Neural network prediction of fluid flow in porous media

By generating simplified network representations and training neural network models, the problem of high computing resource consumption in porous medium fluid flow analysis is solved, and efficient prediction of fluid flow attributes is achieved.

CN119013670APending Publication Date: 2024-11-22INTERNATIONAL BUSINESS MACHINE CORPORATION
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202380030223.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-03-30
Filing Date
2023-03-28
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

Existing methods for fluid flow analysis of porous media require a large amount of computing resources, especially when generating flow properties directly from the capillary network representation, and it is difficult to efficiently predict the fluid flow properties of porous media.

Method used

By generating a simplified network representation from a capillary network representation and performing fluid flow simulations with a physics-based simulator, the simulated flow properties of the simplified network representation are determined, and then using these properties to train a neural network model to generate a model for predicting fluid flow properties.

Benefits of technology

It significantly reduces the demand for computing resources, improves the prediction efficiency and accuracy of fluid flow properties of porous media, and provides a more efficient digital imaging system to predict the flow properties of porous media samples.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119013670A_ABST
    Figure CN119013670A_ABST
Patent Text Reader

Abstract

Systems and methods for predicting fluid flow of a porous medium are provided. In an embodiment, a method includes accessing, by a computing device, a capillary network representation of a porous media sample; generating, by the computing device, a set of simplified network representations from the capillary network representation; performing, by the computing device, a fluid flow simulation using a simulator to determine simulated fluid flow properties for each of the simplified network representations; and training, by the computing device, a neural network (NN) model with the set of simplified network representations as inputs and with the simulated fluid flow properties as model goals, thereby generating a trained NN model for predicting fluid flow properties of the porous medium.
Need to check novelty before this filing date? Find Prior Art

Description

Background Art

[0001] Aspects of the present invention generally relate to fluid flow analysis of porous media, and more particularly, to neural network (NN) prediction of fluid flow in porous media.

[0002] Permeability is a key parameter used to quantify fluid flow in porous media (e.g., rock). Various methods have been developed to determine the permeability and other characteristics of geological samples. One method includes x-ray microtomography imaging of porous media samples. Typically, x-ray microtomography or micro-computed tomography (micro-CT) uses x-rays to create images of cross-sections of physical objects, which can be used to reconstruct a virtual three-dimensional (3D) representation without destroying the original physical object. Another method performs image processing of the 3D representation generated from the x-ray microtomography to generate a capillary network representation of the physical object. Fluid flow simulations can be performed on the capillary network representation to predict fluid flow through the physical object. Summary of the invention

[0003] In a first aspect of the present invention, there is a computer-implemented method, comprising: accessing, by a computing device, a capillary network representation of a porous media sample; generating, by the computing device, a set of simplified network representations from the capillary network representation; performing, by the computing device, a fluid flow simulation using a simulator to determine simulated fluid flow properties of each of the simplified network representations; and training, by the computing device, a neural network (NN) model using the set of simplified network representations as input and using the simulated fluid flow properties as model targets, thereby generating a trained NN model for predicting fluid flow properties of porous media.

[0004] In another aspect of the present invention, there is a computer program product comprising one or more computer-readable storage media having program instructions stored together on the one or more computer-readable storage media. The program instructions are executable to: access a capillary network representation, wherein the capillary network representation is a three-dimensional (3D) network of interconnected capillaries representing an interconnected pore structure of a 3D micro-CT scanner model of a porous media sample; generate a set of simplified network representations from the capillary network representation, wherein each simplified network representation in the set of simplified network representations includes a two-dimensional (2D) or 3D network of interconnected capillary structures representing a subset of the interconnected capillary structures of the capillary network representation; perform a fluid flow simulation using a physics-based simulator to determine simulated fluid flow properties for each simplified network representation in the simplified network representations; and train a neural network (NN) model using the set of simplified network representations as input and using the simulated fluid flow properties as model targets, thereby generating a trained NN model for predicting fluid flow properties.

[0005] In another aspect of the present invention, a system is provided, comprising a processor, a computer readable memory, one or more computer readable storage media, and program instructions stored together on the one or more computer readable storage media. The program instructions are executable to: access a capillary network representation, wherein the capillary network representation is a three-dimensional (3D) network of interconnected capillaries representing the interconnected pore structure of a 3D micro-CT scanner model of a porous media sample; generate a set of simplified network representations from the capillary network representation, wherein each simplified network representation in the set of simplified network representations includes a two-dimensional (2D) or 3D network of interconnected capillary structures representing a subset of the interconnected capillary structures of the capillary network representation; perform fluid flow simulations using a physics-based simulator to determine simulated fluid flow properties for each simplified network representation in the simplified network representations; train a neural network (NN) model using the set of simplified network representations as input and using the simulated fluid flow properties as model targets, thereby generating a trained NN model for predicting fluid flow properties; and input a new capillary network representation of a new porous media sample into the trained NN model, thereby generating predicted fluid flow properties for the new porous media sample. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Various aspects of the invention are described in the following detailed description by way of non-limiting examples of exemplary embodiments of the invention, with reference to the various figures mentioned.

[0007] Figure 1 A cloud computing node according to an embodiment of the invention is depicted.

[0008] Figure 2 A cloud computing environment according to an embodiment of the present invention is depicted.

[0009] Figure 3 Abstract model layers according to an embodiment of the present invention are depicted.

[0010] Figure 4 A block diagram of an exemplary environment according to aspects of the invention is shown.

[0011] Figure 5 is a flow chart providing an overview of a porous media modeling system according to an embodiment of the present invention.

[0012] Figure 6 is an illustrative example of generating predicted flow properties according to an embodiment of the present invention.

[0013] Figure 7 is a flowchart illustrating the training of a NN model according to an embodiment of the present invention.

[0014] Figure 8Neural network training for fluid flow prediction according to an embodiment of the present invention is shown.

[0015] Fig. 9 Fluid flow prediction using a trained NN model according to an embodiment of the present invention is shown.

[0016] Fig.10 A flow chart of an exemplary method according to aspects of the invention is shown. DETAILED DESCRIPTION

[0017] Aspects of the present invention generally relate to fluid flow analysis of porous media, and more specifically, to neural network (NN) prediction of fluid flow in porous media. In an embodiment, a porous media (e.g., rock) sample is imaged by x-ray micro-tomography using a micro-computed tomography (micro-CT) scanner to generate a three-dimensional (3D) representation or image cube representing the pore space of the porous media sample. In an embodiment, a computing device uses a network extraction routine to process the 3D representation and generate a capillary network representation of the pore space of the sample, applies a physics-based flow simulator to the capillary network representation to simulate the flow of a selected fluid through the sample under selected conditions (e.g., pressure, temperature, etc.) to determine flow properties of interest (e.g., permeability, flow rate, residual fluid saturation, etc.), and stores the flow properties of interest for later validation of alternative NN models.

[0018] In an embodiment, a computing device applies a network simplification routine to generate a set of simplified network representations (e.g., 50-100 representations) from a capillary network representation, wherein the set of simplified network representations is substantially smaller than the capillary network representation (e.g., 3 or more orders of magnitude smaller). In an implementation, for each simplified network representation, a physics-based simulator performs a flow simulation and obtains simulated flow properties for each simplified network representation. In various aspects of the present invention, a computing device performs a NN training routine using the simplified network representation as input and using its simulated flow properties as model targets to create a trained NN model, which is configured to generate predicted flow properties as output based on the input of the capillary network representation. In an embodiment, the computing device compares the simulated flow properties with the predicted flow properties for NN model validation purposes.

[0019] The execution of flow simulations based on capillary network representations may require significant computational resources. Embodiments of the present invention provide an improved digital image-based modeling system that estimates flow properties of porous media samples using significantly fewer computational resources than would be required to generate flow property estimates directly from a capillary network representation. In an embodiment, an improved digital imaging system (e.g., a micro-CT scanner) is provided to predict flow properties of porous media samples based on digital image processing and the use of a trained NN model.

[0020] The present invention may be a system, method and / or computer program product at any possible level of technical detail integration. The computer program product may include a computer-readable storage medium (or multiple media) having computer-readable program instructions thereon for causing a processor to perform various aspects of the present invention.

[0021] A computer-readable storage medium may be a tangible device that can retain and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of computer-readable storage media includes the following: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD), a memory stick, a floppy disk, a mechanical encoding device (such as a punched card or a raised structure in a groove on which instructions are recorded), and any suitable combination of the foregoing. As used herein, one or more computer-readable storage media should not be interpreted as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber-optic cables), or electrical signals transmitted through wires.

[0022] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to a corresponding computing / processing device, or downloaded to an external computer or external storage device via a network (e.g., the Internet, a local area network, a wide area network, and / or a wireless network). The network can include copper transmission cables, optical transmission fibers, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in a computer-readable storage medium in the corresponding computing / processing device.

[0023] The computer-readable program instructions for performing the operation of the present invention can be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, configuration data of integrated circuits, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, and process programming languages ​​such as "C" programming language or similar programming languages. The computer-readable program instructions can be executed completely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or completely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any type of network (including a local area network (LAN) or a wide area network (WAN)), or can be connected to an external computer (for example, using an Internet service provider through the Internet). In some embodiments, the electronic circuit including, for example, a programmable logic circuit, a field programmable gate array (FPGA) or a programmable logic array (PLA) can execute the computer-readable program instructions to personalize the electronic circuit by utilizing the state information of the computer-readable program instructions, so as to perform various aspects of the present invention.

[0024] Various aspects of the present invention are described herein with reference to the flow charts and / or block diagrams of the methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each frame of the flow chart diagram and / or block diagram and the combination of frames in the flow chart diagram and / or block diagram can be implemented by computer-readable program instructions.

[0025] These computer-readable program instructions can be provided to a processor of a computer or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device create a means for implementing the functions / actions specified in one or more boxes of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium, which can instruct a computer, a programmable data processing device, and / or other equipment to function in a specific manner, so that the computer-readable storage medium having the instructions stored therein includes an article of manufacture, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes of the flowchart and / or block diagram.

[0026] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device, so that a series of operational steps are performed on the computer, other programmable apparatus, or other device to produce a computer-implemented process, so that the instructions executed on the computer, other programmable apparatus, or other device implement the functions / actions specified in one or more boxes of the flowchart and / or block diagram.

[0027] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each frame in the flow chart or block diagram may represent a part of a module, segment or instruction, which includes one or more executable instructions for realizing the specified logical function. In some alternative embodiments, the function indicated in the frame may not occur in the order indicated in the figure. For example, the two frames shown in succession can actually be implemented as a step, and are executed simultaneously, substantially simultaneously, in a partially or completely time-overlapping manner, or the frame can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each frame of the block diagram and / or flow chart illustration and the combination of the frames in the block diagram and / or flow chart illustration can be implemented by a dedicated hardware-based system that performs a specified function or action or performs a combination of dedicated hardware and computer instructions.

[0028] It should be understood that although the present disclosure includes detailed descriptions about cloud computing, the implementation of the teachings described herein is not limited to cloud computing environments. Rather, embodiments of the present invention can be implemented in conjunction with any other type of computing environment now known or later developed.

[0029] Cloud computing is a service delivery model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, network bandwidth, servers, processing, memory, storage, applications, virtual machines, and services) that can be quickly provisioned and released with minimal management effort or interaction with the service provider. The cloud model can include at least five characteristics, at least three service models, and at least four deployment models.

[0030] Features are as follows:

[0031] On-demand self-service: Cloud consumers can unilaterally and automatically provision computing capabilities, such as server time and network storage, as needed, without requiring human interaction with the service provider.

[0032] Broad network access: Capabilities are available over the network and accessed through standard mechanisms that facilitate use with heterogeneous thin-client platforms or thick-client platforms (e.g., mobile phones, laptops, and PDAs).

[0033] Resource pooling: A provider's computing resources are pooled to serve multiple consumers using a multi-tenant model, where different physical and virtual resources are dynamically allocated and reallocated based on demand. There is a sense of location independence, as consumers typically do not have control or knowledge of the exact location of the provided resources, but may be able to specify a location at a higher level of abstraction (e.g., country, state, or data center).

[0034] Rapid elasticity: Capacity can be quickly and elastically (in some cases automatically) provided to scale up quickly and quickly released to scale down quickly. To the consumer, the capacity available for provisioning often appears to be unlimited and can be purchased in any quantity at any time.

[0035] Metered Services: Cloud systems automatically control and optimize resource usage by leveraging metering capabilities at some level of abstraction appropriate to the type of service (e.g., storage, processing, bandwidth, and active user accounts). Resource usage can be monitored, controlled, and reported, providing transparency to both the provider and consumer of the utilized service.

[0036] The service model is as follows:

[0037] Software as a Service (SaaS): The capability provided to consumers is to use the provider's applications running on a cloud infrastructure. Applications can be accessed from a variety of client devices through a thin client interface such as a web browser (e.g., web-based email). Consumers do not manage or control the underlying cloud infrastructure, including the network, servers, operating systems, storage, or even individual application capabilities, with the possible exception of limited user-specific application configuration settings.

[0038] Platform as a Service (PaaS): The capability provided to consumers is to deploy applications created or acquired by consumers onto cloud infrastructure, where the applications are created using programming languages ​​and tools supported by the provider. Consumers do not manage or control the underlying cloud infrastructure including networks, servers, operating systems, or storage, but have control over deployed applications and possibly the configuration of the application hosting environment.

[0039] Infrastructure as a Service (IaaS): The capabilities provided to consumers are the provision of processing, storage, networking, and other basic computing resources, where consumers can deploy and run arbitrary software, which may include operating systems and applications. Consumers do not manage or control the underlying cloud infrastructure, but have control over the operating system, storage, deployed applications, and may have limited control over selected networking components (e.g., host firewalls).

[0040] The deployment model is as follows:

[0041] Private Cloud: The cloud infrastructure is operated only for the organization. It can be managed by the organization or a third party and can exist on-premises or off-premises.

[0042] Community cloud: The cloud infrastructure is shared by several organizations and supports a specific community with common concerns (e.g., mission, security requirements, policies, and compliance considerations). It can be managed by the organization or a third party and can exist on-premises or off-premises.

[0043] Public Cloud: Cloud infrastructure is available to the general public or large industrial groups and is owned by the organization that sells cloud services.

[0044] Hybrid cloud: A cloud infrastructure consisting of two or more clouds (private, community, or public) that remain unique entities but are bound together by standardized or proprietary technologies that enable data and application portability (e.g., cloud bursting for load balancing between clouds).

[0045] The cloud computing environment is service-oriented, focusing on statelessness, low coupling, modularity, and semantic interoperability. The core of cloud computing is the infrastructure consisting of a network of interconnected nodes.

[0046] Reference now Figure 1 , a schematic diagram of an example of a cloud computing node is shown. Cloud computing node 10 is only one example of a suitable cloud computing node and is not intended to impose any limitations on the scope of use or functionality of the embodiments of the present invention described herein. Regardless, cloud computing node 10 can be implemented and / or perform any of the functions set forth above.

[0047] In cloud computing node 10, there is a computer system / server 12, which can operate with many other general or special computing system environments or configurations. Examples of well-known computing systems, environments, and / or configurations that may be suitable for computer system / server 12 include, but are not limited to, personal computer systems, server computer systems, thin clients, fat clients, handheld or laptop devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputer systems, mainframe computer systems, and distributed cloud computing environments including any of the above systems or devices, etc.

[0048] Computer system / server 12 may be described in the general context of computer system executable instructions (such as program modules) executed by a computer system. Generally, program modules may include routines, programs, objects, components, logic, data structures, etc. that perform specific tasks or implement specific abstract data types. Computer system / server 12 may be practiced in a distributed cloud computing environment, where tasks are performed by remote processing devices linked through a communication network. In a distributed cloud computing environment, program modules may be located in local and remote computer system storage media including memory storage devices.

[0049] like Figure 1 As shown, the computer system / server 12 in the cloud computing node 10 is shown in the form of a general-purpose computing device. The components of the computer system / server 12 may include, but are not limited to, one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 to the processor 16.

[0050] Bus 18 represents any one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example and not limitation, such architectures include an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MCA) bus, an Enhanced ISA (EISA) bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0051] Computer system / server 12 typically includes a variety of computer system readable media. Such media can be any available media that can be accessed by computer system / server 12, and it includes volatile and nonvolatile media, removable and non-removable media.

[0052] The system memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The computer system / server 12 may also include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, a storage system 34 may be provided for reading and writing from a non-removable non-volatile magnetic medium (not shown and generally referred to as a "hard drive"). Although not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") may be provided, as well as an optical drive for reading or writing to a removable non-volatile optical disk (such as a CD-ROM, DVD-ROM or other optical media). In this case, each may be connected to the bus 18 via one or more data media interfaces. As will be further depicted and described below, the memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of an embodiment of the present invention.

[0053] A program / utility 40 having a set (at least one) of program modules 42, as well as an operating system, one or more application programs, other program modules, and program data may be stored in memory 28 as an example and not by way of limitation. Each or some combination of the operating system, one or more application programs, other program modules, and program data may include an implementation of a networking environment. The program modules 42 generally perform the functions and / or methods of embodiments of the present invention as described herein.

[0054] The computer system / server 12 may also communicate with one or more external devices 14, such as a keyboard, a pointing device, a display 24, etc.; one or more devices that enable a user to interact with the computer system / server 12; and / or any device that enables the computer system / server 12 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may occur via an input / output (I / O) interface 22. In addition, the computer system / server 12 may communicate with one or more networks such as a local area network (LAN), a general wide area network (WAN), and / or a public network (e.g., the Internet) via a network adapter 20. As shown, the network adapter 20 communicates with other components of the computer system / server 12 via a bus 18. It should be understood that, although not shown, other hardware and / or software components may be used in conjunction with the computer system / server 12. Examples include, but are not limited to, microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data archival storage systems, etc.

[0055] Reference now Figure 2, depicts an illustrative cloud computing environment 50. As shown, the cloud computing environment 50 includes one or more cloud computing nodes 10 with which a local computing device used by a cloud consumer (e.g., a personal digital assistant (PDA) or cellular phone 54A, a desktop computer 54B, a laptop computer 54C, and / or an automobile computer system 54N) can communicate. The nodes 10 can communicate with each other. They can be physically or virtually grouped (not shown) in one or more networks (such as a private cloud, community cloud, public cloud, or hybrid cloud, or a combination thereof, as described above). This allows the cloud computing environment 50 to provide infrastructure, platforms, and / or software as a service that cloud consumers do not need to maintain resources on local computing devices. It should be understood that Figure 2 The types of computing devices 54A-54N shown in are intended to be illustrative only, and computing node 10 and cloud computing environment 50 may communicate with any type of computerized device over any type of network and / or network addressable connection (eg, using a web browser).

[0056] Reference now Figure 3 , showing the cloud computing environment 50 ( Figure 2 ) provides a set of functional abstraction layers. It should be understood in advance that Figure 3 The components, layers and functions shown in are intended to be illustrative only, and embodiments of the present invention are not limited thereto. As shown, the following layers and corresponding functions are provided:

[0057] The hardware and software layer 60 includes hardware and software components. Examples of hardware components include: host 61; server 62 based on RISC (Reduced Instruction Set Computer) architecture; server 63; blade server 64; storage device 65; and network and networking components 66. In some embodiments, software components include network application server software 67 and database software 68.

[0058] Virtualization layer 70 provides an abstraction layer from which the following examples of virtual entities can be provided: virtual servers 71 ; virtual storage 72 ; virtual networks 73 , including virtual private networks; virtual applications and operating systems 74 ; and virtual clients 75 .

[0059] In one example, the management layer 80 may provide the functionality described below. Resource provisioning 81 provides dynamic procurement of computing resources and other resources for performing tasks within a cloud computing environment. Metering and pricing 82 provides cost tracking when utilizing resources within a cloud computing environment, as well as billing or invoicing for consuming these resources. In one example, these resources may include application software licenses. Security provides authentication for cloud consumers and tasks, as well as protection for data and other resources. User portal 83 provides access to the cloud computing environment for consumers and system administrators. Service level management 84 provides cloud computing resource allocation and management so that the required service levels are met. Service level agreement (SLA) planning and fulfillment 85 provides pre-arrangement and procurement of cloud computing resources in anticipation of future needs according to the SLA.

[0060] The workload layer 90 provides examples of functions that can utilize a cloud computing environment. Examples of workloads and functions that can be provided from this layer include: mapping and navigation 91; software development and lifecycle management 92; virtual classroom education delivery 93; data analysis processing 94; transaction processing 95; and fluid flow prediction 96.

[0061] Embodiments of the present invention may include Figure 1 A computer system / server 12, wherein one or more program modules 42 are configured to execute (or cause the computer system / server 12 to execute) Figure 3 One or more functions of fluid flow prediction 96 of the present invention. For example, one or more of the program modules 42 can be configured to: access a capillary network representation, wherein the capillary network representation is a three-dimensional (3D) network of interconnected capillaries representing the interconnected pore structure of a 3D micro-CT scanner model of a porous media sample; generate a set of simplified network representations from the capillary network representation, wherein each simplified network representation in the set of simplified network representations includes a two-dimensional (2D) or 3D network of interconnected capillary structures representing a subset of the interconnected capillary structures of the capillary network representation; perform fluid flow simulations using a physics-based simulator to determine simulated fluid flow properties for each simplified network representation in the simplified network representations; train a NN model using the set of simplified network representations as input and using the simulated fluid flow properties as model targets, thereby generating a trained NN model for predicting fluid flow properties; and input a new capillary network representation of a new porous media sample into the trained NN model, thereby generating predicted fluid flow properties for the new porous media sample.

[0062] Figure 4A block diagram of an exemplary environment 400 according to aspects of the present invention is shown. In an embodiment, the environment 400 includes a network 402 that enables communication between one or more of the following: a server 404, a micro-computed tomography (micro-CT) scanner 406, a micro-CT scanner 406', and one or more client devices 408. The server 404, the micro-CT scanner 406, the micro-CT scanner 406', and the one or more client devices 408 may each include Figure 1 The server 404, the microCT scanner 406, the microCT scanner 406' and the one or more client devices 408 may each be a Figure 2 The computing node 10 in the cloud computing environment 50 of FIG. 4 is a computing node 10 in the cloud computing environment 50 of FIG. 4. For example, one or more client devices 408 can be provided by Figure 2 A local computing device (e.g., a PDA or cellular phone 54A, a desktop computer 54B, or a laptop computer 54C) used by a cloud consumer in the cloud computing environment 50.

[0063] Embodiments of the present invention utilize a micro-CT scanner 406 that communicates with a remote server 404 via a network 402. In an embodiment, the server 404 provides a cloud-based digital image analysis service to users of the environment 400 based on x-ray micro-tomography results from one or more remote micro-CT scanners 406. Other embodiments of the present invention utilize a micro-CT scanner 406' to generate x-ray micro-tomography results and perform digital image analysis according to the methods of the present invention.

[0064] In an embodiment, the server 404 includes one or more modules, each of which may include one or more program modules, such as Figure 1 The program module 42 is described. Figure 4 In the example of , the server 404 includes a data collection module 410, a network extraction module 411, a network simplifier module 412, a neural network (NN) module 413, a flow simulation module 414, a flow prediction module 415, and a data storage module 416, each of which may include, for example Figure 1 One or more program modules 42.

[0065] In an embodiment, the micro CT scanner 406 and / or the micro CT scanner 406′ includes one or more modules, each of which may include one or more program modules, such as Figure 1 The program module 42 is described. Figure 4 In the example of FIG. 4 , the micro CT scanner 406 includes an imaging module 420 and a data storage module 421 , each of which may include, for example Figure 1One or more program modules 42. In an embodiment, the imaging module 420 is configured to generate digital image data using an x-ray micro-tomography method and store the digital image data (e.g., a 3D model of the pore space within the physical porous medium) in the data storage module 421.

[0066] In an embodiment, the micro-CT scanner 406 and / or micro-CT scanner 406' comprises a dedicated computing device configured to generate and process a 3D model of a physical object derived from x-ray micro-tomography. Typically, x-ray micro-tomography or micro-CT uses x-rays to create a digital image of a cross-section of a physical object, which can be used to reconstruct a virtual model (3D model) without destroying the original object. The prefix "micro" is used to indicate that the pixel size of the cross-section is in the micron range. Various micro-CT scanning systems may be utilized in accordance with embodiments of the present invention, and embodiments of the present invention are not intended to be limited to a particular micro-CT scanning system.

[0067] In an embodiment, the data collection module 410 of the server 404 is configured to obtain digital image data (e.g., a 3D model) from the micro-CT scanner 406 (e.g., from the imaging module 420 or the data storage module 421). In an embodiment, the network extraction module 411 of the server 404 is configured to generate a digital capillary network representation of the 3D model using a network extraction routine, wherein the digital capillary network representation includes a network of interconnected capillaries based on pore spaces in the 3D model.

[0068] In an embodiment, the network simplifier module 412 of the server 404 is configured to generate multiple simplified network representations based on the capillary network representation using a network simplification routine, wherein each simplified network representation includes a plurality of interconnected capillaries, the number of which is less than the number of interconnected capillaries in the capillary network representation.

[0069] In various aspects, the NN module 413 is configured to create and train a NN model using a plurality of simplified networks as inputs and using simulated flow properties determined from the simplified networks as model targets. In an embodiment, the flow simulation module 414 of the server 404 is configured to determine one or more simulated flow properties of a porous media sample by simulating fluid flow through pore spaces of the porous media under predetermined conditions.

[0070] In an embodiment, the flow prediction module 415 is configured to generate predicted flow properties of the porous media sample as an output of the NN model by inputting the capillary network representation of the porous media sample into the trained NN model. In an embodiment, the NN module 413 is also configured to verify the NN model by comparing the simulated flow properties with the predicted flow properties, wherein the divergence of the simulated flow properties from the predicted flow properties indicates inaccuracies in the NN model. In an embodiment, the server 404 stores the generated data in the data storage module 416, such as the predicted flow properties, the simulated flow properties, the 3D model, and the NN model.

[0071] In an embodiment, the micro-CT scanner 406' performs the functions of both the micro-CT scanner 406 and the server 404. In an embodiment, the micro-CT scanner 406' includes one or more modules, each of which may include one or more program modules, such as Figure 1 The program modules 42 are described. In aspects of the present invention, the micro-CT scanner 406' includes modules corresponding to the modules of the micro-CT scanner 406 and the server 404. Figure 4 In the example of FIG. 4 , the micro CT scanner 406′ includes an imaging module 420′ and a data storage module 421′, whose functions correspond to the corresponding modules 420 and 421 of the micro CT scanner 406. In addition, the micro CT scanner 406′ includes a data collection module 410′, a network extraction module 411′, a network simplifier module 412′, a neural network (NN) module 413′, a flow simulation module 414′, a flow prediction module 415′, and a data storage module 416′, whose functions correspond to the corresponding modules 410-416 of the server 404.

[0072] The server 404, the micro-CT scanner 406, the micro-CT scanner 406' and the one or more client devices 408 may each include Figure 4 The modules shown in the embodiment may be more or less than the modules shown in the embodiment. In an embodiment, the individual modules may be integrated into a single module. Additionally or alternatively, a single module may be implemented as multiple modules. In addition, the number of devices and / or networks in the environment 400 is not limited to Figure 4 In practice, environment 400 may include additional devices and / or networks; fewer devices and / or networks; different devices and / or networks; or Figure 4 Different arrangements of devices and / or networks are shown.

[0073] Figure 5 is a flow chart providing an overview of a porous media modeling system according to an embodiment of the present invention. Figure 5 The steps shown in Figure 4 Executed in an environment and referenced Figure 4The elements depicted in are described.

[0074] In an embodiment, a porous media sample (e.g., a rock sample) 500 is scanned by a micro-CT scanner 406 to generate a 3D model 501 of the sample. In an embodiment, a network extraction module 411 of a server 404 generates a capillary network representation 502 of the 3D model 501 using a network extraction routine. The term capillary network representation as used herein refers to a 3D network of interconnected capillary structures, wherein the interconnected capillary structures represent the interconnected pore structure of the micro-CT scanner 3D model (e.g., 3D model 501). The network simplifier module 412 of the server 404 utilizes the capillary network representation 502 to generate a plurality of simplified network representations indicated at 503, wherein the number of capillaries each of the simplified network representations 503 includes is less than the number of capillaries in the capillary network representation 502. The number of simplified network representations generated can vary. In an embodiment, the number of simplified network representations generated depends on how many simplified network representations are required to generate accurate flow predictions, which is determined by validating different iterations of the flow prediction method. As an example, according to an embodiment of the present invention, 50-100 simplified network representations 503 may be generated to obtain accurate flow predictions.

[0075] As used herein, the term simplified network representation refers to a two-dimensional (2D) or 3D network of interconnected capillary structures that represents a subset of the interconnected capillary structures of the complete capillary network representation 502. In an embodiment, each simplified network representation 503 includes a capillary network that is much smaller than the capillary network representation 502 and is optimized to match the fluid-related morphological properties of the capillary network representation 502 (e.g., porosity, capillary diameter distribution, etc.).

[0076] Continue to refer Figure 5 , the flow simulation module 414 of the server 404 determines the simulated flow properties 504 of each simplified network representation 503. The NN module 413 trains the NN model using the simplified network representation 503 and the simulated flow properties 504. The flow prediction module 415 determines the predicted flow properties 505 of the porous media sample 500 by inputting the capillary network representation 502 into the trained NN model. In an embodiment, the predicted flow properties 505 are output from the NN model as a numerical representation of the flow properties of the porous media sample 500. Figure 6 600 depicts a visualization of numerical representations of flow properties.

[0077] Figure 6 is an illustrative example of predicted flow properties generated according to an embodiment of the present invention. Figure 6In the example of FIG. 5 , a 3D model 501 of the porosity of a porous media sample 500 is used to generate a capillary network representation 502, which includes capillaries with diameters between 5 μm and 125 μm generated based on the porosity of the 3D model 501. According to an embodiment of the present invention, the capillary network representation 502 is input into a trained NN model, which generates predicted flow properties in the form of a permeability and porosity map 600.

[0078] Figure 7 is a flowchart illustrating the training of a NN model according to an embodiment of the present invention. Figure 7 The steps can be Figure 4 Executed in an environment and referenced Figure 4 and Figure 5 The elements depicted in are described.

[0079] like Figure 7 As shown, the porous media sample 500 is input into and scanned by the microCT scanner 406, which outputs a 3D model 501 of the porous media sample 500, which is input into the network extraction module 411 to generate a capillary network representation 502. During the training of the NN model 700, the network simplifier module 412 generates a plurality of simplified network representations indicated at 503A-503N. In an embodiment, for each simulated network representation 503A-503N, a fluid flow simulation is performed at 701A-701N (e.g., by the flow simulation module 414) to generate simulated flow properties 504A-504N of the corresponding simulated network representation 503A-503N. In one example, fifty (50) simulated network representations 503A-503N are generated by the network simplifier module 412 and analyzed by the flow simulation module 414 to determine the simulated permeability and simulated porosity of each simulated network representation 503A-503N.

[0080] In an embodiment, the capillary network representation 502 is used as an input to the flow simulation module 414 to generate simulated flow properties 704 of the capillary network representation. The capillary network representation 502 is also used as an input to the trained NN model 700 by the flow prediction module 415 to generate predicted flow properties 505 of the capillary network representation 502.

[0081] Continue to refer Figure 7, in an embodiment, each simplified network representation 503A-503N is used as an input to training the NN model 700 at 703. Additionally, the simulated flow properties 504A-504N are used as model target inputs during model training 703. Thus, embodiments of the present invention utilize multiple simplified network representations 503A-503N (e.g., 50-100 representations) and their associated simulated flow properties to train the NN model 700. At step 706, validation of the trained NN model 700 is performed (e.g., by the NN module 413) by comparing the simulated flow properties 704 of the capillary network representation 502 to the predicted flow properties 505 to see how much they vary, where smaller variations indicate a more accurate NN model 700.

[0082] Figure 8 Neural network training for fluid flow prediction according to an embodiment of the present invention is shown. Figure 8 The steps can be Figure 4 Executed in an environment and referenced Figure 4 and Figure 5 The elements depicted in are described.

[0083] Figure 8 A set of N simplified network representations 503A generated by the server 404 or the micro-CT scanner 406' from the capillary network representation 502A is shown. In this example, the number of nodes of the capillary network representation 502A is 1,885,585, and the number of links of the capillary network representation 502A is 2,763,271, where a node is an intersection between capillaries and a link is a capillary connecting two nodes in the capillary network. The number of nodes of the set of N simplified network representations 503A is 1394, and the number of links is 1846. It can be seen that the number of nodes in the capillary network representation 502A is 3 orders of magnitude greater than the number of nodes in the set of N simplified network representations 503A. Similarly, the number of links in the capillary network representation 502A is 3 orders of magnitude greater than the number of links in the set of N simplified network representations 503A.

[0084] Simulated flow properties 800 are generated by the server 404 or the micro-CT scanner 406' from a plurality of network representations 503A of a first fluid (fluid 1) and a second fluid (fluid 2). The server 404 or the micro-CT scanner 406' identifies morphological properties of interest 801, as well as flow conditions 802 and flow properties 803 (simulated flow properties). The flow properties 803 are used as model targets in the training of the NN model 700. In this example, the kth network representation is set to k=1...N, and each jth network link 804 is represented by a vector Characterization. According to vector, using the surrounding link vectors (where i=1...M) performs M convolutions To generate the convolutional layer 805, which is then performed by The pooling operation represented by 404 is used to generate the pooling layer 806. The server 404 or the microCT scanner 406' is convolutional and pooling The NN model 700 is then generated, as shown in 807, where k=1...N is the number of simplified networks, and j=1...P is the number of pooling elements of each network.

[0085] Fig. 9 Fluid flow prediction using a trained NN model according to an embodiment of the present invention is shown. Fig. 9 The steps can be Figure 4 Executed in an environment and referenced Figure 4 and Figure 5 The elements depicted in are described.

[0086] Fig. 9 A set of N simplified network representations 503B generated by the server 404 or the micro-CT scanner 406' from the capillary network representation 502B of a new porous rock sample is shown. In this example, the capillary network representation 502B includes 1,885,585 nodes and 2,763,271 links, and the set of N simplified capillary network representations 503B includes 1394 nodes and 1846 links. The simplified network representation 503B depicts morphological properties 900 based on certain flow conditions 901. In this example, the kth network representation is set to k=1...N, and each jth network link is represented by the vector Characterization. The application of the trained NN model 700 to the set of N simplified capillary network representations 503B is depicted at 902 (convolutions and pooling not shown), where each network has k=1...N simplified networks and j=1...P pooling elements. The predicted flow properties 905 of a new rock sample are determined by aggregating the results (predicted flow property outputs) 904 from the trained NN model 700, as shown at 906.

[0087] Fig.10 A flow chart of an exemplary method according to various aspects of the present invention is shown. The steps of the method may be Figure 4 Executed in an environment and referenced Figure 4 , Figure 5 and Figure 7 Initially, it should be noted that Fig.10 The steps may be implemented by a micro-CT scanner 406' that has been configured to perform the steps according to an embodiment of the present invention, or may be implemented by a server 404 in communication with a micro-CT scanner 406 (eg, a conventional micro-CT scanner).

[0088] In step 1000, the micro-CT scanner 406 or the micro-CT scanner 406' generates a 3D image representing the spatial distribution of the pore space of the porous media sample (e.g., rock sample) 500. In an embodiment, the 3D image is a 3D image cube 501, which shows the micron-scale morphology (e.g., pores) of the interior of the porous media sample 500. In an embodiment, the micro-CT scanner 406 or 406' reconstructs the 3D image from consecutive cross-sectional images of the porous media sample 500, where the cross-sections are generated by transmission x-rays that pass through the porous media sample 500 to a 2D area detector. Various types of micro-CT scanners may be utilized in the embodiment of step 1000, and embodiments of the present invention are not intended to be limited to a particular micro-CT scanner. In an embodiment, the imaging module 420 of the micro-CT scanner 406 or the imaging module 420' of the micro-CT scanner 406' implements step 1000.

[0089] At step 1001, the server 404 or the micro-CT scanner 406' generates a complete capillary network representation 502 of the pore space of the porous media sample 500 based on the 3D image (e.g., 3D image cube 501) generated at step 1000. Generally, the complete capillary network representation 502 represents the pore space of the porous media sample 500 through geometric primitives (e.g., cylinders and spheres), and its flow properties can be obtained. In an embodiment, the server 404 receives the 3D image from the micro-CT scanner 406 before performing step 1001. For example, the data collection module 410 of the server 404 can collect the 3D image generated by the imaging module 420 from the data storage module 421 of the micro-CT scanner 406 via the network 402.

[0090] In an embodiment, the network extraction module 411 or the network extraction module 411' utilizes a network extraction routine to process the 3D image generated in step 1000 to produce a complete capillary network representation 502 that represents the complete pore space of the 3D image. Typically, the network extraction routine utilizes image processing techniques to process the 3D image to remove pores that are not connected to the network of interconnected pores comprising the flow path, and depicts the network of interconnected pores of the 3D image as capillaries having a diameter.

[0091] An example of a network extraction routine includes the steps of converting the 3D image to 8-bit grayscale, reducing image noise using a filter, and segmenting the noise-reduced grayscale image into solid and void spaces to obtain a binary image. In an embodiment, the network extraction routine also includes processing the binary image for morphological analysis and eliminating pore voxels that are not connected to the percolation network. In an embodiment, the processed binary image data containing connected pore voxels is then input into a network extraction algorithm, such as a pore network model (PNM), a simplified maximum sphere model (RMB), or a capillary network model (CNM), to generate a complete capillary network representation 502.

[0092] At step 1002, the server 404 or the micro-CT scanner 406' determines and stores at least one simulated flow property of the porous media sample 500 based on the capillary network representation 502. In an embodiment, the server 404 or the micro-CT scanner 406' initiates a physics-based flow simulation routine on the capillary network representation 502 using the flow simulation module 414 or 414' to simulate the flow of a predetermined fluid through the capillaries of the capillary network representation 502 under predetermined conditions (e.g., pressure, temperature, etc.) and generate flow properties of interest, such as porosity, permeability, flow rate, residual fluid saturation, etc. The at least one simulated flow property may be stored by the server 404 or the micro-CT scanner 406' for later use during validation of the trained NN module 700 (e.g., in the data storage module 416 or 416').

[0093] The server 404 or the micro-CT scanner 406' can utilize various physics-based flow simulation routines to implement step 1002, and the present invention is not intended to be limited to a particular flow simulation routine. In an embodiment, the physics-based flow simulation routine operates on the nodes and links of the capillary network representation 502. In one example, the flow simulation routine applies Poiseuille's law to the links and applies the law of conservation of mass to the internal nodes while maintaining a fixed pressure difference between the inlet and outlet boundary nodes. For all nodes i, this is given by the mass conservation equation Σ j Qi j = 0 system representation, where Qi j =(πRi j 4 / 8μLi j )(Pi-P j ) is the flow rate in the capillary connecting node i to node j. The geometric parameters R and L represent the radius and length of the capillary (link) connecting two nodes of the network, respectively, and μ is the dynamic viscosity of the fluid. In one example, a viscosity of μ=1 cP is used, and a pressure gradient of 10 kPa / m is applied along the flow direction in addition to atmospheric pressure. In this embodiment, permeability is calculated using Darcy's law Q=K(A / μL)ΔP.

[0094] At step 1003, the server 404 or the micro-CT scanner 406' generates a set of simplified network representations 503 based on the complete capillary network representation 502, wherein each simplified network representation 503 includes fewer capillaries than the complete capillary network representation 502. In an embodiment, at step 1003, the network simplifier module 412 of the server 404 or the network simplifier module 412' of the micro-CT scanner 406' implements a network simplification routine to generate the plurality of simplified network representations 502. In an embodiment, the plurality of simplified network representations 502 are substantially smaller than the complete capillary network representation 502.

[0095] Examples of network simplification routines that may be utilized by the server 404 or the micro-CT scanner 406 ′ include a capillary bundle routine, a regular capillary network routine, and a random capillary network routine.

[0096] Capillary Bundle Routine

[0097] Capillary bundles of length Lx are distributed in a sample of dimensions Lx, Ly, and Lz. The sample volume is defined as the product LxLyLz. Capillary diameters are assigned by randomly selecting diameters from a diameter probability distribution function. The diameter probability distribution function is obtained from the complete capillary network representation 502. Once the above steps are completed, the porosity of each simplified network representation is calculated in step 1004 by dividing the capillary volume by the sample volume. The capillary bundle routine repeats these steps until a porosity that is very similar to the porosity of the complete capillary network representation 502 is reached. In this sense, the set of simplified network representations 503 retains the following properties: (i) porosity; and (ii) capillary diameter distribution.

[0098] Regular Capillary Network Routine

[0099] In the regular capillary network routine, a 2D or 3D cubic complete capillary network representation 502 is created. The intersections between the capillaries (called nodes) display 4 or 8 in 2D, or 6 or 12 coordination numbers (the number of intersecting capillaries) in 3D, depending on the regular network selected. In these simplified capillary networks, the capillary length is fixed (because it is a regular network). Other parameters such as capillary diameter and node coordination number follow the same probability distribution function as the complete capillary network representation 502 of the sample medium. In order to meet these requirements, the capillary diameter is randomly assigned by selecting the diameter from the probability distribution function. A similar process is completed to obtain the node coordination number. Since the coordination number is a priori fixed by selecting a simplified regular network, some capillaries are deleted to match the coordination number probability distribution function. Once all the above steps are completed, the porosity of the group of simplified network representations 503 is calculated by dividing the capillary volume by the sample volume. The regular capillary network routine repeats these steps until a porosity is achieved that is very similar to the porosity of the original complete capillary network representation 502 of the sample medium. In this sense, the simplified network retains the following properties: (i) porosity; (ii) capillary diameter distribution; and (iii) node coordination number distribution.

[0100] Random Capillary Network Routine

[0101] In the random capillary network routine, a set of 2D or 3D points (or nodes) are randomly distributed in a 2D or 3D space defined by Lx, Ly and Lz (if 3D). Then, these nodes are connected. The connection between these nodes is called a capillary. Once the nodes are randomly distributed, the capillary length does not follow the capillary length frequency distribution function as the original complete capillary network representation 502 of the sample medium. On the other hand, the parameters capillary diameter and node coordination number follow the same frequency distribution function as the original complete capillary network representation 502 of the sample medium. In order to generate a simplified random capillary network, the capillary diameter is assigned by randomly selecting the diameter from the diameter frequency distribution function. A similar process is completed to obtain the node coordination number. Since the coordination number is a priori fixed by the number of connections (or capillaries) between nodes, some capillaries are deleted to match the coordination number frequency distribution function. Once all the above steps are completed, the porosity of each simplified network is calculated by dividing the capillary volume by the sample volume. The random capillary network routine repeats these steps until a porosity value is reached that is within a predetermined error range of the porosity value of the original complete capillary network representation 502 of the sample medium. In this sense, the simplified network retains the following properties: (i) porosity; (ii) capillary diameter distribution; and (iii) node coordination number distribution.

[0102] Continue to refer Fig.10At step 1004, the server 404 or the micro-CT scanner 406' determines and stores at least one simulated flow property 504 for each simplified network representation in the set of simplified network representations 503. In an embodiment, the server 404 or the micro-CT scanner 406' initiates a physics-based flow simulation routine on each simplified network representation using the flow simulation module 414 or 414' to simulate the flow of a predetermined fluid through the capillaries of the corresponding simplified network representation under predetermined conditions (e.g., pressure, temperature, etc.) and generate flow properties of interest, such as porosity, permeability, flow rate, and residual fluid saturation. The at least one simulated flow property 504 may be stored by the server 404 or the micro-CT scanner 406'. The server 404 or the micro-CT scanner 406' may utilize a variety of physics-based flow simulation routines to implement step 1004, and the present invention is not intended to be limited to a particular flow simulation routine.

[0103] At step 1005, the server 404 or the micro-CT scanner 406' uses the set of simplified network representations 503 and the determined simulated flow properties 504 of the simplified network representations as training inputs to create and / or train the NN model 700. In an embodiment, the NN module 413 of the server 404 or the NN module 413' of the micro-CT scanner 406' creates and / or trains the NN model 700 using the set of simplified network representations 503 as model inputs and using their simulated flow properties as model targets.

[0104] It should be understood that it is impractical to require thousands of complete capillary network flow simulations for training purposes to determine the flow characteristics of the NN model 700, because it would be cheaper and faster (from a computer resource perspective) to perform the flow simulation directly on the capillary network (e.g., the capillary network representation 502). Embodiments of the present invention solve this technical problem by providing a method for training the NN model 700 using a smaller simplified capillary network (e.g., the simplified network representation 503) to construct a training data set for the NN model 700, which can later be applied to the complete capillary network representation at runtime. From a computer resource perspective, performing flow simulations in a simplified capillary network is much cheaper than running flow simulations directly in the complete capillary network.

[0105] At step 1006, the server 404 or the micro-CT scanner 406' generates predicted flow properties of the porous media sample 500 by inputting the complete capillary network representation 502 into the NN model 700. In an embodiment, the flow prediction module 415 of the server 404 or the flow prediction model 415' of the micro-CT scanner 406' feeds the capillary network representation 502 to the NN model 700, and the evaluation routine of the NN model 700 outputs the predicted flow properties.

[0106] At step 1007, the server 404 or the micro-CT scanner 406' validates the NN model 700 by comparing the simulated flow properties determined at step 1004 and the predicted flow properties determined at step 1006 to determine the change. In an embodiment, when the change between the simulated flow properties and the predicted flow properties exceeds a predetermined threshold, the server 404 or the micro-CT scanner 406' determines that the NN model 700 needs to be modified and / or further trained. In an embodiment, the NN module 413 of the server 404 or the NN module 413' of the micro-CT scanner 406' implements step 1007.

[0107] In an embodiment, if the server 404 or the micro-CT scanner 406' determines that the NN model 700 is invalid, steps 1003-1007 may be repeated until the change between the simulated flow properties and the predicted flow properties determined at step 1004 is below a predetermined threshold. Once the NN model 700 is determined to be valid, the method may proceed to step 1008.

[0108] At step 1008 , the server 404 or the micro-CT scanner 406 ′ generates a new capillary network representation of the new porous media sample. Step 1008 may be performed using the same method as step 1001 .

[0109] At step 1009 , the server 404 or the micro-CT scanner 406 ′ inputs the new capillary network representation generated at step 1008 into the trained NN model 700 to generate predicted flow properties of a new porous media sample as output.

[0110] At step 1010 , the server 404 or the micro CT scanner 406 ′ aggregates the predicted flow properties of the new porous media sample to generate a report of the flow properties of the new porous media sample. In an embodiment, the flow prediction module 415 of the server 404 or the flow prediction module 415 ′ of the micro CT scanner 406 ′ implements step 1010 .

[0111] At step 1011, the server 404 or the micro-CT scanner 406' displays the report to the user via the user interface of the server 404 or the user interface of the micro-CT scanner 406'. In an embodiment, the flow prediction module 415 of the server 404 or the flow prediction module 415' of the micro-CT scanner 406' implements step 1011. In aspects of the present invention, the server 404 or the micro-CT scanner 406' provides the NN model results to one or more client devices 408 via the network 402 for user interface display.

[0112] Now refer to Fig.10Discussing an exemplary use scenario. According to step 1000, the porous medium sample (rock sample) 500 is imaged at a sub-microscopic resolution by the imaging module 420 of the micro-CT scanner 406, thereby capturing a volume of 10 mm 3 According to step 1001, the server 404 obtains the 3D model 501, and the network extraction module 411 generates a capillary network representation 502 based on the 3D model 501. In this example, the capillary network representation 502 contains millions (10 6 -10 8 ) capillaries, making the complete flow simulation of the capillary network representation 502 by the flow simulation module 414 infeasible or computationally expensive. According to step 1003, the network simplifier module 412 of the server 404 generates a network representation including thousands (10 3 -10 5 ) simplified capillary networks, which contain a reduced number of capillaries (10 2 -10 4 ). The flow simulation module 414 of the server 404 performs flow simulation on the set of simplified network representations 503 according to step 1004 , which has a much smaller computational cost than performing flow simulation on the capillary network representation 502 .

[0113] In this example, multiphase fluid properties (such as viscosity, density, interfacial tension, contact angle, temperature, pressure, and boundary conditions) are provided as input to the flow simulation algorithm of the flow simulation module 414. For each instance of thousands of simplified network representations 503, a flow simulation is performed by the flow simulation module 414, and the relevant properties of interest, such as absolute permeability or relative permeability, are extracted. Based on this compilation of simplified networks (a set of simplified network representations 503) and their simulated flow properties 504, the NN model 700 is trained according to step 1005. A suitable architecture for the NN model 700 is a graph convolutional neural network (GCNN). In an embodiment, the trained GCNN model can be used to replace the complete flow simulation performed by the flow simulation module 414, and the desired flow properties (e.g., predicted flow properties 505) will be predicted based on the complete capillary network representation 502 and multiphase fluid properties or based on a set of simplified network representations 503 of the complete capillary network representation 502.

[0114] In an embodiment, a service provider may provide for performing the processes described herein. In this case, the service provider may create, maintain, deploy, support, etc., a computer infrastructure that performs the processing steps of the present invention for one or more customers. These customers may be, for example, any business that uses the technology. In return, the service provider may receive payment from the customer under a subscription and / or fee agreement, and / or the service provider may receive payment from the sale of advertising content to one or more third parties.

[0115] In another embodiment, the present invention provides a computer-implemented method via a network. In this case, a computer infrastructure, such as a computer system / server 12 ( Figure 1 ), and one or more systems for performing the processes of the present invention may be obtained (e.g., created, purchased, used, modified, etc.) and deployed to a computer infrastructure. To this end, deployment of the system may include one or more of the following: (1) downloading from a computer readable medium to a computer system / server 12 (e.g., Figure 1 (a) installing program code on a computing device (as shown); (2) adding one or more computing devices to a computer infrastructure; and (3) combining and / or modifying one or more existing systems of the computer infrastructure to enable the computer infrastructure to perform the process of the present invention.

[0116] The description of various embodiments of the present invention has been presented for the purpose of illustration, but is not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, practical applications, or technical improvements over technologies found in the marketplace, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method comprising: A capillary network representation of a porous media sample is accessed by a computing device; generating, by the computing device, a set of simplified network representations from the capillary network representation; performing, by the computing device, a fluid flow simulation using a simulator to determine simulated fluid flow properties for each of the simplified network representations; as well as The computing device trains a neural network (NN) model using the set of simplified network representations as input and the simulated fluid flow properties as model targets, thereby generating a trained NN model for predicting the fluid flow properties of the porous media sample.

2. The method of claim 1 , further comprising obtaining, by the computing device, a 3D micro-CT scanner model from a remote micro-CT scanner via a network connection.

3. The method of claim 1, further comprising generating, by the computing device, a 3D micro-CT scanner model by imaging the porous media sample using micro-CT.

4. The method of claim 1 , further comprising generating, by the computing device, the capillary network representation from a 3D micro-CT scanner model.

5. The method according to claim 1, further comprising: determining, by the computing device, predicted fluid flow properties of the porous media sample by inputting the capillary network representation into the trained NN model; aggregating, by the computing device, the simulated fluid flow properties of each of the simplified network representations to obtain a simulated fluid flow property result of the porous medium sample; as well as The trained NN model is validated by the computing device by comparing the simulated fluid flow property results of the porous media sample with the predicted fluid flow properties of the porous media sample.

6. The method according to claim 1, further comprising: The computing device inputs a new capillary network representation of the new porous media sample into the trained NN model to generate predicted fluid flow properties of the new porous media sample, wherein the new capillary network representation includes a single capillary network representation of the complete new porous media sample, or a set of simplified capillary network representations of the new porous media sample.

7. The method of claim 1, wherein the simulated fluid flow properties and the predicted fluid flow properties include one or more fluid flow properties selected from the group consisting of: porosity, permeability, flow rate, and residual fluid saturation.

8. The method of claim 1, wherein the computing device includes software provided as a service in a cloud environment.

9. A computer program product comprising one or more computer-readable storage media having program instructions stored together on the one or more computer-readable storage media, the program instructions being executable to: accessing a capillary network representation, wherein the capillary network representation is a three-dimensional 3D network of interconnected capillaries representing an interconnected pore structure of a 3D micro-CT scanner model of a porous media sample; generating a set of simplified network representations from the capillary network representation, wherein each simplified network representation in the set of simplified network representations comprises a two-dimensional (2D) or 3D network of interconnected capillary structures representing a subset of the interconnected capillary structures of the capillary network representation; performing a fluid flow simulation using a physics-based simulator to determine simulated fluid flow properties for each of the simplified network representations; and A neural network (NN) model is trained using the set of simplified network representations as input and using the simulated fluid flow properties as model targets, thereby generating a trained NN model for predicting fluid flow properties.

10. The computer program product of claim 9, wherein the program instructions are further executable to obtain the 3D micro-computed tomography (micro-CT) scanner model from a remote micro-CT scanner via a network connection.

11. The computer program product of claim 9, wherein the program instructions are further executable to generate the 3D micro-CT scanner model by imaging the porous media sample with micro-CT.

12. The computer program product of claim 9, wherein the program instructions are further executable to generate the capillary network representation from the 3D micro-CT scanner model.

13. The computer program product of claim 9, wherein the program instructions are further executable to: determining predicted fluid flow properties of the porous media sample by inputting the capillary network representation into the trained NN model; aggregating the simulated fluid flow properties of each of the simplified network representations to obtain simulated fluid flow property results of the porous medium sample; as well as The trained NN model is validated by comparing the simulated fluid flow property results of the porous media sample with the predicted fluid flow properties of the porous media sample.

14. A computer program product according to claim 9, wherein the program instructions are further executable to input a new capillary network representation of a new porous media sample into the trained NN model to generate predicted fluid flow properties of the new porous media sample, wherein the new capillary network representation includes a single capillary network representation of the complete new porous media sample or a set of simplified capillary network representations of the new porous media sample.

15. The computer program product of claim 9, wherein the set of simplified network representations has a plurality of interconnected capillary structures that is an order of magnitude smaller than the interconnected capillary structures of the capillary network representation.

16. A system comprising: A processor, a computer readable memory, one or more computer readable storage media, and program instructions stored together on the one or more computer readable storage media, the program instructions being executable to: accessing a capillary network representation, wherein the capillary network representation is a three-dimensional 3D network of interconnected capillaries representing an interconnected pore structure of a 3D micro-CT scanner model of a porous media sample; generating a set of simplified network representations from the capillary network representation, wherein each simplified network representation in the set of simplified network representations comprises a two-dimensional (2D) or 3D network of interconnected capillary structures representing a subset of the interconnected capillary structures of the capillary network representation; performing a fluid flow simulation using a physics-based simulator to determine simulated fluid flow properties for each of the simplified network representations; Training a neural network (NN) model using the set of simplified network representations as input and using the simulated fluid flow properties as model targets, thereby generating a trained NN model for predicting fluid flow properties; as well as A new capillary network representation of a new porous media sample is input into the trained NN model to generate predicted fluid flow properties of the new porous media sample.

17. The system of claim 16, wherein the program instructions are further executable to display, by the computing device, the predicted fluid flow property to a user via a user interface.

18. The system of claim 16, wherein the program instructions are further executable to: generating the 3D micro-CT scanner model by imaging the porous media sample using micro-CT; and The capillary network representation is generated from the 3D micro-CT scanner model.

19. The system of claim 16, wherein the program instructions are further executable to: determining predicted fluid flow properties of the porous media sample by inputting the capillary network representation into the trained NN model; aggregating the simulated fluid flow properties of each of the simplified network representations to obtain simulated fluid flow property results of the porous medium sample; as well as The trained NN model is validated by comparing the simulated fluid flow property results of the porous media sample with the predicted fluid flow properties of the porous media sample.

20. The system of claim 16, wherein the set of simplified network representations has a plurality of interconnected capillary structures that are an order of magnitude smaller than the interconnected capillary structures of the capillary network representation.

Citation Information

Cited By

  • Neural network predictions of fluid flow in porous media

    US12657358B2