Gas-electricity combined system safety risk assessment and emergency co-processing method and system

Abstract

The invention relates to the technical field of comprehensive energy system safety operation, in particular to a gas-electricity combined system safety risk assessment and emergency co-processing method and system, and the method comprises the steps: constructing a multi-level risk assessment index system; based on a multi-level risk assessment index system, a static index and a dynamic index are fused through an improved entropy weight-TOPSIS algorithm, and a comprehensive risk level is calculated; presetting a risk level threshold value of the safety of the gas-electricity combined system, judging whether the comprehensive risk level reaches the risk level threshold value or not, and if the comprehensive risk level does not reach the risk level threshold value, performing normal operation of the gas-electricity combined system; if so, triggering a multi-agent deep reinforcement learning MADRL framework, and generating an optimal collaborative regulation strategy; and carrying out emergency coordination treatment on the gas-electricity combined system according to the optimal coordination regulation strategy. According to the invention, the problems of one-sided gas-electricity coupling risk assessment, lagging emergency response and lack of dynamic indexes are effectively solved, and the real-time quantification, rapid positioning and coordinated regulation of the gas-electricity coupling risk are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of safe operation of integrated energy systems, and in particular to a method and system for safety risk assessment and emergency coordinated disposal of a gas-electricity combined system. Background Art

[0002] As energy structures transform, combined gas-electricity systems (such as gas turbines and electricity-to-hydrogen) are increasingly used in new power systems. However, due to the differences in the dynamic characteristics of natural gas and electricity networks, the operational risks of coupled systems have increased significantly, making traditional safety assessment methods for single-energy systems difficult to apply.

[0003] There is an incomplete risk assessment problem in existing technologies. Existing methods mostly target single energy systems (such as pure electric systems) and do not fully consider the chain failure propagation mechanism of gas-electricity coupling; emergency response is delayed, accident handling relies on manual experience, and there is a lack of rapid decision-making support for multi-energy collaboration; there is a lack of indicator system, and there is a lack of dynamic assessment indicators to quantify gas-electricity interaction risks. Key parameters such as fault propagation delay and energy conversion efficiency attenuation rate have not been effectively modeled. Therefore, there is an urgent need for a gas-electricity combined system safety risk assessment and emergency collaborative disposal method. Summary of the Invention

[0004] The present invention provides a method and system for safety risk assessment and emergency coordinated disposal of a gas-electricity combined system, which can effectively solve the problems in the background technology.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] A method for safety risk assessment and emergency coordinated disposal of a gas-electricity combined system, the method comprising:

[0007] Constructing a multi-level risk assessment indicator system, wherein the multi-level risk assessment indicator system includes a static indicator layer and a dynamic indicator layer;

[0008] Based on the multi-level risk assessment indicator system, the static and dynamic indicators are integrated through the improved entropy weight-TOPSIS algorithm to calculate the comprehensive risk level;

[0009] Presetting a risk level threshold for the safety of the gas-electricity combined system, determining whether the comprehensive risk level reaches the risk level threshold, and if not, operating the gas-electricity combined system normally;

[0010] If it is reached, the multi-agent deep reinforcement learning MADRL framework is triggered to generate the optimal collaborative control strategy;

[0011] The gas-electricity combined system is coordinated and handled in an emergency according to the optimal coordinated control strategy.

[0012] Furthermore, the static indicator layer includes network topology coupling and device N-1 margin, and the dynamic indicator layer includes power flow matching deviation rate and fault propagation delay coefficient.

[0013] Furthermore, the fault propagation delay coefficient is calculated using a directed graph model and a Djkstra algorithm.

[0014] Furthermore, the improved entropy weight-TOPSIS algorithm is used to integrate static indicators and dynamic indicators, including dynamically allocating weights based on the information entropy of risk assessment indicators. The weight allocation formula is:

[0015]

[0016] Among them, H j is the information entropy indicator.

[0017] Furthermore, triggering the multi-agent deep reinforcement learning (MADRL) framework to generate the optimal collaborative control strategy includes:

[0018] Privacy-preserving feature extraction of electric power EMS and natural gas SCADA data through federated learning;

[0019] Based on the extracted features, the MADRL framework is used to define the action space of the power-side and natural gas-side agents, and a cross-system collaborative control strategy is generated with the minimization of comprehensive cost as the objective function.

[0020] A digital twin platform was constructed using Modelica and OPAL-RT to simulate and rehearse the cross-system collaborative control strategy and generate the optimal collaborative control strategy.

[0021] Furthermore, the cross-system collaborative control strategy is simulated and rehearsed to generate the optimal collaborative control strategy, including:

[0022] Setting target risk level thresholds and conducting simulation previews of the cross-system collaborative control strategy;

[0023] If the risk level in the simulation result drops to the target risk level threshold, outputting the cross-system collaborative control strategy as the optimal collaborative control strategy;

[0024] If the risk level in the simulation result does not drop to the target risk level threshold, the cross-system collaborative control strategy is optimized through the Stackelberg game model to generate an optimal collaborative control strategy.

[0025] Furthermore, the accuracy error of the gas-electric coupling model of the digital twin platform is ≤2%, and the hardware-in-the-loop simulation step is ≤1ms.

[0026] Furthermore, taking minimizing the comprehensive cost as the objective function, the expression is:

[0027] min(C elec +λC gas +μLoadShed);

[0028] Among them, C elec is the operating cost of the power system, C gas is the operating cost of the natural gas system, LoadShed is the load reduction, and λ and μ are weight coefficients.

[0029] A gas-electricity combined system safety risk assessment and emergency coordinated disposal system, the system comprising:

[0030] An indicator system construction module is used to construct a multi-level risk assessment indicator system, wherein the multi-level risk assessment indicator system includes a static indicator layer and a dynamic indicator layer;

[0031] A risk level calculation module, based on the multi-level risk assessment indicator system, calculates the comprehensive risk level by fusing static indicators and dynamic indicators through an improved entropy weight-TOPSIS algorithm;

[0032] a risk level judgment module, which presets a risk level threshold for the safety of the gas-electricity combined system and judges whether the comprehensive risk level reaches the risk level threshold; if not, the gas-electricity combined system operates normally;

[0033] The control strategy optimization module, if achieved, triggers the multi-agent deep reinforcement learning MADRL framework to generate the optimal collaborative control strategy;

[0034] The combined system coordination module performs emergency coordination and disposal on the combined gas-electricity system according to the optimal coordinated control strategy.

[0035] Furthermore, the control strategy optimization module includes:

[0036] Feature extraction unit, which performs privacy-preserving feature extraction of power EMS and natural gas SCADA data through federated learning;

[0037] The strategy generation unit, based on the extracted features, uses the MADRL framework to define the action space of the power-side and natural gas-side agents, and generates a cross-system collaborative control strategy with the objective function of minimizing the comprehensive cost;

[0038] The strategy optimization unit uses Modelica and OPAL-RT to build a digital twin platform, simulates and rehearses the cross-system collaborative control strategy, and generates the optimal collaborative control strategy.

[0039] A computer-readable storage medium storing a computer program, characterized in that when the program is executed, the method according to any one of claims 1 to 8 is implemented.

[0040] The technical solution of the present invention can achieve the following technical effects:

[0041] It effectively solves the problems of one-sided gas-electricity coupling risk assessment, delayed emergency response and missing dynamic indicators. By constructing a multi-dimensional dynamic risk assessment indicator system and combining digital twins with multi-agent reinforcement learning technology, it realizes real-time quantification, rapid positioning and coordinated control of gas-electricity coupling risks, effectively reducing risk assessment errors and accident response time. At the same time, it has good compatibility and scalability, and supports seamless docking with existing EMS / SCADA systems without the need for additional hardware modification.

[0042] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 A flow chart of the safety risk assessment and emergency coordinated disposal method for the gas-electricity combined system;

[0045] Figure 2 It is a diagram of the multi-level risk assessment indicator system structure;

[0046] Figure 3 This is a structural diagram of the gas-electricity combined system safety risk assessment and emergency coordinated disposal system. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0049] Example 1

[0050] like Figure 1 As shown in FIG, the gas-electricity combined system safety risk assessment and emergency coordinated disposal method includes:

[0051] S1: Construct a multi-level risk assessment indicator system, which includes a static indicator layer and a dynamic indicator layer;

[0052] In this embodiment, the gas-electricity combined system is a complex coupling of electricity and natural gas networks. A single indicator cannot fully reflect the risks. This embodiment combines static indicators with dynamic indicators to cover multi-dimensional risks. Static indicators are used to evaluate the stability of the system structure, and dynamic indicators are used to capture real-time operational anomalies. The combination of the two can avoid one-sided risk assessment.

[0053] S2: Based on a multi-level risk assessment indicator system, the static and dynamic indicators are integrated through the improved entropy weight-TOPSIS algorithm to calculate the comprehensive risk level;

[0054] Specifically, when dealing with complex coupled systems such as gas-electricity systems, single static or dynamic indicators alone cannot fully reflect the system's risks. This step integrates static and dynamic indicators, combining them with structural safety and real-time status to avoid a one-sided risk assessment. Due to the different dimensions and ranges of variation between static and dynamic indicators, direct integration can lead to bias. Therefore, this step uses an improved entropy-weighted TOPSIS algorithm to achieve a scientific integration of static and dynamic indicators. The improved entropy-weighted TOPSIS algorithm combines the entropy-weighted and TOPSIS methods, assigning weights based on the discreteness of the indicator data. It then evaluates each solution's overall performance by calculating its weighted distance from the positive / negative ideal solution, such as outputting a risk rating of low, medium, high, or urgent. Because the traditional entropy-weighted TOPSIS algorithm suffers from static weights that cannot adapt to dynamic operating scenarios and standardization distortion caused by differences in indicator dimensions, this embodiment improves the algorithm by introducing a dynamic weighting correction based on the gas-electricity coupling factor, enhancing the weights of key indicators, and employing dynamic time warping (DTW) to align the heterogeneous time series data of the power and natural gas systems to eliminate sampling frequency differences.

[0055] S3: Preset the risk level threshold for the safety of the gas-electricity combined system and determine whether the comprehensive risk level reaches the risk level threshold. If not, the gas-electricity combined system operates normally.

[0056] S4: If it is reached, the multi-agent deep reinforcement learning MADRL framework is triggered to generate the optimal collaborative control strategy;

[0057] S5: Conduct emergency coordination and disposal of the gas-electricity combined system according to the optimal coordinated control strategy.

[0058] As a preferred embodiment of this invention, it is determined whether cross-system coordinated regulation of the gas-electricity combined system is required based on the comprehensive risk level obtained in the above steps. In order to avoid response delays or inconsistencies caused by manual experience, this embodiment realizes automatic triggering through thresholds. In this step, a level is preset as the risk level threshold in the divided risk levels. If the risk level is greater than or equal to the level threshold, a coordinated regulation strategy is generated to coordinate the gas-electricity combined system. High-cost emergency measures are only initiated when the risk exceeds the conventional control capability to avoid excessive regulation under low risk.

[0059] Through the present invention, the problems of one-sided gas-electricity coupling risk assessment, delayed emergency response and missing dynamic indicators are effectively solved. By constructing a multi-dimensional dynamic risk assessment indicator system and combining digital twins with multi-agent reinforcement learning technology, real-time quantification, rapid positioning and coordinated regulation of gas-electricity coupling risks are achieved, effectively reducing risk assessment errors and accident response time. At the same time, it has good compatibility and scalability, and supports seamless docking with existing EMS / SCADA systems without the need for additional hardware modification.

[0060] Based on the above embodiments, Figure 2 As shown in Figure 1, the static indicator layer includes network topology coupling and device N-1 margin, and the dynamic indicator layer includes power flow matching deviation rate and fault propagation delay coefficient.

[0061] Furthermore, the fault propagation delay coefficient is calculated using a directed graph model and the Djkstra algorithm.

[0062] Specifically, the network topology coupling where N coupling is the number of gas-electric direct coupling nodes, N total is the total number of system nodes; N-1 fault margin of key equipment P max is the maximum capacity of the equipment, P load is the real-time load;

[0063] Dynamic indicators:

[0064] Gas-electricity flow matching deviation rate F gas is the natural gas flow rate, F elec is the equivalent power flow; fault propagation time delay coefficient Solve the maximum delay t of the critical path through the directed graph model and Dijkstra algorithm max , t threshold =300 seconds.

[0065] Furthermore, the static and dynamic indicators are integrated by improving the entropy weight-TOPSIS algorithm, including dynamically allocating weights through the information entropy of risk assessment indicators. The weight allocation formula is:

[0066]

[0067] Among them, H j is the information entropy indicator.

[0068] As a preferred embodiment of this invention, triggering the multi-agent deep reinforcement learning (MADRL) framework to generate the optimal collaborative control strategy includes:

[0069] S41: Privacy-preserving feature extraction of electric power EMS and natural gas SCADA data via federated learning;

[0070] Specifically, real-time access to power EMS and natural gas SCADA data is required. These data belong to different systems, posing data privacy issues and potentially risky when directly shared. This embodiment uses federated learning to extract useful features while protecting data privacy, thereby preventing data privacy leaks. To optimize resource allocation through hierarchical response, this embodiment implements a three-level early warning mechanism. The trigger conditions are: when ΔF > 15% or α > 1.2, a three-level warning (yellow / orange / red) is triggered. This mechanism captures system anomalies in real time through key indicators, and the three-level warning can allocate response resources based on risk level, avoiding over-regulation in low-risk scenarios.

[0071] S42: Based on the extracted features, the MADRL framework is used to define the action space of the power-side and gas-side agents, and a cross-system collaborative control strategy is generated with the minimization of comprehensive cost as the objective function;

[0072] To address the challenges of dynamic coupling of multiple systems, this embodiment achieves global optimization through multi-agent collaboration. It adopts the multi-agent deep reinforcement learning (MADRL) framework to define the action space of the power-side agent, including gas unit output adjustment and demand response priority, and the action space of the natural gas-side agent, including compressor flow regulation and gas storage supply rate.

[0073] S43: Use Modelica and OPAL-RT to build a digital twin platform, simulate and rehearse cross-system collaborative control strategies, and generate the optimal collaborative control strategy.

[0074] Specifically, directly deploying unverified strategies may lead to secondary risks, such as gas unit overload and pipeline pressure collapse. To meet the strategy safety verification requirements, this embodiment verifies the feasibility and effectiveness of the strategy through high-precision simulation, establishes a multi-physics field model of the gas-electricity combined system, covering the power network, natural gas network and coupling equipment, imports the Modelica model into OPAL-RT, and implements hardware-in-the-loop (HIL) simulation. Specifically, the control strategy generated by MADRL is input, the system status after simulation execution is simulated, and the simulation risk level is output.

[0075] Based on the above embodiment, a simulation preview of the cross-system collaborative control strategy is performed to generate the optimal collaborative control strategy, including:

[0076] Set target risk level thresholds and conduct simulation previews of cross-system collaborative control strategies;

[0077] If the risk level in the simulation results drops to the target risk level threshold, the output cross-system collaborative control strategy is the optimal collaborative control strategy;

[0078] If the risk level in the simulation results does not drop to the target risk level threshold, the cross-system collaborative control strategy is optimized through the Stackelberg game model to generate the optimal collaborative control strategy.

[0079] On the basis of the above embodiment, the optimization goal is clarified, and the target risk level threshold is set to provide a quantitative standard for strategy optimization to avoid blind adjustment. The target threshold is used to judge whether the risk level meets the standard. If the risk level after simulation drops to the target threshold, the current strategy is accepted as the optimal one. If it does not meet the standard, a secondary optimization is performed through the Stackelberg game model, with the power grid company as the leader and the gas grid operator as the follower, to reallocate the control resources and generate a new optimal strategy.

[0080] Furthermore, the accuracy error of the gas-electric coupling model of the digital twin platform is ≤2%, and the hardware-in-the-loop simulation step is ≤1ms.

[0081] Furthermore, taking minimizing the comprehensive cost as the objective function, the expression is:

[0082] min(C elec +λC gas +μLoadShed);

[0083] Among them, C elec is the operating cost of the power system, C gas is the operating cost of the natural gas system, LoadShed is the load reduction, and λ and μ are weight coefficients.

[0084] Gas-electricity combined system safety risk assessment and emergency coordinated disposal system, such as Figure 3 As shown, the system includes:

[0085] The indicator system construction module builds a multi-level risk assessment indicator system, which includes a static indicator layer and a dynamic indicator layer;

[0086] The risk level calculation module is based on a multi-level risk assessment indicator system and uses an improved entropy weight-TOPSIS algorithm to integrate static and dynamic indicators to calculate the comprehensive risk level.

[0087] The risk level judgment module presets the risk level threshold of the gas-electricity combined system safety and judges whether the comprehensive risk level reaches the risk level threshold. If not, the gas-electricity combined system operates normally.

[0088] The control strategy optimization module, if achieved, triggers the multi-agent deep reinforcement learning MADRL framework to generate the optimal collaborative control strategy;

[0089] The joint system coordination module coordinates and handles emergencies of the gas-electricity combined system based on the optimal coordinated control strategy.

[0090] The above-mentioned adjustment system in the present invention can effectively realize the safety risk assessment and emergency coordinated disposal method of the gas-electricity combined system. The technical effects that can be achieved are as described in the above-mentioned embodiments and will not be repeated here.

[0091] Furthermore, the control strategy optimization module includes:

[0092] Feature extraction unit, which performs privacy-preserving feature extraction of power EMS and natural gas SCADA data through federated learning;

[0093] The strategy generation unit, based on the extracted features, uses the MADRL framework to define the action space of the power-side and natural gas-side agents, and generates a cross-system collaborative control strategy with the objective function of minimizing the comprehensive cost;

[0094] The strategy optimization unit uses Modelica and OPAL-RT to build a digital twin platform, simulate and rehearse cross-system collaborative control strategies, and generate the optimal collaborative control strategy.

[0095] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the corresponding optimization effects of the method in Example 1, which will not be repeated here.

[0096] A computer-readable storage medium stores a computer program, which implements any one of the methods of claims 1 to 8 when the program is executed.

[0097] Example 2: A regional gas-electricity combined system (including 12 gas-fired power plants and 8 compressor units).

[0098] Data collection and preprocessing were performed, with real-time data collected for grid frequency (50 ± 0.2 Hz), node voltage (220 kV ± 5%), and natural gas pipeline pressure (4.0 MPa ± 0.3). Dynamic time warping (DTW) was used to align the time series data of the power and natural gas systems to eliminate sampling frequency discrepancies. A risk assessment was conducted, calculating static indicators: topological coupling CT = 0.32, gas turbine N-1 margin MN-1 = 0.25. Dynamic indicators were calculated: at t = 10:15, ΔF = 18% and α = 1.35 were detected, triggering a red alert. Emergency coordinated response was implemented, and a MADRL strategy was generated: reducing the output of Clusters 1-3 gas generators by 15% and activating the emergency gas supply rate from the gas storage facility to 120%. This strategy was verified through a digital twin rehearsal: the risk level was reduced from emergency to medium, and load losses were reduced by 62%. The strategy was validated, and cross-system coordinated adjustments were implemented within the deployed gas-power integrated system. The traditional method has a response time of 22 minutes and a load loss of 28%. The present invention has a response time of 4 minutes and 50 seconds and a load loss of 10.5%. The system recovery speed is increased by 40%, and it can be embedded in the existing energy management system (EMS) or natural gas SCADA system without the need for additional hardware investment.

[0099] Although the present application has been described with reference to specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and drawings are merely illustrative of the present application as defined herein and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. It is apparent that various modifications and variations of the present application may be made by those skilled in the art without departing from the scope of the present application. Thus, the present application is intended to include such modifications and variations as fall within the scope of the present application and its equivalents.

Claims

1. A method for safety risk assessment and emergency coordinated disposal of a gas-electricity combined system, characterized in that: The method comprises: constructing a multi-level risk assessment indicator system, wherein the multi-level risk assessment indicator system comprises a static indicator layer and a dynamic indicator layer; Based on the multi-level risk assessment indicator system, the static and dynamic indicators are integrated through the improved entropy weight-TOPSIS algorithm to calculate the comprehensive risk level; Presetting a risk level threshold for the safety of the gas-electricity combined system, determining whether the comprehensive risk level reaches the risk level threshold, and if not, operating the gas-electricity combined system normally; If it is reached, the multi-agent deep reinforcement learning MADRL framework is triggered to generate the optimal collaborative control strategy; The gas-electricity combined system is coordinated and handled in an emergency according to the optimal coordinated control strategy.

2. The gas-electricity combined system safety risk assessment and emergency coordinated disposal method according to claim 1 is characterized in that: The static indicator layer includes network topology coupling and device N-1 margin, and the dynamic indicator layer includes power flow matching deviation rate and fault propagation delay coefficient.

3. The gas-electricity combined system safety risk assessment and emergency coordinated disposal method according to claim 2, characterized in that: The fault propagation delay coefficient is calculated using a directed graph model and the Djkstra algorithm.

4. The gas-electricity combined system safety risk assessment and emergency coordinated disposal method according to claim 1, characterized in that: The improved entropy weight-TOPSIS algorithm integrates static indicators and dynamic indicators, including dynamically allocating weights through the information entropy of risk assessment indicators. The weight allocation formula is: Among them, H j is the information entropy indicator.

5. The gas-electricity combined system safety risk assessment and emergency coordinated disposal method according to claim 1, characterized in that: The triggering of the multi-agent deep reinforcement learning (MADRL) framework to generate the optimal collaborative control strategy includes: Privacy-preserving feature extraction of electric power EMS and natural gas SCADA data through federated learning; Based on the extracted features, the MADRL framework is used to define the action space of the power-side and natural gas-side agents, and a cross-system collaborative control strategy is generated with the minimization of comprehensive cost as the objective function. A digital twin platform was constructed using Modelica and OPAL-RT to simulate and rehearse the cross-system collaborative control strategy and generate the optimal collaborative control strategy.

6. The gas-electricity combined system safety risk assessment and emergency coordinated disposal method according to claim 5, characterized in that: The cross-system collaborative control strategy is simulated and rehearsed to generate the optimal collaborative control strategy, including: Setting target risk level thresholds and conducting simulation previews of the cross-system collaborative control strategy; If the risk level in the simulation result drops to the target risk level threshold, outputting the cross-system collaborative control strategy as the optimal collaborative control strategy; If the risk level in the simulation result does not drop to the target risk level threshold, the cross-system collaborative control strategy is optimized through the Stackelberg game model to generate an optimal collaborative control strategy.

7. The method for safety risk assessment and emergency coordinated disposal of a gas-electricity combined system according to claim 5, characterized in that: The gas-electric coupling model accuracy error of the digital twin platform is ≤2%, and the hardware-in-the-loop simulation step length is ≤1ms.

8. The gas-electricity combined system safety risk assessment and emergency coordinated disposal method according to claim 5, characterized in that: Taking minimizing the comprehensive cost as the objective function, the expression is: min(C elec +λC gas +μLoadShed); Among them, C elec is the operating cost of the power system, C gas is the operating cost of the natural gas system, LoadShed is the load reduction, and λ and μ are weight coefficients.

9. Gas-electricity combined system safety risk assessment and emergency coordinated disposal system, characterized by: The system includes: an indicator system construction module, which constructs a multi-level risk assessment indicator system, wherein the multi-level risk assessment indicator system includes a static indicator layer and a dynamic indicator layer; A risk level calculation module, based on the multi-level risk assessment indicator system, calculates the comprehensive risk level by fusing static indicators and dynamic indicators through an improved entropy weight-TOPSIS algorithm; a risk level judgment module, which presets a risk level threshold for the safety of the gas-electricity combined system and judges whether the comprehensive risk level reaches the risk level threshold; if not, the gas-electricity combined system operates normally; The control strategy optimization module, if achieved, triggers the multi-agent deep reinforcement learning MADRL framework to generate the optimal collaborative control strategy; The combined system coordination module performs emergency coordination and disposal on the combined gas-electricity system according to the optimal coordinated control strategy.

10. The gas-electricity combined system safety risk assessment and emergency coordinated disposal system according to claim 9, characterized in that: The control strategy optimization module includes: Feature extraction unit, which performs privacy-preserving feature extraction of power EMS and natural gas SCADA data through federated learning; The strategy generation unit, based on the extracted features, uses the MADRL framework to define the action space of the power-side and natural gas-side agents, and generates a cross-system collaborative control strategy with the objective function of minimizing the comprehensive cost; The strategy optimization unit uses Modelica and OPAL-RT to build a digital twin platform, simulates and rehearses the cross-system collaborative control strategy, and generates the optimal collaborative control strategy.

11. A computer-readable storage medium storing a computer program, characterized in that: When the program is executed, the method according to any one of claims 1 to 8 is implemented.

Images