An integrated dispatching system for realizing PCS, EMS and BMS

Abstract

The application discloses an integrated scheduling system for realizing PCS, EMS and BMS, and relates to the technical field of power control, which comprises a multi-dimensional performance evaluation module, a multi-objective optimization module and a dynamic de-rating module.The multi-dimensional performance evaluation module constructs a battery aging dynamic model and trains the same, generates a multi-dimensional performance evaluation index through the battery aging dynamic model based on a standardized state vector and a health state prediction, and the multi-objective optimization module generates a collaborative scheduling strategy set through a multi-objective optimization solver of an improved genetic algorithm in combination with a fuzzy analytic hierarchy process based on the multi-dimensional performance evaluation index.The dynamic de-rating module generates an executable instruction queue with safety constraints through an industrial internet of things protocol stack in combination with a dynamic de-rating coefficient algorithm based on the collaborative scheduling strategy set.The application realizes nonlinear coupling modeling of a cycle attenuation and a calendar aging mechanism in a battery aging dynamic model through a physical driving characteristic layer and a dynamic parameter calibration layer.

Description

Technical Field

[0001] This invention relates to the field of power control technology, and in particular to an integrated dispatching system for realizing PCS, EMS and BMS. Background Technology

[0002] With the increasing penetration of renewable energy and the large-scale deployment of energy storage systems, the coordinated scheduling technology of power conversion systems, energy management systems, and battery management systems has become a research hotspot in power system automation. In existing technologies, EMS (Energy Management System) mainly relies on load forecasting and electricity price signal generation to optimize scheduling strategies, such as using linear programming or dynamic programming algorithms to achieve peak-valley arbitrage; BMS (Battery Management System) estimates the state of charge and health based on battery equivalent circuit models or electrochemical models, and ensures battery safety through charge and discharge current limiting strategies; PCS (Power Control System), as the execution unit, uses PI control or direct power control to achieve grid interaction.

[0003] In recent years, digital twin technology has been introduced into the field of energy storage systems to predict battery aging trends or the impact of grid fluctuations through offline simulation models. For example, a temperature-dependent lifetime model based on the Arrhenius equation is combined with a cyclic decay model based on rainflow counting to predict battery life. In addition, multi-objective optimization algorithms are used to balance economic efficiency, battery life, and grid stability, but their weighting coefficients are usually statically set, making it difficult to adapt to dynamic operating conditions. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an integrated scheduling system for realizing PCS, EMS and BMS to solve the problem of insufficient multi-objective collaborative optimization in energy storage systems.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides an integrated scheduling system for PCS, EMS, and BMS, comprising: a data preprocessing module for real-time acquisition and preprocessing of equipment operating status data to generate standardized state vectors; a multi-dimensional performance evaluation module for constructing and training a battery aging dynamic model, and generating multi-dimensional performance evaluation indicators based on the standardized state vectors through dynamic correction and digital twin mapping of the battery aging dynamic model; a multi-objective optimization module for generating a collaborative scheduling strategy set based on the multi-dimensional performance evaluation indicators using a multi-objective optimization solver combining fuzzy hierarchical analysis and an improved genetic algorithm; a dynamic derating module for generating an executable instruction queue with safety constraints based on the collaborative scheduling strategy set, using an industrial IoT protocol stack combined with a dynamic derating coefficient algorithm; a circuit breaker monitoring module for generating an execution feature dataset with anomaly codes based on the executable instruction queue, using a real-time circuit breaker mechanism and a power quality monitoring algorithm; and a closed-loop optimization module for updating battery aging model parameters using an incremental online learning algorithm based on the execution feature dataset, and generating a power grid fluctuation feature library using an Arrhenius equation parameter optimizer.

[0008] As a preferred embodiment of the integrated scheduling system for PCS, EMS and BMS described in this invention, the equipment operating status data includes PCS power and voltage, EMS load prediction data and BMS state of charge and health parameters.

[0009] As a preferred embodiment of the integrated scheduling system for PCS, EMS and BMS described in this invention, the preprocessing includes denoising, normalization, time alignment of multi-source heterogeneous data and feature fusion.

[0010] As a preferred embodiment of the integrated scheduling system for PCS, EMS, and BMS described in this invention, the specific steps for constructing and training the battery aging dynamic model are as follows:

[0011] An input layer is constructed based on a multi-source heterogeneous data fusion method to convert standardized state vectors into time series tensors;

[0012] By using the Arrhenius multi-mechanism coupling analysis method, joint features of cyclic decay and calendar aging are extracted from the input data, and a shared physical driving feature layer is constructed.

[0013] A parameter calibration layer is constructed by performing nonlinear parameter fitting on shared features using an improved particle swarm optimization algorithm.

[0014] A dynamic state estimation layer is constructed by fusing the calibrated cyclic decay factor with real-time data using the extended Kalman filter algorithm.

[0015] A dynamic battery aging model is constructed based on an input layer, a physical driving feature layer, a parameter calibration layer, and a dynamic state estimation layer, and trained using a hybrid collaborative training method.

[0016] As a preferred embodiment of the integrated scheduling system for PCS, EMS, and BMS described in this invention, the step of generating multi-dimensional performance evaluation indicators based on standardized state vectors through dynamic correction and digital twin mapping using a battery aging dynamic model is as follows:

[0017] Based on standardized state vectors, a battery aging dynamic model is used to predict the health status and generate health status indicators.

[0018] Based on health status indicators, dynamic status correction is performed using an extended Kalman filter algorithm to generate real-time health status estimates.

[0019] Based on real-time health status estimates, multi-dimensional performance quantification is performed through a digital twin mapping engine to generate multi-dimensional performance evaluation indicators.

[0020] As a preferred embodiment of the integrated scheduling system for PCS, EMS, and BMS described in this invention, the step of generating a collaborative scheduling strategy set based on multi-dimensional performance evaluation indicators using a multi-objective optimization solver combining fuzzy hierarchical analysis and an improved genetic algorithm is as follows:

[0021] Based on multidimensional performance evaluation indicators, multi-objective weight allocation is performed using fuzzy hierarchical analysis to generate priority weight vectors.

[0022] Based on the priority weight vector, a multi-objective optimization solution is obtained by using an improved non-dominated sorting genetic algorithm to generate the Pareto optimal solution set;

[0023] Based on the Pareto optimal solution set, the TOPSIS decision method is used to select the best strategy and generate the final collaborative scheduling strategy set.

[0024] As a preferred embodiment of the integrated scheduling system for PCS, EMS, and BMS described in this invention, the step of generating an executable instruction queue with security constraints based on a collaborative scheduling strategy set, using an industrial IoT protocol stack combined with a dynamic derating algorithm, comprises the following specific steps:

[0025] Based on the collaborative scheduling strategy set, instructions are securely encapsulated through the industrial IoT protocol stack to generate protocol-compatible instruction frames.

[0026] Based on protocol-compatible instruction frames, security constraints are injected through a dynamic derating factor algorithm to generate risk response instructions;

[0027] Based on risk response instructions, an executable instruction queue with safety constraints is generated using an instruction priority sorting algorithm.

[0028] As a preferred embodiment of the integrated scheduling system for PCS, EMS, and BMS described in this invention, the step of generating an execution feature dataset with anomaly codes based on an executable instruction queue, through a real-time circuit breaker mechanism and a power quality monitoring algorithm, comprises the following specific steps:

[0029] Based on the executable instruction queue, overload risk monitoring is performed through a real-time circuit breaker mechanism, generating circuit breaker events and abnormal tags.

[0030] Based on the execution process data of the executable instruction queue, dynamic parameter analysis is performed through power quality monitoring algorithms to generate quality anomaly waveforms;

[0031] Based on circuit breaker events and quality anomaly waveforms, feature fusion is performed using an anomaly coding rule base to generate an execution feature dataset with security labels.

[0032] As a preferred embodiment of the integrated scheduling system for PCS, EMS, and BMS described in this invention, the steps for updating the battery aging model parameters based on the execution feature dataset using an incremental online learning algorithm are as follows:

[0033] Based on the execution feature dataset, dynamic parameter updates are performed using an online sequence extreme learning machine to generate preliminary calibration parameters;

[0034] Based on the initial calibration parameters, the physical range is corrected using a regularized constraint optimizer to generate updated battery aging model parameters.

[0035] As a preferred embodiment of the integrated scheduling system for PCS, EMS, and BMS described in this invention, the specific steps for generating the power grid fluctuation characteristic library using the Arrhenius equation parameter optimizer are as follows:

[0036] Based on the execution feature dataset, temperature-dependent aging features are generated using a wavelet transform-Arrhenius temperature feature extractor.

[0037] Based on temperature-dependent aging characteristics, a power grid fluctuation feature library is generated by multi-dimensional feature clustering using a spectrum clustering analyzer.

[0038] The beneficial effects of this invention are as follows: by using a physical driving feature layer and a dynamic parameter calibration layer, nonlinear coupling modeling of cyclic decay and calendar aging mechanism in the dynamic model of battery aging is realized, thereby improving the accuracy of health status prediction and model adaptability. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a schematic diagram of an integrated scheduling system for implementing PCS, EMS, and BMS.

[0041] Figure 2 This is a flowchart for implementing integrated scheduling of PCS, EMS, and BMS.

[0042] Figure 3 A flowchart for building a dynamic model of battery aging.

[0043] Figure 4 The flowchart shows the dual-channel processing of the closed-loop optimization module. Detailed Implementation

[0044] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0045] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0046] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0047] Reference Figures 1-4 This is one embodiment of the present invention, which provides an integrated scheduling system for implementing PCS, EMS and BMS, including the following steps:

[0048] The data preprocessing module collects equipment operating status data in real time, performs preprocessing, and generates standardized status vectors.

[0049] Equipment operating status data includes PCS power and voltage, EMS load forecast data, and BMS state of charge and health parameters.

[0050] Preprocessing includes denoising, normalization, time alignment of multi-source heterogeneous data, and feature fusion.

[0051] Furthermore, during the denoising process, wavelet thresholding denoising is employed to address high-frequency noise interference in the PCS power parameters, EMS load forecast data, and BMS state of charge and health parameters. Specifically, this includes: multi-scale decomposition of transient power fluctuations in the PCS power parameters based on the Daubechies wavelet basis function, and filtering out anomalous pulse components in high-frequency detail coefficients using a soft threshold function; suppressing non-physical fluctuations in low-frequency approximation coefficients for random drift noise in the BMS state of charge and health parameters using a fixed threshold rule; and eliminating periodic background noise in the EMS load forecast data using translation-invariant wavelet denoising while preserving the principal components of the load trend.

[0052] Furthermore, during the normalization process, to address the dimensional differences in the power parameters of the PCS, the load forecast data of the EMS, and the state of charge and health parameters of the BMS, the Z-score standardization method is used to eliminate data distribution bias. Specifically, this includes: for the power parameter series of the PCS, converting the instantaneous power values ​​into a zero-mean, unit-variance distribution based on the sliding window statistical mean and standard deviation; for the electricity price time series in the load forecast data of the EMS, linear scaling is performed using the global mean and standard deviation to ensure the comparability of electricity price weights at different time scales; the state of charge parameters of the BMS are dynamically standardized based on the range within the charge-discharge cycle, and the health parameters are bias-corrected by fitting a benchmark value based on the historical capacity decay trajectory.

[0053] Furthermore, during the time alignment of multi-source heterogeneous data, to address the differences in acquisition frequency and timestamps of the PCS power parameters, EMS load forecast data, and BMS state of charge and health parameters, a timestamp alignment method based on a precise clock synchronization protocol is employed to achieve microsecond-level synchronization. Specifically, this includes: linearly interpolating and resampling the PCS power parameters to generate a power parameter sequence aligned with the timestamps of the BMS state of charge parameters; aligning the EMS load forecast data to the PCS power parameter time axis using a sliding window mean aggregation method; and using a forward padding method to fill in the timestamp gaps for the BMS health parameters. The time-aligned PCS power parameters, EMS load forecast data, and BMS state of charge and health parameters form a time-synchronized multi-source data stream, providing a temporally consistent input for feature fusion.

[0054] Furthermore, during the feature fusion process, principal component analysis (PCMA) is employed to achieve cross-domain feature correlation and dimensionality reduction for the time-aligned PCS power parameters, EMS load forecast data, and BMS state of charge and health parameters. Specifically, this includes: extracting power volatility, charge / discharge efficiency, and voltage deviation characteristics from the PCS power parameters; extracting electricity price sensitivity coefficients and load change gradient characteristics from the EMS load forecast data; and extracting capacity decay slope and internal resistance temperature correlation coefficient characteristics from the BMS state of charge and health parameters. Through PCMA, covariance matrix calculations and orthogonal transformations are performed on these features, retaining the principal component vector with the highest cumulative contribution rate to generate a three-dimensional standardized state vector that integrates the PCS dynamic response characteristics, EMS economic preferences, and BMS health characteristics.

[0055] The multidimensional performance evaluation module constructs and trains a dynamic battery aging model. Based on standardized state vectors, it dynamically corrects and maps the battery aging dynamic model to digital twins to generate multidimensional performance evaluation indicators.

[0056] An input layer is constructed based on a multi-source heterogeneous data fusion method to convert standardized state vectors into time series tensors;

[0057] It should be noted that, in constructing the input layer based on the multi-source heterogeneous data fusion method, the process of converting the standardized state vector into a time-series tensor is as follows: Sliding window sampling is performed on the time-aligned PCS power parameters, EMS load prediction data, and BMS state of charge and health parameters to extract the standardized state vector within a continuous time window; the sampling results are organized using a three-dimensional tensor structure, where the first dimension is the time step (representing historical time-series dependencies), the second dimension is the number of features (corresponding to the principal component vectors of PCS dynamic response characteristics, EMS economic preferences, and BMS health status characteristics), and the third dimension is the number of devices (distinguishing independent data sources for PCS, EMS, and BMS). The resulting time-series-feature-device three-dimensional tensor serves as the input layer of the battery aging dynamic model, preserving the temporal evolution characteristics and cross-device correlations of the multi-source data.

[0058] By using the Arrhenius multi-mechanism coupling analysis method, joint features of cyclic decay and calendar aging are extracted from the input data, and a shared physical driving feature layer is constructed.

[0059] It should be noted that cross-mechanism feature extraction is performed on the PCS power parameters, EMS load forecast data, and BMS state of charge and health parameters contained in the time series tensor. Specifically, this includes: calculating the calendar aging rate of BMS state of charge parameters under different temperature gradients based on the temperature dependence characteristics of the Arrhenius equation, generating a temperature-time correlated aging feature vector; statistically analyzing the charge-discharge cycle depth and number of cycles in the PCS power parameters using the rainflow counting method, extracting a cycle stress-related capacity decay feature vector; and jointly modeling the temperature-time correlated aging feature vector and the cycle stress capacity decay feature vector through a physical driving coupling function (in the form of a product of calendar aging rate and cycle decay factor), generating a shared physical driving feature layer that integrates temperature, cycle stress, and time evolution. This feature layer preserves the coupling influence of PCS dynamic power behavior on BMS health status, providing cross-domain correlated input features for the parameter calibration layer.

[0060] A parameter calibration layer is constructed by performing nonlinear parameter fitting on shared features using an improved particle swarm optimization algorithm.

[0061] It should be noted that, in the process of nonlinearly fitting the temperature-time correlated aging feature vector and cyclic stress capacity decay feature vector in the shared physical driving feature layer using the improved particle swarm optimization algorithm, a dynamic inertia weight adjustment strategy is adopted to optimize the particle search process. Specifically, this includes: initializing the particle swarm position and velocity, where the particle position vector corresponds to the combination of cyclic decay factor and calendar aging coefficient parameters in the battery aging dynamic model; constructing a fitness function based on the temperature gradient, cycle depth, and capacity decay observation data in the shared physical driving feature layer, and calculating the fitting error of each individual particle in the swarm; dynamically adjusting the inertia weight according to the particle's historical optimal position and global optimal position during iteration to balance local search and global exploration capabilities; and imposing constraint penalty terms on out-of-bounds cyclic decay factor and calendar aging coefficient parameters to ensure the reasonable physical meaning of the parameters. By constructing the optimal solution set of cyclic decay factor and calendar aging coefficient output by the improved particle swarm optimization algorithm, a parameter calibration layer is built, mapping the multi-mechanism coupling features in the shared physical driving feature layer to the physical parameter space of the battery aging dynamic model, providing the calibrated cyclic decay factor and calendar aging coefficient input to the dynamic state estimation layer.

[0062] A dynamic state estimation layer is constructed by fusing the calibrated cyclic decay factor with real-time data using the extended Kalman filter algorithm.

[0063] It should be noted that, in the process of constructing a dynamic state estimation layer by fusing the calibrated cycle decay factor with real-time data using the extended Kalman filter algorithm, the state vector is defined to include the capacity retention rate in the BMS health state parameters, the calibrated cycle decay factor, and the charge / discharge efficiency deviation in the PCS power parameters. A state equation is established to describe the dynamic relationship between the capacity retention rate and the cycle number and temperature gradient, where the cycle decay factor is used as a time-varying parameter in the calculation of the state transition matrix. The observation equation generates a capacity decay observation sequence based on the measured AC output power of the PCS and the real-time values ​​of the BMS state of charge parameters. Through iterative calculations of the prediction and update steps of the extended Kalman filter, combined with the calibrated cycle decay factor, real-time power fluctuations of the PCS, and BMS temperature parameters, the estimated capacity retention rate and its covariance matrix in the state vector are dynamically corrected. The final output dynamic state estimation layer includes the real-time updated BMS health state parameter estimation results, reflecting the coupled impact of PCS operating mode switching and EMS scheduling strategies on the battery aging process.

[0064] A dynamic battery aging model is constructed based on an input layer, a physical driving feature layer, a parameter calibration layer, and a dynamic state estimation layer, and trained using a hybrid collaborative training method.

[0065] It should be noted that, based on the PCS charge / discharge cycle count, BMS temperature gradient and duration, and EMS load prediction data in the input layer, a standardized state vector is generated and converted into a time series tensor. The physical driving feature layer extracts cycle decay features and calendar aging features through a multi-mechanism coupling analysis method. The cycle decay feature is based on the rainflow counting method to count the equivalent cycle count and combines the joint influence function of discharge depth and temperature to quantify the decay rate. The calendar aging feature is based on the temperature data to calculate the aging rate. The parameter calibration layer uses an improved particle swarm optimization algorithm to perform offline global parameter identification of the cycle decay factor and calendar aging coefficient. The constraints include the activation energy range and the physical upper limit of the cycle decay factor. The dynamic state estimation layer uses an extended Kalman filter algorithm to fuse the calibrated cycle decay factor with the real-time acquired capacity deviation and temperature data, and iteratively updates the health state estimate. The hybrid collaborative training method optimizes the physical driving parameters offline using an improved particle swarm optimization algorithm, and then uses an extended Kalman filter algorithm to fuse the real-time PCS charge / discharge data, BMS temperature monitoring values, and EMS load duration information online to achieve continuous adaptive calibration of model parameters and dynamic tracking of health state.

[0066] Based on standardized state vectors, a battery aging dynamic model is used to predict the health status and generate health status indicators.

[0067]

[0068] Where H is the standardized state vector, α is the cycle decay factor, ΔQ is the single-cycle capacity throughput, D is the depth of discharge, T is the battery operating temperature, β is the calendar aging coefficient, E is the activation energy, R is the gas constant, t is the number of cycles, f(D,T) is the function representing the joint influence of the depth of discharge D and temperature T on the battery cycle aging rate, correcting the actual capacity loss in each cycle, K(ΔQ,T) is the error correction term representing the dynamic calibration of the battery state of health (SOH) model based on the extended Kalman filter algorithm using the observed values ​​of capacity change ΔQ and temperature T, and S(t) is the battery health status index at time t;

[0069] It should be noted that, through the physical-driven feature layer, the cyclic decay feature and the calendar aging feature are modeled as the cyclic decay term α∑ΔQ and the calendar aging term, respectively. The discharge depth D and temperature T were obtained through rainflow counting and time-series sampling. The joint influence function f(D,T) quantifies the nonlinear superposition effect of different discharge depths and temperatures on cycle decay. The Arrhenius equation terms... The system reflects the accelerating effect of temperature on calendar aging. The parameter calibration layer dynamically updates the calibrated cycle decay factor α and calendar aging coefficient β based on the execution feature dataset using an online sequence extreme learning machine. It also uses an extended Kalman filter algorithm to generate an error correction function K(ΔQ,T) based on the observed residuals of the real-time capacity deviation ΔQ and temperature T, compensating for unmodeled dynamics. The dynamic state estimation layer uses the initial health state H as a baseline and iteratively calculates the real-time health state estimate S(t) by fusing the cycle decay term, calendar aging term, and error correction term using an extended Kalman filter algorithm. Specifically, the calculation process involves: weighting and summing the PCS charge / discharge cycle number ΔQ according to the joint influence function f(D,T) of the discharge depth D and temperature T, and then multiplying it by the calibrated cycle decay factor α to generate the cycle decay amount; combining the BMS temperature parameter T, activation energy E, and EMS runtime t, the calendar aging amount is calculated. After superimposing the error term K(ΔQ,T) of the extended Kalman filter dynamic correction, the cyclic decay and calendar aging are subtracted from the initial health state H to output the real-time health state index S(t), realizing multi-source data-driven dynamic modeling of battery aging and accurate state estimation.

[0070] Based on health status indicators, dynamic status correction is performed using an extended Kalman filter algorithm to generate real-time health status estimates.

[0071] It should be noted that, based on health status indicators (including capacity retention offset and calendar aging loss), during the dynamic state correction process using the extended Kalman filter algorithm, the state vector is defined as the battery health status capacity retention estimate, the calibrated cycle decay factor, and the calendar aging coefficient. A state transition equation is established to describe the dynamic evolution of capacity retention with the number of PCS charge-discharge cycles and BMS temperature parameters, where the calibrated cycle decay factor is used as a time-varying parameter in the calculation of the state transition matrix. The observation equation generates a capacity decay observation sequence based on the real-time measured values ​​of the BMS's state of charge parameters and the AC-side output power of the PCS. Through the prediction step (updating the prior estimate and covariance matrix based on the state transition equation) and the update step (calculating the Kalman gain and correcting the posterior estimate based on the observation equation) of the extended Kalman filter, the real-time charge-discharge power fluctuation of the PCS, the temperature gradient change of the BMS, and the cumulative operating time data of the EMS are iteratively fused to dynamically adjust the confidence interval of the capacity retention estimate and generate a real-time health status estimate. The real-time health status estimate reflects the synergistic effect of the PCS charge-discharge strategy and the BMS temperature control, providing dynamic correction input for the digital twin mapping engine.

[0072] Based on real-time health status estimates, multi-dimensional performance quantification is performed through a digital twin mapping engine to generate multi-dimensional performance evaluation indicators.

[0073] It should be noted that, based on real-time health status estimates, the real-time charging and discharging power data of the PCS and the time-of-use electricity price weighting coefficient of the EMS are first input into the economic quantification model. The economic benefit within the current scheduling window is calculated by integrating the product of charging and discharging power and electricity price, generating an economic score. Subsequently, the capacity retention rate and BMS temperature parameters from the real-time health status estimates are input into the lifetime degradation model. The temperature-dependent aging rate is calculated based on the Arrhenius equation, and the lifetime degradation rate index is generated by combining the number of charge-discharge cycles counted by the rainflow counting method and the calibrated cycle decay factor. Simultaneously, the AC voltage RMS deviation of the PCS, the load tracking error rate of the EMS, and the BMS state of charge fluctuation gradient are input into the stability assessment model. The grid stability index is generated by weighted fusion of the root mean square value of the voltage deviation, the cumulative absolute value of the load error, and the slope of the state of charge fluctuation. The final output multi-dimensional performance evaluation indicators (economic score, lifetime degradation rate, and grid stability index) comprehensively characterize the overall performance of the energy storage system in terms of benefit, battery life, and grid interaction, providing a quantitative decision-making basis for multi-objective collaborative optimization.

[0074] The multi-objective optimization module, based on multi-dimensional performance evaluation indicators, generates a set of cooperative scheduling strategies by combining fuzzy hierarchical analysis with an improved genetic algorithm multi-objective optimization solver.

[0075] Based on multidimensional performance evaluation indicators, multi-objective weight allocation is performed using fuzzy hierarchical analysis to generate priority weight vectors.

[0076] It should be noted that the target weight allocation logical framework is constructed using fuzzy hierarchical analysis. Its criterion layer consists of economic dimension, lifetime degradation dimension, and grid stability dimension. The data-driven fuzzy rule base is dynamically triggered by the real-time electricity price range, BMS temperature parameter, and PCS voltage deviation in the EMS load forecast data: when the real-time electricity price is higher than the historical average, the fuzzy importance level of the economic score relative to the lifetime degradation rate is set to "strongly correlated"; when the BMS temperature parameter exceeds the preset safety threshold, the fuzzy importance level of the lifetime degradation rate relative to the grid stability index is set to "extremely strongly correlated"; when the PCS AC voltage effective value deviation exceeds the grid allowable range, the fuzzy importance level of the grid stability index relative to the economic score is set to "moderately correlated". The aforementioned fuzzy rules are mapped to triangular fuzzy numbers to construct a fuzzy judgment matrix for the criterion layer. The fuzzy weight intervals for each dimension are calculated using the geometric mean method, and the fuzzy weight intervals are converted into deterministic scalar weight values ​​using defuzzification processing (centroid method). Priority weight vectors (economic weight, lifetime weight, and stability weight) are generated, providing a dynamic quantitative decision basis for the multi-objective optimization of the improved non-dominated sorting genetic algorithm, ensuring that the weight allocation strictly matches the real-time running state.

[0077] Based on the priority weight vector, a multi-objective optimization solution is obtained by using an improved non-dominated sorting genetic algorithm to generate the Pareto optimal solution set;

[0078] It should be noted that the improved non-dominated sorting genetic algorithm enhances performance based on the traditional non-dominated sorting genetic algorithm through dynamic crossover and mutation probability adjustment strategies and a crowding-based elite retention strategy. The dynamic crossover probability is adaptively adjusted based on the population's genetic diversity (based on gene sequence dissimilarity). When the similarity between individuals in the population is too high, the crossover probability is automatically increased to enhance global search capabilities. The mutation probability is adjusted inversely based on the individual's fitness ranking; individuals with lower fitness are assigned higher mutation probabilities to avoid local convergence. The crowding-based elite retention strategy, after non-dominated sorting, prioritizes individuals in sparsely distributed regions of the target space. By calculating the neighborhood density of individuals in the target space, solutions in highly crowded regions are retained to improve the uniformity of the Pareto front. These improved strategies, through dynamic parameter adjustment and solution set distribution optimization, enhance the algorithm's ability to collaboratively solve multi-objective problems under PCS operation modes, EMS scheduling requirements, and BMS health status constraints.

[0079] Furthermore, based on the priority weight vector, the process of performing multi-objective optimization solutions using an improved non-dominated sorting genetic algorithm is integrated as follows: First, in the population initialization stage, the PCS charging and discharging power setpoint sequence, EMS scheduling time period division strategy, and BMS state-of-charge operation interval constraints are encoded into individual genes, covering all feasible scheduling schemes. Subsequently, the fitness function is generated by weighted summation of multi-dimensional performance evaluation indicators using the priority weight vector, quantifying the comprehensive performance of individuals in multi-objective collaborative optimization. In this process, individuals in the population are divided according to Pareto levels through non-dominated sorting, and highly distributed individuals are selected as parents by combining crowding distance calculation. The offspring population is generated by performing multi-point crossover and boundary mutation operations on the parent gene sequence according to dynamically adjusted crossover and mutation probabilities. After merging the parent and offspring populations, the next generation of individuals is selected based on the crowding elite retention strategy to ensure that the solution set is evenly distributed in the target space and close to the Pareto front. Finally, the non-dominated solution set that satisfies the PCS power constraints, EMS economic requirements, and BMS health status protection is output, providing multi-dimensional optimization scheme input for the TOPSIS decision-making method.

[0080] Based on the Pareto optimal solution set, the TOPSIS decision method is used to select the best strategy and generate the final collaborative scheduling strategy set.

[0081] It should be noted that when using the TOPSIS decision-making method to select the optimal strategy based on the Pareto optimal solution set, a decision matrix is ​​first constructed. For each individual solution in the Pareto solution set, corresponding multi-dimensional performance evaluation indicators are listed, such as economic score, lifetime decay rate, and grid stability index. The decision matrix is ​​then normalized to eliminate dimensional differences. A weighted decision matrix is ​​generated by weighting the normalized matrix based on priority weight vectors. The positive ideal solution for each solution is calculated, for example, the solution with the highest economic score, lowest lifetime decay rate, and highest grid stability index; the negative ideal solution is calculated, for example, the solution with the lowest economic score, highest lifetime decay rate, and lowest grid stability index. The distance from each solution to the positive and negative ideal solutions is calculated using the Euclidean distance formula, generating a relative proximity score. The solutions are sorted in descending order of proximity score, and the solution with the highest score is selected as the final coordinated scheduling strategy set to ensure the optimal balance between PCS power output, EMS scheduling plan, and BMS health status protection.

[0082] The dynamic derating module, based on a collaborative scheduling strategy set, generates an executable instruction queue with security constraints by combining an industrial IoT protocol stack with a dynamic derating coefficient algorithm.

[0083] Based on the collaborative scheduling strategy set, instructions are securely encapsulated through the industrial IoT protocol stack to generate protocol-compatible instruction frames.

[0084] It should be noted that, based on the scheduling period division strategy of the collaborative scheduling strategy set and the BMS charge state, when encapsulating instructions securely through the Industrial Internet of Things (IIoT) protocol stack, the PCS charging and discharging power setting value sequence in the collaborative scheduling strategy set is first parsed into a power command code, the EMS scheduling period division strategy into a timestamp control code, and the BMS charge state operating interval constraint into a threshold check code. Using the message encapsulation rules of the IIoT protocol stack, the power command code, timestamp control code, and threshold check code are converted into protocol data units. Security fields are added to the protocol data units, such as encrypted message headers based on the TLS protocol, hash check codes, and device identifiers, generating protocol-compatible command frames that conform to IIoT security standards. These protocol-compatible command frames are adapted through the physical layer and data link layer interfaces of the IIoT protocol stack, ensuring that the instructions from the PCS, EMS, and BMS can be securely parsed and executed by the target device.

[0085] Based on protocol-compatible instruction frames, security constraints are injected through a dynamic derating factor algorithm to generate risk response instructions;

[0086] It should be noted that, based on protocol-compatible command frames, when injecting safety constraints using a dynamic derating factor algorithm, the process begins by parsing the PCS charging / discharging power command code within the protocol-compatible command frame to obtain the initial power setting value. Real-time monitoring of BMS temperature parameters, PCS DC-side current RMS value, and EMS load prediction error rate is then performed. Based on overload risk levels, such as temperature exceeding limits, current exceeding limits, or load deviation exceeding thresholds, a derating factor is dynamically calculated. This factor is then multiplied by the initial power setting value to generate a drated power command code. Simultaneously, a logical AND operation is performed between the derating factor and the BMS state-of-charge threshold check code to generate a derating constraint check code. Finally, the drated power command code, EMS scheduling period division strategy, and derating constraint check code are fused and re-encapsulated using the security fields (hash check code and device identifier) ​​of the industrial IoT protocol stack to generate a risk response command. This risk response command ensures that the PCS charging / discharging power dynamically adapts to the BMS health status and real-time operational risks, achieving collaborative scheduling under safety constraints.

[0087] Based on risk response instructions, an executable instruction queue with safety constraints is generated using an instruction priority sorting algorithm.

[0088] It should be noted that the specific process of generating an executable instruction queue with safety constraints based on risk response instructions and using an instruction priority sorting algorithm is as follows: The PCS derating power setting sequence, EMS scheduling time period division strategy, and BMS derating constraint check code in the risk response instructions are parsed. The security level identifier (based on the matching degree between the hash check code integrity verification result and the device identity identifier), timestamp control code urgency, and constraint check code compliance of each instruction are extracted. Through multi-level priority determination rules, the PCS charging and discharging power instructions are divided into basic priorities according to the security level identifier (high / medium / low risk). The priority weight of the EMS scheduling instructions is adjusted according to the time period urgency of the timestamp control code. The priority offset of the BMS constraint check code is dynamically corrected according to the degree of deviation of the state of charge from the threshold. A comprehensive priority score is generated by integrating the security level, urgency, and offset. The instruction sequence is arranged in descending order of the score to form an executable instruction queue with safety constraints. Finally, the queue is re-encapsulated through the security fields (hash check code and device identity identifier) ​​of the industrial IoT protocol stack to ensure that PCS charging and discharging operations, EMS scheduling execution, and BMS health status protection are implemented safely in priority order.

[0089] The circuit breaker monitoring module, based on an executable instruction queue, generates an execution feature dataset with anomaly codes through a real-time circuit breaker mechanism and power quality monitoring algorithm.

[0090] Based on the executable instruction queue, overload risk monitoring is performed through a real-time circuit breaker mechanism, generating circuit breaker events and abnormal tags.

[0091] It should be noted that, based on the executable instruction queue with safety constraints, when monitoring overload risks through a real-time fuse mechanism, the effective value of the PCS DC-side current, BMS temperature parameters, and EMS load prediction error rate are collected in real time and dynamically compared with the safety constraint thresholds (PCS charging and discharging power limit, BMS temperature protection threshold, and EMS load deviation allowable range) in the executable instruction queue. When the effective value of the PCS DC-side current exceeds the charging and discharging power limit, it is marked as an overcurrent fuse event and an overcurrent anomaly label is generated. When the BMS temperature parameter exceeds the temperature protection threshold, it is marked as an overtemperature fuse event and an overtemperature anomaly label is generated. When the EMS load prediction error rate exceeds the allowable range, it is marked as a load deviation fuse event and a load anomaly label is generated. The fuse events and anomaly labels are associated with timestamps, device identifiers, and the type of parameters exceeding the limits to form structured anomaly records, providing risk monitoring result input for the execution feature dataset.

[0092] Based on the execution process data of the executable instruction queue, dynamic parameter analysis is performed through power quality monitoring algorithms to generate quality anomaly waveforms;

[0093] Based on the execution process data of the executable instruction queue, when performing dynamic parameter analysis using the power quality monitoring algorithm, the following steps are taken: First, the effective value deviation of the PCS AC voltage, voltage harmonic distortion rate, and frequency offset are extracted. Combined with the active / reactive power fluctuation gradient in the EMS load tracking response data, the high-frequency noise and transient disturbance components of the voltage / current waveforms are decomposed using wavelet transform. Time-frequency joint analysis is performed on the transient fluctuations of the BMS state of charge to detect state of charge jump characteristics caused by abnormal charging and discharging. Waveform segments with voltage harmonic distortion rate exceeding a preset harmonic distortion rate threshold are marked as harmonic distortion anomalies, waveform segments with effective voltage deviation exceeding a preset voltage deviation threshold are marked as voltage sag anomalies, and waveform segments with frequency offset exceeding a preset frequency offset threshold are marked as frequency fluctuation anomalies. Based on the wavelet coefficient energy distribution and transient feature matching rules, comprehensive quality anomaly waveform data containing anomaly type, timestamp, and waveform segment is generated, providing power quality dimension anomaly feature input for the execution feature dataset.

[0094] Based on circuit breaker events and quality anomaly waveforms, feature fusion is performed using an anomaly coding rule base to generate an execution feature dataset with security labels.

[0095] It should be noted that, based on the circuit breaker event and the abnormal waveform, when performing feature fusion through the abnormal coding rule base, the excessive parameter types of the circuit breaker event, such as overcurrent, overtemperature, and load deviation, are first mapped to safety risk level codes, such as overcurrent being coded as "OL", overtemperature as "OT", and load deviation as "LD". Simultaneously, the abnormal types of the abnormal waveform, such as harmonic distortion, voltage sag, and frequency fluctuation, are mapped to quality defect level codes, such as harmonic distortion being coded as "HD", voltage sag as "VS", and frequency fluctuation as "FF". The circuit breaker event and the abnormal waveform are then aligned based on the timestamp. During the occurrence of constant waveforms, synchronous operating status data of PCS AC voltage / current waveforms, EMS load tracking response data, and BMS temperature parameters are extracted. Through the association rules of the anomaly coding rule base, such as mapping "OL+HD" to "high safety risk" and "OT+VS" to "medium safety risk", the safety risk level code and quality defect level code are combined to generate a composite safety label. The timestamp, device identifier, composite safety label, and synchronous operating status data are integrated to generate an execution feature dataset with safety labels, providing multi-dimensional anomaly feature input for the battery aging dynamic model.

[0096] The closed-loop optimization module updates the battery aging model parameters based on the execution feature dataset using an incremental online learning algorithm, and generates a power grid fluctuation feature library through the Arrhenius equation parameter optimizer.

[0097] Based on the execution feature dataset, dynamic parameter updates are performed using an online sequence extreme learning machine to generate preliminary calibration parameters;

[0098] It should be noted that when dynamically updating parameters based on the health status parameters of the execution feature dataset using an online sequence extreme learning machine, the PCS AC voltage RMS deviation, EMS load tracking error rate, and BMS temperature parameters from the execution feature dataset are first used as input feature vectors, with a safety label as a supervision signal. The hidden layer node weight matrix of the online sequence extreme learning machine is initialized, and time-series samples from the execution feature dataset are read batch by batch based on an incremental learning mode. The connection weights from the hidden layer to the output layer are updated using a recursive least squares method. For the dynamic offset of the BMS health status parameters, the PCS power correction coefficient and EMS scheduling compensation factor are adjusted through an error feedback mechanism to generate preliminary calibration parameters (including PCS charge / discharge efficiency compensation, EMS load prediction error correction weights, and BMS health status estimation offset). After verifying the data integrity using a hash checksum, the preliminary calibration parameters are input into the battery aging dynamic model to complete the iterative parameter update.

[0099] Based on the initial calibration parameters, the physical range is corrected by a regularized constraint optimizer to generate updated battery aging model parameters.

[0100] It should be noted that, based on the initial calibration parameters, when performing physical range correction using a regularized constraint optimizer, the physical feasible interval constraints of the battery aging dynamic model parameters are first defined: the PCS charge / discharge efficiency compensation is limited to a preset percentage fluctuation range of the rated efficiency, the EMS load prediction error correction weight is limited to the quantile interval of the historical error statistical distribution, and the BMS health state estimation offset is limited to the coupling correlation threshold between the state of charge and the health state. The parameter update magnitude is constrained by adding an L2 norm regularization term to prevent overfitting. A loss function with a regularization term is constructed, and the optimal parameter combination satisfying the physical feasible interval constraints is iteratively solved using the gradient descent algorithm. Updated battery aging model parameters are generated, ensuring a balance between physical interpretability and model accuracy.

[0101] Based on the execution feature dataset, temperature-dependent aging features are generated using a wavelet transform-Arrhenius temperature feature extractor.

[0102] It should be noted that, based on the execution feature dataset, the specific process of generating temperature-dependent aging features using wavelet transform-Arrhenius temperature feature extractor is as follows: Wavelet transform multi-scale decomposition is performed on the time-series records of BMS temperature parameters to extract the low-frequency trend component and high-frequency fluctuation component of the temperature waveform, identifying transient temperature shock events, such as rapid heating or cooling periods and steady-state temperature rise periods; the low-frequency trend component is input into the Arrhenius equation to calculate the average aging rate coefficient under steady-state temperature rise periods, and the high-frequency fluctuation component is calculated through time-frequency energy integration to determine the cumulative contribution of transient temperature shocks to the aging rate; the difference between the ambient temperature gradient in the PCS operating temperature waveform and the internal temperature distribution of the BMS is fused to generate a temperature spatial non-uniformity compensation factor; combining the steady-state aging rate coefficient, transient shock contribution, and temperature non-uniformity compensation factor, a temperature-dependent aging feature vector is generated through a weighted fusion formula to characterize the coupling effect of temperature at different time scales and spatial dimensions on the battery aging dynamic model, providing physical driving feature input for the parameter calibration layer.

[0103] Based on temperature-dependent aging characteristics, a power grid fluctuation feature library is generated by multi-dimensional feature clustering using a spectrum clustering analyzer.

[0104] It should be noted that the specific process of generating a power grid fluctuation feature library based on temperature-dependent aging characteristics through multi-dimensional feature clustering using a spectrum clustering analyzer is as follows: Extracting the time series of steady-state aging rate coefficients, the frequency domain energy distribution of transient impact contribution, and the spatial gradient parameters of temperature non-uniformity compensation factors from the temperature-dependent aging characteristics to construct a multi-dimensional feature matrix; calculating the similarity matrix of the feature matrix using a Gaussian kernel function, generating a Laplace matrix, and performing feature decomposition to obtain a low-dimensional embedding vector representation; based on the correlation constraint between PCS AC voltage harmonic distortion rate and EMS frequency offset, performing K-means clustering on the low-dimensional embedding vector to classify high-frequency harmonic fluctuation modes, low-frequency oscillation fluctuation modes, and transient impact fluctuation modes; associating the clustering results with PCS voltage waveform segments, EMS load dispatch periods, and BMS temperature transient events to generate a power grid fluctuation feature library containing fluctuation mode labels, feature vectors, and timestamps, providing temperature-grid interactive coupling feature input for the battery aging dynamic model.

[0105] In summary, this invention achieves nonlinear coupling modeling of cyclic decay and calendar aging mechanisms in the dynamic battery aging model through a physical driving feature layer and a dynamic parameter calibration layer, thereby improving the accuracy of health status prediction and model adaptability.

[0106] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An integrated dispatch system for implementing PCS, EMS and BMS, characterized in that: include, The data preprocessing module collects equipment operating status data in real time, performs preprocessing, and generates standardized status vectors. The multidimensional performance evaluation module constructs and trains a dynamic battery aging model. Based on standardized state vectors, it dynamically corrects and maps the battery aging dynamic model to digital twins to generate multidimensional performance evaluation indicators. The multi-objective optimization module, based on multi-dimensional performance evaluation indicators, generates a set of cooperative scheduling strategies by combining fuzzy hierarchical analysis with an improved genetic algorithm multi-objective optimization solver. The dynamic derating module, based on a collaborative scheduling strategy set, generates an executable instruction queue with security constraints by combining an industrial IoT protocol stack with a dynamic derating coefficient algorithm. The circuit breaker monitoring module, based on an executable instruction queue, generates an execution feature dataset with anomaly codes through a real-time circuit breaker mechanism and power quality monitoring algorithm. The closed-loop optimization module updates the battery aging model parameters based on the execution feature dataset using an incremental online learning algorithm, and generates a power grid fluctuation feature library through the Arrhenius equation parameter optimizer. The process involves generating multi-dimensional performance evaluation metrics based on standardized state vectors, dynamically correcting and mapping the data using a battery aging dynamic model, and then mapping the data to a digital twin. The specific steps are as follows: Based on standardized state vectors, a battery aging dynamic model is used to predict the health status and generate health status indicators. Based on health status indicators, dynamic status correction is performed using an extended Kalman filter algorithm to generate real-time health status estimates. Based on real-time health status estimates, multi-dimensional performance quantification is performed through a digital twin mapping engine to generate multi-dimensional performance evaluation indicators. The process involves generating a collaborative scheduling strategy set based on multidimensional performance evaluation metrics, using a multi-objective optimization solver combining fuzzy hierarchical analysis and an improved genetic algorithm. The specific steps are as follows: Based on multidimensional performance evaluation indicators, multi-objective weight allocation is performed using fuzzy hierarchical analysis to generate priority weight vectors. Based on the priority weight vector, a multi-objective optimization solution is obtained by using an improved non-dominated sorting genetic algorithm to generate the Pareto optimal solution set; Based on the Pareto optimal solution set, the TOPSIS decision method is used to select the best strategy and generate the final collaborative scheduling strategy set.

2. The integrated dispatch system for implementing PCS, EMS and BMS as claimed in claim 1 wherein: The equipment operating status data includes PCS power and voltage, EMS load prediction data, and BMS state of charge and health parameters.

3. The integrated dispatch system for implementing PCS, EMS and BMS as claimed in claim 2 wherein: The preprocessing includes denoising, normalization, time alignment of multi-source heterogeneous data, and feature fusion.

4. The integrated dispatch system for implementing PCS, EMS and BMS as claimed in claim 3 wherein: The specific steps for constructing and training a dynamic battery aging model are as follows. An input layer is constructed based on a multi-source heterogeneous data fusion method to convert standardized state vectors into time series tensors; By using the Arrhenius multi-mechanism coupling analysis method, joint features of cyclic decay and calendar aging are extracted from the input data, and a shared physical driving feature layer is constructed. A parameter calibration layer is constructed by performing nonlinear parameter fitting on shared features using an improved particle swarm optimization algorithm. A dynamic state estimation layer is constructed by fusing the calibrated cyclic decay factor with real-time data using the extended Kalman filter algorithm. The battery aging dynamic model is constructed based on an input layer, a physical driving feature layer, a parameter calibration layer and a dynamic state estimation layer, and is trained by a hybrid co-training method.

5. The integrated dispatch system for implementing PCS, EMS and BMS as claimed in claim 1 wherein: The collaborative scheduling strategy set is combined with the industrial Internet of Things protocol stack and a dynamic derating coefficient algorithm to generate an executable instruction queue with safety constraints, and the specific steps are as follows, Based on the collaborative scheduling strategy set, the instruction safety encapsulation is performed through the industrial Internet of Things protocol stack to generate a protocol compatible instruction frame. Based on the protocol compatible instruction frame, the safety constraint injection is performed through the dynamic derating coefficient algorithm to generate a risk response instruction. Based on the risk response instruction, an executable instruction queue with safety constraints is generated through an instruction priority sorting algorithm.

6. The integrated dispatch system for implementing PCS, EMS and BMS as claimed in claim 1 wherein: Based on the executable instruction queue, an execution feature data set with abnormal codes is generated through a real-time fusing mechanism and a power quality monitoring algorithm, and the specific steps are as follows, Based on the executable instruction queue, overload risk monitoring is performed through a real-time fusing mechanism to generate a fusing event and an abnormal label. Based on the execution process data of the executable instruction queue, dynamic parameter analysis is performed through a power quality monitoring algorithm to generate a quality abnormal waveform. Based on the fusing event and the quality abnormal waveform, feature fusion is performed through an abnormal coding rule library to generate an execution feature data set with a safety label.

7. The integrated dispatch system for implementing PCS, EMS and BMS as claimed in claim 1 wherein: Based on the execution feature data set, the battery aging model parameters are updated through an incremental online learning algorithm, and the specific steps are as follows, Based on the execution feature data set, dynamic parameter updating is performed through an online sequence extreme learning machine to generate preliminary calibration parameters. Based on the preliminary calibration parameters, physical range correction is performed through a regularization constraint optimizer to generate updated battery aging model parameters.

8. The integrated dispatch system for implementing PCS, EMS and BMS as claimed in claim 1 wherein: The power grid fluctuation feature library is generated through an Arrhenius equation parameter optimizer, and the specific steps are as follows, Based on the execution feature data set, temperature-dependent aging features are generated through a wavelet transform-Arrhenius temperature feature extractor. Based on the temperature-dependent aging features, multi-dimensional feature clustering is performed through a spectral clustering analyzer to generate a power grid fluctuation feature library.

Images