Industrial and commercial energy storage ems scheduling system based on multi-objective optimization

Abstract

The application discloses a commercial and industrial energy storage EMS scheduling system based on multi-objective optimization and relates to the intelligent control field of energy storage systems. The application is used for solving the problem of insufficient adaptability of battery life attenuation, electricity price fluctuation and load mutation in energy storage scheduling. A life management module analyzes ion migration state to generate a dynamic power safety envelope line as a hierarchical optimization hard constraint; a feature extraction module constructs an electricity price correlation network, combines quantile regression to predict photovoltaic / load and extracts net load steep change features; a hierarchical optimization module compresses the depth of charging and discharging and activates load adjustment when demand risk occurs, optimizes economy, life loss and power grid constraints through tabu search multi-objective evolution synchronization, an execution arbitration module monitors the distance between power and the envelope line in real time, and when the distance is below the threshold, rolling optimization is started to improve the life weight and feedback lithium feature driven envelope line update, forming a closed loop protection.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent control of energy storage systems, in particular to an EMS scheduling system for industrial and commercial energy storage based on multi-objective optimization. BACKGROUND

[0002] With large-scale access of renewable energy to the power grid, industrial and commercial energy storage systems play an important role in load balancing, grid stability and economic scheduling. Through adjusting the charging and discharging behavior, the energy storage system smooths the output fluctuation of photovoltaic and other new energy, reduces the cost of electricity, and improves the reliability of the power grid. However, the complex electrochemical process inside the battery in actual operation leads to safety risks and life attenuation, and the variability of load fluctuation and price mechanism puts higher real-time response and adaptability requirements on the scheduling strategy.

[0003] Existing energy storage scheduling systems are mostly based on a single economic indicator or a simplified battery model, ignoring the influence of the internal ion migration dynamics of the battery on the safety boundary. The charging and discharging strategy cannot be adjusted in real time when facing battery performance degradation, and there is a risk of accelerating battery aging and safety hazards. At the same time, the load prediction model responds slowly to sudden load changes, lacks fusion of multi-source information and dynamic feature extraction, resulting in a lack of robustness of the scheduling strategy. At the level of scheduling algorithm, the existing methods are difficult to effectively balance the economic benefit, life protection and grid constraints, and there are problems such as low efficiency of scheduling decision, incomplete constraint conditions, and lack of dynamic adjustment mechanism for strategy execution, which limits the improvement of the overall performance and safety guarantee capability of industrial and commercial energy storage systems. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides an EMS scheduling system for industrial and commercial energy storage based on multi-objective optimization, which solves the problems in the above background art.

[0005] In order to achieve the above object, the application is realized by the following technical scheme: the EMS scheduling system for industrial and commercial energy storage based on multi-objective optimization comprises the following modules: a life management module, a feature extraction module, a hierarchical optimization module and an execution arbitration module; the life management module is used for real-time analysis of the ion migration state inside the battery, calculation of the lithium precipitation risk coefficient according to the ion concentration gradient distribution and dynamic generation of the charge and discharge power safety envelope line, and embedding of the envelope line into the hierarchical optimization module as a hard constraint condition; the feature extraction module is used for dynamic construction of the electricity price rule association network through transfer learning, generation of the confidence interval prediction curve of photovoltaic and load combined with the quantile regression forest, and extraction of the steep feature sequence of the net load change rate based on the dynamic time warping algorithm; the hierarchical optimization module is used for dynamic compression of the upper limit of the battery discharge depth and activation of the load hierarchical regulation when it is detected that the net load change rate in the continuous period exceeds the set threshold value and the contract demand margin is lower than the safety margin, synchronous optimization of the economy target, the life loss target and the power grid constraint target in the safety envelope line constraint range through the tabu search guided multi-objective evolutionary algorithm, and generation of the three-target balanced charge and discharge strategy; and the execution arbitration module is used for real-time monitoring of the distance between the charge and discharge power and the safety envelope line, starting of the rolling optimization control when the distance is lower than the warning threshold, increasing of the life protection weight and relaxation of the economy constraint according to the preset proportion, and feedback of the actual lithium precipitation risk feature to the life management module to drive the online update of the safety envelope line.

[0006] Further, the specific process of real-time analysis of the ion migration state inside the battery, calculation of the lithium precipitation risk coefficient according to the ion concentration gradient distribution and dynamic generation of the charge and discharge power safety envelope line is as follows: real-time acquisition of the internal sensor data stream of the battery, generation of the positive electrode region ion concentration gradient distribution graph through the ion transport equation, dynamic calculation of the local ion flux reference value; analysis of the concentration mutation nodes in the gradient distribution graph, calculation of the real-time deviation degree of the measured ion flux in the region from the reference value; mapping of the deviation degree to the lithium precipitation risk coefficient, adjustment of the upper limit boundary of the charge and discharge power according to the risk growth rate when the risk coefficient exceeds the threshold value, generation of the power safety envelope line continuously evolving with time, and positive correlation of the envelope line boundary slope with the risk coefficient increment.

[0007] Further, the specific process of embedding the envelope line into the hierarchical optimization module as a hard constraint condition is as follows: loading of the safety envelope line parameters in the initialization stage of the hierarchical optimization module, construction of the charge and discharge power constraint matrix, real-time detection of the power trajectory of the candidate strategy in the optimization iteration process, marking of the trajectory exceeding the envelope line boundary as an infeasible solution through the tabu search mechanism; starting of the adaptive penalty function for the boundary adjacent solution, increasing of the life loss penalty term in the objective function in inverse proportion to the distance from the boundary; and performing the envelope line compliance verification before outputting the final strategy, and performing secondary smoothing processing on the boundary point strategy to ensure the continuity of the power change rate.

[0008] Further, the specific process of dynamically constructing the electricity price rule association network through transfer learning is as follows: the electricity price formation mechanism data is extracted from the historical database of the electricity market, the correlation characteristics between different electricity price mechanisms are analyzed through the pre-trained prediction model, and the electricity price rule knowledge framework is constructed; based on the price structure characteristics of the target region, the localized rule network is adapted through graph structure transfer learning, multi-source operation data is integrated, and the prediction model is trained at multiple confidence levels through the quantile regression forest algorithm to generate the probability distribution curve of photovoltaic output and load.

[0009] Further, and based on the dynamic time warping algorithm, the specific process of extracting the steep change feature sequence of the net load change rate is as follows: the net load mutation events are selected from the historical data, the standardized waveform is extracted as the steep change benchmark template, the dynamic time warping distance between the current curve and the template is calculated in real time, and when the distance is lower than the similarity threshold, the candidate feature area is marked; the change rate derivative is calculated in the feature area, and if the absolute value of the change rate of a plurality of consecutive sampling points exceeds the historical fluctuation range, it is determined as an effective steep change feature, and the timestamp, duration, and maximum change rate of the feature area are encapsulated as a feature vector sequence; wherein the steep change includes steep rise and steep fall.

[0010] Further, when it is detected that the net load change rate in the continuous period exceeds the set threshold and the contract demand reserve is lower than the safety margin, the specific process of dynamically compressing the upper limit of battery discharge depth and activating load hierarchical regulation is as follows: based on the steep change feature sequence output by the feature extraction module, the net load change rate of a plurality of consecutive sampling points exceeding the upper limit of the historical normal fluctuation range is identified, the consumed demand in the contract period is calculated by synchronous real-time integration, the remaining demand occupation trend is deduced in combination with the prediction curve, when the reserve is lower than the safety margin and there is a steep change feature, the demand risk state is triggered; the upper limit of discharge depth is compressed in stages according to the risk duration, the initial stage reduces the fixed proportion of the basic discharge depth, and each additional high-risk period adds the compression amplitude, the load priority management mechanism is activated synchronously, the preset non-critical load category is immediately cut off, the activation time window of the adjustable load is delayed, and the peak pressure of the power grid is reduced through the coordination of the load side.

[0011] Further, within the safety envelope constraint range, the specific process of synchronously optimizing the economy target, life loss target and power grid constraint target by tabu search guided multi-objective evolutionary algorithm to generate a three-objective balanced charging and discharging strategy is as follows: the power safety envelope is converted into a time-varying inequality constraint set, and the charging and discharging power feasible region boundary is defined; the initial solution set is generated by tabu search algorithm, ensuring that the candidate strategy is distributed inside the feasible region, and the three objective function values are synchronously evaluated for each candidate scheme, the economy target calculates the total of time-of-use electricity cost and demand penalty, the life loss target quantifies the cumulative effect of charging and discharging rate stress, and the power grid constraint target evaluates the peak power and fluctuation rate penalty, and the non-dominated sorting is used to screen the Pareto optimal solution set; the solution approaching the safety envelope boundary is injected with an adaptive penalty term inversely proportional to the boundary distance, and the tabu list is started for the out-of-bound solution to block repeated generation, and the output charging and discharging strategy is a structured instruction sequence, including: the charging and discharging power value corresponding to the time stamp sequence, and the load grading regulation action time sequence list.

[0012] Further, when the distance is less than the warning threshold, the specific process of starting the rolling optimization control and increasing the life protection weight by a preset proportion and relaxing the economy constraint is as follows: the minimum Euclidean distance from the actual power point to the envelope boundary is continuously calculated, and the rolling optimization mechanism is triggered when the distance value falls within the preset warning interval; within the rolling time window, the multi-objective function is reconstructed, the life protection target weight is increased to a fixed multiple of the baseline value, and the allowed deviation range of the economy constraint is simultaneously relaxed, the power derating instruction is dynamically generated by the quadratic programming solver, the derating amplitude is negatively exponential to the distance value, and the power trajectory is ensured to converge safely along the envelope.

[0013] Further, the actual lithium precipitation risk characteristics are fed back to the life management module to drive the specific process of online updating of the safety envelope as follows: real-time acquisition of charging and discharging current waveform, temperature distribution and voltage fluctuation data, reconstruction of the battery internal ion concentration gradient field based on impedance spectrum analysis, positioning of the diffusion delay abnormal area and quantification of the local ion flux deviation; the measured ion flux deviation is input into the electrochemical parameter estimation algorithm, the mobility and diffusion coefficient in the ion transport equation are updated, the lithium precipitation risk level is recalculated according to the updated ion transport equation, and the power safety boundary curve continuously evolving with time is generated; the new envelope parameters are input into the constraint library of the hierarchical optimization module in real time through a dynamic interface, the original boundary values are replaced, and the strategy re-verification is triggered.

[0014] The present application has the following beneficial effects:

[0015] (1) Based on multi-objective optimization of industrial and commercial energy storage EMS scheduling system, the life management module analyzes the ion migration state in real time to generate a dynamic power safety envelope, accurately quantifies the lithium precipitation risk coefficient, and converts the invisible concentration gradient anomaly in the battery into an executable charge and discharge power boundary constraint, fundamentally avoiding the cumulative problem of implicit damage caused by the lag of traditional voltage / temperature protection. The feature extraction module fuses transfer learning to build a price rule correlation network, solving the cross-regional price policy adaptation problem; the quantile regression forest generates a photovoltaic / load prediction curve with a confidence interval, quantifying the prediction uncertainty boundary; the dynamic time warping algorithm accurately captures the steep change characteristics of the net load, and the steep rise and steep fall event identification accuracy is improved by more than 40% compared with the traditional threshold method, providing reliable input for risk response.

[0016] (2) Based on multi-objective optimization of industrial and commercial energy storage EMS scheduling system, in the hierarchical optimization module, for the risk state that the continuous net load change rate exceeds the threshold and the contract demand reserve is insufficient, the upper limit of the battery discharge depth is dynamically compressed and the load hierarchical regulation is activated, effectively relieving the peak pressure of the power grid and the demand risk, realizing the multi-objective balanced optimization of economy, life loss and grid constraints, and guaranteeing the safety and economy of the charge and discharge strategy. The execution arbitration module monitors the distance between the charge and discharge power and the safety envelope in real time, starts the rolling optimization control, dynamically adjusts the life protection weight and economic constraints, combines the real-time feedback of the lithium precipitation risk characteristics, realizes the online adaptive correction of power scheduling, further improves the robustness and reliability of the system, prolongs the service life of the battery, and optimizes the overall scheduling benefit.

[0017] Of course, implementing any product of the present application does not necessarily require all the advantages described above to be achieved simultaneously. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 The flowchart of the industrial and commercial energy storage EMS scheduling system based on multi-objective optimization of the present application. DETAILED DESCRIPTION

[0019] The industrial and commercial energy storage EMS scheduling system based on multi-objective optimization of the present application effectively solves the problems of large battery life attenuation risk, insufficient economy and load fluctuation adaptability in the charge and discharge scheduling process of the energy storage system, overcomes the core problems of difficult control of battery implicit damage, multi-objective conflict leading to optimization rigidity and disconnection between prediction and execution in traditional energy storage scheduling, and realizes the coordinated optimization of safety, economy and grid constraints.

[0020] The scheme in the embodiments of the present application has the following overall idea:

[0021] Firstly, the life management module is used to analyze the ion migration state in the battery in real time, the lithium precipitation risk coefficient is dynamically calculated based on the ion concentration gradient, the charging and discharging power safety envelope is generated, and the envelope is embedded as a hard constraint into the scheduling optimization process. Secondly, the feature extraction module uses transfer learning and quantile regression forest to dynamically predict the electricity price rule and photovoltaic load, and extracts the steep change feature of the net load change rate based on the dynamic time warping algorithm, to provide accurate data support for the scheduling strategy. On this basis, the hierarchical optimization module combines the tabu search guided multi-objective evolutionary algorithm to optimize the economy, life loss and grid constraint targets simultaneously under the constraint of the safety envelope, dynamically adjusts the battery discharge depth and activates the load hierarchical regulation to relieve the grid pressure. Finally, the execution arbitration module monitors the distance between the power and the safety envelope in real time, starts the rolling optimization control, dynamically adjusts the weight proportion and feeds back the lithium precipitation risk to realize online adaptive scheduling and strategy correction.

[0022] Please refer to Figure 1 The embodiment of the present application provides a technical scheme: a multi-objective optimization-based industrial and commercial energy storage EMS scheduling system, comprising the following modules: a life management module, a feature extraction module, a hierarchical optimization module, and an execution arbitration module. The life management module is used to analyze the ion migration state in the battery in real time, calculate the lithium precipitation risk coefficient according to the ion concentration gradient distribution, and dynamically generate a charging and discharging power safety envelope. The envelope is embedded into the hierarchical optimization module as a hard constraint condition. The feature extraction module is used to dynamically construct an electricity price rule association network through transfer learning, generate photovoltaic and load confidence interval prediction curves in combination with quantile regression forest, and extract a steep change feature sequence of the net load change rate based on the dynamic time warping algorithm. The hierarchical optimization module is used to dynamically compress the upper limit of the battery discharge depth and activate the load hierarchical regulation when it is detected that the net load change rate in a continuous period exceeds a set threshold and the contract demand margin is lower than a safety margin. Within the constraint range of the safety envelope, the multi-objective evolutionary algorithm guided by tabu search is used to simultaneously optimize the economy target, the life loss target, and the grid constraint target, to generate a three-target balanced charging and discharging strategy. The execution arbitration module is used to monitor the distance between the charging and discharging power and the safety envelope in real time. When the distance is lower than an alarm threshold, the rolling optimization control is started, the life protection weight is increased by a preset proportion, the economy constraint is relaxed, and the actual lithium precipitation risk feature is fed back to the life management module to drive the online update of the safety envelope.

[0023] In this embodiment, the life management module is responsible for real-time analysis of the ion migration state inside the battery. During charging and discharging, the migration of lithium ions in electrode materials can trigger a series of physical and chemical changes, which may lead to battery performance degradation or even safety hazards. This module collects internal sensor data and uses ion transport equations to calculate the ion concentration gradient distribution in the positive electrode region, and then obtains the risk coefficient of lithium precipitation. Lithium precipitation refers to the deposition of metallic lithium on the surface or inside of the electrode, which may cause short circuits and life loss. Based on the risk coefficient, a safe envelope of charging and discharging power is dynamically generated, which is a safety boundary that limits the power output of the battery, ensuring that the charging and discharging process meets performance requirements while avoiding accelerating battery aging. This safety envelope is embedded as a hard constraint in the hierarchical optimization module, ensuring that the optimization results are within the battery safety range. The feature extraction module is used to analyze the external operating environment and load characteristics in depth, improving the prediction accuracy and response capability of the scheduling strategy. Through the transfer learning method, a price rule correlation network is dynamically constructed. Transfer learning is a machine learning technique that can quickly adapt to changes in electricity price rules in the target region using existing regional or scenario data experience, enhancing the model's generalization ability. Combined with quantile regression forest algorithm, the confidence interval prediction curve of photovoltaic power generation and load is generated. Quantile regression forest is a statistical learning method that can predict the distribution interval of output variables at different confidence levels, reflecting the fluctuation range of future power and load. At the same time, based on the dynamic time warping (DTW) algorithm, the steep feature sequence of the net load change rate is extracted. DTW is a method for measuring the similarity of time series, which can identify rapid and significant changes in net load and warn of grid operation risks. The hierarchical optimization module is responsible for generating multi-objective scheduling strategies. When the net load change rate exceeds the preset threshold for consecutive time periods and the contract demand reserve is below the safety margin, the upper limit of battery discharge depth is dynamically compressed, limiting the maximum discharge amount of the battery and reducing the risk of battery loss. At the same time, load grading adjustment is activated to manage non-critical loads and adjustable loads through hierarchical management, achieving coordinated response on the load side and reducing grid peak pressure. During optimization, a multi-objective evolutionary algorithm guided by tabu search is used to simultaneously optimize economy (such as electricity cost), life loss, and grid constraints (such as peak-valley load balance), generating a charging and discharging strategy that meets the safety envelope constraint and achieves a balance of the three objectives. Tabu search is a local search-based optimization algorithm that avoids getting stuck in local optima by remembering taboo solutions. The execution arbitration module is responsible for real-time monitoring of the distance between the current charging and discharging power and the safety envelope, and dynamically adjusting the scheduling strategy. When the power trajectory distance from the envelope line boundary is below the warning threshold, a rolling optimization control mechanism is started, which performs short-period optimization adjustments based on the latest state.By increasing the lifetime protection weight according to a preset ratio, the module enhances the priority protection of battery safety while appropriately relaxing economic constraints, ensuring that the scheduling strategy remains economical under the premise of ensuring safety. In addition, the module feeds back the actual lithium plating risk characteristics collected in real time to the lifetime management module, driving the online dynamic update of the safety envelope, forming a closed-loop control to ensure the dynamic safety of charge and discharge scheduling.

[0024] Specifically, the process of real-time analysis of the ion migration state inside the battery, calculation of the lithium plating risk coefficient based on the ion concentration gradient distribution, and dynamic generation of the charge / discharge power safety envelope is as follows: Real-time acquisition of sensor data streams inside the battery; generation of an ion concentration gradient distribution map in the positive electrode region through the ion transport equation; dynamic calculation of the local ion flux benchmark value; analysis of concentration abrupt change nodes in the gradient distribution map; calculation of the real-time deviation between the measured ion flux in the region and the benchmark value; mapping the deviation to a lithium plating risk coefficient; when the risk coefficient exceeds the threshold, adjusting the upper limit boundary of the charge / discharge power according to the risk growth rate; generating a power safety envelope that evolves continuously over time; the slope of the envelope boundary is positively correlated with the increment of the risk coefficient.

[0025] In this implementation scheme, real-time data acquisition of internal battery sensor data streams is achieved through multiple sensors deployed inside the positive electrode of the battery to collect real-time battery operating status data, including but not limited to: ion concentration. Voltage distribution Temperature field Data is transmitted to the computing unit via a high-speed data acquisition system to ensure that the time resolution meets the dynamic response requirements. An ion concentration gradient distribution map of the cathode region is generated using ion transport equations. Based on electrochemical kinetics and diffusion theory, a one-dimensional ion transport-diffusion equation is used to describe the migration process of lithium ions in the cathode particles. ;in: Electrode position lithium ion concentration; :time; Effective diffusion coefficient (m² / s), considering electrode material porosity and barrier factors; Faraday constant; Local current density describes the driving force of ion migration. Boundary conditions are determined based on the battery current input and reaction rate. The spatial-temporal distribution of ion concentration is obtained by numerically solving the partial differential equation. A baseline value for local ion flux is dynamically calculated. Defined as: ; benchmark value Obtained through statistical analysis of historical steady-state operating data, the calculation formula is as follows: ;in: Number of sampling points within the statistical period; :No. ion flux at a sampling time. The benchmark value reflects the ion flux distribution under normal working conditions, used for comparison cases. Analyzing the concentration mutation nodes in the gradient distribution map, calculating the real-time deviation degree of the measured ion flux and the benchmark value, and aiming at the mutation nodes of the ion concentration gradient , the deviation degree is defined as: ; wherein: is a dimensionless deviation degree, reflecting the relative deviation of the current flux and the normal flux. Map the deviation degree to the risk coefficient of lithium precipitation, define the risk coefficient of lithium precipitation as the weighted cumulative function of the deviation degree: ; wherein: : the number of concentration mutation nodes; : the weight coefficient of the th node, reflecting the contribution of the node to the overall risk, the weight coefficient is determined by historical corrosion experiment data, satisfying the normalization condition ; : threshold function, used for nonlinear mapping of deviation degree, defined as: ; : deviation threshold value, the threshold determination method is based on statistical analysis of the distribution of deviation degree corresponding to the safety boundary in historical data, usually selecting a deviation critical value that can effectively distinguish between normal and risk states. According to the risk coefficient, adjust the upper limit of the charge and discharge power to generate a dynamic power safety envelope. The power safety envelope adjusts the benchmark maximum power : ; wherein: : the maximum charge and discharge power of the battery design; : adjustment coefficient, controls the influence intensity of risk on power limit, determined based on experimental fitting; : exponential weight, reflecting the nonlinear relationship between risk growth rate and power boundary shrinkage, also obtained by fitting historical risk event data. This formula ensures that the upper limit of charge and discharge power gradually shrinks as the risk of lithium precipitation increases, preventing excessive damage to the battery. The relationship between the slope of the envelope boundary and the increment of the risk coefficient, the change rate (slope) of the power envelope boundary is defined as: ; wherein, : power safety envelope; : the maximum charge and discharge power of the battery design; : lithium precipitation risk coefficient; the increment of the boundary slope and the risk coefficient are positively correlated, reflecting the rapid tightening of power limit when risk accelerates, ensuring the safety of the battery. Through the combination of physical modeling, statistical analysis and dynamic adjustment, real-time evaluation of the micro risk inside the battery and dynamic mapping of external scheduling constraints are realized, ensuring the safe operation of the energy storage system and the optimization of the battery life.

[0026] Specifically, the specific process of embedding the envelope line into the hierarchical optimization module as a hard constraint condition is as follows: the safety envelope line parameters are loaded in the initialization stage of the hierarchical optimization module, the charge and discharge power constraint matrix is constructed, the power trajectory of the candidate strategy is detected in real time in the optimization iteration process, and if the trajectory exceeds the envelope line boundary, it is marked as an infeasible solution through the taboo search mechanism; an adaptive penalty function is started for the boundary-adjacent solution, and a life loss penalty term proportional to the distance from the boundary is added in the objective function; the final strategy is output before the envelope line compliance verification is performed, and the boundary point strategy is subjected to secondary smoothing processing to ensure the continuity of the power change rate.

[0027] In the embodiment, the safety envelope line parameters are loaded in the initialization stage of the hierarchical optimization module, and when the system is started, the power safety envelope line parameter sequence at the current time is obtained from the life management module, denoted as , wherein is the number of discrete time steps in the optimization period, represents the maximum allowable power boundary corresponding to the time step. The charge and discharge power constraint matrix is constructed, the constraint matrix is constructed based on the envelope line parameters , the constraint vector is constructed, and the charge and discharge power trajectory is ensured to satisfy: ; specifically: ; wherein: is the negative transformation matrix of the identity matrix; is the boundary negative vector, which ensures that the power trajectory does not exceed the safety envelope line. In the optimization iteration process, the power trajectory of the candidate strategy is detected in real time, and in the multi-objective optimization iteration, the power sequence corresponding to the candidate solution is generated each time , and it is judged in real time whether the hard constraint is satisfied: ; if there is any time such that: , the candidate solution is marked as an infeasible solution and is tabooed in the taboo search list. An adaptive penalty function is started for the boundary-adjacent solution, the boundary distance vector is defined as: ; wherein the judgment condition for the boundary adjacency is: ; the parameters are explained as follows: : the distance of the candidate solution at the time from the envelope line boundary; : distance threshold, representing the critical distance range. The penalty function is defined as: ; wherein: : the penalty term value, the greater the value, the closer to the boundary and the higher the risk; Time step weight coefficient, reflecting the difference of the influence of different time periods on the life, the weight is determined based on the battery load sensitivity analysis, meeting the normalization condition ; Penalty intensity coefficient, controlling the weight of the penalty term in the objective function, adjusted through cross-validation; Small constant to prevent division by zero. The penalty function is inversely proportional to the boundary distance, the smaller the distance, the greater the penalty, effectively inhibiting the strategy from approaching the safety envelope too much. Add a life loss penalty term to the objective function, and the objective function of multi-objective optimization is adjusted as: ; Where: is the original multi-objective weighted function of economy, life and grid constraints; is the boundary distance penalty term, which is an approximate expression of the life loss risk. Perform envelope compliance verification before executing the final strategy, and perform boundary compliance detection on the final optimized strategy power sequence to ensure that: ; For boundary near points, use a quadratic smoothing algorithm to adjust the power trajectory to ensure continuous power change rate and prevent rapid jumps from causing battery and grid shocks. The smoothing process satisfies: ; The condition is: ; Where, represents the sum of the squares of the differences between the power sequence at adjacent times, which is used to measure the smoothness of the power change, and the smaller the value, the more gradual the change; K: total time step, representing the number of discrete divisions of the entire optimization period. : the charge and discharge power value at the final optimization time k, representing the actual dispatching power output of the energy storage system at that time; : the upper limit value of the power safety envelope at time k, derived from the dynamic calculation of the life management module, representing the maximum safe charge and discharge power allowed at that time. Key parameters determination, distance threshold δ: set according to the fatigue characteristics of battery materials and historical failure data, usually in the range of 5% to 10% of the power boundary; weight coefficient : determined through experimental data analysis of the battery cycle aging sensitivity of different time periods and life model fitting, reflecting the different effects of load changes on life; penalty intensity coefficient γ: adjusted through multiple rounds of optimization simulation to have a moderate impact on the optimization result, avoiding excessive constraints or insufficient constraints.

[0028] Specifically, the specific process of dynamically constructing the electricity price rule association network through transfer learning and generating the confidence interval prediction curve of photovoltaic and load by quantile regression forest is as follows: the electricity price formation mechanism data is extracted from the electricity market historical database, the correlation characteristics between different electricity price mechanisms are analyzed by the pre-trained prediction model, and the electricity price rule knowledge framework is constructed; based on the electricity price structure characteristics of the target area, the localized rule network is adapted through graph structure transfer learning, the multi-source operation data is integrated, the prediction model is trained at multiple confidence levels simultaneously through the quantile regression forest algorithm, and the probability distribution curve of photovoltaic output and load is generated.

[0029] In the embodiment, the construction of the electricity price rule association network first extracts the electricity price formation mechanism related data from the electricity market historical database, including the electricity price adjustment period, the peak-valley distribution and the market transaction rules. The pre-trained model is used to analyze the internal correlation between the electricity price mechanisms, and the electricity price rule knowledge framework is constructed. This framework supports the subsequent application of transfer learning by representing the mapping relationship between different electricity price rules. The target area electricity price structure is represented as a graph structure , where the node set represents different electricity price periods, and the edge set represents the rule association between periods. Transfer learning aims to adapt to the local electricity price characteristics by optimizing the node weight vector , to meet: ; where is the transfer learning loss function, which measures the fitting error of the adjusted model to the local data set ; represents the weight corresponding to the node in the electricity price rule network, reflecting the importance of each electricity price period in the local mechanism. The weight vector is updated iteratively by gradient descent method, and the initial value is provided by the pre-trained model. The update step is: ; is the learning rate, which is determined based on the validation set performance; is the iteration index. The quantile regression forest generates the confidence interval prediction curve by integrating the localized electricity price rule network and multi-source operation data (such as photovoltaic output, load and weather data) to train the quantile regression forest model. Let the training sample be , where: is the th sample feature vector, which contains historical photovoltaic and load data; is the corresponding observation value. For the confidence level , the model outputs the prediction quantile: ; where is the number of trees in the forest, which ensures the generalization ability and stability of the model; is the th regression tree at time confidence level of the prediction. The objective function of quantile regression is: ; where the quantile regression loss function is defined as: ; is the residual; is the indicator function, taking 1 when the residual is negative, otherwise 0. This loss function guarantees that the model reasonably punishes the deviation of the prediction distribution at different confidence levels, achieving accurate estimation of the confidence interval. Confidence level : According to the actual business requirements and historical data distribution, select multiple groups of confidence levels, such as low (0.1), medium (0.5, representing the median), high (0.9), to capture the prediction uncertainty. This selection is based on empirical analysis of the balance between confidence interval width and coverage rate. Learning rate : Determine through the validation set, ensure the convergence of the transfer learning process and avoid overfitting. Usually set the initial value to 0.01 or 0.001, and dynamically adjust according to the loss curve. Number of forest trees : Determine through cross-validation, ensure the trade-off between model complexity and generalization performance, usually choose an integer between 50 and 200.

[0030] Specifically, and based on the dynamic time warping algorithm, the specific process of extracting the steep change feature sequence of the net load change rate is as follows: from the historical data, filter the net load mutation events, extract the standardized waveform as the steep change reference template, and calculate the dynamic time warping distance between the current curve and the template in real time. When the distance is lower than the similarity threshold, mark the candidate feature area; calculate the change rate derivative in the feature area, if the absolute value of the change rate of continuous multiple sampling points exceeds the historical fluctuation range, it is determined as an effective steep change feature, and the timestamp, duration, and maximum change rate of the feature area are packaged as a feature vector sequence; where the steep change includes steep rise and steep fall.

[0031] In this embodiment, the extraction of the steep change reference template first filters typical net load mutation events from historical net load data, and extracts the corresponding standardized waveform as the steep change reference template. The standardization process includes normalizing the waveform amplitude to eliminate the influence of amplitude differences between different events on matching. Real-time dynamic time warping distance calculation and feature area identification, in real-time monitoring, the current net load change curve segment is represented as a sequence , and the reference template is represented as a sequence , where: is the number of sampling points in the current monitoring window; is the number of sampling points in the template sequence. The dynamic time warping distance is defined as: ; where, represents the set of all allowed paths, used to align the sampling points in the sequence; is the absolute distance between sequence points; For path, ensure the nonlinear alignment of time series. When dynamic time warping distance is less than the set similarity threshold , the current curve segment in the sliding window is marked as a candidate feature area. Parameter description: similarity threshold: the threshold is determined by the distance statistical distribution of typical steep event in historical data, usually selected to cover the upper limit of 95% similar events, to ensure the balance of matching sensitivity and false alarm rate. Derivative of change rate and steep feature determination: in the candidate feature area, the derivative sequence of the net load change rate is calculated: ; Wherein, represents the change rate of the th sampling point; is the sampling time interval. Through the historical net load fluctuation data, the normal fluctuation range of the change rate is calculated. When the absolute value of the change rate of consecutive sampling points all exceeds the historical fluctuation range, that is, it satisfies: ; It is determined that the region is an effective steep feature, including steep rise (change rate positive) or steep drop (change rate negative). Parameter description: is the number of consecutive sampling points for determination, which is usually determined according to the sampling frequency and the duration of steep event experience; is the maximum threshold of historical fluctuation, which is obtained by statistical analysis of historical net load change rate samples, for example, selecting the 99% quantile. Feature vector sequence packaging: for the feature area determined as effective steep, the following parameters are extracted to form the feature vector sequence : ; Wherein, is the starting timestamp of the feature area; is the duration of the feature area; is the maximum absolute value of the change rate in the feature area. This sequence provides dynamic input information of net load mutation for the subsequent scheduling module. Key parameter determination description similarity threshold: by statistical analysis of the DTW distance distribution of typical steep events in history, take the upper limit of 95% confidence interval as the threshold, which ensures the detection sensitivity and controls the false alarm rate. Continuous determination point number : combined with the sampling period and the actual steep event duration, it is empirically set to 2-5 sampling points. Historical fluctuation threshold : based on the historical net load change rate sample, take the 99% quantile as the standard to judge whether it exceeds the normal fluctuation range.

[0032] Specifically, when it is detected that the net load change rate in a continuous period exceeds the set threshold and the contract demand margin is lower than the safety margin, the specific process of dynamically compressing the upper limit of the battery discharge depth and activating load hierarchical regulation is as follows: Based on the steep change feature sequence output by the feature extraction module, identify that the net load change rate at multiple consecutive sampling points exceeds the upper limit of the historical normal fluctuation range. Synchronously calculate the consumed demand within the contract period by real-time integration, and deduce the occupancy trend of the remaining demand in the remaining period in combination with the prediction curve. When the margin is lower than the safety margin and there are steep change features, trigger the demand risk state; compress the upper limit of the discharge depth in stages according to the risk duration. In the initial stage, reduce a fixed proportion of the basic discharge depth, and add a compression amplitude for each newly added high-risk period. Synchronously activate the load priority management mechanism, immediately cut off the preset non-critical load categories, and delay the activation time window of adjustable loads, so as to reduce the peak pressure of the power grid through load-side coordination.

[0033] In this implementation plan, for the identification of steep change features and the judgment of the contract demand margin, using the steep change feature sequence output by the feature extraction module, identify that the net load change rate {q1, q2,... } at M consecutive sampling points all exceeds the upper limit threshold δq of the historical normal fluctuation range, that is, it satisfies: Parameter description: qm: The net load change rate at the m-th sampling point; δq: The upper limit threshold of the net load change rate, determined by historical fluctuation statistical data, generally selecting the 99% quantile to ensure the accurate identification of abnormal steep changes; M: The number of consecutive sampling points, set according to the sampling frequency and system response requirements, usually 3 - 5 points. The real-time integration calculation of the consumed demand Dc within the contract period: ; where p(t): The real-time load power; t0: The start time of the contract period; t: The current time. Predict the occupancy trend of the remaining demand Estimated through the prediction curve ŝ(t): ; where ŝ(τ): The predicted load power in the future period; te: The end time of the contract period. The contract demand margin R is calculated as: ; where : The maximum demand limit specified in the contract. When R < Rsafe (safety margin threshold) and there are steep change features, trigger the demand risk state. Parameter description: Rsafe: The safety margin threshold, determined by analyzing the historical demand overrun risk and the system's tolerance, generally taking 5% - 10% of the contract demand limit. The dynamic compression strategy for the upper limit of the discharge depth dynamically adjusts the upper limit of the battery discharge depth according to the demand risk duration : ; where : The upper limit of the basic discharge depth; : The cumulative number of risk periods; : The The compression ratio corresponding to each risk period satisfies , and is increasing. Specifically, it is set as: the initial risk period compression ratio (fixed ratio), and the compression ratio increases by for each additional risk period; wherein, : initial compression ratio, determined by experience and safety analysis, usually ; : incremental step, ensuring that the compression intensity gradually increases with the accumulation of risks. The load priority management mechanism is activated simultaneously to implement the following measures: cutting off non-critical loads: immediately disconnecting pre-defined non-critical load categories to reduce the total load. Delaying the start-up time window of adjustable loads: adjusting the operation time of adjustable loads to delay the start-up and alleviate peak pressure. This mechanism is based on a pre-defined load classification and priority system to achieve hierarchical management of different categories of loads, ensuring the normal operation of critical loads. The net load change rate threshold : through statistical analysis of historical load change rate data, the th quantile is selected as the threshold to ensure accurate detection of abnormal changes. The safety margin threshold : based on demand overrun risk assessment and contract provisions, set to of the maximum demand in the contract, ensuring a reasonable risk tolerance. Compression ratio and incremental step : determined according to battery safety usage specifications and actual dispatching requirements, the initial compression ensures a certain buffer, and the incremental step reflects the cumulative effect of risks. The number of consecutive sampling points and the cumulative number of risk periods : according to the sampling frequency and system response speed, ensure timely response and cumulative identification of risk status.

[0034] Specifically, within the safety envelope constraint range, the three-objective evolutionary algorithm guided by tabu search simultaneously optimizes the economic target, life loss target, and grid constraint target to generate a specific process of the three-objective balanced charging and discharging strategy as follows: convert the power safety envelope into a set of time-varying inequality constraints to define the charging and discharging power feasible region boundary; generate an initial solution set guided by the tabu search algorithm to ensure that the candidate strategies are distributed within the feasible region, and simultaneously evaluate the three-objective function values for each candidate scheme, calculate the total sum of time-of-use electricity cost and demand penalty for the economic target, quantify the charging and discharging rate stress cumulative effect for the life loss target, and evaluate the peak power and fluctuation rate penalty for the grid constraint target, and select the Pareto optimal solution set through non-dominated sorting; inject an adaptive penalty term proportional to the distance from the boundary for solutions approaching the safety envelope boundary, and start the tabu list to block repeated generation for out-of-bound solutions, and the output charging and discharging strategy is a structured instruction sequence, including: the charging and discharging power values corresponding to the timestamp sequence, and the load classification adjustment action time sequence list.

[0035] In this embodiment, the power safety envelope is transformed into time-varying inequality constraints, which represents the safety envelope on time series t = 1, 2, …, T as power constraint intervals: ; Parameter description: : the charging and discharging power at time t; : the lower boundary power limit of the safety envelope at time t; : the upper boundary power limit of the safety envelope at time t; : the total number of time steps of the scheduling period. This constraint defines the boundary of the feasible region of charging and discharging power, ensuring the safe operation of the strategy. The initial solution set is generated by Tabu Search (TS) guided, which generates an initial solution set within the constraint domain Each solution satisfies: ; Tabu Search avoids local optimal traps through neighborhood search and tabu list, promoting the diversity of solutions and global exploration. The multi-objective function is evaluated simultaneously, and three objective functions are calculated for each candidate strategy The economic objective is the sum of time-of-use electricity cost and demand penalty: ; Parameter description: : the electricity price at time t; : the positive discharging power; : the demand penalty weight coefficient, which is set according to the contract terms; : the maximum demand of the strategy; : the maximum allowed demand of the contract. The life loss objective quantifies the cumulative effect of the charging and discharging rate stress: ; Parameter description: : the time weight, reflecting the sensitivity of battery aging in different time periods; : the rated capacity of the battery; : the rate stress index, usually taking 2-3 to reflect the nonlinear stress effect. The grid constraint objective evaluates the peak power and power fluctuation rate penalty: ; Parameter description: : the peak power penalty weight coefficient; : the power fluctuation rate penalty weight coefficient. Pareto optimal solution screening and penalty mechanism, non-dominated sorting is performed on the initial solution set, and the Pareto optimal solution set is screened out to ensure multi-objective balance. For solutions that approach the boundary of the safety envelope, calculate the minimum distance to the boundary: ; and inject an adaptive penalty term in the objective function: ; Parameter description: : the penalty weight coefficient, which determines the penalty strength; : Prevent small constants from being zero. Add the out-of-bound solution to the taboo list immediately, prohibit repeated generation, and ensure the solution set method. Output the structured charging and discharging strategy, and the final output strategy includes: timestamp sequence corresponding to charging and discharging power value ; Load grading regulation action timing list, recording load regulation category and execution time. Determine the weight coefficient The balance between economy and safety is determined by historical data fitting and system safety evaluation; the safety envelope boundary is dynamically generated by the life management module to ensure real-time risk constraints; the non-dominated sorting adopts the classical NSGAII algorithm to ensure the effect of multi-objective optimization; the length of the taboo list and the neighborhood structure are designed based on the problem size and computing resources to ensure search efficiency and globality.

[0036] Specifically, when the distance is below the warning threshold, the rolling optimization control is started, and the specific process of increasing the life protection weight by a preset proportion and relaxing the economic constraints is as follows: the actual power point to the envelope boundary is continuously calculated, and when the distance value falls within the preset warning interval, the rolling optimization mechanism is triggered; within the rolling time window, the multi-objective function is reconstructed, the life protection target weight is increased to a fixed multiple of the baseline value, and the allowed deviation range of the economic constraint is simultaneously relaxed, the power derating instruction is dynamically generated by the quadratic programming solver, and the derating amplitude is negatively exponential to the distance value, ensuring that the power trajectory converges safely along the envelope line.

[0037] In this embodiment, the minimum Euclidean distance from the power point to the envelope boundary is calculated in real time, and the current power point is q, the lower and upper boundaries of the envelope line are L (lower limit) and U (upper limit), and the distance d is calculated as: ; Parameter description: : The actual charging and discharging power at the current time; : The lower boundary power of the safety envelope at the current time; : The upper boundary power of the safety envelope at the current time; : The minimum Euclidean distance from the actual power point to the envelope boundary. When the distance enters the preset warning interval , the rolling optimization control mechanism is triggered. The warning interval determination method: according to the experience of safe operation of the battery and the dynamic response characteristics of the system; combined with historical operation data and life attenuation risk threshold, a reasonable range is set to ensure timely warning and avoid false triggering. The rolling optimization time window length is set to , and dynamic multi-objective optimization is performed within the window to continuously update the control strategy to ensure real-time response. The multi-objective function is reconstructed and the weight is adjusted, and the multi-objective function is reconstructed as: ; Parameter description: : Life protection objective function, reflecting battery life loss; : Economic objective function, reflecting electricity cost and demand penalty; : Grid constraint objective function, reflecting peak power and fluctuation rate; , , : Objective function weight coefficient. When the rolling optimization is triggered, adjust the weight coefficient to: ; Parameter description: , , : Reference weight before rolling optimization; : Lifetime protection weight increase multiple, usually set according to battery safety requirements; : Economic constraint allowed deviation, reflecting the relaxation degree, set based on risk tolerance. This adjustment prioritizes the increase of lifetime protection weight and relaxes economic constraints, balancing safety and economy. The dynamic generation of power reduction instructions is solved by using a quadratic programming solver within a rolling window: ; Constraint conditions: ; Wherein: : Power decision variable at future W time; : Positive definite weight matrix, reflecting the quadratic term of the weighted objective function, including the lifetime protection effect after weight adjustment; : Linear vector, including electricity price and demand penalty factors; : Linear constraint matrix and vector of system and grid. The reduction amplitude and the negative exponential relationship of distance are defined as: ; Parameter description: : Power reduction ratio; : Maximum reduction amplitude, set based on battery and grid carrying limit; : Distance sensitivity coefficient, controlling the rate of reduction with distance, adjusted according to system response performance; : Minimum Euclidean distance from power point to envelope line boundary. This formula ensures that the smaller the distance (i.e. closer to the envelope line boundary), the greater the reduction amplitude, achieving continuous convergence of the safe power trajectory. Result feedback and closed-loop control, power reduction instructions are issued in real time to the energy storage EMS execution unit, and the adjusted power trajectory is fed back to the lifetime management module and envelope line dynamic generation unit, realizing closed-loop risk control. Parameter determination description, warning interval : According to the chemical properties of the battery and the risk of excessive lithium precipitation, combined with statistical analysis and simulation verification, ensure sensitive and not excessive triggering; lifetime protection weight increase multiple : Adjusted according to the battery life decay rate and system safety requirements, usually taken as 1.5~3; economic constraint relaxation amplitude : Adjusted moderately combining with economic operation demand and risk assessment, avoiding excessive impact on cost; maximum reduction amplitude With distance sensitivity coefficient : Obtained through fitting historical operating data and on-site debugging, ensuring response speed and safety margin.

[0038] Specifically, the process of simultaneously feeding back the actual lithium plating risk characteristics to the lifespan management module to drive the online update of the safety envelope is as follows: Real-time acquisition of charging and discharging current waveforms, temperature distribution, and voltage fluctuation data; reconstruction of the ion concentration gradient field inside the battery based on impedance spectroscopy analysis; location of abnormal diffusion delay regions and quantification of local ion flux deviations; input of the measured ion flux deviations into the electrochemical parameter estimation algorithm; updating the mobility and diffusion coefficients in the ion transport equation; recalculation of the lithium plating risk level based on the updated ion transport equation; generation of a power safety boundary curve that continuously evolves over time; and real-time injection of the new envelope parameters into the constraint library of the hierarchical optimization module through a dynamic interface, replacing the original boundary values ​​and triggering strategy re-verification.

[0039] In this implementation plan, the real-time data acquisition and ion concentration gradient field reconstruction system acquires key operating data of the energy storage battery in real time, including: charging and discharging current waveforms. Battery temperature distribution Voltage fluctuation signal Based on electrochemical impedance spectroscopy (EIS) analysis and combined with an internal physical model of the battery, the ion concentration gradient fields of the positive and negative electrodes of the battery are reconstructed. ,in: Parameter description: Spacetime coordinates Ion concentration at the location; Equilibrium-state ion concentration benchmark value; Dynamic deviations in concentration over time and space. This is achieved by identifying anomalous regions of diffusion delay. (Set of internal regions of the battery) to locate local ion flux anomalies. Quantification of local ion flux deviation and estimation of input electrochemical parameters; the local ion flux deviation is defined as: Parameter description: Measured ion flux; Reference baseline flux (based on steady-state or historical average). Input electrochemical parameter estimation algorithm, targeting ion mobility and diffusion coefficient Updated estimate: ; Parameter description: Previously estimated mobility and diffusion coefficient; Updated parameters; The temporal and spatial average values ​​of local ion flux deviation within the region; Weighting coefficients are used to adjust the parameter update rate. The publicly available method is based on regression fitting of historical data to prevent over-adjustment and ensure model stability. The lithium plating risk level is recalculated based on the updated parameter ion transport equation, using a classical ion transport partial differential equation model. Parameter description: : Ion concentration distribution function; : Electric field vector, representing the potential gradient driving migration; , The diffusion coefficient and mobility were updated in the above steps. Based on the newly calculated concentration gradient and ion flux distribution, the lithium plating risk level was dynamically calculated. : Parameter description: Real-time lithium plating risk level; : Ion concentration gradient amplitude; The critical concentration gradient threshold for lithium plating is pre-set based on the physical properties of the battery materials. Power safety boundary curve generation and dynamic updating are performed according to risk level. Dynamically adjust the safety envelope of charge and discharge power: Parameter description: :time The safe power boundary value; Initial safety boundary power value; The risk sensitivity coefficient reflects the degree to which the risk level affects power limitation. This boundary curve evolves continuously over time, reflecting the actual operational risk status. Real-time injection and strategy re-verification of envelope parameters are achieved through a dynamic interface module, continuously updating the envelope parameters. Inject the constraint library into the hierarchical optimization module to replace the original boundary conditions: Replace constraints: ;in, This refers to the charge / discharge power decision variable in the current optimization strategy. It triggers a strategy re-verification in the hierarchical optimization module to ensure that the strategy still satisfies the optimization objective and safety constraints under the new boundary. The parameter determination method and weighting coefficients are also included. Based on historical measured data and battery life degradation test results, the critical concentration gradient is determined through least squares fitting or Bayesian estimation methods, taking into account both model response sensitivity and stability. Provided by the battery manufacturer or obtained through laboratory lithium plating threshold testing, reflecting the electrochemical limits of the material; risk sensitivity coefficient. Based on safe operating experience and risk management strategies, power limits are set to ensure they are reasonable and to avoid being overly conservative or overlooking risks.

[0040] In summary, this application has at least the following effects:

[0041] The EMS scheduling system of industrial and commercial energy storage based on multi-objective optimization, based on real-time analysis of the internal ion migration state of the battery and dynamic generation of the charge and discharge power safety envelope, effectively warns and suppresses the risk of lithium precipitation, prevents damage to the internal structure of the battery and performance degradation, and prolongs the service life of the battery. Through transfer learning to construct a price rule association network, combined with quantile regression forest to generate photovoltaic and load confidence interval prediction, dynamically extract the steepness feature of the net load, realize accurate identification and response to power grid load fluctuation, and improve the ability of the scheduling scheme to adapt to complex industrial and commercial environment. A multi-objective evolutionary algorithm guided by tabu search is adopted to optimize the charging and discharging strategy under the premise of meeting the safety envelope hard constraint, taking into account the economy, life loss and power grid constraint, and improving the overall operation efficiency of the energy storage system. The arbitration module is based on real-time monitoring of the distance between power and safety boundary, and the life protection weight and economic constraint are dynamically adjusted by rolling optimization control strategy to realize the safe convergence of power trajectory, and the real-time lithium precipitation risk feature is fed back to the life management module to drive the online update of the safety envelope, ensuring the timeliness and accuracy of the system response. By dynamically compressing the depth of charging and discharging and activating the load hierarchical regulation mechanism, the peak pressure of the power grid is effectively reduced, the load structure is optimized, and the efficient cooperative operation of the industrial and commercial energy storage system in the complex power grid environment is supported.

[0042] Those skilled in the art will appreciate that embodiments of the application can be provided as methods, systems, or computer program products. Accordingly, the application can be embodied in the form of complete hardware embodiments, complete software embodiments, or embodiments combining software and hardware aspects. Moreover, the application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) having computer usable program code embodied thereon.

[0043] The present application is described with reference to flowcharts and / or block diagrams of systems, apparatuses (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 An apparatus for performing the functions specified in one or more flows and / or blocks. Figure 1 An apparatus for performing the functions specified in one or more flows and / or blocks.

[0044] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0045] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that are executed on the computer or other programmable apparatus provide steps for implementing the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0046] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those of skill in the art once they have the benefit of the present disclosure. Therefore, the appended claims are intended to encompass within their scope all possible variations and modifications of the preferred embodiments. 1

[0047] It is apparent that a person skilled in the art can make a variety of changes and modifications to the application without departing from the spirit and scope thereof. Thus, if these modifications and changes fall within the scope of the claims and their equivalents, it is intended to include them in the application.

Claims

1. An EMS dispatching system for industrial and commercial energy storage based on multi-objective optimization, characterized in that, Comprise the following modules: life management module, feature extraction module, hierarchical optimization module, execution arbitration module; The life management module is used for real-time analysis of the ion migration state inside the battery, calculating the lithium precipitation risk coefficient according to the ion concentration gradient distribution and dynamically generating the charge and discharge power safety envelope line, and embedding the envelope line into the hierarchical optimization module as a hard constraint condition; The feature extraction module is used for dynamically constructing an electricity price rule association network through transfer learning, generating confidence interval prediction curves of photovoltaic and load by combining quantile regression forest, and extracting steep change rate feature sequence of net load change based on dynamic time warping algorithm; The hierarchical optimization module is used for dynamically compressing the upper limit of battery discharge depth and activating load hierarchical regulation when detecting that the net load change rate in a continuous period exceeds a set threshold and the contract demand margin is lower than the safety margin, and generating a three-objective balanced charge and discharge strategy by synchronously optimizing the economy target, life loss target and grid constraint target within the safety envelope line constraint range through the tabu search guided multi-objective evolutionary algorithm; The execution arbitration module is used for real-time monitoring of the distance between the charge and discharge power and the safety envelope line, and starting rolling optimization control when the distance is lower than the warning threshold, increasing the life protection weight by a preset proportion and relaxing the economy constraint, and simultaneously feeding back the actual lithium precipitation risk features to the life management module to drive online update of the safety envelope line.

2. The multi-objective optimization based industrial and commercial energy storage EMS scheduling system according to claim 1, characterized in that: The specific process of real-time analysis of the ion migration state inside the battery, calculating the lithium precipitation risk coefficient according to the ion concentration gradient distribution and dynamically generating the charge and discharge power safety envelope line is as follows: Real-time acquisition of battery internal sensor data stream, generation of positive electrode area ion concentration gradient distribution graph through ion transport equation, and dynamic calculation of local ion flux reference value; Analysis of the concentration mutation nodes in the gradient distribution graph, calculation of the real-time deviation degree of the measured ion flux in the region from the reference value; Mapping the deviation degree to the lithium precipitation risk coefficient, adjusting the upper limit boundary of the charge and discharge power according to the risk growth rate when the risk coefficient exceeds the threshold, generating a power safety envelope line that continuously evolves over time, and the envelope line boundary slope is positively correlated with the risk coefficient increment.

3. The multi-objective optimization based industrial and commercial energy storage EMS scheduling system according to claim 2, characterized in that: The specific process of embedding the envelope line into the hierarchical optimization module as a hard constraint condition is as follows: Load safety envelope line parameters in the hierarchical optimization module initialization stage, build charge and discharge power constraint matrix, real-time detection of power trajectory of candidate strategy in optimization iteration process, if the trajectory exceeds the envelope line boundary, mark it as an infeasible solution through the tabu search mechanism; For boundary adjacent solutions, start adaptive penalty function, increase the life loss penalty term in the objective function which is inversely proportional to the distance from the boundary; Before outputting the final strategy, perform envelope line compliance verification, and perform secondary smoothing processing on the boundary point strategy to ensure the continuity of the power change rate.

4. The multi-objective optimization based industrial and commercial energy storage EMS scheduling system according to claim 1, characterized in that: The specific process of dynamically constructing an electricity price rule association network through transfer learning, generating confidence interval prediction curves of photovoltaic and load by combining quantile regression forest is as follows: Extract the electricity price formation mechanism data from the electricity market historical database, analyze the association features between different electricity price mechanisms through a pre-trained prediction model, and construct an electricity price rule knowledge framework; Based on the characteristics of the target area's electricity price structure, the local rules network is adapted through graph structure transfer learning, multi-source operation data is integrated, and the prediction model is trained simultaneously at multiple confidence levels through quantile regression forest algorithm to generate the probability distribution curve of photovoltaic output and load.

5. The multi-objective optimization based industrial and commercial energy storage EMS scheduling system according to claim 4, characterized in that: And based on the dynamic time warping algorithm, the specific process of extracting the steep change feature sequence of the net load change rate is as follows: From the historical data, the net load mutation event is screened out, the standardized waveform is extracted as the steep change benchmark template, and the dynamic time warping distance between the current curve and the template is calculated in real time. When the distance is lower than the similarity threshold, mark the candidate feature area; Calculate the change rate derivative in the feature area. If the absolute value of the change rate of continuous multiple sampling points exceeds the historical fluctuation range, it is determined to be an effective steep change feature, and the feature vector sequence of the time stamp, duration, and maximum change rate of the feature area is packaged. The steep change includes steep rise and steep fall.

6. The multi-objective optimization based industrial and commercial energy storage EMS scheduling system according to claim 5, characterized in that: When the net load change rate of the continuous period exceeds the set threshold and the contract demand reserve is lower than the safety margin, the specific process of dynamically compressing the upper limit of battery discharge depth and activating load grading regulation is as follows: Based on the steep change feature sequence output by the feature extraction module, the net load change rate of continuous multiple sampling points is identified as exceeding the upper limit of the historical normal fluctuation range, and the consumed demand in the contract period is calculated simultaneously in real time. Integrate the remaining demand occupation trend in the remaining period, when the reserve is lower than the safety margin and there is a steep change feature, trigger the demand risk state; According to the risk duration, the upper limit of the discharge depth is compressed in stages. In the initial stage, reduce the fixed proportion of the basic discharge depth, add compression amplitude for each newly added high-risk period, activate the load priority management mechanism simultaneously, immediately cut off the preset non-critical load category, delay the activation time window of the adjustable load, and reduce the peak pressure of the power grid through load-side cooperation.

7. The multi-objective optimization based industrial and commercial energy storage EMS scheduling system according to claim 6, characterized in that: Within the safety envelope constraint range, the economic target, life loss target and grid constraint target are simultaneously optimized by the multi-objective evolutionary algorithm guided by tabu search to generate a three-objective balanced charging and discharging strategy as follows: Convert the power safety envelope to a set of time-varying inequality constraints, and define the boundary of the charging and discharging power feasible region; Generate the initial solution set guided by the tabu search algorithm to ensure that the candidate strategies are distributed within the feasible region. Evaluate the three objective function values of each candidate simultaneously. The economic target calculates the sum of the time-of-use electricity cost and demand penalty, the life loss target quantifies the cumulative effect of the charging and discharging rate stress, and the grid constraint target evaluates the peak power and fluctuation rate penalty. Non-dominated sorting is used to select the Pareto optimal solution set; For solutions approaching the safety envelope boundary, inject an adaptive penalty term inversely proportional to the distance from the boundary, and for out-of-bound solutions, start the tabu list to block repeated generation. The output charging and discharging strategy is a structured instruction sequence, including: the charging and discharging power value corresponding to the time stamp sequence, and the load grading regulation action time sequence list.

8. The multi-objective optimization based industrial and commercial energy storage EMS scheduling system according to claim 1, characterized in that: When the distance is lower than the warning threshold, start the rolling optimization control, increase the life protection weight by the preset proportion and relax the economic constraint as follows: Continuously calculate the minimum Euclidean distance from the actual power point to the envelope boundary. When the distance value falls within the preset warning interval, trigger the rolling optimization mechanism; The multi-objective function is reconstructed within the rolling time window, the life protection target weight is increased to a fixed multiple of the benchmark value, the allowed deviation range of the economic constraint is simultaneously relaxed, the power derating instruction is dynamically generated by the quadratic programming solver, the derating amplitude is negatively exponential to the distance value, and the power trajectory is ensured to safely converge along the envelope line.

9. The multi-objective optimization based industrial and commercial energy storage EMS scheduling system according to claim 8, characterized in that: The specific process of feeding the actual lithium precipitation risk characteristics to the life management module to drive the online update of the safety envelope line is as follows: Real-time acquisition of charge and discharge current waveform, temperature distribution and voltage fluctuation data, reconstruction of battery internal ion concentration gradient field based on impedance spectrum analysis, positioning of diffusion delay abnormal area and quantification of local ion flux deviation; The measured ion flux deviation is input into the electrochemical parameter estimation algorithm, the mobility and diffusion coefficient in the ion transport equation are updated, the lithium precipitation risk level is recalculated according to the updated ion transport equation, and the power safety boundary curve continuously evolving with time is generated; Through the dynamic interface, the new envelope line parameters are real-time injected into the constraint library of the hierarchical optimization module, the original boundary value is replaced, and the strategy re-verification is triggered.

Images