Patent Application
20060196621
Embodiments of the invention relate generally to methods and systems for automated measurement of pulp and sheet properties. More specifically, embodiments of the invention relate to methods and systems for determining pulp and sheet properties.
The manufacture of paper involves many variables. Wood itself is not a uniform raw material, and grinding of wood yields a heterogeneous pulp, which can include a liquid slurry and fibers (e.g., wood chips, wood fibers, cotton, cloth, and the like) suspended in the slurry. In addition, the process of refining pulp may destroy a portion of the fibers therein. Because of these and other variables, pulp is typically characterized before significant quantities of paper are produced. Such characterization may involve determining drainage and/or strength properties, which can be used to predict properties of paper to be manufactured from the pulp. When characterization indicates that the pulp is unsuitable-for instance, the fibers are not sufficiently long and slender-the pulp may be further refined.
Existing offline measurement technologies typically require significant amounts of time in order to adequately characterize a pulp for production purposes (e.g., 8 to 12 hours). The reason is that a physical sheet of paper, termed a “hand sheet,” must be manually made from pulp samples before actual testing can begin. Typically, a predetermined weight of fibers (e.g., 60 g/m2) is needed to make a hand sheet. Once made, the hand sheet is tested to determine such properties as tensile index, tear burst, or the like. Robotic machines can reduce the time to make and test hand sheets to between approximately 20 and 40 minutes.
With existing online measurement technologies, which interface with a production line of a mill, only basic information on fiber morphological characteristics is provided (e.g., freeness, fiber coarseness, fiber length, fiber distribution), and at relatively low measurement frequencies (e.g., 1 to 6 samples per hour). Moreover, such technologies involve costly measurement units, often on the order of hundreds of thousands of dollars. As such, the cost is often prohibitive, with relatively few process points being sampled in a mill.
The following summary sets forth certain exemplary embodiments of the invention described in greater detail below. It does not set forth all such embodiments and should in no way be construed as limiting of the invention.
Embodiments of the invention relate to methods, systems, and devices for estimating pulp and sheet properties.
In one embodiment, a method of estimating a property of paper comprises receiving image data depicting a plurality of pulp fibers, virtually distributing the pulp fibers over a predetermined area to create a virtual distribution of the pulp fibers, and determining the property using the virtual distribution.
In one embodiment, a system for pulp fiber analysis comprises an image capture device and an analyzer. The image capture device is configured to send image data depicting a plurality of pulp fibers. The analyzer is configured to receive the image data; create a virtual distribution of the pulp fibers depicted in the received image data, over a predetermined area; and estimate a property of paper by analyzing the virtual distribution.
Other embodiments herein relate to pulp analyzers, image capture devices, and graphical user interfaces.
Before any embodiments of the invention are explained in detail, it is to be understood that embodiments of the invention are not limited in application to the details of operation and construction, the arrangement of components, and the pseudocode set forth in the following description or illustrated in the following drawings. Embodiments are capable of being practiced or of being carried out in various ways. Also, it is to be understood that the phraseology and terminology used herein are for the purpose of description and should not be regarded as limiting. The use of “including,” “comprising,” or “having” and variations thereof herein is meant to encompass the items listed thereafter and equivalents thereof as well as additional items. The order of limitations specified in any method claims does not imply that the steps or acts set forth therein must be performed in that order, unless an order is explicitly identified in the specification as essential. In addition, it should be noted that a plurality of hardware and software based devices, as well as a plurality of different structural components, may be utilized to implement embodiments of the invention.
Embodiments of the invention provide methods, systems, and devices for measuring fiber morphological characteristics. In one embodiment, images of sampled pulp fibers are processed to create an artificial computer model of a distribution of fibers. This virtual hand sheet is analyzed to calculate pulp properties, such as freeness, drainage time, density, and strength properties. Embodiments herein bypass the time-consuming, costly phases of physical hand sheet preparation, conditioning, and testing typically employed in the art. Embodiments herein provide results at frequencies that are significantly higher than those of existing technologies (e.g., 5 to 20 times higher).
In particular, embodiments herein can provide drainage and strength properties in minutes instead of hours. Moreover, instead of implied properties, sheet properties can be determined. Embodiments can be used, for example, to optimize quality of paper or medium-density fiberboard (MDF) produced by a mill, or to selectively control components in a mill.
Although embodiments herein generally focus on pulp properties in the paper industry, embodiments can be applied to substances that have similarities to pulp, such as textile and food constituents. For instance, embodiments can be applied in settings in which particles have a dimension of interest of approximately 2 to 5 micrometers.
The modules of the pulp analyzer 200 can be implemented in various combinations of software and/or hardware. In one embodiment, the pulp analyzer 200 is implemented as a computer running one or more software applications. For instance, the pulp analyzer 200 may be implemented on a computer having 1 GB of random access memory (RAM) and running the Windows XP operating system. To improve performance, the computer may include multiple processors. The pulp analyzer 200 optionally may have connectivity to one or more networks (e.g., a Megabit network and/or Gigabit Ethernet network). In one embodiment, the pulp analyzer 200 utilizes the MATLAB tool of The MathWorks, Inc. (Natick, Mass.) to perform certain image analysis operations described below. In another embodiment, the pulp analyzer is programmed in C and utilizes the library of the Lua programming language (designed by Tecgraf-Computer Graphics Technology Group, Rio de Janeiro, Brazil).
The image data processor 220 receives image data 210 depicting pulp fibers. For example, the image data 210 may be digital data representing a series of images that depict pulp fibers. The image data processor 220 processes the image data 210 to produce descriptive information for the respective pulp fibers, such as image representations and/or geometric parameters. In some embodiments, the image data processor 220 may be omitted from the pulp analyzer 200, such as embodiments wherein the pulp analyzer 200 receives data suitably preprocessed for the virtual distributor 230. The image data 210 may be received, for example, in real time (e.g., from an image capture device over a network) or from a storage device.
The virtual distributor 230 uses some or all the descriptive information to create a virtual distribution of the pulp fibers. In one embodiment, the virtual distributor 230 virtually distributes the pulp fibers over a predetermined area.
The virtual distribution analyzer 240 analyzes the virtual distribution to determine a property 250. For instance, the virtual distribution analyzer 240 may determine a drainage property or a strength property (e.g., tensile index or tear index).
The pulp analyzer 200 optionally includes a controller or controller interface (not shown), which sends information that controls, or can be used by a controller to control, one or more components of a paper production system, such as, for example, a refiner, a screen, or a paper machine. The paper production system may include a distributed control system (DCS) that interfaces with such components.
Additional examples of techniques that can be employed within or in connection with the pulp analyzer 200 are set forth below.
Task T310 receives image data depicting pulp fibers. In one embodiment, the image data is digitally received as a series of RGB images taken by camera(s) remote from the pulp analyzer 200. Each image frame may include, for instance, 10 to 15 fibers. In other embodiments, grayscale or CMYK images are received.
Task T320 processes the image data to produce descriptive information for the respective pulp fibers. The descriptive information can include, for example, image representations and/or geometric parameters for the fibers. The descriptive information can be used immediately and/or stored in a memory. Task T320 may be omitted in some embodiments, such as embodiments wherein a pulp analyzer receives data that has been suitably preprocessed for task T330 (described below).
In particular, thresholding techniques can be applied to the received images to produce binary images, wherein pixels of value 1 correspond to fibers, and pixels of value 0 correspond to a background (or vice versa). For example, thresholding can be applied to one extracted color plane of a received RGB image. Each binary image can be further processed to identify respective fibers in that image. For each identified fiber, a binary image can be produced and stored as a numerical array.
By applying image analysis techniques to the binary images, geometric parameters for the fibers can be determined and stored in a matrix. Exemplary determined geometric parameters include:
In one embodiment, task T320 is performed until it is determined that no additional useful information is gained from processing of additional images. Statistical analysis techniques (e.g., involving standard deviation) may be applied to help make this determination. For instance, when the standard error with respect to one or more geometric parameters is sufficiently small, no additional fibers need to be considered. In one embodiment, a user can specify a standard error (or range) that is considered acceptable. Accordingly, superfluous analysis of data is avoided, and properties can be determined more rapidly.
Task T330 virtually distributes the pulp fibers over a predetermined area to create a virtual distribution of the pulp fibers. The virtual distribution may be a computerized representation analogous to a physical hand sheet.
In one embodiment, the binary images of each fiber are randomly distributed over a field of known radius, wherein the position of each fiber is randomly oriented. The binary images can be randomly distributed as the images are produced. Accordingly, assuming 10 to 15 fibers per received image frame, the virtual distribution gradually grows by groups of 10 to 15 fibers. In other embodiments, the virtual distribution is not created until all fibers of interest are identified and associated image representations are stored in memory. In an exemplary embodiment, the amount of fibers virtually distributed is roughly equivalent to about 3 grams of actual fibers. Although embodiments described herein involve a two-dimensional field, in alternative embodiments, pulp fibers can be virtually distributed in a three-dimensional space. Moreover, the area over which the fibers are distributed need not be circular or otherwise regular in shape, extent, or volume.
Task T340 determines a property using the virtual distribution. In one embodiment, drainage properties, such as flow, freeness (e.g., Canadian Standard Freeness, CSF), and/or drainage time, can be determined by modeling drainage of water through the set of orifices in the virtual distribution. Each of the orifices contributes to the drainage of the virtual distribution by leaking water. Accordingly, drainage properties can be determined by summing up the flow through the orifices over time.
More specifically, an exemplary model can involve measuring the drainage of water in a water column sitting atop the virtual distribution.
In one embodiment, the model involves, at each instant of time, determining the total flow through the orifices of the virtual distribution, and integrating the flow over the drainage time to determine freeness. The drainage time may be defined as ranging from an initial time to a time when a minimum velocity has been reached.
Below is pseudocode setting forth basic elements of an exemplary procedure for determining drainage properties. The pseudocode is not intended to be exhaustive, and variations of this example are within the scope of embodiments of the invention.
In one embodiment, drainage properties are determined by dividing the virtual distribution into subregions, considering the flow in those subregions, and ignoring the flow of subregions that is statistically far different from that of other regions. Such an approach can prevent subregions that have unusually large orifices, and therefore, measurably higher flow, from dominating the determination of drainage properties. In an exemplary implementation, a virtual distribution is divided into 30 subregions for purposes of determining drainage properties. Statistical methods (e.g., standard deviation) may be used to identify subregions to ignore.
In addition to drainage properties, strength properties can be determined using a virtual distribution. In an exemplary implementation, two-dimensional finite element analysis (FEA) methods are applied. Each fiber is modeled as one or more beam elements having an associated Young's modulus E, moment of inertia I, and cross-sectional area A, which can be determined using image analysis techniques, as described above. More specifically, binary images of each fiber (see task T320 of
After a mesh representation of each fiber is created, and the associated length and width are determined (e.g., by using the geometric parameters that can be determined in task T320 of
FEA techniques can be applied to the generated mesh of elements to solve for deflections and stresses in the virtual distribution according to the equation KU=F, where K is the stiffness matrix, U is the displacement, and F is the applied force. The nodes are considered to be fixed. Different loads (forces and/or displacements) are applied, and the mechanical behavior of the sheet is used to compute tensile stress, bonding, and the like. In one embodiment, a matrix is created that includes, for example, the elements, properties of the elements, and boundary conditions. Reaction forces and stresses can be determined using the matrix. In one embodiment, the mesh of elements is cut into smaller sections, which are analyzed individually as test specimens.
Once the reaction forces, stresses, and displacements have been determined for the virtual distribution, the data can be calibrated against a set of experimentally collected data, such as measured data corresponding to physical hand sheets created from pulp samples. For instance, relationships for tensile index, tear burst, and the like can be identified.
In one embodiment, one or more of the following exemplary strength properties are determined for the virtual distribution or portions thereof: tear index (mN-m2/g), burst index (kPa-m2/g), sheet density (kg/m3), air resistance (s/100 mL), tensile index (Nm/g), stretch (%), tensile energy index (J/kg), and tensile stiffness (kNm/g).
In one embodiment, a pulp analyzer provides a graphical user interface (GUI) that presents useful information for a user and/or receives user inputs.
The image capture device 1410 sends image data 1440 depicting pulp fibers over the network 1430. In one embodiment, the image capture device 1410 interfaces with a source of pulp to be analyzed and captures images of the pulp.
The pulp analyzer 1420 receives the image data 1440 from the image capture device 1410. Using the image data 1440, the pulp analyzer 1420 creates a virtual distribution 1450 of the pulp fibers. For instance, the pulp analyzer 1420 may employ approaches described above, or other suitable approaches. The pulp analyzer 1420 determines at least one property 1460, such as a drainage or strength property, by analyzing the virtual distribution 1450. The pulp analyzer 1420 optionally sends data 1470 to the image capture device 1410, which data may be used, for example, to calibrate the image capture device 1410, request data from the image capture device 1410, or direct the image capture device 1410 to perform an operation.
The pulp analyzer 1420 optionally generates control information 1480 that controls, or can be used by a controller to control, one or more components of a paper production system, such as, for example, a refiner, a screen, or a paper machine. For instance, control information 1480 can be sent to a central computer of a mill to control (1) the energy input, plate gaps, and/or feed of a refiner; (2) the rotor speed and/or delta p (i.e., the difference between infeed and outlet pressure) of a screen (to achieve desired fractionation or freeness); or (3) mixtures of different pulp streams to, and feed consistency of, a paper machine.
The image capture devices 1410 are placed at various process points of a mill 1530. In particular, the image capture devices 1410 are appropriately interfaced at those points in order to sample pulp, capture images of the pulp, and send captured images to the pulp analyzer 1420 over the network 1430. In one embodiment, each image capture device 1410 has an IP address, and sends packetized data to the pulp analyzer 1410, including the IP address and image data. In an exemplary implementation, up to ten image capture devices 1410 send image data to the pulp analyzer 1420. In other embodiments, data from multiple image capture devices 1410 is multiplexed to the pulp analyzer 1420.
Using the image data received from each image capture device 1410, the pulp analyzer 1420 creates a respective virtual distribution and determines properties as described above. In one embodiment, the pulp analyzer 1420 employs an event-driven architecture, wherein the pulp analyzer 1420 is responsive to the occurrence of events, such as the receipt of new image data. Because each image capture device 1410 in the system 1500 can have an identity (e.g., a unique IP address), the pulp analyzer 1420 can identify image data received from respective image capture devices 1410. As such, the pulp analyzer 1420 can in parallel create separate virtual distributions for pulp sampled by respective image capture devices 1410. Multithreading may be used to facilitate the necessary separate processing. Alternatively or additionally, parallel processors can be incorporated in a pulp analyzer 1420. It is to be appreciated that multiple pulp analyzers 1420 can be implemented in a particular setting, and can create virtual distributions for a subset of the image capture devices 1410 in that setting. Further, a pulp analyzer 1420 can be remote from, or proximate to, an image capture device 1410. If implemented on a portable computer (e.g., laptop), the pulp analyzer 1420 can be physically moved.
Various benefits can be realized using embodiments herein. For example, by incorporating image capture devices 1410 at multiple process points of a mill, the pulp sampling frequency is greatly increased, enabling both short- and long-term variations in pulp properties to be detected and responded to. In addition, because a greater body of information is available, problems can be pinpointed with greater accuracy, and control decisions can be more readily made, simplifying associated control systems. Moreover, the architecture is less sensitive to malfunctions; if one image capture device 1410 goes offline, many process points remain available for sampling.
In another embodiment, modules of the pulp analyzer 1420 are written in a scripting language to provide an open source architecture. Accordingly, the behavior and/or rules of the pulp analyzer 1420 can be customized to meet the particular needs of a user or implementation context. For instance, the pulp analyzer 1420 can be customized such that tear index is determined for samples from one process point of a mill, while tensile index is determined for samples from another process point. Moreover, scripts can be written to tailor operation of the control system of the mill. Embodiments herein enhance the flexibility of the control system, as well as reduce maintenance costs in a mill.
The following exemplary disc gap controller rule may be implemented in the pulp analyzer 1420. The rule has logic that controls the disc gap depending on, among other things, freeness calculations made by the pulp analyzer 1420. More particularly, in the rule below, Chxx[t] is a time-based array used to store values; t=0 is the most recent time, t=1 is the time one second ago, and so on. Ch 82[0] is the most recent freeness (CSF) value. Ch83 is an internal timer; no control occurs until 90 seconds have elapsed (i.e., Ch83=90). Ch 26 is the consistency. Depending on the consistency and freeness values, freeness can be controlled by making adjustments to one or both disc gaps (aveCZ and aveFZ) by respectively adjusting the distributed control system parameters Ch84 and/or Ch85. In the rule, control is not allowed to occur unless, for example, the gaps are in range, production (e.g., the consist controller) is on, and/or vibration is low. Permit is a flag that, when set (Permit=1), indicates that these conditions (interlocks) are met and that the rule can be run.
Image capture devices can be implemented in various ways.
The image capture device 1700 includes a circulation pump 1705, a dilution solenoid valve 1710, a camera interface module 1715, a light 1720, a camera 1725, a sample solenoid valve 1730, a drain solenoid valve 1735, a pump motor drive 1740, a network module 1745, an input/output module 1750, and a power supply 1755. Variations of the configuration of the image capture device 1700 are within the scope of embodiments of the invention. For instance, in an alternative embodiment, one or more of the components of the image capture device 1700 can be external to, and interfaced with, the image capture device 1700.
The image capture device 1700 interfaces with a source of pulp that is to be sampled, such as pulp transported in a supply pipe or hose. The pump 1705, pump motor drive 1740 (e.g., a variable speed drive), dilution solenoid valve 1710, sample solenoid valve 1730, and drain solenoid valve 1735 together may form a fluid subsystem within the image capture device 1700. This fluid subsystem interfaces with the source of pulp and circulates pulp to be sampled so that images thereof can be captured.
In one embodiment, the sampled pulp passes through a glass cell (not shown) about 10 to 30 mm wide and 0.5 to 3 mm between the walls thereof. As the pulp passes through the cell, the light 1720 (e.g., an LED) illuminates the fibers therein, creating a large contrast between the fibers and the associated background. The camera 1725 captures images of the fibers, which appear as darker regions in the captured images. In an exemplary implementation, the camera 1725 is a high-resolution CMOS or CCD camera compatible with the Camera Link connectivity standard for vision applications, which defines a communications interface. The camera 1725 may capture, for instance, between 10 to 100 frames per second. An exemplary image resolution may be 1200 by 1000 pixels or larger (e.g., 10,000 by 10,000 pixels), with a pixel size of 3 to 10 micrometers.
The camera interface module 1715 provides an interface between the camera 1725 and, for example, a network. In one embodiment, the camera interface module 1715 converts captured image data to packetized data, such as TCP/IP data. In one embodiment, the camera interface module 1715 can directly connect to a network. The camera interface module 1715 can be implemented with the iPORT PT1000-CL IP Engine offered by Pleora Technologies Inc. (Kanata, Ontario, Canada) or with other suitable hardware and/or software. Use of a camera interface module 1715 and associated software may allow compatible cameras to be interchangeably incorporated in the image capture device 1700. In particular, if a higher performance camera becomes available, that camera can replace the existing camera in the image capture device 1700.
The network module 1745 enables network communication between the image capture device 1700 and device(s) on the network. In one embodiment, the network module 1745 connects to a Megabit network. Control information, such as information to control the camera 1725 and/or the pump motor drive 1740, can be sent from a pulp analyzer to the network module 1745 over the Megabit network, while image data can be sent by the camera interface module 1715 to the pulp analyzer over a second (e.g., Gigabit) network. In such an embodiment, the network module 1745 and the camera interface module 1715 each may have a respective IP address. In other embodiments, a single network transports both control and image data transmitted between an image capture device and a pulp analyzer. In such an embodiment, fewer network-related modules may be implemented in the image capture device.
The input/output module 1750 may include one or more analog and/or digital input/output modules. The modules can be used to control, for example, the on/off state of the solenoid valves 1710, 1730, 1735; the on/off state of the pump 1705; the on/off state of the light 1720; and level indications from a sample reservoir (e.g., if the image capture device 1700 is operating in a laboratory setup). The modules may also provide an analog signal for the controlled speed of the pump 1705 and an analog signal for the actual speed of the pump 1705, for example. Depending on the particular implementation, some digital switches may be used to select other sample points if sample lines of the image capture device 1700 are multiplexed to multiple different sample points (e.g., inlet and outlet of a screen or a refiner).
The housing 1810 receives the sampling module 1820 and electronics module 1830, which in one embodiment are separate self-contained modules. In the illustrated embodiment, the sampling module 1820 and electronics module 1830 are contained in separate drawers that are receivable by the housing 1810.
In one embodiment, the sampling module 1820 includes certain fluid-related components of the image capture device 1800, and the electronics module 1830 includes certain electronics-based components of the image capture device 1800. For instance, the sampling module 1820 can include a pump, light, camera, camera interface, and solenoid valves. The electronics module 1830 can include a pump motor drive, input/output modules, power supply, and network modules. It can be appreciated that, by segregating certain fluid-related and electronics components into separate compartments, replacement of modules can more readily and cost effectively occur in the event of a malfunction. Moreover, the need to repair or replace the entire image capture device 1800 is reduced.
An image capture device can be used in a laboratory setting.
The embodiments described above are merely exemplary embodiments, and other embodiments can be practiced that fall within the scope of embodiments of the invention. For instance, three-dimensional techniques may be used in connection with processing of image data, creation of a virtual distribution, and/or analysis of a virtual distribution. In addition, finite difference or other appropriate analytical approaches may be used in lieu of, or in addition to, finite element analytical approaches such as described herein. Further, image capture devices described herein can be outfitted with additional functionality, such as freeness testers, if desired. It can also be appreciated that embodiments herein can be implemented in conjunction with other technologies used to improve processes in a mill setting.
As should also be apparent to one of ordinary skill in the art, the systems shown in the figures are models of what actual systems can be like. As noted, many of the modules and logical structures described are capable of being implemented in software executed by a microprocessor or a similar device or of being implemented in hardware using a variety of components including, for example, ASICs. Terms like “processor” may include or refer to both hardware and/or software. Furthermore, throughout the specification, where capitalized terms are used, such terms are used to conform to common practices and to help correlate the description with the examples and drawings. However, no specific meaning is implied or should be inferred simply due to the use of capitalization. Thus, the claims should not be limited to the specific examples or terminology or to any specific hardware or software implementation or combination of software or hardware.
Various features and advantages of the invention are set forth in the following claims.
