CN-121983175-A - Main control factor and collaborative mechanism analysis method of multi-field coupling thermochemical energy storage system

CN121983175ACN 121983175 ACN121983175 ACN 121983175ACN-121983175-A

Abstract

The invention discloses a method for analyzing a main control factor and a cooperative mechanism of a multi-field coupling thermochemical energy storage system, which comprises the steps of obtaining multi-field coupling historical data of a reaction unit in the thermochemical energy storage system, determining a structured multi-field data matrix, constructing a structured variable set comprising a target variable set and an observation variable set, forming an observation information expansion matrix, calculating mutual information and specific mutual information based on the target variable set and the observation variable set, carrying out fine causal decomposition on the specific mutual information, and carrying out cooperative-unique-redundant decomposition on the specific mutual information based on incremental sequencing and recursive allocation of the specific mutual information for each target state to obtain a decomposition result, decomposing the decomposition result into redundant causal contribution, unique causal contribution, causal contribution and causal leakage components, integrating causal contribution of a full-state space based on the decomposition result, and carrying out normalization processing, and identifying main control factors and cooperative mechanisms of the system to generate an optimization suggestion. Systems, electronic devices, and computer-readable storage media are also disclosed.

Inventors

HE XIN
LI XUGUANG
ZHONG HONGWEI
ZHAO JUAN
ZHANG RU
JIA GUANGCAI
ZHANG HENG
HOU XIAODONG
ZHAO TING

Assignees

鄂尔多斯市腾远煤炭有限责任公司
清华大学山西清洁能源研究院

Dates

Publication Date: 20260505
Application Date: 20251230

Claims (10)

1. A main control factor and cooperative mechanism analysis method of a multi-field coupling thermochemical energy storage system realizes causal inference and quantification based on causal component decomposition, and is characterized by comprising the following steps: S1, acquiring multi-field coupling historical data of a reaction unit in a thermochemical energy storage system, and determining a structured multi-field data matrix based on the multi-field coupling historical data, wherein the multi-field coupling historical data is derived from a numerical simulation output or experiment monitoring system and comprises time sequence data under one or more working conditions; S2, constructing a structured variable set, wherein the structured variable set comprises a target variable and an observation variable set, and an observation information expansion matrix is formed based on the observation variable set, and the observation information expansion matrix comprises an observation variable at the current moment and a history state of the observation variable at the current moment; s3, calculating mutual information and specific mutual information based on the target variable and the observation variable set, wherein the specific mutual information is input of fine causal decomposition; S4, performing fine causal decomposition, including performing collaborative-unique-redundancy decomposition on the specific mutual information based on increment ordering and recursion allocation of the specific mutual information to obtain decomposition results, decomposing the specific mutual information into redundant causal contribution, unique causal contribution, collaborative causal contribution and causal leakage component, wherein the redundant causal contribution represents causal contribution of all possible non-empty subsets in an observation information expansion matrix to a target variable in a certain possible state, the unique causal contribution represents information about the certain possible state which is unique by a single variable and cannot be obtained from any other single variable in an observation variable set, the collaborative causal contribution represents additional information about the certain possible state which cannot be obtained from any subset in a variable combination, and the leakage component represents information about the target variable which is commonly participated but cannot be interpreted in the whole observation information expansion matrix in a certain possible state, and the information part is derived from potential or random fluctuation which is not included in the current observation; s5, integrating causal contribution of a full state space based on the decomposition result; s6, carrying out normalization processing on the global causal contribution; And S7, identifying a system main control factor and a cooperative mechanism based on the global causal contribution after normalization processing, and generating an optimization suggestion.
2. The method for analyzing the master factors and the collaborative mechanism of the multi-field-coupled thermochemical energy storage system of claim 1, characterized in that the S1 comprises: S11, acquiring multi-field coupling historical data of a reaction unit in a thermochemical energy storage system, wherein the multi-field coupling historical data comprises temperature field information, water vapor concentration field information, flow velocity field information, pore structure parameters, reaction progress and external thermal mass input conditions, and the external thermal mass input conditions comprise boundary heat flux density, boundary steam pressure, energy storage efficiency and energy release power; S12, preprocessing the multi-field coupling historical data, wherein the preprocessing comprises the steps of deriving evolution data of each physical field on discrete space-time nodes from an established flow-heat-chemical multi-field coupling trans-scale model for data outputted by numerical simulation, analyzing a key process from a pore scale to a reaction unit scale by the flow-heat-chemical multi-field coupling trans-scale model, synchronously acquiring time sequence signals for data of an experiment monitoring system through a sensor network arranged in the reaction unit, performing high-dimensional data characteristic extraction operation on the evolution data of each physical field on the discrete space-time nodes and the time sequence signals, compressing high-dimensional data with spatial distribution of a temperature field, a concentration field and a flow velocity field by adopting a statistical method, and extracting characteristic values capable of representing the overall state of the physical field; S13, determining a structured multi-field data matrix based on the preprocessed multi-field coupling historical data, wherein the method comprises the steps of preprocessing each sensor data in the sensor network through time interpolation and spatial alignment to form the structured multi-field data matrix with space-time matching, and carrying out independent normalization processing on each dimension variable data in the structured multi-field data matrix.
3. The method of analyzing a master factor and a collaborative mechanism of a multi-field coupled thermochemical energy storage system of claim 2, wherein S2 comprises: S21, selecting a core performance index as the target variable; s22, selecting related physical field variables to construct the observation variable set, and forming an observation information matrix containing history information based on the observation variable set.
4. A method of analyzing a master factor and a co-mechanism of a multi-field coupled thermochemical energy storage system as recited in claim 3 wherein S3 comprises: s31, calculating mutual information between the target variable and the observation variable set based on the information theory; S32, further calculating specific mutual information for each possible state of the target variable based on the mutual information.
5. The method of analyzing the master factors and co-mechanisms of a multi-field coupled thermochemical energy storage system of claim 4, wherein S3 further comprises: S33, obtaining the total mutual information through restoration based on the specific mutual information of all possible states of the target variable.
6. A multi-field coupled thermochemical process according to claim 5 the analysis method of the main control factors and the cooperative mechanism of the energy storage system, characterized in that S4 comprises: S41, initializing a residual variable set containing all the observation variable indexes, including creating a list containing all the observation variable indexes, called a residual variable set, wherein the initialized residual variable set is dynamically updated along with the distribution process; s42, calculating specific mutual information of each non-empty subset in the observation information expansion matrix about the target state; S43, arranging each non-empty subset in an ascending order according to the numerical value of specific mutual information, and calculating the information increment between adjacent ordered non-empty subsets in a mode of subtracting the information quantity of the previous subset from the information quantity of the next subset, wherein the information increment is used for representing the information gain brought by introducing new variables or variable combinations on the basis of the existing information; S44, recursively distributing the information increment, wherein the recursively distributing the information increment comprises the steps of sequentially processing each non-empty subset and the corresponding information increment according to the sorting order in S43; S45, defining unique contributions, wherein after the recursion allocation is completed according to S44, items, of which keys are single variables, in all redundant contribution dictionaries are redefined as unique contributions corresponding to the variables; s46, subtracting the sum of all allocated redundancy contributions and the cooperative contribution from the total specific mutual information to obtain the causal leakage component, wherein all allocated redundancy contributions comprise items converted into unique contributions.
7. A multi-field coupled thermochemical process according to claim 6 the analysis method of the main control factors and the cooperative mechanism of the energy storage system, characterized in that S5 comprises: S51, integrating all possible target states into a full state space; s52, carrying out probability weighted average on each possible target state in the full state space and the corresponding decomposition result to obtain a global causal contribution and a global causal leakage reflecting the average causal strength of the multi-field coupling thermochemical energy storage system, wherein the global causal contribution comprises global redundancy, uniqueness and synergistic contribution.
8. A multi-field coupled thermochemical process according to claim 7 the analysis method of the main control factors and the cooperative mechanism of the energy storage system, characterized in that the step S6 includes: s61, normalizing the global causal contribution to a ratio relative to total mutual information; s62, normalizing the causal leakage to a ratio relative to the entropy of the target variable.
9. A multi-field coupled thermochemical according to claim 8 the analysis method of the main control factors and the cooperative mechanism of the energy storage system, characterized in that S7 comprises: S71, identifying unique combinations of the main control factors, the cooperative reinforcement mechanisms and the redundant variables by comparing the magnitudes of the normalized global causal contributions, and evaluating modeling completeness of the thermochemical energy storage system, wherein the criteria for identifying the unique main control factors and the cooperative mechanisms comprise deducing that the identified variables are unique main control factors if the normalized unique contributions of the variables are significantly higher than other variables, identifying that the combination has the cooperative reinforcement mechanism if the normalized cooperative contributions of the variable combinations are significantly higher than zero, indicating that the variable information is overlapped if the normalized redundant contributions of the variable combinations are significantly higher, and being capable of being used for guiding model simplification, prompting that important unobserved variables exist and further perfecting model or experimental observation if the normalized causal leakage is higher; S72, providing optimization suggestions for the optimization design and operation strategy formulation of the thermochemical energy storage reaction unit based on the normalized global causal component, the quantitative recognition conclusion of the main control factors and the cooperative mechanism and combining with the physical mechanism of the reaction unit, wherein the optimization suggestions comprise implementing accurate regulation and control on unique main control factors, designing engineering schemes for promoting cooperative variable group combined action, simplifying a monitoring system or a simulation model based on redundancy analysis, and supplementing key variable observation or deepening mechanism research for the high causal leakage indication direction.
10. A master factor and collaborative mechanism analysis system for a multi-field coupled thermochemical energy storage system for performing the method of any of claims 1-9, comprising: The system comprises a structured multi-field data matrix construction module (101) and a thermal storage system, wherein the structured multi-field data matrix construction module is used for acquiring multi-field coupling historical data of a reaction unit in the thermal chemical energy storage system and determining a structured multi-field data matrix based on the multi-field coupling historical data, wherein the multi-field coupling historical data is derived from a numerical simulation output or experiment monitoring system and comprises time sequence data under one or more working conditions; a structured variable set construction module (102) configured to construct a structured variable set, where the structured variable set includes a target variable and an observation variable set, and form an observation information expansion matrix based on the observation variable set, where the observation information expansion matrix includes an observation variable at a current time and a history state of the observation variable at the current time; the mutual information and specific mutual information calculation module (103) is used for calculating the mutual information and the specific mutual information based on the target variable and the observation variable set, wherein the specific mutual information is input of fine causal decomposition; A fine causal decomposition module (104) for performing fine causal decomposition, comprising, for each target state, decomposing the specific mutual information into redundant causal contributions, unique causal contributions, collaborative causal contributions and causal leakage components based on performing a collaborative-unique-redundancy decomposition on the specific mutual information based on an incremental ordering and a recursive allocation of the specific mutual information, wherein the redundant causal contributions represent causal contributions of all possible non-empty subsets of the observation information extension matrix to the target variable in a certain possible state, the unique causal contributions represent information about the certain possible state that is unique by a single variable and that cannot be obtained from any other single variable in the set of observation variables, the collaborative causal contributions represent additional information about the certain possible state that cannot be obtained from any subset of the set of variables when the plurality of variables in the set of variables are jointly observed, the causal leakage components represent information portions of the entire observation information extension matrix that are jointly involved but unexplained about the target variable in a certain possible state, the information portions originating from potential or purely random observation fluctuation that is not included by the current observation variable; A full state space causal contribution integration module (105) for integrating causal contributions of a full state space based on the decomposition result; -a global causal contribution processing module (106) for normalizing the global causal contribution; And the main control factor and collaboration mechanism analysis module (107) is used for identifying the main control factor and collaboration mechanism of the system based on the global causal contribution after normalization processing and generating an optimization suggestion.

Description

Main control factor and collaborative mechanism analysis method of multi-field coupling thermochemical energy storage system Technical Field The invention relates to the technical field of analysis and optimization of thermochemical energy storage systems, in particular to a method and a system for analyzing main control factors and a collaborative mechanism of a multi-field coupling thermochemical energy storage system, and relates to a technology for identifying main control factors and quantifying collaborative mechanisms of the multi-field coupling thermochemical energy storage system based on causal component decomposition, which is suitable for analyzing main control factors and collaborative strengthening mechanisms of a thermochemical multi-field coupling process in a calcium oxide/calcium hydroxide reaction unit. Background Under the background of energy structure transformation and 'double carbon' targets, the efficient and high-density thermal energy storage technology is a key for realizing renewable energy source absorption and industrial waste heat utilization. Thermochemical energy storage has become a research hot spot by virtue of the advantages of high energy density, long-term storage, low heat loss and the like, wherein a reversible reaction system based on CaO/Ca (OH) 2 is regarded as a very potential technical route due to rich raw materials, moderate reaction temperature and high energy storage density. In practical engineering applications, the energy storage/release performance of the CaO/Ca (OH) 2 reaction unit is significantly affected by the thermal-fluidization multi-physical-field coupling effect. The interior of the reaction unit involves the migration of water vapor in the porous medium, the non-uniform temperature field distribution, the strong coupling of chemical reaction kinetics and heat and mass transfer processes, forming a trans-scale (micro-pore-mesoscopic cluster-macro-unit) nonlinear dynamic system. At present, the performance analysis and design optimization for the system mainly depend on experimental trial-and-error and numerical simulation based on simplified assumptions, and the following prominent bottlenecks exist: 1. The multi-field coupling mechanism is unknown, the synergy and redundancy effect are difficult to quantify, and the traditional analysis method focuses on the influence of a single physical field or a few variables, and lacks systematic quantification of synergy enhancement, unique contribution or redundancy counteracting effect among heat-flow-chemical multi-fields. For example, it is not clear whether a temperature gradient and water vapor diffusion are synergistic to promote the reaction or information redundancy exists, and it is difficult to identify what factors among heat transfer, mass transfer and reaction resistance dominate the energy storage/release rate under specific working conditions. 2. The cross-scale association is complex, and the main control factor identification relies on experience that a cross-scale association mechanism from a micro pore structure to macroscopic energy storage/release performance is not fully disclosed. The traditional method is often used for factor screening through parameter sensitivity analysis or statistical regression, and the real contribution and coupling effect of each factor cannot be distinguished from the causal inference angle, so that the design direction of the energy storage power density enhancement of the reaction unit is lack of pertinence, and the accurate enhancement is difficult to realize. 3. The nonlinear and unsteady process lacks a dynamic causal analysis tool, namely the reaction unit is coupled with multiple time scales in the transient process of energy storage/release, the traditional steady-state or quasi-steady-state model is difficult to capture the evolution rule of dominant factors in the dynamic process, and the dynamic adjustment of a multi-field cooperative mechanism under different time scales cannot be quantized. In view of the above challenges, fusing causal inference of information theory with multi-field coupled system analysis is an important way to break through bottlenecks. In recent years, causal discovery methods such as glabellar causality, transfer entropy, etc. have been introduced into engineering system analysis, but it is still difficult to deal with synergistic and redundant effects between multiple variables, and the robustness to non-linear, non-gaussian noise and unobserved variables (such as microstructure evolution) is insufficient. Therefore, in order to deeply disclose the collaborative enhancement mechanism of the heat-flow-chemical multi-field coupling in the CaO/Ca (OH) 2 reaction unit, and realize the spanning from qualitative description to quantitative attribution, a technology capable of systematically decomposing and quantifying the collaborative, unique and redundant causal contribution of each physical field is urgently needed, a c