A multi-objective optimization based integrated energy system collaborative scheduling method

Abstract

The application provides a kind of comprehensive energy system collaborative scheduling method based on multi-objective optimization, comprising: according to bottleneck channel set, the real-time demand and priority configuration of key load in comprehensive energy system are obtained, the load node set needing to be guaranteed in priority is determined, the energy type that can be allocated is obtained from multi-energy supply chain, and energy supply list is generated;Energy supply unit and load node are constructed as weighted directed graph, energy supply unit, energy storage node and load node are mapped as directed graph vertex, the weight of edge represents the remaining available capacity of transmission path, the output distribution of each power generation unit and energy storage unit is coordinated, and short-time resource scheduling scheme is generated;The output adjustment parameter of each energy supply unit of multi-energy supply chain and the flow control parameter of power transmission channel are obtained, control instruction is issued to energy management system, energy supply unit output and power transmission channel flow are dynamically adjusted, the stable operation state of comprehensive energy system under extreme weather impact is judged, and the final dynamic operation scheme is obtained.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for coordinated scheduling of integrated energy systems based on multi-objective optimization. Background Technology

[0002] Integrated energy systems play a crucial role in ensuring socio-economic development and meeting people's livelihood needs, and their dispatching methods directly impact the stability and reliability of energy supply. With the increasing frequency of extreme weather events, energy systems face unprecedented challenges, necessitating research into dispatching methods that enhance system resilience to cope with sudden shocks and reduce social losses. Currently, many dispatching methods rely primarily on static planning or single-energy optimization, making it difficult to adapt to the dynamic changes in multi-energy supply chains under extreme conditions. These methods often overlook the complexity of the overall system behavior after damage to critical infrastructure, resulting in the system's inability to quickly adjust to ensure the supply of critical loads when faced with sudden capacity reductions or transmission bottlenecks. Energy system shocks triggered by extreme weather events expose several core challenges. Sudden reductions in the available capacity of energy supply units make it difficult for the system to maintain its original energy supply plan, forcing dispatching strategies to reallocate resources in a short period to address the gap. This need for resource reallocation further exacerbates the bottleneck problem of cross-regional energy transmission channels, as limited channel capacity cannot promptly meet the surge in energy allocation demands. The existence of transmission bottlenecks directly affects the reconfiguration of energy supply priorities for critical loads, requiring the system to find a balance between ensuring the needs of important users and maintaining overall stability. This series of problems makes the trade-off between system robustness and resilience a core challenge in scheduling method design. Traditional static optimization methods struggle to achieve rapid response and effective recovery in dynamic environments. Therefore, how to dynamically adjust the trade-off strategy between robustness and resilience to cope with the cascading effects of declining energy supply capacity, exacerbated transmission bottlenecks, and the reprioritization of critical loads has become a key issue in the research of resilient scheduling methods for integrated energy systems. Summary of the Invention

[0003] This invention provides a comprehensive energy system coordinated scheduling method based on multi-objective optimization, mainly comprising:

[0004] The system obtains the real-time available capacity of each energy supply unit in the multi-energy supply chain, compares it with a preset normal capacity threshold, determines the capacity reduction of the energy supply unit, identifies the list of affected energy supply units, obtains the current flow and maximum capacity limit of its cross-regional energy transmission channels, calculates the remaining transmission capacity of each channel, determines the location of the bottleneck channel, and obtains the bottleneck channel set.

[0005] Based on the bottleneck channel set, obtain the real-time demand and priority configuration of key loads within the integrated energy system, determine the set of load nodes that need to be prioritized, obtain the types of energy that can be allocated from the multi-energy supply chain, and generate an energy supply list.

[0006] The power supply units and load nodes are constructed as a weighted directed graph, and the power supply units, energy storage nodes and load nodes are mapped as vertices of the directed graph. The weight of the edge represents the remaining available capacity of the transmission path. The output allocation of each power generation unit and energy storage unit is coordinated to generate a short-term resource scheduling scheme.

[0007] The adjustment requirements of energy transmission paths are extracted from short-term resource scheduling schemes, the redundancy capacity of transmission channels is iteratively optimized, the backup supply capacity of multi-energy supply chains is adjusted in combination with the allocation of redundant resources, Monte Carlo simulation is used to judge the stability of the scheduling scheme under random failures, and a preliminary scheduling scheme data table is generated based on the stability judgment results.

[0008] Acquire historical recovery data of integrated energy systems under extreme weather scenarios, analyze the risk distribution of multi-energy supply chains, dynamically adjust and optimize the energy allocation ratio of each load node, determine the adaptability of the scheduling scheme to maintain system stability, iteratively optimize the emergency response speed and magnitude of key nodes based on the adaptability judgment results, and form an emergency scheduling scheme that meets time constraints.

[0009] Based on the emergency dispatch plan that meets time constraints, obtain dynamic flow change data of cross-regional transmission channels in the multi-energy supply chain, update the distribution path of energy resources between supply and demand nodes, dynamically adjust the power supply priority of critical load nodes in combination with load importance, and dynamically reduce the power supply priority of non-critical load nodes based on their capacity requirements and load characteristics, and generate dynamic dispatch instruction sets for each time period.

[0010] The system acquires the output adjustment parameters of each energy supply unit in the multi-energy supply chain and the flow control parameters of the transmission channel, issues control commands to the energy management system, dynamically adjusts the output of the energy supply units and the flow of the transmission channel, judges the stable operating status of the integrated energy system under extreme weather impacts, and obtains the final dynamic operation plan.

[0011] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0012] This invention discloses a collaborative scheduling method for integrated energy systems based on multi-objective optimization. It generates short-term resource scheduling schemes through steps such as acquiring real-time capacity of energy supply units, identifying bottleneck channels, determining critical loads, and constructing a weighted directed graph. By analyzing risk distribution using historical recovery data, it optimizes energy allocation ratios and emergency response strategies, forming a scheduling scheme with fault recovery time window constraints. Furthermore, it updates resource allocation paths based on dynamic path optimization, adjusts load priorities, and generates a dynamic scheduling instruction set. Finally, it issues control commands through regional interconnection scheduling to dynamically adjust the output of energy supply units and the flow of transmission channels, achieving stable system operation under extreme weather impacts. This invention effectively addresses the impact of extreme weather on multi-energy supply chains, ensures the supply of critical loads, and improves the overall stability and resilience of the system. Attached Figure Description

[0013] Figure 1 This is a flowchart of a comprehensive energy system collaborative scheduling method based on multi-objective optimization according to the present invention. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0015] like Figure 1 This embodiment of a comprehensive energy system collaborative scheduling method based on multi-objective optimization may specifically include:

[0016] Step S101: Obtain the real-time available capacity of each energy supply unit in the multi-energy supply chain, compare it with a preset normal capacity threshold, determine the capacity reduction of the energy supply unit, determine the list of affected energy supply units, obtain the current flow and maximum capacity limit of its cross-regional energy transmission channels, calculate the remaining transmission capacity of each channel, determine the location of the bottleneck channel, and obtain the bottleneck channel set.

[0017] Obtain the operating capacity value corresponding to the power supply unit identifier, and calculate the capacity availability rate based on the operating capacity value and a preset normal capacity threshold, storing it in the capacity status data table; for the power supply unit identifier in the capacity status data table, obtain the corresponding channel identifier, calculate the actual flow rate of the channel using channel monitoring parameters, and if the actual flow rate of the channel divided by the rated flow rate of the channel is greater than a preset bottleneck judgment threshold, record it in the bottleneck channel data table; for the channel identifier in the bottleneck channel data table, obtain the channel operating data from the historical database, and establish a channel flow prediction model using a linear regression method to obtain the predicted flow rate; generate a bottleneck channel set based on the predicted flow rate and the rated flow rate of the channel.

[0018] For example, based on the energy supply unit identification number, the current operating capacity value of each energy supply unit is obtained from the real-time database. If the actual capacity value of an energy supply unit is less than 70% of the preset normal capacity threshold, the energy supply unit identification number and the current capacity value are recorded in the affected energy supply unit data table. The capacity availability rate is calculated by dividing the current capacity value by the preset normal capacity value and stored in the capacity status data table. For the energy supply unit identification number in the capacity status data table, the corresponding cross-regional energy transmission channel identification number and channel rated flow value are obtained. Real-time operating parameters such as current, voltage, and power factor are read from the channel monitoring database to calculate the current actual flow value. The remaining available transmission capacity value is calculated based on the channel rated flow value and the current actual flow value. The channel load rate is calculated by dividing the current actual flow value by the channel rated flow value. If the channel load rate is greater than the preset bottleneck judgment threshold of 0.8, the channel identification number is added to the bottleneck channel data table, and the identification numbers of the energy supply units connected upstream and downstream of the channel are obtained and recorded in the channel influence chain data table. For each channel in the bottleneck channel data table, historical 24-hour channel operation data is retrieved from the database. Average flow rates are calculated at 15-minute intervals, and a linear regression model is used to establish a channel flow prediction model. Based on this model, the predicted flow rate for the next four hours is calculated. If the predicted flow rate exceeds 90% of the channel's rated flow rate, the remaining transmission capacity of the channel is updated, and the channel identifier and predicted flow rate are stored in the channel early warning data table. For each channel identifier in the early warning data table, the associated power supply unit identifier is retrieved from the channel impact chain data table. The capacity availability of the corresponding power supply unit is queried from the capacity status data table, and a weighted average method is used to calculate the channel impact level. During extreme weather events, the capacity of power supply units is significantly affected. For example, heavy rain can cause a sharp drop in photovoltaic power generation. A photovoltaic power station with an installed capacity of 100 MW experiences a real-time power generation drop to 30 MW due to continuous heavy rain, resulting in a capacity availability of only 0.3, far below the preset normal capacity threshold of 0.7. The system automatically records this photovoltaic power station information in the affected power supply unit data table. Meanwhile, the operational status of inter-regional energy transmission channels is also facing severe challenges. Taking a 500 kV ultra-high voltage transmission line as an example, the line has a rated transmission capacity of 2000 MW. By collecting real-time data from current transformers and voltage transformers, it is calculated that the current actual transmission power reaches 1800 MW, with only 200 MW of remaining available transmission capacity. The channel load rate is as high as 0.9, exceeding the bottleneck judgment threshold of 0.8, and the transmission channel is identified as a bottleneck channel. After identifying the bottleneck channel, the system automatically traces the upstream and downstream energy supply units of the transmission channel. Upstream, it connects 3 hydropower stations and 2 photovoltaic power stations; downstream, it connects 2 energy storage power stations and 3 large power loads. The operational status of these related units is constrained by the bottleneck channel, and the system records these relationships in the channel influence chain data table.By analyzing the historical operating data of the transmission channel over the past 24 hours and calculating the average transmission power at a 15-minute time granularity, a historical peak of 1900 MW was found between 13:00 and 15:00. After establishing a prediction model through linear regression, the maximum transmission power is predicted to reach 1850 MW in the next 4 hours, exceeding 90% of the rated capacity. The system automatically adds this channel information to the early warning data table. Considering the capacity status of upstream and downstream energy supply units, the availability rates of the three upstream hydropower stations are 0.85, 0.82, and 0.78, respectively, and the availability rates of the two photovoltaic power stations are 0.3 and 0.35, respectively. Taking into account the importance of different types of energy supply units, the weight of hydropower stations is taken as 0.6, and the weight of photovoltaic power stations is taken as 0.4. The overall impact of the channel is calculated to be 0.67, indicating that the restricted state of this transmission channel has a significant impact on regional energy supply. This multi-dimensional analysis method not only enables the timely detection of abnormal states in power supply units and transmission channels, but also assesses the propagation range and impact of these abnormal states, providing crucial information for subsequent dispatch decisions and thus improving the safe and stable operation of the power system under extreme weather conditions. Specifically, in this case, the system identified issues such as a sudden drop in photovoltaic power plant capacity and excessive load on transmission channels, both of which require timely intervention from the dispatching department.

[0019] Step S102: Based on the bottleneck channel set, obtain the real-time demand and priority configuration of key loads within the integrated energy system, determine the set of load nodes that need to be prioritized, obtain the types of energy that can be allocated from the multi-energy supply chain, and generate an energy supply list.

[0020] The load node identifier is obtained from the channel identifier in the bottleneck channel set. The load nodes are sorted in descending order using priority values. For load nodes with the same priority value, they are sorted in descending order according to their real-time energy demand to obtain a priority load data table. For the load nodes in the priority load data table, the real-time total demand is calculated according to the energy type (electricity, natural gas, heat). The category supply and demand matching rate is calculated by dividing the total dispatchable capacity of the same type of energy supply unit by the total load demand. The energy supply unit identifier is obtained from the energy type identifier that is greater than the preset supply and demand matching threshold. The power generation type identifier of the energy supply unit is read to determine whether the power generation type identifier belongs to the green energy type code set to obtain a green energy candidate data table. For the energy supply units in the green energy candidate data table, the environmental score is calculated using the unit power generation efficiency index and pollutant emission index. The energy supply unit identifier is selected in descending order according to the environmental score. The remaining available capacity of the transmission channel is read to determine whether the real-time energy demand of the load node is met, and a power supply scheme identifier is generated to obtain an energy supply list.

[0021] For example, downstream load node identifiers are obtained based on channel identifiers in the bottleneck channel set. Real-time energy demand values ​​and preset priority values ​​of load nodes are read from the load database. Load nodes are sorted in descending order using priority values. For load nodes with the same priority value, they are sorted in descending order based on their real-time energy demand values. Load nodes with priority values ​​higher than the preset load priority threshold are selected and written into the priority guaranteed load data table. For load nodes in the priority guaranteed load data table, energy types are categorized as electricity, natural gas, and heat. The total real-time demand for each type of energy is calculated. The real-time dispatchable capacity of each energy supply unit corresponding to its energy type is obtained from the energy supply database. The categorized supply-demand matching rate is calculated by dividing the sum of dispatchable capacities of the same type of energy supply units by the total load demand and stored in the supply-demand matching data table. Energy type identifiers with categorized supply-demand matching rates greater than the preset supply-demand matching threshold are filtered from the supply-demand matching data table. All energy supply unit identifiers under this type are obtained. The power generation type identifier of the energy supply unit is read. If the power generation type identifier belongs to the green energy type code set, the energy supply unit is written into the green energy candidate data table. For each energy supply unit in the green energy candidate data table, the unit's power generation efficiency and pollutant emission indicators are read. An environmental score is calculated using a weighted summation method. After sorting by environmental score in descending order, the top-ranked energy supply units are written into the green energy supply data table. The corresponding transmission channel identifier is obtained based on the energy supply unit identifier in the green energy supply data table. The remaining available capacity of the transmission channel is read. If the remaining available capacity is greater than the real-time energy demand of the load node, a unique power supply scheme identifier is generated. The energy supply unit identifier, transmission channel identifier, load node identifier, and power supply capacity are written into the power supply scheme data table. For each record in the power supply scheme data table, the operating parameters of the energy supply unit and the transmission channel are read. The power supply reliability index of the power supply scheme is calculated. If the power supply reliability index meets a preset threshold, the status of the record is updated to "activated." In integrated energy systems, the priority configuration of load nodes directly affects power supply reliability. Taking an industrial park as an example, the park contains various energy-consuming units such as a medical center, a data center, and general industrial loads. The priority of the medical center is set to 5, the priority of the data center to 4, and the priority of general industrial loads to 2. When the load priority preset threshold is set to 3, the medical center and the data center are identified as priority loads. When calculating the energy demand of priority loads, it is necessary to distinguish between different energy categories. In addition to a 10 MW power demand, the medical center also has a 5 MW heating demand, while the data center mainly has an 8 MW power load. Through summarization and calculation, the total power demand is 18 MW, and the total heating demand is 5 MW.Meanwhile, the park possesses various energy supply facilities, including photovoltaic power stations, gas-fired combined heat and power (CHP) units, and gas-fired boilers. The photovoltaic power station currently has a dispatchable capacity of 12 MW, the CHP unit has a dispatchable power capacity of 10 MW, and the heating capacity has a dispatchable capacity of 8 MW. The calculated power supply-demand matching rate is 1.22, and the heating supply-demand matching rate is 1.6. When selecting green energy supply solutions, the energy supply units within the park were categorized into solar and gas-fired energy based on power generation type. The photovoltaic power station, identified as solar power, falls under the green energy category and was included as a candidate solution. Further evaluation revealed that the photovoltaic power station achieves a power generation efficiency of 20%, with virtually no pollutant emissions, ranking among the top in environmental protection scores, and was ultimately selected for the green energy supply list. Analysis of the transmission line status revealed that the remaining available capacity of the 10 kV distribution line connecting the photovoltaic power station and the medical center is 15 MW, exceeding the medical center's 10 MW power demand. Furthermore, this line is a direct connection, with a short power supply distance, high line stability, and a power supply reliability index of 0.99, meeting the preset threshold requirement of 0.95. In actual operation, the complementary nature of energy types must be considered when determining the energy supply plan for the park. Although photovoltaic power generation is green and environmentally friendly, it is greatly affected by weather. By complementing it with gas-fired combined heat and power (CHP) units, both power supply reliability and full coverage of heating demand are ensured. The 5 MW heating load of the medical center can be stably supplied by the gas-fired CHP units. This energy supply plan makes full use of the existing green energy resources in the park and meets the energy needs of different types of loads through multi-energy complementarity, reflecting the characteristics of green energy priority and multi-energy synergy in energy supply.

[0022] Step S103: Construct a weighted directed graph of energy supply units and load nodes, map energy supply units, energy storage nodes and load nodes as vertices of the directed graph, the weight of the edge represents the remaining available capacity of the transmission path, coordinate the output allocation of each power generation unit and energy storage unit, and generate a short-term resource scheduling scheme.

[0023] A vertex data table is generated based on the energy supply unit identifier and the energy storage node identifier. The vertex data table constructs a power grid topology map using the remaining available capacity values ​​of the transmission channels. Based on the power grid topology map, the real-time power generation output values ​​of the energy supply units and the charging and discharging power values ​​of the energy storage nodes are obtained. The maximum flow algorithm is used to calculate the path set between the energy supply units and the load nodes. For the path set, the channel tripping rate and fault recovery time are read, and a weighted summation method is used to obtain the path reliability score. The path identifier with the highest path reliability score is selected. For the path identifier, the power output values ​​of the energy supply units and the charging and discharging values ​​of the energy storage nodes are obtained. The minimum deviation method is used to calculate the node power allocation scheme. If the power allocation scheme meets the node power balance constraint condition, the scheme status is updated to "activated", and a short-term resource scheduling scheme is obtained.

[0024] For example, based on the priority load node set and energy supply list, the vertex data table is generated by reading the energy supply unit identifier, energy storage node identifier, and load node identifier. The remaining available capacity of the transmission channel is obtained from the transmission channel database as the edge weight value. The power grid topology is constructed by using zero values ​​to represent unconnected nodes and actual capacity to represent connected nodes. For the energy supply unit identifier, the real-time power generation output value and output adjustment range value are read. The current state of charge, maximum charging and discharging power value, and remaining available capacity value of the energy storage node are obtained. Based on the remaining available capacity value of the transmission channels between nodes, the maximum flow algorithm is used to calculate the set of all feasible paths between the energy supply unit and the load node. The path identifier and the corresponding maximum transmission capacity value are written into the path data table. The path identifier is read from the path data table, and the channel identifier contained in the path is obtained. The channel tripping rate, fault recovery time, maintenance interval and other operating indicators are read. The channel commissioning time, operating hours and number of faults and other life indicators are read. The path reliability score is calculated using a weighted summation method. The path identifiers are sorted in descending order according to the score value. The paths with the scores in the top 20% are written into the preferred path data table. For each path identifier in the preferred path data table, the rated and real-time output values ​​of the starting power supply unit are obtained. The remaining charging / discharging time and power values ​​of the energy storage nodes along the route are read. Based on the remaining available capacity of the channel, the minimum deviation method is used to calculate the power allocation values ​​for each node in the next four hours. The path identifier, node identifier, and power allocation values ​​are then written into the initial scheduling scheme data table. Power allocation values ​​are read from the initial scheduling scheme data table, and constraints on the ramp rate of the power supply unit and the rate of change of the charging / discharging power of the energy storage nodes are obtained. The dynamic feasibility of the power allocation scheme is verified using a time series verification method. If the constraints are met, the scheme is written into the short-term resource scheduling scheme. For each record in the short-term resource scheduling scheme, the node identifier and power allocation value are read. The node connection relationship is obtained from the power grid topology diagram. It is verified that the difference between the node power injection and power outflow is less than a preset balance threshold. The status flag of the scheme that meets the power balance constraints is updated to "activated". In the process of power system dispatching, constructing a power grid topology map is the foundation for achieving optimal resource allocation. Taking a regional power grid as an example, the region contains 2 thermal power plants, 1 photovoltaic power station, 3 energy storage power stations and 5 important load nodes. By reading the identification information of each node, a topology map with 11 vertices is constructed. The remaining available capacity of the transmission channels between nodes is used as the weight value of the edge. The edge weight of connected nodes is the actual capacity value, and the edge weight of non-connected nodes is recorded as 0, thus forming an 11th-order adjacency matrix.When acquiring node operating parameters, the current output of thermal power unit 1 is 200 MW, with an adjustment range of 120 to 300 MW; the output of unit 2 is 180 MW, with an adjustment range of 100 to 280 MW; and the real-time output of the photovoltaic power station is 80 MW. Energy storage power station 1 currently has a state of charge of 75%, a maximum charge / discharge power of 50 MW, and a remaining usable capacity of 200 MW. Using the maximum flow algorithm, 12 feasible transmission paths from the power supply to the load node are calculated. When evaluating the reliability of the transmission paths, multiple operating indicators need to be considered comprehensively. A certain transmission path includes three transmission lines. Line 1 has an annual trip rate of 0.8 times / year, an average fault recovery time of 2 hours, and a maintenance interval of 180 days. Line 2 has been in operation for 5 years, accumulating 40,000 hours of operation, and has experienced two serious faults. Line 3 is in good operating condition, with all indicators exceeding the average level. Through weighted calculation, the reliability score of this path is 0.85, ranking in the top 20% of all paths. When allocating power based on the preferred path, considering the regulation characteristics of thermal power units, the output of Unit 1 is gradually increased from 200 MW to 250 MW, with the ramp rate controlled within 2 MW per minute. Energy storage power station 1 is planned to continuously discharge over the next 4 hours, with its power gradually ramping up from 0 to 30 MW, fully utilizing its remaining available capacity while ensuring that the discharge power change rate does not exceed 5 MW per minute. In verifying the power allocation scheme, taking a transmission node as an example, this node connects one generator unit and two transmission lines. The generator unit injects 200 MW of power, and the two transmission lines outflow 120 MW and 78 MW respectively. The difference between the node's power injection and outflow is 2 MW, less than the preset balance threshold of 5 MW, thus satisfying the node's power balance constraint. In this way, the power balance of all nodes is verified sequentially, and finally, a feasible scheduling scheme that meets all constraints is selected.

[0025] Step S104: Extract the adjustment requirements of energy transmission paths from the short-term resource scheduling scheme, iteratively optimize the redundancy capacity of the transmission channels, adjust the backup supply capacity of the multi-energy supply chain in combination with the allocation of redundant resources, use Monte Carlo simulation to judge the stability of the scheduling scheme under random failures, and generate a preliminary scheduling scheme data table based on the stability judgment results.

[0026] The remaining available capacity of the transmission channel is read based on the path identifier. The channel reliability index is calculated by weighting the fault frequency and fault recovery time. For the energy supply unit identifier and energy storage node identifier connected at both ends of the channel, the rated capacity and current output value of the energy supply unit are obtained, and the rated capacity and current state of charge of the energy storage node are obtained to calculate the remaining available capacity value of the node. The remaining available capacity value of the node is classified and statistically analyzed using a hierarchical recursive method to obtain the total reserve capacity value of the three energy types: electricity, natural gas, and heat. If the fault duration exceeds a preset threshold, a random fault scenario is generated using the Monte Carlo method. The node power adjustment scheme is calculated based on the total reserve capacity value and the channel reliability index. The node stability coefficient is determined by the difference between the power injection and outflow. The power allocation value and node stability coefficient in the short-term resource scheduling scheme are written into the preliminary scheduling scheme data table.

[0027] For example, based on the path identifier and power allocation value in the short-term resource scheduling scheme, the remaining available capacity of the transmission channel on each path is read. The frequency of fault occurrence and fault recovery time of each channel are obtained from the historical database. The channel reliability index is calculated according to the weighted sum of fault frequency and recovery time. The channel identifier and reliability index are written into the channel evaluation data table. For the channel identifier in the channel evaluation data table, the identifiers of the energy supply units and energy storage nodes connected at both ends of the channel are obtained. The rated capacity and current output value of the energy supply unit, and the rated capacity and current state of charge of the energy storage node are read. The remaining available capacity value of each node is calculated. The node identifier and remaining available capacity value are written into the reserve resource data table. The remaining available capacity value of the nodes in the reserve resource data table is read. The energy supply units are classified according to electricity, natural gas, and heat. The capacity regulation scheme of each type of node is calculated using a hierarchical recursive method. The total reserve capacity value of each energy type is calculated. The energy type identifier and total reserve capacity value are written into the regulation scheme data table. Fault records occurring within the past year and lasting longer than a preset threshold are selected from the historical fault database. Fault location and impact range parameters are extracted, and a Monte Carlo method is used to generate 100 random fault scenarios, which are then written into a fault scenario data table. For each scenario in the fault scenario data table, the total reserve capacity of each energy type is read from the regulation scheme data table. Combined with the channel reliability index from the channel evaluation data table, the node power regulation scheme is calculated, and the difference between power injection and outflow is verified to be less than a preset balance threshold. Based on the power balance verification results under the fault scenarios, the fault adaptation times of each node are counted. The total reserve capacity value is read from the regulation scheme data table, and the node stability coefficient is calculated by dividing the fault adaptation times by the total number of scenarios. The power allocation values ​​and node stability coefficients in the short-term resource scheduling scheme are written into the preliminary scheduling scheme data table. In power system scheduling, channel reliability assessment is a crucial step in ensuring the safe and stable operation of the power grid.

[0028] For example, in a certain regional power grid, the transmission channel numbered T105 connects the No. 1 thermal power plant and the No. 3 load center. By querying the historical database, it was found that the channel had experienced 12 failures in the past three years, with an average recovery time of 1.8 hours per failure. The channel reliability index was calculated to be 0.82 using the weighted method of "0.7 × failure frequency + 0.3 × recovery time", and was written into the channel evaluation data table.

[0029] In one possible implementation, the standby resource assessment process examines the remaining adjustment capacity of each node.

[0030] For example, the No. 1 thermal power plant connected to both ends of the T105 channel has a rated capacity of 300 MW, a current output of 220 MW, and a remaining available capacity of 80 MW; the No. 2 energy storage power station connected to it has a rated capacity of 100 MW, a current state of charge of 65%, and a maximum discharge power of 50 MW. These data are recorded in the standby resource data table to provide a basis for subsequent capacity regulation scheme calculations.

[0031] It should be noted that the hierarchical recursive method has a significant advantage in calculating capacity adjustment schemes.

[0032] Specifically, the system first categorizes energy supply units into three types: electricity (thermal power, photovoltaic), natural gas (gas turbine), and heat (combined heat and power). Then, it recursively calculates regulation schemes for each type. For example, the total reserve capacity for electricity nodes is 180 MW, for natural gas nodes it is 120 MW, and for heat nodes it is 90 MW. These values ​​are written into a regulation scheme data table to address potential failure scenarios. In failure scenario simulations, the Monte Carlo method effectively assesses the system's resilience.

[0033] For example, 15 fault records with durations exceeding 2 hours within the past year were selected from the historical fault database. After extracting the fault location and impact range parameters, 100 sets of random fault scenarios were generated. In one scenario, assuming a complete interruption of channel T105, the system needed to utilize the 80 MW reserve capacity of Power Plant No. 1 and the 40 MW discharge capacity of Energy Storage Station No. 2. Verification showed that the power balance deviation of each node under this scheme was less than the preset threshold of 3 MW. By statistically analyzing the fault adaptability of each node in the 100 scenarios, Power Plant No. 1 was found to be able to effectively cope in 92 scenarios, with a stability coefficient of 0.92. This value, along with the power allocation scheme, was written into the preliminary dispatch scheme data table, providing reliable assurance for system operation.

[0034] Step S105: Obtain historical recovery data of the integrated energy system under extreme weather scenarios, analyze the risk distribution of the multi-energy supply chain, dynamically adjust and optimize the energy allocation ratio of each load node, determine the adaptability of the scheduling scheme to maintain system stability, iteratively optimize the emergency response speed and magnitude of key nodes based on the adaptability judgment results, and form an emergency scheduling scheme that meets time constraints.

[0035] Based on the node identifier, the database retrieves extreme weather occurrence times and node failure records from the historical database. Fault response curves and fault recovery curves are obtained through least squares fitting, and the average response time and average recovery time are written into the fault feature data table. For the node identifier in the fault feature data table, the real-time inventory and rated inventory capacity of the energy supply unit are retrieved. A Bayesian network is used to calculate the probability of a node experiencing an energy supply interruption, and the node risk level is written into the risk assessment data table. If the real-time inventory of a node is lower than the preset safety inventory, the inventory replenishment priority is set according to the risk level, and a multi-energy collaborative replenishment scheme is used to obtain the replenishment scheme in the inventory adjustment data table. For the replenishment scheme in the inventory adjustment data table, the output adjustment rate of the energy supply unit and the charging and discharging rate of the energy storage unit are retrieved. A piecewise linear programming method is used to optimize the node response rate and adjustment range to obtain an emergency dispatch scheme that meets time constraints.

[0036] For example, based on the key node identifiers in the preliminary scheduling plan, extreme weather occurrence times and node failure records are read from the historical database. Three timestamps are extracted: the time of failure occurrence, the start time of failure handling, and the time of failure repair completion. The least squares method is used to fit the failure response curve and failure recovery curve, and the average response time and average recovery time are calculated. The node identifiers and time characteristic parameters are written into the failure characteristic data table. For the node identifiers in the failure characteristic data table, the real-time inventory and rated inventory capacity of the energy supply unit are obtained. The energy demand curve and historical fluctuation curve of the load node are read. A Bayesian network is used to calculate the probability of a node experiencing an energy supply interruption. The risk level is classified based on the scope of the failure's impact and the difficulty of recovery. The node identifiers and risk levels are written into the risk assessment data table. The node risk level is read from the risk assessment data table, and the real-time inventory and preset safety inventory of the node are obtained. If the real-time inventory is lower than the preset safety inventory, the inventory replenishment priority is set based on the risk level. The minimum replenishment amounts for electricity, natural gas, and heat are calculated, and a multi-energy coordinated replenishment plan is generated and written into the inventory adjustment data table. For the replenishment plans in the inventory adjustment data table, the output adjustment rate of the energy supply unit and the charging and discharging rate of the energy storage unit are read. Based on the transmission speed and conversion efficiency of each type of energy, the execution cycle of the replenishment plan is calculated. The execution cycle value is compared with the average recovery time of the nodes to determine the time feasibility of the replenishment plan. Based on the time feasibility judgment result of the replenishment plan, the dynamic response characteristics of the energy supply unit and the energy storage unit are read. Piecewise linear programming method is used to optimize the node response rate and adjustment range, and an emergency dispatch plan that meets the time constraints is generated and written into the response optimization data table. For the dispatch plan in the response optimization data table, the supply and demand balance constraints of each energy category are read, the node inventory change curve is verified to meet the safety stock constraint, the stability margin of the plan is calculated, and an emergency dispatch plan that meets multiple constraints is obtained. In extreme weather events, the timing characteristics of fault response and recovery processes have a significant impact on system stability. Taking a large industrial park as an example, when a strong typhoon struck, a 110 kV substation in the park experienced a fault trip at 10:30 AM on August 15th. Maintenance personnel arrived on-site at 10:45 AM to begin handling the situation and finally completed fault repair at 1:30 PM. By fitting the fault response curve, the average response time of this node was found to be 15 minutes, and the average recovery time was 165 minutes. When analyzing the node's risk characteristics, this substation, as the park's main power supply facility, bears the power load of four chemical plants. Real-time operation data showed that the substation's current load rate was 85%, and historical load fluctuations ranged from 75% to 90%. Considering the extremely high requirements for power supply reliability in chemical production, and combining the substation's equipment status and historical fault data, a power outage probability of 0.15 was calculated using a Bayesian network, classifying it as the highest risk level.Based on the risk assessment results, conduct an inventory analysis of the energy storage facilities supporting the substation. The rated capacity of the energy storage power station is 100 megawatt-hours, and the current remaining power is 30 megawatt-hours, which is lower than the safety inventory requirement of 50 megawatt-hours. Considering the high risk level of the substation, set the energy storage replenishment task as the highest priority. The calculated required replenishment power is 30 megawatt-hours. At the same time, the natural gas inventory of the gas turbine in the park also needs to be replenished by 200 cubic meters. When planning the multi-energy collaborative replenishment plan, it is necessary to consider the transmission characteristics of different energy categories. The maximum transmission power of the 220 kV transmission line in the park is 300 megawatts, and the maximum charging power of the energy storage power station is 50 megawatts. It takes 36 minutes to complete the power replenishment of 30 megawatt-hours, and the conveying capacity of the natural gas pipeline network is 500 cubic meters per hour. It takes 24 minutes to replenish the gas turbine storage, both of which are less than the average fault recovery time of 165 minutes. Combining the dynamic characteristics of each energy supply facility, the charging power adjustment rate of the energy storage power station is 10 megawatts per minute, and the output ramp rate of the gas turbine is 5 megawatts per minute. By optimizing the charge-discharge strategy, at the initial stage of the fault, the energy storage power station supplies power to the load at a power of 50 megawatts, while receiving charging replenishment from the grid side. After the gas turbine is connected to the grid, gradually reduce the energy storage discharge power to achieve a smooth transition of power supply. This multi-energy collaborative scheduling plan based on risk assessment not only ensures that the inventory levels of the energy storage power station and the gas turbine are always higher than the safety threshold, but also realizes the uninterrupted power supply to the chemical load during the fault, reflecting the complementary advantages and collaborative effects of the integrated energy system.

[0037] Step S106, according to the emergency scheduling plan that meets the time constraint, obtain the dynamic flow change data of the cross-regional transmission channels in the multi-energy supply chain, update the distribution paths of energy resources between supply and demand nodes, and dynamically adjust the power supply priorities of key load nodes in combination with the importance of the load. For non-key load nodes, dynamically reduce the power supply priority based on their capacity requirements and load characteristics, and generate a dynamic scheduling instruction set for each period.

[0038] Read the real-time flow values and channel rated capacities of each channel from the transmission channel monitoring database, and use a deep neural network model to predict the channel flow prediction data for future periods; obtain the identification numbers of supply and demand nodes at both ends of the channel for the channel flow prediction data, and use the shortest path algorithm to calculate the transmission path data with the minimum loss between supply and demand nodes; obtain the time-sharing energy consumption data of load nodes according to the transmission path data, and use the fuzzy comprehensive evaluation method to calculate the energy consumption characteristic score data of load nodes; read the identification numbers of non-key load nodes for the energy consumption characteristic score data, and generate a load regulation plan data sorted from low to high according to the energy consumption characteristic scores, forming a dynamic scheduling instruction set for each period.

[0039] For example, based on an emergency dispatch plan that meets time constraints, the real-time flow values ​​and rated capacity of each transmission channel for the past 24 hours are read from the transmission channel monitoring database. The flow change rate is obtained by calculating the flow difference between adjacent time periods. A flow prediction model is constructed using a deep neural network. The input features include historical flow values, meteorological data, and load curves. The predicted channel flow values ​​for the next 4 hours are written into the flow prediction data table. For the channel identifier in the flow prediction data table, the identifiers of the supply and demand nodes connected at both ends of the channel are obtained. The transmission distance and line loss rate between the nodes are read. The transmission loss coefficient is calculated based on the transmission line impedance and power factor. The shortest path algorithm is used to calculate the transmission path combination between the supply and demand nodes. The top three paths with the lowest overall loss are selected and written into the path optimization data table. The path identifier and node identifier are read from the path optimization data table. The time-of-use energy consumption data of the load nodes is obtained. The peak-period energy consumption ratio and energy consumption fluctuation rate are calculated. The fuzzy comprehensive evaluation method is used to calculate the energy consumption characteristic score of the load nodes. Nodes with scores higher than a preset threshold are marked as critical loads and written into the load classification data table. For critical load nodes in the load classification data table, fault records for the past year are read to analyze the frequency and impact of faults. Combined with energy consumption characteristic scores, a comprehensive priority coefficient for each node is calculated, and the node identifier and priority coefficient are written into the priority data table. Based on the node priority coefficients in the priority data table, non-critical load node identifiers are read from the load classification data table. Non-critical loads are then sorted by energy consumption characteristic scores from low to high, and the capacity adjustment space for each load level is calculated. Load control schemes are generated and written into the load response data table. For the control schemes in the load response data table, combined with the transmission paths in the path optimization data table, time-segmented dispatch instructions containing path identifiers, start and end timestamps, load identifiers, and power supply capacity values ​​are generated. These dispatch instructions are recorded and written into the instruction set data table. In the dynamic scheduling of multi-energy supply chains, transmission channel flow forecasting is crucial for optimizing resource allocation. Taking a cross-regional transmission channel as an example, with a rated capacity of 500 MW, real-time flow data collected over the past 24 hours revealed flow fluctuations between 200 and 450 MW, with a maximum flow change rate of 50 MW per hour between adjacent hours. Combining local meteorological data such as temperature and solar irradiance, as well as downstream load power consumption curves, the peak flow of the channel is predicted to reach 420 MW within the next 4 hours. Based on the flow forecast results, the transmission loss of the transmission channel is assessed. The channel is 150 km long, with an average line impedance of 0.3 ohms per km and a power factor of 0.85. The line loss rate at a transmission power of 400 MW is approximately 3%. Using the shortest path algorithm, three optional transmission paths are calculated: path 1 has a comprehensive loss of 2.8%, path 2 has 3.2%, and path 3 has 3.5%. Path 1, with the lowest loss, is selected as the primary transmission channel.During the load node classification process, an industrial park had 15 energy-consuming loads. Analysis of their time-of-use energy consumption data revealed that the data center accounted for 65% of its daily energy consumption during peak power hours, with a consumption fluctuation rate of 15%. In contrast, the peak energy consumption of ordinary office buildings accounted for only 40%, with a consumption fluctuation rate of 35%. After fuzzy comprehensive evaluation, the data center's energy consumption characteristic score was 0.85, higher than the preset threshold of 0.75, and it was marked as a critical load. Further analysis revealed that the data center had experienced two power outages in the past year, affecting the normal operation of four downstream users in the park. Combined with its energy consumption characteristic score of 0.85, its comprehensive priority coefficient was calculated to be 0.92, ranking first among all loads. In contrast, the ordinary office buildings had an energy consumption characteristic score of 0.6, had not experienced any power outages in the past year, and had a comprehensive priority coefficient of only 0.55, classifying them as non-critical loads. When generating the load control plan, for the lower-priority office building load, peak shaving and valley filling measures were implemented during peak electricity consumption periods. The originally planned 100 kW load was adjusted to 80 kW, and the reduced power supply was supplemented by the energy storage system, thereby ensuring the data center's full-load power supply needs. The generated dispatch instructions clearly stipulated that during the peak period from 14:00 to 16:00, 350 kW of power would be continuously supplied to the data center through path 1, while the power supply to the office building would be controlled within 80 kW, thus ensuring the power supply guarantee for critical loads.

[0040] Step S107: Obtain the output adjustment parameters of each energy supply unit in the multi-energy supply chain and the flow control parameters of the transmission channel, issue control commands to the energy management system, dynamically adjust the output of the energy supply units and the flow of the transmission channel, determine the stable operating status of the integrated energy system under extreme weather impact, and obtain the final dynamic operation plan.

[0041] Based on the power supply unit identifier and transmission channel identifier, the rated output parameters of the power supply unit and the rated capacity parameters of the transmission channel are read to obtain the adjustment rate parameters and ramp limit parameters of the power supply unit. For the power supply unit identifier, the output adjustment parameters are read, and the real-time output values ​​of the unit are obtained from the operating database. A model predictive control method is used to obtain the output adjustment curve. Based on the output adjustment curve, a demand response controller is constructed using a deep reinforcement learning algorithm. The power balance index is obtained by calculating the difference between the power supply and power consumption within the region. For the power balance index, a distributed cooperative control method is used to coordinate the power balance between regions. A regional interconnection scheduling scheme is generated by calculating the power limit value of the inter-regional transmission channel. The regional interconnection scheduling scheme is verified to meet the fault recovery time window constraint. Scheduling schemes that meet the constraint requirements are marked as activated and written into the final scheme data table.

[0042] For example, based on the dynamic scheduling instruction set, the power supply unit identifier and transmission channel identifier are read to obtain the rated output, regulation rate, and ramp limit parameters of the power supply unit, and the rated capacity, flow rate change rate, and transmission delay parameters of the transmission channel. The controlled objects are grouped according to the area identifier, and partition instruction data containing control parameters is generated and written to the area control data table. For the power supply unit identifier in the area control data table, the output regulation parameters are read, the real-time output values ​​of the unit are obtained from the operating database, and the output regulation curve is calculated using a model predictive control method. Based on the ramp limit parameters, the maximum regulation step size is set, and the regulation curve is piecewise linearized to generate an output regulation scheme that meets dynamic constraints, which is written to the regulation scheme data table. The output regulation values ​​are read from the regulation scheme data table, control instructions are generated according to the power supply unit identifier, a demand response controller is constructed using a deep reinforcement learning algorithm, the real-time energy consumption data of the load nodes and the real-time output data of the power supply units are read, the difference between the power supply and consumption in the area is calculated to obtain the power balance index, and it is determined whether the operating status meets the stability margin requirements. Based on the operational status assessment results, the response time and adjustment accuracy parameters of the power supply unit are read. Power compensation is calculated based on real-time response deviation, and power adjustment parameters are adjusted in conjunction with extreme weather warning information. New control commands are generated and written into the optimization command data table. According to the control commands in the optimization command data table, the power exchange plans of adjacent regions are read. A distributed cooperative control method is used to coordinate power balance between regions, calculate the power limit value of inter-regional transmission channels, and generate regional interconnection scheduling schemes, which are written into the cooperative scheduling data table. For the scheduling schemes in the cooperative scheduling data table, the power adjustment time and response delay time of each region are calculated. It is verified whether the scheduling schemes meet the fault recovery time window constraints. Scheduling schemes that meet the constraints are marked as activated and written into the final scheme data table. In multi-regional coordinated dispatching, the operating characteristics of energy supply units in different regions vary significantly. Taking a provincial power grid as an example, this grid is divided into two dispatching regions, north and south. The southern region is mainly powered by thermal power, with a rated output of 350 MW and a minimum output of 175 MW, and a ramp rate limit of 7 MW per minute. The northern region is mainly powered by wind power, with an installed capacity of 500 MW and an output fluctuation range of 0 to 450 MW. The regulation rate is significantly affected by wind speed. When generating the output regulation scheme, considering the operating characteristics of thermal power units, the regulation curve is divided into control cycles with 5-minute intervals. The maximum regulation amplitude within each cycle does not exceed 35 MW. Starting from the current output level of 280 MW, it gradually increases to 320 MW according to a linear change law. At the same time, considering the output characteristics of wind farms, they are used as peak-shaving power sources. When the wind speed reaches 12 m / s, the actual output of wind farms can reach 400 MW, providing sufficient regulation margin for the system.Dynamic supply and demand balance is achieved through a demand response controller. The southern region currently has a total power generation of 1200 MW and a load of 1100 MW, resulting in a power surplus of 100 MW. The northern region has a power generation of 800 MW and a load of 900 MW, resulting in a power deficit of 100 MW. The system calculates a power balance index of 0, meeting the requirements for stable operation. Under extreme weather conditions, dynamic adjustment of control parameters is necessary. On a certain day, the meteorological department issued a typhoon warning, predicting a maximum wind speed of 25 m / s, exceeding the wind turbine cutoff speed of 20 m / s, posing a risk of complete shutdown for wind farms in the northern region. Based on this warning, the system proactively increased the standby capacity of thermal power units, adjusting their output to 300 MW, while also preparing 70 MW of power compensation. At the regional interconnection and dispatch level, the northern and southern regions are connected by a 500 kV transmission channel with a rated transmission capacity of 600 MW. Considering line stability margins, the actual transmission power is limited to within 480 MW. When wind power output in the northern region drops to 200 MW, the southern region provides 250 MW of support power through this channel. The transmission channel's power is approximately 42% of its rated capacity, meeting safe operation requirements. Throughout the execution of the dispatching scheme, the system continuously monitors various time parameters. The response delay time for thermal power units is 30 seconds, the AGC adjustment time is 3 minutes, and the establishment time for cross-regional power support is 5 minutes, all less than the 15-minute fault recovery time window constraint. The resulting dispatching scheme ensures real-time system balance while reserving sufficient adjustment margin to cope with the impact of extreme weather.

[0043] The above are only some preferred embodiments of the present invention, but the present invention is not limited thereto, and many improvements and modifications can be made. Any improvements and modifications made based on the basic principles of the present invention should be considered to fall within the protection scope of the present invention.

Claims

1. A method for coordinated scheduling of integrated energy systems based on multi-objective optimization, characterized in that, The method includes: The system obtains the real-time available capacity of each energy supply unit in the multi-energy supply chain, compares it with a preset normal capacity threshold, determines the capacity reduction of the energy supply unit, identifies the list of affected energy supply units, obtains the current flow and maximum capacity limit of its cross-regional energy transmission channels, calculates the remaining transmission capacity of each channel, and determines the channel identifier of the bottleneck channel based on the comparison results of the channel load rate of each channel with the bottleneck determination threshold, thus obtaining the bottleneck channel set. Based on the channel identifiers in the bottleneck channel set, obtain the downstream load node identifiers, obtain the real-time demand and priority configuration of key loads within the integrated energy system, determine the set of load nodes that need to be prioritized, obtain the types of energy that can be allocated from the multi-energy supply chain, and generate an energy supply list. The power supply units and load nodes are constructed into a weighted directed graph, and the power supply units, energy storage nodes and load nodes are mapped to the vertices of the directed graph. The weight of the edge represents the remaining available capacity of the transmission path. The output allocation of each power generation unit and energy storage unit is coordinated to generate a short-term resource scheduling scheme. The adjustment requirements of energy transmission paths are extracted from short-term resource scheduling schemes, the redundancy capacity of transmission channels is iteratively optimized, the backup supply capacity of multi-energy supply chains is adjusted in combination with the allocation of redundant resources, Monte Carlo simulation is used to judge the stability of the scheduling scheme under random failures, and a preliminary scheduling scheme data table is generated based on the stability judgment results. Acquire historical recovery data of the integrated energy system under extreme weather scenarios, analyze the risk distribution of the multi-energy supply chain, dynamically adjust and optimize the energy allocation ratio of each load node, determine the adaptability of the short-term resource scheduling scheme in the scheduling scheme data table to maintain system stability, iteratively optimize the emergency response speed and magnitude of key nodes based on the adaptability judgment results, and form an emergency scheduling scheme that meets time constraints.

2. The integrated energy system collaborative scheduling method based on multi-objective optimization according to claim 1, characterized in that, The process involves obtaining the real-time available capacity of each energy supply unit in the multi-energy supply chain, comparing it with a preset normal capacity threshold to determine the capacity decline of the energy supply unit, identifying a list of affected energy supply units, obtaining the current flow and maximum capacity limit of its cross-regional energy transmission channels, calculating the remaining transmission capacity of each channel, and determining the channel identifier of the bottleneck channel based on the comparison result of the channel load rate of each channel with the bottleneck determination threshold, thus obtaining a set of bottleneck channels, including: Obtain the operating capacity value corresponding to the power supply unit identifier. If the operating capacity value of the power supply unit is less than the preset ratio of the preset normal capacity threshold, then record the power supply unit in the affected power supply unit data table. Calculate the capacity availability rate based on the operating capacity value and the preset normal capacity threshold and store it in the capacity status data table. For the power supply unit identifier in the capacity status data table, obtain the corresponding channel identifier, calculate the actual flow value of the channel using the channel monitoring parameters, and if the actual flow value of the channel divided by the rated flow value of the channel is greater than the preset bottleneck judgment threshold, then record it in the bottleneck channel data table. For the channel identifier in the bottleneck channel data table, channel operation data is obtained from the historical database, and a channel flow prediction model is established using the linear regression method to obtain the predicted flow value. A bottleneck channel set is generated based on the predicted flow rate and the channel's rated flow rate.

3. The integrated energy system collaborative scheduling method based on multi-objective optimization according to claim 1, characterized in that, The process involves obtaining downstream load node identifiers based on the channel identifiers in the bottleneck channel set, acquiring the real-time demand and priority configuration of key loads within the integrated energy system, determining the set of load nodes requiring priority protection, obtaining available energy types from the multi-energy supply chain, and generating an energy supply list, including: The load node identifier is obtained from the channel identifier in the bottleneck channel set. The load nodes are sorted in descending order using priority values. For load nodes with the same priority value, they are sorted in descending order according to the real-time energy demand value to obtain the priority guaranteed load data table. For the load nodes in the priority load data table, the real-time total demand is calculated according to the energy type of electricity, natural gas and heat. The supply and demand matching rate is calculated by dividing the total dispatchable capacity of the same type of energy supply units by the total load demand. Based on the energy type identifiers whose supply-demand matching rate is greater than the preset threshold for supply-demand matching, obtain the energy supply unit identifier, read the power generation type identifier of the energy supply unit to determine whether the power generation type identifier belongs to the green energy type code set, and obtain the green energy candidate data table. For the energy supply units in the green energy candidate data table, an environmental score is calculated using the unit power generation efficiency index and pollutant emission index. The energy supply unit identifier is selected according to the environmental score in descending order. The remaining available capacity value of the transmission channel is read to determine if it meets the real-time energy demand value of the load node, and a power supply scheme identifier is generated to obtain the energy supply list.

4. The integrated energy system collaborative scheduling method based on multi-objective optimization according to claim 1, characterized in that, The process involves constructing a weighted directed graph of power supply units and load nodes, mapping power supply units, energy storage nodes, and load nodes to vertices of the directed graph, with edge weights representing the remaining available capacity of the transmission path, coordinating the output allocation of each power generation unit and energy storage unit, and generating a short-term resource scheduling scheme, including: A vertex data table is generated based on the energy supply unit identifier and the energy storage node identifier. The vertex data table is used to construct a power grid topology map through the remaining available capacity values ​​of the power transmission channels. Based on the power grid topology, the real-time power generation output of the energy supply unit and the charging and discharging power of the energy storage node are obtained, and the maximum flow algorithm is used to calculate the set of paths between the energy supply unit and the load node. For the path set, the tripping rate and fault recovery time of the read channel are used to obtain the path reliability score by weighted summation method, and the path identifier number with the highest path reliability score is selected. For the path identifier, the output value of the power supply unit and the charging and discharging value of the energy storage node are obtained. The node power allocation scheme is calculated using the minimum deviation method. If the power allocation scheme meets the node power balance constraint, the scheme status identifier is updated to activated, and a short-term resource scheduling scheme is obtained.

5. The integrated energy system collaborative scheduling method based on multi-objective optimization according to claim 1, characterized in that, The process involves extracting adjustment requirements for energy transmission paths from short-term resource scheduling schemes, iteratively optimizing the redundancy capacity of transmission channels, adjusting the backup supply capacity of multi-energy supply chains in conjunction with the allocation of redundant resources, using Monte Carlo simulation to determine the stability of the scheduling scheme under random failures, and generating a preliminary scheduling scheme data table based on the stability determination results. This table includes: The remaining available capacity of the transmission channel is read based on the path identifier, and the channel reliability index is calculated by weighting the fault frequency and fault recovery time. For the energy supply unit identification number and energy storage node identification number connected at both ends of the channel, obtain the rated capacity and current output value of the energy supply unit, obtain the rated capacity and current state of charge of the energy storage node, and calculate the remaining available capacity value of the node. A hierarchical recursive method is used to classify and statistically analyze the remaining available capacity of the nodes to obtain the total reserve capacity of the three energy types: electricity, natural gas and heat. If the duration of the fault exceeds the preset threshold, a random fault scenario is generated using the Monte Carlo method. The node power adjustment scheme is calculated based on the total reserve capacity and channel reliability index. The node stability coefficient is determined by the difference between the power injection and outflow. The power allocation value and node stability coefficient in the short-term resource scheduling scheme are written into the preliminary scheduling scheme data table.

6. The integrated energy system collaborative scheduling method based on multi-objective optimization according to claim 1, characterized in that, The process involves acquiring historical recovery data of the integrated energy system under extreme weather scenarios, analyzing the risk distribution of multiple energy supply chains, dynamically adjusting and optimizing the energy allocation ratio of each load node, determining the adaptability of short-term resource scheduling schemes in the scheduling scheme data table to maintain system stability, iteratively optimizing the emergency response speed and magnitude of key nodes based on the adaptability assessment results, and forming an emergency scheduling scheme that meets time constraints. This includes: Based on the node identifier, the extreme weather occurrence time and node failure record are read from the historical database. The fault response curve and fault recovery curve are obtained by fitting with the least squares method. The average response time and average recovery time are obtained and written into the fault feature data table. For the node identifier in the fault feature data table, read the real-time inventory and rated inventory capacity of the power supply unit, use a Bayesian network to calculate the probability of power supply interruption of the node, obtain the node risk level, and write it into the risk assessment data table. If the real-time inventory of a node is lower than the preset safety inventory, the inventory replenishment priority is set according to the risk level, and the replenishment plan in the inventory adjustment data table is obtained by adopting a multi-energy collaborative replenishment scheme. Based on the replenishment scheme in the inventory adjustment data table, the power supply unit output adjustment rate and the energy storage unit charging and discharging rate are read. The node response rate and adjustment range are optimized by piecewise linear programming to obtain an emergency dispatch scheme that meets time constraints.

7. The integrated energy system collaborative scheduling method based on multi-objective optimization according to claim 1, characterized in that, The method further includes: acquiring dynamic flow change data of cross-regional transmission channels in the multi-energy supply chain according to an emergency dispatch plan that meets time constraints, updating the distribution path of energy resources between supply and demand nodes, dynamically adjusting the power supply priority of critical load nodes in combination with load importance, and dynamically reducing the power supply priority of non-critical load nodes based on their capacity requirements and load characteristics, and generating dynamic dispatch instruction sets for each time period. The system acquires the output adjustment parameters of each energy supply unit in the multi-energy supply chain and the flow control parameters of the transmission channel, issues control commands to the energy management system, dynamically adjusts the output of the energy supply units and the flow of the transmission channel, judges the stable operating status of the integrated energy system under extreme weather impacts, and obtains the final dynamic operation plan.

8. The integrated energy system collaborative scheduling method based on multi-objective optimization according to claim 7, characterized in that, The process involves acquiring dynamic flow change data of cross-regional transmission channels in a multi-energy supply chain based on an emergency dispatch plan that meets time constraints, updating the allocation path of energy resources between supply and demand nodes, dynamically adjusting the power supply priority of critical load nodes based on load importance, and dynamically reducing the power supply priority of non-critical load nodes based on their capacity requirements and load characteristics. This generates a dynamic dispatch instruction set for each time period, including: Real-time flow data and rated capacity of each transmission channel are read from the transmission channel monitoring database, and the channel flow prediction data for future periods are obtained by using a deep neural network model. Based on the channel flow prediction data, the identification numbers of the supply and demand nodes at both ends of the channel are obtained, and the shortest path algorithm is used to calculate the transmission path data with the minimum loss between the supply and demand nodes. Based on the transmission path data, the load node time-of-use energy consumption data is obtained, and the load node energy consumption characteristic score data is calculated using the fuzzy comprehensive evaluation method. The non-critical load node identifiers are read from the energy consumption characteristic score data, and load control scheme data is generated by sorting the energy consumption characteristic scores from low to high, forming a dynamic scheduling instruction set for each time period.

9. A method for coordinated scheduling of integrated energy systems based on multi-objective optimization according to claim 7, characterized in that, The process of acquiring the output adjustment parameters of each energy supply unit in the multi-energy supply chain and the flow control parameters of the transmission channel, issuing control commands to the energy management system, dynamically adjusting the output of the energy supply units and the flow of the transmission channel, determining the stable operating status of the integrated energy system under extreme weather impacts, and obtaining the final dynamic operation plan includes: Based on the power supply unit identification number and the power transmission channel identification number, read the rated output parameters of the power supply unit and the rated capacity parameters of the power transmission channel, and obtain the regulation rate parameters and ramp limit parameters of the power supply unit. The output adjustment parameters are read according to the identification number of the power supply unit, the real-time output value of the unit is obtained from the operation database, and the output adjustment curve is obtained by model predictive control method. Based on the output adjustment curve, a demand response controller is constructed using a deep reinforcement learning algorithm, and a power balance index is obtained by calculating the difference between the power supply and the power consumption in the region. A distributed collaborative control method is used to coordinate the power balance between regions for the power balance index. The regional interconnection scheduling scheme is generated by calculating the power limit value of the cross-regional transmission channel. The regional interconnection scheduling scheme is verified to meet the fault recovery time window constraint. The scheduling scheme that meets the constraint requirement is marked as activated and used as the final scheme data table.

Images