Source-load-storage collaborative optimization system and method for flexibility improvement of industrial park

Abstract

The application belongs to the technical field of source-load-storage collaborative optimization, and particularly relates to an industrial park source-load-storage collaborative optimization system and method for flexibility improvement. The system is provided with a dynamic resource portrait modeling module, a space-time elasticity evaluation solving module, a quantum collaborative optimization module, a virtual-real linkage verification module and a collaborative optimization instruction updating module, and solves the problem of uncontrolled range of traditional industrial park supply-demand balance. The system helps to break through the limitation of traditional static equipment model, more accurately depict source-load-storage resource characteristics, consider the elasticity of resources in time and space dimensions and the game relationship among different subjects (source, load and storage), is more in line with the complex actual operation scene of the industrial park compared with the traditional optimization mode from a single dimension or ignoring the game of subjects, and guarantees the reliability of the optimization instruction. The application realizes accurate depiction, collaborative optimization and reliable execution of the source-load-storage resources of the industrial park.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of source-load-storage collaborative optimization, and specifically relates to an industrial park source-load-storage collaborative optimization system and method for flexibility improvement. BACKGROUND

[0002] Renewable energy generation accounts for an increasing proportion of total power generation in the power system. Renewable energy will inevitably become an important mainstay of the power system, and the construction of a new type of power system has become an important development direction. Industrial parks are places where industrial production is concentrated, and energy consumption and carbon emissions are large. Therefore, achieving source-load-storage collaborative optimization can improve energy utilization efficiency, reduce the use of fossil energy, reduce carbon emissions, and help green and low-carbon development of the park, which is the development demand in the future.

[0003] In existing research on source-load-storage collaborative optimization, a "source follows load" dispatching mode is adopted, such as CN115833255A, a source-load collaborative optimization dispatching method and terminal for flexibility supply and demand balance, which analyzes the source-load collaborative optimization dispatching of flexibility supply and demand balance, but considers how to realize the flexibility of both sides of source and load, and does not consider the uncertainty factors when analyzing the flexibility supply and demand balance of the new type of power system.

[0004] For example, CN115222195A considers the optimization dispatching method of distribution network of source-grid-load-storage flexibility resources, which mentions the up-regulation flexibility deficiency index and the down-regulation flexibility deficiency index, but does not consider the characteristics of flexibility balance in specific scenarios, so it cannot guarantee that the flexibility supply and demand balance of the new type of power system is within the controllable range. SUMMARY

[0005] In view of the deficiencies of the prior art, the application provides an industrial park source-load-storage collaborative optimization system and method for flexibility improvement, which realizes accurate characterization, collaborative optimization and reliable execution of industrial park source-load-storage resources through dynamic modeling, quantum optimization and virtual-real linkage verification, and significantly improves the system flexibility and operation efficiency.

[0006] To achieve the above purpose, the application is implemented through the following technical solutions:

[0007] The industrial park source-load-storage collaborative optimization system for flexibility improvement comprises:

[0008] A dynamic resource portrait modeling module is used to collect industrial park source-load-storage data, dynamically model the resource portrait, and output a dynamic device performance set.

[0009] The spatio-temporal elasticity evaluation solving module is configured to acquire power grid constraints and market price signals, establish constraints and calculate a dynamic cost function in combination with a dynamic device performance set, perform spatio-temporal elasticity evaluation, and output an elastic boundary surface through dynamic game solving;

[0010] The quantum collaborative optimization module is configured to consider the elastic boundary surface and collect real-time monitoring data of the industrial park, and obtain an optimal strategy vector through quantum collaborative optimization;

[0011] The virtual-real linkage verification module is configured to perform virtual-real linkage verification based on the optimal strategy vector, and obtain a collaborative optimization verification instruction set and a collaborative optimization credibility label;

[0012] The collaborative optimization instruction updating module is configured to acquire and analyze industrial park data collection completeness features and industrial park data transmission reliability features, eliminate untrustworthy collaborative optimization instructions in combination with the collaborative optimization credibility label, and perform collaborative optimization adjustment on the source-load-storage of the industrial park based on the remaining trustworthy collaborative optimization instructions.

[0013] Preferably, in the dynamic resource portrait modeling module, industrial park source-load-storage data is collected, and dynamic resource portrait modeling is performed, and the specific process is as follows:

[0014] The industrial park source-load-storage data includes industrial park photovoltaic output, industrial park environmental temperature, industrial park energy storage SOC, industrial park device rated power, industrial park device actual power, industrial park device rated efficiency, industrial park photovoltaic panel installation area, and industrial park irradiance.

[0015] The dynamic resource portrait modeling is performed in the following manner:

[0016] The long short-term memory network (LSTM) formula is set as follows:

[0017] ;

[0018] In the formula, h t is a hidden state, is the industrial park photovoltaic output at time t, is the industrial park environmental temperature at time t, is the industrial park energy storage SOC at time t. For the energy storage SOC, a particle swarm optimization is used to select features and maximize a particle swarm optimization objective function, wherein the particle swarm optimization objective function is as follows:

[0019]

[0020] ;

[0021] In the formula, η t is the efficiency of the energy storage system for charging and discharging operation at time t, ​​is the time decay factor, e is the natural constant;

[0022] Normalizing the actual power of the industrial park equipment:

[0023] ;

[0024] wherein, is the normalized power of the industrial park equipment at time t, is the actual power of the industrial park equipment at time t, is the rated power of the industrial park equipment;

[0025] Establishing an efficiency decay model, combining the effects of aging and load, to calculate the real-time efficiency of the industrial park equipment:

[0026] ;

[0027] wherein, is the real-time efficiency of the industrial park equipment at time t, is the rated efficiency of the industrial park equipment, is the aging factor stored in the database, is the first correction coefficient of load fluctuation on efficiency, is the second correction coefficient of load fluctuation on efficiency, is the historical load average of the industrial park equipment;

[0028] Correction of photovoltaic output:

[0029] Based on the industrial park irradiance and the industrial park environmental temperature, the weather deviation of the historical data is corrected:

[0030] ;

[0031] wherein, is the industrial park irradiance, A is the installation area of the industrial park photovoltaic panel, is the photoelectric conversion efficiency.

[0032] Preferably, in the dynamic resource portrait modeling module, a set of dynamic equipment performance is output, and the specific process is:

[0033] Obtaining the maximum power data of the industrial park equipment , the response time data of the industrial park equipment , the maintenance cost data of the industrial park equipment ;

[0034] Based on the power data of the past 24 hours, the equipment health index is analyzed, wherein the power data includes the actual power of the industrial park equipment and the efficiency of the industrial park equipment:

[0035] ;

[0036] In the formula, The health index of equipment in the industrial park at time k. The actual power of the equipment in the industrial park at time k. Let k be the energy conversion efficiency of the equipment in the industrial park at time k, which is the ratio of output energy to input energy.

[0037] Will , , and Combined into a dynamic device performance set .

[0038] Preferably, in the spatiotemporal elasticity assessment and solution module, the grid constraints and market price signals are obtained, and constraints are established by combining the dynamic equipment efficiency set and the dynamic cost function is calculated. The specific process is as follows:

[0039] Response time constraints, based on equipment response time data from the industrial park. Extracting the energy storage response time of industrial parks and the response time of gas turbines in industrial parks :

[0040] ;

[0041] In the formula, This refers to the response time for scheduling instructions.

[0042] Data on equipment maintenance costs in industrial parks Extracting the highest maintenance cost of each piece of equipment in the industrial park The maintenance cost constraint is the maintenance cost of the equipment under the dispatch command. Not exceeding the maximum maintenance cost of the corresponding industrial park equipment ;

[0043] Obtain the grid constraints and voltage limits, and transform them into quadratic programming conditions:

[0044] ;

[0045] In the formula, The reference voltage, For the first Line impedance at each node, For the first The injection power of each node, where n represents the number of nodes participating in the calculation;

[0046] Define the feasible power region: ;

[0047] In the formula, the minimum power value allowed to be transmitted by the line, the power actually transmitted by the line, the maximum power value allowed to be transmitted by the line;

[0048] The market price signal is obtained, and the dynamic cost function is:

[0049] ;

[0050] In the formula, the dynamic cost at time t, the peak electricity price, the power actually transmitted by the line at time t, the time-of-use electricity price, the start time of the time-of-use electricity price calculation, the end time of the time-of-use electricity price calculation.

[0051] Preferably, in the space-time elasticity evaluation solving module, the space-time elasticity evaluation is performed, and the dynamic game solving outputs an elasticity boundary surface, and the specific process is as follows:

[0052] Elasticity margin calculation:

[0053] ;

[0054] Wherein, the source side adjustment capability is calculated as follows: ;

[0055] The load elasticity is calculated as follows: ;

[0056] In the formula, the flexibility evaluation index, T is the total time period, and m is the number of source side devices, the maximum power output value of the s-th source side device, the minimum power output value of the s-th source side device, the number of load sides, the power change amount of the l-th load, the time decay factor, e is the natural constant, the partial derivative of the s-th source side efficiency to the power , the inverse of the response time, the load transfer cost, the maximum power of the l-th load, and t is the time index;

[0057] Dynamic game solving Nash equilibrium, source, load, and storage through virtual iteration to reach the optimal strategy:

[0058] ;

[0059] In the formula, is the strategy variable of the f party in the K+1 iteration, is the strategy variable of the f party, is the strategy combination of other parties except the f party in the K iteration, is the utility function of the f party;

[0060] The convergence condition is that the strategy change rate is less than 1%;

[0061] Output elastic boundary surface ;

[0062] is the time boundary, that is, the updated allowable adjustment time window is solved, is the power boundary, that is, the updated maximum schedulable range of power is solved, is the cost boundary, that is, the updated cost constraint is solved.

[0063] Preferably, in the quantum collaborative optimization module, the elastic boundary surface is considered and the real-time monitoring data of the industrial park is collected, and the quantum collaborative optimization obtains the optimal strategy vector, and the specific process is:

[0064] Input the elastic boundary surface, map the optimal solution set to the quantum state amplitude Provide an initial population for quantum collaborative optimization quantum chromosome:

[0065] ;

[0066] In the formula, is the power of the jth strategy of the elastic boundary surface, is the total power of the elastic boundary surface strategy, j is the index of the elastic boundary surface strategy, and J is the number of elastic boundary surface strategies;

[0067] Real-time monitoring data, establish energy storage SOC constraint:

[0068] ;

[0069] In the formula, is the highest value of the power of the industrial park equipment, is the rated power of the industrial park equipment, and SOC is the state of charge of the industrial park equipment;

[0070] Quantum collaborative optimization:

[0071] For chromosome encoding, the rotation angle formula for the dynamic rotation door strategy is:

[0072] ;

[0073] In the formula, For the current iteration number The rotation angle at that time, The maximum number of iterations, For the first The optimal fitness of the next iteration The optimal fitness for the 0th iteration;

[0074] Fitness assessment:

[0075] ;

[0076] In the formula, This is the fitness assessment value. Power generation capacity of the equipment, To reduce the maintenance costs of the dispatching command equipment, As a flexibility assessment indicator, It is the hyperbolic tangent function. This represents the change in carbon dioxide emissions. This serves as a baseline value for carbon dioxide emissions;

[0077] The policy corresponding to the highest fitness evaluation value is recorded as the optimal policy, and the optimal policy vector is output.

[0078] Preferably, in the virtual-real linkage verification module, virtual-real linkage verification is performed based on the optimal strategy vector to obtain a collaborative optimization verification instruction set and a collaborative optimization credibility label. The specific process is as follows:

[0079] The optimal strategy vector is decomposed into instructions, and a digital twin is built for verification, including photovoltaic and energy storage models:

[0080] The photovoltaic model is a single diode model, and the energy storage model is an RC equivalent circuit.

[0081] Deviation calculation is performed by comparing virtual and real data:

[0082] ;

[0083] In the formula, The overall bias is given by Q, where Q is the sample size. This represents the actual measurement value of the q-th sample. Let be the virtual model prediction value for the q-th sample. Rated power;

[0084] If the overall deviation exceeds the deviation threshold stored in the database, a correction is triggered.

[0085] The sliding window least square method corrects the parameters, ensuring that the digital twin is consistent with the entity:

[0086]

[0087] wherein, is the updated model parameter vector, is the model parameter vector, is the actual measurement value at time t, is the model prediction value at time t based on the model parameter ;

[0088] After correction, the comprehensive deviation degree is recalculated, the comprehensive deviation degree-collaborative optimization credibility mapping set stored in the database is obtained, and the matching collaborative optimization credibility is found based on the comprehensive deviation degree, which is denoted as a collaborative optimization credibility label;

[0089] The collaborative optimization verification instruction set and the corresponding collaborative optimization credibility label are output.

[0090] Preferably, in the collaborative optimization instruction updating module, the industrial park data collection completeness feature and the industrial park data transmission reliability feature are obtained and analyzed, and the specific process is as follows:

[0091] The industrial park data collection completeness feature and the industrial park data transmission reliability feature are obtained, including the industrial park data collection completeness ratio, the industrial park data collection error rate, and the industrial park data transmission packet loss rate.

[0092] The industrial park data collection completeness feature and the industrial park data transmission reliability feature are analyzed to obtain a credibility verification index, which serves as an analysis basis for eliminating untrustworthy collaborative optimization instructions.

[0093] The credibility verification index, and the specific process is as follows:

[0094]

[0095] wherein, is the credibility verification index, is the industrial park data collection completeness ratio, is the industrial park data collection error rate, is the industrial park data transmission packet loss rate, and e is a natural constant.

[0096] A credibility verification index-collaborative optimization credibility update value mapping set stored in the database is obtained, and a matching collaborative optimization credibility update value is found based on the credibility verification index.

[0097] ​​Preferably, in the synergistic optimization instruction updating module, untrusted synergistic optimization instructions are removed in combination with the synergistic optimization credibility label, and the remaining trusted synergistic optimization instructions are used to perform synergistic optimization adjustment on the source-load-storage of the industrial park.

[0098] The synergistic optimization credibility update value is accumulated with the synergistic optimization credibility to obtain a synergistic optimization credibility feature, and the synergistic optimization credibility feature is compared with a synergistic optimization credibility threshold value stored in a database.

[0099] If the synergistic optimization credibility feature is not lower than the synergistic optimization credibility threshold value, the synergistic optimization instruction corresponding to the synergistic optimization credibility feature is trusted.

[0100] If the synergistic optimization credibility feature is lower than the synergistic optimization credibility threshold value, the synergistic optimization instruction corresponding to the synergistic optimization credibility feature is untrusted, and the untrusted synergistic optimization instruction is removed.

[0101] The remaining trusted synergistic optimization instructions are used to perform synergistic optimization adjustment on the source-load-storage of the industrial park.

[0102] The industrial park source-load-storage synergistic optimization method for flexibility improvement is based on the above system and includes the following steps.

[0103] Industrial park source-load-storage data are collected, dynamic resource portrait modeling is performed, and a dynamic device performance set is output.

[0104] Power grid constraints and market price signals are obtained, constraint conditions are established in combination with the dynamic device performance set, a dynamic cost function is calculated, time-space elasticity is evaluated, and an elastic boundary surface is output by dynamic game solution.

[0105] The elastic boundary surface is considered, and real-time monitoring data of the industrial park are collected, and an optimal strategy vector is obtained by quantum synergistic optimization.

[0106] Based on the optimal strategy vector, virtual-real linkage verification is performed to obtain a synergistic optimization verification instruction set and a synergistic optimization credibility label.

[0107] Industrial park data collection completeness features and industrial park data transmission reliability features are obtained and analyzed, untrusted synergistic optimization instructions are removed in combination with the synergistic optimization credibility label, and the remaining trusted synergistic optimization instructions are used to perform synergistic optimization adjustment on the source-load-storage of the industrial park.

[0108] The present application has the following beneficial effects:

[0109] This invention's system comprises multiple functional modules, including dynamic resource profiling and modeling, spatiotemporal elasticity assessment and solution, quantum collaborative optimization, virtual-real linkage verification, and collaborative optimization instruction updates. Breaking away from traditional single-optimization methods, it achieves full-process collaboration from data modeling, assessment and solution, strategy optimization to verification and updates, representing an innovation in the source-load-storage collaborative optimization system architecture. It collects source-load-storage data from industrial parks for dynamic modeling and outputs a dynamic equipment performance set, overcoming the limitations of traditional static equipment models. This allows for real-time reflection of equipment performance under different operating conditions and at different times, more accurately characterizing the resource characteristics of source-load-storage. It combines grid constraints and market price signals to establish constraints and dynamic cost functions for spatiotemporal elasticity assessment and dynamic game theory solutions. Considering the elasticity of resources in both time and space, as well as the game relationships between different entities (source, load, and storage), it better aligns with the complex actual operating scenarios of industrial parks compared to traditional optimization methods that focus on a single dimension or ignore the game between entities.

[0110] This invention utilizes quantum correlation technology for collaborative optimization to obtain the optimal strategy vector. Quantum computing has unique advantages in handling complex optimization problems, enabling more efficient search of the solution space and finding better strategies, providing a new technical path for source-load-storage collaborative optimization. A collaborative optimization verification instruction set and trustworthiness labels are obtained through virtual-real linkage verification, and untrustworthy instructions are eliminated by combining data feature analysis. The reliability of the optimization instructions is ensured by using a virtual model linked with the actual system for verification and introducing a trustworthiness management mechanism. Attached Figure Description

[0111] Figure 1 This is a schematic diagram of the system module connections of the present invention;

[0112] Figure 2 This is a schematic diagram of the method steps of the present invention;

[0113] Figure 3 This is a schematic diagram of the complete process of the method of the present invention. Detailed Implementation

[0114] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0115] Example 1: As Figure 1 As shown, the industrial park source-load-storage collaborative optimization system for improving flexibility includes: a dynamic resource profile modeling module, a spatiotemporal elasticity assessment and solution module, a quantum collaborative optimization module, a virtual-real linkage verification module, and a collaborative optimization instruction update module.

[0116] The dynamic resource profiling and modeling module is used to collect source, load, and storage data from industrial parks, create dynamic resource profiling models, and output dynamic equipment performance sets.

[0117] The specific process is: collecting industrial park source and load storage data, including industrial park photovoltaic output, industrial park environmental temperature, industrial park energy storage SOC, industrial park equipment rated power, industrial park equipment actual power, industrial park equipment rated efficiency, industrial park photovoltaic panel installation area, and industrial park irradiance;

[0118] Dynamic resource portrait modeling is performed, and the modeling mode is:

[0119] The long short-term memory network LSTM formula is set:

[0120] In the formula, is a hidden state, is the industrial park photovoltaic output at time t, is the industrial park environmental temperature at time t, is the industrial park energy storage SOC at time t;

[0121] For energy storage SOC: use particle swarm optimization to select key features and maximize the particle swarm optimization objective function, wherein the particle swarm optimization objective function is:

[0122] In the formula, is the efficiency of the energy storage system at time t, is a time decay factor, and e is a natural constant;

[0123] The industrial park equipment actual power is normalized to ensure that the outputs of different equipment are comparable:

[0124] In the formula, is the normalized industrial park equipment power at time t, is the industrial park equipment actual power at time t, is the industrial park equipment rated power;

[0125] 20% safety margin is reserved, is to avoid over-limit;

[0126] An efficiency decay model is established to calculate the real-time efficiency of the industrial park equipment considering the effects of aging and load:

[0127] In the formula, is the real-time efficiency of the industrial park equipment at time t, is the industrial park equipment rated efficiency, is an aging factor stored in the database, is a first correction coefficient of load fluctuation on efficiency, is a second correction coefficient of load fluctuation on efficiency, is the average of historical load of the industrial park equipment; ln is the logarithm with base e;

[0128] Correction of photovoltaic output:

[0129] Correction of weather bias of historical data based on industrial park irradiance and industrial park ambient temperature:

[0130] wherein, is the industrial park irradiance, A is the installation area of the industrial park photovoltaic panel, is the photoelectric conversion efficiency.

[0131] Collecting multi-dimensional data such as industrial park photovoltaic output, ambient temperature, energy storage SOC, and equipment power, the key information of source, load, and storage is comprehensively covered, providing rich and accurate data basis for subsequent analysis and optimization, and avoiding analysis deviation caused by data loss.

[0132] Using the LSTM formula combined with photovoltaic output, temperature, and energy storage SOC information, the dynamic characteristics of photovoltaic affected by weather and season can be effectively captured, such as photovoltaic output changes under different time periods and different weather conditions, providing a basis for reasonable scheduling and prediction of photovoltaic power.

[0133] Through particle swarm optimization to screen energy storage SOC key features, considering the time decay of charging and discharging efficiency, it is helpful to accurately evaluate the performance and health status of the energy storage system, reasonably arrange the charging and discharging strategy of the energy storage, and improve the utilization efficiency and life of the energy storage.

[0134] Normalizing the actual power of the equipment ensures the comparability of the output of different equipment, facilitating unified analysis and scheduling; establishing an efficiency decay model to calculate the real-time efficiency considering the effects of aging and load, which can more accurately evaluate the actual performance of the equipment in different operating stages, and provide a reference for equipment maintenance and scheduling.

[0135] Based on irradiance and temperature to correct photovoltaic output, considering the influence of meteorological factors on photovoltaic, making photovoltaic output prediction more realistic, and improving the rationality of photovoltaic energy consumption and utilization.

[0136] Through accurate data processing and resource characteristic description, the dynamic resource portrait modeling module can more clearly understand the resource state and change law of each link of source, load, and storage, and then realize the optimal configuration of power supply, load, and energy storage resources in subsequent links such as spatiotemporal flexibility evaluation solution and quantum synergistic optimization, improve energy utilization efficiency, and reduce resource waste. Accurately grasping the dynamic characteristics of photovoltaic, energy storage, and other resources enables the system to better cope with the intermittency and volatility of renewable energy and load changes, and improves the flexibility and response capability of the system; at the same time, based on reliable data and models for decision-making, the reliability of system operation is guaranteed, and the operation risk caused by inaccurate understanding of resource characteristics is reduced.

[0137] The process of outputting a dynamic set of device performance metrics is as follows:

[0138] Obtain maximum power data of equipment in industrial parks Industrial park equipment response time data Industrial park equipment maintenance cost data ;

[0139] Equipment health index is analyzed based on power data from the past 24 hours, including the actual power and efficiency of equipment in the industrial park.

[0140] In the formula, The health index of equipment in the industrial park at time k. The actual power of the equipment in the industrial park at time k. Let k be the energy conversion efficiency of the equipment in the industrial park at time k, which is the ratio of output energy to input energy.

[0141] Will , , and Combined into a dynamic device performance set .

[0142] The dynamic equipment performance dataset encompasses multiple dimensions of information, including maximum power, response time, maintenance costs, and health index. Maximum power reflects the equipment's maximum power generation or work output capability; response time reflects the equipment's speed of responding to commands, impacting the overall system response efficiency; maintenance costs measure the equipment's operational economy; and the health index, combining actual power and efficiency, assesses the equipment's health status. This multi-dimensional approach allows for a comprehensive and accurate understanding of equipment status, providing a reliable basis for subsequent decision-making.

[0143] Based on dynamic equipment performance sets, scheduling can be rationally arranged according to equipment capabilities and status, prioritizing the use of equipment with high maximum power, fast response speed, and good health status to improve the overall system operating efficiency; for equipment with high maintenance costs and low health index, usage strategies can be adjusted, such as reducing usage frequency or arranging maintenance to optimize resource utilization.

[0144] The spatiotemporal elasticity assessment and solution module is used to obtain power grid constraints and market price signals, establish constraints by combining dynamic equipment efficiency sets, calculate dynamic cost functions, perform spatiotemporal elasticity assessment, and output elastic boundary surfaces through dynamic game theory solutions.

[0145] The specific process is as follows: Response time constraints, based on the response time data of equipment in the industrial park. Extracting the energy storage response time of industrial parks and the response time of gas turbines in industrial parks , ensure the feasibility of scheduling instructions:

[0146] ; wherein, is the response time of the scheduling instruction;

[0147] Extract the highest maintenance cost of each device in the industrial park from the industrial park equipment maintenance cost data , the maintenance cost constraint is the device maintenance cost of the scheduling instruction not higher than the corresponding industrial park equipment highest maintenance cost ;

[0148] Get grid constraints, voltage limits, and convert to quadratic programming conditions:

[0149] ; wherein, is the reference voltage, is the line impedance of the node, is the injection power of the node, and n is the number of nodes participating in the calculation;

[0150] Define the power feasible region: ; wherein, is the minimum power value allowed to be transmitted by the line, is the actual power transmitted by the line, is the maximum power value allowed to be transmitted by the line;

[0151] Get market price signals, and the dynamic cost function is:

[0152] ; wherein, is the dynamic cost at time t, is the peak electricity price, is the actual power transmitted by the line at time t, is the time-of-use electricity price, is the start time of time-of-use electricity price calculation, is the end time of time-of-use electricity price calculation.

[0153] Through the response time constraint, the response time of devices such as energy storage and gas turbines is extracted to ensure that the response time of the scheduling instruction meets the requirements, avoid the scheduling instruction cannot be effectively executed due to slow device response, ensure the timeliness and effectiveness of system scheduling operation, and maintain stable operation of the system.

[0154] ​The highest maintenance cost is extracted from the equipment maintenance cost as a constraint to limit the maintenance cost of the scheduling instruction involving the equipment, prevent excessive use of equipment or unreasonable scheduling from causing high maintenance cost, ensure the system runs within the economically feasible range, and improve the service life of the equipment and the overall economy of the system.

[0155] The grid voltage limit is converted into a quadratic programming condition, considering factors such as reference voltage, line impedance and node injection power, to ensure that the grid voltage fluctuates within a safe range. Avoiding high or low voltage damage to grid equipment, ensuring safe and reliable operation of the grid, providing stable power supply for various equipment in the industrial park.

[0156] In combination with the above grid constraints and cost functions, in the process of spatiotemporal flexibility assessment and dynamic game solving, the source, load and storage subjects optimize resource allocation according to their own characteristics and market price signals. For example, energy storage devices charge at low price valleys and discharge at peak times, achieving rational allocation of resources in time and space, and improving the overall efficiency and economy of the system.

[0157] Perform spatiotemporal flexibility assessment and dynamic game solving to output the flexible boundary surface. The specific process is as follows: flexibility margin calculation:

[0158] ;

[0159] The calculation method of the source side regulation capacity is as follows: ;

[0160] The calculation method of the load flexibility is as follows: ;

[0161] In the formula, is the flexibility evaluation index, T is the total time period, m is the number of source side devices, is the maximum power output value of the s-th source side device, is the minimum power output value of the s-th source side device, is the number of load sides, is the power change of the l-th load, is the time decay factor, e is the natural constant, is the partial derivative of the s-th source side efficiency to its power , is the inverse of the response time, is the load transfer cost, is the maximum power of the l-th load, t is the time index;

[0162] Dynamic game solving Nash equilibrium, source, load and storage reach the optimal strategy through virtual iteration:

[0163] wherein, is the strategy variable of the fth party in the (K+1)th iteration, is the strategy variable of the fth party, is the strategy combination of all parties except the fth party in the Kth iteration, is the utility function of the fth party;

[0164] The convergence condition is that the strategy change rate is less than 1%.

[0165] Output elastic boundary surface ; is the time boundary, i.e., the time window of allowed adjustment for solving the update, is the power boundary, i.e., the maximum schedulable range of power for solving the update, is the cost boundary, i.e., the cost constraint for solving the update.

[0166] The flexibility evaluation index is calculated by the elastic margin, which comprehensively considers factors such as source side adjustment capacity and load flexibility, and quantifies the flexibility of the source-load-storage system in the industrial park within the total time period, providing specific numerical basis for system operation state evaluation, and facilitating intuitive understanding of the system flexible adjustment capacity level.

[0167] Through dynamic game solving Nash equilibrium, the source, load and storage three parties continuously adjust the strategy in the virtual iteration process, with the goal of maximizing their own utility function. Finally, the convergence condition that the strategy change rate is less than 1% is reached, and the optimal strategy of the three parties is found, realizing the maximization of the interests of each participant and improving the overall operation benefit of the system. The dynamic game process prompts the source, load and storage three parties to consider the strategies of other parties, and seeks balance in mutual influence and restriction. Break the independent running state of each party, strengthen cooperation, improve the ability of the system to respond to complex working conditions and external changes, and realize the efficient collaborative operation of the source-load-storage system.

[0168] Output elastic boundary surface, and clearly define the time boundary (allowed adjustment time window), power boundary (maximum schedulable range of power) and cost boundary (cost constraint). Provide clear boundary conditions for the operation of the source-load-storage system, guide the system to adjust and optimize within a reasonable range, and avoid risks and losses caused by out-of-range operation.

[0169] The quantum collaborative optimization module is used to consider the elastic boundary surface and collect real-time monitoring data of the industrial park, and the optimal strategy vector is obtained by quantum collaborative optimization.

[0170] The specific process is: input the elastic boundary surface, map the optimal solution set to the quantum state amplitude , provide an initial population for quantum collaborative optimization quantum chromosomes:

[0171] wherein, is the power of the jth strategy of the elastic boundary surface, is the total power of the elastic boundary surface strategy, j is the index of the elastic boundary surface strategy, and J is the number of the elastic boundary surface strategies;

[0172] Real-time monitoring data, establish energy storage SOC constraint:

[0173] wherein, is the highest value of the industrial park equipment power, is the rated power of the industrial park equipment, and SOC is the state of charge of the industrial park equipment;

[0174] Quantum collaborative optimization:

[0175] Carry out chromosome coding, and the dynamic rotating door strategy rotation angle formula is:

[0176] wherein, is the rotation angle at the current iteration number , is the index of the iteration number, is the maximum iteration number, is the optimal fitness of the jth iteration, is the optimal fitness of the 0th iteration (initial state); Fitness evaluation:

[0177] Fitness evaluation:

[0178] wherein, is the fitness evaluation value, is the equipment power generation, is the scheduling instruction equipment maintenance cost, is the flexibility evaluation index, is the hyperbolic tangent function, is the change amount of carbon dioxide emissions, is the reference value of carbon dioxide emissions;

[0179] The strategy corresponding to the highest value of the fitness evaluation value is recorded as the optimal strategy, and the optimal strategy vector is output.

[0180] The quantum collaborative optimization module maps the optimal solution set to the quantum state amplitude by means of the quantum computing characteristics, and provides an initial population for the quantum chromosome. By utilizing the superposition and entanglement characteristics of the quantum state, parallel search is carried out in the solution space. Compared with the traditional optimization algorithm, the optimal solution can be explored more quickly, the optimization solving efficiency is greatly improved, and it is especially suitable for the optimization problem of complex multi-variables such as industrial park source and load storage. ​

[0181] The dynamic rotating door strategy provides a wide range search by a large angle formula in the early stage of algorithm iteration to quickly locate the better solution area, and uses a small angle for fine optimization in the later iteration to further explore more accurate optimal solution in the area found in the early stage. Through the adaptive adjustment of the search range and precision, the quantum collaborative optimization is assisted to converge to the global optimal solution more efficiently.

[0182] The time, power and cost boundaries determined by the elastic boundary surface are considered, and the energy storage SOC constraint is established combined with real-time monitoring data. The upper limit of equipment power is adjusted flexibly according to the state of charge of energy storage to ensure the safe and stable operation of energy storage, and also makes the optimization strategy meet the multiple constraints of actual operation of the system, enhancing the feasibility and practicality of the strategy.

[0183] The fitness evaluation function comprehensively considers factors such as power generation, dispatching instruction equipment maintenance cost, flexibility evaluation index and carbon dioxide emission change. The performance of the system in economy, flexibility, environmental protection and other aspects is comprehensively measured, so that the optimization strategy is not limited to a single target, but pursues the maximization of overall benefit, and promotes the sustainable development of the source-load-storage system in the industrial park.

[0184] The virtual-real linkage verification module is used to verify the virtual-real linkage based on the optimal strategy vector, and obtain the collaborative optimization verification instruction set and the collaborative optimization credibility label.

[0185] The specific process is: the optimal strategy vector is decomposed into instructions, a digital twin is established for verification, including a photovoltaic model and an energy storage model: the photovoltaic model is a single diode model, and the energy storage model is an RC equivalent circuit;

[0186] The deviation degree is calculated by virtual-real comparison:

[0187] ; in the formula, is the comprehensive deviation degree, Q is the number of samples, is the actual measurement value of the qth sample, is the virtual model prediction value of the qth sample, is the rated power;

[0188] If the comprehensive deviation degree is higher than the deviation degree threshold value stored in the database, the correction is triggered; the sliding window least square method is used for parameter correction to ensure that the digital twin is consistent with the entity:

[0189] ; in the formula, is the updated model parameter vector, is the model parameter vector, is the actual measurement value at time t, is the model prediction value at time t based on the model parameter at time t.

[0190] Re-calculate the comprehensive deviation degree after correction, obtain the comprehensive deviation degree-cooperative optimization credibility mapping set stored in the database, find the matching cooperative optimization credibility based on the comprehensive deviation degree, and mark it as the cooperative optimization credibility label; output the cooperative optimization verification instruction set and the corresponding cooperative optimization credibility label.

[0191] By establishing a photovoltaic single-diode model and an energy storage RC equivalent circuit model to construct a digital twin, the optimal strategy vector decomposition instruction is simulated in the virtual model. By using the comprehensive deviation degree calculation formula, the actual measured value is compared with the virtual model prediction value, and the difference between the virtual and actual is accurately measured, providing a quantitative basis for judging the reliability of the instruction.

[0192] According to the comprehensive deviation degree, the cooperative optimization credibility label is determined from the comprehensive deviation degree-cooperative optimization credibility mapping set. It intuitively reflects the reliability of the instruction, which is convenient for subsequent screening and use. Eliminate instructions with low credibility and retain instructions with high credibility to ensure the reliability of the output instruction set and reduce the possibility of system operation risks caused by incorrect instructions.

[0193] When the comprehensive deviation degree is higher than the threshold value, the sliding window least square method is triggered to correct the model parameters. The parameters are continuously adjusted according to the difference between the actual measured value and the model prediction value to ensure that the digital twin is consistent with the actual entity, so that the virtual model can accurately reflect the actual system operation state, providing a reliable simulation environment for instruction verification and optimization.

[0194] Through virtual-real linkage verification and credibility management, reliable instructions are screened to form a cooperative optimization verification instruction set. High-quality instructions are provided for the operation of the source-load-storage system in the industrial park, guiding the rational operation of source-load-storage equipment, achieving optimal resource allocation, and improving the overall stability and operation efficiency of the system.

[0195] The cooperative optimization instruction updating module is used to obtain and analyze the industrial park data collection completeness characteristics and industrial park data transmission reliability characteristics, eliminate untrustworthy cooperative optimization instructions based on the cooperative optimization credibility label, and adjust the source-load-storage in the industrial park based on the remaining trustworthy cooperative optimization instructions.

[0196] The specific process is: obtain the industrial park data collection completeness characteristics and the industrial park data transmission reliability characteristics, including the industrial park data collection completeness ratio, the industrial park data collection error rate, and the industrial park data transmission packet loss rate; analyze the industrial park data collection completeness characteristics and the industrial park data transmission reliability characteristics to obtain the credibility verification index, which serves as the basis for eliminating untrustworthy cooperative optimization instructions;

[0197] The credibility verification index has the following specific process:

[0198] ; wherein, is a credibility verification index, is an industrial park collected data completeness ratio, is an industrial park collected data error rate, is an industrial park data transmission packet loss rate;

[0199] Obtain the credibility verification index stored in the database - the collaborative optimization credibility update value mapping set, based on the credibility verification index, find the matching collaborative optimization credibility update value.

[0200] Add the collaborative optimization credibility update value to the collaborative optimization credibility to obtain the collaborative optimization credibility feature, and compare the collaborative optimization credibility feature with the collaborative optimization credibility threshold stored in the database; if the collaborative optimization credibility feature is not lower than the collaborative optimization credibility threshold, the collaborative optimization instruction corresponding to the collaborative optimization credibility feature is credible; if the collaborative optimization credibility feature is lower than the collaborative optimization credibility threshold, the collaborative optimization instruction corresponding to the collaborative optimization credibility feature is not credible, and the untrusted collaborative optimization instruction is removed; based on the remaining trusted collaborative optimization instruction, the industrial park source load storage is adjusted.

[0201] Obtain the industrial park collected data completeness ratio, error rate, transmission packet loss rate and other features, and evaluate the data quality from multiple dimensions of data integrity, accuracy and transmission reliability. Fully understand the data situation, and provide rich information basis for instruction credibility analysis.

[0202] Calculate the credibility verification index, which comprehensively considers the features of each dimension of data. Based on the index, find the corresponding update value from the credibility verification index - collaborative optimization credibility update value mapping set, and add it to the original credibility to obtain the credibility feature. Scientifically quantify the instruction credibility, provide an objective basis for removing untrusted instructions, and ensure the reliability of the remaining instructions.

[0203] According to the comparison result of the credibility feature and the credibility threshold, accurately remove the untrusted collaborative optimization instruction, and only keep the reliable instruction. Avoid unreliable instructions from interfering with or incorrectly guiding the operation of the source load storage system, improve the instruction execution effect, and ensure the stability and efficiency of the system operation.

[0204] Based on the remaining trusted instructions, the industrial park source load storage is adjusted, so that the source load storage equipment operates reasonably according to the reliable instructions. Realize the effective collaboration of power supply, load and energy storage resources, improve the energy utilization efficiency, reduce the operation cost, and promote the efficient and economic operation of the industrial park energy system.

[0205] Embodiment 2: as Figure 2 , Figure 3As shown, the embodiment provides an industrial park source-load-storage collaborative optimization method for flexibility improvement, based on the system in embodiment 1, including the following steps: collecting industrial park source-load-storage data, dynamic resource portrait modeling, and outputting dynamic device performance set.

[0206] Obtain power grid constraints and market price signals, establish constraint conditions and calculate dynamic cost functions in combination with the dynamic device performance set, perform spatio-temporal elasticity evaluation, and output the elastic boundary surface through dynamic game solution.

[0207] Consider the elastic boundary surface and collect real-time monitoring data of the industrial park, and obtain the optimal strategy vector through quantum collaborative optimization.

[0208] Based on the optimal strategy vector, perform virtual-real linkage verification to obtain a collaborative optimization verification instruction set and a collaborative optimization credibility label.

[0209] Obtain and analyze the completeness characteristics of the industrial park collected data and the data transmission reliability characteristics of the industrial park, eliminate untrustworthy collaborative optimization instructions in combination with the collaborative optimization credibility label, and perform collaborative optimization adjustment on the source-load-storage of the industrial park based on the remaining trustworthy collaborative optimization instructions.

Claims

1. A source-load-storage collaborative optimization system for industrial parks aimed at improving flexibility, characterized in that: include: The dynamic resource profiling and modeling module is used to collect source, load, and storage data from industrial parks, create dynamic resource profiling models, and output a dynamic set of equipment performance data. The specific process of collecting source, load, and storage data from industrial parks and creating dynamic resource profiles is as follows: Collect source-load-storage data for industrial parks, including photovoltaic output, ambient temperature, SOC of energy storage, rated power of equipment, actual power of equipment, rated efficiency of equipment, installed area of ​​photovoltaic panels, and irradiance. Dynamic resource profiling modeling is performed, and the modeling method is as follows: The formula for the Long Short-Term Memory (LSTM) network is defined as follows: ； In the formula, In hidden state, To generate photovoltaic power in the industrial park at time t, Let be the ambient temperature of the industrial park at time t. The SOC of energy storage in the industrial park at time t; For energy storage SOC: Particle swarm optimization is used to screen features and maximize the Particle swarm optimization objective function, where the Particle swarm optimization objective function is: ； In the formula, The efficiency of the energy storage system during charging and discharging operations at time t is given. Here, e is the time decay factor, and e is the natural constant. Normalization of actual power of equipment in industrial parks: ； In the formula, The normalized power of the industrial park equipment at time t. The actual power of the equipment in the industrial park at time t. Rated power of equipment in the industrial park; Establish an efficiency decay model, combining the effects of aging and load, to calculate the real-time efficiency of equipment in the industrial park: ； In the formula, This refers to the real-time efficiency of the equipment in the industrial park at time t. Rated efficiency for equipment in the industrial park. The aging factors are stored in the database. This is the first correction factor for efficiency caused by load fluctuations. This is the second correction factor for efficiency caused by load fluctuations. This represents the historical average load of equipment in the industrial park. Adjustments to photovoltaic output: Based on the industrial park's irradiance and ambient temperature, the weather bias of historical data is corrected: ； In the formula, Let A represent the irradiance of the industrial park, and A represent the area of ​​photovoltaic panels installed in the industrial park. Photoelectric conversion efficiency; The spatiotemporal elasticity assessment and solution module is used to acquire power grid constraints and market price signals, combine dynamic equipment efficiency sets to establish constraint conditions and calculate dynamic cost functions, perform spatiotemporal elasticity assessment, and output elastic boundary surfaces through dynamic game theory solutions. The quantum collaborative optimization module is used to consider elastic boundary surfaces and collect real-time monitoring data of industrial parks. Quantum collaborative optimization yields the optimal strategy vector. The virtual-real linkage verification module is used to perform virtual-real linkage verification based on the optimal strategy vector, and obtain the collaborative optimization verification instruction set and collaborative optimization credibility label. The collaborative optimization instruction update module is used to acquire and analyze the integrity characteristics of the data collected in the industrial park and the reliability characteristics of the data transmission in the industrial park. It combines the collaborative optimization credibility label to remove untrusted collaborative optimization instructions and performs collaborative optimization adjustment on the source, load and storage of the industrial park based on the remaining trustworthy collaborative optimization instructions.

2. The industrial park source-load-storage collaborative optimization system for improving flexibility as described in claim 1, characterized in that: The dynamic resource profiling modeling module outputs a dynamic device performance set, and the specific process is as follows: Obtain maximum power data of equipment in industrial parks Industrial park equipment response time data Industrial park equipment maintenance cost data ; Equipment health index is analyzed based on power data from the past 24 hours, including the actual power and efficiency of equipment in the industrial park. ； In the formula, The health index of equipment in the industrial park at time k. The actual power of the equipment in the industrial park at time k. Let k be the energy conversion efficiency of the equipment in the industrial park at time k, which is the ratio of output energy to input energy. Will , , and Combined into a dynamic device performance set .

3. The industrial park source-load-storage collaborative optimization system for improving flexibility as described in claim 1, characterized in that: In the spatiotemporal elasticity assessment and solution module, grid constraints and market price signals are acquired, and constraints are established by combining dynamic equipment efficiency sets and dynamic cost functions are calculated. The specific process is as follows: Response time constraints, based on equipment response time data from the industrial park. Extracting the energy storage response time of industrial parks and the response time of gas turbines in industrial parks : ； In the formula, This refers to the response time for scheduling instructions. Data on equipment maintenance costs in industrial parks Extracting the highest maintenance cost of each piece of equipment in the industrial park The maintenance cost constraint is the maintenance cost of the equipment under the dispatch command. Not exceeding the maximum maintenance cost of the corresponding industrial park equipment ; Obtain the grid constraints and voltage limits, and transform them into quadratic programming conditions: ； In the formula, The reference voltage, For the first Line impedance at each node, For the first The injection power of each node, where n represents the number of nodes participating in the calculation; Define the feasible power region: ; In the formula, This is the minimum power value that the line is allowed to transmit. This represents the actual power transmitted through the line. This represents the maximum power that the line is allowed to transmit. To obtain market price signals, the dynamic cost function is: ； In the formula, The dynamic cost at time t, Peak electricity price, The actual power transmitted by the line at time t. For time-of-use electricity pricing, The starting time for calculating time-of-use electricity pricing. This is the end time for calculating time-of-use electricity prices.

4. The industrial park source-load-storage collaborative optimization system for improving flexibility as described in claim 3, characterized in that: In the spatiotemporal elasticity assessment and solution module, spatiotemporal elasticity assessment is performed, and the elastic boundary surface is output through dynamic game theory solution. The specific process is as follows: Calculation of elasticity margin: ； Among them, source-side regulation capability The calculation method is as follows: ; Load elasticity The calculation method is as follows: ; In the formula, This is a flexibility assessment metric, where T represents the total time period and m represents the number of source-side devices. Let be the maximum power output value of the s-th source-side device. Let be the minimum power output value of the s-th source-side device. For the number of load sides, Let l be the power change of the l-th load. Here, e is the time decay factor, and e is the natural constant. For the efficiency of the s-th source side Its power The partial derivatives, The reciprocal of the response time, For load transfer costs, Let t be the maximum power of the l-th load, and t be the time index; Dynamic game theory is used to solve for Nash equilibrium, with the source, load, and storage parties reaching the optimal strategy through virtual iteration: ； In the formula, Let f be the policy variable of the f-th party in the (K+1)-th iteration. Let f be the policy variable of the f-th party. Let f be the strategy combination of all parties except f in the k-th iteration. Let f be the utility function of the f-th power; The convergence condition is that the rate of change of the strategy is less than 1%; Output elastic boundary surface ; The time boundary is defined as the allowable adjustment time window for solving the update. This represents the power boundary, i.e., solving for the updated maximum schedulable power range. The cost boundary is the solution to the updated cost constraints.

5. The industrial park source-load-storage collaborative optimization system for improving flexibility as described in claim 1, characterized in that: In the quantum collaborative optimization module, considering the elastic boundary surface and collecting real-time monitoring data from the industrial park, the quantum collaborative optimization obtains the optimal strategy vector. The specific process is as follows: Input an elastic boundary surface and map the optimal solution set to quantum state amplitudes. This provides an initial population for quantum collaborative optimization of quantum chromosomes. ； In the formula, Let the power of the j-th strategy be the elastic boundary surface. Let be the total power of the elastic boundary surface strategy, j be the index of the elastic boundary surface strategy, and J be the number of elastic boundary surface strategies. Real-time monitoring data to establish energy storage SOC constraints: ； In the formula, This represents the highest power output of equipment in the industrial park. The rated power of the equipment in the industrial park is denoted as SOC, which represents the state of charge of the equipment in the industrial park. Quantum cooperative optimization: For chromosome encoding, the rotation angle formula for the dynamic rotation door strategy is: ； In the formula, For the current iteration number The rotation angle at that time, The maximum number of iterations, For the first The optimal fitness of the next iteration The optimal fitness for the 0th iteration; Fitness assessment: ； In the formula, For fitness assessment values, Power generation capacity of the equipment, To reduce the maintenance costs of the dispatching command equipment, As a flexibility assessment indicator, It is the hyperbolic tangent function. This represents the change in carbon dioxide emissions. This serves as a baseline value for carbon dioxide emissions; The policy corresponding to the highest fitness evaluation value is recorded as the optimal policy, and the optimal policy vector is output.

6. The industrial park source-load-storage collaborative optimization system for improving flexibility according to claim 1, characterized in that: In the virtual-real linkage verification module, virtual-real linkage verification is performed based on the optimal strategy vector to obtain a collaborative optimization verification instruction set and a collaborative optimization credibility label. The specific process is as follows: The optimal strategy vector is decomposed into instructions, and a digital twin is built for verification, including photovoltaic and energy storage models: The photovoltaic model is a single diode model, and the energy storage model is an RC equivalent circuit. Deviation calculation is performed by comparing virtual and real data: ； In the formula, The overall bias is given by Q, where Q is the sample size. This represents the actual measurement value of the q-th sample. Let be the virtual model prediction value for the q-th sample. Rated power; If the overall deviation exceeds the deviation threshold stored in the database, a correction is triggered. The sliding window least squares method is used to correct parameters, ensuring that the digital twin is consistent with the physical entity. ； In the formula, This is the updated model parameter vector. For model parameter vectors, This is the actual measured value at time t. To determine the model parameters at time t The model's predicted values; After correction, the overall deviation is recalculated, the overall deviation-cooperative optimization credibility mapping set stored in the database is obtained, and the matching cooperative optimization credibility is found based on the overall deviation, which is recorded as the cooperative optimization credibility label. Output the collaborative optimization verification instruction set and the corresponding collaborative optimization credibility label.

7. The industrial park source-load-storage collaborative optimization system for improving flexibility as described in claim 1, characterized in that: In the collaborative optimization instruction update module, the integrity characteristics of the data collected from the industrial park and the reliability characteristics of the data transmission from the industrial park are obtained and analyzed. The specific process is as follows: Obtain the integrity characteristics of data collected from industrial parks and the reliability characteristics of data transmission from industrial parks, including the integrity ratio of data collected from industrial parks, the error rate of data collected from industrial parks, and the packet loss rate of data transmission from industrial parks. The integrity characteristics of data collected from industrial parks and the reliability characteristics of data transmission in industrial parks are analyzed to obtain a credibility verification index. The credibility verification index serves as the basis for eliminating untrusted collaborative optimization instructions. Credibility verification metrics, the specific process is as follows: ； In the formula, As a credibility verification indicator, The percentage of complete data collected for the industrial park. To reduce the error rate of data collected from industrial parks, The packet loss rate for data transmission in the industrial park is given by denoted as e, where e is a natural constant. Retrieve the credibility verification index-collaborative optimization credibility update value mapping set stored in the database, and find the matching collaborative optimization credibility update value based on the credibility verification index.

8. The industrial park source-load-storage collaborative optimization system for improving flexibility according to claim 7, characterized in that: In the collaborative optimization instruction update module, untrusted collaborative optimization instructions are removed by combining the collaborative optimization trustworthiness label, and the source-load-storage of the industrial park is adjusted for collaborative optimization based on the remaining trustworthy collaborative optimization instructions. The specific process is as follows: The collaborative optimization credibility update value is accumulated with the collaborative optimization credibility to obtain the collaborative optimization credibility feature. The collaborative optimization credibility feature is then compared with the collaborative optimization credibility threshold stored in the database. If the collaborative optimization trust feature is not lower than the collaborative optimization trust threshold, then the collaborative optimization instruction corresponding to the collaborative optimization trust feature is trustworthy. If the trustworthy feature of collaborative optimization is lower than the trustworthy threshold of collaborative optimization, then the collaborative optimization instruction corresponding to the trustworthy feature is untrustworthy, and the untrustworthy collaborative optimization instruction is removed. The source, load, and storage of the industrial park are adjusted collaboratively based on the remaining trusted collaborative optimization instructions.

9. A source-load-storage collaborative optimization method for industrial parks aimed at improving flexibility, based on the system described in any one of claims 1-8, characterized in that, Includes the following steps: Collect source, load and storage data of industrial parks, create dynamic resource profiles and models, and output dynamic equipment efficiency sets; Obtain power grid constraints and market price signals, combine dynamic equipment efficiency sets to establish constraints and calculate dynamic cost functions, perform spatiotemporal elasticity assessment, and output elastic boundary surfaces through dynamic game theory solutions. Considering the elastic boundary surface and collecting real-time monitoring data from the industrial park, the optimal strategy vector is obtained through quantum collaborative optimization. Based on the optimal strategy vector, virtual-real linkage verification is performed to obtain a collaborative optimization verification instruction set and a collaborative optimization credibility label. The integrity characteristics of data collected from industrial parks and the reliability characteristics of data transmission in industrial parks are obtained and analyzed. Untrusted collaborative optimization instructions are eliminated by combining the trustworthiness label of collaborative optimization. Based on the remaining trustworthy collaborative optimization instructions, the source, load and storage of industrial parks are adjusted for collaborative optimization.

Images