Gas quality prediction method and device for gas output outside gas storage and electronic equipment

Abstract

The invention provides a gas quality prediction method and device for gas conveyed outside a gas storage and electronic equipment, and belongs to the technical field of gas storage. The method comprises the following steps: acquiring to-be-tested operating parameters of gas conveyed outside the gas storage; inputting the to-be-measured operation parameters into the gas quality prediction model to obtain a gas quality prediction result, output by the gas quality prediction model, of gas conveyed outside the gas storage; the gas quality prediction model is constructed based on a fuzzy cognitive map model, and the fuzzy cognitive map model is obtained by fusing a first fuzzy cognitive map sub-model and a second fuzzy cognitive map sub-model under different working conditions; the first fuzzy cognitive map sub-model is constructed based on expert experience; the second fuzzy cognitive map sub-model of each working condition is obtained based on training of multiple types of sample operation parameters under a single working condition. The method is used for overcoming the defect that a traditional mechanism modeling method and a data-driven modeling method are difficult to meet actual high-precision gas quality prediction requirements.

Description

Technical Field

[0001] This invention relates to the field of gas storage technology, specifically to a method for predicting the gas quality of gas exported from a gas storage facility, a device for predicting the gas quality of gas exported from a gas storage facility, an electronic device, a machine-readable storage medium, and a computer program product. Background Technology

[0002] Natural gas from gas storage facilities typically begins by producing natural gas at the well site and transporting it to the gathering and injection station via the production trunk line. The natural gas then enters a pre-separator for three-phase separation, followed by cooling in a pre-cooler before entering a production separator for two-phase separation. The separated natural gas then enters a coiled-tube heat exchanger for heat exchange. After pressure regulation, the natural gas enters a cryogenic separator for two-phase separation. The qualified natural gas separated in the cryogenic separator is then cooled in a coiled-tube heat exchanger before being pressure regulated and exported. By modeling the gas quality (GC) of the gas exported from the storage facility, the changing trends of the GC can be predicted, providing a basis for the design, operation, and control of the gas export system and achieving stable GC quality for the gas exported from the storage facility.

[0003] Existing methods for modeling the gas quality of gas transported from gas storage facilities mainly include mechanistic modeling and data-driven modeling. Mechanistic modeling first establishes a physicochemical model describing processes such as gas-liquid-oil separation, gas cooling, gas-liquid separation, and gas compression and diffusion. Then, it uses historical data generated by the gas transport system to identify the parameters of the physicochemical model, enabling gas quality modeling and prediction. Data-driven modeling, as an important approach to gas quality modeling in gas transport from gas storage facilities, focuses on using historical data and machine learning algorithms to predict and analyze gas quality and behavior. Compared to traditional mechanistic modeling based on physicochemical principles, this method can handle large amounts of historical data. By learning patterns and trends from this data, it can predict gas quality even without detailed prior knowledge of the physical processes.

[0004] However, due to deviations in the temperature, pressure, and physical properties of the gas produced from the wellhead of the gas storage external transmission system, and the time-varying characteristics of the gas storage reservoir pressure, temperature, and water content, the input data may have deviations or insufficient coverage. This makes it difficult for traditional mechanism modeling methods and data-driven modeling methods to meet the requirements of actual high-precision gas quality prediction. Summary of the Invention

[0005] The purpose of this invention is to provide a method, device, and electronic equipment for predicting the gas quality of gas transported from a gas storage facility, in order to address the shortcomings of traditional mechanism modeling methods and data-driven modeling methods in meeting the requirements of high-precision gas quality prediction in practice.

[0006] To achieve the above objectives, embodiments of the present invention provide a method for predicting the gas quality of gas transported from a gas storage facility, comprising:

[0007] Obtain the operating parameters to be measured for gas transmission from the gas storage facility;

[0008] The operating parameters to be measured are input into the gas quality prediction model to obtain the gas quality prediction results of the gas storage external gas output by the gas quality prediction model; the gas quality prediction model is constructed based on the fuzzy cognitive graph model, which is obtained by fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model for different operating conditions; the first fuzzy cognitive graph model is constructed based on expert experience; the second fuzzy cognitive graph model for each operating condition is trained based on the operating parameters of multiple types of samples under a single operating condition.

[0009] Optionally, the fuzzy cognitive graph model is obtained based on the following steps:

[0010] Obtain various types of sample operating parameters for external gas transmission from the gas storage facility; the various types of sample operating parameters include a first type of sample operating parameters that can be used to obtain the correlation between sample operating parameters based on expert experience, and a second type of sample operating parameters that are the remaining sample operating parameters excluding the first type of sample operating parameters.

[0011] Preprocess the running parameters of all types of samples;

[0012] Based on expert experience, a fuzzy cognitive graph model is constructed for the operating parameters of the first type of sample to obtain the first fuzzy cognitive graph sub-model.

[0013] Clustering is performed on the second type of sample operating parameters to obtain multiple sample operating parameter sets under different working conditions; each sample operating parameter set corresponds to a working condition and includes multiple types of second type sample operating parameters.

[0014] For each set of sample operating parameters, the weight matrix between various types of sample operating parameters is solved to obtain the second fuzzy cognitive graph sub-model for different working conditions;

[0015] The first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different working conditions are fused to obtain the fuzzy cognitive graph model.

[0016] Optionally, the step of constructing a fuzzy cognitive graph model based on expert experience for the operating parameters of the first type of samples to obtain a first fuzzy cognitive graph sub-model includes:

[0017] Multiple sample operation parameters in the first type of sample operation parameters are determined as concept nodes of the first fuzzy cognitive graph sub-model;

[0018] The correlation between the multiple sample operating parameters is obtained based on expert experience, and the correlation between the multiple sample operating parameters is determined as the weight matrix of the first fuzzy cognitive graph sub-model.

[0019] The first fuzzy cognitive graph sub-model is constructed based on the multiple sample operating parameters and the correlation between the multiple sample operating parameters.

[0020] Optionally, the step of solving for the weight matrix between various types of sample operating parameters for each set of sample operating parameters to obtain the second fuzzy cognitive graph sub-model for different operating conditions includes:

[0021] Multiple types of sample operation parameters from the first sample operation parameter set are identified as concept nodes in the second fuzzy cognitive graph sub-model; the first sample operation parameter set can be any one of multiple sample operation parameter sets;

[0022] The weight matrix of the first sample running parameter set is obtained by solving the weight matrix between various types of sample running parameters based on the least squares method, and the weight matrix of the first sample running parameter set is used as the weight matrix of the second fuzzy cognitive graph sub-model.

[0023] Based on the sample operation parameters and weight matrix of the first sample operation parameter set, a second fuzzy cognitive graph sub-model under a single working condition is constructed.

[0024] Optionally, fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-models for different working conditions to obtain the fuzzy cognitive graph model includes:

[0025] The weight matrix of the first fuzzy cognitive graph sub-model and the weight matrix of the second fuzzy cognitive graph sub-model under different working conditions are fused based on the fuzzy TS model to obtain the fuzzy cognitive graph model.

[0026] Optionally, the fusion of the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model to obtain the fuzzy cognitive graph model includes:

[0027] When the intersection of the concept node sets of fuzzy cognitive sub-model A and fuzzy cognitive sub-model B is an empty set, the weight matrix of the fuzzy cognitive graph model is represented by the following formula:

[0028]

[0029] Among them, W global W represents the weight matrix of the fuzzy cognitive graph model. A ij W represents the weight matrix of the fuzzy cognitive sub-model A.B ij Let A represent the weight matrix of fuzzy cognitive sub-model B; fuzzy cognitive sub-model A represents either the first fuzzy cognitive graph sub-model or the second fuzzy cognitive graph sub-model, and fuzzy cognitive sub-model B represents either the first fuzzy cognitive graph model or the second fuzzy cognitive graph model.

[0030] Optionally, the fusion of the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model to obtain the fuzzy cognitive graph model includes:

[0031] When the concept node sets of the first fuzzy cognitive graph sub-model and the concept node sets of the second fuzzy cognitive graph sub-model intersect, or when the concept node sets of the second fuzzy cognitive graph sub-models under different working conditions intersect, the weight matrix of the fuzzy cognitive graph model is represented by the following formula:

[0032]

[0033] Among them, W global This represents the weight matrix of the fuzzy cognitive graph model, where m is the number of fuzzy cognitive graph sub-models in the fuzzy cognitive graph model, and W... k ij This represents the concept node C in the k-th (k = 1, 2, ..., m) first or second fuzzy cognitive graph sub-model. j Pointing to C i The connection weights.

[0034] Optionally, the preprocessing of the sample running parameters for all types includes:

[0035] Perform data standardization and / or data normalization on the operating parameters of all types of samples.

[0036] Optionally, the various types of sample operating parameters include at least two of the following: temperature, pressure, flow rate, liquid level, gas impurity content, and gas composition of the gas storage external gas transmission system.

[0037] On the other hand, embodiments of the present invention also provide a gas quality prediction device for external gas transmission from a gas storage facility, comprising:

[0038] The acquisition module is used to acquire the operating parameters to be measured for the external gas transmission of the gas storage facility;

[0039] The prediction module is used to input the operating parameters to be measured into the gas quality prediction model to obtain the gas quality prediction result of the gas storage external gas output by the gas quality prediction model; the gas quality prediction model is constructed based on the fuzzy cognitive graph model, which is obtained by fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model for different operating conditions; the first fuzzy cognitive graph model is constructed based on expert experience; the second fuzzy cognitive graph model for each operating condition is trained based on the operating parameters of multiple types of samples under a single operating condition.

[0040] Optionally, the fuzzy cognitive graph model is obtained based on the following steps:

[0041] Obtain various types of sample operating parameters for external gas transmission from the gas storage facility; the various types of sample operating parameters include a first type of sample operating parameters that can be used to obtain the correlation between sample operating parameters based on expert experience, and a second type of sample operating parameters that are the remaining sample operating parameters excluding the first type of sample operating parameters.

[0042] Preprocess the running parameters of all types of samples;

[0043] Based on expert experience, a fuzzy cognitive graph model is constructed for the operating parameters of the first type of sample to obtain the first fuzzy cognitive graph sub-model.

[0044] Clustering is performed on the second type of sample operating parameters to obtain multiple sample operating parameter sets under different working conditions; each sample operating parameter set corresponds to a working condition and includes multiple types of second type sample operating parameters.

[0045] For each set of sample operating parameters, the weight matrix between various types of sample operating parameters is solved to obtain the second fuzzy cognitive graph sub-model for different working conditions;

[0046] The first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different working conditions are fused to obtain the fuzzy cognitive graph model.

[0047] Optionally, the step of constructing a fuzzy cognitive graph model based on expert experience for the operating parameters of the first type of samples to obtain a first fuzzy cognitive graph sub-model includes:

[0048] Multiple sample operation parameters in the first type of sample operation parameters are determined as concept nodes of the first fuzzy cognitive graph sub-model;

[0049] The correlation between the multiple sample operating parameters is obtained based on expert experience, and the correlation between the multiple sample operating parameters is determined as the weight matrix of the first fuzzy cognitive graph sub-model.

[0050] The first fuzzy cognitive graph sub-model is constructed based on the multiple sample operating parameters and the correlation between the multiple sample operating parameters.

[0051] Optionally, the step of solving for the weight matrix between various types of sample operating parameters for each set of sample operating parameters to obtain the second fuzzy cognitive graph sub-model for different operating conditions includes:

[0052] Multiple types of sample operation parameters from the first sample operation parameter set are identified as concept nodes in the second fuzzy cognitive graph sub-model; the first sample operation parameter set can be any one of multiple sample operation parameter sets;

[0053] The weight matrix of the first sample running parameter set is obtained by solving the weight matrix between various types of sample running parameters based on the least squares method, and the weight matrix of the first sample running parameter set is used as the weight matrix of the second fuzzy cognitive graph sub-model.

[0054] Based on the sample operation parameters and weight matrix of the first sample operation parameter set, a second fuzzy cognitive graph sub-model under a single working condition is constructed.

[0055] Optionally, fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-models for different working conditions to obtain the fuzzy cognitive graph model includes:

[0056] The weight matrix of the first fuzzy cognitive graph sub-model and the weight matrix of the second fuzzy cognitive graph sub-model under different working conditions are fused based on the fuzzy TS model to obtain the fuzzy cognitive graph model.

[0057] Optionally, the fusion of the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model to obtain the fuzzy cognitive graph model includes:

[0058] When the intersection of the concept node sets of fuzzy cognitive sub-model A and fuzzy cognitive sub-model B is an empty set, the weight matrix of the fuzzy cognitive graph model is represented by the following formula:

[0059]

[0060] Among them, W global W represents the weight matrix of the fuzzy cognitive graph model. A ij W represents the weight matrix of the fuzzy cognitive sub-model A. B ijLet A represent the weight matrix of fuzzy cognitive sub-model B; fuzzy cognitive sub-model A represents either the first fuzzy cognitive graph sub-model or the second fuzzy cognitive graph sub-model, and fuzzy cognitive sub-model B represents either the first fuzzy cognitive graph model or the second fuzzy cognitive graph model.

[0061] Optionally, the fusion of the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model to obtain the fuzzy cognitive graph model includes:

[0062] When the concept node sets of the first fuzzy cognitive graph sub-model and the concept node sets of the second fuzzy cognitive graph sub-model intersect, or when the concept node sets of the second fuzzy cognitive graph sub-models under different working conditions intersect, the weight matrix of the fuzzy cognitive graph model is represented by the following formula:

[0063]

[0064] Among them, W global This represents the weight matrix of the fuzzy cognitive graph model, where m is the number of fuzzy cognitive graph sub-models in the fuzzy cognitive graph model, and W... k ij This represents the concept node C in the k-th (k = 1, 2, ..., m) first or second fuzzy cognitive graph sub-model. j Pointing to C i The connection weights.

[0065] Optionally, the preprocessing of the sample running parameters for all types includes:

[0066] Perform data standardization and / or data normalization on the operating parameters of all types of samples.

[0067] Optionally, the various types of sample operating parameters include at least two of the following: temperature, pressure, flow rate, liquid level, gas impurity content, and gas composition of the gas storage external gas transmission system.

[0068] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-mentioned method for predicting the gas quality of gas exported from the gas storage facility.

[0069] On the other hand, the present invention also provides a machine-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for predicting the gas quality of gas exported from the gas storage facility.

[0070] On the other hand, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-mentioned method for predicting the gas quality of gas exported from the gas storage facility.

[0071] Through the above technical solution, this invention fuses a first fuzzy cognitive graph sub-model constructed based on expert experience and a second fuzzy cognitive graph sub-model trained with sample operating parameters under different operating conditions to obtain a fuzzy cognitive graph model. Then, based on the fuzzy cognitive graph model, it predicts the gas quality of gas transported from the gas storage facility. This invention utilizes the interpretability of fuzzy cognitive graph model reasoning to reveal the impact of various operating parameters of the gas storage facility's external gas transport system on gas quality. Furthermore, by leveraging the good reasoning ability of the fuzzy cognitive graph model, it achieves high-precision prediction of the gas quality of gas transported from the gas storage facility.

[0072] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0073] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0074] Figure 1 This is one of the flowcharts illustrating the gas quality prediction method for external gas transmission from a gas storage facility provided by the present invention.

[0075] Figure 2 This is the second flowchart of the gas quality prediction method for external gas transmission from a gas storage facility provided by the present invention.

[0076] Figure 3 This is the third flowchart of the gas quality prediction method for external gas transmission from a gas storage facility provided by the present invention;

[0077] Figure 4 This is the fourth flowchart of the gas quality prediction method for external gas transmission from a gas storage facility provided by the present invention.

[0078] Figure 5 This is a schematic diagram of the gas quality prediction device for external gas transmission from a gas storage facility provided by the present invention.

[0079] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0080] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0081] During the gas production period, the gas quality of the external gas transmission system from the gas storage facility experiences significant fluctuations. These fluctuations are caused by several factors, including mismatches between the temperature, pressure, and physical properties of the produced gas from the wellhead and the design specifications; complex geological conditions at the gas storage facility; changes in reservoir pressure, temperature, and water cut over time; and inaccurate control of the transmission system's operating parameters. These fluctuations have a crucial impact on the safe operation of the natural gas pipeline network, the automated control of the gas transmission system, and the end-user experience. Modeling the gas quality of the external transmission system allows for the prediction of its changing trends, providing a basis for the design, operation, and control of the system and ultimately ensuring stable gas quality.

[0082] Existing gas quality modeling methods for external gas transmission from gas storage facilities mainly include mechanistic modeling and data-driven modeling. Mechanistic modeling first establishes a physicochemical model describing processes such as gas-liquid-oil separation, gas cooling, gas-liquid separation, and gas compression and diffusion. Then, it uses historical data generated by the gas transmission system to identify the parameters of the physicochemical model, achieving gas quality modeling and prediction. However, this modeling method is highly dependent on the accuracy of the physicochemical model parameters and requires a large amount of experimental data and complex calculations. In practice, these parameters may be affected by various factors, leading to uncertainty in model predictions. Data-driven modeling, as an important approach to gas quality modeling for external gas transmission from gas storage facilities, focuses on using historical data and machine learning algorithms to predict and analyze gas quality and behavior. Compared to traditional mechanistic modeling based on physicochemical principles, this method can handle large amounts of historical data. By learning patterns and trends in this data, it can predict gas quality even without detailed prior knowledge of physical processes. However, the effectiveness of this model is highly dependent on the quality of the available data. If the input data is biased or has insufficient coverage, the model's predictive ability may be severely affected. Secondly, data-driven models can sometimes become a "black box," even if they can accurately predict the results, it is difficult to explain why such results are obtained, which means that such data-driven models lack interpretability.

[0083] In summary, due to deviations in the temperature, pressure, and physical properties of the produced gas stream from the wellhead of the gas storage external transmission system, and the time-varying characteristics of the gas storage reservoir pressure, temperature, and water content, the input data may have biases or insufficient coverage. As a result, traditional mechanism modeling methods and data-driven modeling methods are difficult to meet the requirements of actual high-precision gas quality prediction.

[0084] The purpose of this invention is to provide a method, device, and electronic equipment for predicting the gas quality of gas transported from a gas storage facility, in order to address the shortcomings of traditional mechanism modeling methods and data-driven modeling methods in meeting the requirements of high-precision gas quality prediction in practice.

[0085] Method Implementation Examples

[0086] Please refer to Figure 1 This invention provides a method for predicting the gas quality of gas transported from a gas storage facility, comprising:

[0087] Step 100: Obtain the operating parameters to be measured for the external gas transmission of the gas storage facility.

[0088] The technical concept of this invention is to establish a gas quality modeling method for gas transmission from a gas storage facility based on fuzzy cognitive map (FCM) reasoning. A fuzzy cognitive map (FCM) model is a graph-based cognitive model capable of describing and reasoning about complex causal relationships between factors in a system. First, electronic equipment acquires the operational parameters (or process parameters) to be measured for the gas transmission from the gas storage facility. These operational parameters can be key factors affecting the gas quality transmitted from the gas storage facility, determined through expert knowledge and historical data analysis, such as temperature, pressure, flow rate, liquid level, and gas impurity content in the gas transmission system.

[0089] Step 200: Input the operating parameters to be measured into the gas quality prediction model to obtain the gas quality prediction results of the gas storage external gas output by the gas quality prediction model.

[0090] The electronic device inputs the operating parameters to be measured into the gas quality prediction model, and obtains the gas quality prediction result of the gas storage facility's external gas output from the gas quality prediction model. The gas quality prediction model is constructed based on a fuzzy cognitive graph model, which is obtained by fusing a first fuzzy cognitive graph sub-model and second fuzzy cognitive graph sub-models for different operating conditions. The first fuzzy cognitive graph model is constructed based on expert experience. The second fuzzy cognitive graph model for each operating condition is trained based on multiple types of sample operating parameters under a single operating condition.

[0091] For example, the first fuzzy cognitive graph sub-model can use the storage temperature, gas composition, and gas impurity content of a gas storage facility as concept nodes. Then, based on domain knowledge and historical data, the causal relationships between these concept nodes are determined. For example, the storage temperature of a gas storage facility may affect the gas composition and impurity content. Next, a weight is assigned to each causal relationship to represent the degree of influence of that relationship on gas quality prediction. For example, the weight between temperature and gas composition is 0.7, and the weight between temperature and gas impurity content is 0.8. The weight values ​​typically range from [-1, 1], where positive numbers represent positive effects and negative numbers represent negative effects.

[0092] The structure of the fuzzy cognitive graph model for gas quality in gas storage facilities under different operating conditions essentially aims to establish the correlation between various operating parameters and gas quality, which forms the basis for subsequent model fusion. However, gas storage facility gas transmission systems often involve numerous operating parameters with extremely complex relationships. Therefore, the fuzzy cognitive graph model constructed for such a complex dynamic subsystem may contain a large number of concept nodes, and the connections between these concept nodes are also quite complex, posing a significant challenge to constructing the corresponding fuzzy cognitive graph model structure. Please refer to... Figure 2 Furthermore, among the various operating parameters of the gas storage facility's external gas transmission system, there exist relationships that can be clearly described by mechanisms and expert experience, as in the first fuzzy cognitive graph sub-model. However, there are also relationships that cannot be clearly explained by mechanisms. For the former, the correlation between conceptual nodes can be directly obtained based on mechanisms and expert experience. However, for the latter, it is necessary to consider how to integrate the fuzzy rules such as causal rules, correlation rules, and temporal rules obtained from the big data analysis of gas quality in the external gas transmission system with the correlations obtained from mechanisms and expert experience. Finally, a complete fuzzy cognitive graph model structure can be formed that describes the essential dynamic characteristics of the gas quality evolution under various operating conditions.

[0093] The second fuzzy cognitive graph sub-model can identify key factors affecting the gas quality of gas supplied from the gas storage facility through expert knowledge and historical data analysis. These factors include sample operating parameters such as temperature, pressure, and gas moisture content in the gas supply system, and are used as concept nodes in the second fuzzy cognitive graph sub-model. The weight matrix of the second fuzzy cognitive graph sub-model for each operating condition is trained based on multiple types of sample operating parameters under a single operating condition. During training, optimization algorithms such as genetic algorithms and particle swarm optimization can be used to search for the optimal weight combination. In this embodiment, the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-models for different operating conditions are fused to obtain a fuzzy cognitive graph model. Through fuzzy inference, the fuzzy cognitive graph model can predict the changing trend of the gas quality of the supplied gas under different operating conditions (different operating parameters), thereby achieving precise control of the gas quality of the supplied gas. This embodiment can more accurately predict gas quality changes, providing a more reliable basis for the design, operation, and control of the gas supply system of the gas storage facility.

[0094] This invention fuses a first fuzzy cognitive graph sub-model constructed based on expert experience and a second fuzzy cognitive graph sub-model trained with sample operating parameters under different operating conditions to obtain a fuzzy cognitive graph model. This fuzzy cognitive graph model is then used to predict the gas quality of gas transported from the gas storage facility. This invention leverages the interpretability of fuzzy cognitive graph model reasoning to reveal the impact of various operating parameters of the gas storage facility's external gas transport system on gas quality. Furthermore, by utilizing the strong reasoning capabilities of the fuzzy cognitive graph model, it achieves high-precision prediction of the gas quality of gas transported from the gas storage facility.

[0095] In other aspects of the embodiments of the present invention, the fuzzy cognitive graph model is obtained based on the following steps:

[0096] Step 10: Obtain various types of sample operating parameters for external gas transmission from the gas storage facility; the various types of sample operating parameters include a first type of sample operating parameters that can be used to obtain the correlation between sample operating parameters based on expert experience, and a second type of sample operating parameters that are the remaining sample operating parameters excluding the first type of sample operating parameters.

[0097] Step 20: Preprocess the running parameters of all types of samples.

[0098] Step 30: Based on expert experience, construct a fuzzy cognitive graph model for the operating parameters of the first type of sample to obtain the first fuzzy cognitive graph sub-model.

[0099] Step 40: Cluster the second type of sample operating parameters to obtain multiple sample operating parameter sets under different operating conditions.

[0100] Step 50: Solve the weight matrix between various types of sample operating parameters for each set of sample operating parameters to obtain the second fuzzy cognitive graph sub-model for different operating conditions.

[0101] Step 60: Fuse the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different working conditions to obtain the fuzzy cognitive graph model.

[0102] In one embodiment, based on historical data of operating parameters generated by the gas storage facility's external gas transmission system, canonical correlation analysis, Granger's theory of causal association between multivariate time series variables, and Pearl's Bayesian causal relationship method are comprehensively applied to discover the relationships between the external gas transmissions of various gas storage facilities within the system, determining their coupling, independence, and causal relationships. Subsequently, a dynamic factor model is used to map the multivariate dynamic data of gas quality under different operating conditions of the external gas transmission system into univariate dynamic data that fully reflects its changing patterns. Clustering and other methods are then used to segment this univariate dynamic data, and the results of the univariate data segmentation are transformed into window partitions of the multivariate dynamic data. Then, relevant rules are extracted from each partition window to describe the dynamic trend relationships between variables within the system. To overcome the large number of redundant and contradictory rules caused by the exponential increase in rules extracted from multivariate data due to the increased observation time of complex dynamic systems, probability statistics and information entropy methods are applied to eliminate redundant and contradictory fuzzy rules, improving the accuracy, scalability, and interpretability of the discovered dynamic trend relationships between operating parameters within the external gas transmission system. Finally, we consider drawing on probability theory to determine the accuracy and importance of the rules describing the causal and dynamic trends among various operating parameters within the gas storage facility's external gas transmission system, based on data obtained through mechanisms and expert evaluation. We then use this as a basis for weighted processing to achieve the integration of knowledge about the gas quality evolution laws of the external gas transmission system obtained from different knowledge sources.

[0103] Specifically, firstly, electronic equipment collects historical data. This involves acquiring various types of sample operating parameters from the historical data of the gas storage facility's external gas transmission. These various types of sample operating parameters include a first category of parameters whose correlations can be determined based on expert experience, and a second category of parameters excluding the first category. The first category of sample operating parameters includes storage temperature, gas composition, and gas impurity content. The second category of sample operating parameters represents the gas storage facility's external gas transmission system operating parameters for which correlations cannot be determined based on expert experience. This includes at least two of the following: temperature, pressure, flow rate, liquid level, gas impurity content, and gas composition. For example, it may include the temperature, pressure, flow rate, liquid level, and gas moisture content of the gas storage facility's external gas transmission system. Based on the key operating parameters of the gas storage facility's external gas transmission, including temperature, pressure, flow rate, liquid level, and gas moisture content, the historical database of the gas storage facility's external gas transmission system is used to extract the gas quality data feature points through subsequent algorithms. This allows for a more organized and effective understanding of the structure and patterns of the gas storage facility's external gas transmission system's operating data, further identifying and revealing potential patterns and trends in the gas quality data, discovering new knowledge and insights that were previously unrecognized in the data, and classifying the operating conditions of the gas storage facility's external gas transmission accordingly.

[0104] Next, the electronic device preprocesses the operating parameters of all types of samples. For example, preprocessing the operating parameters of all types of samples includes data standardization and / or data normalization. In one embodiment, the electronic device performs data standardization and data normalization on the operating parameters of all types of samples to obtain the data needed to construct the fuzzy cognitive graph model. For example, the electronic device scans data in the historical gas transmission database of the gas storage facility to ensure that the data is free of noise or missing values, and selects or extracts the operating parameters that best reflect the data distribution and clusters. Subsequently, the operating parameters are standardized to avoid the influence of features at different scales on the clustering results, i.e.:

[0105]

[0106] Where Z is the standardized sample operating parameter, X is the original data, μ is the mean of the sample operating parameter, and θ is the standard deviation of the sample operating parameter.

[0107] Then, the electronic device constructs a fuzzy cognitive graph model based on expert experience for the operating parameters of the first type of samples, obtaining a first fuzzy cognitive graph sub-model. In one embodiment, constructing a fuzzy cognitive graph model based on expert experience for the operating parameters of the first type of samples, obtaining a first fuzzy cognitive graph sub-model, includes: determining multiple sample operating parameters in the operating parameters of the first type of samples as concept nodes of the first fuzzy cognitive graph sub-model; obtaining the correlation between the multiple sample operating parameters based on expert experience, and determining the correlation between the multiple sample operating parameters as the weight matrix of the first fuzzy cognitive graph sub-model; and constructing the first fuzzy cognitive graph sub-model based on the multiple sample operating parameters and the correlation between the multiple sample operating parameters.

[0108] For example, the first fuzzy cognitive graph sub-model can use the storage temperature, gas composition, and gas impurity content of the gas storage tank as concept nodes. The weight matrix of the first fuzzy cognitive graph sub-model is set as follows: a weight of 0.7 between temperature and gas composition, and a weight of 0.8 between temperature and gas impurity content.

[0109] Next, please refer to Figure 3 The electronic equipment performs clustering processing on the second type of sample operating parameters to obtain multiple sample operating parameter sets under different operating conditions. For example, clustering the second type of sample operating parameters yields sample operating parameter sets for operating condition 1, operating condition 2, ..., operating condition n. Each sample operating parameter set corresponds to one operating condition, and each sample operating parameter set includes multiple types of second type sample operating parameters. For example, the sample operating parameter sets include temperature, pressure, physical properties, and gas impurity content of the gas storage external gas transmission system.

[0110] Specifically, the electronic device can select k sample operating parameters as initial cluster centers. This embodiment of the invention can employ various initialization methods, including random selection and K-means++. Next, the sample operating parameters are assigned to the nearest cluster centers, using the Euclidean distance formula to assign a cluster center to each sample operating parameter:

[0111]

[0112] Where: d(x,y) is the distance between the sample running parameters x and y, n is the number of features, and x is the distance between the sample running parameters. i and y i The value on the i-th feature. Then update the cluster centers: for each cluster, calculate the mean of the running parameters of all its samples and set it as the new cluster center:

[0113]

[0114] Where: μ jIt is the new center of the j-th cluster, n j x is the number of sample runtime parameters in the j-th cluster. i This refers to the running parameters of the i-th sample in the j-th cluster. Finally, check if there are any significant changes in the cluster centers. If so, reassign the cluster centers; otherwise, continue.

[0115] To evaluate clustering results, the sum of distances from each data point within a cluster to its cluster center is first calculated, followed by the silhouette coefficient to assess the quality of the clustering.

[0116]

[0117] Where: s(i) is the silhouette coefficient of sample operating parameter i, a(i) is the average distance from sample operating parameter i to other sample operating parameters in its cluster, and b(i) is the average distance from sample operating parameter i to all points in the nearest other cluster. The gas quality data rules for gas transmission from gas storage facilities discovered through clustering are stored in the knowledge base. This completes the classification of gas quality data operating conditions for gas transmission from gas storage facilities, for use in subsequent modeling.

[0118] Next, the electronic device constructs and trains the second fuzzy cognitive graph sub-model under various operating conditions. The electronic device solves for the weight matrix between various types of sample operating parameters for each set of sample operating parameters, obtaining the second fuzzy cognitive graph sub-model for different operating conditions.

[0119] The step of solving for the weight matrix between multiple types of sample operating parameters in each sample operating parameter set to obtain a second fuzzy cognitive graph sub-model for different working conditions includes: determining multiple types of sample operating parameters in the first sample operating parameter set as concept nodes of the second fuzzy cognitive graph sub-model; the first sample operating parameter set being any one of multiple sample operating parameter sets; solving for the weight matrix between multiple types of sample operating parameters in the first sample operating parameter set using the least squares method to obtain the weight matrix of the first sample operating parameter set as the weight matrix of the second fuzzy cognitive graph sub-model; and constructing a second fuzzy cognitive graph sub-model for a single working condition based on the sample operating parameters and the weight matrix of the first sample operating parameter set.

[0120] In one embodiment, each sample operating parameter set includes the temperature, pressure, physical properties, and gas impurity content of the gas transmission system outside the gas storage facility as concept nodes of the second fuzzy cognitive graph sub-model. For the model parameters of the second fuzzy cognitive graph sub-model under different operating conditions, this embodiment of the invention employs a fast learning method based on least squares to determine its model parameters.

[0121] This invention's embodiment uses a state space partitioning method based on prediction error. Its ultimate goal is to divide the entire sample running parameter space into several subspaces based on the prediction error of the constructed fuzzy cognitive graph model. Then, based on the data in each subspace, it learns the corresponding weight matrix of the fuzzy cognitive graph model, so that the prediction error J of each concept node in the fuzzy cognitive graph model in each subspace is minimized. s Minimum, that is:

[0122]

[0123] Here, s (s = 1, 2, ..., n) is the s-th partitioned subspace. It is the state value of the i-th concept node (i.e., the i-th variable of the system) at time t in the second fuzzy cognitive graph sub-model. N is the true state value of the i-th sample running parameter of the system (i.e., the i-th concept node of the second fuzzy cognitive graph sub-model) in the s-th subspace at time t. s It represents the size of the data in the s-th partitioned subspace. K-1 indicates time step K-1.

[0124] A fast learning algorithm for the weight matrix of a fuzzy cognitive graph model based on least squares for electronic devices: For the learning of the second fuzzy cognitive graph sub-model in each subspace, this embodiment of the invention uses the least squares method. In the constructed second fuzzy cognitive graph sub-model, the iterative calculation formula for each concept node is as follows:

[0125]

[0126] Here, c is the number of concept nodes in the constructed second fuzzy cognitive graph sub-model, and ω is... 0j It is related to concept node C j The relevant coordination coefficient. In formula (5) Treat it as the independent variable. Treating it as the dependent variable, omitting the subspace notation s, and... Let it be x it , Let it be y it , Let it be y i(t+1) ω 0j Let ω0 be the denoted ω0. Thus, formula (5) can be rewritten as:

[0127]

[0128] As can be seen from formula (6), it is essentially a linear equation. For a second fuzzy cognitive graph sub-model containing c nodes, c linear equations of the form formula (6) can be formed in the above manner, which can be solved using the least squares method. Since the least squares method only involves simple matrix operations when solving the system of equations, it is much faster than other existing methods for learning the weight matrix of fuzzy cognitive graphs, especially for learning the weight matrix of large-scale fuzzy cognitive graphs. Thus, this embodiment of the invention realizes the construction of a second fuzzy cognitive graph sub-model under a single working condition based on the sample running parameters and the weight matrix of the first sample running parameter set.

[0129] Finally, the electronic device fuses the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different operating conditions to obtain the fuzzy cognitive graph model. In one embodiment, fusing the first fuzzy cognitive graph model and the second fuzzy cognitive graph model under different operating conditions to obtain the fuzzy cognitive graph model includes: fusing the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrix of the second fuzzy cognitive graph model under different operating conditions based on the fuzzy TS model to obtain the fuzzy cognitive graph model.

[0130] The gas storage facility's external gas transmission system is a complete dynamic system. Comprehensive modeling and analysis of it are conducted after obtaining fuzzy cognitive graph sub-models of the external gas transmission system under different operating conditions. By fusing the fuzzy cognitive graph sub-models obtained under various operating conditions, a fuzzy cognitive graph model of the external gas transmission system is obtained. Then, leveraging the dynamic reasoning mechanism of the graph model, a dynamic analysis of the gas storage facility's external gas transmission system and its gas quality is performed.

[0131] In this embodiment of the invention, firstly, from the perspective of the entire gas storage facility's external gas transmission system, the aforementioned associated concept nodes are described using fuzzy semantics. Then, fuzzy semantic logical relationships are applied to form a corresponding hierarchical structure, which is subsequently measured to obtain the inclusion relationships between related concepts in the entire gas storage facility's external gas transmission system at different granularity levels. At a certain granularity level, based on this obtained inclusion relationship, the various fuzzy cognitive graph sub-models are merged to obtain a global fuzzy cognitive graph model of the entire gas storage facility's external gas transmission system at that granularity level.

[0132] In one embodiment, the fusion of the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model to obtain the fuzzy cognitive graph model includes:

[0133] For the two fuzzy cognitive graph sub-models FCM corresponding to operating conditions A and B of gas transmission from the gas storage facility. A and FCM B FCMA The concept node set is FCM B The concept node set is At a certain granularity level, if When the intersection of the concept node sets of fuzzy cognitive sub-model A and fuzzy cognitive sub-model B is an empty set, the weight matrix of the fuzzy cognitive graph model is expressed by the following formula:

[0134]

[0135] Among them, W global W represents the weight matrix of the fuzzy cognitive graph model. A ij W represents the weight matrix of the fuzzy cognitive sub-model A. B ij Let A represent the weight matrix of fuzzy cognitive sub-model B; fuzzy cognitive sub-model A represents either the first fuzzy cognitive graph sub-model or the second fuzzy cognitive graph sub-model, and fuzzy cognitive sub-model B represents either the first fuzzy cognitive graph model or the second fuzzy cognitive graph model.

[0136] In another implementation, the fuzzy cognitive graph model is obtained by fusing the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model, including:

[0137] if When the concept node sets of the first fuzzy cognitive graph sub-model and the concept node sets of the second fuzzy cognitive graph sub-model intersect, or when the concept node sets of the second fuzzy cognitive graph sub-models under different working conditions intersect, the weight matrix of the fuzzy cognitive graph model is represented by the following formula:

[0138]

[0139] Among them, W global This represents the weight matrix of the fuzzy cognitive graph model, where m is the number of fuzzy cognitive graph sub-models in the fuzzy cognitive graph model, and W... k ij This represents the concept node C in the k-th (k = 1, 2, ..., m) first or second fuzzy cognitive graph sub-model. j Pointing to C i The connection weights.

[0140] Finally, by evaluating the accuracy of the global fuzzy cognitive graph models formed at different granularity levels, the most suitable global fuzzy cognitive graph corresponding to the granularity level is selected, thereby completing the fusion of the corresponding fuzzy cognitive graph sub-models. The global fuzzy cognitive graph model of the gas storage facility's external gas quality formed above can completely depict the overall dynamic characteristics of the gas storage facility, external gas supply, and gas quality evolution. In this embodiment of the invention, the model is applied to actual production. By accessing real-time data through a DCS (Distributed Control System), the system's temperature, pressure, flow rate, liquid level, and gas moisture content, as well as key operating parameters, can be analyzed. Using formula (6), the predicted values ​​of each operating parameter of the gas storage facility's external gas supply system and the quality of the external gas can be obtained.

[0141] Please refer to Figure 4 This invention starts from the characteristics and laws of the operation of the gas storage facility's external gas transmission system. It analyzes the historical operating conditions of the system, using mechanisms and data as a foundation. Through big data analysis, it uncovers the implicit dynamic characteristics and operating patterns, integrating and unifying these with the mechanisms, operating rules, and patterns. This leads to the establishment of fuzzy cognitive graph sub-models for external gas transmission under various operating conditions, learning the connection weights between corresponding nodes (concepts). Subsequently, these fuzzy cognitive graph sub-models under various operating conditions are effectively merged to form a comprehensive operating condition transition model. This paper proposes a fuzzy cognitive graph model for the gas quality of gas exported from gas storage facilities. This model accurately simulates the dynamic characteristics of the gas quality operation of the gas export system and helps decision-makers observe and analyze the gas quality operation (evolution law) under different operating conditions from a global (local) perspective. Furthermore, by leveraging the good reasoning ability and interpretability of the fuzzy cognitive graph, reasonable decisions can be made, and a reasonable and interpretable basis can be provided for the global optimization of the gas quality of gas exported from gas storage facilities.

[0142] By employing the gas quality modeling method for external gas transmission from gas storage facilities according to embodiments of the present invention, the stability of the transmitted gas quality can be significantly improved, reducing the negative impact of gas quality fluctuations. Embodiments of the present invention can accurately predict the changing trend of gas quality transmitted from gas storage facilities. Furthermore, embodiments of the present invention are applicable to different types of external gas transmission systems from gas storage facilities, and the provided modeling method is highly operable and easy to promote and apply.

[0143] Compared with traditional gas quality modeling methods for external gas transmission from gas storage facilities, the embodiments of this invention have the following differences: 1) A gas quality modeling method for external gas transmission from gas storage facilities based on fuzzy cognitive graph reasoning is proposed, which improves the accuracy of gas quality prediction for external gas transmission from gas storage facilities; 2) By utilizing the interpretability of fuzzy cognitive graph reasoning, the influence of various operating parameters of the external gas transmission system of gas storage facilities on gas quality can be revealed.

[0144] This invention, for the first time, applies fuzzy cognitive graphs and reasoning methods to gas quality modeling for external gas transmission from gas storage facilities. The resulting gas quality model based on fuzzy cognitive graph reasoning not only intuitively reveals the operational status of gas quality under different operating conditions but also achieves high-precision prediction of gas quality through the strong reasoning capabilities of fuzzy cognitive graphs. The gas quality modeling method based on fuzzy cognitive graph reasoning proposed in this invention is essentially a data-driven modeling method with good generalization ability. It only requires relevant historical data and does not require prior knowledge such as physical mechanisms to achieve high-precision prediction of gas quality for external gas transmission from gas storage facilities.

[0145] Device Examples

[0146] Please refer to Figure 5 On the other hand, embodiments of the present invention also provide a gas quality prediction device for external gas transmission from a gas storage facility, comprising:

[0147] The acquisition module 501 is used to acquire the operating parameters to be measured for the external gas transmission of the gas storage facility;

[0148] The prediction module 502 is used to input the operating parameters to be measured into the gas quality prediction model to obtain the gas quality prediction result of the gas storage external gas output by the gas quality prediction model; the gas quality prediction model is constructed based on the fuzzy cognitive graph model, which is obtained by fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model for different operating conditions; the first fuzzy cognitive graph model is constructed based on expert experience; the second fuzzy cognitive graph model for each operating condition is trained based on the operating parameters of multiple types of samples under a single operating condition.

[0149] Optionally, the fuzzy cognitive graph model is obtained based on the following steps:

[0150] Obtain various types of sample operating parameters for external gas transmission from the gas storage facility; the various types of sample operating parameters include a first type of sample operating parameters that can be used to obtain the correlation between sample operating parameters based on expert experience, and a second type of sample operating parameters that are the remaining sample operating parameters excluding the first type of sample operating parameters.

[0151] Preprocess the running parameters of all types of samples;

[0152] Based on expert experience, a fuzzy cognitive graph model is constructed for the operating parameters of the first type of sample to obtain the first fuzzy cognitive graph sub-model.

[0153] Clustering is performed on the second type of sample operating parameters to obtain multiple sample operating parameter sets under different working conditions; each sample operating parameter set corresponds to a working condition and includes multiple types of second type sample operating parameters.

[0154] For each set of sample operating parameters, the weight matrix between various types of sample operating parameters is solved to obtain the second fuzzy cognitive graph sub-model for different working conditions;

[0155] The first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different working conditions are fused to obtain the fuzzy cognitive graph model.

[0156] Optionally, the step of constructing a fuzzy cognitive graph model based on expert experience for the operating parameters of the first type of samples to obtain a first fuzzy cognitive graph sub-model includes:

[0157] Multiple sample operation parameters in the first type of sample operation parameters are determined as concept nodes of the first fuzzy cognitive graph sub-model;

[0158] The correlation between the multiple sample operating parameters is obtained based on expert experience, and the correlation between the multiple sample operating parameters is determined as the weight matrix of the first fuzzy cognitive graph sub-model.

[0159] The first fuzzy cognitive graph sub-model is constructed based on the multiple sample operating parameters and the correlation between the multiple sample operating parameters.

[0160] Optionally, the step of solving for the weight matrix between various types of sample operating parameters for each set of sample operating parameters to obtain the second fuzzy cognitive graph sub-model for different operating conditions includes:

[0161] Multiple types of sample operation parameters from the first sample operation parameter set are identified as concept nodes in the second fuzzy cognitive graph sub-model; the first sample operation parameter set can be any one of multiple sample operation parameter sets;

[0162] The weight matrix of the first sample running parameter set is obtained by solving the weight matrix between various types of sample running parameters based on the least squares method, and the weight matrix of the first sample running parameter set is used as the weight matrix of the second fuzzy cognitive graph sub-model.

[0163] Based on the sample operation parameters and weight matrix of the first sample operation parameter set, a second fuzzy cognitive graph sub-model under a single working condition is constructed.

[0164] Optionally, fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-models for different working conditions to obtain the fuzzy cognitive graph model includes:

[0165] The weight matrix of the first fuzzy cognitive graph sub-model and the weight matrix of the second fuzzy cognitive graph sub-model under different working conditions are fused based on the fuzzy TS model to obtain the fuzzy cognitive graph model.

[0166] Optionally, the fusion of the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model to obtain the fuzzy cognitive graph model includes:

[0167] When the intersection of the concept node sets of fuzzy cognitive sub-model A and fuzzy cognitive sub-model B is an empty set, the weight matrix of the fuzzy cognitive graph model is represented by the following formula:

[0168]

[0169] Among them, W global W represents the weight matrix of the fuzzy cognitive graph model. A ij W represents the weight matrix of the fuzzy cognitive sub-model A. B ij Let A represent the weight matrix of fuzzy cognitive sub-model B; fuzzy cognitive sub-model A represents either the first fuzzy cognitive graph sub-model or the second fuzzy cognitive graph sub-model, and fuzzy cognitive sub-model B represents either the first fuzzy cognitive graph model or the second fuzzy cognitive graph model.

[0170] Optionally, the fusion of the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model to obtain the fuzzy cognitive graph model includes:

[0171] When the concept node sets of the first fuzzy cognitive graph sub-model and the concept node sets of the second fuzzy cognitive graph sub-model intersect, or when the concept node sets of the second fuzzy cognitive graph sub-models under different working conditions intersect, the weight matrix of the fuzzy cognitive graph model is represented by the following formula:

[0172]

[0173] Among them, W global This represents the weight matrix of the fuzzy cognitive graph model, where m is the number of fuzzy cognitive graph sub-models in the fuzzy cognitive graph model, and W... k ij This represents the concept node C in the k-th (k = 1, 2, ..., m) first or second fuzzy cognitive graph sub-model. j Pointing to C i The connection weights.

[0174] Optionally, the preprocessing of the sample running parameters for all types includes:

[0175] Perform data standardization and / or data normalization on the operating parameters of all types of samples.

[0176] Optionally, the various types of sample operating parameters include at least two of the following: temperature, pressure, flow rate, liquid level, gas impurity content, and gas composition of the gas storage external gas transmission system.

[0177] The gas quality prediction device for external gas transmission from the gas storage facility includes a processor and a memory. The acquisition module 501 and prediction module 502 are stored in the memory as program units, and the processor executes the program units stored in the memory to achieve the corresponding functions.

[0178] A processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured.

[0179] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0180] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a gas quality prediction method for gas exported from a gas storage facility. The method includes: acquiring the operating parameters to be measured for gas exported from the gas storage facility; inputting the operating parameters to be measured into a gas quality prediction model to obtain the gas quality prediction result of gas exported from the gas storage facility output by the gas quality prediction model; the gas quality prediction model is constructed based on a fuzzy cognitive graph model, which is obtained by fusing a first fuzzy cognitive graph sub-model and a second fuzzy cognitive graph sub-model for different operating conditions; the first fuzzy cognitive graph model is constructed based on expert experience; the second fuzzy cognitive graph sub-model for each operating condition is trained based on multiple types of sample operating parameters under a single operating condition.

[0181] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0182] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a machine-readable storage medium. When the computer program is executed by a processor, the computer can execute a gas quality prediction method for gas exported from a gas storage facility. The method includes: acquiring the operating parameters to be measured for gas exported from the gas storage facility; inputting the operating parameters to be measured into a gas quality prediction model to obtain the gas quality prediction result of gas exported from the gas storage facility output by the gas quality prediction model; the gas quality prediction model is constructed based on a fuzzy cognitive graph model, which is obtained by fusing a first fuzzy cognitive graph sub-model and a second fuzzy cognitive graph sub-model for different operating conditions; the first fuzzy cognitive graph model is constructed based on expert experience; and the second fuzzy cognitive graph model for each operating condition is trained based on multiple types of sample operating parameters under a single operating condition.

[0183] In another aspect, the present invention also provides a machine-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements a method for predicting the gas quality of gas exported from a gas storage facility. The method includes: acquiring the operating parameters to be measured for the gas exported from the gas storage facility; inputting the operating parameters to be measured into a gas quality prediction model to obtain the gas quality prediction result of the gas exported from the gas storage facility output by the gas quality prediction model; the gas quality prediction model is constructed based on a fuzzy cognitive graph model, which is obtained by fusing a first fuzzy cognitive graph sub-model and a second fuzzy cognitive graph sub-model for different operating conditions; the first fuzzy cognitive graph model is constructed based on expert experience; and the second fuzzy cognitive graph model for each operating condition is trained based on multiple types of sample operating parameters under a single operating condition.

[0184] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0185] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0186] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the gas quality of gas transported from a gas storage facility, characterized in that, include: Obtain the operating parameters to be measured for gas transmission from the gas storage facility; The operating parameters to be measured are input into the gas quality prediction model to obtain the gas quality prediction results of the gas storage external gas output by the gas quality prediction model; the gas quality prediction model is constructed based on the fuzzy cognitive graph model, which is obtained by fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different operating conditions; The first fuzzy cognitive graph sub-model is constructed based on expert experience; the second fuzzy cognitive graph model for each working condition is trained based on the operating parameters of multiple types of samples under a single working condition.

2. The gas quality prediction method for external gas transmission from a gas storage facility according to claim 1, characterized in that, The fuzzy cognitive graph model is obtained based on the following steps: Obtain various types of sample operating parameters for external gas transmission from the gas storage facility; the various types of sample operating parameters include a first type of sample operating parameters that can be used to obtain the correlation between sample operating parameters based on expert experience, and a second type of sample operating parameters that are the remaining sample operating parameters excluding the first type of sample operating parameters. Preprocess the running parameters of all types of samples; Based on expert experience, a fuzzy cognitive graph model is constructed for the operating parameters of the first type of sample to obtain the first fuzzy cognitive graph sub-model. Clustering is performed on the second type of sample operating parameters to obtain multiple sample operating parameter sets under different working conditions; each sample operating parameter set corresponds to a working condition and includes multiple types of second type sample operating parameters. For each set of sample operating parameters, the weight matrix between various types of sample operating parameters is solved to obtain the second fuzzy cognitive graph sub-model for different working conditions; The first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different working conditions are fused to obtain the fuzzy cognitive graph model.

3. The gas quality prediction method for external gas transmission from a gas storage facility according to claim 2, characterized in that, The construction of a fuzzy cognitive graph model based on expert experience of the first type of sample operating parameters yields a first fuzzy cognitive graph sub-model, including: Multiple sample operation parameters in the first type of sample operation parameters are determined as concept nodes of the first fuzzy cognitive graph sub-model; The correlation between the multiple sample operating parameters is obtained based on expert experience, and the correlation between the multiple sample operating parameters is determined as the weight matrix of the first fuzzy cognitive graph sub-model. The first fuzzy cognitive graph sub-model is constructed based on the multiple sample operating parameters and the correlation between the multiple sample operating parameters.

4. The gas quality prediction method for external gas transmission from a gas storage facility according to claim 2, characterized in that, The step of solving the weight matrix between various types of sample operating parameters for each set of sample operating parameters to obtain the second fuzzy cognitive graph sub-model for different operating conditions includes: Multiple types of sample operation parameters from the first sample operation parameter set are identified as concept nodes in the second fuzzy cognitive graph sub-model; the first sample operation parameter set can be any one of multiple sample operation parameter sets; The weight matrix of the first sample running parameter set is obtained by solving the weight matrix between various types of sample running parameters based on the least squares method, and the weight matrix of the first sample running parameter set is used as the weight matrix of the second fuzzy cognitive graph sub-model. Based on the sample operation parameters and weight matrix of the first sample operation parameter set, a second fuzzy cognitive graph sub-model under a single working condition is constructed.

5. The gas quality prediction method for external gas transmission from a gas storage facility according to claim 2, characterized in that, The process of fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph models under different working conditions to obtain the fuzzy cognitive graph model includes: The weight matrix of the first fuzzy cognitive graph sub-model and the weight matrix of the second fuzzy cognitive graph sub-model under different working conditions are fused based on the fuzzy TS model to obtain the fuzzy cognitive graph model.

6. The gas quality prediction method for external gas transmission from a gas storage facility according to claim 5, characterized in that, The fuzzy cognitive graph model is obtained by fusing the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model, including: When the intersection of the concept node sets of fuzzy cognitive sub-model A and fuzzy cognitive sub-model B is an empty set, the weight matrix of the fuzzy cognitive graph model is represented by the following formula: Among them, W global W represents the weight matrix of the fuzzy cognitive graph model. A ij W represents the weight matrix of the fuzzy cognitive sub-model A. B ij Let A represent the weight matrix of fuzzy cognitive sub-model B; fuzzy cognitive sub-model A represents either the first fuzzy cognitive graph sub-model or the second fuzzy cognitive graph sub-model, and fuzzy cognitive sub-model B represents either the first fuzzy cognitive graph model or the second fuzzy cognitive graph model.

7. The gas quality prediction method for external gas transmission from a gas storage facility according to claim 5, characterized in that, The fuzzy cognitive graph model is obtained by fusing the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model, including: When the concept node sets of the first fuzzy cognitive graph sub-model and the concept node sets of the second fuzzy cognitive graph sub-model intersect, or when the concept node sets of the second fuzzy cognitive graph sub-models under different working conditions intersect, the weight matrix of the fuzzy cognitive graph model is represented by the following formula: Among them, W global This represents the weight matrix of the fuzzy cognitive graph model, where m is the number of fuzzy cognitive graph sub-models in the fuzzy cognitive graph model, and W... k ij This represents the concept node C in the k-th (k = 1, 2, ..., m) first or second fuzzy cognitive graph sub-model. j Pointing to C i The connection weights.

8. The gas quality prediction method for external gas transmission from a gas storage facility according to claim 2, characterized in that, The preprocessing of the sample operating parameters for all types includes: Perform data standardization and / or data normalization on the operating parameters of all types of samples.

9. The gas quality prediction method for external gas transmission from a gas storage facility according to claim 1, characterized in that, The various types of sample operating parameters include at least two of the following: temperature, pressure, flow rate, liquid level, gas impurity content, and gas composition of the gas storage external gas transmission system.

10. A gas quality prediction device for external gas transmission from a gas storage facility, characterized in that, include: The acquisition module is used to acquire the operating parameters to be measured for the external gas transmission of the gas storage facility; The prediction module is used to input the operating parameters to be measured into the gas quality prediction model to obtain the gas quality prediction result of the gas storage external gas output by the gas quality prediction model; the gas quality prediction model is constructed based on the fuzzy cognitive graph model, which is obtained by fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different operating conditions; The first fuzzy cognitive graph sub-model is constructed based on expert experience; the second fuzzy cognitive graph model for each working condition is trained based on the operating parameters of multiple types of samples under a single working condition.

11. The gas quality prediction device for external gas transmission from a gas storage facility according to claim 10, characterized in that, The fuzzy cognitive graph model is obtained based on the following steps: Obtain various types of sample operating parameters for external gas transmission from the gas storage facility; the various types of sample operating parameters include a first type of sample operating parameters that can be used to obtain the correlation between sample operating parameters based on expert experience, and a second type of sample operating parameters that are the remaining sample operating parameters excluding the first type of sample operating parameters. Preprocess the running parameters of all types of samples; Based on expert experience, a fuzzy cognitive graph model is constructed for the operating parameters of the first type of sample to obtain the first fuzzy cognitive graph sub-model. Clustering is performed on the second type of sample operating parameters to obtain multiple sample operating parameter sets under different working conditions; each sample operating parameter set corresponds to a working condition and includes multiple types of second type sample operating parameters. For each set of sample operating parameters, the weight matrix between various types of sample operating parameters is solved to obtain the second fuzzy cognitive graph sub-model for different working conditions; The first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph sub-model under different working conditions are fused to obtain the fuzzy cognitive graph model.

12. The gas quality prediction device for external gas transmission from a gas storage facility according to claim 11, characterized in that, The construction of a fuzzy cognitive graph model based on expert experience of the first type of sample operating parameters yields a first fuzzy cognitive graph sub-model, including: Multiple sample operation parameters in the first type of sample operation parameters are determined as concept nodes of the first fuzzy cognitive graph sub-model; The correlation between the multiple sample operating parameters is obtained based on expert experience, and the correlation between the multiple sample operating parameters is determined as the weight matrix of the first fuzzy cognitive graph sub-model. The first fuzzy cognitive graph sub-model is constructed based on the multiple sample operating parameters and the correlation between the multiple sample operating parameters.

13. The gas quality prediction device for external gas transmission from a gas storage facility according to claim 11, characterized in that, The step of solving the weight matrix between various types of sample operating parameters for each set of sample operating parameters to obtain the second fuzzy cognitive graph sub-model for different operating conditions includes: Multiple types of sample operation parameters from the first sample operation parameter set are identified as concept nodes in the second fuzzy cognitive graph sub-model; the first sample operation parameter set can be any one of multiple sample operation parameter sets; The weight matrix of the first sample running parameter set is obtained by solving the weight matrix between various types of sample running parameters based on the least squares method, and the weight matrix of the first sample running parameter set is used as the weight matrix of the second fuzzy cognitive graph sub-model. Based on the sample operation parameters and weight matrix of the first sample operation parameter set, a second fuzzy cognitive graph sub-model under a single working condition is constructed.

14. The gas quality prediction device for external gas transmission from a gas storage facility according to claim 11, characterized in that, The process of fusing the first fuzzy cognitive graph sub-model and the second fuzzy cognitive graph models under different working conditions to obtain the fuzzy cognitive graph model includes: The weight matrix of the first fuzzy cognitive graph sub-model and the weight matrix of the second fuzzy cognitive graph sub-model under different working conditions are fused based on the fuzzy TS model to obtain the fuzzy cognitive graph model.

15. The gas quality prediction device for external gas transmission from a gas storage facility according to claim 14, characterized in that, The fuzzy cognitive graph model is obtained by fusing the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model, including: When the intersection of the concept node sets of fuzzy cognitive sub-model A and fuzzy cognitive sub-model B is an empty set, the weight matrix of the fuzzy cognitive graph model is represented by the following formula: Among them, W global W represents the weight matrix of the fuzzy cognitive graph model. A ij W represents the weight matrix of the fuzzy cognitive sub-model A. B ij Let A represent the weight matrix of fuzzy cognitive sub-model B; fuzzy cognitive sub-model A represents either the first fuzzy cognitive graph sub-model or the second fuzzy cognitive graph sub-model, and fuzzy cognitive sub-model B represents either the first fuzzy cognitive graph model or the second fuzzy cognitive graph model.

16. The gas quality prediction device for external gas transmission from a gas storage facility according to claim 14, characterized in that, The fuzzy cognitive graph model is obtained by fusing the weight matrix of the first fuzzy cognitive graph sub-model and the weight matrices of the second fuzzy cognitive graph sub-model under different working conditions based on the fuzzy TS model, including: When the concept node sets of the first fuzzy cognitive graph sub-model and the concept node sets of the second fuzzy cognitive graph sub-model intersect, or when the concept node sets of the second fuzzy cognitive graph sub-models under different working conditions intersect, the weight matrix of the fuzzy cognitive graph model is represented by the following formula: Among them, W global This represents the weight matrix of the fuzzy cognitive graph model, where m is the number of fuzzy cognitive graph sub-models in the fuzzy cognitive graph model, and W... k ij This represents the concept node C in the k-th (k = 1, 2, ..., m) first or second fuzzy cognitive graph sub-model. j Pointing to C i The connection weights.

17. The gas quality prediction device for external gas transmission from a gas storage facility according to claim 11, characterized in that, The preprocessing of the sample operating parameters for all types includes: Perform data standardization and / or data normalization on the operating parameters of all types of samples.

18. The gas quality prediction device for external gas transmission from a gas storage facility according to claim 10, characterized in that, The various types of sample operating parameters include at least two of the following: temperature, pressure, flow rate, liquid level, gas impurity content, and gas composition of the gas storage external gas transmission system.

19. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the gas quality prediction method for external gas transmission from the gas storage facility as described in any one of claims 1 to 9.

20. A machine-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the gas quality prediction method for external gas transmission from the gas storage facility as described in any one of claims 1 to 9.

21. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the gas quality prediction method for external gas transmission from the gas storage facility as described in any one of claims 1 to 9.

Images