A Fast Decision Generation Method, System, Device and Medium for an Electrical-Gas Integrated Energy System Based on Neural Networks

Abstract

The present invention provides a method, system, device and medium for quickly generating decisions for an electric-gas integrated energy system based on a neural network, which relates to the technical field of operation control of electric-gas integrated energy systems. The method includes collecting natural gas system data, gas turbine unit safety domain data and power system data at each time period, constructing a multi-module neural network training set and a power system data set after normalization processing; constructing and training a feedback-compensation multi-module neural network to obtain a safety domain prediction model; constructing a penalty function neural network training set; constructing and training a penalty function neural network to obtain an optimal decision generation model; and jointly operating the two models to generate an optimal scheduling decision for the power system. The present invention can optimize economic scheduling while ensuring the safe operation of the pipeline network, generate an optimal scheduling decision for the power system, improve the decision generation efficiency, and has good safety, stability and adaptability.

Description

Background Art

[0002] With the rapid development of renewable energy, the proportion of renewable energy such as wind power and photovoltaic power in the power system is increasing continuously. However, due to the strong uncertainty and volatility of these renewable energies, the power system is facing increasing challenges.

[0003] In order to effectively cope with the power mismatch caused by the fluctuations of renewable energy, gas turbines have become an important regulating resource in the power system due to their flexible regulating ability and low pollution emissions. With the increase in the installed capacity of gas turbines, the coupling relationship between the power system and the natural gas system has become increasingly close. During the process of participating in the operation and dispatching of the power system, gas turbines can effectively balance the uncertainty of renewable energy, but their dynamic regulation will also affect the pressure and flow of the natural gas pipeline network, bringing certain safety hazards. Therefore, the traditional independent dispatching mode can no longer give full play to the complementary advantages of the power and natural gas systems, and there is an urgent need for coordinated optimization of the power-gas system. Through the research on the coordinated decision-making of the integrated power-gas energy system, the consumption capacity of renewable energy can be improved, and at the same time, the safe, stable and economic operation of the power system can be ensured, which is of great significance for promoting the transformation of the energy system towards clean and low-carbon.

[0004] However, since most gas power plants do not store fuel on site, the natural gas used for power generation is extracted in real time from the natural gas pipeline, and the uncertain gas extraction of gas turbines may cause fluctuations in the node air pressure of the natural gas pipeline network, and even violate the safety constraints of the natural gas pipeline network. In the case of a lack of natural gas supply, the rapid ramping of gas turbines may cause the pressure of the inlet pipeline to be lower than the minimum threshold, resulting in the shutdown of gas turbines and causing safety accidents. Therefore, it is very necessary to study the safe operating region of the pipeline network in the integrated power-gas energy system (that is, the output range of gas turbines that does not violate the natural gas safety constraints, also known as the safety region of gas turbines).

[0005] In the prior art, many studies considering the coordinated decision-making of the integrated power-gas energy system use model-driven methods to model and solve to obtain the optimal strategy, which requires establishing accurate mathematical models for the physical principles, chemical processes, electrical characteristics, etc. of the integrated power-gas energy system to describe the behavior and characteristics of the system. For example, the energy flow characteristics of the system are modeled through power flow equations and natural gas partial differential equations. However, these methods need to use techniques such as finite differences to discretize the partial differential equations, which greatly increases the complexity of system solving. In addition, with the increase in the complexity of modern multi-energy systems, it is difficult to accurately obtain the topological structure of the system and the relevant parameters of the network. Therefore, when facing complex and changing actual operating conditions, the accuracy and adaptability of the existing model-driven methods may be limited. Summary of the Invention

[0006] Aiming at the technical problems in the prior art that the optimal strategy for collaborative decision-making of an integrated electricity-gas energy system obtained by using a model-driven method for modeling and solving is complex, and it is difficult to obtain the topological structure and network-related parameters of a complex system, the invention provides a fast decision-making generation method, system, device and medium for an integrated electricity-gas energy system based on a neural network, which can optimize economic dispatch while ensuring the safe operation of the pipeline network, generate the optimal dispatch decision of the power system, improve the decision-making generation efficiency, and has good safety, stability and adaptability.

[0007] In the first aspect, the invention provides a fast decision-making generation method for an integrated electricity-gas energy system based on a neural network. The steps include:

[0008] S1. Collect the natural gas system data, gas turbine unit safety domain data and power system data of the same integrated electricity-gas energy system at each time period, perform normalization processing, and then construct a multi-module neural network training set and a power system data set;

[0009] S2. Construct a feedback-compensation multi-module neural network and train it using the multi-module neural network training set to obtain a safety domain prediction model. The input of the safety domain prediction model is the natural gas system data, and the output is the predicted value of the gas turbine unit safety domain;

[0010] S3. Input the natural gas system data into the safety domain prediction model, calculate the actual output boundary value of the gas turbine unit according to the predicted value of the gas turbine unit safety domain output and the inherent output boundary value of the gas turbine unit, splice the actual output boundary value with the power system data, and construct a penalty function neural network training set;

[0011] Among them, the inherent output boundary value of the gas turbine unit is the upper and lower boundary values of the output set during the design and manufacture of the gas turbine unit;

[0012] S4. Construct a penalty function neural network and train it using the penalty function neural network training set to obtain an optimal decision-making generation model;

[0013] The input of the optimal decision-making generation model of the power system is the actual output boundary value of the gas turbine unit and the power system data, and the output is the dispatch strategy of the output of the gas turbine unit and the coal-fired unit;

[0014] S5. Input the natural gas system data to be decision-making into the safety domain prediction model, use its output to calculate the actual processing boundary value of the gas group, and then jointly input it with the power system data into the optimal decision-making model to generate the optimal dispatch decision of the power system.

[0015] It should be further noted that in step S1, a multi-module neural network training set is constructed according to the normalized natural gas system data and gas turbine unit safety domain data, and a power system data set is constructed according to the normalized power system data;

[0016] Among them, the natural gas system data includes the gas source pressure, compressor boost, and natural gas load in the natural gas system;

[0017] The safe operating range data of the gas turbine unit includes the upper output limit and the lower output limit for the safe operation of the gas turbine unit;

[0018] The power system data includes the power load and wind power output data in the power system.

[0019] Furthermore, it should be noted that in step S1, for any original data of the natural gas system data or the power system data , the formula for its normalization processing is:

[0020]

[0021] In the formula, represents the value after normalization processing;

[0022] and respectively represent the maximum and minimum values of the data of the corresponding type in the acquisition sample.

[0023] Furthermore, it should be noted that in step S2, the feedback-compensation multi-module neural network includes a prediction module, a feedback module, and a compensation module, where:

[0024] The prediction module includes three single-layer perceptrons with activation functions and a multi-layer perceptron. After inputting the gas source pressure, natural gas load, and compressor boost volume in the natural gas system data into the three single-layer perceptrons respectively, the single-layer perceptrons extract features, and then the multi-layer perceptron obtains the preliminary estimated value of the safe operating range of the gas turbine unit. The expression of the prediction module is:

[0025]

[0026] In the formula, is the set of upper output limits in the preliminary estimated value of the safe operating range of the gas turbine unit;

[0027] is the set of lower output limits in the preliminary estimated value of the safe operating range of the gas turbine unit;

[0028] is the function for calculating the preliminary estimated value of the safe operating range of the gas turbine unit;

[0029] represents the model parameters of the four perceptrons in the prediction module, including the weights and biases of each layer of neurons;

[0030] is the gas source pressure;

[0031] is the natural gas load;

[0032] is the compressor boost volume;

[0033] The feedback module includes a multi-layer perceptron. The feedback module receives the preliminary estimated value of the safe operating range of the gas turbine unit obtained by the prediction module and the corresponding natural gas system data, extracts data features through the multi-layer perceptron, and generates the correction values for the upper and lower limits of the safe output of the gas turbine unit. The expression of the feedback module is:

[0034]

[0035] In the formula, is the correction value for the upper limit of the safe output of the gas turbine unit;

[0036] is the correction value for the lower limit of the safe output of the gas turbine unit;

[0037] is the function for calculating the correction value of the safe output of the gas turbine unit;

[0038] represents the model parameters of the multi-layer perceptron in the feedback module, including the weights and biases of neurons in each layer;

[0039] The compensation module includes a multi-layer perceptron. The compensation module is used to capture the fluctuations in the safe operating range of the gas turbine unit caused by the overall change in the natural gas system data. It receives the total gas source pressure data, total natural gas load data, and total compressor boost data, and generates the compensation amounts for the upper and lower limits of the safe output of the gas turbine unit, which are used to compensate for the fluctuations in the safe operating range of the gas turbine unit caused by the change in the natural gas system input. The expression of the compensation module is:

[0040]

[0041] In the formula, is the compensation amount for the upper limit of the safe output of the gas turbine unit;

[0042] is the compensation amount for the upper limit of the safe output of the gas turbine unit;

[0043] is the function for calculating the compensation amount of the safe output of the gas turbine unit;

[0044] is the model parameters of the multi-layer perceptron in the compensation module, including the weights and biases of neurons in each layer;

[0045] is the total gas source pressure;

[0046] is the total natural gas load;

[0047] is the total compressor boost volume;

[0048] By combining the prediction module, feedback module, and compensation module, a feedback-compensation multi-module neural network is constructed to calculate the predicted value of the safety domain of the gas turbine unit. The expression for the predicted value of the safety domain of the gas turbine unit is:

[0049]

[0050] In the formula, is the function for calculating the predicted value of the safety domain of the gas turbine unit;

[0051] is the upper output limit in the predicted value of the safety domain of the gas turbine unit;

[0052] is the lower output limit in the predicted value of the safety domain of the gas turbine unit.

[0053] It should be further noted that by training the feedback-compensation multi-module neural network with a multi-module neural dataset , the training steps are as follows:

[0054] S201. Initialize the parameters of the feedback-compensation multi-module neural network using the Xavier method, including initializing the weights and biases of each layer of neurons;

[0055] S202. Construct the loss function of the feedback-compensation multi-module neural network using the mean square error MSE. The loss function of the feedback-compensation multi-module neural network is specifically:

[0056]

[0057]

[0058]

[0059]

[0060] In the formula, represents the model prediction error calculated by MSE, which constitutes the main component of the loss function; secondly;

[0061] represents the error between the output of the prediction module and the true value calculated by MSE, which aims to reduce the correction amount from the feedback module and the compensation amount from the compensation module, thereby improving the overall accuracy;

[0062] is the regularization term, and the Manhattan norm is used to limit the magnitude of the compensation amount;

[0063] represents the data batch size;

[0064] and represents the scaling factor;

[0065] and respectively represent the Manhattan norm and the Euclidean norm;

[0066] and represents the set of predicted values of the gas turbine unit safety domain;

[0067] and represents the set of preliminary estimated values of the gas turbine unit safety domain;

[0068] and expresses the true value corresponding to the output of the feedback-compensation multi-module neural network;

[0069] S203. Calculate the loss function according to the multi-module neural network training set and the predicted values of the gas turbine unit safety domain, calculate the gradient of the loss function with respect to the network parameters using the chain rule and automatic differentiation, then update the network parameters through the Adam optimizer, and then calculate the loss function between the prediction and the true value according to the new network parameters. Repeat the iterative update until the preset number of training rounds is reached to obtain the gas turbine unit safety domain prediction model.

[0070] Furthermore, it should be noted that in step S3, the calculation formula for the actual output boundary of the gas turbine unit is:

[0071]

[0072]

[0073] In the formula, is the upper output limit in the actual output boundary value of the gas turbine unit;

[0074] is the lower output limit in the actual output boundary value of the gas turbine unit;

[0075] is the upper output limit in the predicted value of the gas turbine unit safety domain;

[0076] is the lower output limit in the predicted value of the gas turbine unit safety domain;

[0077] is the upper output limit in the inherent output boundary value of the gas turbine unit;

[0078] It is the lower output limit among the inherent output boundary values of the gas turbine unit.

[0079] Furthermore, it should be noted that in step S1, the power system data set is expressed as ， where is the power load, is the output of the wind turbine generator;

[0080] In step S3, the penalty function neural network training set is expressed as .

[0081] Furthermore, it should be noted that step S4 is specifically as follows: Establish an economic dispatch model of the power system and a penalty function neural network. Based on the objectives and constraints defined by the economic dispatch model of the power system, use the penalty function neural network training set to train the penalty function neural network to approximate the optimal solution in a data-driven manner, and finally form a decision-making generation model.

[0082] Furthermore, it should be noted that the economic dispatch model of the power system includes the output constraint of the gas turbine unit, the output constraint of the coal-fired unit, and the source-load power balance constraint. The expression of the economic dispatch model of the power system is;

[0083]

[0084]

[0085]

[0086] In the formula, is the output of the gas turbine unit at time ;

[0087] is the output of the coal-fired unit at time ;

[0088] is the output of the wind turbine generator at time ;

[0089] is the magnitude of the power load at time ;

[0090] is the upper output limit among the actual output boundary values of the gas turbine unit at time ;

[0091] is the gas turbine unit at time The lower output limit among the actual output boundary values;

[0092] and are the upper and lower output boundaries of the coal-fired unit ;

[0093] , , and respectively represent the sets corresponding to the coal-fired unit, gas-fired unit, wind turbine unit, and power load;

[0094] The objective function is to minimize the power generation cost of the coal-fired unit, and the objective function is;

[0095]

[0096] In the formula, , , are the quadratic term, linear term coefficient, and constant term corresponding to the cost function of the coal-fired unit;

[0097] is the total scheduling period.

[0098] It should be further noted that in step S4, the expression of the penalty function neural network is:

[0099]

[0100] In the formula, is the output of the gas-fired unit;

[0101] is the output of the coal-fired unit;

[0102] is the function for calculating the scheduling decision of the power system;

[0103] represents the model parameters of the penalty function neural network, including the weights and biases of each layer of neurons;

[0104] The expression of the loss function of the penalty function neural network is:

[0105]

[0106]

[0107]

[0108]

[0109]

[0110] In the formula, is the weight factor of the loss term for the power balance constraint;

[0111] is the weight factor of the loss term for the output constraint of the gas turbine unit;

[0112] is the weight factor of the loss term for the output constraint of the coal-fired unit.

[0113] Furthermore, it should be noted that in step S4, the training process of the penalty function neural network is as follows:

[0114] Input the penalty function neural network training set into the penalty function neural network, then calculate the loss function value, update the network parameters through the Adam optimizer, and recalculate according to the new parameters. Repeat the iterative update until the preset number of training rounds is reached to obtain the optimal decision-making generation model for the power system.

[0115] In the second aspect, the present invention provides a fast decision-making generation system for an electric-gas integrated energy system based on a neural network, which is used to implement the above-mentioned fast decision-making generation method for an electric-gas integrated energy system based on a neural network, including:

[0116] A data collection and processing module, which is used to collect the natural gas system data, gas turbine unit safety domain data, and power system data of the same electric-gas integrated energy system at each time period, perform normalization processing, and then construct a multi-module neural network training set and a power system data set;

[0117] A feedback-compensation multi-module neural network construction and training module, which is used to construct a feedback-compensation multi-module neural network and train it using the multi-module neural network training set to obtain a safety domain prediction model;

[0118] A gas turbine unit actual output boundary value calculation module, which is used to calculate the actual output boundary value of the gas turbine unit;

[0119] A penalty function neural network training set construction module, which is used to splice the actual output boundary value of the gas turbine unit with the corresponding power system data to construct a penalty function neural network training set;

[0120] A penalty function neural network construction and training module, which is used to construct a penalty function neural network and train it using the penalty function neural network training set to obtain an optimal decision-making generation model;

[0121] A coordinated operation module, which is used to input the natural gas system data to be decided into the safety domain prediction model, use its output to calculate the actual processing boundary value of the gas turbine unit, and then jointly input it with the power system data into the optimal decision model to generate the optimal scheduling decision for the power system.

[0122] In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor is configured to implement the steps of the above-mentioned method for quickly generating a decision for an electric-gas integrated energy system based on a neural network when executing the computer program.

[0123] In a fourth aspect, the present invention provides a storage medium with a computer program stored thereon. The computer program, when executed by a processor, implements the steps of the above-mentioned method for quickly generating a decision for an electric-gas integrated energy system based on a neural network.

[0124] The beneficial effects of the present invention are as follows:

[0125] 1. After collecting natural gas system data, gas turbine unit safety domain data, and power system data, the present invention performs normalization processing and constructs a multi-module neural network data set and a power system data set. A feedback-compensation multi-module neural network is constructed and trained to obtain a gas turbine unit safety domain prediction model. According to the output of the gas turbine unit safety domain prediction model and the inherent output boundary value of the gas turbine unit, combined with the corresponding power system data, a penalty function neural network training set is constructed. The penalty function neural network is trained to obtain an optimal decision generation model for the power system. Finally, the gas turbine unit safety domain prediction model and the optimal decision generation model for the power system are operated in coordination to generate an optimal dispatching decision for the power system. It is possible to propose an optimal dispatching decision for the power system that fully considers economic factors on the premise of considering the actual output boundary value of the gas turbine unit and meeting the safe operation of the system, rationally allocate power resources, reduce the operation cost of the power system, and improve the economy of the electric-gas integrated energy system.

[0126] 2. In the process of generating the optimal dispatching decision for the power system, the present invention processes the economic dispatching problem of the power system by constructing a feedback-compensation multi-module neural network and a penalty function neural network. It is not necessary to model the entire system and does not rely on a complex system model, reducing the time cost of model establishment and solution. Among them, by constructing and training a feedback-compensation multi-module neural network, the operation safety domain of the gas turbine unit can be accurately and real-time predicted, providing guarantee for the safe operation of the pipeline network; by constructing and training a penalty function neural network, the optimal dispatching problem of the power system is integrated into network training. The trained penalty function neural network can provide an optimal dispatching strategy in real time according to the operation safety domain of the gas turbine unit provided by the feedback-compensation multi-module neural network; the coordinated operation of the gas turbine unit safety domain prediction model and the optimal decision generation model for the power system can effectively avoid the influence of inaccurate model parameters and overly complex models on the dispatching problem of the electric-gas integrated energy system, and at the same time greatly shorten the problem-solving time, improve the efficiency of decision generation, and meet the real-time requirements.

[0127] 3. The data sources of the present invention are natural gas system data, gas turbine unit safety domain data, and power system data, covering various aspects of information such as gas source pressure, compressor boosting, natural gas load, safe operation output boundary of gas turbine units, as well as power load and wind power output. Through the normalization processing and comprehensive utilization of these data, the mutual relationships and impacts between different subsystems in the electricity-gas integrated energy system are fully considered, which can more comprehensively ensure the safe operation of the pipeline network, improve the adaptability of the decision-making generation method, and maintain stability under complex and changeable actual operating conditions. Among them, a prediction model for the safety domain of gas turbine units is constructed and trained using a multi-module neural network training set, making the prediction of the safe operation of gas turbine units more accurate, avoiding potential safety problems caused by improper operation of gas turbine units from the source, and further ensuring the safe and stable operation of the entire system.

[0128] 4. The rapid decision-making generation method of the present invention is data-driven. By collecting, processing, and training a large amount of historical data, it can adapt to the electricity-gas integrated energy system under different working conditions and operating conditions. As the system operation data continues to accumulate, the model can be continuously optimized and updated, further improving the accuracy and adaptability of prediction and decision-making, and having strong versatility and scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0129] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0130] Figure 1 It is a flowchart of a rapid decision-making generation method for an electricity-gas integrated energy system based on a neural network in an embodiment of the present invention.

[0131] Figure 2 It is a schematic structural diagram of a feedback-compensation multi-module neural network in an embodiment of the present invention.

[0132] Figure 3 It is a schematic structural diagram of a penalty function neural network in an embodiment of the present invention.

[0133] Figure 4 It is a schematic block diagram of a rapid decision-making generation system for an electricity-gas integrated energy system based on a neural network in an embodiment of the present invention.

[0134] Figure 5 It is a schematic hardware structure diagram of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0135] In order to make the objectives, features, and advantages of the present invention more obvious and understandable, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the specific embodiments. Obviously, the embodiments described below are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0136] The rapid decision-making generation method for the electric-gas integrated energy system based on neural network involved in this application mainly aims at the operation control technology field of the electric-gas integrated energy system. The technical solution includes collecting natural gas system data, gas turbine unit safety domain data, and power system data at each time period, constructing a multi-module neural network training set and a power system data set after normalization processing; constructing and training a feedback-compensation multi-module neural network to obtain a gas turbine unit safety domain prediction model; constructing a penalty function neural network training set; constructing and training a penalty function neural network to obtain an optimal decision-making generation model for the power system; and operating the two models in coordination to generate an optimal scheduling decision for the power system. Compared with the prior art, in this invention, after collecting natural gas system data, gas turbine unit safety domain data, and power system data, they are normalized and a multi-module neural network data set and a power system data set are constructed. A feedback-compensation multi-module neural network is constructed and trained to obtain a gas turbine unit safety domain prediction model. According to the output of the gas turbine unit safety domain prediction model and the inherent output boundary value of the gas turbine unit, combined with the corresponding power system data, a penalty function neural network training set is constructed. The penalty function neural network is trained to obtain an optimal decision-making generation model for the power system. Finally, the gas turbine unit safety domain prediction model and the optimal decision-making generation model for the power system are operated in coordination to generate an optimal scheduling decision for the power system, which can propose an optimal scheduling decision for the power system that fully considers economic factors on the premise of considering the actual output boundary value of the gas turbine unit and meeting the safe operation of the system, rationally allocate power resources, reduce the operation cost of the power system, and improve the economy of the electric-gas integrated energy system; an optimal scheduling decision for the power system is generated by a model-free method. By constructing a feedback-compensation multi-module neural network and a penalty function neural network to handle the economic scheduling problem of the power system, it does not rely on a complex system model and reduces the time cost of model establishment and solution. Among them, by constructing and training a feedback-compensation multi-module neural network, the operation safety domain of the gas turbine unit can be accurately and real-time predicted, providing guarantee for the safe operation of the pipe network. By constructing and training a penalty function neural network, the optimal scheduling problem of the power system is integrated into network training. The trained penalty function neural network can provide an optimal scheduling strategy in real time according to the operation safety domain of the gas turbine unit provided by the feedback-compensation multi-module neural network; operating the gas turbine unit safety domain prediction model and the optimal decision-making generation model for the power system in coordination can effectively avoid the influence on the scheduling problem of the electric-gas integrated energy system caused by inaccurate model parameters and overly complex models, and at the same time greatly shorten the problem-solving time, improve the efficiency of decision-making generation, and meet the real-time requirement. The data sources of this invention are natural gas system data, gas turbine unit safety domain data, and power system data, covering multiple aspects of information such as gas source pressure, compressor boosting, natural gas load, safe operation output boundary of gas turbine units, and power load, wind power output, etc.Through the normalization and comprehensive utilization of these data, the mutual relations and influences among different subsystems in the electric-gas integrated energy system are fully considered, which can more comprehensively ensure the safe operation of the pipe network, improve the adaptability of the decision-making generation method, and maintain stability under complex and changeable actual operating conditions. Among them, a safety domain prediction model for gas turbines is constructed and trained using a multi-module neural network training set, making the prediction of the safe operation of gas turbines more accurate, avoiding potential safety problems caused by improper operation of gas turbines from the source, and further ensuring the safe and stable operation of the entire system. The fast decision-making generation method of the present invention is data-driven. Through the collection, processing, and training of a large amount of historical data, it can adapt to the electric-gas integrated energy system under different working conditions and operating conditions. As the system operation data continues to accumulate, the model can be continuously optimized and updated, further improving the accuracy and adaptability of prediction and decision-making, and having strong generality and scalability.

[0137] The fast decision-making generation method for the electric-gas integrated energy system based on neural network involved in this application mainly aims at the technical problems that the existing technology uses the model-driven method to model and solve to obtain the optimal strategy for collaborative decision-making of the electric-gas integrated energy system, which is complex to solve and difficult to obtain the topological structure and network-related parameters of the complex system.

[0138] The fast decision-making generation method for the electric-gas integrated energy system based on neural network involved in this application will be described in detail below. For the purpose of illustration rather than limitation, specific details such as specific system structures and technologies are proposed to thoroughly understand the embodiments of this application. However, those skilled in the art should clearly understand that this application can also be implemented in other embodiments without these specific details.

[0139] In the fast decision-making generation method for the electric-gas integrated energy system based on neural network involved in this application, the term "including" indicates the existence of the described features, wholes, steps, operations, elements, and / or components, but does not exclude the existence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations. The terms "including", "comprising", "having" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0140] For the convenience of clearly describing the technical solution of this application, terms such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and roles. Those skilled in the art can understand that the terms "first", "second", etc. do not limit the quantity and execution order, and the terms "first", "second", etc. do not necessarily limit to be different.

[0141] Statements such as "an embodiment" or "some embodiments" described in this application mean that the specific features, structures, or characteristics described in the embodiment are included in one or more embodiments of this application. Thus, statements such as "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments", etc. that appear in different places in this application do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways.

[0142] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0143] The method for quickly generating decisions for an electric-gas integrated energy system based on a neural network provided by an embodiment of the present invention is executed by a computer device. Correspondingly, the system for quickly generating decisions for an electric-gas integrated energy system based on a neural network runs in the computer device.

[0144] Figure 1 is a flowchart of the method for quickly generating decisions for an electric-gas integrated energy system based on a neural network according to an embodiment of the present invention. Among them, Figure 1 The execution subject can be a system for quickly generating decisions for an electric-gas integrated energy system based on a neural network. According to different requirements, the order of the steps in this flowchart can be changed, and some can be omitted.

[0145] As Figure 1 shown, the method for quickly generating decisions for an electric-gas integrated energy system based on a neural network includes:

[0146] Step S1, collect the natural gas system data, gas turbine unit safety domain data, and power system data of the same electric-gas integrated energy system at each time period, and after normalization processing, construct a multi-module neural network training set and a power system data set.

[0147] Collecting the natural gas system, gas turbine unit safety domain, and power system data and normalizing them can provide an accurate and standardized data basis for subsequent model training; constructing corresponding training sets according to different data helps to train the model specifically, enabling the model to better learn the data characteristics and laws of each system and improving the accuracy of model prediction and decision-making.

[0148] In some specific embodiments, a multi-module neural network training set is constructed based on the normalized natural gas system data and the gas turbine unit safety domain data, and a power system data set is constructed based on the normalized power system data;

[0149] Among them, the natural gas system data includes the gas source pressure, compressor boost, and natural gas load in the natural gas system;

[0150] The gas turbine unit safety domain data includes the upper output limit and the lower output limit for the safe operation of the gas turbine unit;

[0151] The power system data includes the power load and wind power output data in the power system.

[0152] In some specific embodiments, the power system data set is expressed as , where is the power load, is the output of the wind turbine unit.

[0153] In some specific embodiments, for the original data of any natural gas system data or power system data , the formula for its normalization process is:

[0154]

[0155] In the formula, represents the value after normalization;

[0156] and respectively represent the maximum and minimum values of the data of the corresponding type in the collected samples.

[0157] Step S2, construct a feedback-compensation multi-module neural network and train it using the multi-module neural network training set to obtain a safety domain prediction model. The input of the safety domain prediction model is the natural gas system data, and the output is the predicted value of the gas turbine unit safety domain; the structure of the feedback-compensation multi-module neural network is as Figure 2 shown.

[0158] In some specific embodiments, the feedback-compensation multi-module neural network includes a prediction module, a feedback module, and a compensation module, where:

[0159] The prediction module includes three single-layer perceptrons with activation functions and a multi-layer perceptron. After the gas source pressure, natural gas load, and compressor boost amount in the natural gas system data are respectively input into the three single-layer perceptrons, the single-layer perceptrons extract features, and then the multi-layer perceptron obtains a preliminary estimate of the gas turbine unit safety domain;

[0160] The feedback module includes a multi-layer perceptron. The feedback module receives the preliminary estimated value of the safety domain of the gas turbine unit obtained by the prediction module and the corresponding natural gas system data, extracts data features through the multi-layer perceptron, and generates the correction values of the upper and lower limits of the safe output of the gas turbine unit.

[0161] The compensation module includes a multi-layer perceptron. The compensation module is used to capture the fluctuations in the safety domain of the gas turbine unit caused by the overall change of the natural gas system data. It receives the total gas source pressure data, the total natural gas load data, and the total compressor boost data, and generates the compensation amount of the upper and lower limits of the safe output of the gas turbine unit, which is used to compensate for the fluctuations in the safety domain of the gas turbine unit caused by the change of the natural gas system input.

[0162] By combining the prediction module, the feedback module, and the compensation module, a feedback-compensation multi-module neural network is constructed to calculate the predicted value of the safety domain of the gas turbine unit. The prediction module, the feedback module, and the compensation module cooperate with each other, and can more accurately capture the complex relationship between the natural gas system data and the safety domain of the gas turbine unit. The feedback module can correct the upper and lower limits of the safe output according to the preliminary estimated value and the input data, and the compensation module can capture the fluctuations caused by the overall change of the system data, improving the accuracy of the prediction of the safety domain of the gas turbine unit and providing a reliable basis for subsequent decisions.

[0163] Defining the specific structure and functions of each module of the feedback-compensation multi-module neural network enables the network to predict and correct the safety domain of the gas turbine unit from different perspectives. Each module has a clear division of labor and cooperates with each other, improving the network's ability to process complex data relationships. Compared with a neural network with a single structure, it can more accurately predict the safety domain of the gas turbine unit, enhancing the reliability and adaptability of the model.

[0164] In step S3, the natural gas system data is input into the safety domain prediction model, and the actual output boundary value of the gas turbine unit is calculated according to the predicted value of the safety domain of the gas turbine unit and the inherent output boundary value of the gas turbine unit. The actual output boundary value is spliced with the power system data to construct a penalty function neural network training set.

[0165] Among them, the inherent output boundary value of the gas turbine unit is the upper and lower boundary values of the output set during the design and manufacture of the gas turbine unit.

[0166] Calculating the actual output boundary value through the predicted value of the safety domain of the gas turbine unit and the inherent output boundary value, and constructing a penalty function neural network training set can comprehensively consider the prediction results and the inherent capabilities of the unit itself, making the decisions generated by the subsequent trained penalty function neural network more in line with the actual operating conditions, ensuring that the gas turbine unit operates within a safe and reasonable output range, and guaranteeing the safety and stability of the system.

[0167] In some specific embodiments, the calculation formula for the actual output boundary value of the gas turbine unit is:

[0168]

[0169]

[0170] In the formula, is the upper output limit in the actual output boundary value of the gas turbine unit;

[0171] is the lower output limit in the actual output boundary value of the gas turbine unit;

[0172] is the upper output limit in the predicted value of the safety domain of the gas turbine unit;

[0173] is the lower output limit in the predicted value of the safety domain of the gas turbine unit;

[0174] is the upper output limit in the inherent output boundary value of the gas turbine unit;

[0175] is the lower output limit in the inherent output boundary value of the gas turbine unit.

[0176] The calculation formula of the actual output boundary of the gas turbine unit comprehensively considers the predicted value of the safety domain of the gas turbine unit and the inherent output boundary value, ensuring that the calculated actual output boundary is both based on the prediction result and in line with the capacity limit of the unit itself, providing an accurate basis for determining the actual operating output range of the unit in the future and ensuring the safety and feasibility of the operation of the gas turbine unit.

[0177] In some specific embodiments, the penalty function neural network training set is expressed as .

[0178] The representation form of the penalty function neural network training set facilitates the management, processing and model training of data. During the model training process, a clear data structure helps the neural network accurately learn the data features and relationships, improve the training efficiency and model performance, and enable the model to be better applied to the actual power system decision-making.

[0179] Step S4, construct a penalty function neural network and train it using the penalty function neural network training set to obtain an optimal decision-making generation model. The structure of the penalty function neural network is as Figure 3 shown;

[0180] The input of the optimal decision-making generation model of the power system is the actual output boundary value of the gas turbine unit, the power load, and the wind power output, and the output is the scheduling strategy of the output of the gas turbine unit and the coal-fired unit.

[0181] An economic dispatch model of the power system is established, and a penalty function neural network is constructed and trained to solve this model to obtain an optimal decision-making generation model, which takes into account the output constraints of gas-fired units and coal-fired units and the source-load power balance constraint, with the goal of minimizing the power generation cost of coal-fired units; the penalty function neural network can effectively solve this complex model by reasonably setting the loss function and weight factors, and the obtained optimal decision-making generation model can realize the economic dispatch of the power system, reduce the power generation cost, and improve the economic benefits of the system.

[0182] In some specific embodiments, the economic dispatch model of the power system includes the output constraint of gas-fired units, the output constraint of coal-fired units, and the source-load power balance constraint. The expression of the economic dispatch model of the power system is;

[0183]

[0184]

[0185]

[0186] In the formula, is the output of the gas-fired unit at time ;

[0187] is the output of the coal-fired unit at time ;

[0188] is the output of the wind turbine at time ;

[0189] is the power load at time ;

[0190] is the upper limit output in the actual output boundary value of the gas-fired unit at time ;

[0191] is the lower limit output in the actual output boundary value of the gas-fired unit at time ;

[0192] and are the upper and lower boundaries of the output of the coal-fired unit ;

[0193] , , and respectively represent the sets corresponding to coal-fired units, gas-fired units, wind turbine units, and power loads;

[0194] The objective function is to minimize the power generation cost of coal-fired units, and the objective function is;

[0195]

[0196] In the formula, , , are the quadratic term, linear term coefficient, and constant term corresponding to the cost function of coal-fired units;

[0197] is the total scheduling period.

[0198] The constraint conditions and objective function of the power system economic dispatch model consider the output limits of gas-fired units and coal-fired units as well as the power balance between sources and loads, with the goal of minimizing the power generation cost of coal-fired units. These constraints and objectives meet the actual operation requirements of the power system. By optimizing this model, the reasonable allocation of power system resources can be achieved, the energy utilization efficiency can be improved, the power generation cost can be reduced, and the economic and stable operation of the power system can be guaranteed.

[0199] In some specific embodiments, the expression of the penalty function neural network is:

[0200] ;

[0201] In the formula, is the output of the gas-fired unit;

[0202] is the output of the coal-fired unit;

[0203] is the function for calculating the power system dispatch decision;

[0204] represents the model parameters of the penalty function neural network, including the weights and biases of neurons in each layer;

[0205] The expression of the loss function of the penalty function neural network is:

[0206]

[0207]

[0208]

[0209]

[0210]

[0211] In the formula, is the weight factor of the loss term for the power balance constraint;

[0212] is the weight factor of the loss term for the output constraint of the gas turbine unit;

[0213] is the weight factor of the loss term for the output constraint of the coal-fired unit.

[0214] By setting a reasonable loss function and comprehensively considering various factors such as power balance constraints, output constraints of gas turbine units and coal-fired units, etc., it can effectively guide the neural network to learn the optimal solution that meets the actual operation requirements; the weight factor can adjust the importance of each constraint according to the actual situation, improve the flexibility and adaptability of the model, and enable the penalty function neural network to better solve the economic dispatch model of the power system.

[0215] In some specific embodiments, the training process of the penalty function neural network is as follows:

[0216] Input the training set of the penalty function neural network into the penalty function neural network, then calculate the loss function value, update the network parameters through the Adam optimizer, and then recalculate according to the new parameters, and repeat the iterative update until the preset number of training rounds is reached to obtain the optimal decision-making model for the power system.

[0217] Using the Adam optimizer for iterative update can continuously adjust the network parameters, gradually reduce the loss function value, so that the model is closer to the optimal solution; the preset number of training rounds ensures the controllability of the training process, ensures that the model reaches better performance within a reasonable time and computing resources, improves the efficiency and quality of model training, and obtains a more reliable optimal decision-making model for the power system.

[0218] Step S5: Input the data of the natural gas system to be decided into the safety domain prediction model, use its output to calculate the actual processing boundary value of the gas group, and then input it together with the power system data into the optimal decision-making model to generate the optimal dispatch decision of the power system.

[0219] Running the two models in sequence to generate the optimal dispatch decision of the power system, combining the safety domain prediction of the gas turbine unit with the economic dispatch decision of the power system, not only ensures the safe operation of the gas turbine unit, but also realizes the economic optimal dispatch of the power system, and comprehensively improves the operation performance and benefits of the electric-gas integrated energy system.

[0220] In a specific embodiment, the method for quickly generating decisions for an electric-gas integrated energy system based on a neural network includes:

[0221] Step S1, collect the natural gas system data, gas turbine unit safety domain data, and power system data for each time period of the same electrical-gas integrated energy system, perform normalization processing, construct a multi-module neural network training set based on the normalized natural gas system data and gas turbine unit safety domain data, and construct a power system data set based on the normalized power system data;

[0222] Among them, the natural gas system data includes the gas source pressure, compressor boost, and natural gas load in the natural gas system;

[0223] The gas turbine unit safety domain data includes the upper output limit and lower output limit for the safe operation of the gas turbine unit;

[0224] The power system data includes the power load and wind power output data in the power system;

[0225] For the original data of any natural gas system data or power system data , the formula for its normalization processing is:

[0226]

[0227] In the formula, represents the value after normalization processing;

[0228] and respectively represent the maximum and minimum values of the data of the corresponding type in the collected samples;

[0229] Step S2, construct a feedback-compensation multi-module neural network, use the multi-module neural network training set to train the parameters of the feedback-compensation multi-module neural network, obtain a gas turbine unit safety domain prediction model, and the input of the gas turbine unit safety domain prediction model is the natural gas system data, and the output is the predicted value of the gas turbine unit safety domain;

[0230] The feedback-compensation multi-module neural network includes a prediction module, a feedback module, and a compensation module, where:

[0231] The prediction module includes three single-layer perceptrons with activation functions and a multi-layer perceptron. Among them, the single-layer perceptron for processing the gas source pressure contains 5 neurons, uses the ReLU activation function, the initial value of the weight matrix is set by the Xavier initialization method, the initialization range is [-0.5, 0.5], and the initial value of the bias vector is set to 0.1;

[0232] The single-layer perceptron for processing the natural gas load has 8 neurons, uses the ReLU activation function, the initialization range of the weight matrix is [-0.5, 0.5], and the initial value of the bias vector is set to 0.2;

[0233] The single-layer perceptron for processing the compressor boost volume contains 6 neurons, uses the ReLU activation function, the initial range of the weight matrix is [-0.5, 0.5], and the bias vector is 0.15;

[0234] The multi-layer perceptron has 2 hidden layers. The first hidden layer contains 10 neurons, the second hidden layer contains 8 neurons, and the output layer has 2 neurons, corresponding to the upper output and lower output of the preliminary estimate of the safety domain of the gas turbine unit respectively; the activation function of the hidden layer is ReLU, and the output layer uses the linear activation function; the initial value of the weight matrix is determined by Xavier initialization, and the value range is [-0.5, 0.5], and the initial value of the bias vector is set in [0.1, 0.3];

[0235] After the gas source pressure, natural gas load, and compressor boost volume in the natural gas system data are respectively input into the corresponding single-layer perceptron, the single-layer perceptron extracts features, and then the preliminary estimate value of the safety domain of the gas turbine unit is obtained through the multi-layer perceptron. The expression of the prediction module is:

[0236]

[0237] In the formula, is the set of upper outputs in the preliminary estimate value of the safety domain of the gas turbine unit;

[0238] is the set of lower outputs in the preliminary estimate value of the safety domain of the gas turbine unit;

[0239] is the function for calculating the preliminary estimate value of the safety domain of the gas turbine unit;

[0240] represents the model parameters of the four perceptrons in the prediction module, including the weights and biases of each layer of neurons;

[0241] is the gas source pressure;

[0242] is the natural gas load;

[0243] is the compressor boost volume;

[0244] The feedback module includes a multi-layer perceptron with three hidden layers. The number of neurons in each layer is 12, 10, and 8 in sequence, and the output layer has two neurons, which are used to generate the correction values of the upper and lower limits of the safe output of the gas turbine unit. The activation function of the hidden layer is ReLU, and the output layer is a linear activation function. The initial values of the weight matrices are obtained by Xavier initialization, with a range of [-0.5, 0.5], and the initial values of the bias vectors are in [0.1, 0.3]. The feedback module receives the preliminary estimated value of the safe domain of the gas turbine unit obtained by the prediction module and the corresponding natural gas system data, extracts data features through the multi-layer perceptron, and generates the correction values of the upper and lower limits of the safe output of the gas turbine unit. The expression of the feedback module is:

[0245]

[0246] In the formula, is the correction value of the upper limit of the safe output of the gas turbine unit;

[0247] is the correction value of the lower limit of the safe output of the gas turbine unit;

[0248] is the function for calculating the correction value of the safe output of the gas turbine unit;

[0249] represents the model parameters of the multi-layer perceptron in the feedback module, including the weights and biases of the neurons in each layer;

[0250] The compensation module includes a multi-layer perceptron with two hidden layers. The first hidden layer contains 10 neurons, the second hidden layer contains 6 neurons, and the output layer has two neurons, corresponding to the compensation amounts of the upper and lower limits of the safe output of the gas turbine unit. The activation function of the hidden layer is ReLU, and the output layer is a linear activation function. The initial values of the weight matrices are initialized by Xavier, with a value range of [-0.5, 0.5], and the initial values of the bias vectors are [0.1, 0.2]. The compensation module is used to capture the fluctuations in the safe domain of the gas turbine unit caused by the overall change of the natural gas system data. It receives the total gas source pressure data, the total natural gas load data, and the total compressor boost data, and generates the compensation amounts of the upper and lower limits of the safe output of the gas turbine unit, which are used to compensate for the fluctuations in the safe domain of the gas turbine unit caused by the change of the natural gas system input. The expression of the compensation module is:

[0251]

[0252] In the formula, is the compensation amount of the upper limit of the safe output of the gas turbine unit;

[0253] is the compensation amount of the upper limit of the safe output of the gas turbine unit;

[0254] A function for calculating the safety output compensation of a gas turbine unit;

[0255] Are the model parameters of the multi-layer perceptron in the compensation module, including the weights and biases of neurons in each layer;

[0256] Is the total gas source pressure;

[0257] Is the total natural gas load;

[0258] Is the total compressor boost;

[0259] By combining the prediction module, feedback module and compensation module, a feedback-compensation multi-module neural network is constructed to calculate the predicted value of the safety domain of the gas turbine unit. The expression for the predicted value of the safety domain of the gas turbine unit is:

[0260]

[0261] In the formula, Is a function for calculating the predicted value of the safety domain of the gas turbine unit;

[0262] Is the upper limit output in the predicted value of the safety domain of the gas turbine unit;

[0263] Is the lower limit output in the predicted value of the safety domain of the gas turbine unit;

[0264] By training the feedback-compensation multi-module neural network with a multi-module neural dataset , the data batch size is set to 32, the learning rate is set to 0.001, the Adam optimizer is used for parameter update, and the preset number of training epochs is 500. The training steps are:

[0265] S201. Initialize the parameters of the feedback-compensation multi-module neural network using the Xavier method, including initializing the weights and biases of neurons in each layer;

[0266] S202. Construct the loss function of the feedback-compensation multi-module neural network using the mean squared error MSE. The loss function of the feedback-compensation multi-module neural network is specifically:

[0267]

[0268]

[0269]

[0270]

[0271] In the formula, represents the model prediction error calculated by MSE, which constitutes the main component of the loss function; secondly;

[0272] represents the error between the output of the prediction module calculated by MSE and the true value, which aims to reduce the correction amount from the feedback module and the compensation amount from the compensation module, thereby improving the overall accuracy;

[0273] is a regularization term, and the Manhattan norm is adopted to limit the magnitude of the compensation amount;

[0274] represents the data batch size;

[0275] and represent scale factors;

[0276] and represent the Manhattan norm and the Euclidean norm respectively;

[0277] and represent the set of predicted values of the gas turbine unit safety domain;

[0278] and represent the set of preliminary estimated values of the gas turbine unit safety domain;

[0279] and express the true values corresponding to the output of the feedback-compensation multi-module neural network;

[0280] S203. Calculate the loss function according to the multi-module neural network training set and the predicted values of the gas turbine unit safety domain, calculate the gradient of the loss function with respect to the network parameters using the chain rule and automatic differentiation, then update the network parameters through the Adam optimizer, and then calculate the loss function between the prediction and the true value according to the new network parameters, and repeat the iterative update until the preset number of training rounds is reached to obtain the gas turbine unit safety domain prediction model;

[0281] During the training process, monitor the model performance with the loss function value of the validation set. If the loss of the validation set does not decrease for 20 consecutive rounds, terminate the training in advance to prevent overfitting;

[0282] Step S3: Input the natural gas system data corresponding to each group of power system data in the power system dataset into the security domain prediction model, output the corresponding predicted values of the gas turbine unit security domain, then calculate the actual output boundary value of the gas turbine unit according to the predicted value of the gas turbine unit security domain and the inherent output boundary value of the gas turbine unit, and splice the actual output boundary value of the gas turbine unit with the corresponding power system data to construct a penalty function neural network training set;

[0283] Among them, the inherent output boundary value of the gas turbine unit is the upper and lower boundary values of the output set during the design and manufacture of the gas turbine unit;

[0284] The calculation formula for the actual output boundary of the gas turbine unit is:

[0285]

[0286]

[0287] In the formula, is the upper limit output in the actual output boundary value of the gas turbine unit;

[0288] is the lower limit output in the actual output boundary value of the gas turbine unit;

[0289] is the upper limit output in the predicted value of the gas turbine unit security domain;

[0290] is the lower limit output in the predicted value of the gas turbine unit security domain;

[0291] is the upper limit output in the inherent output boundary value of the gas turbine unit;

[0292] is the lower limit output in the inherent output boundary value of the gas turbine unit;

[0293] Step S4: Construct a penalty function neural network and train it using the penalty function neural network training set to obtain an optimal decision-making generation model;

[0294] The input of the optimal decision-making generation model for the power system is the actual output boundary value of the gas turbine unit and the power system data, and the output is the dispatching strategy for the output of the gas turbine unit and the coal-fired unit;

[0295] The economic dispatching model of the power system includes the output constraint of the gas turbine unit, the output constraint of the coal-fired unit, and the source-load power balance constraint. The expression of the economic dispatching model of the power system is;

[0296]

[0297]

[0298]

[0299] In the formula, is the output of the gas turbine unit at time ;

[0300] is the output of the coal-fired power unit at time ;

[0301] is the output of the wind turbine unit at time ;

[0302] is the magnitude of the electrical load at time ;

[0303] is the upper output limit in the actual output boundary of the gas turbine unit at time ;

[0304] is the lower output limit in the actual output boundary of the gas turbine unit at time ;

[0305] and are the upper and lower output boundaries of the coal-fired power unit ;

[0306] , , and respectively represent the sets corresponding to the coal-fired power unit, gas turbine unit, wind turbine unit, and electrical load;

[0307] The objective function is to minimize the power generation cost of the coal-fired power unit, and the objective function is;

[0308]

[0309] In the formula, , , are the quadratic term, linear term coefficient, and constant term corresponding to the cost function of the coal-fired power unit;

[0310] is the total scheduling period;

[0311] The expression of the penalty function neural network is:

[0312]

[0313] In the formula, is the output of the gas turbine unit;

[0314] is the output of the coal-fired power unit;

[0315] is a function for calculating the dispatching decision of the power system;

[0316] represents the model parameters of the penalty function neural network, including the weights and biases of neurons in each layer;

[0317] The expression of the loss function of the penalty function neural network is:

[0318]

[0319]

[0320]

[0321]

[0322]

[0323] In the formula, is the weight factor of the loss term of the power balance constraint;

[0324] is the weight factor of the loss term of the output constraint of the gas turbine unit;

[0325] is the weight factor of the loss term of the output constraint of the coal-fired power unit;

[0326] The training process of the penalty function neural network is as follows:

[0327] Input the training set of the penalty function neural network into the penalty function neural network, then calculate the loss function value, update the network parameters through the Adam optimizer, and recalculate according to the new parameters. Repeat the iterative update until the preset number of training rounds is reached to obtain the optimal decision-making model of the power system;

[0328] Step S5: Input the data of the natural gas system to be decided into the security domain prediction model, use its output to calculate the actual processing boundary value of the gas group, and then input it together with the power system data into the optimal decision-making model to generate the optimal dispatching decision of the power system.

[0329] The following are embodiments of a fast decision-making generation system for an electric-gas integrated energy system based on a neural network. This active power load shedding optimization system and the fast decision-making generation method for an electric-gas integrated energy system in each of the above embodiments belong to the same inventive concept. Details not described in detail in the embodiments of the fast decision-making generation system for an electric-gas integrated energy system based on a neural network can refer to the embodiments of the fast decision-making generation method for an electric-gas integrated energy system based on a neural network.

[0330] Now, mobile terminals implementing various embodiments of the present invention will be described with reference to the accompanying drawings. In the following description, suffixes such as "module", "component", or "unit" used to represent elements are only for the convenience of describing embodiments of the present invention and have no specific meaning in themselves. Therefore, "module" and "component" can be used interchangeably.

[0331] As Figure 4 shown, the fast decision-making generation system for an electric-gas integrated energy system based on a neural network includes:

[0332] A data collection and processing module, configured to collect natural gas system data, gas turbine unit safety domain data, and power system data for each period of the same electric-gas integrated energy system, perform normalization processing, and then construct a multi-module neural network training set and a power system data set;

[0333] A feedback-compensation multi-module neural network construction and training module, configured to construct a feedback-compensation multi-module neural network and train it using the multi-module neural network training set to obtain a safety domain prediction model;

[0334] A gas turbine unit actual output boundary value calculation module, configured to calculate the actual output boundary value of the gas turbine unit;

[0335] A penalty function neural network training set construction module, configured to splice the actual output boundary value of the gas turbine unit with the corresponding power system data to construct a penalty function neural network training set;

[0336] A penalty function neural network construction and training module, configured to construct a penalty function neural network and train it using the penalty function neural network training set to obtain an optimal decision-making generation model;

[0337] A coordinated operation module, configured to input the natural gas system data to be decision-making into the safety domain prediction model, use its output to calculate the actual processing boundary value of the gas turbine unit, and then jointly input it with the power system data into the optimal decision-making model to generate an optimal power system scheduling decision.

[0338] The fast decision-making generation system for an electric-gas integrated energy system in this embodiment is used to implement the fast decision-making generation method for an electric-gas integrated energy system based on a neural network, and the steps include:

[0339] S1. Collect the natural gas system data, gas turbine unit safety domain data, and power system data of the integrated electricity-gas energy system at each time period, perform normalization processing, and construct a multi-module neural network training set and a power system data set.

[0340] S2. Construct a feedback-compensation multi-module neural network and train it using the multi-module neural network training set to obtain a safety domain prediction model. The input of the safety domain prediction model is the natural gas system data, and the output is the predicted value of the gas turbine unit safety domain.

[0341] S3. Input the natural gas system data into the safety domain prediction model, calculate the actual output boundary value of the gas turbine unit according to the predicted value of the gas turbine unit safety domain output and the inherent output boundary value of the gas turbine unit, splice the actual output boundary value with the power system data, and construct a penalty function neural network training set.

[0342] Among them, the inherent output boundary value of the gas turbine unit is the upper and lower boundary values of the output set during the design and manufacture of the gas turbine unit.

[0343] S4. Construct a penalty function neural network and train it using the penalty function neural network training set to obtain an optimal decision-making generation model.

[0344] The input of the power system optimal decision-making generation model is the actual output boundary value of the gas turbine unit and the power system data, and the output is the dispatching strategy of the output of the gas turbine unit and the coal-fired unit.

[0345] S5. Input the natural gas system data to be decision-making into the safety domain prediction model, use its output to calculate the actual processing boundary value of the gas group, and then jointly input it with the power system data into the optimal decision-making model to generate an optimal dispatching decision for the power system.

[0346] This application also provides an electronic device for implementing various embodiments of the present invention. The electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor.

[0347] Those skilled in the art can understand that the structure of the electronic device involved in the embodiments of the present invention does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine some components, or have different component arrangements.

[0348] Figure 5 Schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0349] The electronic device includes, but is not limited to, components such as a processor and a memory. Those skilled in the art can understand that the structure of the electronic device involved in the embodiments of the present invention does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0350] In the embodiments of the present invention, the electronic device includes, but is not limited to, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device may also represent various forms of mobile devices and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described herein and / or claimed.

[0351] In the embodiments of the present application, the processor may be implemented by using at least one of an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a processor, a controller, a microcontroller, a microprocessor, and an electronic unit designed to perform the functions described herein. In some cases, such an implementation may be implemented in the controller. For a software implementation, an implementation of a process or function may be implemented with a separate software module that permits performing at least one function or operation. The software code may be implemented by a software application (or program) written in any suitable programming language. The software code may be stored in the memory and executed by the controller.

[0352] In addition, the electronic device includes some functional modules not shown herein, which will not be elaborated herein.

[0353] Those skilled in the art to which the present application pertains can understand that various aspects of the electronic device provided by the present application can be implemented as a system, a method, or a program product. Therefore, various aspects of the present disclosure can be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which can be collectively referred to herein as "circuit", "module", or "system".

[0354] The present application also provides a storage medium. In the storage medium, a program product is stored that can implement the method for quickly generating decisions for an electric-gas integrated energy system based on a neural network. In some possible implementation manners, various aspects of the present disclosure can also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present disclosure described in the above "Exemplary Method" section of this specification.

[0355] The storage medium may adopt any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may, for example, but not be limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0356] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A rapid decision-making generation method for an electric-gas integrated energy system based on a neural network, characterized in that the steps Including: S1. Collect the natural gas system data, gas turbine unit safety domain data, and power system data of each period of the same integrated electricity-gas energy system, perform normalization processing, and construct a multi-module neural network training set and a power system data set; S2. Construct a feedback-compensation multi-module neural network and train it using the multi-module neural network training set to obtain a safety domain prediction model; The feedback-compensation multi-module neural network includes a prediction module, a feedback module, and a compensation module, where: The prediction module includes three single-layer perceptrons with activation functions and a multi-layer perceptron. After the gas source pressure, natural gas load, and compressor boost in the natural gas system data are respectively input into the three single-layer perceptrons, the single-layer perceptrons extract features, and then the multi-layer perceptron obtains a preliminary estimate of the gas turbine unit safety domain; The feedback module includes a multi-layer perceptron. The feedback module receives the preliminary estimate of the gas turbine unit safety domain obtained by the prediction module and the corresponding natural gas system data, extracts data features through the multi-layer perceptron, and generates correction values for the upper and lower limits of the safe output of the gas turbine unit; The compensation module includes a multi-layer perceptron. The compensation module is used to capture the fluctuations in the gas turbine unit safety domain caused by changes in the overall natural gas system data, receives the total gas source pressure data, total natural gas load data, and total compressor boost data, and generates compensation amounts for the upper and lower limits of the safe output of the gas turbine unit to compensate for the fluctuations in the gas turbine unit safety domain caused by changes in the natural gas system input; By combining the prediction module, the feedback module, and the compensation module, a feedback-compensation multi-module neural network is formed to calculate the predicted value of the gas turbine unit safety domain; S3. Input the natural gas system data into the safety domain prediction model, calculate the actual output boundary value of the gas turbine unit according to the predicted value of the gas turbine unit safety domain output and the inherent output boundary value of the gas turbine unit, splice the actual output boundary value with the power system data, and construct a penalty function neural network training set; S4. Construct a penalty function neural network and train it using the penalty function neural network training set to obtain an optimal decision-making generation model; S5. Input the natural gas system data to be decided into the safety domain prediction model, use its output to calculate the actual processing boundary value of the gas turbine group, and then input it together with the power system data into the optimal decision-making model to generate an optimal scheduling decision for the power system.

2. The rapid decision-making generation method for the electrical-gas integrated energy system according to claim 1, characterized in that In step S1, a multi-module neural network training set is constructed according to the normalized natural gas system data and gas turbine unit safety domain data, and a power system data set is constructed according to the normalized power system data; Among them, the natural gas system data includes the gas source pressure, compressor boost, and natural gas load in the natural gas system; The gas turbine unit safety domain data includes the upper limit output and the lower limit output of the gas turbine unit operating safely; The power system data includes the power load and wind power output data in the power system.

3. The rapid decision-making generation method for the electric-gas integrated energy system according to claim 1, wherein In step S3, the calculation formula for the actual output boundary of the gas turbine unit is: In the formula, is the upper output limit among the actual output boundaries of the gas turbine unit; is the lower output limit among the actual output boundary values of the gas turbine unit; is the upper output limit in the predicted value of the safety domain of the gas turbine unit; is the lower output limit in the safety domain prediction value of the gas turbine unit; is the upper limit output among the inherent output boundary values of the gas turbine unit; It is the lower limit output among the inherent output boundary values of the gas turbine unit.

4. The rapid decision-making generation method for the electric-gas integrated energy system according to claim 3, wherein In step S1, the power system data set is represented as , where is the power load, is the output of the wind turbine generator; In step S3, the penalty function neural network training set is expressed as .

5. The rapid decision-making generation method for the electric-gas integrated energy system according to claim 1, characterized in that Step S4 is specifically: establish a power system economic dispatch model and a penalty function neural network, and based on the objectives and constraints defined by the power system economic dispatch model, use the penalty function neural network training set to train the penalty function neural network to approximate the optimal solution in a data-driven manner, and finally form a decision-making generation model; The economic dispatch model of the power system includes the output constraints of gas-fired units, the output constraints of coal-fired units, and the power balance constraints between the source and the load. The expression of the economic dispatch model of the power system is: Wherein, is the gas turbine unit at time output; for a coal-fired unit at the moment output for a wind turbine at time output is the electrical load at the moment magnitude; For a gas turbine unit At the moment The upper output limit in the actual output boundary value for a gas turbine unit at the moment the lower output limit in the actual output boundary value and are the upper and lower boundaries of the output of the coal-fired unit respectively. , , and represent the sets corresponding to coal-fired units, gas-fired units, wind turbine units, and power loads, respectively; The objective function is to minimize the power generation cost of coal-fired units. The objective function is: In the formula, , , are the quadratic term, the linear term coefficient and the constant term corresponding to the cost function of the coal-fired unit; is the total scheduling period.

6. The rapid decision-making generation method for the electric-gas integrated energy system according to claim 5, wherein In step S4, the expression of the penalty function neural network is: ； Wherein, is the output of the gas turbine unit; is the output of the coal-fired unit; A function for calculating power system dispatching decisions; Denote the model parameters of the penalty function neural network, including the weights and biases of neurons in each layer; The expression of the loss function of the penalty function neural network is: In the formula, is the weight factor of the loss term of the power balance constraint; is the weight factor of the loss term for the output constraint of the gas turbine unit; It is the weight factor of the loss item for the output constraint of the coal-fired unit.

7. A fast decision-making generation system for an electric-gas integrated energy system based on a neural network, characterized in that, A method for quickly generating decisions for an electric-gas integrated energy system as described in any one of claims 1-6, including: A data collection and processing module for collecting natural gas system data, gas unit safety domain data, and power system data for each period of the same electric-gas integrated energy system, performing normalization processing, and constructing a multi-module neural network training set and a power system data set; A feedback-compensated multi-module neural network construction and training module for constructing a feedback-compensated multi-module neural network and training it using the multi-module neural network training set to obtain a safety domain prediction model; A gas unit actual output boundary value calculation module for calculating the actual output boundary value of the gas unit; A penalty function neural network training set construction module for splicing the actual output boundary value of the gas unit with the corresponding power system data to construct a penalty function neural network training set; A penalty function neural network construction and training module for constructing a penalty function neural network and training it using the penalty function neural network training set to obtain an optimal decision generation model; A coordinated operation module for inputting the natural gas system data to be decided into the safety domain prediction model, using its output to calculate the actual processing boundary value of the gas unit, and then jointly inputting it with the power system data into the optimal decision model to generate an optimal dispatching decision for the power system.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor is used to execute the computer program to implement the steps of the method for quickly generating decisions for the electric-gas integrated energy system as described in any one of claims 1-6.

9. A storage medium, characterized in that, A computer program is stored on a storage medium. When the computer program is executed by a processor, it implements the steps of the method for quickly generating decisions for the electric-gas integrated energy system as described in any one of claims 1-6.

Images