Model-constrained multi-phase virtual flow metering and forecasting with machine learning

Abstract

A computer-implemented method for constrained multi-phase virtual flow metering and forecasting is described. The method includes predicting instantaneous flow rates and forecasting future target flow rates and well dynamics. The method includes constructing a virtual sensing model trained using forecasted target flow rates and well dynamics. The method includes building a constrained forecasting model by combining unconstrained flow forecasting models, well dynamics models, and virtual sensing models, wherein the constrained forecasting model forecasts multi-phase flow rates.

Description

TECHNICAL FIELD

This disclosure relates generally to forecasting flow rates, and more particularly to forecasting multi-phase flow rates with machine learning.

BACKGROUND

Generally, a multi-phase flow includes oil, gas, and water flow and is frequently encountered in the production of oil and gas. Accurate multi-phase flow rate measurement plays a significant role in production optimization from oil and gas fields, especially in an offshore environment. In addition, production forecasts can provide important information to significantly influence field development planning and economic evaluation, including production optimization, rate allocation, reservoir management and predict the future performance of the field.

SUMMARY

An embodiment described herein provides a computer-implemented method for model constrained multi-phase virtual flow metering and forecasting. The method includes forecasting future target flow rates using unconstrained flow models trained using current and historic multi-phase flow rates. The method also includes forecasting auxiliary sensor data using trained well dynamics models. Further, the method includes constructing virtual sensing models trained using predicted target flow rates and well dynamics, wherein the virtual sensing model predicts real-time multi-phase flow rates. The method includes building a constrained forecasting model by combining the unconstrained flow forecasting models, well dynamics models, and virtual sensing models, wherein the constrained forecasting model forecasts multi-phase flow rates.

An embodiment described herein provides an apparatus comprising a non-transitory, computer readable, storage medium that stores instructions that, when executed by at least one processor, cause the at least one processor to perform operations. The operations include forecasting future target flow rates using unconstrained flow models trained using current and historic multi-phase flow rates. The operations also include forecasting auxiliary sensor data using trained well dynamics models. Further, the operations include constructing virtual sensing models trained using predicted target flow rates and well dynamics, wherein the virtual sensing model predicts real-time multi-phase flow rates. The operations include building a constrained forecasting model by combining the unconstrained flow forecasting models, well dynamics models, and virtual sensing models, wherein the constrained forecasting model forecasts multi-phase flow rates.

An embodiment described herein provides a system comprising one or more memory modules and one or more hardware processors communicably coupled to the one or more memory modules. The one or more hardware processors are configured to execute instructions stored on the one or more memory models to perform operations. The operations include forecasting future target flow rates using unconstrained flow models trained using current and historic multi-phase flow rates. The operations also include forecasting auxiliary sensor data using trained well dynamics models. Further, the operations include constructing virtual sensing models trained using predicted target flow rates and well dynamics, wherein the virtual sensing model predicts real-time multi-phase flow rates. The operations include building a constrained forecasting model by combining the unconstrained flow forecasting models, well dynamics models, and virtual sensing models, wherein the constrained forecasting model forecasts multi-phase flow rates.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1 is an illustration of a workflow that enables model-constrained multi-phase virtual flow metering and forecasting with a machine learning approach.

FIG. 2 is a block diagram illustrating an unconstrained flow forecasting model.

FIG. 3 is an illustration of flow rate production over time.

FIG. 4 is a block diagram illustrating a well dynamics forecasting model.

FIG. 5 is a block diagram illustrating a virtual flow metering model.

FIG. 6 is an illustration of production flow rate ground truth label and model predictions made using a virtual flow metering model.

FIG. 7 is a block diagram illustrating model-constrained flow forecasting.

FIG. 8 is an illustration of model-constrained flow forecasting over time.

FIG. 9 is a process flow diagram of a process for model-constrained multi-phase virtual flow metering and forecasting with a machine learning approach.

FIG. 10 is a schematic illustration of an example controller (or control system) for model-constrained multi-phase virtual flow metering and forecasting with machine learning according to the present disclosure.

DETAILED DESCRIPTION

Embodiments described herein enable model-constrained multi-phase virtual flow metering and forecasting with machine learning approach. In particular, the present techniques estimate flow rates (for example, oil/water/gas flow rates) and forecast production rates using flow models combined with well dynamics models. The virtual flow metering according to the present techniques incorporates the internal changes in various formation and fluid factors as well as external influences due to manual operations. Internal changes in formation and fluid factors generally refers to the internal characteristics of each well, such as geological heterogeneity of the reservoir, well structure, boundary conditions, fluid properties, and the like. The internal characteristics of the wells vary from well to well, but are not directly impacted by manual operations. The external influences refer to the operation management events and manual operations, such as gas injection, choke valve, and the like. Traditional models generally do not provide combined multi-phase flow estimation and forecasting in a model-constrained manner, using machine learning.

FIG. 1 is an illustration of a workflow 100 that enables model-constrained multi-phase virtual flow metering and forecasting with machine learning. Generally, the models and algorithms according to the present techniques enable predicting and forecasting multi-phase flow rates to assist production planning and optimization. The system consists of flow models 102, well dynamics models 104, virtual sensing models 106, and flow forecasting 108. The flow models 102, well dynamics models 104, virtual sensing models 106, and flow forecasting 108 are individually executable for each of their respective purposes. In embodiments, the flow models 102, well dynamics model 104, virtual sensing model 106, and flow forecasting 108 are combined as an integrated artificial intelligence based virtual flow metering system. Flow models 102 include unconstrained forecasting of flow rates. The flow models 102 directly forecast multi-phase flow rates and build models from historical flow rates. The well dynamics model 104 forecasts well dynamics directly from auxiliary sensor data. The well dynamics model 104 is built using historical sensor data. The virtual sensing model 106 enables virtual flow metering. The virtual sensing model 106 predicts instantaneous multi-phase flow rates. In some examples, the predicted instantaneous multi-phase flow rates are at a current time. In some examples, the virtual sensing model can predict flow rates from the sensor measurements forecasting at a future time with the forecasting of well dynamics model. Models are built from historical and/or forecast previous time data. Forecast previous time data includes estimates associated with a previous time. Generally, flow forecasting 108 enables model-constrained flow forecasting. In particular, multi-phase flow rates are forecast with constraints from models built using historical data or forecast previous time data. In embodiments, virtual flow metering models are inferred using flow models 102 and well dynamics model 104 as an input to a virtual sensing model 106.

In embodiments, the flow models 102, well dynamics model 104, virtual sensing model 106, and flow forecasting 108 of FIG. 1 are executed individually for each of the respective purposes. Alternatively, or additionally, the flow models 102, well dynamics model 104, virtual sensing model 106, and flow forecasting 108 operate together as an integrated artificial intelligence (AI) based virtual flow metering system. The models used are selected according to a particular use case. In a first example, in resource-limited or low data availability/quality scenarios a flow rate forecasting application based on flow models 102 may be more feasible to business stakeholder when compared to full virtual flow metering from the flow models 102, well dynamics model 104, virtual sensing model 106, and flow forecasting 108. Thus, unconstrained forecasting using the flow models 102 provides a viable solution to determine individual flow rates. In a second example, virtual flow metering is implemented by a business stakeholder as the constantly and lively monitored flow rates are the major focus. In this example, unconstrained forecasting using the flow models 102, well dynamics model 104, and virtual sensing model 106 provides real time monitoring of multi-phase flows. In a third example with resource-rich and high-data availability/quality scenarios, flow forecasting 108 (based on the flow models 102, well dynamics model 104, virtual sensing model 106) includes a high-confident flow rates forecasting application that aligns with the underlying well dynamics model. In this example, flow models 102, well dynamics model 104, virtual sensing model 106, and flow forecasting 108 together provide a complete solution. In embodiments, the constrained models obtained by flow forecasting 108 be generated using various inputs. In an example, the constrained models are built using modeled historical sensor data (for example, data output by a well dynamics model). In another example, the constrained models are built using both modeled historical sensor data and modeled historical flow rates (for example, data output by a flow model). In embodiments, a constrained model based on historical sensor data is built to focus on flow rates monitoring. A constrained model based on both modeled historical sensor data and modeled historical flow rates provides complete virtual flow metering and forecasting.

Generally, oil and gas production is affected by various factors, such as complexity of geological and reservoir data, dynamic operation management events, and rapid production rate changes due to development method diversity and manual interference. As a result, accurate production forecasting is complicated and challenging. Traditional techniques for building well production forecasting models suffer from requiring extensive domain knowledge. These techniques also do not have much flexibility due to their basis on complex reservoir and multiphase flow physics. Traditional techniques include numerical simulation models, analytical techniques, and decline curve analysis (DCA) models.

Numerical simulation models seek to fully describe the geological heterogeneity of the reservoir and can generally provide good results. However, the models are often tedious and time-consuming to build, where building generally involves establishing a reservoir geological model, numerical model, and history matching. Numerical simulation models also require various types of information and data about formation and multiphase flow fluid, such as logging, permeability, porosity, and saturation, which may not be available in reality or may be unreliable when available. Analytical techniques simplify complex reservoir models based on some assumptions regarding formation heterogeneity, well structure, boundary conditions and normal operation conditions. However, these analytical assumptions and results may not match those with actual production changes, especially frequent manual operations and dynamically varying underground multi-phase flow. Additionally, information regarding the correct formation and fluid data often requires long-lasting and expensive physical experiments. DCA models involve matching actual historical production rate/time data with a model, such as an exponential, harmonic, or hyperbolic model. The matched models conceptually fit abstract curves without considering the actual formation factors in general. Hence, it is difficult to guarantee the correct performance using DCA models.

In embodiments, the present techniques process production data as time series data. Machine learning, data-driven techniques are applied to time-series forecasting related applications. In embodiments, the time series analysis methods extract hidden information from historical data to predict future behavior of well production. The extracted information is combined with the reservoir and geological model and operation interferences. In some examples, the extracted hidden information includes a pattern of the time series, trend, seasonality, cycles, signal/noise ratio, frequency spectrum, similarity, etc. In some examples, the geological model may include formations, locations, rock properties, etc. Additionally, in examples the operation interferences includes operation management events and manual operations, such as gas injection, choke valve, etc. The present techniques use flow models 102 and a well dynamics model 104 as input to a virtual sensing model 106. Flow forecasting 108 is based on the virtual sensing model 106 constructed from historical and/or forecasted time data.

The block diagram of FIG. 1 is not intended to indicate that the workflow 100 is to include all of the components shown in FIG. 1. Rather, the workflow 100 can include fewer or additional components not illustrated in FIG. 1 (for example, additional models, sensor data, and the like). The workflow 100 may include any number of additional components not shown, depending on the details of the specific implementation. Furthermore, any of the functionalities of the workflow 100 may be partially, or entirely, implemented in hardware and/or in a processor. For example, the functionality may be implemented with an application specific integrated circuit, in logic implemented in a processor, in logic implemented in a specialized graphics processing unit, or in any other device.

FIG. 2 is a block diagram illustrating an unconstrained flow forecasting model 200. The flow model 200 may be, for example, a flow model 102 of FIG. 1. Generally, the flow model 200 enables unconstrained flow forecasting. Unconstrained refers to building the flow models to forecast target flow rates based on flow rate data at present and past time stamps, without constraints from auxiliary data of well dynamics. In the example of FIG. 2, historical flow rates 202 are provided as training data for unconstrained forecasting 204. The historical flow rates 202 include past gas rates, oil rates, and water rates of a well or reservoir. Unconstrained forecasting 204 outputs a forecast of flow rates 206. Unconstrained forecasting 204 is realized using machine learning models trained using historical flow rates. In embodiments, the unconstrained forecasting 204 is based on physics models, machine learning models, or a hybrid of the two. The forecasted flow rates 206 are inferred from the unconstrained forecasting 204. The forecasted flow rates include gas rates, oil rates, and water rates. In some examples, additional post-processing steps, such as smoothing are applied to the forecasted flow rates.

In the example of FIG. 2, one or multiple models are built to forecast the gas/oil/water multi-phase flow rates produced by a well or reservoir, without constraints from auxiliary data or well dynamics (for example, well dynamics model 104 of FIG. 1). The flow forecasting models 200 are trained from historical flow rates 202 of each individual target flow (for example, gas rate, oil rate, water rate), or a combination of selected flows, or all flows. The modeling approach can be based on physics models, machine learning models, or a combination of physics models and machine learning models. In examples, the unconstrained forecasting 204 is based on modeling technology such as Auto Regressive Integrated Moving Average (ARIMA) and its variants, exponential smoothing, DeepAR, and the like. For example, the model can be expressed as ARIMA(p, D, q), where the parameters p, D and q denote the structure of the forecasting model, which is a combination of auto-regression AR(p), moving average MA(q) and differencing degree D. The mathematical formula of the ARIMA (p, D, q) can be described as formula below.

$\left(1-\sum _{i=1}^p{\phi }_i{L}^i\right){\left(1-L\right)}^D{\chi }_t=\left(1+\sum _{i=1}^q{\theta }_i{L}^i\right){ϵ}_t$

Where L denotes the lag operator, φ_i are the parameters of the autoregressive part of the model, θ_i are the parameters of the MA part, and ϵ_t are error terms.

The trained models are then inferred on a future time from the training time period to predict the unconstrained target flow rates. In some embodiments, machine learning models (for example ARIMA) enable additional forecasting information associated with the flow rates, such as a prediction of confidence in user-defined intervals. An example of flow rate prediction from the unconstrained flow forecasting model is shown in FIG. 3. FIG. 3 is graph 300 illustrating flow rate production over time. The y-axis 302 provides flow rates while the x-axis 304 provides time.

The block diagram of FIG. 2 is not intended to indicate that the flow model 200 is to include all of the components shown in FIG. 2. Rather, the flow model 200 can include fewer or additional components not illustrated in FIG. 2 (for example, additional flow rates, forecasting, and the like). The flow model 200 may include any number of additional components not shown, depending on the details of the specific implementation. Furthermore, any of the functionalities of the flow model 200 may be partially, or entirely, implemented in hardware and/or in a processor. For example, the functionality may be implemented with an application specific integrated circuit, in logic implemented in a processor, in logic implemented in a specialized graphics processing unit, or in any other device.

FIG. 4 is a block diagram illustrating a well dynamics forecasting model 400. The well dynamics forecasting model 400 may be, for example, a well dynamics model 104 of FIG. 1. Generally, the well dynamics model 104 enables well dynamics forecasting. In the example of FIG. 4, historical sensor data 402 are provided as training data for a well dynamics forecasting model 404. The historical sensor data 402 include downhole pressure, downhole temperature, wellhead pressure, and wellhead temperature. The well dynamics forecasting model 404 is realized using machine learning models trained using historical sensor data 402. In embodiments, the well dynamics forecasting 404 is based on physics models, machine learning models, or a hybrid of the two. The forecasted sensor data 206 is inferred from the well dynamics forecasting 404. The forecasted sensor data include same sensors types as training data, for example, downhole pressure, downhole temperature, wellhead pressure, and wellhead temperature.

In the example of FIG. 4, one or multiple models are built to forecast the auxiliary sensor data that represent well dynamics. As used herein, well dynamics refers to data that characterizes physical conditions at a well. Instead of direct measurement of flow rates, the well dynamics model is a collection of well characteristics, such as downhole pressure, downhole temperature, wellhead pressure, wellhead temperature, and the like. In some examples, any sensors available that are related to the well characteristics can be modeled. The forecasting models are trained from historical sensor data 402 of one or more individual target sensors, or a combination of selected sensors, or all sensors. The modeling approach can be based on physics models, machine learning models, or combination of physics models and machine learning models. A few examples of modeling technology include ARIMA and its variants, exponential smoothing, DeepAR, and the like. For example, the model can be expressed as ARIMA(p, D, q), where the parameters p, D and q denote the structure of the forecasting model, which is a combination of auto-regression AR(p), moving average MA(q) and differencing degree D. The mathematical formula of the ARIMA (p, D, q) can be described as formula below.

$\left(1-\sum _{i=1}^p{\phi }_i{L}^i\right){\left(1-L\right)}^D{\chi }_t=\left(1+\sum _{i=1}^q{\theta }_i{L}^i\right){ϵ}_t$

Where L denotes the lag operator, φ_i are the parameters of the autoregressive part of the model, θ_i are the parameters of the MA part, and ϵ_t are error terms.

The trained models are then inferred on a future time from the training time period to predict the target sensor data. In some embodiments, machine learning models (for example ARIMA) enable additional forecasting information associated with sensor data, such as a prediction of confidence in user-defined intervals.

The block diagram of FIG. 4 is not intended to indicate that the well dynamics forecasting model 400 is to include all of the components shown in FIG. 4. Rather, the well dynamics forecasting model 400 can include fewer or additional components not illustrated in FIG. 4 (for example, additional sensor data, forecasting, and the like). The well dynamics forecasting model 400 may include any number of additional components not shown, depending on the details of the specific implementation. Furthermore, any of the functionalities of the well dynamics forecasting model 400 may be partially, or entirely, implemented in hardware and/or in a processor. For example, the functionality may be implemented with an application specific integrated circuit, in logic implemented in a processor, in logic implemented in a specialized graphics processing unit, or in any other device.

FIG. 5 is a block diagram illustrating a virtual flow metering model 500. The virtual flow metering model 500 may be, for example, a virtual sensing model 106 of FIG. 1. Generally, the virtual sensing model 106 enables virtual flow metering. Virtual flow metering measures, in real time, one or more flow rates without data obtained from a direct, physical measurement of the flow rates. In the example of FIG. 5, the virtual sensing model is built to predict the gas/oil/water multi-phase flow rates into the near future using virtual flow metering (VFM). The VFM model 500 is trained from historical sensor data as input features and historical flow rates as ground truth labels. In examples, the historical flow rates are used to label features extracted from the historical sensor data. The modeling approach is a supervised regression, by either physics-driven modeling such as first principle simulation by a dynamic multi-fluid model (for example, OLGA, and the like), data-driven modeling including machine learning models, or a combination of physics driven modeling and data-driven modeling. A few examples of modeling technology include random forest regression, XGBoost regression, deep fully connected network (DFCN) regression, long short-term memory (LSTM) regression, and the like. In the example of FIG. 5, historical flow rates 502 and historical sensor data 504 are provided as training data to virtual sensing models 506. Virtual sensing 506 is realized using machine learning models such as random forest (RF), XGBoost, DFCN, LTSM, and the like. In embodiments, the virtual sensing 506 is based on physics models, machine learning models, or a hybrid of the two. Flow rates 508 are inferred from the virtual sensing 506. The inferred flow rates include oil, water, and gas flow rates.

The block diagram of FIG. 5 is not intended to indicate that the virtual flow metering model 500 is to include all of the components shown in FIG. 5. Rather, the virtual flow metering model 500 can include fewer or additional components not illustrated in FIG. 5 (for example, additional flow rates, sensor data, forecasting, and the like). The virtual flow metering model 500 may include any number of additional components not shown, depending on the details of the specific implementation. Furthermore, any of the functionalities of the virtual flow metering model 500 may be partially, or entirely, implemented in hardware and/or in a processor. For example, the functionality may be implemented with an application specific integrated circuit, in logic implemented in a processor, in logic implemented in a specialized graphics processing unit, or in any other device.

FIG. 6 is an illustration of production flow rate 600 ground truth label and model prediction made using a virtual flow metering model (for example, model 500 of FIG. 5). FIG. 6 is an illustration of true flow rates and predicted flow rates over time. The y-axis 602 provides flow rates while the x-axis 604 provides time. In the example of FIG. 6, the trained model (for example, virtual flow metering model 500) represents a correlation model between the multi-phase flow rates and the auxiliary sensors data, and thus reflects the characteristics of the well performance. It is then inferred on a next time stamp, to predict the instantaneous target flow rates. An example of instantaneous flow rate prediction from the virtual flow metering model 500 is shown in FIG. 6.

FIG. 7 is a block diagram illustrating a model-constrained flow forecasting model 700. The model-constrained flow forecasting 700 may be, for example, a flow forecasting model 108 of FIG. 1. Generally, the model-constrained flow forecasting model 700 is based on unconstrained forecasting 702 (for example, model 200 of FIG. 2) to generate forecast flow rates 706 and well dynamics 704 (for example, model 400 of FIG. 4) to generate forecast sensor data 708. In the example of FIG. 7, the model-constrained flow forecasting model 700 provides the forecast flow rates 706 and the forecast sensor data 708 as input to trained virtual sensing models 710. Virtual sensing 710 is realized using machine learning models such as random forest (RF), XGBoost, DFCN, LTSM, and the like. In embodiments, the virtual sensing 710 is based on physics models, machine learning models, or a hybrid of the two. Constrained flow rates are inferred from the virtual sensing 710. The inferred constrained flow rates include oil, water, and gas flow rates. In embodiments, the virtually sensed multi-phase flow rates enable constrained forecasting 712 of future multi-phase flow rates as constrained by one or more models.

In the example of FIG. 7, the model-constrained forecasting model 700 is built on top of the other models, i.e. unconstrained flow forecasting models, well dynamics models, and virtual flow metering models, to forecast the gas/oil/water multi-phase flow rates. The outputs of unconstrained flow forecasting models and well dynamics models are combined as input features to the virtual flow metering model. The virtual flow metering model reflects the characteristics of the well performance and constrains the correlation between the forecasted flow rates and sensor data.

The block diagram of FIG. 7 is not intended to indicate that the model-constrained flow forecasting model 700 is to include all of the components shown in FIG. 7. Rather, the model-constrained flow forecasting model 700 can include fewer or additional components not illustrated in FIG. 7 (for example, additional models, sensor data, and the like). The model-constrained flow forecasting model 700 may include any number of additional components not shown, depending on the details of the specific implementation. Furthermore, any of the functionalities of the virtual flow metering model 700 may be partially, or entirely, implemented in hardware and/or in a processor. For example, the functionality may be implemented with an application specific integrated circuit, in logic implemented in a processor, in logic implemented in a specialized graphics processing unit, or in any other device.

FIG. 8 is a graph 800 illustrating model-constrained flow forecasting over time. The y-axis 802 provides flow rates while the x-axis 804 provides time. In the example of FIG. 8, the workflow of the model-constrained flow forecasting model 700 is inferred on the future time from the training time period, to predict the target flow rates in a model-constrained manner.

FIG. 9 is a process flow diagram of a process 900 for model-constrained multi-phase virtual flow metering and forecasting with machine learning. The process 900 may include the workflow 100 of FIG. 1. In embodiments, the process 900 is implemented using the controller 1000 of FIG. 10.

At block 902, at least one unconstrained flow model is built to forecast target flow rates based on current and historic multi-phase flow data. At block 904, a well dynamics model is built that captures well dynamics (for example, auxiliary sensor data, target sensor data). Generally, the well dynamics models are independent from the unconstrained flow model, and vice versa. In some use cases, only unconstrained forecasting models are applied. At block 906, a virtual sensing model is constructed to predict the gas/oil/water multi-phase flow rates. At block 906, a constrained forecasting model is built by combining the unconstrained flow forecasting model, well dynamics model, and virtual sensing model.

The process flow diagram of FIG. 9 is not intended to indicate that the process 900 is to include all of the steps shown in FIG. 9. Rather, the process 900 can include fewer or additional components not illustrated in FIG. 9 (for example, additional micro-containers, applications, and the like). The process 900 of FIG. 9 may include any number of additional components not shown, depending on the details of the specific implementation.

FIG. 10 is a schematic illustration of an example controller 1000 (or control system) for model-constrained multi-phase virtual flow metering and forecasting with machine learning according to the present disclosure. For example, the controller 1000 may be operable according to the workflow 100 of FIG. 1 or the process 900 of FIG. 9. The controller 1000 is intended to include various forms of digital computers, such as printed circuit boards (PCB), processors, digital circuitry, or otherwise parts of a system for supply chain alert management. Additionally the system can include portable storage media, such as, Universal Serial Bus (USB) flash drives. For example, the USB flash drives may store operating systems and other applications. The USB flash drives can include input/output components, such as a wireless transmitter or USB connector that may be inserted into a USB port of another computing device.

The controller 1000 includes a processor 1010, a memory 1020, a storage device 1030, and an input/output interface 1040 communicatively coupled with input/output devices 1060 (for example, displays, keyboards, measurement devices, sensors, valves, pumps). Each of the components 1010, 1020, 1030, and 1040 are interconnected using a system bus 1050. The processor 1010 is capable of processing instructions for execution within the controller 1000. The processor may be designed using any of a number of architectures. For example, the processor 1010 may be a CISC (Complex Instruction Set Computers) processor, a RISC (Reduced Instruction Set Computer) processor, or a MISC (Minimal Instruction Set Computer) processor.

In one implementation, the processor 1010 is a single-threaded processor. In another implementation, the processor 1010 is a multi-threaded processor. The processor 1010 is capable of processing instructions stored in the memory 1020 or on the storage device 1030 to display graphical information for a user interface on the input/output interface 1040.

The memory 1020 stores information within the controller 1000. In one implementation, the memory 1020 is a computer-readable medium. In one implementation, the memory 1020 is a volatile memory unit. In another implementation, the memory 1020 is a nonvolatile memory unit.

The storage device 1030 is capable of providing mass storage for the controller 1000. In one implementation, the storage device 1030 is a computer-readable medium. In various different implementations, the storage device 1030 may be a floppy disk device, a hard disk device, an optical disk device, or a tape device.

The input/output interface 1040 provides input/output operations for the controller 1000. In one implementation, the input/output devices 1060 includes a keyboard and/or pointing device. In another implementation, the input/output devices 1060 includes a display unit for displaying graphical user interfaces.

There can be any number of controllers 1000 associated with, or external to, a computer system containing controller 1000, with each controller 1000 communicating over a network. Further, the terms “client,” “user,” and other appropriate terminology can be used interchangeably, as appropriate, without departing from the scope of the present disclosure. Moreover, the present disclosure contemplates that many users can use one controller 1000 and one user can use multiple controllers 1000.

Implementations of the subject matter and the functional operations described in this specification can be implemented in digital electronic circuitry, in tangibly embodied computer software or firmware, in computer hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. Software implementations of the described subject matter can be implemented as one or more computer programs. Each computer program can include one or more modules of computer program instructions encoded on a tangible, non-transitory, computer-readable computer-storage medium for execution by, or to control the operation of, data processing apparatus. Alternatively, or additionally, the program instructions can be encoded in/on an artificially generated propagated signal. The example, the signal can be a machine-generated electrical, optical, or electromagnetic signal that is generated to encode information for transmission to suitable receiver apparatus for execution by a data processing apparatus. The computer-storage medium can be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or a combination of computer-storage mediums.

The terms “data processing apparatus,” “computer,” and “electronic computer device” (or equivalent as understood by one of ordinary skill in the art) refer to data processing hardware. For example, a data processing apparatus can encompass all kinds of apparatus, devices, and machines for processing data, including by way of example, a programmable processor, a computer, or multiple processors or computers. The apparatus can also include special purpose logic circuitry including, for example, a central processing unit (CPU), a field programmable gate array (FPGA), or an application specific integrated circuit (ASIC). In some implementations, the data processing apparatus or special purpose logic circuitry (or a combination of the data processing apparatus or special purpose logic circuitry) can be hardware- or software-based (or a combination of both hardware- and software-based). The apparatus can optionally include code that creates an execution environment for computer programs, for example, code that constitutes processor firmware, a protocol stack, a database management system, an operating system, or a combination of execution environments. The present disclosure contemplates the use of data processing apparatuses with or without conventional operating systems, for example, LINUX, UNIX, WINDOWS, MAC OS, ANDROID, or IOS.

A computer program, which can also be referred to or described as a program, software, a software application, a module, a software module, a script, or code, can be written in any form of programming language. Programming languages can include, for example, compiled languages, interpreted languages, declarative languages, or procedural languages. Programs can be deployed in any form, including as stand-alone programs, modules, components, subroutines, or units for use in a computing environment. A computer program can, but need not, correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data, for example, one or more scripts stored in a markup language document, in a single file dedicated to the program in question, or in multiple coordinated files storing one or more modules, sub programs, or portions of code. A computer program can be deployed for execution on one computer or on multiple computers that are located, for example, at one site or distributed across multiple sites that are interconnected by a communication network. While portions of the programs illustrated in the various figures may be shown as individual modules that implement the various features and functionality through various objects, methods, or processes, the programs can instead include a number of sub-modules, third-party services, components, and libraries. Conversely, the features and functionality of various components can be combined into single components as appropriate. Thresholds used to make computational determinations can be statically, dynamically, or both statically and dynamically determined.

The methods, processes, or logic flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform functions by operating on input data and generating output. The methods, processes, or logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, for example, a CPU, an FPGA, or an ASIC.

Computers suitable for the execution of a computer program can be based on one or more of general and special purpose microprocessors and other kinds of CPUs. The elements of a computer are a CPU for performing or executing instructions and one or more memory devices for storing instructions and data. Generally, a CPU can receive instructions and data from (and write data to) a memory. A computer can also include, or be operatively coupled to, one or more mass storage devices for storing data. In some implementations, a computer can receive data from, and transfer data to, the mass storage devices including, for example, magnetic, magneto optical disks, or optical disks. Moreover, a computer can be embedded in another device, for example, a mobile telephone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device such as a universal serial bus (USB) flash drive.

Computer readable media (transitory or non-transitory, as appropriate) suitable for storing computer program instructions and data can include all forms of permanent/non-permanent and volatile/non-volatile memory, media, and memory devices. Computer readable media can include, for example, semiconductor memory devices such as random access memory (RAM), read only memory (ROM), phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), and flash memory devices. Computer readable media can also include, for example, magnetic devices such as tape, cartridges, cassettes, and internal/removable disks. Computer readable media can also include magneto optical disks and optical memory devices and technologies including, for example, digital video disc (DVD), CD ROM, DVD+/−R, DVD-RAM, DVD-ROM, HD-DVD, and BLURAY. The memory can store various objects or data, including caches, classes, frameworks, applications, modules, backup data, jobs, web pages, web page templates, data structures, database tables, repositories, and dynamic information. Types of objects and data stored in memory can include parameters, variables, algorithms, instructions, rules, constraints, and references. Additionally, the memory can include logs, policies, security or access data, and reporting files. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.

Implementations of the subject matter described in the present disclosure can be implemented on a computer having a display device for providing interaction with a user, including displaying information to (and receiving input from) the user. Types of display devices can include, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), a light-emitting diode (LED), and a plasma monitor. Display devices can include a keyboard and pointing devices including, for example, a mouse, a trackball, or a trackpad. User input can also be provided to the computer through the use of a touchscreen, such as a tablet computer surface with pressure sensitivity or a multi-touch screen using capacitive or electric sensing. Other kinds of devices can be used to provide for interaction with a user, including to receive user feedback including, for example, sensory feedback including visual feedback, auditory feedback, or tactile feedback. Input from the user can be received in the form of acoustic, speech, or tactile input. In addition, a computer can interact with a user by sending documents to, and receiving documents from, a device that is used by the user. For example, the computer can send web pages to a web browser on a user's client device in response to requests received from the web browser.

The term “graphical user interface,” or “GUI,” can be used in the singular or the plural to describe one or more graphical user interfaces and each of the displays of a particular graphical user interface. Therefore, a GUI can represent any graphical user interface, including, but not limited to, a web browser, a touch screen, or a command line interface (CLI) that processes information and efficiently presents the information results to the user. In general, a GUI can include a plurality of user interface (UI) elements, some or all associated with a web browser, such as interactive fields, pull-down lists, and buttons. These and other UI elements can be related to or represent the functions of the web browser.

Implementations of the subject matter described in this specification can be implemented in a computing system that includes a back end component, for example, as a data server, or that includes a middleware component, for example, an application server. Moreover, the computing system can include a front-end component, for example, a client computer having one or both of a graphical user interface or a Web browser through which a user can interact with the computer. The components of the system can be interconnected by any form or medium of wireline or wireless digital data communication (or a combination of data communication) in a communication network. Examples of communication networks include a local area network (LAN), a radio access network (RAN), a metropolitan area network (MAN), a wide area network (WAN), Worldwide Interoperability for Microwave Access (WIMAX), a wireless local area network (WLAN) (for example, using 802.11 a/b/g/n or 802.20 or a combination of protocols), all or a portion of the Internet, or any other communication system or systems at one or more locations (or a combination of communication networks). The network can communicate with, for example, Internet Protocol (IP) packets, frame relay frames, asynchronous transfer mode (ATM) cells, voice, video, data, or a combination of communication types between network addresses.

The computing system can include clients and servers. A client and server can generally be remote from each other and can typically interact through a communication network. The relationship of client and server can arise by virtue of computer programs running on the respective computers and having a client-server relationship. Cluster file systems can be any file system type accessible from multiple servers for read and update. Locking or consistency tracking may not be necessary since the locking of exchange file system can be done at application layer. Furthermore, Unicode data files can be different from non-Unicode data files.

While this specification contains many specific implementation details, these should not be construed as limitations on the scope of what may be claimed, but rather as descriptions of features that may be specific to particular implementations. Certain features that are described in this specification in the context of separate implementations can also be implemented, in combination, in a single implementation. Conversely, various features that are described in the context of a single implementation can also be implemented in multiple implementations, separately, or in any suitable sub-combination. Moreover, although previously described features may be described as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can, in some cases, be excised from the combination, and the claimed combination may be directed to a sub-combination or variation of a sub-combination.

Particular implementations of the subject matter have been described. Other implementations, alterations, and permutations of the described implementations are within the scope of the following claims as will be apparent to those skilled in the art. While operations are depicted in the drawings or claims in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed (some operations may be considered optional), to achieve desirable results. In certain circumstances, multitasking or parallel processing (or a combination of multitasking and parallel processing) may be advantageous and performed as deemed appropriate.

Moreover, the separation or integration of various system modules and components in the previously described implementations should not be understood as requiring such separation or integration in all implementations, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

Accordingly, the previously described example implementations do not define or constrain the present disclosure. Other changes, substitutions, and alterations are also possible without departing from the spirit and scope of the present disclosure.

Furthermore, any claimed implementation is considered to be applicable to at least a computer-implemented method; a non-transitory, computer-readable medium storing computer-readable instructions to perform the computer-implemented method; and a computer system comprising a computer memory interoperably coupled with a hardware processor configured to perform the computer-implemented method or the instructions stored on the non-transitory, computer-readable medium.

Particular embodiments of the subject matter have been described. Other embodiments are within the scope of the following claims. For example, the actions recited in the claims can be performed in a different order and still achieve desirable results. As one example, some processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results.

Claims

1. A computer-implemented method for model constrained multi-phase virtual flow metering and forecasting, the method comprising: forecasting, with one or more hardware processors, target flow rates for real-time monitoring using unconstrained flow models trained using current and historic multi-phase flow rates;forecasting, with the one or more hardware processors, auxiliary sensor data for real-time monitoring using trained well dynamics models;constructing, with the one or more hardware processors, at least one virtual sensing model trained using the forecast target flow rates and the forecast auxiliary sensor data, wherein the at least one virtual sensing model predicts real-time constrained multi-phase flow rates; andmonitoring, with the one or more hardware processors, the predicted real-time constrained multi-phase flow rates output by the at least one virtual sensing model of a virtual flow metering system that enables production planning based on the real-time constrained multi-phase flow rates.

2. The computer-implemented method of claim 1, comprising combining the trained unconstrained flow models, the trained well dynamics models, and the at least one trained virtual sensing model by providing outputs of the trained unconstrained flow forecasting models and the trained well dynamics models as input features to the at least one trained virtual sensing model.

3. The computer-implemented method of claim 1, wherein the unconstrained flow models are trained from historical flow rates of each individual target flow, or a combination of selected or all flows.

4. The computer-implemented method of claim 1, wherein the unconstrained flow models are based on physics models, machine learning models, or any combinations thereof.

5. The computer-implemented method of claim 1, wherein the well dynamics models are trained using historical sensor data.

6. The computer-implemented method of claim 1, wherein the at least one trained virtual sensing model predicts instantaneous multi-phase flow rates based on historical flow rates and sensor data.

7. The computer-implemented method of claim 1, wherein the virtual flow metering system enables virtual flow metering that predicts constraint flow rates as a function of time.

8. An apparatus comprising a non-transitory, computer readable, storage medium that stores instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising: predicting target flow rates for real time monitoring using unconstrained flow models trained using current and historic multi-phase flow rates;predicting future auxiliary sensor data for real-time monitoring using trained well dynamics models;constructing at least one virtual sensing model trained using predicted target flow rates and predicted auxiliary sensor data, wherein the at least one virtual sensing model predicts real-time constrained multi-phase flow rates; andmonitoring the predicted real-time constrained multi-phase flow rates output by the at least one virtual sensing model of a virtual flow metering system that enables production planning based on the real-time constrained multi-phase flow rates.

9. The apparatus of claim 8, comprising combining the trained unconstrained flow models, the trained well dynamics models, and the at least one trained virtual sensing model by providing outputs of the trained unconstrained flow forecasting models and the trained well dynamics models as input features to the at least one trained virtual sensing model.

10. The apparatus of claim 8, wherein the unconstrained flow is models are trained from historical flow rates of each individual target flow, or a combination of selected or all flows.

11. The apparatus of claim 8, wherein the unconstrained flow is models are based on physics models, machine learning models, or any combinations thereof.

12. The apparatus of claim 8, wherein the well dynamics models are trained using historical sensor data.

13. The apparatus of claim 8, wherein the at least one trained virtual sensing model predicts instantaneous multi-phase flow rates based on historical flow rates and sensor data.

14. The apparatus of claim 8, wherein the virtual flow metering system enables virtual flow metering that predicts constraint flow rates as a function of time.

15. A system, comprising: one or more memory modules;one or more hardware processors communicably coupled to the one or more memory modules, the one or more hardware processors configured to execute instructions stored on the one or more memory models to perform operations comprising:predicting target flow rates for real-time monitoring using unconstrained flow models trained using current and historic multi-phase flow rates;predicting auxiliary sensor data for real-time monitoring using trained well dynamics models;constructing at least one virtual sensing model trained using the predicted target flow rates and the predicted auxiliary sensor data, wherein the at least one virtual sensing model predicts real-time constrained multi-phase flow rates; andmonitoring the predicted real-time constrained multi-phase flow rates output by the at least one virtual sensing model of a virtual flow metering system that enables production planning based on the real-time constrained multi-phase flow rates.

16. The system of claim 15, comprising the operations comprise combining the trained unconstrained flow models, the trained well dynamics models, and the at least one trained virtual sensing model by providing outputs of the trained unconstrained flow forecasting models and the trained well dynamics models as input features to the at least one trained virtual sensing model.

17. The system of claim 15, wherein the operations comprise training the unconstrained flow models from historical flow rates of each individual target flow, or a combination of selected or all flows.

18. The system of claim 15, wherein the operations comprise generating the unconstrained flow models based on physics models, machine learning models, or any combinations thereof.

19. The system of claim 15, wherein the well dynamics models are trained using historical sensor data.

20. The system of claim 15, wherein the at least one trained virtual sensing model predicts instantaneous multi-phase flow rates based on historical flow rates and sensor data.