A method for simulating multi-field evolution of oil-rich coal in-situ self-supply heat targeted cracking

Abstract

The application provides a kind of oil-rich coal in-situ self-heat targeted cleavage multi-field evolution simulation method, first through the dynamic assimilation of multi-source discrete data and self-adaptive inversion operator, reconstruct three-dimensional dynamic control field, accurately capture the real-time evolution state of oxygen-consuming isolated boundary manifold surface, realize the leap from single-point monitoring to global visualization accurate perception, provide reliable data basis for subsequent regulation and control. In the regulation and control strategy, the physical convection damping and chemical acceleration consumption double-track collaborative mode are adopted, the pressure difference gradient and the proportion of oxidizing agent are reduced at the same time, the critical boundary is locked from the two dimensions of fluid dynamics and chemical reaction kinetics, the pyrolysis zone oxidation out-of-control hidden danger is effectively eliminated, and the working condition of the targeted cleavage area is stable.

Description

Technical Field

[0001] This invention relates to the field of unconventional oil and gas in-situ extraction and industrial intelligent measurement and control technology, and in particular to a multi-field evolution simulation method for targeted pyrolysis of oil-rich coal with in-situ self-heating. Background Technology

[0002] Traditional oil-rich coal mining often employs a segmented process of first mining the coal and then pyrolyzing it on the surface. This model not only causes severe ecological disturbance and incurs high mining costs, but also makes it difficult to economically and effectively utilize deep coal seam resources. In contrast, underground in-situ pyrolysis technology eliminates the need to bring coal to the surface, possessing disruptive advantages such as greater resource development potential and a smaller environmental footprint, and is widely regarded as an important direction for the future development of this field. However, this technology still faces three critical technical bottlenecks that urgently need to be overcome in its journey towards industrial application: (i) Lack of boundary dynamic perception and locking capabilities, resulting in a significant risk of secondary oxidation and burnout of light oil: Due to the heterogeneity of the coal body and the development of natural fractures, the oxygen injection process is prone to creating dominant air leakage channels, leading to airflow short-circuiting. Because it is impossible to detect and accurately locate the oxygen-consuming boundary in the complex porous medium in real time, oxygen often intrudes into the downstream pyrolysis reaction zone, causing secondary oxidation or even combustion of the newly precipitated high-value light coal tar, severely reducing the yield of liquid products. The current lack of dynamic monitoring methods adapted to the boundaries of deep, hidden physical fields is a key weakness restricting yield improvement.

[0003] (ii) Insufficient in-situ thermodynamic control capability, frequent side reactions, and high dependence on secondary surface upgrading: Existing oxygen injection self-ignition technology struggles to maintain the critical thermodynamic parameters required for the spatiotemporal scale of directional pyrolysis of light oils, and it also fails to fully consider the constraining effect of the in-situ high-pressure environment on the chemical reaction equilibrium. When the high-temperature CO2 gas stream generated in the combustion zone migrates downstream, it easily induces a violent Boudouard carbon dissolution reaction (CO2 + C → 2CO). This strongly endothermic process not only consumes a large amount of the system's internal heat energy but also promotes the deterioration and heavyening of the in-situ precipitated tar. Current solutions can only rely on energy-intensive and costly secondary hydrotreating processes on the surface as a remedy, and cannot yet achieve highly selective, low-carbon cleavage conversion in-situ downhole.

[0004] (III) Lack of precision measurement and control hardware and algorithm carriers that integrate microscopic mechanism verification and physical constraint control: Current engineering practices largely rely on on-site experience for debugging, lacking precise experimental reconstruction devices capable of reproducing multi-field coupled evolution processes in large-scale three-dimensional space. Furthermore, traditional pure data-driven prediction models, under unsteady conditions such as ground disturbances and abrupt flow field boundary changes, are prone to algorithm divergence or even collapse due to the lack of rigid constraints from underlying physical laws, failing to effectively guide actuators such as mass flow meters and electronic back pressure valves in adaptive control. Moreover, under sparse sensor deployment conditions, existing technologies struggle to accurately invert and analyze the three-dimensional continuous temperature field and critical oxygen-consuming isolation boundary within a completely black-box region, severely hindering the realization of intelligent closed-loop control. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a multi-field evolution simulation method for in-situ self-heating targeted pyrolysis of oil-rich coal. This method overcomes the limitations of traditional techniques that rely solely on discrete point monitoring data, enabling the acquisition of the full-domain, continuous, and dynamic multi-physics field distribution patterns of porous coal within the reactor, and achieving accurate prediction of the CODB critical boundary location and its evolution trend. Simultaneously, this invention also solves the problem of existing single-parameter control methods failing to simultaneously coordinate the imbalance between flow field convection turbulence and oxidation reaction, effectively avoiding a series of chain reactions caused by the spread of the CODB boundary into the pyrolysis zone, such as oxidation interference in the targeted pyrolysis zone, tar burn-off, and heating imbalance.

[0006] To achieve the above objectives, the technical solution of this invention is as follows: In a first aspect, the present invention provides a multi-field evolution simulation method for in-situ self-heating targeted pyrolysis of oil-rich coal, the method comprising: Multi-source discrete monitoring data of oil-rich coal specimens in a true triaxial reactor are obtained. The multi-source discrete monitoring data is input into a multi-field adaptive collaborative inversion operator deployed in an edge computing hub. Dynamic assimilation is performed on the black-box flow field of the heterogeneous porous medium inside the reactor. A three-dimensional dynamic control field with global continuity is output. The spatial coordinate manifold of the critical oxygen-consuming isolation boundary CODB is extracted based on the three-dimensional dynamic control field. Based on the temperature distribution of the three-dimensional dynamic control field and the evolution trend of the spatial coordinate manifold of the CODB, a dual-track PID control law is used to generate control commands. When the uncontrolled spread of the CODB boundary to the pyrolysis zone is detected, a first sub-command to reduce the global pressure gradient and a second sub-command to reduce the proportion of oxidant intake are generated simultaneously. The spatial position of the CODB is forcibly locked through the linkage of physical convection damping and chemical acceleration consumption. The theoretical heat supply required to maintain the target temperature of the targeted pyrolysis zone is calculated. Based on the thermochemical stoichiometry of the carbon oxidation exothermic reaction, the instantaneous volume fraction of oxidant required to maintain the self-heating source is calculated. The first control command is generated and sent to the front-end multi-channel mass flow meter to adjust the instantaneous volume ratio of oxygen in the mixed gas injected into the reactor. When the pyrolysis zone is detected to have entered the light tar precipitation stage, the electronic back pressure valve at the output end is controlled to construct an ultra-high pressure physical constraint environment in the reactor. The Le Chatelier principle is used to suppress the carbon dissolution endothermic side reaction in situ, thus completing the closed-loop intelligent measurement and control of the in-situ self-heating targeted pyrolysis of oil-rich coal.

[0007] In some embodiments, the multi-field adaptive cooperative inversion operator is constructed through the following steps: constructing a joint evaluation cost function that includes sensor data residual terms and physical partial differential equation penalty terms, wherein the physical partial differential equation penalty terms explicitly characterize the energy conservation equation, the component convection-diffusion equation, and Darcy's law of percolation; using automatic differentiation technology, randomly arranging virtual spatiotemporal configuration points in the internal space of the reactor, directly differentiating the predicted temperature field, concentration field, and pressure field output by the network and substituting them into the THMC coupled partial differential equation, calculating the mean square error of the residuals at each point as the physical constraint loss; performing inversion optimization by minimizing the joint evaluation cost function, outputting a solution that rigidly obeys the physical conservation law, and using discrete measurement point data to inversely analyze the continuous physical field.

[0008] In some embodiments, the process of generating the first sub-instruction to reduce the global differential pressure gradient specifically includes: real-time monitoring of the oxygen volume fraction fed back by the axial component sensor; when the oxygen volume fraction exceeds a preset warning threshold and the temperature is in the critical transition zone of oxidation-pyrolysis, determining that the CODB boundary is unstable; immediately sending a voltage signal to reduce the opening of the electronic back pressure valve at the output end, smoothly adjusting the global differential pressure gradient from the injection end to the output end in the reactor to the range of 0.5~0.8 MPa, relying on physical resistance to slow down the gas convection mass transfer rate, and in conjunction with the instruction to reduce the proportion of oxidant inlet gas, forcibly locking the CODB boundary within the safe limit coordinates.

[0009] In some embodiments, the process of generating a second sub-instruction to reduce the oxidant intake ratio specifically includes: calculating the theoretical total heat required to heat the specimen at the current time step and maintain self-heating based on the calorific value of the exothermic carbon oxidation reaction and the target maintenance temperature of the targeted pyrolysis zone; calculating the theoretical oxygen mass flow rate required to maintain the theoretical total heat, and introducing carbon dioxide as an inert heat transfer medium and dilution medium; adaptively iteratively calculating the carbon dioxide mass flow rate based on the transient thermal balance constraint equation, dynamically determining the instantaneous volume fraction of the oxidant, and generating a control signal to adjust the output of the front-end multi-channel mass flow meter.

[0010] In some embodiments, the specific process for calculating the instantaneous volume fraction of the oxidant is as follows: based on the theoretical heat supply and the standard enthalpy of the carbon oxidation exothermic reaction, calculate the theoretical oxygen mass flow rate required to maintain the target temperature of the targeted pyrolysis zone; introduce carbon dioxide as an inert heat transfer and dilution medium, and based on the measured temperature at the center of the combustion zone at the front end of the reactor and the specific heat capacity of carbon dioxide at the current operating pressure, adaptively iterate and calculate the carbon dioxide mass flow rate based on the transient thermal balance constraint equation; determine the target volume fraction of oxygen and carbon dioxide in the mixed gas at the current time step based on the theoretical oxygen mass flow rate and the carbon dioxide mass flow rate, and send the corresponding flow control word to each gas channel of the front-end multi-channel mass flow meter.

[0011] In some embodiments, the specific execution method of the dual-track PID control law includes: continuously monitoring the oxygen volume fraction at each measuring point along the reactor axis; when the oxygen volume fraction exceeds a preset warning threshold, determining that the CODB boundary is spreading uncontrollably into the pyrolysis zone; sending a voltage signal to the electronic back pressure valve at the output end to reduce the opening, smoothly adjusting the global differential pressure gradient of the reactor to a preset low differential pressure safety threshold; and simultaneously sending an oxygen flow rate reduction command to the front-end multi-channel mass flow meter to reduce the proportion of oxygen intake flow rate, thereby accelerating chemical oxygen consumption in the non-target area and forcibly locking the CODB boundary within the safety limit coordinates.

[0012] In some embodiments, the in-situ suppression of the carbon dissolution endothermic side reaction using the Le Chatelier principle specifically includes: controlling the electronic back pressure valve at the output end to increase the pressure inside the reactor and maintain it above a preset high pressure threshold; according to the principle of chemical equilibrium shift, using a preset ultra-high pressure physical constraint environment to force the chemical equilibrium of the carbon dissolution endothermic side reaction to shift in the reverse direction, so that the high-temperature carbon dioxide fluid only acts as a physical heat transfer medium to transport heat, blocking the endothermic path of carbon dioxide reacting with carbon to generate carbon monoxide.

[0013] In some embodiments, the specific thermodynamic control process for in-situ suppression of the endothermic side reaction of carbon dissolution is as follows: real-time acquisition of the reaction temperature and partial pressure of carbon dioxide in the pyrolysis zone, and construction of a dynamic comparison mechanism between the reaction quotient and the standard equilibrium constant at that temperature; when it is determined that the reaction is in the endothermic expansion range, the electronic back pressure valve at the output end is controlled to raise the total system pressure to above the critical suppression threshold, so that the critical suppression threshold increases rapidly due to the pressure quotient effect and approaches or exceeds it, and the principle of chemical equilibrium shift is used to force the carbon dissolution reaction equilibrium to shift in the reverse direction, thereby blocking the endothermic side reaction of carbon dioxide and carbon at the thermodynamic source.

[0014] The multi-field evolution simulation method for targeted pyrolysis of oil-rich coal with in-situ self-heating provided by this invention first reconstructs a three-dimensional dynamic control field through dynamic assimilation and adaptive inversion operators of multi-source discrete data. This accurately captures the real-time evolution state of the oxygen-consuming isolation boundary manifold, achieving a leap from single-point monitoring to precise perception of the entire domain through visualization, providing a reliable data foundation for subsequent regulation. In terms of regulation strategy, a dual-track synergistic mode of physical convection damping and chemical acceleration consumption is adopted, while simultaneously reducing the pressure gradient and oxidant ratio. This locks the critical boundary from both fluid dynamics and chemical reaction kinetics dimensions, effectively eliminating the risk of runaway oxidation in the pyrolysis zone and ensuring stable operating conditions in the targeted pyrolysis region.

[0015] Based on this, the instantaneous volume fraction of the oxidant is precisely calculated using thermochemical stoichiometry, and the oxygen ratio is adjusted as needed. This allows for a dynamic and precise match between the exothermic oxidation and the heat loss and endothermic demand in the pyrolysis zone, achieving an adaptive fit for in-situ self-heating. This significantly reduces experimental energy consumption and improves the stability of the pyrolysis reaction. Simultaneously, during the critical window of product precipitation, an ultra-high pressure physical constraint environment is constructed using an electronic back pressure valve. The Le Chatelier principle is then used to directionally suppress ineffective endothermic side reactions, maximizing the retention of light tar components, significantly improving tar yield and product purity, and optimizing the resource utilization efficiency of oil-rich coal. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the multi-field evolution simulation method for in-situ self-heating targeted pyrolysis of oil-rich coal provided in this embodiment of the invention. Figure 2 This is a schematic diagram of the spatial partitioning and multi-field sensing monitoring of the specimen inside the simulated reactor provided in this embodiment of the invention; Figure 3 This is the in-situ thermodynamic inversion and closed-loop control logic diagram based on a multi-field adaptive assimilation operator provided in this embodiment of the invention; Figure 4 This is a diagram illustrating the multi-field evolution of the isolation barrier and the carbon suppression reaction mechanism provided in this embodiment of the invention. Figure 5 This is a planar diagram of the heterogeneous evolution and regulation of the hidden boundary of CODB provided in this embodiment of the invention; Figure 6 This is a diagram showing the relationship between multidimensional parameter optimization and light oil yield / risk assessment provided in an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0018] In the following description, references to "some embodiments" refer to a subset of all possible embodiments; however, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of the invention have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of the invention pertain. The terminology used in the embodiments of the invention is for the purpose of describing the embodiments of the invention only and is not intended to limit the invention.

[0019] The following describes an exemplary application of the oil-rich coal in-situ self-heating targeted pyrolysis multi-field evolution simulation device according to embodiments of the present invention. This device can be implemented as a terminal or a server. In one implementation, the device can be implemented as a laptop, tablet, desktop computer, mobile device, or other types of terminal. In another implementation, it can also be implemented as a server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The terminal and server can be directly or indirectly connected via wired or wireless communication, which is not limited in the embodiments of the present invention. The following will illustrate an exemplary application of the multi-field evolution simulation device for in-situ self-heating targeted pyrolysis of oil-rich coal as a server.

[0020] This invention provides a multi-field evolution simulation method for targeted pyrolysis of oil-rich coal with in-situ self-heating, see [link to relevant documentation]. Figure 1 , Figure 1 This is a flowchart illustrating a multi-field evolution simulation method for in-situ self-heating targeted pyrolysis of oil-rich coal provided in an embodiment of the present invention, which will be combined with... Figure 1 The steps shown are explained.

[0021] S110: Obtain multi-source discrete monitoring data of oil-rich coal specimens in a true triaxial reactor, input the multi-source discrete monitoring data into a multi-field adaptive collaborative inversion operator deployed in an edge computing hub, perform dynamic assimilation on the black-box flow field of the heterogeneous porous medium inside the reactor, output a global continuous three-dimensional dynamic control field, and extract the spatial coordinate manifold of the critical oxygen-consuming isolation boundary CODB based on the three-dimensional dynamic control field.

[0022] Here, the oil-rich coal specimen refers to a coal sample selected from oil-rich coal, which has undergone standardized sampling and processing to form a shape, and is used for indoor true triaxial physical simulation tests. It serves as the test and control carrier for this invention. Oil-rich coal is a special coal resource with low ash content, high volatile matter, and high tar yield, possessing excellent pyrolysis oil production characteristics.

[0023] Here, multi-source discrete monitoring data refers to the multi-dimensional raw monitoring data collected in real time during the experiment by various types of sensors deployed inside the true triaxial reactor and on the surface of the specimen. The data is independent in time and spatially discrete. The data sources are diverse, the sampling points are scattered, and the initial state is non-continuous discrete data. It is the basic data source for subsequent field inversion calculation.

[0024] In this invention, the edge computing hub carries the deployment and operation of multiple adaptive collaborative inversion operators, ensuring the real-time dynamic control of the fragmentation process and avoiding the control lag problem caused by cloud transmission delay.

[0025] Here, the multi-field adaptive collaborative inversion operator is an algorithmic computation model exclusively constructed in this invention. It is an intelligent inversion computation unit that integrates the coupling laws of multiple physical fields, including temperature, pressure, flow, and stress fields. It can adaptively correct the multi-physical field coupling parameters based on discrete monitoring data and reconstruct the complex medium flow field within the reactor through a data assimilation algorithm. It has adaptive correction capabilities and can dynamically adjust the inversion weights according to parameter changes at different stages of coal pyrolysis, adapting to the flow field evolution laws of heterogeneous porous media, and achieving accurate conversion from discrete data to a continuous field.

[0026] It should be noted that the characteristics of heterogeneous porous media can cause gas flow, heat transfer, and oxidation reactions in the reactor to exhibit non-uniform and non-linear features. Conventional uniform field calculation models cannot accurately adapt to these features, and accurate reconstruction can only be achieved by relying on the inversion operator of this technology.

[0027] Here, the black-box flow field refers to the actual fluid motion field in the oil-rich coal-containing medium region inside a true triaxial reactor. Due to the heterogeneity of the coal body and the interference of multiple field couplings, the internal gas flow direction, velocity, pressure distribution, and mass transfer laws cannot be obtained through direct observation. It belongs to an unknown field with complex internal mechanisms and no intuitive visualization data, hence it is defined as a black-box flow field. Conventional monitoring can only obtain discrete point data and cannot obtain the full-domain, continuous, and dynamic fluid distribution state inside the reactor, i.e., the black-box state. Algorithms are needed to assimilate and decipher the unknown nature of the field.

[0028] Here, dynamic assimilation is a multi-source data fusion and correction operation method. It refers to the process of dynamically fusing and iteratively correcting the discrete monitoring and measured data collected in real time with the multiphysics field theoretical evolution model, continuously optimizing the model calculation error, correcting the flow field evolution parameters, and transforming discrete, single-point, and time-series fragmented data into continuous and dynamically updated field data across the entire domain.

[0029] Here, the global continuous three-dimensional dynamic control field refers to the visualized, quantifiable, and controllable three-dimensional spatial field model output after dynamic assimilation calculation. It is a multi-physics field fusion control field that covers the entire reaction area of ​​the oil-rich coal specimen in the reactor, has no spatial discontinuities, and is updated in real time over time. It includes coupled parameters such as global temperature, pressure, flow rate, and oxygen concentration.

[0030] Here, CODB (Critical Oxygen Depletion Boundary) refers to the critical oxygen depletion boundary in the in-situ pyrolysis process of oil-rich coal, which is the spatial interface between the coal body's oxidation heating zone and the pure pyrolysis zone. This boundary is a critical state interface; the area inside the boundary is the oxidation exothermic heating zone, while the area outside the boundary is the targeted pyrolysis zone where no oxidation occurs and only pyrolysis occurs. It is the critical interface that distinguishes the self-heating zone from the targeted oil-producing zone.

[0031] Here, the spatial coordinate manifold refers to the continuous surface formed by the CODB critical oxygen consumption isolation boundary in three-dimensional space. It is composed of a large number of three-dimensional spatial coordinate points and can accurately characterize the spatial shape, position distribution and surface curvature of the isolation boundary. It also evolves dynamically with the reaction process and is a geometric carrier for quantifying the boundary migration law and judging the stability of the operating condition.

[0032] S120, based on the temperature distribution of the three-dimensional dynamic control field and the evolution trend of the spatial coordinate manifold of the CODB, a control command is generated using a dual-track PID control law; when the uncontrolled spread of the CODB boundary to the pyrolysis zone is detected, a first sub-command to reduce the global differential pressure gradient and a second sub-command to reduce the proportion of oxidant intake are generated simultaneously, and the spatial position of the CODB is forcibly locked through the linkage of physical convection damping and chemical acceleration consumption.

[0033] Here, the dual-track PID control law is a proprietary dual-path parallel PID intelligent control algorithm of this invention. Unlike traditional single-parameter PID control, it can simultaneously perform proportional, integral, and derivative coordinated control on two control tracks: the pressure gradient field and the oxidant concentration field. It synchronously outputs pressure control commands and gas ratio control commands, achieving precise parameter control through dual-dimensional linkage. It can simultaneously correct physical flow field parameters and chemical reaction parameters, overcoming the technical deficiency of single control methods in locking the boundaries.

[0034] Here, the pyrolysis zone is a dedicated reaction area in the oil-rich coal specimen where macromolecular chemical bonds break, the coal body decomposes under heat, and tar, coal gas, and semi-coke are generated. This zone does not require oxygen to participate in the oxidation reaction, and its goal is to directionally precipitate light tar products.

[0035] Here, the global pressure gradient refers to the rate of change of the difference in fluid pressure at different points within the entire space inside the true triaxial reactor as a function of spatial distance. It characterizes the overall driving force of the fluid inside the reactor. The larger the pressure gradient, the faster the gas convection and diffusion rate, and the easier it is for the CODB boundary to spread rapidly into the pyrolysis zone.

[0036] Here, the first sub-instruction is the physical field control instruction output by the dual-track PID control law. Its function is to reduce the pressure difference gradient across the entire reactor, weaken the gas convection driving force inside the reactor, achieve physical-level convection damping constraint, and suppress the diffusion of oxygen into the pyrolysis zone.

[0037] Here, the oxidant intake gas ratio refers to the volume ratio of oxygen in the mixed process gas injected into the reactor. It directly determines the intensity and heating rate of the exothermic carbon oxidation reaction in the coal body and is a core chemical parameter for regulating the self-heating intensity and locking the CODB boundary.

[0038] Here, the second sub-instruction is the chemical reaction control instruction output by the dual-track PID control law. Its function is to reduce the proportion of oxidant in the mixed gas, thereby weakening the intensity of the oxidation reaction and accelerating local oxygen consumption, thus inhibiting the uncontrolled spread of the boundary from a chemical perspective.

[0039] Here, physical convection damping refers to a physical constraint mechanism that reduces the gas convection diffusion dynamics by lowering the global pressure gradient inside the vessel, thereby delaying the migration of oxygen and high-temperature gas flow to the target pyrolysis zone. It is a passive flow field control method.

[0040] S130: Calculate the theoretical heat supply required to maintain the target temperature of the targeted pyrolysis zone. Based on the thermochemical stoichiometry of the exothermic carbon oxidation reaction, calculate the instantaneous volume fraction of oxidant required to maintain the self-heating source. Generate the first control command and send it to the front-end multi-channel mass flow meter to adjust the instantaneous volume ratio of oxygen in the mixed gas injected into the reactor.

[0041] Here, the targeted pyrolysis zone is a pre-defined coal pyrolysis zone for efficient production of light tar, which is the optimal pyrolysis reaction zone after boundary locking, free from oxidation interference, and with stable and controllable temperature.

[0042] Here, theoretical heat supply refers to the precise heat supply value required to maintain the temperature stability of the cracking zone, calculated by the heat balance equation based on the preset target temperature of the targeted cracking zone, the heat absorption rate of coal pyrolysis, and the heat loss rate. It is the theoretical benchmark for regulating the intensity of oxidation heating.

[0043] Here, the carbon oxidation exothermic reaction refers to the exothermic chemical reaction of fixed carbon and oxygen in oil-rich coal. It is the core heat source for in-situ self-heating in this invention. The exothermic reaction provides all the heat required for coal pyrolysis without the need for external heating equipment.

[0044] Here, the instantaneous volume fraction of oxidant refers to the instantaneous volume ratio of oxygen in the gas mixture entering the reactor at a certain real time. Unlike the average ratio, it can achieve millisecond-level dynamic and precise control to adapt to the dynamic heating needs of different stages of the pyrolysis process.

[0045] Here, the first control command is a precise control command for gas ratio based on thermochemical stoichiometry, which is sent to the multi-channel mass flow meter to adjust the oxygen ratio of the mixed gas in real time, accurately match the self-heating demand, and achieve a dynamic balance between heat supply and demand.

[0046] S140, when the pyrolysis zone enters the light tar precipitation stage, controls the electronic back pressure valve at the output end to construct an ultra-high pressure physical constraint environment in the reactor, and uses Le Chatelier's principle to suppress the carbon dissolution endothermic side reaction in situ, thus completing the fully closed-loop intelligent measurement and control of the in-situ self-heating targeted pyrolysis of oil-rich coal.

[0047] Here, the ultra-high pressure physical confinement environment refers to the high-pressure closed environment inside the reactor, which is higher than the conventional pyrolysis pressure, formed by electronic back pressure valve regulation during the light tar precipitation stage. The pressure confinement changes the equilibrium direction of the pyrolysis reaction and suppresses ineffective side reactions.

[0048] The multi-field evolution simulation method for targeted pyrolysis of oil-rich coal with in-situ self-heating provided by this invention first reconstructs a three-dimensional dynamic control field through dynamic assimilation and adaptive inversion operators of multi-source discrete data. This accurately captures the real-time evolution state of the oxygen-consuming isolation boundary manifold, achieving a leap from single-point monitoring to precise perception of the entire domain through visualization, providing a reliable data foundation for subsequent regulation. In terms of regulation strategy, a dual-track synergistic mode of physical convection damping and chemical acceleration consumption is adopted, while simultaneously reducing the pressure gradient and oxidant ratio. This locks the critical boundary from both fluid dynamics and chemical reaction kinetics dimensions, effectively eliminating the risk of runaway oxidation in the pyrolysis zone and ensuring stable operating conditions in the targeted pyrolysis region.

[0049] Based on this, the instantaneous volume fraction of the oxidant is precisely calculated using thermochemical stoichiometry, and the oxygen ratio is adjusted as needed. This allows for a dynamic and precise match between the exothermic oxidation and the heat loss and endothermic demand in the pyrolysis zone, achieving an adaptive fit for in-situ self-heating. This significantly reduces experimental energy consumption and improves the stability of the pyrolysis reaction. Simultaneously, during the critical window of product precipitation, an ultra-high pressure physical constraint environment is constructed using an electronic back pressure valve. The Le Chatelier principle is then used to directionally suppress ineffective endothermic side reactions, maximizing the retention of light tar components, significantly improving tar yield and product purity, and optimizing the resource utilization efficiency of oil-rich coal.

[0050] In summary, this invention achieves fully automated closed-loop control of the entire process from data acquisition, field inversion, boundary judgment, command output, parameter adjustment to side reaction suppression. It not only has a fast response speed and high control accuracy, but also requires no manual intervention, comprehensively improving the standardization and intelligence level of in-situ pyrolysis tests of oil-rich coal, and providing a strong guarantee for the accuracy and repeatability of test results.

[0051] In some embodiments, the multi-field adaptive cooperative inversion operator is constructed through the following steps: constructing a joint evaluation cost function that includes sensor data residual terms and physical partial differential equation penalty terms, wherein the physical partial differential equation penalty terms explicitly characterize the energy conservation equation, the component convection-diffusion equation, and Darcy's law of percolation; using automatic differentiation technology, randomly arranging virtual spatiotemporal configuration points in the internal space of the reactor, directly differentiating the predicted temperature field, concentration field, and pressure field output by the network and substituting them into the THMC coupled partial differential equation, calculating the mean square error of the residuals at each point as the physical constraint loss; performing inversion optimization by minimizing the joint evaluation cost function, outputting a solution that rigidly obeys the physical conservation law, and using discrete measurement point data to inversely analyze the continuous physical field.

[0052] Here, the joint evaluation cost function refers to the comprehensive loss function used to quantify the deviation between the neural network's predicted field and the real physical field, and to integrate the data fitting accuracy and compliance with physical rules. It is the core evaluation criterion for training, iteration, and optimization of multi-field adaptive collaborative inversion operators. This function simultaneously considers the fitting error of measured sensor data and the constraint error of physical equations, achieving dual correction based on data-driven and physical constraints.

[0053] In this embodiment, the multi-field inversion scenario of oil-rich coal pyrolysis includes the residual loss between the measured and predicted values ​​of temperature, concentration, and pressure measurement points, as well as the physical penalty loss due to energy conservation, component diffusion, and seepage mechanics equations. The joint evaluation cost function comprehensively assesses the accuracy and physical rationality of the model output field.

[0054] Here, the sensor data residual term refers to the data-driven component of the joint evaluation cost function, which is used to characterize the deviation between the physical field prediction value output by the model network and the discrete monitoring data measured by the field sensors. It is used to constrain the model output results to fit the real experimental conditions and ensure that the inversion results fit the measured data.

[0055] Here, the physical partial differential equation penalty term refers to the physical constraint term of the joint evaluation cost function. It is an explicitly embedded residual constraint of the classical multi-field coupled control equation. It is used to penalize model prediction results that violate the fundamental physical conservation laws, and to force the network output field to satisfy the real fluid, heat transfer, and mass transfer physical laws, so as to avoid non-physical interpretation and extrapolation distortion problems in the pure data-driven model.

[0056] Here, virtual spatiotemporal configuration points refer to virtual computational sampling points that are randomly generated by an algorithm within the three-dimensional spatial domain and pyrolysis time-series domain inside a true triaxial reactor. These points are different from the discrete measured points of physical sensors and are used for global dense physical residual calculations to fill the gaps in the sparseness and insufficient spatial coverage of physical measurement points.

[0057] Here, inverse optimization is a parameter optimization process that aims to minimize the joint evaluation cost function, continuously update the neural network weight parameters through intelligent iterative algorithms such as gradient descent, and continuously reduce the data residual and physical residual to solve the optimal global physical field distribution.

[0058] The multi-field adaptive collaborative inversion operator construction method disclosed in this embodiment constructs a joint cost function that fuses data residuals and physical equation penalties. Combined with automatic differentiation and iterative optimization of virtual configuration points across the entire physical constraint domain, this method offers the following advantages compared to traditional pure data-driven interpolation, single physical simulation, and sparse measurement point fitting field reconstruction techniques: Firstly, by forcing the model output to conform to the real multi-field coupling mechanism through the penalty term, the realism and reliability of black-box flow field inversion are significantly improved. Secondly, by deploying globally random virtual spatiotemporal configuration points and combining them with automatic differentiation technology to complete dense physical residual calculation, the method fills the spatial gaps of physical measurement points, achieving a leap from sparse discrete measurement point monitoring to accurate analysis of the entire continuous physical field. Thirdly, by dynamically balancing data fitting accuracy and physical constraint strength through the joint cost function and adaptively correcting field parameters through iterative optimization, this method can adapt to the multi-field evolution laws of different pyrolysis stages, different seepage states, and different oxidation reaction intensities, exhibiting strong resistance to data noise and a wide range of operating conditions.

[0059] In summary, this embodiment does not rely on high-precision, high-density physical sensor arrays. It can complete the full-domain high-precision field reconstruction using only conventional multi-source discrete monitoring data, effectively reducing equipment costs and experimental difficulty. At the same time, the unified physical constraint standard ensures the stability and repeatability of the field inversion results under different test batches and different coal sample conditions, greatly improving the engineering applicability of the intelligent monitoring and control system for in-situ pyrolysis of oil-rich coal.

[0060] In some embodiments, the process of generating the first sub-instruction to reduce the global differential pressure gradient specifically includes: real-time monitoring of the oxygen volume fraction fed back by the axial component sensor; when the oxygen volume fraction exceeds a preset warning threshold and the temperature is in the critical transition zone of oxidation-pyrolysis, determining that the CODB boundary is unstable; immediately sending a voltage signal to reduce the opening of the electronic back pressure valve at the output end, smoothly adjusting the global differential pressure gradient from the injection end to the output end in the reactor to the range of 0.5~0.8 MPa, relying on physical resistance to slow down the gas convection mass transfer rate, and in conjunction with the instruction to reduce the proportion of oxidant inlet gas, forcibly locking the CODB boundary within the safe limit coordinates.

[0061] Here, oxygen volume fraction refers to the percentage of oxygen component volume in the mixed gas at a specified monitoring point in the reactor, which is the core quantitative parameter characterizing the enrichment degree of oxidant, the intensity of oxidation reaction, and the migration trend of CODB boundary. It directly reflects the boundary equilibrium state between the oxidation region and the pyrolysis region.

[0062] Here, the preset warning threshold is a critical oxygen concentration value pre-calibrated based on the oil-rich coal pyrolysis test conditions and the CODB boundary stability critical conditions. It serves as a built-in safety judgment standard for the system. When the monitored oxygen volume fraction exceeds this threshold, it indicates that oxygen is abnormally diffusing into the pyrolysis zone, posing a risk of uncontrolled spread at the boundary.

[0063] Here, the oxidation-pyrolysis critical transition zone refers to the narrow transition space between the oil-rich coal oxidation heating zone and the targeted pyrolysis zone, which is the dynamic attachment area of ​​the CODB critical oxygen consumption isolation boundary. The temperature and oxygen concentration in this region are in a dynamic critical equilibrium state, and the temperature value is between the high-temperature oxidation zone and the isothermal pyrolysis zone. It is the core sensitive area where the boundary is most prone to deviation, instability, and spread.

[0064] Here, CODB boundary instability refers to the equilibrium failure state of the critical oxygen-consuming isolation boundary. Specifically, it refers to the unstable working condition state in which the originally stable isolation zone between the oxidation zone and the pyrolysis zone is directionally shifted, distorted, and spreads outward due to the influence of pressure difference, oxygen concentration, and temperature field fluctuations inside the reactor. Oxygen abnormally invades the targeted pyrolysis zone, resulting in an imbalance of pyrolysis conditions and the risk of product loss.

[0065] Here, the safety limit coordinates refer to the pre-calibrated three-dimensional spatial safety coordinate range of the CODB boundary. It is the optimal isolation position between the pyrolysis zone and the oxidation zone, which can ensure stable heating in the oxidation heating zone and no oxygen interference in the pyrolysis zone. It is the target spatial position of the system's locking boundary.

[0066] This embodiment refines the generation and control process of the first sub-instruction. Compared with traditional coarse pressure control and passive operating condition adjustment technology, it solves industry pain points such as delayed prediction of CODB boundary instability, low control accuracy, large operating condition fluctuations, and poor boundary locking effect. It has the following beneficial effects: This invention abandons the traditional coarse mode of judging by a single temperature or a single pressure. It combines the axial oxygen volume fraction threshold and the temperature of the oxidation-pyrolysis critical transition zone as dual-dimensional parameters for joint judgment. It can accurately identify the small instability trend of CODB boundary and trigger control in the early stage of large-scale oxygen invasion into the pyrolysis zone. It realizes the technical upgrade from post-remediation to pre-judgment and real-time correction, and completely avoids the risk of oxidation runaway in the pyrolysis zone. Meanwhile, by finely adjusting the opening of the electronic back pressure valve through voltage signals, the global differential pressure gradient is precisely locked within the optimal range of 0.5~0.8 MPa. Furthermore, by adopting a continuous and smooth adjustment method, problems such as sudden rises and falls in pressure, flow field disturbances, and temperature field jumps caused by large opening and closing of traditional valves are avoided. This effectively protects the continuity and stability of oil-rich coal pyrolysis and prevents external disturbances from affecting the quality of light tar precipitation.

[0067] In addition, by constructing physical resistance to the fluid at the output end, the gas convection mass transfer rate inside the reactor is directly reduced, weakening the migration force of oxygen to the pyrolysis zone from the source of fluid dynamics. Combined with chemical control measures that reduce the proportion of oxidant inlet gas, a dual constraint of physical diffusion resistance and chemical oxygen consumption is formed. Compared with single chemical control, the boundary locking speed is faster and the stability is stronger.

[0068] In some embodiments, the process of generating a second sub-instruction to reduce the oxidant intake ratio specifically includes: calculating the theoretical total heat required to heat the specimen at the current time step and maintain self-heating based on the calorific value of the exothermic carbon oxidation reaction and the target maintenance temperature of the targeted pyrolysis zone; calculating the theoretical oxygen mass flow rate required to maintain the theoretical total heat, and introducing carbon dioxide as an inert heat transfer medium and dilution medium; adaptively iteratively calculating the carbon dioxide mass flow rate based on the transient thermal balance constraint equation, dynamically determining the instantaneous volume fraction of the oxidant, and generating a control signal to adjust the output of the front-end multi-channel mass flow meter.

[0069] Here, the heat required for specimen heating refers to the sensible heat consumed in the current time step for the temperature of the oil-rich coal specimen body to rise, the temperature of the porous medium to rise, and the heat storage of the solid structure. It is an important component of the heat demand in the early stage of pyrolysis and the stage of fluctuating operating conditions.

[0070] Here, the theoretical oxygen mass flow rate refers to the standard oxygen supply mass flow rate calculated based on the thermochemical stoichiometry of carbon oxidation and matching the theoretical total heat demand at the current time step, which is exactly the amount required to maintain thermal equilibrium. This flow rate is the optimal theoretical supply value, ensuring that the heat released by oxidation just offsets the heat consumption of the system, with no excess oxygen and no insufficient heat supply.

[0071] Here, the inert heat carrier and dilution medium is an inert gas medium introduced in this embodiment that does not participate in the exothermic carbon oxidation reaction and does not interfere with the pyrolysis reaction. It also has two core functions: first, heat carrier and energy storage, uniformly carrying and transferring heat in the reactor to balance the temperature field; second, oxygen dilution, precisely reducing the volume ratio of oxidant, suppressing the intensity of the oxidation reaction, and constraining the migration of the CODB boundary.

[0072] Here, the control signal refers to the analog control command generated by the edge computing center, which is used to drive the multi-channel mass flow meter to correct the output flow of each gas in real time, and ultimately realize the electronic control signal of dynamically reducing the oxidant intake ratio and adaptively maintaining thermal balance.

[0073] This embodiment defines the precise generation and dynamic control process of the second sub-instruction. Through transient thermal balance iteration, inert medium synergistic proportioning, and dynamic oxygen dilution control, it achieves refined constraints on the oxidation reaction intensity. Compared to traditional fixed oxygen ratio and open-loop gas intake control technology, it has the following advantages: The theoretical total heat is calculated in real-time on a single time step basis, and the oxygen supply flow is dynamically matched to achieve on-demand heating. This ensures that the heat release from oxidation always precisely matches the heat consumption of pyrolysis, significantly improving the stability of the pyrolysis operation. The innovative introduction of carbon dioxide as a heat carrier and dilution medium, unlike the traditional method of single nitrogen dilution, can precisely reduce the oxidant intake ratio, weaken the oxidation reaction intensity, and inhibit the spread of the CODB boundary to the pyrolysis zone. Furthermore, the high heat capacity of carbon dioxide carries uniform heat, solving the problem of localized high-temperature hotspots within the reactor, balancing the overall temperature field, and providing a superior temperature environment for the stable precipitation of light tar.

[0074] Meanwhile, this embodiment relies on the transient thermal balance constraint equation to adaptively iterate and solve the medium flow rate and oxygen volume fraction. It does not require a large number of manual calibration empirical parameters and can automatically adapt to the strong nonlinear characteristics of oil-rich coal heterogeneous media, multi-field coupled fluctuations, and dynamic evolution of reactions. It has strong control robustness, small error, and wide range of adaptable working conditions.

[0075] In some embodiments, the specific process for calculating the instantaneous volume fraction of the oxidant is as follows: based on the theoretical heat supply and the standard enthalpy of the carbon oxidation exothermic reaction, calculate the theoretical oxygen mass flow rate required to maintain the target temperature of the targeted pyrolysis zone; introduce carbon dioxide as an inert heat transfer and dilution medium, and based on the measured temperature at the center of the combustion zone at the front end of the reactor and the specific heat capacity of carbon dioxide at the current operating pressure, adaptively iterate and calculate the carbon dioxide mass flow rate based on the transient thermal balance constraint equation; determine the target volume fraction of oxygen and carbon dioxide in the mixed gas at the current time step based on the theoretical oxygen mass flow rate and the carbon dioxide mass flow rate, and send the corresponding flow control word to each gas channel of the front-end multi-channel mass flow meter.

[0076] This embodiment further refines the complete calculation and execution process for the instantaneous volume fraction of the oxidant, constructing a standardized proportioning control logic that includes quantitative calculation of oxygen based on enthalpy, calculation of carrier gas through temperature-pressure coupling, iterative thermal balance proportioning, and precise execution by digital control. Compared to existing technologies that rely on fixed-proportion gas supply, empirical manual parameter adjustment, and constant physical property parameter calculations, this embodiment offers the following advantages: It accurately calculates the theoretical oxygen mass flow rate based on the standard reaction enthalpy of carbon oxidation, using thermochemical stoichiometry as the underlying basis, ensuring a strict match between oxygen supply and pyrolysis heat consumption. This upgrades the self-heating system from empirical fuzzy control to precise mechanistic quantitative control. Furthermore, this embodiment dynamically corrects the carbon dioxide specific heat capacity parameter by combining the measured temperature of the front-end combustion zone with the current operating pressure inside the reactor, closely reflecting real-world property changes and effectively eliminating calculation deviations caused by varying operating conditions, thus improving the accuracy of transient thermal balance iterative solutions.

[0077] In some embodiments, the specific execution method of the dual-track PID control law includes: continuously monitoring the oxygen volume fraction at each measuring point along the reactor axis; when the oxygen volume fraction exceeds a preset warning threshold, determining that the CODB boundary is spreading uncontrollably into the pyrolysis zone; sending a voltage signal to the electronic back pressure valve at the output end to reduce the opening, smoothly adjusting the global differential pressure gradient of the reactor to a preset low differential pressure safety threshold; and simultaneously sending an oxygen flow rate reduction command to the front-end multi-channel mass flow meter to reduce the proportion of oxygen intake flow rate, thereby accelerating chemical oxygen consumption in the non-target area and forcibly locking the CODB boundary within the safety limit coordinates.

[0078] This embodiment employs a continuous multi-point oxygen concentration monitoring method along the reactor's axial direction, unlike traditional single-point discrete monitoring. It can continuously characterize the oxygen axial diffusion gradient and boundary migration trajectory. By setting a pre-defined warning threshold, it achieves early instability identification, intervening and controlling the situation before large-scale spread at the boundary and oxidative contamination in the pyrolysis zone, thus upgrading the technology from post-event remediation to pre-event control. This embodiment simultaneously performs physical flow field control and chemical reaction control: by reducing the pressure difference gradient through the back pressure valve, the gas convection mass transfer dynamics are weakened, physically hindering diffusion; by reducing the oxygen flow rate and enhancing chemical oxygen consumption in non-targeted areas, residual oxygen is removed chemically. This dual-track synergy significantly improves the correction efficiency. Furthermore, by smoothly locking the global pressure difference gradient within a low-pressure-difference safety threshold range, problems such as airflow turbulence, localized surges, and severe boundary distortion caused by high pressure differences are avoided. This ensures that the gas inside the reactor remains in a stable, low-speed seepage state, stabilizing the overall temperature and concentration fields, and providing a stable flow environment for the slow, sufficient, and high-quality precipitation of light tar.

[0079] In some embodiments, the in-situ suppression of the carbon dissolution endothermic side reaction using the Le Chatelier principle specifically includes: controlling the electronic back pressure valve at the output end to increase the pressure inside the reactor and maintain it above a preset high pressure threshold; according to the principle of chemical equilibrium shift, using a preset ultra-high pressure physical constraint environment to force the chemical equilibrium of the carbon dissolution endothermic side reaction to shift in the reverse direction, so that the high-temperature carbon dioxide fluid only acts as a physical heat transfer medium to transport heat, blocking the endothermic path of carbon dioxide reacting with carbon to generate carbon monoxide.

[0080] In some embodiments, the specific thermodynamic control process for in-situ suppression of the endothermic side reaction of carbon dissolution is as follows: real-time acquisition of the reaction temperature and partial pressure of carbon dioxide in the pyrolysis zone, and construction of a dynamic comparison mechanism between the reaction quotient and the standard equilibrium constant at that temperature; when it is determined that the reaction is in the endothermic expansion range, the electronic back pressure valve at the output end is controlled to raise the total system pressure to above the critical suppression threshold, so that the critical suppression threshold increases rapidly due to the pressure quotient effect and approaches or exceeds it, and the principle of chemical equilibrium shift is used to force the carbon dissolution reaction equilibrium to shift in the reverse direction, thereby blocking the endothermic side reaction of carbon dioxide and carbon at the thermodynamic source.

[0081] This embodiment utilizes a real-time dynamic comparison mechanism of the standard equilibrium constant correlated with the reaction quotient and temperature to accurately capture the endothermic expansion reaction range, thereby suppressing side reactions as soon as they begin. The identification logic relies on strict thermodynamic mechanisms, ensuring no empirical bias and high reliability. By precisely increasing the total system pressure above the critical suppression threshold, the thermodynamic equilibrium state of the reaction is directly altered, forcing the carbon dissolution endothermic side reaction equilibrium to shift in the reverse direction. This blocks the consumption of carbon matrix and ineffective heat absorption at the reaction mechanism level, thoroughly resolving the core pain points of organic matter loss, thermal balance imbalance, and product destruction during the pyrolysis of oil-rich coal. Furthermore, employing a pure pressure thermodynamic control method, without altering the heating temperature, oxygen ratio, or flow field structure, it specifically suppresses the carbon dissolution endothermic side reaction while fully preserving the optimal operating conditions of the main pyrolysis reaction, achieving a precise control effect of "preserving the main reaction and suppressing the side reaction."

[0082] The following will describe an exemplary application of the embodiments of the present invention in a practical application scenario.

[0083] To address the shortcomings of existing in-situ pyrolysis technologies for rich oil-bearing coal, such as high external heating energy consumption, significant waste of residual coke, easy secondary oxidation of light oil, and lack of intelligent control methods, this invention proposes a multi-field evolution simulation device and intelligent measurement and control system for in-situ self-heating targeted pyrolysis of rich oil-bearing coal. This system utilizes the controllable combustion of in-situ pyrolysis coke to provide a self-sufficient heat source, constructs a pressure-difference-driven combustion-pyrolysis oxygen-isolated dual-zone system, and introduces a multi-field adaptive reverse assimilation operator with embedded physical mechanisms to achieve dynamic parameter optimization, thereby achieving the goal of low-energy consumption and high-purity targeted extraction of light oil. The core hardware architecture of the system consists of the following four collaborative subsystems: (1) Three-dimensional in-situ high-temperature and high-pressure true triaxial reactor main system: The reactor body is integrally forged from high-nickel alloy steel that is resistant to high temperature, high pressure and strong corrosion, and has a sealed cavity inside that can accommodate large-scale specimens. The reactor body is surrounded by a true triaxial hydraulic servo loading frame, which applies three-dimensional independent confining pressure to the specimen through independent axial and opposing flexible pressure heads to accurately simulate the original stress field of deep reservoirs. The injection end of the reactor body is embedded with a micro high-power electric heating matrix, which serves as a controlled heat source for the initial pyrolysis start-up.

[0084] (2) Dynamic precision gas mixing and high-pressure injection system: The system is connected in parallel to high-pressure O2, CO2 and inert N2 protective gas sources at the gas inlet. Each gas source pipeline is connected in series with a multi-stage precision pressure regulating valve and a high-precision digital mass flow meter (MFC). Each channel MFC realizes bidirectional data communication with the intelligent measurement and control center through the fieldbus. It can dynamically adjust the instantaneous flow rate of each component within a millisecond response time according to the analog voltage or digital signal output by the center, so as to realize the precise on-demand ratio of the volume fraction of the mixed oxidant.

[0085] (3) Multi-field, all-time, and distributed sensing and monitoring system: To break the black box limitation of the reaction process in a closed porous medium, an array of sensors is arranged alternately along the fluid axis (migration direction) and radial direction inside the reactor. This array includes densely distributed miniature armored thermocouple groups, high-frequency miniature pressure transmitters, and high-temperature resistant miniature gas component sampling probes. This sensor array can capture multimodal discrete evolution data such as temperature, pressure, and concentration of characteristic components (O2, CO2, CO, CH4) at different spatial coordinates inside the specimen in real time and directly, and synchronously assimilate and transmit all information to the measurement and control center.

[0086] (4) High-pressure back pressure control and separation collection system: The fluid output pipeline of the reactor is sequentially connected to a pneumatic high-pressure condenser, a gas-liquid-solid three-phase separator and collector, and an electrically controlled high-precision electronic back pressure valve (BPR). As the core actuator of the system pressure, the back pressure valve is subject to closed-loop feedback control from the intelligent measurement and control center. Its valve core opening can achieve micron-level displacement fine adjustment, and the dynamic adjustment accuracy of the operating pressure inside the reactor reaches 0.05 MPa, ensuring the precise stability of the pressure field during the reaction process.

[0087] This invention provides a dynamic isolation barrier construction and adaptive measurement and control closed-loop device for in-situ self-heating targeted pyrolysis of oil-rich coal, which specifically includes the following steps: Step S1: Seal and install the oil-rich coal specimen in the main inner cavity of the three-dimensional in-situ true triaxial reactor. Activate the true triaxial hydraulic servo loading frame and apply triaxial stress according to the set depth gradient. The monitoring and control center triggers the micro-electric heating matrix at the injection end to rapidly heat the near-end boundary coal body to the targeted pyrolysis temperature range (400~450 ℃) and maintain a constant temperature. The specimen undergoes initial pyrolysis, and the primary product is extracted from the output end. When the sensor array inside the reactor detects that the gas evolution rate drops below the preset threshold, the electric heating matrix is ​​turned off, leaving a section of initial pyrolysis coke matrix with high porosity and high reactivity in situ at the specimen injection end as a precursor for subsequent self-heating.

[0088] Step S2: Dynamic and precise oxidant gas injection based on the thermal fluid balance equation: The intelligent monitoring and control center abandons the traditional fixed-ratio gas injection mode and controls the mass flow meter (MFC) output based on the system's built-in dynamic thermal fluid balance algorithm. This maintains the target temperature required for the pyrolysis zone. Using the setpoints, based on the real-time thermal conductivity and temperature rise rate of the specimen collected by the sensor array, the theoretical total heat absorption required for the specimen to rise and maintain its temperature at the current time step is calculated. The monitoring and control center operates based on the exothermic reaction of carbon oxidation (C + O2 → CO2 + ... Solve for the theoretical oxygen mass flow rate required per unit time to maintain this self-supplied heat source. : ; In the formula, The heat of reaction is the exothermic reaction of carbon oxidation.

[0089] To prevent excessive localized reactions from causing heterogeneous thermal explosion damage to the specimen, CO2 was introduced as an inert heat transfer medium and dilution medium. The theoretical carbon dioxide mass flow rate... Adaptive iterative solution is performed using the system's built-in transient thermal equilibrium constraint equations: ; In the formula, The specific heat capacity of CO2 at the current operating pressure of the system. This is the measured temperature at the center of the combustion zone at the front end of the vessel.

[0090] After real-time calculation, the control center determines the target volume fractions of O2 and CO2, and sends control commands to the MFC of the corresponding channel to accurately output the instantaneous proportioned mixed gas, ignite the initial pyrolysis, and establish a controlled self-heating oxidation front.

[0091] Step S3: Locking the Critical Oxygen Depletion Barrier (CODB) Space Based on Differential Pressure PID Closed-Loop: During operation, the monitoring and control center continuously monitors the gas composition and temperature data of the axial sensor array. The physical spatial section where the O2 volume fraction decays below the safe threshold and the temperature is in the critical transition zone of oxidation-pyrolysis is defined as the critical oxygen depletion barrier. When the axial component sensor detects that O2 exceeds the set warning measurement point, indicating that the CODB boundary is spreading uncontrollably to the downstream targeted pyrolysis zone and that light tar faces the risk of secondary oxidation and burnout, the monitoring and control center activates the dual-track PID control law: immediately sends a voltage signal to the electronic back pressure valve (BPR) at the output end to reduce the opening, simultaneously reduces the output frequency of the high-pressure injection pump, and smoothly adjusts the global differential pressure gradient ∆P in the reactor to the range of 0.5~0.8 MPa, relying on physical resistance to slow down the gas convection mass transfer rate; at the same time, it reduces the proportion of O2 inlet flow in the MFC to accelerate chemical oxygen consumption. By employing a hardware-integrated strategy of physical convection damping and chemical acceleration consumption, the CODB spatial boundary is forcibly locked within a safe limit coordinate, thereby achieving dynamic isolation and protection of the oxygen-free targeted pyrolysis zone.

[0092] Step S4: When the entire pyrolysis zone at the rear end of the reactor enters the stage of dense precipitation of light coal tar, the intelligent monitoring and control center intervenes in the thermodynamic reaction path, forcibly raising the pressure threshold of the electronic back pressure valve (BPR) at the output end to the set high pressure state (the pressure inside the reactor is not lower than 8~10 MPa). According to Le Chatelier's equilibrium shift principle, the carbon dissolution side reaction (CO2+C→2CO) under high temperature conditions is a strongly endothermic process with an increase in the number of gas molecules. The high-pressure physical constraint environment can force this chemical equilibrium to shift in reverse, thereby blocking the side reaction from plundering the self-supplied heat energy and degrading the quality of the tar in situ inside the reactor. At this time, the high-temperature CO2 fluid only serves as a purely physical heat transfer medium to transport heat, ensuring the stable precipitation of high-purity light coal tar and the gas-liquid-solid separation and collection.

[0093] Step S5: Dynamic Assimilation and Hidden Multi-Field Reverse Reconstruction of Multi-Source Data: Real-time acquisition of discrete multimodal datasets such as distributed temperature, in-situ micro-pressure difference, and multi-component concentration from sensor array feedback, and inputting them into an embedded physical mechanism multi-field adaptive collaborative inversion operator deployed in an edge computing hub. This operator constructs a joint evaluation cost function that includes sensor data residuals and physical partial differential equation penalty terms. : ; in, For sensor data residuals; This is a penalty term in physical partial differential equations; , These are the weighting coefficients for the sensor data residual term and the penalty term of the physical partial differential equation, respectively.

[0094] Energy conservation equation: ; Where ρ is the density of the fluid or porous coal skeleton inside the reactor (kg / m³). 3 ); c p t is the specific heat capacity at constant pressure (J / (kg·K); T is the global three-dimensional continuous temperature field (K); t is the reaction time (s); v is the gas-fluid seepage velocity vector (m / s). Q is the equivalent thermal conductivity of the coal (W / (m·K)); rxn The endogenous heat source term (W / m²) generated by chemical reactions such as carbon oxidation exothermic reaction in the self-heating combustion zone. 3 ).

[0095] Component convection-diffusion equation: ; in, Porosity (dimensionless) of the heterogeneous porous coal medium inside the reactor. The instantaneous component concentration of oxygen in the three-dimensional space inside the reactor (mol / m³) 3 ); t is the reaction time (s); v is the gas-fluid seepage velocity vector flowing through the pore and fracture network of the coal body (m / s); D is the effective diffusion coefficient tensor of oxygen in the porous medium (m 2 / s); R cons The local oxygen consumption rate term (mol / (m) due to the exothermic carbon oxidation reaction during pyrolysis coke) 3 ·s)).

[0096] Darcy's law of seepage: ; Where v is the gas seepage velocity vector (m / s) flowing through the pore and fracture network of the coal body; K is the absolute permeability tensor of the porous coal medium (m 2 ); Let be the dynamic viscosity of the gas mixture (Pa·s); P be the global three-dimensional continuous pressure field distribution. This is the spatial pressure gradient vector (Pa / m) inside the reactor.

[0097] The inversion operator utilizes automatic differentiation to convert the predicted temperature field T output by the network into a single value. pred Concentration field C pred and pressure field P pred By directly differentiating and substituting into the above partial differential equations, explicit residual operators for each field are formed. and NPDE virtual spatiotemporal collocation points are randomly placed inside the reactor. The mean square error of the residuals at each point is calculated. : ; in, The total number of virtual spatiotemporal configuration points deployed within the interior space of the reactor; This represents the spatial three-dimensional coordinate vector and time coordinate of the j-th placement point; The explicit physical residual value generated at this placement point after substituting the predicted value from the neural network inversion into the partial differential equation of energy conservation; The explicit physical residual value generated after substituting the predicted value into the component convection-diffusion equation; 2 This represents the residual norm (i.e., the squared error).

[0098] In the optimization iteration, this measurement and control operator transforms the thermal convection-diffusion equation and the fluid continuity partial differential equation of the heterogeneous medium into underlying forced mathematical constraints, and minimizes them. This forces the mathematical solution of the black box region to rigidly obey the above physical conservation law during the optimization process, thereby enabling high-precision reverse analysis of the global continuous temperature field and CODB spatial coordinate manifold.

[0099] Step S6: The edge monitoring and control center aims to maintain the stability of the CODB barrier within the reactor and maximize the coal sample's position within the targeted pyrolysis temperature range. It constructs a cost function and calculates the optimal hardware execution action for the next time step. It automatically issues digital commands, synchronously adjusting the front-end multi-channel MFC to reduce the instantaneous gas inlet integral by 2%~5%. Simultaneously, it coordinates with the output end's high-precision BPR to smoothly lower the global differential pressure gradient within the reactor by 0.2 MPa. Using physical convection damping, it forces the CODB spatial surface back within safe physical limits, forming a negative feedback closed loop across the entire hardware actuator to ensure long-term stable system operation.

[0100] according to Figure 2 As can be seen, the schematic diagram of the spatial partitioning and multi-field sensing monitoring of the internal specimen of the simulated reactor provided in this embodiment of the invention illustrates the physicochemical decoupling state within the reactor. The left side of the diagram represents the self-heating combustion zone, and the right side represents the oxygen-free targeted pyrolysis zone. The solid line in the middle represents the direction of the global fluid pressure differential gradient established by the mixed gas entering at the injection end, while the critical oxygen-consuming isolation boundary (CODB) serves as the interface between the two zones. The diagram also indicates the precision gas mixing system, the three-dimensional servo hydraulic loading device, the edge monitoring and control center, and the injection and output ends (including the electronic back pressure valve BPR) at the left and right ends.

[0101] according to Figure 3As can be seen, the in-situ thermodynamic inversion and closed-loop control logic diagram based on a multi-field adaptive assimilation operator provided in this embodiment of the invention reveals the hidden thermodynamic field inversion architecture with embedded physical manifold laws. The left side inputs distributed temperature, in-situ dynamic pressure difference, and multi-component concentration data; the core module in the middle, within the "multi-dimensional feature assimilation mapping domain," compares the residuals with the physical penalty gradient, utilizing the THMC multi-field coupled partial differential constraint, namely the joint loss function LOSS of energy conservation, component convection and diffusion, and Darcy flow. PDE The inversion is performed; the right side outputs a 3D continuous temperature field and a dynamic isolation boundary of CODB, and the optimal gas mixture ratio and injection-production pressure difference gradient are obtained by combining the objective function gradient optimization; finally, a closed loop of physical field dynamic evolution reconstruction is formed through hardware execution terminals (MFC and BPR).

[0102] according to Figure 4 As can be seen from the multi-field evolution and carbon suppression reaction mechanism diagram of the isolation barrier provided in this embodiment of the invention; the diagram uses the distance from the injection end as the abscissa to show the spatial evolution of the temperature field distribution (solid line, left vertical axis) and oxygen concentration (dashed line, right vertical axis). It can be seen that the temperature reaches its peak in the combustion zone and then decreases, while the oxygen concentration drops sharply to zero when it reaches the "critical oxygen-consuming isolation boundary". The diagram specifically marks the "in-situ high-temperature carbon suppression mechanism" and its corresponding chemical equation in the high-temperature plateau region (pyrolysis zone) on the right. This explains the thermodynamic mechanism by which the endothermic side reaction is suppressed under high pressure.

[0103] according to Figure 5 As can be seen, the CODB hidden boundary heterogeneous evolution and control planar diagram provided in this embodiment of the invention simulates the 3D distortion of an ideal, straight CODB boundary under heterogeneous disturbance. The arrows in the diagram indicate the BPR correction of the global pressure difference in conjunction with the measurement and control center, demonstrating how the system uses a dual-track PID control law to force correction through physical convection damping, thereby relocking the distorted boundary that has become unstable and spread within the safe limit coordinates.

[0104] according to Figure 6 As can be seen from the multi-dimensional parameter optimization and light oil yield / risk assessment relationship diagram provided in this embodiment of the invention; the diagram uses the oxygen volume fraction at the injection end as the horizontal axis to show the correspondence between light oil yield / purity (solid parabola, left vertical axis) and oxidation damage risk index (dashed line, right vertical axis). The gray shaded area in the diagram indicates the optimal ratio range for gradient optimization in the monitoring and control center, illustrating that by precisely controlling the oxygen ratio, it is possible to effectively control the risk of secondary oxidation while ensuring sufficient heat supply, thereby maximizing the yield of light tar.

[0105] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of the present invention are included within the scope of protection of the present invention.

[0106] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of the invention, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the invention. The sequence numbers of the above-described embodiments of the invention are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0107] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, or apparatus. Without further limitations, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. In the several embodiments provided by this invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components may be combined, or integrated into another system, or some features may be ignored or not performed.

[0108] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A multi-field evolution simulation method for in-situ self-heating targeted pyrolysis of oil-rich coal, characterized in that, The method includes: Multi-source discrete monitoring data of oil-rich coal specimens in a true triaxial reactor are obtained. The multi-source discrete monitoring data is input into a multi-field adaptive collaborative inversion operator deployed in an edge computing center. Dynamic assimilation is performed on the black box flow field of the heterogeneous porous medium inside the reactor. A three-dimensional dynamic control field with global continuity is output. The spatial coordinate manifold of CODB is extracted based on the three-dimensional dynamic control field. Based on the temperature distribution of the three-dimensional dynamic control field and the evolution trend of the spatial coordinate manifold of the CODB, control commands are generated using a dual-track PID control law. When the uncontrolled spread of the CODB boundary to the pyrolysis zone is detected, a first sub-command to reduce the global differential pressure gradient and a second sub-command to reduce the proportion of oxidant intake are generated simultaneously. The spatial position of the CODB is forcibly locked through the linkage of physical convection damping and chemical acceleration consumption. The theoretical heat supply required to maintain the target temperature of the targeted pyrolysis zone is calculated. Based on the thermochemical stoichiometry of the carbon oxidation exothermic reaction, the instantaneous volume fraction of oxidant required to maintain the self-heating source is calculated. The first control command is generated and sent to the front-end multi-channel mass flow meter to adjust the instantaneous volume ratio of oxygen in the mixed gas injected into the reactor. When the pyrolysis zone is detected to have entered the light tar precipitation stage, the electronic back pressure valve at the output end is controlled to construct an ultra-high pressure physical constraint environment in the reactor. The Le Chatelier principle is used to suppress the carbon dissolution endothermic side reaction in situ, thus completing the closed-loop intelligent measurement and control of the in-situ self-heating targeted pyrolysis of oil-rich coal.

2. The method according to claim 1, characterized in that, The multi-field adaptive cooperative inversion operator is constructed through the following steps: A joint evaluation cost function is constructed, which includes sensor data residual terms and physical partial differential equation penalty terms. The physical partial differential equation penalty terms explicitly characterize the energy conservation equation, the component convection-diffusion equation, and Darcy's law of seepage. Using automatic differentiation technology, virtual spatiotemporal configuration points are randomly arranged in the internal space of the reactor. The predicted temperature field, concentration field and pressure field output by the network are directly differentiated and substituted into the THMC coupled partial differential equation to calculate the mean square error of the residuals at each point as the physical constraint loss. By minimizing the joint evaluation cost function, inversion optimization is performed, and the output solution rigidly obeys the physical conservation law. The continuous physical field is then analyzed in reverse using discrete measurement point data.

3. The method according to claim 1, characterized in that, The process of generating the first sub-instruction to reduce the global differential pressure gradient specifically includes: The oxygen volume fraction fed back by the axial component sensor is monitored in real time. When the oxygen volume fraction exceeds the preset warning threshold and the temperature is in the critical transition zone of oxidation-pyrolysis, the CODB boundary is determined to be unstable. Immediately send a voltage signal to the electronic back pressure valve at the output end to reduce the opening, smoothly adjust the global pressure difference gradient from the injection end to the output end in the reactor to the range of 0.5~0.8 MPa, rely on physical resistance to slow down the gas convection mass transfer rate, and in conjunction with the command to reduce the proportion of oxidant inlet gas, forcibly lock the CODB boundary within the safe limit coordinates.

4. The method according to claim 1, characterized in that, The process of generating the second sub-instruction to reduce the oxidant intake ratio specifically includes: Based on the calorific value of the exothermic carbon oxidation reaction and the target maintenance temperature of the targeted pyrolysis zone, the theoretical total heat absorbed by the specimen for heating and self-heating at the current time step is calculated. The theoretical oxygen mass flow rate required to maintain the theoretical total heat is calculated, and carbon dioxide is introduced as an inert heat transfer and dilution medium; The carbon dioxide mass flow rate is adaptively and iteratively calculated based on the transient thermal balance constraint equation, the instantaneous volume fraction of oxidant is dynamically determined, and a control signal is generated to adjust the output of the front-end multi-channel mass flow meter.

5. The method according to claim 1, characterized in that, The specific process for calculating the instantaneous volume fraction of the oxidant is as follows: Based on the theoretical heat supply and the standard enthalpy of reaction of the exothermic carbon oxidation reaction, calculate the theoretical oxygen mass flow rate required to maintain the target temperature of the targeted pyrolysis zone. Carbon dioxide is introduced as an inert heat transfer and dilution medium. Based on the measured temperature at the center of the combustion zone at the front end of the reactor and the specific heat capacity of carbon dioxide under the current operating pressure, the carbon dioxide mass flow rate is adaptively iteratively calculated based on the transient thermal balance constraint equation. Based on the theoretical oxygen mass flow rate and the carbon dioxide mass flow rate, the target volume fractions of oxygen and carbon dioxide in the mixed gas at the current time step are determined, and the corresponding flow control words are sent to each gas channel of the front-end multi-channel mass flow meter.

6. The method according to claim 1, characterized in that, The specific execution method of the dual-track PID control law includes: The oxygen volume fraction at each measuring point is continuously monitored along the axial direction of the reactor. When the oxygen volume fraction exceeds a preset warning threshold, it is determined that the CODB boundary is spreading uncontrollably into the pyrolysis zone. A voltage signal is sent to the electronic back pressure valve at the output end to reduce the opening degree, so as to smoothly adjust the global differential pressure gradient of the reactor to a preset low differential pressure safety threshold. Simultaneously, an oxygen flow rate reduction command is sent to the front-end multi-channel mass flow meter to reduce the proportion of oxygen intake flow rate, thereby accelerating chemical oxygen consumption in the non-target area and forcibly locking the CODB boundary within the safe limit coordinates.

7. The method according to claim 1, characterized in that, The in-situ suppression of endothermic side reactions of carbon dissolution using the Le Chatelier principle specifically includes: The electronic back pressure valve at the output end is controlled to increase the pressure inside the reactor and maintain it above the preset high pressure threshold. Based on the principle of chemical equilibrium shift, the chemical equilibrium of the carbon dissolution endothermic side reaction is forced to shift in the reverse direction by using a preset ultra-high pressure physical constraint environment, so that the high-temperature carbon dioxide fluid can only act as a physical heat transfer medium to transport heat, thus blocking the endothermic path of carbon dioxide reacting with carbon to produce carbon monoxide.

8. The method according to claim 1, characterized in that, The specific thermodynamic control process for in-situ suppression of endothermic side reactions of carbon dissolution is as follows: Real-time acquisition of reaction temperature and carbon dioxide partial pressure in the pyrolysis zone; and construction of a dynamic comparison mechanism between the reaction quotient and the standard equilibrium constant at that temperature. When the reaction is determined to be in the endothermic expansion range, the electronic back pressure valve at the output end is controlled to raise the total system pressure to above the critical inhibition threshold. This causes the critical inhibition threshold to increase rapidly due to the pressure quotient effect and approach or exceed it. By utilizing the principle of chemical equilibrium shift, the equilibrium of the carbon dissolution reaction is forced to shift in the reverse direction, thereby blocking the endothermic side reaction between carbon dioxide and carbon at the thermodynamic source.

Images