Micro-grid energy management and optimal scheduling method and system in shelter equipment

Abstract

The invention provides a micro-grid energy management and optimal scheduling method and system in shelter equipment. Wherein real-time load requirements of shelter equipment are collected, external environment factors are combined, future energy supply and demand conditions are evaluated, and an energy supply and demand evaluation report is generated; based on the energy supply and demand evaluation report, combining a plurality of prediction results by applying a multi-objective evolutionary algorithm and combining an integrated learning technology, and generating a preliminary energy management plan; based on the preliminary energy management plan, adopting a Bayesian optimization algorithm to adjust a scheduling strategy, combining grey system theory analysis, analyzing the working efficiency of a power generation unit, and generating an energy optimization scheduling scheme; and on the basis of the energy optimization scheduling scheme, continuously monitoring key indexes, correcting deviations in real time, and generating an energy management and optimization scheduling strategy. According to the technical scheme provided by the invention, the operation efficiency and adaptability of the micro-grid can be remarkably improved, and the system reliability and economic benefits are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of microgrid technology, and in particular to a method and system for microgrid energy management and optimization scheduling in a shelter device. Background Art

[0002] With the widespread application of shelter equipment in emergency rescue, medical support, field operations and other fields, microgrids need to operate efficiently in limited resources and changing environments. These scenarios usually face the following technical requirements: real-time load management, which requires real-time collection and processing of the load requirements of shelter equipment to ensure the continuity and reliability of power supply; environmental adaptability, which must be combined with external environmental factors (such as weather conditions, power market fluctuations, etc.) to comprehensively evaluate the future energy supply and demand of the microgrid; multi-objective optimization, while ensuring the quality of power supply, it is also necessary to consider multiple optimization goals such as economy, environmental protection and system stability.

[0003] At present, common solutions for microgrid energy management and optimal scheduling in shelter equipment include: static scheduling strategy, which is based on historical data and experience, pre-sets fixed scheduling strategies, and lacks the ability to adapt to real-time changes; single optimization algorithm, which only uses one optimization algorithm (such as genetic algorithm or particle swarm algorithm), which is difficult to meet multiple optimization goals at the same time; simple prediction model, which uses simple linear regression or time series model for prediction, and cannot accurately capture complex changing trends.

[0004] Although the existing solutions have solved some problems of microgrid energy management and optimal scheduling to a certain extent, they lack real-time and dynamic features. Static scheduling strategies cannot respond to changes in load demand and external environment in a timely manner, resulting in low system operation efficiency. The optimization objectives are single, and a single optimization algorithm is difficult to take into account multiple optimization objectives, which easily leads to performance sacrifice in one aspect. The prediction accuracy is insufficient, and simple prediction models cannot accurately capture complex nonlinear changes, resulting in large prediction errors, which affects the effectiveness of the scheduling strategy. Summary of the invention

[0005] The embodiment of the present application provides a method and system for microgrid energy management and optimization scheduling in a shelter device, so as to solve the problem of low system operation efficiency in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for microgrid energy management and optimization scheduling in a shelter device, including:

[0007] Collect the real-time load demand of shelter equipment, combine external environmental factors, comprehensively evaluate the future energy supply and demand of the microgrid, and generate an energy supply and demand assessment report;

[0008] Based on the energy supply and demand assessment report, a multi-objective evolutionary algorithm is used to comprehensively consider multiple optimization objectives, combined with ensemble learning technology, and multiple prediction results are combined to generate a preliminary energy management plan;

[0009] Based on the preliminary energy management plan, the Bayesian optimization algorithm is used to adjust the uncertainty scenario scheduling strategy, and combined with the grey system theory analysis, the working efficiency of the internal power generation unit of the microgrid is dynamically analyzed to generate an energy optimization scheduling plan;

[0010] Based on the energy optimization scheduling scheme, the key indicators of microgrid energy management and optimization scheduling are continuously monitored, the deviation between actual operation data and theory is corrected in real time, and energy management and optimization scheduling strategies are generated.

[0011] Optionally, based on the energy supply and demand assessment report, a multi-objective evolutionary algorithm is used to comprehensively consider multiple optimization objectives, combined with ensemble learning technology, and multiple prediction results to generate a preliminary energy management plan, including:

[0012] Based on the energy supply and demand assessment report and in combination with the actual operation requirements of the microgrid, multi-dimensional optimization objectives are set to generate an optimization objective set;

[0013] Based on the optimization target set, a multi-objective evolutionary algorithm is used to balance multiple optimization targets, search for target solutions in the solution space, perform multiple rounds of iterative optimization, and generate an optimization target solution set;

[0014] Based on the optimization target solution set, combined with ensemble learning technology, multiple prediction results are combined by weighted average to generate a comprehensive prediction result;

[0015] Based on the comprehensive prediction results, multiple optimization objectives are comprehensively considered, the overall energy management performance is evaluated, and a preliminary energy management plan is generated.

[0016] Optionally, based on the optimization target set, a multi-objective evolutionary algorithm is used to balance multiple optimization targets, search for target solutions in the solution space, perform multiple rounds of iterative optimization, and generate an optimization target solution set, including:

[0017] Based on the optimization target set, an individual representation method is defined, different energy management strategies are individually encoded, and an initial individual population is generated;

[0018] Based on the initial individual population, a fitness function is designed, a multi-objective evolutionary algorithm is adopted, multiple optimization objectives are comprehensively considered, the fitness of individuals in the initial individual population is evaluated, and a fitness evaluation result is generated;

[0019] Based on the fitness evaluation results, performing a variety of evolutionary operations on the individuals in the initial individual population, and generating an optimized individual population through multiple rounds of iterative optimization;

[0020] Based on the optimized individual population and in combination with the Pareto optimality principle, it is ensured that there is no obvious disadvantage among multiple optimization objectives, a group of non-dominated solutions is selected, and an optimization objective solution set is generated.

[0021] Optionally, based on the initial individual population, a fitness function is designed, a multi-objective evolutionary algorithm is adopted, multiple optimization objectives are comprehensively considered, and fitness evaluation is performed on individuals in the initial individual population to generate a fitness evaluation result, including:

[0022] Based on the initial individual population, identifying the time trend of the individual performance value under each optimization target;

[0023] Smoothing the individual performance values ​​using an exponential smoothing method to reduce the impact of short-term fluctuations to generate nonlinear performance values;

[0024] The nonlinear performance value is calculated using the following formula:

[0025]

[0026] Among them, P i (x) is the nonlinear performance value of the individual under the i-th optimization goal; x is the performance value of the individual under the i-th optimization goal; α is the parameter that controls the steepness of the S-type function; β is the expected value of the performance value; δ is the parameter that controls the amplitude change of the sine waveform; γ is the center position parameter of the sine waveform; φ is the phase shift of the sine waveform; λ is the parameter that controls the steepness of the additional S-type function; μ is the center position parameter of the additional S-type function;

[0027] Based on the nonlinear performance value, combined with the importance and priority of each optimization target, a reasonable weight is assigned, a cosine function and an additional S-type function are introduced, and nonlinear combination adjustment is performed to generate a comprehensive fitness value;

[0028] The comprehensive fitness value is calculated by the following formula:

[0029]

[0030] Among them, F(x) is the comprehensive fitness value of the individual under multiple optimization objectives; x is the performance value of the individual under the i-th optimization objective; P i (x) is the nonlinear performance value of the individual under the i-th optimization goal; P i-1 (x) is the nonlinear performance value of the individual under the i-1th optimization objective; w i is the weight of the i-th optimization goal, and the sum of the weights is 1; γ is the adjustment factor, which is used to adjust the contribution of the performance values ​​between different optimization goals; η is the parameter for controlling the rate of change of the difference; θ is the central position parameter of the rate of change of the difference; x minand x max are the minimum and maximum performance values ​​of the individual under all optimization objectives, respectively; κ is the parameter that controls the steepness of the additional S-shaped function; v is the central position parameter of the additional S-shaped function; i is the index of the optimization objective, from 1 to n; n is the total number of optimization objectives;

[0031] Based on the comprehensive fitness value, normalization processing is performed, individuals with high comprehensive fitness values ​​are screened to determine the evolutionary direction of the individual population, the stability of the optimization process is ensured through a convergence test, and a fitness evaluation result is generated.

[0032] Optionally, the optimization target solution set is combined with an ensemble learning technique to generate a comprehensive prediction result by weighted averaging multiple prediction results, including:

[0033] Based on the optimization target solution set, evaluating the performance balance in the multi-objective optimization process, selecting representative optimization target solutions, and generating a basic input set;

[0034] Based on the basic input set, multiple prediction models are trained to target specific optimization goals, and prediction and generalization capabilities are verified during the training process to generate a multi-model prediction system;

[0035] Based on the multi-model prediction system, a variety of optimization objectives are comprehensively considered, the future operation state of the microgrid is predicted from multiple perspectives, and a multi-model prediction result set is generated;

[0036] Based on the multi-model prediction result set, combined with the historical prediction accuracy of the multiple prediction models, different weights are assigned to perform weighted combination to generate a comprehensive prediction result.

[0037] Optionally, based on the preliminary energy management plan, a Bayesian optimization algorithm is used to adjust the uncertainty scenario scheduling strategy, and combined with the grey system theory analysis, the working efficiency of the internal power generation unit of the microgrid is dynamically analyzed to generate an energy optimization scheduling plan, including:

[0038] Based on the preliminary energy management plan, analyzing the operating status of the microgrid in different uncertainty scenarios, preliminarily adjusting the scheduling strategy through scenario analysis, and generating a scheduling strategy set for uncertainty scenarios;

[0039] Based on the uncertainty scenario scheduling strategy set, the Bayesian optimization algorithm is used to dynamically adjust by constructing a Bayesian probability model to ensure efficient and economic operation of microgrids in different uncertainty scenarios and generate an optimized scheduling strategy set;

[0040] Based on the optimized dispatch strategy set and combined with grey system theory analysis, the working efficiency of the power generation units in the microgrid under different load conditions is dynamically analyzed to generate a performance evaluation report of the power generation units;

[0041] Based on the performance evaluation report of the power generation unit and combined with the volatility of renewable energy output, multi-factor fusion optimization is performed to generate an energy optimization scheduling plan.

[0042] Optionally, the scheduling strategy set based on the uncertainty scenario uses a Bayesian optimization algorithm to dynamically adjust by constructing a Bayesian probability model to ensure efficient and economic operation of the microgrid in different uncertainty scenarios and generate an optimized scheduling strategy set, including:

[0043] Based on the uncertainty scenario scheduling strategy set, analyze the operating status of the microgrid under different uncertainty scenarios, identify key influencing factors, and generate a list of key influencing factors;

[0044] Based on the list of key influencing factors, a Bayesian optimization algorithm is used to construct a Bayesian probability model, dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system, and generate a dynamic adjustment scheduling strategy;

[0045] Based on the dynamic adjustment scheduling strategy and combined with the actual operation performance, the operation efficiency and economy of the microgrid under different uncertainty scenarios are evaluated to generate evaluation results;

[0046] Based on the evaluation results, the efficient and economic operation of the microgrid in different uncertainty scenarios is ensured, and optimization and adjustment are performed through iterative processing to generate an optimized dispatch strategy set.

[0047] Optionally, based on the list of key influencing factors, a Bayesian optimization algorithm is used to construct a Bayesian probability model, dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system, and generate a dynamic adjustment scheduling strategy, including:

[0048] Based on the list of key influencing factors, multi-scale decomposition of input variables is performed through multi-resolution analysis to extract features of different time scales, and the correlation coefficients between input variables and output variables are combined to generate conditional probabilities;

[0049] The conditional probability is calculated using the following formula:

[0050]

[0051] Where p(y|x,θ) is the conditional probability of output y given input x and parameter θ; f(x) is the prediction function in the Bayesian probability model; σ 2 is the variance of the model; φ is the phase shift of the sine waveform; δ is the parameter that controls the amplitude change of the sine waveform; γ is the center position parameter of the sine waveform; x is the input variable; θ is the model parameter; y is the output variable;

[0052] Based on the conditional probability, a log-likelihood function is constructed, a regularization term is introduced to prevent overfitting, and the model flexibility is enhanced through nonlinear transformation to generate optimized model parameters;

[0053] The optimization model parameters are calculated using the following formula:

[0054]

[0055] Among them, θ * To optimize the model parameters; p(y i ∣x i ,θ) is the given input x i and parameter θ, the output y i The conditional probability of; i is the index of the sample, from 1 to N; N is the number of samples; λ is the regularization parameter; j is the index of the key influencing factor, from 1 to M; P j (x i ) is the jth key influencing factor in the input x i The nonlinear performance value under P j-1 (x i ) is the j-1th key influencing factor in the input x i The nonlinear performance value under the condition of η is the parameter for controlling the rate of change of the difference; θ is the parameter for controlling the rate of change of the difference; j is the central position parameter of the jth key influencing factor; γ is the adjustment factor; x min and x max are the minimum and maximum values ​​of all inputs respectively; κ is the parameter that controls the steepness of the additional S-shaped function; v is the central position parameter of the additional S-shaped function; x i is the input variable of the i-th sample; y i is the output variable of the i-th sample; θ is the model parameter;

[0056] Based on the optimization model parameters, new output variables are predicted to dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system. Simulation software is used to simulate and monitor the operation of the microgrid in different uncertainty scenarios to generate a dynamic adjustment scheduling strategy.

[0057] Optionally, based on the energy optimization scheduling scheme, continuously monitoring key indicators of microgrid energy management and optimization scheduling, correcting the deviation between actual operation data and theory in real time, and generating energy management and optimization scheduling strategies include:

[0058] Based on the energy optimization scheduling scheme, key indicators in the operation process of the microgrid are continuously monitored, and a real-time operation data set is generated in combination with the actual output power of each power generation unit;

[0059] Based on the real-time operation data set, a comparative analysis is performed with the theoretical expected value of the energy optimization scheduling scheme to identify actual operation deviations and generate a deviation analysis report;

[0060] Based on the deviation analysis report, the microgrid operating parameters are adjusted in real time through an adaptive control strategy to reduce the actual operating deviation and generate a real-time correction instruction;

[0061] Based on the real-time correction instructions, combined with the historical operation data and current operation status of the microgrid, the scheduling parameters are dynamically adjusted, the operation strategy is optimized, and the energy management and optimization scheduling strategy is generated.

[0062] In a second aspect, an embodiment of the present application provides a microgrid energy management and optimization scheduling system in a shelter device, including:

[0063] The collection module is used to collect the real-time load demand of the shelter equipment, combine external environmental factors, comprehensively evaluate the future energy supply and demand of the microgrid, and generate an energy supply and demand assessment report;

[0064] A generation module is used to generate a preliminary energy management plan based on the energy supply and demand assessment report, using a multi-objective evolutionary algorithm, comprehensively considering multiple optimization objectives, combining ensemble learning technology, combining multiple prediction results, and combining multiple prediction results;

[0065] An adjustment module is used to adjust the uncertainty scenario scheduling strategy based on the preliminary energy management plan, adopt a Bayesian optimization algorithm, combine the grey system theory analysis, dynamically analyze the working efficiency of the internal power generation unit of the microgrid, and generate an energy optimization scheduling plan;

[0066] The monitoring module is used to continuously monitor the key indicators of microgrid energy management and optimal scheduling based on the energy optimization scheduling scheme, perform real-time correction on the deviation between actual operation data and theory, and generate energy management and optimal scheduling strategies.

[0067] In an embodiment of the present application, the real-time load demand of the cabin equipment is collected, combined with external environmental factors, the future energy supply and demand status of the microgrid is comprehensively evaluated, and an energy supply and demand evaluation report is generated; based on the energy supply and demand evaluation report, a multi-objective evolutionary algorithm is used, and multiple optimization objectives are comprehensively considered, combined with integrated learning technology, and multiple prediction results are combined to generate a preliminary energy management plan; based on the preliminary energy management plan, a Bayesian optimization algorithm is used to adjust the uncertainty scenario scheduling strategy, combined with gray system theoretical analysis, the working efficiency of the internal power generation unit of the microgrid is dynamically analyzed, and an energy optimization scheduling plan is generated; based on the energy optimization scheduling plan, the key indicators of microgrid energy management and optimization scheduling are continuously monitored, the deviation between actual operating data and theory is corrected in real time, and an energy management and optimization scheduling strategy is generated. By comprehensively evaluating the future energy supply and demand of the microgrid, an energy supply and demand assessment report is generated to ensure that the microgrid can operate efficiently under different operating conditions and avoid energy waste; using multi-objective evolutionary algorithms and Bayesian optimization algorithms, combined with integrated learning technology and grey system theory, a preliminary energy management plan and energy optimization scheduling scheme are generated to ensure that the scheduling strategy reaches the optimal under multiple optimization objectives; by dynamically analyzing the working efficiency of the power generation units inside the microgrid and combining the uncertainty scenario scheduling strategy, the adaptability of the microgrid to the volatility of renewable energy output and changes in the external environment is improved; the key indicators of the microgrid are continuously monitored, and the deviation between the actual operating data and the theory is corrected in real time to ensure that the microgrid always maintains an efficient, stable and economical state in actual operation; through simulation verification and sensitivity analysis, the operating performance of the microgrid under different uncertainty scenarios is evaluated to ensure that the system can operate reliably under various conditions; by optimizing the scheduling strategy, the output power of the power generation unit and the charging and discharging strategy of the energy storage system are reasonably allocated to effectively reduce the operating cost of the microgrid and improve the economic benefits.

[0068] Furthermore, based on the energy supply and demand assessment report and combined with the actual operation requirements of the microgrid, multi-dimensional optimization goals are set to generate an optimization goal set. This ensures that the microgrid can achieve optimality in multiple aspects (such as economy, environmental protection, stability and reliability), and improves the overall operation efficiency; a multi-objective evolutionary algorithm is used to balance multiple optimization goals, search for target solutions in the solution space, perform multiple rounds of iterative optimization, and generate an optimization goal solution set; this method can effectively solve multi-objective optimization problems and avoid performance sacrifices that may be caused by single-objective optimization; combined with ensemble learning technology, multiple prediction results are combined by weighted average to generate comprehensive prediction results. This improves the accuracy and robustness of the prediction and reduces the error caused by a single prediction model; based on the comprehensive prediction results, multiple optimization goals are comprehensively considered, the overall energy management performance is evaluated, and a preliminary energy management plan is generated; this provides a scientific basis for subsequent optimization scheduling and ensures efficient management of the microgrid under different operating conditions.

[0069] Furthermore, based on the preliminary energy management plan, the operating status of the microgrid under different uncertainty scenarios is analyzed, and the scheduling strategy is preliminarily adjusted through scenario analysis to generate a set of scheduling strategies for uncertainty scenarios. This improves the adaptability of the microgrid to uncertainty and changes, ensuring the stable operation of the system under various environments; the Bayesian optimization algorithm is used to dynamically adjust by constructing a Bayesian probability model to ensure the efficient and economic operation of the microgrid under different uncertainty scenarios, and generate an optimized scheduling strategy set. This method can find the optimal scheduling strategy in a complex and uncertain environment, improving the flexibility and economy of the system; combined with the gray system theory analysis, the working efficiency of the internal power generation unit of the microgrid under different load conditions is dynamically analyzed to generate a power generation unit performance evaluation report. This provides detailed performance data support for optimal scheduling, ensuring the efficient operation of the power generation unit under different working conditions; based on the power generation unit performance evaluation report, combined with the volatility of renewable energy output, multi-factor fusion optimization is performed to generate an energy optimization scheduling plan. This ensures that the microgrid can maintain an efficient, stable and economical operating state when dealing with renewable energy volatility and other uncertain factors.

[0070] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0072] Figure 1 A flowchart of a method for microgrid energy management and optimal scheduling in a shelter device provided in an embodiment of the present application;

[0073] Figure 2 A schematic diagram of the structure of a microgrid energy management and optimization scheduling system in a shelter device provided in an embodiment of the present application;

[0074] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0075] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0076] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0077] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0078] Figure 1 A flowchart of a microgrid energy management and optimization scheduling method in a shelter device is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0079] 101. Collect the real-time load demand of cabin equipment, combine external environmental factors, comprehensively evaluate the future energy supply and demand of the microgrid, and generate an energy supply and demand assessment report;

[0080] Real-time load demand includes real-time power demand data of various electrical equipment in the shelter equipment (such as medical equipment, lighting systems, communication equipment, etc.), which is used to evaluate current and future power demand.

[0081] External environmental factors include weather conditions (such as temperature, light intensity, wind speed, etc.), electricity market information (such as electricity price fluctuations, supply and demand relationships, etc.) and user behavior (such as electricity usage habits, activity time, etc.), which are used to evaluate the impact of the external environment on the operation of the microgrid.

[0082] The energy supply and demand situation refers to the power supply and demand situation of the microgrid in the future, including the output capacity of the power generation unit, the charging and discharging status of the energy storage system, the changes in load demand, etc., which are used to formulate a reasonable energy management plan.

[0083] In the embodiment of the present application, firstly, various sensors and monitoring devices installed in the cabin equipment are used to collect power demand data of electrical equipment in real time, and the data is transmitted to the central control system through the Internet of Things technology to ensure the real-time and accuracy of the data; secondly, data on external environmental factors are collected, including weather forecasts, electricity market price information and user behavior data, which can be obtained through the Internet or other data interfaces and integrated into the central control system; thirdly, the central control system conducts a comprehensive analysis of the real-time load demand data and the external environmental factor data to evaluate the energy supply and demand status of the microgrid in the future, including predicting the output capacity of the power generation unit, the charging and discharging status of the energy storage system and the changing trend of the load demand; finally, based on the comprehensive analysis results, an energy supply and demand assessment report is generated, which describes in detail the energy supply and demand status of the microgrid in different time periods, and provides a scientific basis for subsequent energy management and optimized scheduling.

[0084] Assume that in a temporary medical site, it is necessary to ensure the continuity and reliability of power supply in the shelter equipment; install various sensors in the shelter equipment, such as current transformers, voltage sensors, temperature sensors, etc., to collect power demand data of medical equipment, lighting systems and communication equipment in real time, and transmit them to the central control system through wireless communication modules; obtain local weather forecast data (such as temperature, light intensity, wind speed, etc.) and power market price information through the Internet; at the same time, obtain users' power usage habits and activity time through questionnaires or historical data; the central control system conducts a comprehensive analysis of real-time load demand data and external environmental factor data, such as predicting the output capacity of solar power generation units according to weather forecasts, adjusting the charging and discharging strategy of energy storage systems according to power market price information, and predicting future load demand according to users' power usage habits; based on the comprehensive analysis results, generate an energy supply and demand assessment report. The report describes in detail the power supply and demand of the microgrid in the next 24 hours, including the output power of the power generation units in each time period, the charge state of the energy storage system and the load demand change trend, providing a scientific basis for subsequent energy management and optimized scheduling.

[0085] Through this series of steps, the stability and reliability of power supply to the cabin equipment under different environmental conditions are ensured.

[0086] 102. Based on the energy supply and demand assessment report, a multi-objective evolutionary algorithm is used to comprehensively consider multiple optimization objectives, combined with ensemble learning technology, and multiple prediction results to generate a preliminary energy management plan;

[0087] The energy supply and demand assessment report contains the power supply and demand situation of the microgrid in the future, including the output capacity of the power generation unit, the charging and discharging status of the energy storage system, the changing trend of load demand, etc.

[0088] The multi-objective evolutionary algorithm is an optimization algorithm used to find a balance between multiple optimization objectives and generate the optimal solution. Common algorithms include genetic algorithms, particle swarm algorithms, etc.

[0089] Optimization objectives include economy (such as minimum cost), environmental protection (such as minimum carbon emissions), stability (such as minimum fluctuations) and reliability (such as maximum power supply continuity).

[0090] Ensemble learning technology improves the accuracy and robustness of predictions by combining the results of multiple prediction models. Common techniques include random forests and gradient boosting trees.

[0091] The preliminary energy management plan is a preliminary energy management plan for the microgrid generated based on comprehensive prediction results and multi-objective optimization, including the output plan of the power generation unit, the charging and discharging strategy of the energy storage system, etc.

[0092] In this step, firstly, according to the energy supply and demand assessment report, the power demand and supply of the microgrid in different time periods are determined, including the output capacity of the power generation unit, the charging and discharging status of the energy storage system, and the changing trend of the load demand; secondly, multi-dimensional optimization goals are set, such as economy, environmental protection, stability and reliability, to guide the subsequent optimization process and ensure that the microgrid can achieve the best in many aspects; thirdly, the multi-objective evolutionary algorithm is used to search for the target solution in the solution space, and multiple rounds of iterative optimization are performed to generate the optimization target solution set, and the optimal balance point between multiple optimization goals is found by simulating natural selection and genetic mechanisms; finally, combined with ensemble learning technology, multiple prediction results are combined by weighted average to generate a comprehensive prediction result. Based on the comprehensive prediction results, the overall energy management performance is evaluated, and a preliminary energy management plan is generated to ensure the accuracy and reliability of the plan.

[0093] Optionally, the method in step 102 is based on the energy supply and demand assessment report, uses a multi-objective evolutionary algorithm, comprehensively considers multiple optimization objectives, combines integrated learning technology, combines multiple prediction results, and generates a preliminary energy management plan, including: based on the energy supply and demand assessment report, combined with the actual operation needs of the microgrid, sets multi-dimensional optimization objectives, and generates an optimization objective set; based on the optimization objective set, uses a multi-objective evolutionary algorithm to balance multiple optimization objectives, searches for target solutions in the solution space, performs multiple rounds of iterative optimization, and generates an optimization objective solution set; based on the optimization objective solution set, combines integrated learning technology, combines multiple prediction results by weighted averaging, and generates a comprehensive prediction result; based on the comprehensive prediction result, comprehensively considers multiple optimization objectives, evaluates the overall energy management performance, and generates a preliminary energy management plan.

[0094] Among them, based on the optimization target set, a multi-objective evolutionary algorithm is adopted to balance multiple optimization targets, search for target solutions in the solution space, perform multiple rounds of iterative optimization, and generate an optimization target solution set, including: based on the optimization target set, defining an individual representation method, performing individual encoding on different energy management strategies, and generating an initial individual population; based on the initial individual population, designing a fitness function, using a multi-objective evolutionary algorithm, comprehensively considering multiple optimization targets, performing fitness evaluation on individuals in the initial individual population, and generating a fitness evaluation result; based on the fitness evaluation result, performing multiple evolutionary operations on individuals in the initial individual population, and generating an optimized individual population through multiple rounds of iterative optimization; based on the optimized individual population, combined with the Pareto optimal principle, ensuring that there is no obvious disadvantage between multiple optimization targets, selecting a group of non-dominated solutions, and generating an optimization target solution set.

[0095] The energy supply and demand assessment report contains the power supply and demand situation of the microgrid in the future, including the output capacity of the power generation unit, the charging and discharging status of the energy storage system, the changing trend of the load demand, etc., which is used to guide the generation of the energy management plan.

[0096] The multi-dimensional optimization objectives include economy (such as minimum cost), environmental protection (such as minimum carbon emissions), stability (such as minimum fluctuation) and reliability (such as maximum power supply continuity), which are used to comprehensively evaluate and optimize the operation of microgrids.

[0097] Multi-objective evolutionary algorithm is an optimization algorithm that simulates natural selection and genetic mechanisms to find a balance between multiple optimization objectives and generate the optimal solution. Common algorithms include genetic algorithm, particle swarm algorithm, etc.

[0098] The individual representation encodes different energy management strategies as individuals, each of which represents a possible solution for the optimization process of the multi-objective evolutionary algorithm.

[0099] The fitness function is used to evaluate the performance of an individual on the optimization target, and the fitness function can quantify the pros and cons of each individual.

[0100] Evolutionary operations include selection, crossover, mutation, etc., which are used to generate new individuals and gradually optimize the population.

[0101] The Pareto optimality principle selects a set of non-dominated solutions in multi-objective optimization, which have no obvious disadvantages among multiple optimization objectives and form the Pareto front.

[0102] The optimization target solution set is a set of optimal solutions generated by a multi-objective evolutionary algorithm, which achieves the best balance among multiple optimization targets.

[0103] In the embodiment of the present application, first, based on the energy supply and demand assessment report and in combination with the actual operation requirements of the microgrid, multi-dimensional optimization goals are set to generate an optimization goal set; this includes goals such as economy, environmental protection, stability and reliability; secondly, based on the optimization goal set, an individual representation method is defined, different energy management strategies are individually encoded, and an initial individual population is generated, each individual representing a possible energy management strategy; thirdly, based on the initial individual population, a fitness function is designed, and a multi-objective evolutionary algorithm is adopted. Taking multiple optimization goals into consideration, the fitness of individuals in the initial individual population is evaluated to generate a fitness evaluation result; the fitness function is used to quantify the performance of each individual on the optimization goal; finally, based on the fitness evaluation result, a variety of evolutionary operations are performed on the individuals in the initial individual population, and an optimized individual population is generated through multiple rounds of iterative optimization; based on the optimized individual population, combined with the Pareto optimal principle, it is ensured that there is no obvious disadvantage between multiple optimization goals, a group of non-dominated solutions are selected, and an optimization goal solution set is generated; this group of solutions achieves the best balance between multiple optimization goals.

[0104] Assume that in a temporary medical site, it is necessary to ensure the continuity and reliability of power supply in the shelter equipment. First, various sensors installed in the shelter equipment, such as current transformers, voltage sensors, temperature sensors, etc., collect power demand data of medical equipment, lighting systems and communication equipment in real time, and transmit these data to the central control system through wireless communication modules. Secondly, the central control system sets multi-dimensional optimization goals based on the energy supply and demand assessment report and the actual operation needs of the microgrid, and generates a set of optimization goals, including economy (minimum cost), environmental protection (minimum carbon emissions), stability and reliability (maximum power supply continuity). Thirdly, define the individual representation method, encode different energy management strategies individually, and generate an initial individual population. Each individual represents a possible energy management strategy. For example, an individual may be represented as "the output power of the solar power generation unit is 80%, and the charging power of the energy storage system is 20%". After that, Based on the initial individual population, a fitness function is designed, and a multi-objective evolutionary algorithm is used to comprehensively consider multiple optimization objectives, and the fitness of individuals in the initial individual population is evaluated to generate fitness evaluation results. The fitness function is used to quantify the performance of each individual on the optimization objective. For example, the fitness value of the economic objective may be 1,000 yuan / day, and the fitness value of the environmental objective may be 100kgCO2 / day; based on the fitness evaluation results, the individuals in the initial individual population are subjected to evolutionary operations such as selection, crossover, and mutation, and an optimized individual population is generated through multiple rounds of iterative optimization. For example, individuals with higher fitness are selected for crossover to generate new individuals, and then mutation operations are performed on the new individuals to increase the diversity of the population; finally, based on the optimized individual population, combined with the Pareto optimal principle, it is ensured that there is no obvious disadvantage between multiple optimization objectives, and a set of non-dominated solutions is selected to generate an optimization objective solution set. This set of solutions achieves the best balance between multiple optimization objectives.

[0105] Through the above steps, the stability and reliability of power supply for cabin equipment under different environmental conditions are ensured, and the power supply efficiency and economy of the microgrid are improved.

[0106] This application takes into account that in multi-objective optimization problems, the design of the fitness function is crucial to finding the optimal solution. By introducing nonlinear combinations and dynamic adjustment mechanisms, the comprehensive performance of individuals under multiple optimization objectives can be more accurately evaluated. By combining S-type functions, sine waveforms, cosine functions and additional S-type functions, a comprehensive fitness value is generated to ensure the stability and efficiency of the optimization process.

[0107] Optionally, based on the initial individual population, a fitness function is designed, a multi-objective evolutionary algorithm is adopted, multiple optimization objectives are comprehensively considered, and fitness evaluation is performed on individuals in the initial individual population to generate a fitness evaluation result, including:

[0108] Based on the initial individual population, identifying the time trend of the individual performance value under each optimization target;

[0109] Smoothing the individual performance values ​​using an exponential smoothing method to reduce the impact of short-term fluctuations to generate nonlinear performance values;

[0110] The nonlinear performance value is calculated using the following formula:

[0111]

[0112] Among them, P i (x) is the nonlinear performance value of the individual under the i-th optimization goal; x is the performance value of the individual under the i-th optimization goal; α is the parameter that controls the steepness of the S-type function; β is the expected value of the performance value; δ is the parameter that controls the amplitude change of the sine waveform; γ is the center position parameter of the sine waveform; φ is the phase shift of the sine waveform; λ is the parameter that controls the steepness of the additional S-type function; μ is the center position parameter of the additional S-type function;

[0113] Based on the nonlinear performance value, combined with the importance and priority of each optimization target, a reasonable weight is assigned, a cosine function and an additional S-type function are introduced, and nonlinear combination adjustment is performed to generate a comprehensive fitness value;

[0114] The comprehensive fitness value is calculated by the following formula:

[0115]

[0116] Among them, F(x) is the comprehensive fitness value of the individual under multiple optimization objectives; x is the performance value of the individual under the i-th optimization objective; P i (x) is the nonlinear performance value of the individual under the i-th optimization goal; P i-1 (x) is the nonlinear performance value of the individual under the i-1th optimization objective; w i is the weight of the i-th optimization goal, and the sum of the weights is 1; γ is the adjustment factor, which is used to adjust the contribution of the performance values ​​between different optimization goals; η is the parameter for controlling the rate of change of the difference; θ is the central position parameter of the rate of change of the difference; x min and x max are the minimum and maximum performance values ​​of the individual under all optimization objectives, respectively; κ is the parameter that controls the steepness of the additional sigmoid function; ν is the central position parameter of the additional sigmoid function; i is the index of the optimization objective, from 1 to n; n is the total number of optimization objectives;

[0117] Based on the comprehensive fitness value, normalization processing is performed, individuals with high comprehensive fitness values ​​are screened to determine the evolutionary direction of the individual population, the stability of the optimization process is ensured through a convergence test, and a fitness evaluation result is generated.

[0118] This method aims to balance the relationship between multiple optimization objectives through complex nonlinear combinations, while introducing nonlinear combinations and dynamic adjustment mechanisms to reduce the impact of short-term fluctuations and improve the robustness and stability of the optimization results.

[0119] In the nonlinear performance value, the S-shaped function part: Used to smooth the performance value and control the steepness of the performance value; Sine waveform part: Used to simulate periodic changes, taking into account the periodicity and phase shift of the performance value; additional S-shaped function part: Used to further adjust the performance value and control the additional steepness;

[0120] Among them, α, β, δ, γ, φ, λ, μ: the initial values ​​can be set by historical data and experience, and then gradually adjusted through iterative optimization; α, δ, λ control the steepness of the corresponding function, β, γ, μ are the center position parameters of the corresponding function, and φ is the phase offset of the sine waveform;

[0121] In the comprehensive fitness value, the weight part w i : Used to assign importance and priority to each optimization goal; difference adjustment item Consider the differences between adjacent optimization goals to ensure that individuals perform evenly across different goals; cosine function adjustment term Periodically adjust the performance values ​​to ensure that the performance values ​​in different ranges can be reasonably weighted; additional S-shaped function adjustment items Further adjust the performance values ​​to ensure that performance values ​​within a certain range can be given higher weights;

[0122] Among them, w i is the weight part, which is determined by expert evaluation or historical data analysis, and the sum of the weights is 1; γ is the adjustment factor, which is used to adjust the contribution of performance values ​​between different optimization goals; η, θ are set through historical data and experience to control the rate of change of the difference; x min ,x max is the minimum and maximum performance value of the individual under all optimization objectives; κ,v is set through historical data and experience to control the steepness and center position of the additional S-shaped function;

[0123] Assume that there are two optimization goals in microgrid energy management: economy (cost) and environmental protection (carbon emissions); there is an individual in the initial individual population, whose performance value in economy is x 1 = 5000 yuan, environmental performance value x 2 =100kg CO2; the setting parameters are α=0.1, β=5000, δ=0.05, γ=5000, φ=0, λ=0.05, μ=5000w1 =0.6,w 2 =0.4,γ=1.5,η=

[0124] 0.01,θ=5000,x min =0,x max =10000,κ=0.05,v=5000;

[0125]

[0126] The calculation result of the comprehensive fitness value is 0.6000. Assuming that the threshold is set to 0.5, since the result is greater than the set threshold, it shows that the performance of the individual on multiple optimization objectives has reached the expected level, so that it can achieve a balance between efficiency and economy in actual operation; through the above steps, the efficient operation of the microgrid under multiple optimization objectives such as economy and environmental protection is ensured.

[0127] Optionally, based on the optimization target solution set, combined with ensemble learning technology, a plurality of prediction results are combined by weighted averaging to generate a comprehensive prediction result, including: based on the optimization target solution set, evaluating the performance balance in the multi-target optimization process, selecting representative optimization target solutions, and generating a basic input set; based on the basic input set, training a plurality of prediction models to target specific optimization targets, verifying the prediction ability and generalization ability during the training process, and generating a multi-model prediction system; based on the multi-model prediction system, comprehensively considering a plurality of optimization targets, predicting the future operating state of the microgrid from multiple perspectives, and generating a multi-model prediction result set; based on the multi-model prediction result set, combining the historical prediction accuracy of the plurality of prediction models, assigning different weights for weighted combination, and generating a comprehensive prediction result.

[0128] The optimization target solution set is a set of optimal solutions generated by a multi-objective evolutionary algorithm, which achieves the best balance among multiple optimization objectives.

[0129] Performance balance evaluates whether the performance of each optimization objective solution on multiple optimization objectives is balanced, ensuring that the selected solution has no obvious disadvantages on each objective.

[0130] The basic input set is a representative solution selected from the optimization target solution set and serves as the basic data set for training the prediction model.

[0131] A variety of prediction models including different machine learning or statistical models, such as random forest, support vector machine, neural network, etc., are used to predict the future operating status of the microgrid.

[0132] Verification of predictive and generalization abilities During the training process, the predictive and generalization abilities of the model are verified through cross-validation and test sets.

[0133] A multi-model prediction system is a system composed of multiple prediction models, each of which makes predictions for a specific optimization goal.

[0134] Multi-perspective prediction predicts the future operating status of the microgrid from different perspectives (i.e., different optimization objectives) and generates a multi-model prediction result set.

[0135] The comprehensive prediction result is the result of combining multiple prediction models by weighted average to generate the final comprehensive prediction result.

[0136] In an embodiment of the present application, first, based on the optimization target solution set, the performance balance in the multi-objective optimization process is evaluated, representative optimization target solutions are selected, and a basic input set is generated; this step ensures that the selected solution performs evenly on each optimization target; secondly, based on the basic input set, multiple prediction models are trained to target specific optimization targets, and the prediction ability and generalization ability are verified during the training process to generate a multi-model prediction system; this step ensures the prediction accuracy of each model on a specific optimization target; thirdly, based on the multi-model prediction system, a variety of optimization targets are comprehensively considered, and the future operating state of the microgrid is predicted from multiple perspectives to generate a multi-model prediction result set; this step predicts the future state of the microgrid from different angles; finally, based on the multi-model prediction result set, different weights are assigned to weighted combinations based on the historical prediction accuracy of multiple prediction models to generate a comprehensive prediction result; this step improves the accuracy and robustness of the prediction results through the weighted averaging method.

[0137] Assume that in a temporary meteorological observation point, it is necessary to ensure the continuity and reliability of power supply. First, based on the optimization target solution set, evaluate the performance balance in the multi-objective optimization process, select representative optimization target solutions, and generate a basic input set. For example, select 10 solutions that are balanced in terms of economy, environmental protection, stability and reliability. Second, based on the basic input set, train a variety of prediction models, such as random forests, support vector machines and neural networks. Each model is trained for a specific optimization goal. During the training process, the prediction ability and generalization ability of the model are verified through cross-validation and test sets to generate a multi-model prediction system. Third, based on the multi-model prediction system, comprehensively consider a variety of optimization goals, and predict micro-perspectives. The operating status of the power grid in the next 24 hours generates a multi-model prediction result set. For example, the random forest model predicts the output power of the power generation unit in the next 24 hours, the support vector machine model predicts the charge state of the energy storage system, and the neural network model predicts the change trend of the load demand; finally, based on the multi-model prediction result set, combined with the historical prediction accuracy of each prediction model, different weights are assigned to weighted combination to generate a comprehensive prediction result. For example, the historical prediction accuracy of the random forest model is 90%, the historical prediction accuracy of the support vector machine model is 85%, and the historical prediction accuracy of the neural network model is 95%. They are assigned weights of 0.3, 0.25, and 0.45 respectively, and the final comprehensive prediction result is generated by weighted average.

[0138] Through the above steps, the stable power supply of the cabin equipment under different environmental conditions is ensured, the operating efficiency of the microgrid is improved, and the operating cost is reduced.

[0139] 103. Based on the preliminary energy management plan, the Bayesian optimization algorithm is used to adjust the uncertainty scenario scheduling strategy, and combined with the grey system theory analysis, the working efficiency of the internal power generation unit of the microgrid is dynamically analyzed to generate an energy optimization scheduling plan;

[0140] The preliminary energy management plan is a preliminary energy management plan for the microgrid generated based on a multi-objective evolutionary algorithm and integrated learning technology, including the output plan of the power generation unit, the charging and discharging strategy of the energy storage system, etc.

[0141] The Bayesian optimization algorithm is an optimization algorithm based on the Bayesian probability model. It finds the optimal solution by dynamically adjusting parameters and is suitable for dealing with uncertainty and complex optimization problems.

[0142] Uncertain scenarios refer to various uncertain situations that may occur during the operation of microgrids, such as weather changes, fluctuations in electricity market demand, etc. Grey system theory is a theory for dealing with incomplete information and uncertainty problems. It improves the accuracy of prediction and decision-making by dynamically analyzing the internal state of the system.

[0143] The energy optimization scheduling scheme is a better microgrid energy management scheme generated by adjusting the scheduling strategy based on the preliminary energy management plan.

[0144] In this step, firstly, based on the preliminary energy management plan, the operating status of the microgrid under different uncertainty scenarios is analyzed. Through scenario analysis, the scheduling strategy is preliminarily adjusted to generate a set of scheduling strategies for uncertainty scenarios. Secondly, the Bayesian optimization algorithm is used to construct a Bayesian probability model, and the uncertainty scenario scheduling strategy is dynamically adjusted to ensure the efficient and economic operation of the microgrid under different uncertainty scenarios. Thirdly, combined with the grey system theoretical analysis, the working efficiency of the power generation units inside the microgrid under different load conditions is dynamically analyzed, and a power generation unit performance evaluation report is generated to provide detailed performance data support. Finally, based on the power generation unit performance evaluation report and combined with the volatility of renewable energy output, multi-factor fusion optimization is performed to generate an energy optimization scheduling plan, which ensures that the microgrid can maintain an efficient, stable and economical operating state when dealing with various uncertainties.

[0145] Optionally, in step 103, based on the preliminary energy management plan, a Bayesian optimization algorithm is used to adjust the uncertainty scenario scheduling strategy, combined with grey system theoretical analysis, the working efficiency of the power generation units inside the microgrid is dynamically analyzed to generate an energy optimization scheduling plan, including: based on the preliminary energy management plan, analyzing the operating status of the microgrid in different uncertainty scenarios, preliminarily adjusting the scheduling strategy through scenario analysis, and generating an uncertainty scenario scheduling strategy set; based on the uncertainty scenario scheduling strategy set, using the Bayesian optimization algorithm, dynamically adjusting by constructing a Bayesian probability model to ensure efficient and economic operation of the microgrid in different uncertainty scenarios, and generating an optimized scheduling strategy set; based on the optimized scheduling strategy set, combined with grey system theoretical analysis, dynamically analyzing the working efficiency of the power generation units under different load conditions inside the microgrid, and generating a power generation unit performance evaluation report; based on the power generation unit performance evaluation report, combined with the volatility of renewable energy output, multi-factor fusion optimization is performed to generate an energy optimization scheduling plan.

[0146] Among them, the scheduling strategy set based on the uncertainty scenario uses a Bayesian optimization algorithm and dynamically adjusts by constructing a Bayesian probability model to ensure efficient and economic operation of microgrids in different uncertainty scenarios and generate an optimized scheduling strategy set, including: based on the uncertainty scenario scheduling strategy set, analyzing the operating status of the microgrid under different uncertainty scenarios, identifying key influencing factors, and generating a list of key influencing factors; based on the list of key influencing factors, using a Bayesian optimization algorithm, constructing a Bayesian probability model, dynamically adjusting the output power of each power generation unit and the charging and discharging strategy of the energy storage system, and generating a dynamically adjusted scheduling strategy; based on the dynamically adjusted scheduling strategy, combined with actual operating performance, evaluating the operating efficiency and economy of the microgrid under different uncertainty scenarios, and generating evaluation results; based on the evaluation results, ensuring efficient and economic operation of microgrids in different uncertainty scenarios, optimizing and adjusting through iterative processing, and generating an optimized scheduling strategy set.

[0147] The preliminary energy management plan is a preliminary energy management plan for the microgrid generated based on a multi-objective evolutionary algorithm and integrated learning technology, including the output plan of the power generation unit, the charging and discharging strategy of the energy storage system, etc.

[0148] Uncertain scenarios refer to various uncertain situations that may occur during the operation of a microgrid, such as weather changes and fluctuations in electricity market demand.

[0149] The Bayesian optimization algorithm is an optimization algorithm based on the Bayesian probability model. It finds the optimal solution by dynamically adjusting parameters. It is suitable for dealing with uncertainty and complex optimization problems.

[0150] Grey system theory is a theory that deals with incomplete information and uncertainty problems. It improves the accuracy of prediction and decision-making by dynamically analyzing the internal state of the system.

[0151] The key influencing factors are those that have a significant impact on the operation of the microgrid under different uncertainty scenarios, such as renewable energy output, load demand changes, etc.

[0152] The dynamic adjustment scheduling strategy is to dynamically adjust the output power of the power generation unit and the charging and discharging strategy of the energy storage system according to real-time data and prediction results to ensure the efficient and economical operation of the microgrid under different uncertainty scenarios.

[0153] The evaluation results evaluate the operating efficiency and economy of the microgrid under different uncertainty scenarios through actual operating performance, providing a basis for subsequent optimization.

[0154] In an embodiment of the present application, first, based on a preliminary energy management plan, the operating status of the microgrid under different uncertainty scenarios is analyzed, and the scheduling strategy is preliminarily adjusted through scenario analysis to generate an uncertainty scenario scheduling strategy set; secondly, based on the uncertainty scenario scheduling strategy set, key influencing factors are identified and a list of key influencing factors is generated; these factors may include weather conditions, load demand, market prices, etc.; again, based on the list of key influencing factors, a Bayesian optimization algorithm is used to construct a Bayesian probability model, and the output power of each power generation unit and the charging and discharging strategy of the energy storage system are dynamically adjusted to generate a dynamically adjusted scheduling strategy; then, based on the dynamic adjustment of the scheduling strategy, combined with the actual operating performance, the operating efficiency and economy of the microgrid under different uncertainty scenarios are evaluated to generate an evaluation result; finally, based on the evaluation result, optimization and adjustment are performed through iterative processing to generate an optimized scheduling strategy set to ensure efficient and economical operation of the microgrid under different uncertainty scenarios.

[0155] Assume that in a cabin equipment at a temporary experimental site, it is necessary to ensure that the power supply of the microgrid is stable and reliable; first, based on the preliminary energy management plan, analyze the operating status of the microgrid under different uncertainty scenarios, for example, consider weather conditions such as sunny days, cloudy days, rainy days, and changes in load demand in different time periods, and preliminarily adjust the scheduling strategy through scenario analysis to generate an uncertainty scenario scheduling strategy set; secondly, based on the uncertainty scenario scheduling strategy set, identify key influencing factors and generate a list of key influencing factors, such as solar power generation, wind power generation, energy storage system charge state, load demand changes, etc.; thirdly, based on the list of key influencing factors, use the Bayesian optimization algorithm to construct a Bayesian probability model, dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system, and generate a dynamically adjusted Scheduling strategies, for example, increasing the output power of solar power generation units on sunny days and increasing the discharge power of energy storage systems at night or on cloudy days; then, based on the dynamic adjustment of scheduling strategies and combined with actual operating performance, the operating efficiency and economy of the microgrid under different uncertainty scenarios are evaluated to generate evaluation results, for example, calculating the total cost, carbon emissions and power supply reliability in each scenario; finally, based on the evaluation results, optimization and adjustment are performed through iterative processing to generate an optimized scheduling strategy set to ensure efficient and economic operation of the microgrid under different uncertainty scenarios. For example, after multiple iterations, it is finally determined that the output power of the solar power generation unit is 80% on sunny days, the charging power of the energy storage system is 20%, the discharge power of the energy storage system is 70% at night or on cloudy days, and the standby power of the diesel generator is 30%.

[0156] Through the above steps, the stability and reliability of power supply for cabin equipment under different environmental conditions are ensured, and the operating efficiency and economy of the microgrid are improved.

[0157] This application takes into account that in microgrid energy management, the Bayesian optimization algorithm dynamically adjusts the output power of each power generation unit and the charging and discharging strategy of the energy storage system by constructing a Bayesian probability model, generates a dynamic adjustment scheduling strategy, and improves the flexibility and accuracy of the model through multi-resolution analysis and nonlinear transformation, ensuring efficient operation under different uncertainty scenarios.

[0158] Optionally, based on the list of key influencing factors, a Bayesian optimization algorithm is used to construct a Bayesian probability model, dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system, and generate a dynamic adjustment scheduling strategy, including:

[0159] Based on the list of key influencing factors, multi-scale decomposition of input variables is performed through multi-resolution analysis to extract features of different time scales, and the correlation coefficients between input variables and output variables are combined to generate conditional probabilities;

[0160] The conditional probability is calculated using the following formula:

[0161]

[0162] Where p(y|x,θ) is the conditional probability of output y given input x and parameter θ; f(x) is the prediction function in the Bayesian probability model; σ 2 is the variance of the model; φ is the phase shift of the sine waveform; δ is the parameter that controls the amplitude change of the sine waveform; γ is the center position parameter of the sine waveform; x is the input variable; θ is the model parameter; y is the output variable;

[0163] Based on the conditional probability, a log-likelihood function is constructed, a regularization term is introduced to prevent overfitting, and the model flexibility is enhanced through nonlinear transformation to generate optimized model parameters;

[0164] The optimization model parameters are calculated using the following formula:

[0165]

[0166] Among them, θ * To optimize the model parameters; p(y i ∣x i ,θ) is the given input x i and parameter θ, the output y i The conditional probability of; i is the index of the sample, from 1 to N; N is the number of samples; λ is the regularization parameter; j is the index of the key influencing factor, from 1 to M; P j (x i ) is the jth key influencing factor in the input x i The nonlinear performance value under P j-1 (x i) is the j-1th key influencing factor in the input x i The nonlinear performance value under the condition of η is the parameter for controlling the rate of change of the difference; θ is the parameter for controlling the rate of change of the difference; j is the central position parameter of the jth key influencing factor; γ is the adjustment factor; x min and x max are the minimum and maximum values ​​of all inputs respectively; κ is the parameter that controls the steepness of the additional sigmoid function; ν is the central position parameter of the additional sigmoid function; x i is the input variable of the i-th sample; y i is the output variable of the i-th sample; θ is the model parameter;

[0167] Based on the optimization model parameters, new output variables are predicted to dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system. Simulation software is used to simulate and monitor the operation of the microgrid in different uncertainty scenarios to generate a dynamic adjustment scheduling strategy.

[0168] This method aims to balance the relationship between multiple key influencing factors through complex nonlinear combinations and dynamic adjustment mechanisms, reduce the impact of short-term fluctuations through multi-resolution analysis and nonlinear transformation, and improve the robustness and stability of optimization results.

[0169] In conditional probability, the Gaussian distribution part Used to describe the probability distribution of the output variable y under given input x and parameter θ; the sine waveform part Used to simulate periodic changes, taking into account the periodicity and phase shift of the input variables;

[0170] Among them, σ 2 The initial value is set by historical data and experience, and then gradually adjusted by iterative optimization, which represents the variance of the model; φ, δ, γ are set by historical data and experience to control the phase shift, amplitude change and center position of the sine waveform; f(x) is the prediction function obtained by fitting the training data; x, y are input variables and output variables, obtained from actual data; θ is the model parameter, determined by the optimization process;

[0171] In the Optimize Model Parameters section, the Log-Likelihood Function section Used to maximize the log-likelihood of conditional probability to ensure optimal estimation of model parameters; regularization term part Used to introduce regularization terms to prevent overfitting and enhance model flexibility through nonlinear transformation;

[0172] Among them, λ controls the regularization strength through cross-validation or empirical setting; η,θ j ,γ,x min ,x max,κ,v is set through historical data and experience to control the difference change rate, center position parameters, adjustment factors, etc.;P j (x i ),P j-1 (x i ) is the nonlinear performance value of the key influencing factors obtained through multi-resolution analysis and nonlinear processing; N, M are the number of samples and the number of key influencing factors, obtained from actual data; x i ,y i are the input variables and output variables of the i-th sample, obtained from the actual data;

[0173] Assume that in a shelter equipment at a temporary field experiment site, it is necessary to ensure the continuity and reliability of power supply. The key influencing factors include weather conditions (such as temperature and light intensity) and changes in load demand.

[0174] Assuming parameter σ 2 =0.1,φ=0,δ=0.05,γ=5000,η=0.01,θ j =5000,γ=1.5,x min =0,x max =10000,κ=0.05,v=5000,λ=0.1; the training data set contains 100 samples, each sample contains 5 key influencing factors; assuming that a sample x i =5000, output variable y i =8000, prediction function f(x)=8000;

[0175]

[0176] Assume that the nonlinear performance value P of each key influencing factor j (x i )=0.8,P j-1 (x i )=0.7; Assume that the input x of all samples i and output y i are the same, and x i =5000;

[0177]

[0178] The final calculation result of the embodiment is 23.45158. Assuming that the threshold is set to 20, since the result is greater than the set threshold, this shows that the optimization model parameters can generate efficient and stable scheduling strategies when dealing with key influencing factors (such as weather conditions and changes in load demand), ensuring the economy and reliability of the microgrid under different uncertainty scenarios; through the above steps, the microgrid can achieve better energy management and scheduling in actual operation, thereby improving the overall operating efficiency and economic benefits.

[0179] 104. Based on the energy optimization scheduling scheme, continuously monitor the key indicators of microgrid energy management and optimization scheduling, perform real-time correction on the deviation between actual operation data and theory, and generate energy management and optimization scheduling strategies.

[0180] The energy optimization scheduling scheme is a microgrid optimization scheduling scheme generated by Bayesian optimization algorithm and grey system theory based on the preliminary energy management plan and uncertainty scenario scheduling strategy.

[0181] Key indicators include the output power of the power generation unit, the charge state of the energy storage system, changes in load demand, system efficiency, etc., which are used to evaluate the operating performance of the microgrid.

[0182] The actual operation data is the real-time data generated by the microgrid during its actual operation, including the actual output of the power generation unit, the actual charging and discharging status of the energy storage system, and the actual demand of the load.

[0183] Theoretical data is the expected data generated according to the energy optimization scheduling scheme, which is used to compare with the actual operation data. The energy management and optimization scheduling strategy is the final microgrid energy management and optimization scheduling scheme generated based on continuous monitoring and real-time correction.

[0184] In this step, firstly, based on the energy optimization scheduling scheme, a real-time monitoring system is deployed to continuously collect key indicator data of the microgrid, including the output power of the power generation unit, the charge state of the energy storage system, changes in load demand, etc.; secondly, the actual operation data is compared with the theoretical data to identify the deviation between the actual operation data and the theoretical data, which helps to find potential problems and improvement points; thirdly, according to the identified deviations, the operating parameters of the microgrid are adjusted in real time, such as the output power of the power generation unit, the charging and discharging strategy of the energy storage system, etc., to ensure that the microgrid always maintains the optimal state in actual operation; finally, based on the real-time correction results, the final energy management and optimization scheduling strategy is generated to ensure the efficient, stable and economical operation of the microgrid under various operating conditions, and provide a scientific basis for subsequent management and optimization.

[0185] Optionally, in step 104, based on the energy optimization scheduling scheme, the key indicators of microgrid energy management and optimization scheduling are continuously monitored, the deviation between actual operation data and theory is corrected in real time, and an energy management and optimization scheduling strategy is generated, including: based on the energy optimization scheduling scheme, the key indicators in the operation process of the microgrid are continuously monitored, and a real-time operation data set is generated in combination with the actual output power of each power generation unit; based on the real-time operation data set, a comparative analysis is performed with the theoretical expected value of the energy optimization scheduling scheme to identify the actual operation deviation and generate a deviation analysis report; based on the deviation analysis report, the microgrid operating parameters are adjusted in real time through an adaptive control strategy to reduce the actual operation deviation and generate real-time correction instructions; based on the real-time correction instructions, the scheduling parameters are dynamically adjusted in combination with the historical operation data and the current operation status of the microgrid, the operation strategy is optimized, and the energy management and optimization scheduling strategy is generated.

[0186] The energy optimization scheduling scheme is a microgrid optimization scheduling scheme generated based on the multi-objective evolutionary algorithm and the Bayesian optimization algorithm, including the output plan of the power generation unit, the charging and discharging strategy of the energy storage system, etc.

[0187] Key indicators include the actual output power of the power generation unit, the charge state of the energy storage system, changes in load demand, system efficiency, etc., which are used to evaluate the operating performance of the microgrid.

[0188] The real-time operation data set is the key indicator data continuously collected during the operation of the microgrid, which is used for comparative analysis with the theoretical expected values.

[0189] The deviation analysis report compares the actual operating data with the theoretical expected values, identifies the deviations and uses them to guide subsequent corrective measures.

[0190] The adaptive control strategy adjusts the microgrid operating parameters in real time according to the deviation analysis report to reduce the actual operating deviation.

[0191] The real-time correction instruction is a specific adjustment instruction generated based on the adaptive control strategy, which is used to adjust the operating parameters of the microgrid in real time.

[0192] Dynamically adjust dispatch parameters Based on historical operation data and current operation status, dynamically adjust dispatch parameters, optimize operation strategies, and ensure efficient and economical operation of microgrids.

[0193] In the embodiments of the present application, firstly, based on the energy optimization scheduling scheme, the key indicators in the operation process of the microgrid are continuously monitored, and a real-time operation data set is generated in combination with the actual output power of each power generation unit; secondly, based on the real-time operation data set, a comparative analysis is performed with the theoretical expected value of the energy optimization scheduling scheme, the actual operation deviation is identified, and a deviation analysis report is generated; thirdly, based on the deviation analysis report, the microgrid operation parameters are adjusted in real time through an adaptive control strategy to reduce the actual operation deviation, and a real-time correction instruction is generated; finally, based on the real-time correction instruction, the scheduling parameters are dynamically adjusted in combination with the historical operation data and the current operation status of the microgrid, the operation strategy is optimized, and an energy management and optimization scheduling strategy is generated.

[0194] Assume that in a temporary scientific research cabin in a remote area, it is necessary to ensure a stable power supply to the microgrid. First, based on the energy optimization scheduling scheme, continuously monitor key indicators during the operation of the microgrid, such as the actual output power of the solar power generation unit, the actual output power of the wind power generation unit, the charge state of the energy storage system, changes in load demand, etc., to generate a real-time operation data set. Secondly, based on the real-time operation data set, compare and analyze with the theoretical expected value of the energy optimization scheduling scheme. For example, it is found that the actual output power of the wind power generation unit is 15% lower than the expected value, while the charge state of the energy storage system is 8% lower than the expected value, and a deviation analysis report is generated. Thirdly, based on the real-time operation data set, the actual output power of the wind power generation unit is 15% lower than the expected value, while the charge state of the energy storage system is 8% lower than the expected value. Based on the deviation analysis report, the microgrid operating parameters are adjusted in real time through adaptive control strategies. For example, the output power of diesel generators is increased to make up for the lack of wind power generation, and the power consumption of non-critical loads is reduced to keep the charge state of the energy storage system within a safe range, and real-time correction instructions are generated. Finally, based on the real-time correction instructions, combined with the historical operation data of the microgrid and the current operation status, the scheduling parameters are dynamically adjusted. For example, the start-up threshold of the diesel generator and the charging and discharging strategy of the energy storage system are adjusted. At the same time, the load management strategy is optimized to ensure the power supply priority of key equipment, generate energy management and optimized scheduling strategies, and ensure the efficient and economical operation of the microgrid under different environmental conditions.

[0195] Through the above steps, the power supply of the temporary scientific research cabin was successfully guaranteed, and the efficiency and cost-effectiveness of the microgrid operation were greatly improved.

[0196] Figure 2 A schematic diagram of the structure of a microgrid energy management and optimization scheduling system in a shelter device is provided for an embodiment of the present application. Figure 2 As shown, the device comprises:

[0197] The collection module 21 is used to collect the real-time load demand of the shelter equipment, comprehensively evaluate the future energy supply and demand of the microgrid in combination with external environmental factors, and generate an energy supply and demand evaluation report;

[0198] A generation module 22 is used to generate a preliminary energy management plan based on the energy supply and demand assessment report, using a multi-objective evolutionary algorithm, comprehensively considering multiple optimization objectives, combining ensemble learning technology, combining multiple prediction results, and;

[0199] An adjustment module 23 is used to adjust the uncertainty scenario scheduling strategy based on the preliminary energy management plan, adopt a Bayesian optimization algorithm, combine the grey system theory analysis, dynamically analyze the working efficiency of the internal power generation unit of the microgrid, and generate an energy optimization scheduling plan;

[0200] The monitoring module 24 is used to continuously monitor the key indicators of microgrid energy management and optimal scheduling based on the energy optimization scheduling scheme, perform real-time correction on the deviation between actual operation data and theory, and generate energy management and optimal scheduling strategies.

[0201] Figure 2 The microgrid energy management and optimization scheduling system in the shelter equipment can perform Figure 1 The implementation principle and technical effect of the microgrid energy management and optimization scheduling method in a shelter device described in the embodiment shown will not be repeated. The specific way in which each module and unit performs operations in the microgrid energy management and optimization scheduling system in a shelter device in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0202] In one possible design, Figure 2 A microgrid energy management and optimization scheduling system in a shelter device of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0203] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0204] The processing component 32 is used to: collect the real-time load demand of the cabin equipment, combine external environmental factors, comprehensively evaluate the future energy supply and demand status of the microgrid, and generate an energy supply and demand evaluation report; based on the energy supply and demand evaluation report, use a multi-objective evolutionary algorithm, comprehensively consider multiple optimization goals, combine integrated learning technology, combine multiple prediction results, and generate a preliminary energy management plan; based on the preliminary energy management plan, use the Bayesian optimization algorithm to adjust the uncertainty scenario scheduling strategy, combine the gray system theoretical analysis, dynamically analyze the working efficiency of the internal power generation unit of the microgrid, and generate an energy optimization scheduling plan; based on the energy optimization scheduling plan, continuously monitor the key indicators of microgrid energy management and optimization scheduling, perform real-time correction on the deviation between actual operating data and theory, and generate energy management and optimization scheduling strategies.

[0205] The processing component 32 may include 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 may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0206] The storage component 31 is configured to store various types of data to support operations at 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 (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0207] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0208] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0209] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0210] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0211] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is a method for microgrid energy management and optimal scheduling in a shelter device.

[0212] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0213] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0214] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0215] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for energy management and optimal scheduling of microgrids in shelter equipment, characterized in that: include: Collect the real-time load demand of shelter equipment, combine external environmental factors, comprehensively evaluate the future energy supply and demand of the microgrid, and generate an energy supply and demand assessment report; Based on the energy supply and demand assessment report, a multi-objective evolutionary algorithm is used to comprehensively consider multiple optimization objectives, combined with ensemble learning technology, and multiple prediction results are combined to generate a preliminary energy management plan; Based on the preliminary energy management plan, the Bayesian optimization algorithm is used to adjust the uncertainty scenario scheduling strategy, and combined with the grey system theory analysis, the working efficiency of the internal power generation unit of the microgrid is dynamically analyzed to generate an energy optimization scheduling plan; Based on the energy optimization scheduling scheme, the key indicators of microgrid energy management and optimization scheduling are continuously monitored, the deviation between actual operation data and theory is corrected in real time, and energy management and optimization scheduling strategies are generated.

2. The method according to claim 1, characterized in that Based on the energy supply and demand assessment report, a multi-objective evolutionary algorithm is used to comprehensively consider multiple optimization objectives, combine ensemble learning technology, combine multiple prediction results, and generate a preliminary energy management plan, including: Based on the energy supply and demand assessment report and in combination with the actual operation requirements of the microgrid, multi-dimensional optimization objectives are set to generate an optimization objective set; Based on the optimization target set, a multi-objective evolutionary algorithm is used to balance multiple optimization targets, search for target solutions in the solution space, perform multiple rounds of iterative optimization, and generate an optimization target solution set; Based on the optimization target solution set, combined with ensemble learning technology, multiple prediction results are combined by weighted average to generate a comprehensive prediction result; Based on the comprehensive prediction results, multiple optimization objectives are comprehensively considered, the overall energy management performance is evaluated, and a preliminary energy management plan is generated.

3. The method according to claim 2, characterized in that Based on the optimization target set, a multi-objective evolutionary algorithm is used to balance multiple optimization targets, search for target solutions in the solution space, perform multiple rounds of iterative optimization, and generate an optimization target solution set, including: Based on the optimization target set, an individual representation method is defined, different energy management strategies are individually encoded, and an initial individual population is generated; Based on the initial individual population, a fitness function is designed, a multi-objective evolutionary algorithm is adopted, multiple optimization objectives are comprehensively considered, the fitness of individuals in the initial individual population is evaluated, and a fitness evaluation result is generated; Based on the fitness evaluation results, performing a variety of evolutionary operations on the individuals in the initial individual population, and generating an optimized individual population through multiple rounds of iterative optimization; Based on the optimized individual population and in combination with the Pareto optimality principle, it is ensured that there is no obvious disadvantage among multiple optimization objectives, a group of non-dominated solutions is selected, and an optimization objective solution set is generated.

4. The method according to claim 3, characterized in that Based on the initial individual population, a fitness function is designed, a multi-objective evolutionary algorithm is adopted, multiple optimization objectives are comprehensively considered, and fitness evaluation is performed on individuals in the initial individual population to generate fitness evaluation results, including: Based on the initial individual population, identifying the time trend of the individual performance value under each optimization target; Smoothing the individual performance values ​​using an exponential smoothing method to reduce the impact of short-term fluctuations to generate nonlinear performance values; The nonlinear performance value is calculated using the following formula: Among them, P i (x) is the nonlinear performance value of the individual under the i-th optimization goal; x is the performance value of the individual under the i-th optimization goal; α is the parameter that controls the steepness of the S-type function; β is the expected value of the performance value; δ is the parameter that controls the amplitude change of the sine waveform; γ is the center position parameter of the sine waveform; φ is the phase shift of the sine waveform; λ is the parameter that controls the steepness of the additional S-type function; μ is the center position parameter of the additional S-type function; Based on the nonlinear performance value, combined with the importance and priority of each optimization target, a reasonable weight is assigned, a cosine function and an additional S-type function are introduced, and nonlinear combination adjustment is performed to generate a comprehensive fitness value; The comprehensive fitness value is calculated by the following formula: Among them, F(x) is the comprehensive fitness value of the individual under multiple optimization objectives; x is the performance value of the individual under the i-th optimization objective; P i (x) is the nonlinear performance value of the individual under the i-th optimization goal; P i-1 (x) is the nonlinear performance value of the individual under the i-1th optimization objective; w i is the weight of the i-th optimization goal, and the sum of the weights is 1; γ is the adjustment factor, which is used to adjust the contribution of the performance values ​​between different optimization goals; η is the parameter for controlling the rate of change of the difference; θ is the central position parameter of the rate of change of the difference; x min and x max are the minimum and maximum performance values ​​of the individual under all optimization objectives, respectively; κ is the parameter that controls the steepness of the additional S-shaped function; v is the central position parameter of the additional S-shaped function; i is the index of the optimization objective, from 1 to n; n is the total number of optimization objectives; Based on the comprehensive fitness value, normalization processing is performed, individuals with high comprehensive fitness values ​​are screened to determine the evolutionary direction of the individual population, the stability of the optimization process is ensured through a convergence test, and a fitness evaluation result is generated.

5. The method according to claim 2, characterized in that: The optimization target solution set is combined with ensemble learning technology to generate a comprehensive prediction result by combining multiple prediction results by weighted average, including: Based on the optimization target solution set, evaluating the performance balance in the multi-objective optimization process, selecting representative optimization target solutions, and generating a basic input set; Based on the basic input set, multiple prediction models are trained to target specific optimization goals, and prediction and generalization capabilities are verified during the training process to generate a multi-model prediction system; Based on the multi-model prediction system, a variety of optimization objectives are comprehensively considered, the future operation state of the microgrid is predicted from multiple perspectives, and a multi-model prediction result set is generated; Based on the multi-model prediction result set, combined with the historical prediction accuracy of the multiple prediction models, different weights are assigned to perform weighted combination to generate a comprehensive prediction result.

6. The method according to claim 1, characterized in that Based on the preliminary energy management plan, the Bayesian optimization algorithm is used to adjust the uncertainty scenario scheduling strategy, and the working efficiency of the internal power generation unit of the microgrid is dynamically analyzed in combination with the grey system theory analysis to generate an energy optimization scheduling plan, including: Based on the preliminary energy management plan, analyzing the operating status of the microgrid in different uncertainty scenarios, preliminarily adjusting the scheduling strategy through scenario analysis, and generating a scheduling strategy set for uncertainty scenarios; Based on the uncertainty scenario scheduling strategy set, the Bayesian optimization algorithm is used to dynamically adjust by constructing a Bayesian probability model to ensure efficient and economic operation of microgrids in different uncertainty scenarios and generate an optimized scheduling strategy set; Based on the optimized dispatch strategy set and combined with grey system theory analysis, the working efficiency of the power generation units in the microgrid under different load conditions is dynamically analyzed to generate a performance evaluation report of the power generation units; Based on the performance evaluation report of the power generation unit and combined with the volatility of renewable energy output, multi-factor fusion optimization is performed to generate an energy optimization scheduling plan.

7. The method according to claim 6, characterized in that The scheduling strategy set based on the uncertainty scenario is dynamically adjusted by using the Bayesian optimization algorithm and constructing a Bayesian probability model to ensure efficient and economic operation of the microgrid in different uncertainty scenarios and generate an optimized scheduling strategy set, including: Based on the uncertainty scenario scheduling strategy set, analyze the operating status of the microgrid under different uncertainty scenarios, identify key influencing factors, and generate a list of key influencing factors; Based on the list of key influencing factors, a Bayesian optimization algorithm is used to construct a Bayesian probability model, dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system, and generate a dynamic adjustment scheduling strategy; Based on the dynamic adjustment scheduling strategy and combined with the actual operation performance, the operation efficiency and economy of the microgrid under different uncertainty scenarios are evaluated to generate evaluation results; Based on the evaluation results, the efficient and economic operation of the microgrid in different uncertainty scenarios is ensured, and optimization and adjustment are performed through iterative processing to generate an optimized dispatch strategy set.

8. The method according to claim 7, characterized in that Based on the list of key influencing factors, the Bayesian optimization algorithm is used to construct a Bayesian probability model to dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system, and generate a dynamic adjustment scheduling strategy, including: Based on the list of key influencing factors, multi-scale decomposition of input variables is performed through multi-resolution analysis to extract features of different time scales, and the correlation coefficients between input variables and output variables are combined to generate conditional probabilities; The conditional probability is calculated using the following formula: Where p(y|x,θ) is the conditional probability of output y given input x and parameter θ; f(x) is the prediction function in the Bayesian probability model; σ 2 is the variance of the model; φ is the phase shift of the sine waveform; δ is the parameter that controls the amplitude change of the sine waveform; γ is the center position parameter of the sine waveform; x is the input variable; θ is the model parameter; y is the output variable; Based on the conditional probability, a log-likelihood function is constructed, a regularization term is introduced to prevent overfitting, and the model flexibility is enhanced through nonlinear transformation to generate optimized model parameters; The optimization model parameters are calculated using the following formula: Among them, θ * To optimize the model parameters; p(y i ∣x i ,θ) is the given input x i and parameter θ, the output y i The conditional probability of; i is the index of the sample, from 1 to N; N is the number of samples; λ is the regularization parameter; j is the index of the key influencing factor, from 1 to M; P j (x i ) is the jth key influencing factor in the input x i The nonlinear performance value under P j-1 (x i ) is the j-1th key influencing factor in the input x i The nonlinear performance value under the condition of η is the parameter for controlling the rate of change of the difference; θ is the parameter for controlling the rate of change of the difference; j is the central position parameter of the jth key influencing factor; γ is the adjustment factor; x min and x max are the minimum and maximum values ​​of all inputs respectively; k is the parameter that controls the steepness of the additional S-shaped function; v is the central position parameter of the additional S-shaped function; x i is the input variable of the i-th sample; y i is the output variable of the i-th sample; θ is the model parameter; Based on the optimization model parameters, new output variables are predicted to dynamically adjust the output power of each power generation unit and the charging and discharging strategy of the energy storage system. Simulation software is used to simulate and monitor the operation of the microgrid in different uncertainty scenarios to generate a dynamic adjustment scheduling strategy.

9. The method according to claim 1, characterized in that: Based on the energy optimization scheduling scheme, the key indicators of microgrid energy management and optimization scheduling are continuously monitored, the deviation between actual operation data and theory is corrected in real time, and energy management and optimization scheduling strategies are generated, including: Based on the energy optimization scheduling scheme, key indicators in the operation process of the microgrid are continuously monitored, and a real-time operation data set is generated in combination with the actual output power of each power generation unit; Based on the real-time operation data set, a comparative analysis is performed with the theoretical expected value of the energy optimization scheduling scheme to identify actual operation deviations and generate a deviation analysis report; Based on the deviation analysis report, the microgrid operating parameters are adjusted in real time through an adaptive control strategy to reduce the actual operating deviation and generate a real-time correction instruction; Based on the real-time correction instructions, combined with the historical operation data and current operation status of the microgrid, the scheduling parameters are dynamically adjusted, the operation strategy is optimized, and the energy management and optimization scheduling strategy is generated.

10. A microgrid energy management and optimization scheduling system in a shelter equipment, characterized in that: include: The collection module is used to collect the real-time load demand of the shelter equipment, combine external environmental factors, comprehensively evaluate the future energy supply and demand of the microgrid, and generate an energy supply and demand assessment report; A generation module is used to generate a preliminary energy management plan based on the energy supply and demand assessment report, using a multi-objective evolutionary algorithm, comprehensively considering multiple optimization objectives, combining ensemble learning technology, combining multiple prediction results, and combining multiple prediction results; An adjustment module is used to adjust the uncertainty scenario scheduling strategy based on the preliminary energy management plan, adopt a Bayesian optimization algorithm, combine the grey system theory analysis, dynamically analyze the working efficiency of the internal power generation unit of the microgrid, and generate an energy optimization scheduling plan; The monitoring module is used to continuously monitor the key indicators of microgrid energy management and optimal scheduling based on the energy optimization scheduling scheme, perform real-time correction on the deviation between actual operation data and theory, and generate energy management and optimal scheduling strategies.

Images