to the tangent vector t, whereas the unit binormal vector b is perpendicular to both t and n. The vectors x, y and z point to the East, North, and Up, respectively.
Drill tool assemblies can be described as a function of arc length, s, along their centerline in the unstressed state. In the stressed condition the drill tool assembly is stretched and twisted relative to the unstressed condition. The differences between the stressed and unstressed conditions are discussed further below. For the purposes of the present systems and methods, the drill tool assembly is assumed to consist of elements attached rigidly end-to-end along a common axis of rotational symmetry, each element having a uniform cross-section along its length, free of bend and twist in its unstressed state. The description of each drill tool assembly element includes information about the material (elastic modulus, E, shear modulus, G, density, ρ) and geometrical properties (area, A, moment of inertia, I, polar moment of inertia, J). This information can typically be obtained from drill tool assembly descriptions and technical specifications of the drill tool assembly components.
When the drill tool assembly is in the borehole, it is constrained by the forces imparted to it by the borehole walls, such that its shape closely follows the trajectory of the borehole, which can be tortuous in complex borehole trajectories. It is presently understood that it may be possible to improve the accuracy of the model by using a stiff-string model and resolving bending moments at the BHA, or possibly along the entire drill tool assembly if necessary. Examples of such models have been disclosed at least in “Drillstring Solutions Improve the Torque-Drag Model,” Robert F. Mitchell, SPE 112623. Use of such improvements in the base model are within the scope of the present disclosure. For example, while some of the discussion herein will reference assumptions regarding equations that can be simplified or solved by using this soft-string approximation, any one or more of these assumptions could be replaced utilizing appropriate stiff-string models.
In some implementations, the exemplary base model considers the motion of the drill tool assembly while it is rotating at a particular bit depth (BD), WOB, and nominal rotation speed. The lateral displacement constraint leaves only two kinematic degrees of freedom for the drill tool assembly; stretch h and twist α. The overall motion of the drill tool assembly can be described by:
where, h0 and α0 represent the “baseline solution”—the amount of stretch and twist present in the drill tool assembly when it is rotating smoothly, and hdyn and αdyn represent the solutions to the dynamic motion of the drill tool assembly relative to the baseline solution. The model considers only small deviations around the baseline solution, allowing dynamic motions at different frequencies to be decoupled from each other.
The motions of the drill tool assembly are accompanied by internal tension, T, and torque, τ, transmitted along the drill tool assembly, which can be likewise described as:
where Tdyn and τdyn represent the solutions to the dynamic motion of the drill tool assembly relative to the baseline solution. In the linear elastic regime and within the soft-string approximation, these are given in terms of the drill tool assembly configuration as:
The drill tool assembly elements are also subject to a variety of external forces, fbody, and torques, θbody, per unit length that affect their motion. The axial equation of motion is obtained by equating the net axial force to the force associated with the axial acceleration of the element mass:
ρA{umlaut over (h)}=T′+fbody·t, (34)
where t is the unit vector along the tangent direction. The torsional equation of motion is obtained by equating the net torque along the tangent vector to the torsional moment times angular acceleration of the element:
−ρJ{umlaut over (α)}=−τζ+θbody·t. (35)
External Forces and Torques
At the junction of two drill tool assembly elements, the stretch, h, and twist, α, are continuous. Since no concentrated forces or torques are present, the tension, T, and torque, τ, are also continuous across these boundaries. The partial differential equations (PDE's), constitutive relations, and external forces and torques fully describe the dynamics along the drill tool assembly once appropriate boundary conditions are specified at the ends of the drill tool assembly.
Three types of external forces, f, and torques, θ, are considered: gravitational (fg, θg), mud (fmud, θmud), and borehole (fbh, θbh). The body force and torque is a composite sum of these three forces and torques,
fbody=fmud+fbhfg, (36)
θbody=θmud+θbh+θg. (37)
Gravitational forces set up the characteristic tension profile along the drill tool assembly, which further affects torque, drag and drill tool assembly dynamics. The gravitational force per unit length acting on an element is
fg=−(ρ−ρmud)Agz, (38)
where z is a unit vector that points upward and which takes into account the buoyancy associated with the mud density ρmud. Since the elements have an axis of symmetry, no torque is generated by gravity: θg.=0.
During drilling operations, the drilling mud shears against both the inside and the outside of the drill tool assembly, and creates forces, fmud, and torques, θmud, per unit length that resist motion. In the absence of lateral motion according to the constraints described above, no lateral forces are generated by the mud. Also, any torque that is not along the local tangent will be cancelled out by borehole torques, so we need only consider the component of torque along the tangent vector. The mud forces and torques are then obtained as
fmud≡fmudt, (39)
θmud·t≡θmud. (40)
These forces and torques can be separated into a steady-state portion associated with the steady-state rotation of the drill tool assembly and circulation of the mud at average pump pressure, and a dynamic portion associated with dynamic variations in the mud pressure and the relative motion of the drill tool assembly with respect to steady-state.
For the purposes of the presently described implementation, it is assumed that the borehole forces dominate the steady-state force balance. The hook load differences between pumps-off and pumps-on and the effects of mud pump strokes and active components such as MWD systems that generate axial forces are assumed to be negligible in this exemplary embodiment. These assumptions simplify the solution but are not required for implementation of the present systems and methods. The only mud effects that the model takes into account are those associated with the dynamic motion of the drill tool assembly with respect to its steady-state rotation. Since axial and torsional movements of the elements do not displace any mud, their main effect is to create a shearing motion of the mud adjacent to the drill tool assembly surface and to dampen dynamic vibrations around the steady-state.
There may be several possible dynamic models of the mud system that may be considered to be within the scope of this model. For example, one or more of the assumptions described above may be made differently, thereby altering the formulation of the model. One example of a suitable dynamic model of the mud system comprises the superposition of the dynamic effects of the mud system on the baseline solution using a model for shear stress on an infinite plane. The amplitude of the shear stress acting on an infinite plane immersed in a viscous fluid and undergoing an oscillatory motion parallel to its own surface at an angular frequency ω is given by:
where αω is the displacement amplitude of the plane motion, ρmud is the mud density, j is an imaginary number, and δω, the frequency-dependent depth of penetration, is given by
δω=√{square root over (2ηpl/ωρmud)}, (42)
where ηpl is the plastic viscosity of the drilling mud under pumps-on conditions.
For the typical mud plastic viscosities ηpl densities ρmud, and frequencies ω of interest, the penetration depth is small compared to the inner and outer radii of the element; δω<<ID, OD. The mud plastic viscosity term is not restricted to the Bingham model and can be easily generalized to include other rheological models, in which the viscosity term varies with rotary speed. In the high-frequency limit, Eq. 41 can be used to approximate the shear stress on an annular object. For axial motion at frequency ω, this term results in a mud-related axial force per unit length:
fmud,ω≈σmud,ω(πID+πOD), (43)
where the axial displacement amplitude is given by αω=hω. Similarly, the torque per unit length associated with torsional oscillations is given by:
where the torsional displacement amplitudes at the ID and OD are given by αω(ID)=αω·ID/2 and αω(OD)=αω·OD/2, respectively. The total mud force for a general motion can be obtained by summing over all frequencies.
The borehole walls exert forces and torques that keep the drill tool assembly along the borehole trajectory. The currently described model assumes that each element has continuous contact with the borehole, consistent with the soft-string approximation, and that no concentrated forces are present. Other models that may be implemented within the scope of the present systems and methods may make different assumptions. For example, as discussed above, other models may use stiff-string approximations for some or all of the drill tool assembly. The contact with the borehole is localized somewhere along the circumference of the element, and rc denotes the vector that connects the centerline to the contact point within the local normal plane, whose magnitude, rc, is equal to half the “torque OD” of the element. The borehole force per unit length, fbh, can then be decomposed into axial, radial and tangential components as follows:
fbh≡fat+fn=fat−frrc/rc+fτ(t×rc)/rc. (45)
Here, a sign convention is used such that fr and fτ are always positive, provided that the drill tool assembly rotates in a clockwise manner when viewed from above. fn is the total borehole force in the local normal plane, with magnitude fn.
Four equations are needed to determine the three force components and direction of rc in the local normal plane. Since no lateral motion is allowed in the presently described implementation, imposing a force balance in the local normal plane yields two equations. Collecting borehole forces on one side of the equation and noting that there are no lateral mud forces present, gives,
fn=κbT+fg−(fg·t)t. (46)
Next, enforcing Coulomb friction against the borehole wall with a friction angle ψC provides two additional equations,
In general, ψC can be a function of the relative velocity, νrel=√{square root over (h{dot over (h)}2+{dot over (α)}2rc2)}, of the element with respect to the borehole. The dependence of the friction angle, ψC, on the relative velocity of the element, νrel, with respect to the borehole can be expressed in terms of a logarithmic derivative,
A negative value for Cμ represents a reduction of friction with increasing velocity, which may be referred to as velocity-weakening friction. Such a situation can have a significant impact on the stability of torsional vibrations and stick-slip behavior of the drill tool assembly. This equation represents one manner in which a velocity-dependent damping relationship may be incorporated into the models utilized in the present systems and methods. Other equations and/or relationships may be incorporated as appropriate.
The constraint on lateral motion also implies that there is no net torque in the local normal plane, so any applied torque that is not along the tangent vector will be cancelled out by the borehole. Thus, the equations of motion are obtained by considering the component of torque that is along the local tangent direction, which is responsible for rotating the drill tool assembly. This component of torque per unit length exerted by the borehole is given by:
θbh·t=rcfr. (50)
The baseline solution is a particular solution of the equations of motion that corresponds to smooth drilling with no vibration, at a particular bit depth, weight on bit, and specified drill tool assembly rotary speed that results in a rate of penetration. The equations of motion are then linearized around this baseline solution to study harmonic deviations from this baseline solution. An exemplary baseline solution is described below. As described above, a variety of equations could be used to describe the motion of the drill tool assembly considering the multitude of relationships and interactions in the borehole.
Baseline Solution
In the baseline solution, every point along the drill tool assembly has a steady downward velocity equal to the ROP. Deviations in this motion are very small over the typical vibration profiles of interest (smooth drilling with no vibration); hence these will be ignored during this steady downward motion. The drill tool assembly also rotates at a steady rotary speed dictated by the imposed surface rotary speed. It is also assumed that positive RPM corresponds to clockwise rotation of the drill tool assembly when viewed from the top. The baseline solution can be written as,
h(l,t)=h0(l), (51)
α(l,t)=ΩRPMt+α0(l), (52)
such that the baseline displacement h0 and twist α0 do not change with time. From the constitutive relations, it follows that baseline tension T0 and torque τ0 also do not change with time and are function of position l only. The subscript “0” is used to denote the baseline values of all variables and parameters.
First, the axial forces and displacements are obtained. It is seen that fa0=0, that is, the borehole does not exert any axial forces on the drill tool assembly. Then, the axial baseline solution for the composite drill tool assembly and the boundary conditions at the bit (T0(0)=−WOB, h0(0)=0) can be computed from:
Next, the tangential borehole force is obtained assuming no axial borehole forces:
fr0=fn0 sin ψC0. (55)
This enables computation of the baseline twist and torque along the drill tool assembly, ignoring the contribution of the mud torque, θmud, to the baseline torque. The result is another set of first-order ODEs:
Based on the boundary conditions at the bit (τ0(0)=τbit, α0(0)=0), the baseline solution for the twist and torque can be obtained by integration, just as in the axial case. In general, the torque generated at the bit cannot be controlled independently of the WOB; the two quantities are related through bit aggressiveness. The present model relates the bit torque to WOB through an empirical bit friction coefficient, μb,
The model uses the input parameter μb to compute the baseline solution. The torque at the bit enters the baseline torque solution only additively, and does not influence the dynamic linear response of the drill tool assembly; it is there mainly to enable calibration of the model with surface measurements.
For the numerical implementation of this solution scheme, the model interpolates the inclination, cos θ, and curvature, κb, from survey points to the midpoint of each element. The expressions, A, E and ρ are piece-wise constant over each drill tool assembly element. Also, the stretch of the drill tool assembly elements is ignored during the integration where dl=ds is assumed. Since all other drill tool assembly properties are constants within each element, the solution at each element boundary is obtained by applying the following recursive sums:
where fn0,i is the borehole force of the ith element of the drill tool assembly, T0,i-1/2 is the arithmetic average tension of the (i−1)th and ith elements of the drill tool assembly, and τ0,i-1/2 is the arithmetic average torque of the (i−1)th and ith elements of the drill tool assembly. Note that the tension along the drill tool assembly is needed for all of the computations in the above implementation and is the first quantity to be computed.
Harmonic Wave Equation
Having computed the baseline solution for a particular bit depth, WOB, and rotary speed, small motions hdyn and αdyn of an individual element may be calculated around this solution along with the associated forces (Tdyn) and torques (τdyn) to model the vibrations of the drill tool assembly.
Beginning with the axial equations, the change in axial borehole force is obtained to linear order in dynamic variables as,
Substitutions and rearrangement yields:
for each frequency component ω where
This second-order linear ODE has the following solution:
hω(l)=hωuejk
where hωu and hωd are arbitrary constants that represent the complex amplitude of upwards and downwards traveling axial waves along the elements of the drill tool assembly, respectively. The associated wave vector, ka, at frequency ω is given by:
In the absence of mud and borehole effects, this dispersion relation reduces to the well-known non-dispersive longitudinal wave along a uniform rod. Due to the large wavelengths associated with the frequency range of interest, these waves typically travel along the entire drill tool assembly. The corresponding tension amplitude is given by:
The state of the axial wave at each frequency is uniquely described by hωu and hωd. However, it is more convenient to represent the state of the axial wave by the axial displacement hω and tension Tω instead, since these have to be continuous across element boundaries. The modified expression is obtained by combining equations in matrix form at two ends (locations l and l−L) of an element of length L,
Thus, as a first step in obtaining the dynamic response of the drill tool assembly at a given frequency ω, the present model computes the transfer matrix for each element:
where ka,i is obtained using previous equations. For an axial vibration at that frequency, the state vector between any two points along the drill tool assembly can be related to each other through products of these transfer matrices:
The transfer matrix Eq. (70) can be used to relate the axial vibration state anywhere along the drill tool assembly to, for example, the state at the surface end of the drill tool assembly. However, in order to solve for the response of the drill tool assembly to a particular excitation, it is necessary to specify the relationship between the displacement and tension amplitudes at the surface. The simplest boundary condition is to assume that the rig is axially rigid and has perfect rotary speed control, such that
hrig≡hdyn(MD)=0, αrig≡αdyn(MD)=0, (71)
where MD denotes the position of the rig along the drill tool assembly. In general, a rig should have finite compliance against the axial and torsional modes. The response of a drilling rig is dependent on the rig type and configuration and can change rapidly as the frequency of the vibration mode sweeps through a resonant mode of the rig. The response of the drilling rig can be modeled and incorporated into the present systems and methods in a variety of manners, including the approach described below.
For the case of axial motion, the drill tool assembly can be assumed to be rigidly attached to the top drive block, which can be approximated as a large point mass Mrig. This block is free to move up and down along the elevators, and is held in place by a number of cables that carry the hook load. There are also some damping forces present, which are assumed to be proportional to the velocity of the block. Thus, for small amplitude vibrations, a simple representation of the dynamics of this system is a mass-spring-dashpot attached to a rigid end, with a spring associated with the hoisting cables and a dashpot representing the damping. Here, Thook reflects the upwards force exerted on the block by the rig, including the spring and the damping force. Imposing force balance for the baseline solution yields:
Thook,0=T0(MD)+Mrigg. (72)
The hoisting cable length is adjusted to achieve the desired hook load; therefore the position of the baseline axial displacement is immaterial and is not needed to compute the baseline solution. However, this length sets the equilibrium position of the spring. When the block mass moves away from the baseline position, a net force is exerted on it by the drill tool assembly and the rig. The dynamic hook load is given by:
Tend=−krighrig−γrig{dot over (h)}rig. (73)
Newton's equation of motion for the block mass yields the following relation between vibration amplitudes at each frequency:
−Mrigω2hrig,ω=−Trig,ω+Tend,ω=−Trig,ω−(krig−jωγrig)hrig,ω. (74)
Thus, the axial rig compliance, based on a reference frame fixed at the rig, is given by:
This quantity measures the amount of axial movement the block mass will exhibit for a unit axial force at a particular frequency ω. It is a complex-valued function whose magnitude gives the ratio of the displacement magnitude to force magnitude, and whose phase gives the phase lag between the forcing function and the resulting displacement.
The dynamic response of the mass-spring-dashpot system is well known and will only be described briefly. Three parameters are needed to fully describe this simple dynamic rig model. The block mass is typically estimated from the hook load reading with no drill tool assembly attached. The spring constant can be estimated from the length, number and cross-sectional area of the hoisting cables. These two parameters define a characteristic rig frequency, ωrig,α≡√{square root over (krig/Mrig)}, for which the displacement of the block is 90° out of phase with the dynamic force. The severity of the rig response at this frequency is controlled by the rig damping coefficient; critical damping occurs for γrig=γcrit≡2Mrigωrig. Since the rig frequency and the amount of damping relative to the critical damping is more intuitive and easier to observe, the current model uses Mrig, ωrig and γrig/γcrit as inputs in order to compute the dynamic response. The “stiff-rig” limit can be recovered by considering the limit ωrig→∞, where the compliance vanishes. At this limit, the rig end does not move regardless of the tension in the drill tool assembly.
In general, the dynamic response of the rig is much more complicated. However, all the information that is necessary to analyze vibration response is embedded in the compliance function, and the model framework provides an easy way to incorporate such effects. If desired, it is possible to provide the model with any compliance function, possibly obtained from acceleration and strain data from a measurement sub.
As a practical matter, the effective compliance of the rig will vary with the traveling block height and the length and number of the cables between the crown block and traveling block. In the drilling of a well, the traveling block height varies continuously as a joint or stand is drilled down and the next section is attached to continue the drilling process. Also, the number of such cable passes may vary as the drilling load changes. The derrick and rig floor is a complex structure that is likely to have multiple resonances which may have interactions with the variable natural frequency of the traveling equipment. For these reasons, in addition to a well-defined resonance with specified mass, stiffness, and damping, and in addition to the “stiff rig” limit or alternatively a fully compliant rig, it is within the scope of this invention to consider that the surface system may be near resonance for any rotary speed under consideration. Then some desirable configurations and operating conditions may be identified as having desired index values despite possible resonance conditions in the rig surface equipment.
Eqs. (46) and (51) can be combined to obtain the vibration response everywhere along the drill tool assembly, associated with unit force amplitude at the surface:
Due to the linearity of the equations, the actual dynamic motion of the drill tool assembly at a given point is given by a linear superposition of these state vectors with different amplitudes at different frequencies. The main interest will be the dynamic linear response of the system to excitations at a given point along the drill tool assembly. The response of the system to multiple excitations can likewise be analyzed using the superposition principle.
In defining the vibration performance of the drill tool assembly, the primary quantity of interest is described by the way it responds to excitations at different frequencies caused by the drill bit. The effective drill tool assembly compliance at the bit can be defined as:
which is given by the ratio of the elements of {tilde over (S)}a at the bit. General linear response functions that relate amplitudes at different positions along the drill tool assembly can also be defined.
Turning now to the torsional equations, the methodology used for obtaining the expressions for torsional waves is similar to that described above for axial waves. As suggested above and throughout, while particular equations are provided as exemplary equations and expressions, the methodology used for obtaining these equations and expressions is included within the scope of the present disclosure regardless of the selected starting equations, boundary conditions, or other factors that may vary from the implementations described herein. Similar to the methodology used for axial waves, the dynamic torque associated with the borehole forces is computed using the lateral motion constraint and the Coulomb criterion. Expanding the lateral force balance to linear order in dynamic variables and eliminating the baseline terms to obtain:
fn0fn,dyn=└κb2T0+(ρ−ρmud)g(κb·z)┘Tdyn. (78)
To linear order, the change in the instantaneous friction coefficient can be obtained as
Thus, expanding to linear order and eliminating baseline terms yields:
The borehole torque associated with each torsional frequency component is:
The dynamic variation in the tension, associated with axial waves, couples linearly to the dynamic torque in the curved section of the borehole. The present model currently decouples these effects and explores axial and torsional modes independently. The decoupling is accomplished by setting the tension, Tdyn, to zero while analyzing torsional modes.
For each frequency component, substituting these into the torsional equation of motion and eliminating baseline terms yields:
This equation can be rearranged to yield:
where
This equation has exactly the same form as the axial equation, with the solution:
αω(l)=αωuejk
where the associated wave vector, kτ, at frequency, ω, is given by:
In the absence of mud and borehole effects, this dispersion relation reduces to the well-known non-dispersive torsional wave along a uniform rod. Once again, borehole and mud damping is typically relatively small, resulting in a weakly damped, nearly non-dispersive wave along the drill tool assembly. These waves typically travel along the entire drill tool assembly rather than just in the bottom hole assembly. One significant difference is that the effective damping associated with the borehole can be negative when the friction law has velocity-weakening characteristics, that is, Cμ<0. This has important implications for stick-slip behavior of the drill tool assembly.
As discussed above, the velocity-dependent damping relationships incorporated into the models of the present systems and methods provide models that are more reliable and more accurate than prior models. More specifically, it has been observed that the mud damping effect increases with increasing velocity whereas the borehole damping effect actually decreases with increasing velocity. Accordingly, in some implementations, models that incorporate both mud effects and borehole effects may be more accurate than models that neglect these effects. While the mud effects and borehole effects may be relatively small, the appropriate modeling of these effects will increase the model accuracy to enable drilling at optimized conditions. Because the costs of drilling operations and the risks and costs associated with problems are so high, misunderstandings of the drilling operations, whether for over-prediction or under-prediction, can result in significant economic impacts on the operations, such as in additional days of drilling or in additional operations to recover from complications.
The torque amplitude is given by:
As in the axial case, the transfer matrix formalism can be used to relate twist and torque amplitudes at the two ends of an element:
The rest of the torsional formulation precisely follows the axial case, with the appropriate substitution of variables and parameters. The torsional compliance at the surface is defined similarly, using appropriate torsional spring, damping and inertial parameters.
In addition to the elements of the drill tool assembly, the model can accommodate special elements, in its general framework. In general, these can be accommodated as long as expressions relating the baseline solution across the two ends, as well as its associated dynamic transfer matrix, can be described.
Many tubular components of the drill tool assembly, especially the drill pipes, do not have a uniform cross-sectional profile along their length. They tend to be bulkier near the ends (tool joints) where connections are made, and slimmer in the middle. Heavy weight drill pipe and other non-standard drill pipe can also have reinforced sections where the cross-sectional profile is different from the rest of the pipe. Many drill pipes also have tapered cross sections that connect the body of the pipe to the tool joints at the ends, rather than a piecewise constant cross-sectional profile. To construct a drill tool assembly, many nearly identical copies of such tubular components are connected end-to-end to create a structure with many variations in cross-section along its length. Representing each part with a different cross-section as a separate element is tedious and computationally costly. It is desirable to use a simpler effective drill tool assembly description to speed up the computation and reduce the complexity of the model. This can be achieved by taking advantage of the fact that for a section of the drill tool assembly consisting of a series of tubulars of nominally the same design and length, typically around 10 m (30 ft), the variations in cross-section are nearly periodic, with a period (˜10 m) that is much smaller than the wavelengths associated with axial and torsional vibrations of interest. Thus, a method of averaging can be employed to simplify the equations to be solved. This method, as it applies for the problem at hand here, is disclosed below.
Consider a section of the drill tool assembly consisting of a number of nominally identical components of length, L, attached end-to-end, for which the cross-sectional area, A, moment of inertia, I, and polar moment of inertia, J, are periodic functions of arc length, l, with a period L that is considered short compared to the characteristic wavelengths of interest. Then, the axial baseline solution can be approximated by:
where the angular brackets denote averaging over one period of the variation:
Similarly, the torsional baseline solution can be obtained by replacing the torque outer diameter, rc, and the inverse of the polar moment of inertia 1/J, by their averaged versions. By replacing the geometrical parameters with their averaged values, it is no longer necessary to break up the drill tool assembly into elements of constant cross-section.
Note that inversion and averaging operations are not interchangeable; for example, (1/A) is not equal to 1/<A> unless A is a constant. For a given drill tool assembly component of specified cross-sectional profile, we can define the following shape factors:
For a component with a general cross-sectional profile, these shape factors are always greater than or equal to one, the equality holding only when the cross-section remains constant along the component.
Now turning to the harmonic wave equations, when the geometry parameters are no longer a constant along the arc length,
After applying the method of averaging to the individual elements of the matrix, and further manipulation of equations familiar to someone skilled in the art, the generalized version of the axial transfer matrix is obtained as:
where the subscript i has been dropped for simplicity. The averaging process also affects the mud and borehole damping parameters as follows:
The averaged torsional equations can be obtained similarly, with the resulting transfer matrix having the same form as above with the appropriate substitutions of torsional quantities:
where, the torsional damping parameters are also appropriately averaged.
The most significant effect of using drill tool assembly components with a non-uniform cross-section is to change the wave vectors associated with axial and torsional waves at a given frequency by a constant shape factor. In other words, the velocities of axial and torsional waves along this section of the drill tool assembly are reduced by sA and sJ, respectively. This causes an associated shift of resonant frequencies of the drill tool assembly to lower values, which can be important if the model is used to identify rotary speed “sweet spots”. As mentioned at various places herein, the costs of drilling operations makes even minor improvements in predictions and corresponding operations efficiencies valuable.
To illustrate the magnitude of this effect, let us consider a typical 5″ OD, 19.50 pound per foot (ppf) high strength drill pipe with an NC50(XH) connection. A section of the drill tool assembly consisting of a number of these drill pipes will have a repeating cross-sectional pattern, consisting of approximately 30 ft of pipe body with an OD=5″ and ID=4.276″, and a tool joint section with a total (pin+box) length of 21″, OD=6.625″ and ID=2.75″. The corresponding shape factors for this pipe are sA=1.09 and sj=1.11, respectively. Thus, if most of the drill tool assembly length consists of this pipe, the tool joints may cause a downward shift of resonant frequencies of up to about 10%, compared to a drill pipe of uniform cross-section. This can be significant depending on the application, and may be included in an exemplary embodiment of the invention.
Drill Tool Assembly Performance Assessment
The baseline solution, frequency eigenstates, and linear response functions provided by the base model may be used to evaluate downhole vibration index that include but are not limited to bit bounce and stick-slip tendencies of drill tool assembly designs, which may be by means of compliances derived from these results. More specifically, downhole vibration index for the drill tool assembly may include but are not limited to bit disengagement index, ROP limit state index, bit bounce compliance index, bit chatter index, relative bit chatter index, stick-slip tendency index, bit torsional aggressiveness index, forced torsional vibration index, relative forced torsional vibration index, axial strain energy index, torsional strain energy index, and combinations thereof. Without restricting the scope of the claimed invention, a few examples are presented below.
In one exemplary embodiment of the claimed subject matter, a downhole vibration index may be determined by the effective compliance (axial and torsional) of the drill tool assembly:
The axial compliance provides the relationship between the axial displacement and tension amplitude at a particular frequency. Similarly, the torsional compliance relates the angular displacement amplitude to the torque amplitude. The compliance is a complex function of ω and has information on both the relative magnitude and phase of the oscillations.
Compliance functions defined at the bit can be referenced to surface parameter measurements using the bit-to-surface transfer functions described in (93) and (96). In the following discussion, certain relationships are discussed which can thereby be referenced to surface measurements. The indices below are exemplary Vibration Amplitude Ratios which may be translated to the surface using the methods taught above, with corresponding reference values translated to reference values of surface parameters for comparison with surface measurements to obtain the desired real-time downhole vibration indices to improve drilling performance.
Axial Compliance—Bit Bounce
In evaluating the drill tool assembly performance considering forced displacement at the bit, the drill bit is assumed to act as a displacement source at certain harmonics of the rotary speed. For roller cone (RC) bits with three cones, the 3xRPM mode is generally implicated in bit bounce, thus it is appropriate to treat n=3 as the most important harmonic mode. For PDC bits, the number of blades is likely to be an important harmonic node. Also, in a laminated formation, any mismatch between the borehole trajectory and the toolface, such as during directional drilling, will give rise to an excitation at the fundamental frequency, thus n=1 should always be considered. Considering the harmonics, n=3 for RC bits and n=1 and blade count for PDC bits, should be used; however, considering other frequencies are within the scope of this invention.
It is assumed that the origin of the displacement excitation is the heterogeneity in the rock, such as hard nodules or streaks, or transitions between different formations. While passing over these hard streaks, the drill bit is pushed up by the harder formation. If the additional axial force that is generated by the drill tool assembly response to this motion exceeds the WOB, the resulting oscillations in WOB can cause the bit to lose contact with the bottom hole. The situation is similar to the case when a car with a stiff suspension gets airborne after driving over a speed bump. The effective spring constant of the drill tool assembly that generates the restoring force is given by:
The worst-case scenario occurs when the strength of the hard portions significantly exceed the average strength of the rock, such that the bit nearly disengages from its bottom hole pattern, resulting in an excitation amplitude equal to the penetration per cycle (PPC), or the amount the drill tool assembly advances axially in one oscillation period; thus, it is assumed that:
The proportionality constant, a, between the PPC and the imposed displacement amplitude can be adjusted from 0 to 1 to indicate rock heterogeneity, with 0 corresponding to a completely homogeneous rock and 1 corresponding to the presence of very hard stringers in a soft rock. A bit bounce index can then be defined by the ratio of the dynamic axial force to the average WOB. Setting the proportionality constant, a, to one corresponds to a worst-case scenario:
“Bit Disengagement Index”
The bit would completely disengage from the rock for part of the cycle if this ratio exceeds one, so the design goal would be to minimize this index; keeping it small compared to one. The index is only relevant when the real part of the compliance is negative, that is, when the drill tool assembly actually pushes back.
The first ratio in this expression depends on the bit and formation characteristics, and this can be obtained from drill-off tests at the relevant rotational speeds. Alternatively, the vibrational performance of an already-run drill tool assembly design can be hindcast using ROP and WOB data in the drilling log.
In a pre-drill situation where ROPs are not known, it may be more advantageous to provide a pre-drill ROP “limit state” estimate associated with a bit bounce index of one: “ROP Limit State Index”
A contour plot of this quantity will indicate, for a given set of drilling conditions, the ROP beyond which bit bounce may become prevalent and the design goal would be to maximize the ROP within an operating window without inducing excessive or undesirable bit bounce.
For the purposes of drill tool assembly design, a comparative bit bounce index that takes into account only drill tool assembly properties can be useful:
“Bit Bounce Compliance Index”
where Db is the bit diameter. The design goal would be to minimize this quantity in the operating window. It is a relative indicator, in that the actual magnitude does not provide any quantitative information; however, it has units of stress and should be small when compared to the formation strength. Only positive values of this parameter pose a potential axial vibration problem.
For cases where the uncertainty in the input parameters does not allow accurate determination of the phase of the compliance, a more conservative index can be used by replacing the real part with the magnitude and disregarding the phase. The discussion above illustrates several available indices that can be developed from the relationships within the borehole. Other suitable indices may be developed applying the systems and methods of the present disclosure and are within the scope of the present disclosure.
Another important potential source of axial vibration is regenerative chatter of the drill bit, which has a more solid foundational understanding. As a source of axial vibration, relationships defining regenerative chatter behavior can be used to provide still additional performance indices. Regenerative chatter is a self-excited vibration, where the interaction between the dynamic response of the drill tool assembly and the bit-rock interaction can cause a bottom hole pattern whose amplitude grows with time. This is a well-known and studied phenomenon in machining, metal cutting and milling, and is referred to as “chatter theory”. In comparison to the earlier discussion, this type of instability can occur in completely homogeneous rock and is more directly tied to the drill tool assembly design.
Linear theories of regenerative chatter were developed in the 1950's and 1960's by various researchers, including Tobias, Tlusty and Merritt. In the years since the introductory theories of regenerative chatter, significant improvements have been made to the theories including theories that feature predictive capabilities. Chatter can occur at frequencies where the real part of the compliance is positive, thus it covers frequencies complementary to the ones considered previously. The sign convention used in the present systems and methods is different from most conventional descriptions of chatter. For these frequencies, chatter can occur if:
For unconditional stability, this inequality needs to be satisfied for any candidate chatter frequency. The penetration per cycle (PPC) can be related to ROP:
Thus, the criterion for unconditional stability can be made into a chatter index:
“Bit Chatter Index”
This quantity needs to be less than one for unconditional stability. If calibration (drill-off) information is not available, it is still possible to construct a relative chatter index:
“Relative Bit Chatter Index”
In reality, requiring unconditional stability is conservative, since the chatter frequency and rotary speed are related. It is possible to compute a conditional stability diagram and locate RPM “sweet spots” by fully employing Tlusty's theory. This computation is complicated by the fact that the effective bit compliance itself is a function of rotary speed, although the dependence is fairly weak. This results in a more computationally intensive analysis, which is not described in detail herein, but which is within the broader scope of the present disclosure.
Torsional Compliance—Stick-Slip
While torsional vibration, also referred to as stick-slip, can be caused or influenced by a number of factors within the borehole, the interaction between the bit and the formation is an important factor. The prevailing explanation of bit-induced stick-slip is that it arises as an instability due to the dependence of bit aggressiveness (Torque/WOB ratio) on RPM. Most bits exhibit reduced aggressiveness at higher RPMs. At constant WOB, the torque generated by the bit actually decreases as the bit speeds up, resulting in rotary speed fluctuations that grow in time. What prevents this from happening at all times is the dynamic damping of torsional motion along the drill tool assembly. Stick-slip behavior can potentially occur at resonant frequencies of the drill tool assembly, where “inertial” and “elastic” forces exactly cancel each other out. When this occurs, the real part of the compliance vanishes:
Re└Cτ,bit(ωres,i)┘=0; i=1,2, . . . (108)
The magnitude of the effective damping at this frequency is given by:
If one assumes that the dynamical response of the bit can be inferred from its steady-state behavior at varying RPMs, then the damping parameter associated with the bit is given by:
Stick-slip instability occurs when the negative bit damping is large enough to make the overall damping of the system become negative:
γbit+γτj<0. (111)
A drill tool assembly has multiple resonant frequencies, but in most cases, the effective drill tool assembly damping is smallest for the lowest-frequency resonance (i=1), unless vibration at this frequency is suppressed by active control such as Soft Torque™. Thus, the presently-described model locates the first resonance and uses it to assess stick-slip performance. Other suitable models used to develop indices may consider other resonances. A suitable stick-slip tendency index can be constructed as:
“Stick-Slip Tendency Index”
The factor multiplying the overall damping coefficient is chosen to non-dimensionalize the index by means of a characteristic torque (rig torque) and angular displacement (encountered at full stick-slip conditions). Another reasonable choice for a characteristic torque would be torque at the bit; there are also other characteristic frequencies such as the stick-slip frequency. Accordingly, the index presented here is merely exemplary of the methodology within the scope of the present disclosure. Other index formulations may be utilized based on the teachings herein and are within the scope of the present invention. The design goal of a drill tool assembly configuration design and/or a drilling operation design would be to primarily avoid regions where this index is negative, and then to minimize any positive values within the operating window.
This index requires information about how the bit torque depends on rotary speed. The exemplary embodiment uses a functional form for the bit aggressiveness as follows:
where Db is the bit diameter. Other implementations may utilize other relationships to describe how the bit torque depends on rotary speed. According to the present implementation, as the rotary speed is increased, the bit aggressiveness goes down from its “static” value μs at low RPMs towards its “dynamic” value μd at high RPMs, with a characteristic crossover RPM associated with rotary speed ΩXO. Then,
Other suitable functional forms can also be used. It should be noted that if a mud motor is present, the rotation speed at the bit should be used to compute the damping of the bit. Mud motor systems operate at higher rotary speeds and tend to have significant torsional damping due to their architecture. Use of mud motors can significantly reduce stick-slip risk; this effect can be accounted for if the dynamic transfer matrix of the mud motor is provided to the model. Other suitable adaptations of the present models to account for various other drill tool assembly elements and configurations are within the scope of the present disclosure.
If no bit characteristic information is available, a relative index can be used for the purposes of side-by-side comparison of drill tool assembly designs by assuming suitable default values, such as 0.3 for bit aggressiveness and no velocity weakening. This index will not allow determination of when stick-slip will occur, but will provide a relative comparison between different drill tool assembly designs meant for the same bit, with the better designs having a lower index:
“Bit Torsional Aggressiveness Index”
In order to evaluate drill tool assembly performance under torsional forcing, the linear response to various types of excitations can be considered, all of which are within the scope of the disclosed invention. In one exemplary embodiment, the drill bit is assumed to act as a source of torque oscillations with a frequency that matches the rotary speed and its harmonics. When one of these harmonics is close to one of the torsional resonant frequencies of the drill tool assembly, severe torsional oscillations can be induced due to the large effective compliance of the drill tool assembly, i.e., a small torque oscillation can result in a large variation in the rotary speed of the bit. The effective torsional compliance at the bit, taking into consideration drill string and bit damping is given by,
where, Cbit*(ω)=1/jωγbit. The * is used to indicate that the term is not a true compliance and only includes the velocity weakening term associated with the bit aggressiveness. A non-dimensionalized forced torsional vibration index for the nth harmonic excitation can then be defined as:
“Forced Torsional Vibration Index”
TT1(n)=nτrig∥Ceff(nΩRPM)∥. (117)
For the desired range of drilling parameters, better drill tool assembly and bit designs result in lower indices. The index is normalized such that it reflects the ratio of a characteristic torque (chosen here as the torque at the surface) to the excitation torque amplitude needed to achieve full stick-slip at the bit. Another reasonable choice for a characteristic torque would be torque at the bit. There are also other characteristic frequencies that can be considered, another example is disclosed below. Accordingly, the index presented here is merely exemplary of the methodology within the scope of the present disclosure. Other index formulations may be utilized based on the teachings herein and are within the scope of the present invention. The design goal would be to minimize the index within the operating window.
If no bit characteristic information is available, suitable default values such as 0.3 for bit aggressiveness and no velocity weakening can be assumed and a relative index similar to the stick slip index can then be defined as:
“Relative Forced Torsional Vibration Index”
TT2(n)=nτrig(μb=0.3)∥Cτ,bit(nΩRPM)∥. (118)
This index can provide a relative comparison between different drill tool assembly designs utilizing the same bit, with the better design having a lower vibration index.
Elastic Energy in the Drill Tool Assembly
The amount of stored elastic energy in the drill tool assembly resulting from dynamic conditions can be an indicator of excessive motion that can lead to drill tool assembly damage, wear of pipe and casing, and perhaps even borehole breakouts and other poor hole conditions. The amount of stored elastic energy in the drill tool assembly may be written in integral form as:
Since the hole curvature can be considered to be pre-determined and not part of the dynamics problem, the first two terms in the integrand, the dynamic axial strain energy and torsional strain energy respectively, may be used as, or considered in, additional vibration indices. Better performance would generally be associated with lower index values calculated as follows:
“Axial Strain Energy Index”
“Torsional Strain Energy Index”
The particular solutions used in computing the indices above can be the baseline solution, the dynamic part of the linear response functions at a relevant frequency (a harmonic of the rotary speed, or a resonant frequency in the case of chatter or stick-slip), or a superposition of the two.
Examples
During the drilling of a well, a downhole vibration sensor sub was used to collect rotary speed at the bit at a rate of 50 samples per second, allowing direct determination of torsional severity. Data from the rig acquisition system was also collected at a rate of one sample per second.
The bottom panel 730 reflects a comparison of torsional severity (here reported as the ratio of rotary speed fluctuation amplitude and average rotary speed, in percent) obtained directly from the downhole data and estimated from the rig torque signal using the method disclosed herein. The two curves track each other very well, except when the top drive RPM is changed to a new value, which is expected. Also illustrated are the ROP and MSE data displayed on the rig during this interval. The ROP and MSE signals do not correlate well with the torsional severity. One exception is the interval around 3700 seconds where high values of both MSE and torsional severity are seen. Further analysis of the downhole data suggests coexisting stick-slip and whirl in this interval. Thus, the ability to monitor both MSE and TSE1 (or TSEu) can provide more insight on downhole behavior, compared to either signal on its own.
In another example,
The dTorque and reference dTorque values may be combined to obtain the TSE. The results may be displayed such as in the set of graphs 800 illustrated in
For the dataset of
For this same dataset used in
The torsional vibration severity was also estimated using a simple model that considers only the length and static torsional stiffness of the drill sting component in the drill tool assembly. This model does not consider certain drill string physics that are present and as such provides a less reliable determination of TSE than the methods such as disclosed herein. The results of this analysis are illustrated in
An additional utility of these methods may be observed with reference to
Consider a drilling program in which many wells may be drilled, and optimization of the drilling process is a primary factor in the development engineering. With a specified bit design and one set of operating parameters, Well A is drilled and the drilling data is recorded. From this data, the torsional severity estimate is calculated with an accurate physical drill string model. These results are displayed in
The calculations of TSE may provide some indication of the relative amounts of stick-slip that were present in the drilling operations of each of Well A and Well B. This valuable information can be used in a continuous optimization process, or “relentless re-engineering” effort to combine this information with other data such as: average ROP, bit dull characteristics, Mechanical Specific Energy (MSE), number of bit runs required to drill the section, and other vibration and drilling performance indicators known to those skilled in the art.
Estimated downhole vibration indices (e.g., torsional, axial, etc.), when divided by their associated reference downhole vibration indices, provide numerical estimates of how close the drilling operation is to the reference state. However, as discussed above, it may be more convenient to provide instead an alarm level associated with the estimates. For instance, low levels of estimated vibration indices may yield a green light, high levels a red light, and intermediate levels a yellow light.
Such a discrete classification scheme may be validated using downhole data with a table similar to that in
Various combinations of the values in this table may be made in order to measure the quality of the estimated classification. Several of these are shown in
The method disclosed herein teaches and enables drilling operations performance engineering methods that were previously not available using previously available methods that relied only upon surface data measurements to estimate or project downhole responses. The presently claimed methodology provides enabling tools and technology to optimize the wellbore drilling process.
While the present techniques of the invention may be susceptible to various modifications and alternative forms, the exemplary embodiments discussed above have been illustrated by way of example. However, it should again be understood that the invention is not intended to be limited to the particular embodiments disclosed herein. Illustrative, non-exclusive, examples of descriptions of some systems and methods within the scope of the present disclosure are presented in the following numbered paragraphs. The preceding paragraphs are not intended to be an exhaustive set of descriptions, and are not intended to define minimum or maximum scopes or required elements of the present disclosure. Instead, they are provided as illustrative examples, with other descriptions of broader or narrower scopes still being within the scope of the present disclosure. Indeed, the present techniques of the invention are to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the description provided herein.
This application is the National Stage of International Application No. PCT/US10/44735, filed Aug. 6, 2010, which is related to and claims benefit of U.S. Provisional Application No. 61/232,275, filed Aug. 7, 2009, and related U.S. Provisional Application No. 61/232,274 filed Aug. 7, 2009, the entirety of each incorporated herein by reference. This application is also related to and claims benefit of U.S. Provisional Application No. 61/364,247, filed Jul. 14, 2010, the entirety of which is incorporated herein by reference.
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/US2010/044735 | 8/6/2010 | WO | 00 | 1/24/2012 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2011/017627 | 2/10/2011 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
4507735 | Moorehead et al. | Mar 1985 | A |
4736297 | LeJeune | Apr 1988 | A |
4903245 | Close et al. | Feb 1990 | A |
5551286 | Booer | Sep 1996 | A |
5730234 | Putot | Mar 1998 | A |
5842149 | Harrell et al. | Nov 1998 | A |
5852235 | Pavone et al. | Dec 1998 | A |
5881310 | Airhart et al. | Mar 1999 | A |
6026912 | King et al. | Feb 2000 | A |
6155357 | King et al. | Dec 2000 | A |
6192998 | Pinckard | Feb 2001 | B1 |
6206108 | MacDonald et al. | Mar 2001 | B1 |
6233524 | Harrell et al. | May 2001 | B1 |
6293356 | King et al. | Sep 2001 | B1 |
6363780 | Rey-Fabret et al. | Apr 2002 | B1 |
6382331 | Pinckard | May 2002 | B1 |
6408953 | Goldman et al. | Jun 2002 | B1 |
6424919 | Moran et al. | Jul 2002 | B1 |
6443242 | Newman et al. | Sep 2002 | B1 |
6480118 | Rao | Nov 2002 | B1 |
6662110 | Bargach et al. | Dec 2003 | B1 |
6732052 | Macdonald et al. | May 2004 | B2 |
6820702 | Niedermayr et al. | Nov 2004 | B2 |
6892812 | Niedermayr et al. | May 2005 | B2 |
6968909 | Aldred et al. | Nov 2005 | B2 |
7020597 | Oliver et al. | Mar 2006 | B2 |
7044238 | Hutchinson | May 2006 | B2 |
7044239 | Pinckard et al. | May 2006 | B2 |
7142986 | Moran | Nov 2006 | B2 |
7172037 | Dashevskiy et al. | Feb 2007 | B2 |
7261167 | Goldman et al. | Aug 2007 | B2 |
7316278 | Hutchinson | Jan 2008 | B2 |
7357196 | Goldman et al. | Apr 2008 | B2 |
7363988 | Jeffryes | Apr 2008 | B2 |
7606666 | Repin et al. | Oct 2009 | B2 |
7610251 | Shayegi et al. | Oct 2009 | B2 |
7645124 | Garlow | Jan 2010 | B2 |
7857047 | Remmert et al. | Dec 2010 | B2 |
7878268 | Chapman et al. | Feb 2011 | B2 |
7896105 | Dupriest | Mar 2011 | B2 |
8014987 | Pabon et al. | Sep 2011 | B2 |
20020104685 | Pinckard et al. | Aug 2002 | A1 |
20020148648 | Schultz et al. | Oct 2002 | A1 |
20030147360 | Nero et al. | Aug 2003 | A1 |
20050096847 | Huang | May 2005 | A1 |
20050197777 | Rodney et al. | Sep 2005 | A1 |
20070289373 | Sugiura | Dec 2007 | A1 |
20080105424 | Remmert et al. | May 2008 | A1 |
20080255817 | Pabon et al. | Oct 2008 | A1 |
20090076873 | Johnson et al. | Mar 2009 | A1 |
20090089227 | Sturrock et al. | Apr 2009 | A1 |
20090090555 | Boone et al. | Apr 2009 | A1 |
20090132458 | Edwards et al. | May 2009 | A1 |
20100108384 | Byreddy et al. | May 2010 | A1 |
20110077924 | Ertas et al. | Mar 2011 | A1 |
20110245980 | Nessjoen et al. | Oct 2011 | A1 |
Number | Date | Country | |
---|---|---|---|
20120123757 A1 | May 2012 | US |
Number | Date | Country | |
---|---|---|---|
61232275 | Aug 2009 | US | |
61232274 | Aug 2009 | US | |
61364247 | Jul 2010 | US |