Method and system for operating a modular energy storage shelter

Abstract

The application provides a kind of modular energy storage shelter operation control method and system, wherein, the comprehensive operation data is obtained by collecting the operation parameters of each energy storage module in the embodiment of the application, the health state evaluation model of energy storage module is established by machine learning algorithm, real-time power demand data is obtained from grid company API interface, user end intelligent electric meter and historical data and current environmental parameter analysis, the charging and discharging strategy of each energy storage module is dynamically adjusted by using genetic algorithm to obtain charging and discharging plan, the prediction result is generated by combining deep learning technology to predict the change trend of power supply and demand, the optimal energy storage module scheduling scheme is confirmed by using mixed integer linear programming algorithm to obtain scheduling instruction, the decision-making process is continuously optimized by online learning, and the best control instruction is generated by using the adaptive control strategy of reinforcement learning.The technical scheme provided by the application effectively reduces energy waste, prolongs service life, reduces maintenance and replacement, thereby reducing operating cost and improving economic benefit.

Description

TECHNICAL FIELD

[0001] The embodiment of the present application relates to the technical field of modular energy storage shelter, and particularly relates to a running control method of modular energy storage shelter. BACKGROUND

[0002] With the wide application of renewable energy and the continuous development of power systems, modular energy storage shelters are widely used in the fields of power grid peak shaving, smoothing renewable energy output, standby power supply, etc. as an efficient energy storage solution. The existing control method relies on traditional scheduling strategies and simple optimization algorithms, including real-time acquisition of operating parameters, simple assessment of health status, prediction of power demand based on historical data, generation of fixed scheduling scheme and execution of charge and discharge operation. Although these methods can meet the basic needs, there are problems such as inaccurate health status assessment, low power demand prediction accuracy, single scheduling scheme and lack of flexibility, and lack of adaptive control capability. Therefore, a more intelligent and efficient control method is needed. SUMMARY

[0003] The embodiment of the present application provides a running control method and system of modular energy storage shelter, to solve the problems of inaccurate health status assessment, low power demand prediction accuracy, single scheduling scheme and lack of flexibility, and lack of adaptive control capability in the prior art.

[0004] In the first aspect, the embodiment of the present application provides a running control method of modular energy storage shelter, comprising:

[0005] Real-time acquisition of operating parameters of each energy storage module to obtain comprehensive operating data, wherein the operating parameters include battery temperature, state of charge, internal resistance, voltage, current and external environmental parameters;

[0006] According to the comprehensive operating data, an energy storage module health status assessment model is established by a machine learning algorithm to obtain a health index, and real-time power demand data is obtained from a power grid company API interface, a user end smart meter and historical data and current environmental parameter analysis;

[0007] Based on the health index and the real-time power demand data, a genetic algorithm is used to dynamically adjust the charge and discharge strategy of each energy storage module to obtain a charge and discharge plan, and a power supply and demand change trend in a preset time range is predicted using the charge and discharge plan combined with deep learning technology to generate a power demand prediction result;

[0008] Based on the power demand prediction result, a mixed integer linear programming algorithm is used to confirm an optimal energy storage module scheduling scheme to obtain a scheduling instruction of the energy storage module;

[0009] Implementing the scheduling instruction of the energy storage module, continuously optimizing the decision-making process through online learning, using the adaptive control strategy of reinforcement learning to generate the best control instruction.

[0010] Optionally, characterized in that, based on the power demand prediction result, the optimal energy storage module scheduling scheme is confirmed by using a mixed integer linear programming algorithm, and the scheduling instruction of the energy storage module is obtained, including:

[0011] Using the power demand prediction result, combining the current health status of the energy storage shelter, the energy storage capacity and the charge and discharge rate limit, a multi-objective optimization model is constructed;

[0012] Using the multi-objective optimization model, a plurality of preset target internal scenes are generated by Monte Carlo simulation to obtain scheduling scheme risk assessment results under a plurality of scenes;

[0013] Based on the scheduling scheme risk assessment results, the multi-objective optimization model is solved by using a mixed integer linear programming algorithm to generate an initial energy storage module scheduling scheme;

[0014] Using the initial energy storage module scheduling scheme, a comprehensive evaluation is performed by using an analytic hierarchy process to determine the best energy storage module scheduling scheme.

[0015] Optionally, characterized in that, based on the scheduling scheme risk assessment results, the multi-objective optimization model is solved by using a mixed integer linear programming algorithm to generate an initial energy storage module scheduling scheme, including:

[0016] Based on the scheduling scheme risk assessment results, the physical limit conditions and operation rules of the energy storage shelter are integrated to obtain a multi-objective optimization model under the constraint conditions;

[0017] Using the multi-objective optimization model, a phased optimization calculation is performed on the scheduling scheme by introducing a time window rolling optimization mechanism to obtain a phased scheduling scheme;

[0018] According to the phased scheduling scheme, a mixed integer linear programming algorithm is used to combine historical operation data and real-time operation status of the energy storage shelter to perform multi-objective optimization solving to generate a preliminary energy storage module scheduling scheme;

[0019] Based on comprehensive operation data, an energy storage module health status evaluation model is established by using a machine learning algorithm to obtain a health status evaluation result;

[0020] Using the preliminary energy storage module scheduling scheme in combination with the health status evaluation result, the preliminary energy storage module scheduling scheme is corrected to obtain a corrected energy storage module scheduling scheme;

[0021] The initial energy storage module scheduling scheme is generated by using the modified energy storage module scheduling scheme, simulating and verifying the scheduling scheme, and adjusting the unfeasible scheduling instructions.

[0022] Optionally, the multi-objective optimization model is used to perform stage-by-stage optimization calculation on the scheduling scheme by introducing a time window rolling optimization mechanism to obtain a stage-by-stage scheduling scheme, including:

[0023] The multi-objective optimization model is used in combination with the real-time running state of the energy storage shelter and the power demand prediction results within a preset time to define a series of time windows to obtain a time window sequence.

[0024] The time window sequence is used to perform local optimization calculation on the scheduling scheme in each time window according to the current health state of the energy storage shelter, the energy storage capacity, the charge and discharge rate limit, and the external environmental parameters to obtain a local optimal scheduling scheme for each time window.

[0025] The local optimal scheduling scheme for each time window is integrated to form a stage-by-stage scheduling scheme covering the entire preset time range.

[0026] The stage-by-stage scheduling scheme is used in combination with the historical running data of the energy storage shelter to predict the performance change trend of each energy storage module in each time window by a machine learning algorithm to obtain a performance prediction result.

[0027] The performance prediction result is used to dynamically adjust the stage-by-stage scheduling scheme to obtain a dynamic stage-by-stage scheduling scheme.

[0028] The dynamic stage-by-stage scheduling scheme is used in combination with the real-time running data of the energy storage shelter to dynamically adjust the scheduling instructions by a rolling optimization mechanism to generate a target stage-by-stage scheduling scheme.

[0029] Optionally, according to the stage-by-stage scheduling scheme, a mixed integer linear programming algorithm is used in combination with the historical running data and the real-time running state of the energy storage shelter to perform multi-objective optimization solving to generate a preliminary energy storage module scheduling scheme, including:

[0030] The stage-by-stage scheduling scheme is used in combination with the historical running data and the real-time running state of the energy storage shelter to analyze the running characteristics of the energy storage shelter to obtain a running characteristic analysis result.

[0031] The running characteristic analysis result is used to customize the objective function of the mixed integer linear programming algorithm to obtain a customized multi-objective optimization model.

[0032] The customized multi-objective optimization model is used in combination with the time windows in the stage-by-stage scheduling scheme to perform multi-objective optimization solving to generate a preliminary multi-objective optimization result.

[0033] Refining the optimization results based on the preliminary multi-objective optimization results and the health state evaluation results of the energy storage shelter to obtain refined multi-objective optimization results;

[0034] Based on the refined multi-objective optimization results, the preliminary scheduling schemes are comprehensively evaluated by the analytic hierarchy process to obtain the evaluated preliminary scheduling schemes;

[0035] Using the evaluated preliminary scheduling schemes, combined with real-time operation data of the energy storage shelter, dynamic adjustment is performed to generate a preliminary energy storage module scheduling scheme.

[0036] Optionally, based on the health index and the real-time power demand data, a genetic algorithm is used to dynamically adjust the charging and discharging strategies of each energy storage module to obtain a charging and discharging plan, and using the charging and discharging plan, combined with deep learning technology, the power supply and demand trend in a preset time range is predicted to generate a power demand prediction result, including:

[0037] Using the health index and the real-time power demand data, the charging and discharging strategies of each energy storage module are dynamically adjusted by a genetic algorithm to obtain a preliminary charging and discharging plan;

[0038] Using the preliminary charging and discharging plan, combined with the real-time operation state and historical operation data of the energy storage shelter, the power supply and demand trend in a preset time range is predicted by deep learning technology to obtain a preliminary power demand prediction result;

[0039] Using the preliminary power demand prediction result, the preliminary charging and discharging plan is optimized to obtain an optimized charging and discharging plan;

[0040] Using the optimized charging and discharging plan, the power supply and demand trend in a preset time range is predicted again by deep learning technology to generate a target power demand prediction result.

[0041] Optionally, the scheduling instructions of the energy storage module are implemented, the decision-making process is continuously optimized through online learning, the adaptive control strategy of reinforcement learning is used to generate optimal control instructions, including:

[0042] Using the scheduling instructions of the energy storage module to start the charging and discharging operation of the energy storage shelter, and collecting real-time operation data of the energy storage shelter during the execution of the scheduling instructions to obtain real-time operation feedback data;

[0043] According to the real-time operation feedback data, combined with the historical operation data of the energy storage shelter, the current decision-making process is evaluated by an online learning algorithm to obtain a decision-making evaluation result;

[0044] The decision evaluation result is used to design a reinforcement learning environment, define a reward function, construct a reinforcement learning model, and train the reinforcement learning model by simulating different charging and discharging strategies and energy storage module scheduling schemes, so as to obtain an enhanced reinforcement learning model.

[0045] The enhanced reinforcement learning model is used to optimize control instructions of the energy storage module, generate new control instructions, and obtain optimized control instructions.

[0046] The optimized control instructions are used to dynamically adjust the charging and discharging strategy in combination with real-time running states and environmental parameters of the energy storage shelter, and generate optimal control instructions.

[0047] In a second aspect, the embodiments of the present application provide a running control system of a modular energy storage shelter, comprising:

[0048] The acquisition module is configured to acquire running parameters of each energy storage module in real time to obtain comprehensive running data, wherein the running parameters include battery temperature, state of charge, internal resistance, voltage, current, and external environmental parameters.

[0049] The construction module is configured to construct an energy storage module health state evaluation model by a machine learning algorithm according to the comprehensive running data to obtain a health index, and obtain real-time power demand data from an API interface of a power grid company, a user-side smart meter, and historical data and current environmental parameter analysis.

[0050] The prediction module is configured to dynamically adjust charging and discharging strategies of each energy storage module based on the health index and the real-time power demand data by using a genetic algorithm to obtain a charging and discharging plan, and predict power supply and demand change trends within a preset time range by using the charging and discharging plan in combination with deep learning technology to generate a power demand prediction result.

[0051] The confirmation module is configured to confirm an optimal energy storage module scheduling scheme by using a mixed integer linear programming algorithm based on the power demand prediction result to obtain a scheduling instruction of the energy storage module.

[0052] The optimization module is configured to implement the scheduling instruction of the energy storage module, optimize a decision-making process by online learning, and generate optimal control instructions by using an adaptive control strategy of reinforcement learning.

[0053] In a third aspect, the embodiments of the present application provide a computing device comprising a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the running control method of the modular energy storage shelter according to any one of the first aspect.

[0054] In a fourth aspect, the embodiments of the present application provide a computer storage medium, which stores computer program instructions, and the computer program instructions are executed by a processor to implement the operation control method of the modular energy storage shelter according to any one of the first aspect.

[0055] In the embodiments of the present application, the operation parameters of each energy storage module are collected in real time to obtain comprehensive operation data, wherein, according to the comprehensive operation data, an energy storage module health state evaluation model is established through a machine learning algorithm to obtain a health index, and real-time power demand data is obtained from an API interface of a power grid company, a user-side smart meter, historical data and current environmental parameter analysis; based on the health index and the real-time power demand data, a genetic algorithm is used to dynamically adjust the charging and discharging strategy of each energy storage module to obtain a charging and discharging plan, the power supply and demand change trend in a preset time range is predicted by using the charging and discharging plan in combination with deep learning technology to generate a power demand prediction result; based on the power demand prediction result, a mixed integer linear programming algorithm is used to confirm an optimal energy storage module scheduling scheme to obtain a scheduling instruction of the energy storage module; the scheduling instruction of the energy storage module is implemented, the decision-making process is continuously optimized through online learning, the adaptive control strategy of reinforcement learning is used, and the best control instruction is generated. The technical scheme provided by the present application improves the accuracy of health state evaluation, improves the accuracy of power demand prediction, optimizes the charging and discharging strategy, generates the optimal scheduling scheme, realizes adaptive control, enhances the flexibility and expansibility of the system, and finally improves the overall operation efficiency and economic benefit.

[0056] Further, by combining the current health status of the energy storage shelter, the energy storage capacity, and the charge and discharge rate limit, a multi-objective optimization model is constructed to ensure that the scheduling scheme is feasible in actual operation and to avoid infeasible schemes caused by ignoring actual limitations; by generating multiple preset target scenarios through Monte Carlo simulation and performing risk assessment, various possible situations can be considered comprehensively to improve the robustness of the scheduling scheme, so that it can remain stable and efficient when facing uncertainty and changes; the mixed integer linear programming algorithm is used to solve the multi-objective optimization model, which can find the optimal balance point between multiple objectives to ensure that the scheduling scheme meets multiple constraint conditions while achieving the best performance, improving the overall efficiency and economy of the system, and optimizing the performance of the scheduling scheme; the initial energy storage module scheduling scheme is comprehensively evaluated using the analytic hierarchy process, which can systematically weigh the importance of each factor to ensure that the final selected scheduling scheme is optimal in all aspects, improving the scientificity and rationality of decision-making; by comprehensively considering multiple factors and scenarios, the generated scheduling scheme can better cope with various operating conditions and external environmental changes, improving the reliability and stability of the system, and reducing downtime; the optimized scheduling scheme can effectively reduce unnecessary energy waste, prolong the service life of the energy storage module, and reduce maintenance and replacement costs, thereby significantly reducing operating costs and improving economic efficiency.

[0057] These aspects or other aspects of the present application will be more apparent in the following description of the embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0059] Figure 1 A flow chart of a modular energy storage shelter operation control method provided by the embodiment of the present application;

[0060] Figure 2 A structural schematic diagram of a modular energy storage shelter operation control system provided by the embodiment of the present application;

[0061] Figure 3 A structural schematic diagram of a computing device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0062] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0063] In some of the flowcharts described in the specification and claims of the present application and in the above-mentioned drawings, a plurality of operations are included which occur in a particular order, but it should be clearly understood that these operations can be performed in the order in which they appear herein or in parallel, and the serial numbers of the operations, such as 101, 102, etc., are merely used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these flowcharts can include more or fewer operations, and the operations can be performed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. herein are used to distinguish different messages, devices, modules, etc., and do not represent the order of precedence, nor do "first" and "second" represent different types.

[0064] The technical solutions in the embodiments of the present application will be described clearly and completely in the specification of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0065] Since the existing control method relies on traditional scheduling strategies and simple optimization algorithms, including real-time acquisition of operating parameters, simple assessment of health status, prediction of power demand based on historical data, generation of fixed scheduling scheme and execution of charge and discharge operation. These methods can meet the basic needs, but there are problems such as inaccurate health status assessment, low power demand prediction accuracy, single scheduling scheme and lack of flexibility, and lack of adaptive control capability. Based on this, the present application provides a running control method for modular energy storage shelter, as shown in Figure 1 , including:

[0066] Step 101: Real-time acquisition of operating parameters of each energy storage module to obtain comprehensive operating data, wherein the operating parameters include battery temperature, state of charge, internal resistance, voltage, current and external environmental parameters;

[0067] In this step, through sensors and monitoring devices, the key operating parameters of each energy storage module are collected in real time, including battery temperature, state of charge (SOC), internal resistance, voltage, current and external environmental parameters (such as temperature, humidity, etc.);

[0068] The collected various operating parameters are integrated to form comprehensive operating data. These data reflect the operating conditions of the energy storage module at the current time and the external environmental conditions.

[0069] Step 102: According to the comprehensive operation data, a health state evaluation model of the energy storage module is established by a machine learning algorithm to obtain a health index, and real-time power demand data is obtained from the API interface of the power grid company, the user-side smart meter, and historical data and current environmental parameter analysis;

[0070] In this step, a health state evaluation model is trained based on comprehensive operation data using a machine learning algorithm such as support vector machine, random forest, neural network, etc. The model can predict the health state of the energy storage module according to the input operation parameters and generate a health index;

[0071] Real-time power demand data is obtained from the API interface of the power grid company, combined with data from the user-side smart meter and historical data and current environmental parameters (such as weather forecasts, holiday effects, etc.), and data analysis is performed to generate real-time power demand data.

[0072] Step 103: Based on the health index and the real-time power demand data, a genetic algorithm is used to dynamically adjust the charging and discharging strategies of each energy storage module to obtain a charging and discharging plan. Using the charging and discharging plan, deep learning technology is used to predict the power supply and demand trend within a preset time range to generate a power demand prediction result;

[0073] In this step, based on the health index and real-time power demand data, a genetic algorithm is used to dynamically adjust the charging and discharging strategies of each energy storage module. The genetic algorithm simulates natural selection and genetic mechanisms to optimize the charging and discharging strategies to achieve the best energy utilization efficiency;

[0074] According to the optimized charging and discharging strategy, a specific charging and discharging plan is generated;

[0075] Using deep learning technology (such as recurrent neural network RNN, long short-term memory network LSTM, etc.), the power supply and demand trend within a preset time range is predicted. Combined with the generated charging and discharging plan, a high-precision power demand prediction result is generated.

[0076] Step 104: Based on the power demand prediction result, a mixed integer linear programming algorithm is used to confirm the optimal energy storage module scheduling scheme to obtain the scheduling instructions of the energy storage module;

[0077] In this step, based on the power demand prediction result, a multi-objective optimization model is constructed combining the current health state of the energy storage shelter, the energy storage capacity, and the charging and discharging rate limit;

[0078] The mixed integer linear programming algorithm (MILP) is used to solve the multi-objective optimization model to generate an initial energy storage module scheduling scheme. MILP can find the optimal balance point between multiple objectives to ensure that the scheduling scheme meets multiple constraint conditions while achieving the best performance.

[0079] According to the solving result, a specific energy storage module scheduling instruction is generated to guide the charging and discharging operation of the energy storage module.

[0080] Step 105: Implement the scheduling instruction of the energy storage module, continuously optimize the decision-making process through online learning, and generate the optimal control instruction by using the adaptive control strategy of reinforcement learning;

[0081] In this step, the charging and discharging operation of the energy storage module is performed according to the generated scheduling instruction.

[0082] Through the online learning method (such as recurrent neural network, incremental learning, etc.), actual operation data is continuously collected, model parameters are updated, and the decision-making process is optimized.

[0083] By using the adaptive control strategy of reinforcement learning (such as Q-learning, DQN, etc.), the control strategy is dynamically adjusted according to the real-time operation data and environmental changes, and the optimal control instruction is generated. Reinforcement learning continuously learns and optimizes the control strategy through interaction with the environment, improves the adaptive ability and overall performance of the system.

[0084] Based on this, the application provides a specific embodiment, and the step 104 comprises the following steps:

[0085] Step 201: Use the power demand prediction result, combine the current health status of the energy storage shelter, the energy storage capacity, and the charging and discharging rate limit, and construct a multi-objective optimization model.

[0086] In this step, the power demand prediction result refers to the high-precision power demand prediction result obtained in step 103; the current health status of the energy storage shelter refers to the health index obtained in step 102; the energy storage capacity refers to the current available capacity of each energy storage module; and the charging and discharging rate limit refers to the maximum charging and discharging rate of each energy storage module.

[0087] Define multiple optimization objectives, such as minimizing operating cost, maximizing system stability, and minimizing energy storage module loss; constraint conditions include energy storage capacity limit, charging and discharging rate limit, and health status limit; model formula refers to converting the above objectives and constraint conditions into mathematical expressions to form a multi-objective optimization model; generate multiple preset target internal scenes through Monte Carlo simulation, each scene represents different operating conditions and external environmental changes; and the scheduling scheme risk assessment result under each scene refers to calculating the risk value (such as cost, stability, and loss) of the scheduling scheme under each scene.

[0088] Step 202: Using the multi-objective optimization model, generate multiple preset target scenarios through Monte Carlo simulation to obtain risk assessment results of scheduling schemes under multiple scenarios;

[0089] In this step, multiple possible operating scenarios are generated through random sampling methods, considering different power demands, weather conditions, holiday effects, and other factors. Under each scenario, the risk assessment results of the scheduling scheme are calculated according to the multi-objective optimization model;

[0090] The performance of the scheduling scheme under each scenario is calculated in terms of different objectives, such as operating cost, system stability, and energy storage module wear and tear. A risk assessment report is generated, including risk values and corresponding scheduling schemes under each scenario.

[0091] Step 203: Based on the risk assessment results of the scheduling scheme, use a mixed integer linear programming algorithm to solve the multi-objective optimization model and generate an initial energy storage module scheduling scheme;

[0092] In this step, according to the risk assessment results in step 202, select scenarios with lower risk and better performance as references, and input the data of these scenarios into the multi-objective optimization model;

[0093] Use the mixed integer linear programming (MILP) algorithm to solve the multi-objective optimization model. The MILP algorithm can handle multiple objectives and complex constraints, find the optimal scheduling scheme, and generate an initial energy storage module scheduling scheme, including the charging and discharging instructions of each energy storage module at different time points;

[0094] The initial energy storage module scheduling scheme includes the charging and discharging plan of each energy storage module and the corresponding control parameters.

[0095] Step 204: Use the initial energy storage module scheduling scheme to conduct comprehensive evaluation through the analytic hierarchy process (AHP) to determine the best energy storage module scheduling scheme;

[0096] In this step, use the analytic hierarchy process (AHP) to define evaluation criteria such as cost, stability, reliability, and efficiency, and sort the evaluation criteria according to importance to form a hierarchical structure diagram. Through expert scoring or data analysis, assign weights to each criterion;

[0097] Evaluate the initial energy storage module scheduling scheme, calculate the score of each scheme under each criterion, and calculate the comprehensive score of each scheme according to the hierarchical structure and weights;

[0098] Select the scheduling scheme with the highest comprehensive score as the final best energy storage module scheduling scheme; output the best scheduling scheme, including detailed charging and discharging plans and control instructions.

[0099] Based on this, the application provides a specific embodiment, wherein the step 203 comprises the following steps:

[0100] Step 301: Based on the scheduling scheme risk assessment result, the physical limitation conditions and operation rules of the energy storage shelter are integrated to obtain a multi-objective optimization model under constraints.

[0101] In this step, the physical limitation conditions are integrated to determine the physical limitation conditions of the energy storage shelter, such as the maximum charge / discharge rate, upper and lower limits of the energy storage capacity, battery temperature range, etc. These physical limitation conditions are integrated into the optimization model to ensure that the generated scheduling scheme is feasible in actual operation.

[0102] The operation rules are integrated to determine the operation rules of the energy storage shelter, such as the charge / discharge time window, minimum charge / discharge interval, safety operation procedures, etc. These operation rules are also integrated into the optimization model to ensure that the scheduling scheme meets the operation specifications and safety requirements.

[0103] After integrating the physical limitation conditions and operation rules, a multi-objective optimization model is constructed, which considers multiple optimization objectives such as cost minimization, system stability maximization, and energy storage module loss minimization. The model includes all the above constraints to ensure that the generated scheduling scheme meets various limitations while achieving optimal performance.

[0104] Step 302: Using the multi-objective optimization model, a phased optimization calculation is performed on the scheduling scheme by introducing a time window rolling optimization mechanism to obtain a phased scheduling scheme.

[0105] In this step, the entire scheduling period is divided into multiple time windows (e.g., every hour or every half hour). The scheduling scheme within each time window is independently optimized while considering the connection between adjacent time windows.

[0106] In each time window, the multi-objective optimization model is used for optimization calculation to generate the scheduling scheme within that time period. The time window is gradually advanced to continuously update and optimize the scheduling scheme, forming a complete phased scheduling scheme.

[0107] Through the rolling optimization mechanism, a series of phased scheduling schemes are finally generated, covering the entire scheduling period. Each phase of the scheduling scheme considers the actual situation in the current time window and the predicted data in the future, ensuring the continuity and feasibility of the scheduling scheme.

[0108] Step 303: According to the phased scheduling scheme, use the mixed integer linear programming algorithm, combined with the historical operation data and real-time operation state of the energy storage shelter, to perform multi-objective optimization solution, and generate a preliminary energy storage module scheduling scheme;

[0109] In this step, the historical operation data of the energy storage shelter is collected, including the past charging and discharging records, health status, environmental parameters, etc.; the current real-time operation state is obtained, including the current state of charge, battery temperature, external environmental parameters, etc.;

[0110] Use the mixed integer linear programming (MILP) algorithm to further optimize the phased scheduling scheme combined with historical operation data and real-time operation state; MILP algorithm can handle multiple objectives and complex constraints, and find the optimal scheduling scheme;

[0111] According to the optimization result, a preliminary energy storage module scheduling scheme is generated, including the charging and discharging instructions of each energy storage module at different time points; this scheme takes into account historical data and real-time state, ensuring the feasibility and efficiency of the scheduling scheme in actual operation.

[0112] Step 304: Based on the comprehensive operation data, a health status evaluation model of the energy storage module is established through a machine learning algorithm, and the health status evaluation result is obtained;

[0113] In this step, the comprehensive operation data is collected, including battery temperature, state of charge, internal resistance, voltage, current, and external environmental parameters; a health status evaluation model is trained using a machine learning algorithm (such as support vector machine, random forest, neural network, etc.), and the training data includes historical operation data and known health status labels; the current comprehensive operation data is input into the trained health status evaluation model to obtain the health status evaluation result of each energy storage module.

[0114] Step 305: Use the preliminary energy storage module scheduling scheme combined with the health status evaluation result to modify the preliminary energy storage module scheduling scheme, and obtain the modified energy storage module scheduling scheme;

[0115] In this step, the preliminary energy storage module scheduling scheme is combined with the health status evaluation result to modify the scheduling scheme, if the health status of some energy storage modules is not good, the charging and discharging load of these modules is reduced, or the charging and discharging strategy of other energy storage modules is adjusted, to ensure the stability and reliability of the overall system; after modification, the modified energy storage module scheduling scheme is generated, ensuring that the health status requirements are met while achieving the best scheduling effect.

[0116] Step 306: Using the modified energy storage module scheduling scheme, verify the scheduling scheme through simulation technology, adjust the unfeasible scheduling instructions, and generate an initial energy storage module scheduling scheme;

[0117] In this step, the simulation test is performed on the modified energy storage module scheduling scheme using simulation technology (such as MATLAB / Simulink, Python simulation library, etc.). During the simulation process, various possible operating conditions and external environmental changes are simulated. During the simulation process, the feasibility of the scheduling scheme is checked, and unfeasible scheduling instructions are identified and adjusted. For example, if the charge and discharge rate of a certain energy storage module exceeds its physical limit in a certain time period, the charge and discharge plan for that time period needs to be adjusted. After simulation verification and adjustment, an initial energy storage module scheduling scheme is generated, which has high feasibility and reliability in actual operation.

[0118] Based on this, the present application provides a specific embodiment, wherein step 302 utilizes the multi-objective optimization model to perform phased optimization calculation on the scheduling scheme by introducing a time window rolling optimization mechanism to obtain a phased scheduling scheme, which specifically includes the following steps:

[0119] Step 401: Using the multi-objective optimization model, combining the real-time operating state of the energy storage shelter and the power demand prediction results within a preset time, a series of time windows are defined to obtain a time window sequence;

[0120] In this step, the entire time range is divided into a series of time windows according to the preset time range (such as 24 hours). Each time window can be fixed (such as every hour or every half hour). The current operating state of the energy storage shelter is considered, including battery temperature, state of charge, internal resistance, voltage, current, etc. The power demand prediction results generated in step 103 are used to provide power demand data for each time window. According to the above information, a series of time windows are defined to form a time window sequence. Each time window has its specific power demand and operating state.

[0121] Step 402: Using the time window sequence, according to the current health state of the energy storage shelter, the energy storage capacity, the charge and discharge rate limit, and the external environmental parameters, the scheduling scheme in each time window is locally optimized and calculated to obtain a locally optimal scheduling scheme for each time window;

[0122] In this step, local optimization refers to performing local optimization calculation for each time window using a multi-objective optimization model; considering health status refers to considering the current health status of the energy storage shelter to ensure that the scheduling scheme does not cause excessive wear of the energy storage module; considering energy storage capacity refers to ensuring that the scheduling scheme within each time window is within the energy storage capacity range; considering charge and discharge rate limit refers to ensuring that the scheduling scheme meets the maximum charge and discharge rate limit of the energy storage module; considering external environmental parameters refers to combining external environmental parameters (such as temperature, humidity, etc.) to ensure the feasibility of the scheduling scheme under actual operating conditions; generating a locally optimal scheme refers to generating a locally optimal scheduling scheme for each time window through local optimization calculation;

[0123] wherein the optimal charge and discharge power Opt ima l_Charge(t) in the time window t is calculated by the following formula:

[0124]

[0125] wherein Opt ima l_Charge(t) is the optimal charge and discharge power; P demand (t) is the power demand prediction value in the time window t, H(t) is the energy storage module health status evaluation value in the time window t, E(t) is the external environmental parameter in the time window t, T(t) is the current temperature in the time window t, C(t) is the charge and discharge cost in the time window t, D(t) is the power grid stability index in the time window t, F(t) is the future power demand fluctuation prediction in the time window t, and α, β, γ, λ, δ, η, θ, φ are weight coefficients.

[0126] Step 403: Utilize the locally optimal scheduling scheme of each time window to perform integration processing to form a phased scheduling scheme covering the entire preset time range;

[0127] In this step, the locally optimal scheduling scheme of each time window is integrated to form a phased scheduling scheme covering the entire preset time range; ensure smooth transition between adjacent time windows to avoid system instability caused by sudden changes; finally generate a complete phased scheduling scheme covering all time windows to ensure that the scheduling plan within the entire preset time range is coherent and consistent.

[0128] Step 404: Utilize the phased scheduling scheme and combine historical operation data of the energy storage shelter to predict the performance change trend of each energy storage module in each time window through a machine learning algorithm to obtain a performance prediction result;

[0129] In this step, historical operation data of the energy storage shelter is collected, including battery temperature, state of charge, internal resistance, voltage, current, etc.; using machine learning algorithms (such as random forest, support vector machine, neural network, etc.), based on historical operation data and current phased scheduling scheme, the performance change trend of each energy storage module in each time window is predicted; the performance prediction results of each energy storage module in different time windows are output, including health status, charging and discharging efficiency, etc.

[0130] Step 405: dynamically adjusting the phased scheduling scheme using the performance prediction results to obtain a dynamic phased scheduling scheme;

[0131] In this step, according to the performance prediction results, the phased scheduling scheme is dynamically adjusted. If it is predicted that the performance of some energy storage modules will decrease in a future time window, the scheduling instructions in the time window are adjusted to reduce the burden of these modules; the adjusted scheduling scheme should ensure that the power demand is met while maintaining the stability and reliability of the system; after dynamic adjustment, a new dynamic phased scheduling scheme is generated to ensure that the best scheduling effect is achieved in each time window.

[0132] Step 406: using the dynamic phased scheduling scheme, combining the real-time operation data of the energy storage shelter, and dynamically adjusting the scheduling instructions through a rolling optimization mechanism to generate a target phased scheduling scheme;

[0133] In this step, the real-time operation data of the energy storage shelter is continuously collected, including battery temperature, state of charge, internal resistance, voltage, current, etc.; using a rolling optimization mechanism, according to the real-time operation data and the dynamic phased scheduling scheme, the time window is gradually advanced, and each time window is optimized and adjusted; in each time window, according to the latest real-time data, the scheduling instructions are dynamically adjusted to ensure that the scheduling scheme is always optimal; after rolling optimization and dynamic adjustment, the final target phased scheduling scheme is generated. This scheme has high feasibility and reliability in actual operation and can cope with changing operating conditions and power demand.

[0134] Based on this, the present application provides a specific embodiment, wherein the step 303, according to the phased scheduling scheme, using a mixed integer linear programming algorithm, combining the historical operation data and real-time operation state of the energy storage shelter, performing multi-objective optimization solution to generate a preliminary energy storage module scheduling scheme, specifically including the following steps:

[0135] Step 501: using the phased scheduling scheme, combining the historical operation data and real-time operation state of the energy storage shelter, analyzing the operation characteristics of the energy storage shelter to obtain operation characteristic analysis results;

[0136] In this step, the phased scheduling scheme refers to a scheduling plan covering the entire preset time range, divided into multiple time windows; the historical operation data refer to the operation records of the energy storage shelter in the past period of time, including battery temperature, state of charge, internal resistance, voltage, current, etc.; the real-time operation state refers to the operation parameters of the current energy storage shelter, reflecting its immediate working condition; the operation characteristic analysis result refers to the performance characteristics and behavior patterns of the energy storage shelter under different conditions obtained through data analysis;

[0137] The data of the phased scheduling scheme, the historical operation data and the real-time operation state are integrated together, and the key operation characteristics of the energy storage shelter such as charging and discharging efficiency, response speed and stability are extracted using data analysis tools such as statistical analysis and cluster analysis, and according to the above analysis, detailed operation characteristic analysis results are generated to provide a basis for subsequent optimization.

[0138] Step 502: Customizing the objective function of the mixed integer linear programming algorithm using the operation characteristic analysis result to obtain a customized multi-objective optimization model;

[0139] In this step, the operation characteristic analysis result refers to the performance characteristics and behavior patterns of the energy storage shelter obtained in step 501; the mixed integer linear programming algorithm refers to an optimization algorithm that can handle discrete variables and continuous variables, suitable for multi-objective optimization problems under complex constraint conditions; the customized multi-objective optimization model refers to an optimization model adjusted according to the actual operation characteristics of the energy storage shelter, which is more in line with actual needs;

[0140] Based on the operation characteristic analysis result, the objective function of the optimization model is redefined, for example, if the analysis result shows that the system stability is particularly important in some time periods, the weight of stability can be increased, and the constraint conditions in the model are adjusted according to the operation characteristics to ensure that the optimization result is more in line with the actual situation. After customization, a new multi-objective optimization model is constructed and prepared for solution.

[0141] Step 503: Using the customized multi-objective optimization model, combining the time windows in the phased scheduling scheme, performing multi-objective optimization solution to generate a preliminary multi-objective optimization result;

[0142] In this step, the customized multi-objective optimization model refers to the optimization model adjusted according to the operation characteristics of the energy storage shelter; the time window refers to the time period divided in the phased scheduling scheme; the preliminary multi-objective optimization result refers to the preliminary scheduling scheme obtained after optimization calculation; the staged optimization: for each time window, use the customized multi-objective optimization model for local optimization calculation.

[0143] The optimization results of each time window are aggregated to form a preliminary multi-objective optimization result covering the entire preset time range, and a preliminary multi-objective optimization result is generated as the basis for subsequent refinement.

[0144] Step 504: Refine the optimization result using the preliminary multi-objective optimization result in combination with the health state assessment result of the energy storage shelter to obtain a refined multi-objective optimization result.

[0145] In this step, the preliminary multi-objective optimization result refers to the preliminary scheduling scheme generated in step 503; the health state assessment result refers to the health state information of the energy storage module obtained through the machine learning algorithm; and the refined multi-objective optimization result refers to the detailed optimization result considering the health state.

[0146] Combining the preliminary multi-objective optimization result with the health state assessment result ensures that the scheduling scheme does not cause excessive burden on modules with poor health states. Based on the health state assessment result, the preliminary multi-objective optimization result is corrected, the charging and discharging strategies of each energy storage module are adjusted, and a refined multi-objective optimization result is generated to ensure that the health state requirements are met while achieving the best scheduling effect.

[0147] Step 505: Based on the refined multi-objective optimization result, a comprehensive evaluation of multiple preliminary scheduling schemes is performed using the analytic hierarchy process to obtain an evaluated preliminary scheduling scheme.

[0148] In this step, the refined multi-objective optimization result refers to the detailed optimization result generated in step 504; the analytic hierarchy process refers to a decision analysis method that ranks the importance of multiple criteria by constructing a hierarchical structure to determine the optimal solution; and the evaluated preliminary scheduling scheme refers to the best scheduling scheme selected after comprehensive evaluation.

[0149] Define evaluation criteria such as cost, stability, reliability, efficiency, etc., and rank them by importance to form a hierarchical structure diagram for comprehensive evaluation of multiple preliminary scheduling schemes. Calculate the score of each scheme under each criterion, and select the scheduling scheme with the highest comprehensive score as the evaluated preliminary scheduling scheme based on the hierarchical structure and weights.

[0150] Step 506: Use the evaluated preliminary scheduling scheme in combination with real-time operation data of the energy storage shelter to perform dynamic adjustment and generate a preliminary energy storage module scheduling scheme.

[0151] In this step, the evaluated preliminary scheduling scheme refers to the best scheduling scheme selected in step 505; the real-time operation data refers to the current operation parameters of the energy storage shelter; and the preliminary energy storage module scheduling scheme refers to the final generated scheduling instructions.

[0152] Real-time operation data of the energy storage shelter is continuously collected, including battery temperature, state of charge, internal resistance, voltage, current, etc., and the preliminary scheduling scheme after evaluation is dynamically adjusted according to the real-time operation data, so that the scheduling scheme is always optimal. After dynamic adjustment, a preliminary energy storage module scheduling scheme is generated to guide the specific charging and discharging operation of the energy storage module.

[0153] The embodiment of the present application improves the operation efficiency, reliability and economy of the energy storage shelter through the above steps, not only solves the defects in the prior art, but also provides an effective solution for intelligent management of the energy storage system, and promotes the effective use of renewable energy and the sustainable development of the power system.

[0154] Based on this, a specific embodiment of the present application is provided, and the step 103 comprises the following steps:

[0155] Step 601: dynamically adjust the charging and discharging strategy of each energy storage module by genetic algorithm based on the health index and real-time power demand data, and obtain a preliminary charging and discharging plan;

[0156] In this step, first, the health index of the energy storage module and the real-time power demand data obtained from the API interface of the power grid company, the intelligent electric meter of the user end and the analysis of historical data and current environmental parameters are combined. Then, the genetic algorithm (GA) is used to dynamically adjust the charging and discharging strategy of each energy storage module. Genetic algorithm optimizes the charging and discharging strategy to achieve the best energy utilization efficiency by simulating natural selection and genetic mechanism. According to the optimized charging and discharging strategy, a specific charging and discharging plan is generated as a preliminary charging and discharging plan.

[0157] Step 602: use the preliminary charging and discharging plan, combine the real-time running state and historical running data of the energy storage shelter, and use deep learning technology to predict the power supply and demand change trend in the preset time range to obtain a preliminary power demand prediction result;

[0158] In this step, based on the preliminary charging and discharging plan, the real-time running state (such as battery temperature, state of charge, internal resistance, voltage, current, etc.) and historical running data of the energy storage shelter are combined, and deep learning technology (such as long short-term memory network LSTM) is used to predict the power supply and demand change trend in the preset time range. Deep learning model can capture the complex change trend of power demand, combine charging and discharging plan, historical data and current environmental parameters, and generate high-precision preliminary power demand prediction result.

[0159] Step 603: Use the preliminary power demand prediction result to optimize the preliminary charging and discharging plan to obtain an optimized charging and discharging plan;

[0160] In this step, based on the preliminary power demand prediction result, the preliminary charging and discharging plan is further optimized. By introducing a multi-objective optimization method, factors such as power demand prediction, energy storage capacity limit, charging and discharging rate limit, and health status are considered comprehensively, and each detail in the charging and discharging plan is adjusted to ensure that the scheduling scheme is more reasonable and efficient. After this round of optimization, the optimized charging and discharging plan is generated, improving the feasibility and adaptability of the plan.

[0161] Step 604: Use the optimized charging and discharging plan to predict the power supply and demand trend in the preset time range again through deep learning technology to generate a target power demand prediction result;

[0162] In this step, based on the optimized charging and discharging plan, deep learning technology is used again to predict the power supply and demand trend in the preset time range. This prediction not only refers to the latest charging and discharging plan, but also combines more accurate power demand prediction results, further improving the accuracy and reliability of the prediction, and finally generating a target power demand prediction result.

[0163] The embodiment of the present application improves the operation efficiency, stability and economy of the modular energy storage shelter through the above steps.

[0164] Based on this, the present application provides a specific embodiment, the step 105, implements the scheduling instruction of the energy storage module, continuously optimizes the decision-making process through online learning, generates the best control instruction by using the adaptive control strategy of reinforcement learning, and specifically includes the following steps:

[0165] Step 701: Start the charging and discharging operation of the energy storage shelter using the scheduling instruction of the energy storage module, and collect real-time operation data of the energy storage shelter during the execution of the scheduling instruction to obtain real-time operation feedback data;

[0166] In this step, according to the optimized control instruction generated in the previous step, the charging and discharging operation of the energy storage shelter is started. The system will monitor and record various operation parameters of the energy storage shelter during the execution of these scheduling instructions in real time, including battery temperature, state of charge (SOC), internal resistance, voltage, current, etc., as well as external environmental parameters such as temperature and humidity. These data will be integrated into real-time operation feedback data for subsequent evaluation and optimization.

[0167] Step 702: According to the real-time operation feedback data, combine the historical operation data of the energy storage shelter, and evaluate the current decision-making process through online learning algorithm to obtain a decision-making evaluation result;

[0168] In this step, based on real-time operation feedback data and historical operation data of the energy storage shelter, an online learning algorithm is used to comprehensively evaluate the current decision-making process. The online learning algorithm can continuously update model parameters to ensure that the evaluation results reflect the latest operation status. The evaluation content includes the execution effect of the scheduling instruction, the response speed of the system, the energy consumption situation, etc., and finally generates a decision evaluation result to provide a basis for the next optimization.

[0169] Step 703: Using the decision evaluation result, design a reinforcement learning environment, define a reward function, build a reinforcement learning model, and train the reinforcement learning model by simulating different charging and discharging strategies and energy storage module scheduling schemes to obtain an enhanced reinforcement learning model;

[0170] In this step, first, according to the decision evaluation result, a reinforcement learning environment suitable for the operation characteristics of the energy storage shelter is designed. This environment defines various possible states, actions, and reward functions, where the reward function aims to encourage the system to adopt more efficient and stable charging and discharging strategies. Then, a reinforcement learning model such as a deep Q network (DQN) or a policy gradient method is constructed. By simulating different charging and discharging strategies and energy storage module scheduling schemes, the model is trained to make optimal decisions in complex and variable environments. After sufficient training, an enhanced reinforcement learning model is generated, which has stronger learning and adaptation capabilities.

[0171] Step 704: Using the enhanced reinforcement learning model, optimize the control instructions of the energy storage module to generate new control instructions, and obtain optimized control instructions;

[0172] In this step, the enhanced reinforcement learning model is used to optimize the existing control instructions. The model predicts the effects of different control instructions based on the current real-time operation status and historical data, and selects the optimal control instruction. These optimized control instructions not only consider the current power demand and health status, but also take into account the long-term operation efficiency and stability, thereby generating more intelligent and efficient control instructions.

[0173] Step 705: Using the optimized control instructions, combined with the real-time operation status and environmental parameters of the energy storage shelter, dynamically adjust the charging and discharging strategy to generate the best control instructions;

[0174] In this step, the optimized control instructions are combined with the real-time operation status and environmental parameters of the energy storage shelter for dynamic adjustment. Through this dynamic adjustment mechanism, it is ensured that each charging and discharging operation is optimized, meeting the immediate power demand while considering the long-term system health and economic benefits. The best control instructions generated finally will guide the operation of the energy storage shelter in the future, ensuring its efficient and stable operation.

[0175] The energy storage shelter can better cope with unexpected situations and changing needs through continuous optimization and dynamic adjustment.

[0176] Figure 2 A structural schematic diagram of an operation control system of a modular energy storage shelter is provided for the embodiments of the present application, as shown in the figure, the system comprises: Figure 2

[0177] The acquisition module 21 is configured to acquire operation parameters of each energy storage module in real time to obtain comprehensive operation data, wherein the operation parameters include battery temperature, state of charge, internal resistance, voltage, current, and external environmental parameters.

[0178] The construction module 22 is configured to construct an energy storage module health state evaluation model by a machine learning algorithm according to the comprehensive operation data to obtain a health index, and obtain real-time power demand data from an API interface of a power grid company, a user-side smart meter, and historical data and current environmental parameter analysis.

[0179] The prediction module 23 is configured to dynamically adjust the charging and discharging strategy of each energy storage module based on the health index and the real-time power demand data by using a genetic algorithm to obtain a charging and discharging plan, and use the charging and discharging plan to predict the power supply and demand change trend in a preset time range by using a deep learning technology to generate a power demand prediction result.

[0180] The confirmation module 24 is configured to confirm an optimal energy storage module scheduling scheme based on the power demand prediction result by using a mixed integer linear programming algorithm to obtain a scheduling instruction of the energy storage module.

[0181] The optimization module 25 is configured to implement the scheduling instruction of the energy storage module, optimize the decision-making process through online learning, and generate an optimal control instruction by using an adaptive control strategy of reinforcement learning.

[0182] Figure 2 The operation control system of the modular energy storage shelter can perform the operation control method of the modular energy storage shelter as shown in the embodiments. Figure 1 The implementation principle and technical effects of the operation control method of the modular energy storage shelter are not described again. The specific operation modes of each module and unit of the operation control system of the modular energy storage shelter in the above embodiments have been described in detail in the embodiments related to the method, and will not be described in detail here.

[0183] Figure 2 The operation control system of the modular energy storage shelter as shown in the embodiments can be implemented as a computing device, as shown in the figure, the computing device can include a storage component 31 and a processing component 32. Figure 3 ​​

[0184] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called for execution by the processing component 32.

[0185] The processing component 32 is configured to collect operation parameters of each energy storage module in real time to obtain comprehensive operation data, wherein the operation parameters include battery temperature, state of charge, internal resistance, voltage, current, and external environmental parameters.

[0186] According to the comprehensive operation data, an energy storage module health state evaluation model is established through a machine learning algorithm to obtain a health index, and real-time power demand data is obtained from a power grid company API interface, a user-side smart meter, and historical data and current environmental parameter analysis.

[0187] Based on the health index and the real-time power demand data, a genetic algorithm is used to dynamically adjust the charging and discharging strategy of each energy storage module to obtain a charging and discharging plan, and the charging and discharging plan is used in combination with deep learning technology to predict the power supply and demand trend in a preset time range to generate a power demand prediction result.

[0188] Based on the power demand prediction result, a mixed integer linear programming algorithm is used to confirm an optimal energy storage module scheduling scheme to obtain a scheduling instruction of the energy storage module.

[0189] The scheduling instruction of the energy storage module is implemented, the decision-making process is continuously optimized through online learning, and the best control instruction is generated by using the adaptive control strategy of reinforcement learning.

[0190] The processing component 32 includes one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component can also be one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), controllers, microcontrollers, microprocessors, or other electronic elements for executing the above method.

[0191] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (RAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0192] The computing device also includes other components, such as an input / output interface, a display component, and a communication component.

[0193] The input / output interface provides an interface between the processing component and peripheral interface modules, which can be output devices, input devices.

[0194] The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, etc.

[0195] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform, and the computing device can be a cloud server, and the processing component, the storage component, etc.

[0196] The embodiment of the present application also provides a computer storage medium, which stores a computer program, and the computer program can realize the above-mentioned Figure 1 The embodiment shown in the figure provides a running control method and system of a modular energy storage shelter.

[0197] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-mentioned system, device and unit can refer to the corresponding processes in the foregoing method embodiments, and will not be described here.

[0198] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment scheme. Those skilled in the art can understand and implement without creative labor.

[0199] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be realized by means of software and necessary general hardware platform, and of course, it can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software products, and the computer software product can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including a plurality of instructions to make a computer device (which can be a personal computer, server, or network device, etc.) execute the method described in each embodiment or some part of the embodiment.

[0200] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; and although the present application has been described in detail with reference to the foregoing embodiments, it should be appreciated by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for operating and controlling a modular energy storage container, characterized in that, The method comprises the following steps: Real-time acquisition of operating parameters of each energy storage module to obtain comprehensive operating data, wherein the operating parameters include battery temperature, state of charge, internal resistance, voltage, current, and external environmental parameters; Based on the comprehensive operating data, a health state evaluation model of the energy storage module is established through a machine learning algorithm to obtain a health index, and real-time power demand data is obtained from a power grid company API interface, a user-side smart meter, and historical data and current environmental parameter analysis; Based on the health index and the real-time power demand data, a genetic algorithm is used to dynamically adjust the charging and discharging strategy of each energy storage module to obtain a charging and discharging plan, and a power supply and demand trend in a preset time range is predicted using the charging and discharging plan combined with deep learning technology to generate a power demand prediction result; Based on the power demand prediction result, a mixed integer linear programming algorithm is used to confirm an optimal energy storage module scheduling scheme to obtain a scheduling instruction of the energy storage module; The scheduling instruction of the energy storage module is implemented, the decision-making process is continuously optimized through online learning, an adaptive control strategy of reinforcement learning is used to generate an optimal control instruction; Based on the power demand prediction result, a mixed integer linear programming algorithm is used to confirm an optimal energy storage module scheduling scheme to obtain a scheduling instruction of the energy storage module, comprising: Using the power demand prediction result, a multi-objective optimization model is constructed in combination with the current health state of the energy storage shelter, the energy storage capacity, and the charging and discharging rate limit; Using the multi-objective optimization model, a plurality of preset target internal scenarios are generated through Monte Carlo simulation to obtain scheduling scheme risk assessment results in a plurality of scenarios; Based on the scheduling scheme risk assessment results, a multi-objective optimization model is solved using a mixed integer linear programming algorithm to generate an initial energy storage module scheduling scheme; Using the initial energy storage module scheduling scheme, a comprehensive evaluation is performed through an analytic hierarchy process to determine an optimal energy storage module scheduling scheme.

2. The method of claim 1, wherein, Based on the scheduling scheme risk assessment results, a multi-objective optimization model is solved using a mixed integer linear programming algorithm to generate an initial energy storage module scheduling scheme, comprising: Based on the scheduling scheme risk assessment results, the physical limit conditions and operation rules of the energy storage shelter are integrated to obtain a multi-objective optimization model under constraints; Using the multi-objective optimization model, a phased optimization calculation of the scheduling scheme is performed through the introduction of a time window rolling optimization mechanism to obtain a phased scheduling scheme; Based on the phased scheduling scheme, a multi-objective optimization is solved using a mixed integer linear programming algorithm in combination with historical operating data and real-time operating state of the energy storage shelter to generate a preliminary energy storage module scheduling scheme; Based on comprehensive operating data, a health state evaluation model of the energy storage module is established through a machine learning algorithm to obtain a health state evaluation result; Using the preliminary energy storage module scheduling scheme in combination with the health state evaluation result, the preliminary energy storage module scheduling scheme is corrected to obtain a corrected energy storage module scheduling scheme; Using the corrected energy storage module scheduling scheme, the scheduling scheme is verified through simulation technology to adjust infeasible scheduling instructions to generate an initial energy storage module scheduling scheme.

3. The method of claim 2, wherein, The multi-objective optimization model is used to perform stage-by-stage optimization calculation on the scheduling scheme by introducing a time window rolling optimization mechanism to obtain a stage-by-stage scheduling scheme, including: The multi-objective optimization model is used to define a series of time windows by combining the real-time running state of the energy storage shelter and the power demand prediction results within a preset time to obtain a time window sequence; The time window sequence is used to perform local optimization calculation on the scheduling scheme within each time window according to the current health state of the energy storage shelter, the energy storage capacity, and the charge and discharge rate limit and external environmental parameters to obtain a locally optimal scheduling scheme for each time window.

4. The method of claim 2, wherein, According to the stage-by-stage scheduling scheme, a mixed integer linear programming algorithm is used to combine historical running data and real-time running state of the energy storage shelter to perform multi-objective optimization solving to generate a preliminary energy storage module scheduling scheme, including: The stage-by-stage scheduling scheme is used to analyze the running characteristics of the energy storage shelter by combining historical running data and real-time running state of the energy storage shelter to obtain running characteristic analysis results; The running characteristic analysis results are used to customize the objective function of the mixed integer linear programming algorithm to obtain a customized multi-objective optimization model; The customized multi-objective optimization model is used to combine the time windows in the stage-by-stage scheduling scheme to perform multi-objective optimization solving to generate preliminary multi-objective optimization results; The preliminary multi-objective optimization results are used to refine the optimization results by combining health state evaluation results of the energy storage shelter to obtain refined multi-objective optimization results; Based on the refined multi-objective optimization results, a comprehensive evaluation of multiple preliminary scheduling schemes is performed by an analytic hierarchy process to obtain an evaluated preliminary scheduling scheme; The evaluated preliminary scheduling scheme is used to combine real-time running data of the energy storage shelter to perform dynamic adjustment to generate a preliminary energy storage module scheduling scheme.

5. The method of claim 1, wherein, Based on the health index and the real-time power demand data, a genetic algorithm is used to dynamically adjust the charge and discharge strategies of each energy storage module to obtain a charge and discharge plan, and the charge and discharge plan is used in combination with deep learning technology to predict the power supply and demand trend within a preset time range to generate power demand prediction results, including: The health index and the real-time power demand data are used to dynamically adjust the charge and discharge strategies of each energy storage module by a genetic algorithm to obtain a preliminary charge and discharge plan; The preliminary charge and discharge plan is used in combination with real-time running state and historical running data of the energy storage shelter to predict the power supply and demand trend within a preset time range by deep learning technology to obtain preliminary power demand prediction results; The preliminary power demand prediction results are used to optimize the preliminary charge and discharge plan to obtain an optimized charge and discharge plan; The optimized charge and discharge plan is used to again predict the power supply and demand trend within a preset time range by deep learning technology to generate target power demand prediction results.

6. The method of claim 1, wherein, The scheduling instructions of the energy storage module are implemented to continuously optimize the decision-making process through online learning, and a best control instruction is generated by using a self-adaptive control strategy of reinforcement learning, including: The scheduling instruction of the energy storage module is used to start the charge and discharge operation of the energy storage shelter, and real-time operation data of the energy storage shelter during execution of the scheduling instruction is collected to obtain real-time operation feedback data; According to the real-time operation feedback data, in combination with historical operation data of the energy storage shelter, the current decision-making process is evaluated through an online learning algorithm to obtain a decision-making evaluation result; The decision-making evaluation result is used to design a reinforcement learning environment, define a reward function, construct a reinforcement learning model, and train the reinforcement learning model by simulating different charge and discharge strategies and energy storage module scheduling schemes to obtain an enhanced reinforcement learning model; The enhanced reinforcement learning model is used to optimize the control instruction of the energy storage module, generate a new control instruction, and obtain an optimized control instruction; The optimized control instruction is used to dynamically adjust the charge and discharge strategy in combination with the real-time operation state and environmental parameters of the energy storage shelter to generate an optimal control instruction.

7. A system for operating a modular energy storage shelter according to any one of claims 1 to 6. It comprises: The acquisition module is used to acquire the operation parameters of each energy storage module in real time to obtain comprehensive operation data, wherein the operation parameters include battery temperature, state of charge, internal resistance, voltage, current, and external environmental parameters; The construction module is used to construct an energy storage module health state evaluation model through a machine learning algorithm based on the comprehensive operation data to obtain a health index, and obtain real-time power demand data from a power grid company API interface, a user-side smart meter, and historical data and current environmental parameter analysis; The prediction module is used to dynamically adjust the charge and discharge strategy of each energy storage module based on the health index and real-time power demand data using a genetic algorithm to obtain a charge and discharge plan, and use the charge and discharge plan in combination with deep learning technology to predict the power supply and demand change trend within a preset time range to generate a power demand prediction result; The confirmation module is used to confirm an optimal energy storage module scheduling scheme based on the power demand prediction result using a mixed integer linear programming algorithm to obtain the scheduling instruction of the energy storage module; The optimization module is used to implement the scheduling instruction of the energy storage module, optimize the decision-making process through online learning, and generate an optimal control instruction using an adaptive control strategy of reinforcement learning.

8. A computing device, comprising: It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the operation control method of the modular energy storage shelter according to any one of claims 1-6.

9. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, the operation control method of the modular energy storage shelter according to any one of claims 1-6 is implemented.

Images