Energy storage power station operation efficiency improvement and optimization method based on big data analysis

Abstract

The application discloses a method for improving and optimizing operation efficiency of energy storage power station based on big data analysis, and relates to the technical field of energy storage power station operation management.The application collects full life cycle data of the same type of battery, clusters the data based on a GS-ADC algorithm, obtains three final clusters and determines the stages;for different stage clusters, a segmented linear regression, an adaptive kernel function RBF neural network and a grey prediction GM(1,1) model are respectively constructed; the battery management system collects SOH, capacity attenuation acceleration and internal resistance growth rate in real time, forms a real-time feature vector, calculates the weighted Mahalanobis distance between the real-time feature vector and each cluster center, and judges the stage and calls the model based on the weighted Mahalanobis distance; and based on the predicted residual life in different stages, a battery adjustment strategy is set.The application improves the stage division accuracy and prediction accuracy, and realizes the collaborative optimization of the operation efficiency of the energy storage power station and the battery life.

Description

Technical Field

[0001] This invention relates to the field of energy storage power station operation and management technology, specifically to a method for improving and optimizing the operation efficiency of energy storage power stations based on big data analysis. Background Technology

[0002] As a key component of the new power system, the operating efficiency of energy storage power stations is closely related to battery performance management. The aging stage division of the battery's entire life cycle, the prediction of remaining life, and dynamic adjustment strategies are the core factors affecting the operating efficiency of energy storage power stations. The main shortcomings of existing technologies are as follows:

[0003] Firstly, in terms of dividing battery aging stages, existing technologies mostly rely on fixed empirical thresholds or traditional clustering algorithms, without combining the dynamic correlation of the core features of battery aging for adaptive clustering, resulting in blurred stage boundaries and easy mixing of samples from different aging stages.

[0004] Secondly, the design of the remaining life prediction model lacks stage adaptability. Existing technologies often use a single model to cover the entire battery life cycle, ignoring the different aging mechanisms of stable growth of the SEI film in the early stage, accelerated degradation in the middle stage, and sharp performance drop in the end stage, resulting in low prediction accuracy.

[0005] Third, the real-time operation adjustment strategy lacks dynamism. Existing strategies are mostly based on fixed charging and discharging parameters or single state indicators, without combining real-time feature vectors with the current aging stage for coordinated adjustment, making it difficult to balance operating efficiency and battery life.

[0006] Therefore, there is an urgent need for a method to improve and optimize the operating efficiency of energy storage power stations based on big data analysis. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis, which solves the problems of insufficient battery stage division, lifespan prediction, and strategy adjustment in energy storage power stations.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis, comprising:

[0009] Step 1: Collect full life cycle data of the same type of battery, and cluster them based on the Gravity Search-Adaptive Density Clustering (GS-ADC) algorithm to obtain three final clusters Q1, Q2, and Q3. Then determine the stage of the final clusters, which are the initial stage, the middle stage, and the final stage.

[0010] Step 2: For clusters in the initial, middle and final stages, construct piecewise linear regression, adaptive kernel function RBF neural network and grey prediction GM(1,1) model respectively to predict the remaining life of the battery in each stage.

[0011] Step 3: Real-time data collection of SOH, capacity decay acceleration, and internal resistance growth rate is performed through the Battery Management System (BMS) to form a real-time feature vector. The weighted Mahalanobis distance between the vector and the cluster centers of Q1, Q2, and Q3 is calculated. Based on the distance, stage judgment and prediction model are called. Based on the predicted remaining life (RUL) of different stages, different adjustment strategies for the battery are set.

[0012] As a further aspect of the present invention, the specific steps for clustering the full life cycle data of batteries of the same model based on the Gravity Search-Adaptive Density Clustering (GS-ADC) algorithm are as follows:

[0013] The SOH value, capacity decay acceleration, and internal resistance growth rate were used as the three-dimensional clustering features, and the obtained three-dimensional features were min-max normalized and mapped to the [0,1] interval.

[0014] The sample potential is defined by the following formula: , where w j Let d be the feature weights of sample j. ij Let be the Euclidean distance between samples i and j, and α be the attenuation coefficient;

[0015] Determine the density threshold for each sample. ,in, Let be the average density of the 10 nearest samples around sample i, and max(E) be the maximum gravitational potential energy of all samples.

[0016] Calculate the local density for each sample ,filter And E i Samples with a ≥0.6×max(E) value are used as the initial core samples, where d th Distance threshold;

[0017] For a non-core sample i, calculate the gravitational interaction strength between it and each initial core sample k. The samples are assigned to the cluster of the initial core sample with the strongest gravitational influence, thus forming the initial cluster;

[0018] The intra-cluster eigenvalue CV of the initial cluster is calculated using the formula: CV = std(intra-cluster eigenvalue) / mean(intra-cluster eigenvalue). SOH CV ron If a certain cluster CV SOH >CV th or CV ron >CV th1The cluster is then partitioned according to the local minimum of gravitational potential energy, where std() is the standard deviation function, mean() is the mean function, and CV is the mean value. SOH CV is the fluctuation coefficient of SOH value. ron CV is the capacity decay acceleration fluctuation coefficient. th CV th1 The threshold for the volatility coefficient;

[0019] If the distance between the centers of the two clusters is less than 0.2 × max(d) all If the two clusters are merged into one, then three clusters are ultimately retained, where max(d) all ) represents the maximum distance between all cluster centers;

[0020] Calculate the intra-cluster distance percentage for each sample If r i >r th If it is an isolated point, it is classified as an isolated point and assigned to the cluster with the nearest distance and whose CV meets the requirements, where d cu It is the set of distances between a single sample and all other samples in the cluster;

[0021] Cluster purity is verified based on the existing GS-ADC model. After the verification is successful, the collected full life cycle data of the same type of battery needs to be clustered again based on the GS-ADC model with modified parameters to obtain three final clusters Q1, Q2 and Q3.

[0022] As a further aspect of the present invention, the specific process for splitting the cluster according to the local minimum of gravitational potential energy is as follows:

[0023] Determine the sample set S={s1,s2,...,sn} of the cluster to be split, and calculate the potential energy values ​​{E1,E2,...,En} of all samples, where n is the number of samples in the cluster;

[0024] The samples within a cluster are sorted in descending order according to their Soh values ​​to obtain an ordered sample sequence Sord={sq1,sq2,...,sqn};

[0025] For each sample sqk in the ordered sample sequence Sord, excluding the first and last samples, if Eqk < Eq(k-1) and Eqk < Eq(k+1), that is, the potential energy of the sample is less than the potential energy of the previous and next adjacent samples, then sqk is a local minimum point. This point is used as the split boundary, where 2 ≤ k ≤ n-1.

[0026] If there are multiple local minima, the point with the lowest potential energy is selected as the final split point.

[0027] Using the finally located local minimum point sqk as the boundary, the ordered sample sequence Sord is split into two parts: sub-cluster S1={sq1,...,sq(k-1)} and sub-cluster S2={sq(k+1),...,sqn}, and the local minimum point sqk is removed;

[0028] Calculate the CV of the two subclusters separately. SOH CV ron If both CV values ​​of the two subclusters are not greater than the preset threshold, then the split is valid;

[0029] If a sub-cluster still does not meet the CV threshold, repeat the above splitting steps until the feature fluctuations of all sub-clusters meet the requirements.

[0030] As a further aspect of the present invention, the specific steps for cluster purity verification based on the existing GS-ADC model are as follows:

[0031] Ten groups of batteries with the same model as the batteries used in the energy storage power station to be optimized were selected. Data for each group of batteries was obtained according to the "full life cycle aging test": starting from the new battery, data on SOH, capacity decay acceleration, and internal resistance growth rate were recorded every 100 cycles until the battery was scrapped.

[0032] Each group of batteries is labeled with stage tags for the initial, middle and final stages, ultimately forming 10 validation subsets, each containing 50 samples from the initial stage, 50 samples from the middle stage, and 50 samples from the final stage, for a total of 1500 labeled samples in the validation set.

[0033] Substituting 1500 labeled validation samples into the GS-ADC model with parameters completely consistent with the actual clustering, we obtained 3 validation clusters, denoted as V1, V2, and V3.

[0034] The number of labeled samples belonging to the initial, middle and final stages in each validation cluster is counted. The cluster purity is calculated according to the formula: (number of samples in the stage with the largest number in the cluster / total number of samples in the cluster). The stage with the largest number in the cluster is taken as the stage of the cluster. Clusters with purity <90% are selected as the target clusters for this parameter adjustment.

[0035] For the target cluster, calculate the gravitational potential energy difference between its core sample and misclustered sample. If the difference is less than 0.2 × the average potential energy difference, then α needs to be adjusted. If the purity is still less than 90% after increasing α to 1.0, then adjust the core sample density threshold. The average potential energy difference is the average potential energy difference between the core sample and the samples in the same cluster.

[0036] As a further aspect of the present invention, the specific rules for adjusting α are as follows:

[0037] Set the initial α to 0.5, and increase it stepwise in increments of 0.1 each time. The maximum adjusted α should not exceed 1.0. After each adjustment, the purity of the target cluster needs to be recalculated. If it is ≥90%, stop adjusting α.

[0038] Adjusting the core sample density threshold The specific rules are as follows:

[0039] For the core samples of the target cluster, extract their original adaptive density threshold. ;

[0040] In the core statistical sample, for samples where the proportion of samples from different stages in the neighborhood is greater than 20%, calculate the percentage of these samples. Average threshold of core samples within the cluster ,calculate The ratio of the sample to the total sample is used. If the ratio is ≥80%, the adaptive threshold is judged to be too low. The original dynamic coefficient of 0.3 in the threshold formula is increased stepwise by "+0.1 each time", with a maximum of 0.7.

[0041] The adaptive density threshold for all samples was recalculated based on the new coefficients. The core samples are re-screened, and the purity of the target cluster is calculated. If the purity is ≥90%, the current dynamic coefficient and the corresponding adaptive threshold formula are locked.

[0042] As a further aspect of the present invention, after obtaining the final clusters Q1, Q2, and Q3, it is necessary to determine the stage of each cluster, namely the initial stage, the middle stage, and the final stage: calculate the center distance dis(i,j) between cluster Qi and cluster Vj respectively, find the Vj corresponding to min(dis(i,j)) for Qi, and take the stage of Vj as the stage corresponding to Qi, where i and j ∈ [1,3].

[0043] As a further aspect of the present invention, the specific steps for constructing a piecewise linear regression model are as follows:

[0044] Collect the DOD value, cycle number, and capacity retention rate of the target battery during the initial charge-discharge cycle.

[0045] Two types of abnormal samples were removed: the fluctuation of the DOD value in a single charge / discharge cycle compared to the command value was greater than ΔD; and the fluctuation of the capacity retention rate compared to the previous cycle was greater than ΔR. Here, ΔD is the upper limit of the allowable DOD fluctuation of the charge / discharge system, and ΔR is the upper limit of the accuracy of the detection equipment.

[0046] Two DOD thresholds, D1 and D2, are set, and the valid data is divided into three intervals: if the DOD of a single charge-discharge cycle is less than or equal to D1, it is classified as the low DOD interval; if D1 is less than the DOD of a single charge-discharge cycle and less than or equal to D2, it is classified as the medium DOD interval; if the DOD of a single charge-discharge cycle is greater than or equal to D2, it is classified as the high DOD interval; where D1 is less than D2.

[0047] For each DOD interval, the cumulative number of cycles for the samples within that interval is used as the independent variable, and the capacity retention rate after the corresponding cycle is used as the dependent variable. The least squares method is used to fit the linear equation, resulting in three fitting curves for the "number of cycles - capacity retention rate" of different DOD intervals.

[0048] Substitute the current cycle number into the fitting equation of the corresponding DOD interval to obtain the current capacity retention rate, calculate the decay amount = current capacity retention rate - retirement threshold, and then obtain the predicted remaining lifetime RUL = decay amount / fitting equation slope through the fitted decay rate, i.e. the slope of the fitted equation.

[0049] Calculate the deviation rate between the predicted value and the measured value. If the deviation rate is greater than the threshold, readjust D1 and D2.

[0050] As a further aspect of the present invention, the specific steps for constructing the adaptive kernel function RBF neural network are as follows:

[0051] Collect data on the voltage plateau rate of voltage drop, temperature fluctuation amplitude, and remaining lifetime of the target battery during its mid-term operation.

[0052] The samples are divided into three groups according to the temperature fluctuation range: if the temperature fluctuation range is ≤ T1, it is classified as small fluctuation; if T1 < temperature fluctuation range ≤ T2, it is classified as medium fluctuation; if the temperature fluctuation range is > T2, it is classified as large fluctuation; where T1 and T2 are the lower limit and upper limit of the fluctuation range, respectively.

[0053] Each sample group corresponds to a voltage plateau rate of drop sequence, with a length denoted as m;

[0054] The RBF neural network structure is determined as follows: the number of neurons in the input layer is the length m of the voltage drop rate sequence, the number of neurons in the hidden layer is h, and the number of neurons in the output layer is 1.

[0055] The kernel function used is the Gaussian kernel function, and the specific formula is as follows: Where c is the average pressure drop rate of each group, and σ is dynamically adjusted with temperature fluctuations: small fluctuations correspond to σ1, medium fluctuations correspond to σ2, and large fluctuations correspond to σ3, where σ3 < σ2 < σ1;

[0056] Using the grouped pressure drop rate sequence as input and remaining lifetime as output, the loss function is minimized using gradient descent. If the MSE of all groups is less than or equal to the threshold MSE0, it indicates that all trained models have reached the fitting accuracy.

[0057] Test the trained model with several sets of new mid-term battery data: Input a set of voltage drop rate sequences and compare the deviation between the model's predicted remaining lifetime and the measured value. If the deviation of a certain fluctuation group is greater than ΔL, adjust the σ value or the number of hidden layer neurons h of that group and retrain the model until the deviation is less than or equal to ΔL, where ΔL is the preset upper limit of lifetime deviation.

[0058] As a further aspect of the present invention, the specific steps for constructing the grey prediction GM(1,1) model are as follows:

[0059] Collect data on the instantaneous increase in internal resistance and the range of SOC fluctuation at the end of the target battery period, and remove samples with an instantaneous increase in internal resistance < 0 and a range of SOC fluctuation < ΔSOC0, where ΔSOC0 is the minimum fluctuation range allowed by the charging and discharging strategy.

[0060] Construct a non-negative original sequence X(0)={x(0)(1),x(0)(2),...,x(0)(n)} of instantaneous internal resistance increments, where x(0)(i) is the instantaneous internal resistance increment of the i-th record;

[0061] The original sequence X(0) is summed once to generate the sequence X(1)(k)=sum(X(0)(i)), where sum() is the summation function and i∈[1,n].

[0062] Establish the gray differential equation: X(0)(k)+a×X(1)(k)=b, where a is the development coefficient and b is the gray action quantity;

[0063] Solve for the parameter vector [a, b] using the least squares method. T The prediction equation for the cumulative sequence is obtained as follows: ;

[0064] The predicted value of the original sequence is obtained by cumulative subtraction: That is, the predicted value of the instantaneous increment of internal resistance in the (k+1)th iteration;

[0065] The cumulative internal resistance value Rpred is obtained by successively accumulating the predicted instantaneous increment of internal resistance. The number of cycles when Rpred first ≥ rth is found, and the difference between Rpred and the current number of cycles is calculated to obtain the preliminary remaining lifetime L. At the same time, Rpred / rth is calculated.

[0066] If (Rpred / rth)≤ If L does not need to be corrected; If (Rpred / rth) ≤ 1, the remaining lifespan needs to be adjusted to obtain L × ε; if (Rpred / rth) > 1, the battery needs to be immediately deactivated, and L should be set to 0; where, To approximate the threshold, ε is a correction coefficient.

[0067] As a further aspect of the present invention, the specific operation process for determining the stage based on distance is as follows:

[0068] Extract the cluster centers of clusters Q1, Q2, and Q3, denoted as C1 (early stage), C2 (middle stage), and C3 (late stage). Each center contains three features: SOH (saturation loss), denoted as f1; capacity decay acceleration, denoted as f2; and internal resistance growth rate, denoted as f3, forming a center feature matrix.

[0069] ;

[0070] The stage discrimination variance of each feature among the three centers is calculated using the following formula: ,in, Let j be the mean of this feature at the three centers, j∈[1,3];

[0071] Calculate weights The weight matrix W = diag(wf1, wf2, wf3) is obtained.

[0072] The real-time feature vector Y = [Y(f1), Y(f2), Y(f3)] is collected by BMS, where Y(f1) is the SOH after Kalman filtering correction, Y(f2) is the average capacity decay acceleration of the most recent 100 cycles, and Y(f3) is the average internal resistance growth rate of the day.

[0073] First, calculate the covariance matrix of the central features of the three stages. Construct a weighted covariance matrix Calculate the weighted Mahalanobis distance between the real-time feature vector Y and each center, using the following formula: We obtain D(Y,C1), D(Y,C2), and D(Y,C3) respectively, where, It is the inverse of the weighted covariance matrix;

[0074] Select the minimum value among the three distances and take the corresponding center stage as the candidate stage. Calculate the difference ΔDis between the minimum distance and the second smallest distance. If ΔDis > ΔDth, the candidate stage is the final stage. If ΔDis ≤ ΔDth, the stage at the previous moment is maintained. ΔDth is a preset threshold.

[0075] This invention provides a method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis, which has the following advantages compared with existing technologies:

[0076] (1) This invention adopts the Gravity Search-Adaptive Density Clustering GS-ADC algorithm, with battery SOH, capacity decay acceleration and internal resistance growth rate as three-dimensional core features. Combined with adaptive density threshold setting and cluster purity verification iterative optimization, it effectively solves the sample mixing problem caused by the traditional clustering method relying on fixed thresholds, and ensures that the initial, middle and final cluster structures are highly matched with the actual aging law of the battery.

[0077] (2) In view of the different mechanisms of battery aging at different stages, this invention constructs piecewise linear regression, adaptive kernel function RBF neural network and grey prediction GM(1,1) model respectively, avoiding the problem of insufficient adaptability caused by traditional single model covering the whole life cycle, and significantly improving the accuracy of remaining life prediction.

[0078] (3) Based on the feature vector and stage judgment results collected in real time by BMS, this invention formulates a differentiated adjustment strategy in combination with the remaining life (RUL) tiers to solve the contradiction of "over-limiting efficiency" or "accelerating aging" in the traditional fixed parameter strategy. Furthermore, through the staged coordinated adjustment of parameters such as charge / discharge depth, voltage, and rate, the operating efficiency of the energy storage power station and the utilization of the battery throughout its entire life cycle are balanced. Attached Figure Description

[0079] Figure 1 This is a flowchart of the steps of the present invention;

[0080] Figure 2 This is a flowchart illustrating the steps of splitting clusters in this invention;

[0081] Figure 3 This is a flowchart illustrating the steps involved in verifying cluster purity according to the present invention. Detailed Implementation

[0082] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0083] Example 1

[0084] like Figure 1 This invention provides a method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis, including:

[0085] Step 1: Collect full lifecycle data of batteries of the same model and cluster them based on the Gravity Search-Adaptive Density Clustering (GS-ADC) algorithm. The specific steps are as follows:

[0086] Using SOH value, capacity decay acceleration, and internal resistance growth rate as three-dimensional clustering features, we collect full life cycle data of batteries of the same model, covering new batteries to scrapped state.

[0087] SOH value is a core indicator for measuring the "current state" of battery aging. It quantifies health by the ratio of current capacity to rated capacity, and can intuitively correspond to the stage division of the initial stage (high SOH), the middle stage (medium SOH), and the final stage (low SOH). At the same time, by comparing with the initial state, it can avoid the misjudgment of "false health" caused by looking only at the current capacity.

[0088] Capacity decay acceleration focuses on the "dynamic trend" of aging, which can make up for the static lag of SOH. That is, it can distinguish batteries with the same SOH but different decay rates, avoiding a one-size-fits-all approach to clustering. Its core advantage is to keenly capture the turning points of the aging stage: slow decay in the early stage (acceleration close to 0), accelerated decay in the middle stage (acceleration absolute value increases), and cliff-like decay in the final stage (acceleration increases sharply), making the stage boundaries clearer.

[0089] The calculation is performed every 100 cycles, and the specific formula is: (current capacity decay rate - previous capacity decay rate) / difference in cycle number;

[0090] The internal resistance growth rate reflects the "intrinsic root cause" of aging and is a direct mapping of internal structural degradation. It can penetrate short-term fluctuations in state of harm (SOH), avoiding superficial judgment biases, and can also distinguish different failure modes, allowing cells in the same cluster to have similar aging mechanisms, facilitating subsequent targeted protection. In addition, the exponential increase in the internal resistance growth rate can accurately identify cells nearing failure, avoiding confusion between cells with low SOH in the middle stage and those in the final stage.

[0091] The daily calculation is based on the formula: (current day's internal resistance - yesterday's internal resistance) / yesterday's internal resistance.

[0092] The obtained three-dimensional features are min-max normalized and mapped to the [0,1] interval;

[0093] To avoid bias caused by subjective weight allocation based solely on experience, grey relational analysis was used to calculate the correlation between the three features and the actual cycle life based on the battery aging mechanism, and feature weights were allocated according to the correlation ratio.

[0094] Setting GS-ADC model parameters: Define the sample potential energy, the specific formula is as follows: Higher potential energy indicates a stronger attraction of the sample to its surroundings, making it more likely to become a core sample. Among these, w... j Let d be the feature weights of sample j. ij Let be the Euclidean distance between samples i and j, and α be the attenuation coefficient;

[0095] The density threshold for each sample is determined by multiplying the local density mean by the dynamic coefficient. To ensure higher thresholds in high-density regions and avoid excessive cluster splitting, Let be the average density of the 10 nearest samples around sample i, and max(E) be the maximum gravitational potential energy of all samples.

[0096] Calculate the local density for each sample ,filter And E i Samples with a ≥0.6×max(E) value are used as the initial core samples, where d th Distance threshold;

[0097] For a non-core sample i, calculate the gravitational interaction strength between it and each initial core sample k. The samples are assigned to the cluster of the initial core sample with the strongest gravitational influence, thus forming the initial cluster;

[0098] The intra-cluster eigenvalue CV of the initial cluster is calculated using the formula: CV = std(intra-cluster eigenvalue) / mean(intra-cluster eigenvalue). SOH CV ron If a certain cluster CV SOH >CV th or CV ron >CV th1 The cluster is then partitioned according to the local minimum of gravitational potential energy, where std() is the standard deviation function, mean() is the mean function, and CV is the mean value. SOH CV is the fluctuation coefficient of SOH value. ron CV is the capacity decay acceleration fluctuation coefficient. th CV th1 The fluctuation coefficient thresholds should be set according to the actual situation.

[0099] CV SOH Reflecting the relative fluctuation of SOH values ​​within the cluster, the differences in battery health status at the same stage should be small. If CV SOH >CV th This indicates that samples with different SOH levels may have been mixed in within the cluster (such as samples with high SOH in the early stage and low SOH in the middle stage), which does not conform to the characteristics of a single stage.

[0100] CV ron Reflecting the relative fluctuations in aging rate within the cluster, the aging rate trend should be consistent within the same stage. If CV ron >CV th1 This indicates that the cluster may contain both slow-decaying and fast-decaying samples, which violates the aging pattern within a stage.

[0101] Therefore, when these two CVs exceed the threshold, it indicates that the samples within the cluster are not in the same aging stage and need to be split to ensure the feature stability of each cluster;

[0102] If the distance between the centers of the two clusters is less than 0.2 × max(d) all If the two clusters are merged into one, then three clusters are ultimately retained, where max(d) all ) represents the maximum distance between all cluster centers;

[0103] Calculate the intra-cluster distance percentage for each sample If r i >r th If it is an isolated point, it is classified as an isolated point and assigned to the cluster with the nearest distance and whose CV meets the requirements, where d cu It is the set of distances between a single sample and all other samples in the cluster;

[0104] To further verify the accuracy of clustering, cluster purity needs to be verified based on the existing GS-ADC model;

[0105] After the cluster purity verification is passed, the collected full life cycle data of the same type of battery needs to be clustered again based on the GS-ADC model with modified parameters to obtain three final clusters Q1, Q2 and Q3.

[0106] After obtaining the three final clusters, it is also necessary to determine the stage of these three clusters, namely the initial stage, the middle stage, and the final stage: calculate the center distance dis(i,j) between cluster Qi and cluster Vj respectively, find the Vj corresponding to min(dis(i,j)) for Qi, and take the stage of Vj as the stage corresponding to Qi, where i, j∈[1,3].

[0107] Step 2: For clusters in the initial, middle and final stages, construct piecewise linear regression, adaptive kernel function RBF neural network and grey prediction GM(1,1) model respectively to predict the remaining life of the battery in each stage.

[0108] For clusters in the initial stage, the specific steps for constructing a piecewise linear regression model are as follows:

[0109] Collect the DOD value, cycle number, and capacity retention rate of the target battery during the initial charge-discharge cycle.

[0110] Two types of abnormal samples were removed: ① The fluctuation of the DOD value in a single charge / discharge cycle compared to the command value is greater than △D (△D is the upper limit of the allowable DOD fluctuation of the charge / discharge system); ② The fluctuation of the capacity retention rate compared to the previous cycle is greater than △R (△R is the upper limit of the accuracy of the detection equipment).

[0111] Based on the target battery's conventional charge and discharge strategy, two DOD thresholds, D1 and D2, are set, and the valid data is divided into three intervals: if the single charge and discharge DOD ≤ D1, it is classified as the low DOD interval; if D1 < single charge and discharge DOD ≤ D2, it is classified as the medium DOD interval; if single charge and discharge DOD > D2, it is classified as the high DOD interval; where D1 < D2.

[0112] For each DOD interval, the cumulative number of cycles of the samples within that interval is used as the independent variable (reflecting the cumulative aging process), and the capacity retention rate after the corresponding cycle is used as the dependent variable (reflecting the current aging state). The least squares method is used to fit the linear equation, resulting in three fitting curves of "number of cycles - capacity retention rate" for different DOD intervals, which correspond to the initial capacity decay pattern under different single load intensities.

[0113] Substitute the current cycle number into the fitting equation of the corresponding DOD interval to obtain the current capacity retention rate, calculate the decay amount = current capacity retention rate - retirement threshold, and then combine it with the fitted decay rate, i.e. the slope of the fitted equation, to obtain the predicted remaining lifetime RUL = decay amount / slope of the fitted equation.

[0114] Calculate the deviation rate between the predicted value and the measured value. If the deviation rate is greater than the threshold, readjust D1 and D2.

[0115] For clusters in the intermediate stage, the specific steps for constructing an adaptive kernel function RBF neural network are as follows:

[0116] Collect data on the voltage plateau rate of voltage drop, temperature fluctuation amplitude, and remaining lifetime of the target battery during its mid-term operation.

[0117] The samples were divided into three groups according to the temperature fluctuation range: small fluctuation (temperature fluctuation range ≤ T1), medium fluctuation (T1 < temperature fluctuation range ≤ T2), and large fluctuation (temperature fluctuation range > T2). T1 and T2 were set according to the operating temperature range of the target battery.

[0118] Each sample group corresponds to a voltage plateau rate of drop sequence, with a length denoted as m;

[0119] Determine the RBF neural network structure: the number of neurons in the input layer is the length m of the voltage drop rate sequence, the number of neurons in the hidden layer is h (determined by k-fold cross-validation, usually 5 or 10), and the number of neurons in the output layer is 1 (output remaining lifetime).

[0120] The kernel function used is the Gaussian kernel function, and the specific formula is as follows: Where c is the average pressure drop rate of each group, and σ is dynamically adjusted with temperature fluctuations: small fluctuations correspond to σ1, medium fluctuations correspond to σ2, and large fluctuations correspond to σ3, where σ3 < σ2 < σ1;

[0121] Using the grouped pressure drop rate sequence as input and remaining lifetime as output, the loss function is minimized using gradient descent. If the MSE of all groups is less than or equal to the threshold MSE0, it indicates that all prediction models have reached the fitting accuracy.

[0122] Test the model with several new sets of mid-term battery data (covering small, medium and large temperature fluctuations): Input a set of voltage drop rate sequences, compare the deviation between the model's predicted remaining lifetime and the measured value. If the deviation of a certain fluctuation set is > ΔL (ΔL is the preset upper limit of lifetime deviation), then adjust the σ value or the number of hidden layer neurons h of that set, and retrain the model until the deviation is ≤ ΔL.

[0123] For clusters in the final stage, the specific steps for constructing the grey prediction GM(1,1) model are as follows:

[0124] Collect data on the instantaneous increase in internal resistance and the range of SOC fluctuation at the end of the target battery period, and remove samples with an instantaneous increase in internal resistance < 0 (no negative growth in internal resistance) and a range of SOC fluctuation < ΔSOC0 (ΔSOC0 is the minimum fluctuation range allowed by the charging and discharging strategy).

[0125] Construct a non-negative original sequence X(0)={x(0)(1),x(0)(2),...,x(0)(n)} of instantaneous internal resistance increments, where x(0)(i) is the instantaneous internal resistance increment of the i-th record;

[0126] The original sequence X(0) is accumulated once to generate the sequence X(1)(k)=sum(X(0)(i)), which reduces the randomness of the original data, where i∈[1,n];

[0127] Combining the two, we establish the gray differential equation: X(0)(k)+a×X(1)(k)=b, where a is the development coefficient (reflecting the growth trend of internal resistance) and b is the gray action quantity (reflecting the driving intensity).

[0128] Solve for the parameter vector [a, b] using the least squares method. T The prediction equation for the cumulative sequence is obtained as follows: ;

[0129] The predicted value of the original sequence is obtained by cumulative subtraction: That is, the predicted value of the instantaneous increment of internal resistance in the (k+1)th iteration;

[0130] The cumulative internal resistance value Rpred is obtained by successively accumulating the predicted instantaneous increment of internal resistance. The number of cycles when Rpred first ≥ rth is found, and the difference between Rpred and the current number of cycles is calculated to obtain the preliminary remaining lifetime L. At the same time, Rpred / rth is calculated.

[0131] If (Rpred / rth)≤ If L does not need to be corrected; If (Rpred / rth) ≤ 1, the remaining lifespan needs to be adjusted to obtain L × ε; if (Rpred / rth) > 1, the battery needs to be immediately deactivated, and L should be set to 0; where, To approximate the threshold, ε is a correction coefficient, both of which need to be adjusted according to the actual situation.

[0132] Step 3: Real-time data collection of SOH, capacity decay acceleration, and internal resistance growth rate via the Battery Management System (BMS) to form a real-time feature vector. The weighted Mahalanobis distance between this vector and the cluster centers of Q1, Q2, and Q3 is calculated. Based on this distance, stage judgments and prediction model calls are performed. The specific operation process is as follows:

[0133] Extract the feature data of the cluster centers in the three stages Q1, Q2, and Q3, denoted as centers C1 (early stage), C2 (middle stage), and C3 (late stage). Each center contains three features: SOH (saturation loss), denoted as f1; capacity decay acceleration, denoted as f2; and internal resistance growth rate, denoted as f3, forming a center feature matrix.

[0134] ;

[0135] The stage discrimination variance of each feature among the three centers is calculated using the following formula: The smaller the variance, the more significant the differences in features at different stages, and the higher the discriminative power. Let j be the mean of this feature at the three centers, j∈[1,3];

[0136] Calculate the weights according to the principle that the higher the discrimination, the greater the weight. The weight matrix W = diag(wf1, wf2, wf3) is obtained.

[0137] The real-time feature vector Y = [Y(f1), Y(f2), Y(f3)] is collected by BMS, where Y(f1) is the SOH after Kalman filtering correction, Y(f2) is the average capacity decay acceleration of the most recent 100 cycles, and Y(f3) is the average internal resistance growth rate of the day.

[0138] First, calculate the covariance matrix of the central features of the three stages. Combine it with W to construct a weighted covariance matrix. Calculate the weighted Mahalanobis distance between the real-time feature vector Y and each center, using the following formula: We obtained D(Y,C1) (distance to the initial center), D(Y,C2) (distance to the middle center), and D(Y,C3) (distance to the final center), respectively. It is the inverse of the weighted covariance matrix;

[0139] The minimum value among the three distances is selected, and the corresponding center stage is selected as the candidate stage. The difference between the minimum distance and the second smallest distance, ΔDis, is calculated. If ΔDis > ΔDth, the candidate stage is the final stage. If ΔDis ≤ ΔDth, the stage at the previous moment is maintained. ΔDth is a preset threshold that can be determined based on historical data statistics, and its value range is [0.1, 0.2].

[0140] The specific rules for invoking the prediction model are as follows:

[0141] If the final stage is the initial stage, then the trained piecewise linear regression model is used to predict the remaining lifespan.

[0142] If the final stage is the intermediate stage, then the trained adaptive kernel function RBF neural network model is used to predict the remaining lifetime.

[0143] If the final stage is the end stage, then the trained grey prediction GM(1,1) model is called to predict the remaining lifetime.

[0144] Based on the predicted Remaining Life (RUL) at different stages, different adjustment strategies for the battery are set, as detailed below:

[0145] When the battery is in the initial stage, if RUL>R1, the depth of charge and discharge is relaxed by 5% compared with the normal threshold (the standard depth of charge and discharge for the same model battery in the initial stage), the charging cut-off voltage is maintained at the normal value U1 in the initial stage, and the discharging cut-off voltage is maintained at the normal value U2 in the initial stage.

[0146] If R2 < RUL ≤ R1, the depth of charge and discharge is tightened by 10% compared to the conventional threshold. The charging cutoff voltage is still U1, but the discharging cutoff voltage U2 needs to be increased.

[0147] If RUL≤R2, the depth of charge and discharge is tightened by 15% compared to the conventional threshold, and the charging cut-off voltage needs to be lowered compared to U1 to prevent the SEI film from entering the rupture stage prematurely due to deep cycling. Among them, R1 is the initial long life threshold, and R2 is the initial medium life threshold.

[0148] When the battery is in the middle stage, if RUL>R3 and △T≤T0, the charging rate should be maintained at the normal value C1 for the middle stage, and the discharge cut-off voltage should be increased compared to the normal value U3 for the middle stage to reduce polarization accumulation.

[0149] If RUL > R3 and ΔT > T0, then the charging rate needs to be lower than C1, while the discharge cut-off voltage remains at the increased value.

[0150] If RUL≤R3, regardless of the size of ΔT, the charging rate needs to be lower than C1 and the discharge cutoff voltage needs to be higher than U3. Here, R3 is the medium-term long life threshold, T0 is the temperature stability threshold, and ΔT is the temperature fluctuation value collected by BMS in real time.

[0151] When the battery is in its final stage, if R4 < RUL ≤ R5, the charging cut-off voltage needs to be lower than the normal value U4 in the final stage, the maximum discharge power needs to be lower than the rated power, and the SOC operating range needs to be narrowed by 10% compared with the normal range (the upper limit and the lower limit are each narrowed by 5%).

[0152] If RUL≤R4, the charging cutoff voltage needs to be lower than U4, the maximum discharge power needs to be lower than the rated power, and the SOC operating range needs to be narrowed by 20% compared to the normal range (the upper limit and the lower limit are each narrowed by 10%). Among them, R4 is the end-of-life safety threshold and R5 is the end-of-life critical threshold.

[0153] Example 2

[0154] This embodiment, based on Embodiment 1, further discloses a method for splitting existing clusters based on local minima of gravitational potential energy, such as... Figure 2 As shown, the specific content includes:

[0155] Determine the sample set S={s1,s2,...,sn} of the cluster to be split (n is the number of samples in the cluster), and calculate the potential energy values ​​{E1,E2,...,En} of all samples. The larger the potential energy value, the stronger the correlation between the sample and other samples in the cluster (the more concentrated the attraction of the surrounding samples to it).

[0156] The samples within a cluster are sorted in descending order according to their SoH (Solar Oxygen Detection) values. Since SoH is the core characteristic of the aging stage, the aging degree of adjacent samples is more similar after sorting, resulting in an ordered sample sequence Sord={sq1,sq2,...,sqn}.

[0157] For each sample sqk (2≤k≤n-1) in the ordered sample sequence Sord, excluding the first and last samples, if Eqk<Eq(k-1) and Eqk<Eq(k+1), that is, the potential energy of the sample is less than the potential energy of the previous and next adjacent samples, then sqk is a local minimum point. This point is a weak point in the cluster correlation, and the attraction between the sample and the previous and next adjacent samples is weak, which is suitable as a split boundary.

[0158] If there are multiple local minima, the point with the lowest potential energy is selected as the final split point to ensure that the internal correlation between the two subclusters is the strongest after the split.

[0159] Using the finally located local minimum point sqk as the boundary, the ordered sample sequence Sord is split into two parts: sub-cluster S1={sq1,...,sq(k-1)} (high SOH end, less aged) and sub-cluster S2={sq(k+1),...,sqn} (low SOH end, more aged). The minimum point sqk is removed (if its potential energy is too low, it is judged as a noise sample within the cluster).

[0160] Calculate the CV of the two subclusters separately. SOH CV ron If both CV values ​​of the two subclusters are not greater than the preset threshold, then the split is valid;

[0161] If a sub-cluster still does not meet the CV threshold, repeat the above splitting steps until the feature fluctuations of all sub-clusters meet the requirements.

[0162] Example 3

[0163] This embodiment, based on Embodiments 1 and 2, further discloses a method for verifying cluster purity of the GS-ADC model, such as... Figure 3 As shown, the specific process is as follows:

[0164] Ten groups of batteries with the same model as those used in the energy storage power station to be optimized were selected. Data for each group of batteries was obtained through a "full life cycle aging test": starting from new batteries (SOH=100%), data on SOH, capacity decay acceleration, and internal resistance growth rate were recorded every 100 cycles until the batteries were scrapped (SOH≤40%). Each group of batteries was labeled with stage labels for the initial stage (SOH≥80%), the middle stage (60%≤SOH<80%), and the final stage (SOH<60%), thus forming 10 validation subsets. Each subset contained 50 samples from the initial stage, 50 samples from the middle stage, and 50 samples from the final stage, for a total of 1500 labeled samples in the validation set.

[0165] Substituting 1500 labeled validation samples into the GS-ADC model with parameters completely consistent with the actual clustering, we obtained 3 validation clusters, denoted as V1, V2, and V3.

[0166] The number of labeled samples belonging to the initial, middle and final stages in each validation cluster is counted. The cluster purity is calculated according to the formula: (number of samples in the stage with the largest number in the cluster / total number of samples in the cluster). The stage with the largest number in the cluster is taken as the stage of the cluster. Clusters with purity <90% are selected as the target clusters for this parameter adjustment.

[0167] For the target cluster, calculate the gravitational potential energy difference between its core sample and misclustered samples. If the difference is less than 0.2 × average potential energy difference (the average potential energy difference is the average potential energy difference between the core sample and samples in the same cluster), it is determined that the gravitational discrimination is insufficient. In this case, α needs to be adjusted. The specific adjustment rule is as follows: set the initial α to 0.5, and increase it stepwise in increments of 0.1 each time. The maximum adjusted α should not exceed 1.0. After each adjustment, the purity of the target cluster needs to be recalculated. If it is ≥90%, stop adjusting α. If the purity is still <90% after α is increased to 1.0, adjust the core sample density threshold. The specific operation is as follows:

[0168] For the core samples of the target cluster, extract their original adaptive density threshold. ;

[0169] In the core statistical sample, for samples where the proportion of samples from different stages in the neighborhood is greater than 20%, calculate the percentage of these samples. Average threshold of core samples within the cluster ,calculate The ratio of the sample to the total sample is used to determine whether the core sample is not representative enough due to a low adaptive threshold. If the ratio is ≥80%, it is determined that the dynamic coefficient in the threshold formula needs to be adjusted.

[0170] The original dynamic coefficient is 0.3 (0.3 in the formula), and it is increased in a stepwise manner with an increment of 0.1 each time, with a maximum of no more than 0.7.

[0171] The adaptive density threshold for all samples was recalculated based on the new coefficients. The core samples are re-screened, and the purity of the target cluster is calculated. If the purity is ≥90%, the current dynamic coefficient and the corresponding adaptive threshold formula are locked.

[0172] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0173] The above embodiments are only used to illustrate the technical methods 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 methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for improving and optimizing the operation efficiency of an energy storage power station based on big data analysis, characterized in that, include: Step 1: Collect full life cycle data of the same type of battery, and cluster them based on the Gravity Search-Adaptive Density Clustering (GS-ADC) algorithm to obtain three final clusters Q1, Q2, and Q3. Then determine the stage of the final clusters, which are the initial stage, the middle stage, and the final stage. Step 2: For clusters in the initial, middle and final stages, construct piecewise linear regression, adaptive kernel function RBF neural network and grey prediction GM(1,1) model respectively to predict the remaining life of the battery in each stage. Step 3: Real-time data collection of SOH, capacity decay acceleration, and internal resistance growth rate is performed through the Battery Management System (BMS) to form a real-time feature vector. The weighted Mahalanobis distance between the vector and the cluster centers of Q1, Q2, and Q3 is calculated. Based on the distance, stage judgment and prediction model are called. Based on the predicted remaining life (RUL) of different stages, different adjustment strategies for the battery are set.

2. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 1, characterized in that, The specific steps for clustering the entire lifecycle data of batteries of the same model based on the Gravity Search-Adaptive Density Clustering (GS-ADC) algorithm are as follows: The SOH value, capacity decay acceleration, and internal resistance growth rate were used as the three-dimensional clustering features, and the obtained three-dimensional features were min-max normalized and mapped to the [0,1] interval. The sample potential is defined by the following formula: , where w j Let d be the feature weights of sample j. ij Let be the Euclidean distance between samples i and j, and α be the attenuation coefficient; Determine the density threshold for each sample. ,in, Let be the average density of the 10 nearest samples around sample i, and max(E) be the maximum gravitational potential energy of all samples. Calculate the local density for each sample ,filter And E i Samples with a ≥0.6×max(E) value are used as the initial core samples, where d th Distance threshold; For a non-core sample i, calculate the gravitational interaction strength between it and each initial core sample k. The samples are assigned to the initial core sample cluster with the strongest gravitational influence, forming an initial cluster, where E k Let d be the gravitational potential energy of the initial core sample k. ik Let be the Euclidean distance between non-core sample i and the initial core sample k; The intra-cluster characteristic fluctuation coefficient CV of the initial cluster is calculated according to the formula CV=std(intra-cluster characteristic value) / mean(intra-cluster characteristic value) SOH , CV ron , if the CV of a cluster SOH > CV th or CV ron > CV th1 , the cluster is split according to the local minimum of the gravitational potential energy, wherein std() is a standard deviation function, mean() is a mean function, CV SOH is a SOH value fluctuation coefficient, CV ron is a capacity attenuation acceleration fluctuation coefficient, CV th , CV th1 is a fluctuation coefficient threshold value; If the distance between the centers of two clusters is < 0.2*max(d all ), the two clusters are merged into one cluster, and finally three clusters are reserved, wherein max(d all ) is the maximum distance of all cluster centers. Calculate the intra-cluster distance percentage for each sample If r i >r th If it is an isolated point, it is classified as an isolated point and assigned to the cluster with the nearest distance and whose CV meets the requirements, where d cu Let d be the set of distances between a single sample and all other samples in the cluster. i,center Let be the Euclidean distance from sample i to the cluster center of its cluster; Cluster purity is verified based on the existing GS-ADC model. After the verification is successful, the collected full life cycle data of the same type of battery needs to be clustered again based on the GS-ADC model with modified parameters to obtain three final clusters Q1, Q2 and Q3.

3. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 2, characterized in that, The specific process for splitting the cluster based on the local minimum of gravitational potential energy is as follows: Determine the sample set S={s1,s2,...,sn} of the cluster to be split, and calculate the potential energy values ​​{E1,E2,...,En} of all samples, where n is the number of samples in the cluster; The samples within a cluster are sorted in descending order according to their Soh values ​​to obtain an ordered sample sequence Sord={sq1,sq2,...,sqn}; For each sample sqk in the ordered sample sequence Sord, excluding the first and last samples, if Eqk < Eq(k-1) and Eqk < Eq(k+1), that is, the potential energy of the sample is less than the potential energy of the previous and next adjacent samples, then sqk is a local minimum point. This point is used as the split boundary, where 2 ≤ k ≤ n-1. If there are multiple local minima, the point with the lowest potential energy is selected as the final split point. Using the finally located local minimum point sqk as the boundary, the ordered sample sequence Sord is split into two parts: sub-cluster S1={sq1,...,sq(k-1)} and sub-cluster S2={sq(k+1),...,sqn}, and the local minimum point sqk is removed; Calculate the CV of two sub-clusters respectively SOH , CV ron , if the two CV values of two sub-clusters are not greater than the preset threshold, the splitting is effective; If a sub-cluster still does not meet the CV threshold, repeat the above splitting steps until the feature fluctuations of all sub-clusters meet the requirements.

4. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 1, characterized in that, The specific steps for verifying cluster purity based on the existing GS-ADC model are as follows: Ten groups of batteries with the same model as the batteries used in the energy storage power station to be optimized were selected. Data for each group of batteries was obtained according to the "full life cycle aging test": starting from the new battery, data on SOH, capacity decay acceleration, and internal resistance growth rate were recorded every 100 cycles until the battery was scrapped. Each group of batteries is labeled with stage tags for the initial, middle and final stages, ultimately forming 10 validation subsets, each containing 50 samples from the initial stage, 50 samples from the middle stage, and 50 samples from the final stage, for a total of 1500 labeled samples in the validation set. Substituting 1500 labeled validation samples into the GS-ADC model with parameters completely consistent with the actual clustering, we obtained 3 validation clusters, denoted as V1, V2, and V3. The number of labeled samples belonging to the initial, middle and final stages in each validation cluster is counted. The cluster purity is calculated according to the formula: (number of samples in the stage with the largest number in the cluster / total number of samples in the cluster). The stage with the largest number in the cluster is taken as the stage of the cluster. Clusters with purity <90% are selected as the target clusters for this parameter adjustment. For the target cluster, calculate the gravitational potential energy difference between its core sample and misclustered sample. If the difference is less than 0.2 × the average potential energy difference, then α needs to be adjusted. If the purity is still less than 90% after increasing α to 1.0, then adjust the core sample density threshold. The average potential energy difference is the average potential energy difference between the core sample and the samples in the same cluster.

5. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 4, characterized in that, The specific rules for adjusting α are as follows: Set the initial α to 0.5, and increase it stepwise in increments of 0.1 each time. The maximum adjusted α should not exceed 1.

0. After each adjustment, the purity of the target cluster needs to be recalculated. If it is ≥90%, stop adjusting α. Adjusting the core sample density threshold The specific rules are as follows: For the core samples of the target cluster, extract their original adaptive density threshold. ,in, Let E be the average density of the 10 nearest samples around sample i, and max(E) be the maximum gravitational potential energy of all samples. i Let be the gravitational potential energy of sample i; In the core statistical sample, for samples where the proportion of samples from different stages in the neighborhood is greater than 20%, calculate the percentage of these samples. Average threshold of core samples within the cluster ,calculate The ratio of the sample to the total sample is used. If the ratio is ≥80%, the adaptive threshold is judged to be too low. The original dynamic coefficient of 0.3 in the threshold formula is increased stepwise by "+0.1 each time", with a maximum of 0.

7. The adaptive density threshold for all samples was recalculated based on the new coefficients. The core samples are re-screened, and the purity of the target cluster is calculated. If the purity is ≥90%, the current dynamic coefficient and the corresponding adaptive threshold formula are locked.

6. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 1, characterized in that, After obtaining the final clusters Q1, Q2, and Q3, it is necessary to determine the stage of each cluster, namely the initial stage, the middle stage, and the final stage: calculate the center distance dis(i,j) between cluster Qi and cluster Vj respectively. For Qi, find the Vj corresponding to min(dis(i,j)) and take the stage of Vj as the stage corresponding to Qi, where i and j ∈ [1,3].

7. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 1, characterized in that, The specific steps for constructing a piecewise linear regression model are as follows: Collect the DOD value, cycle number, and capacity retention rate of the target battery during the initial charge-discharge cycle. Two types of abnormal samples were removed: the fluctuation of the DOD value in a single charge / discharge cycle compared to the command value was greater than ΔD; and the fluctuation of the capacity retention rate compared to the previous cycle was greater than ΔR. Here, ΔD is the upper limit of the allowable DOD fluctuation of the charge / discharge system, and ΔR is the upper limit of the accuracy of the detection equipment. Two DOD thresholds, D1 and D2, are set, and the valid data is divided into three intervals: if the DOD of a single charge-discharge cycle is less than or equal to D1, it is classified as the low DOD interval; if D1 is less than the DOD of a single charge-discharge cycle and less than or equal to D2, it is classified as the medium DOD interval; if the DOD of a single charge-discharge cycle is greater than or equal to D2, it is classified as the high DOD interval; where D1 is less than D2. For each DOD interval, the cumulative number of cycles for the samples within that interval is used as the independent variable, and the capacity retention rate after the corresponding cycle is used as the dependent variable. The least squares method is used to fit the linear equation, resulting in three fitting curves for the "number of cycles - capacity retention rate" of different DOD intervals. Substitute the current cycle number into the fitting equation of the corresponding DOD interval to obtain the current capacity retention rate, calculate the decay amount = current capacity retention rate - retirement threshold, and then obtain the predicted remaining lifetime RUL = decay amount / fitting equation slope through the fitted decay rate, i.e. the slope of the fitted equation. Calculate the deviation rate between the predicted value and the measured value. If the deviation rate is greater than the threshold, readjust D1 and D2.

8. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 1, characterized in that, The specific steps for constructing an adaptive kernel function RBF neural network are as follows: Collect data on the voltage plateau rate of voltage drop, temperature fluctuation amplitude, and remaining lifetime of the target battery during its mid-term operation. The samples are divided into three groups according to the temperature fluctuation range: if the temperature fluctuation range is ≤ T1, it is classified as small fluctuation; if T1 < temperature fluctuation range ≤ T2, it is classified as medium fluctuation; if the temperature fluctuation range is > T2, it is classified as large fluctuation; where T1 and T2 are the lower limit and upper limit of the fluctuation range, respectively. Each sample group corresponds to a voltage plateau rate of drop sequence, with a length denoted as m; The RBF neural network structure is determined as follows: the number of neurons in the input layer is the length m of the voltage drop rate sequence, the number of neurons in the hidden layer is h, and the number of neurons in the output layer is 1. The kernel function used is the Gaussian kernel function, and the specific formula is as follows: Where x is the voltage plateau rate of drop sequence, c is the mean of each group of rate of drop, and σ is dynamically adjusted with temperature fluctuations: small fluctuations correspond to σ1, medium fluctuations correspond to σ2, and large fluctuations correspond to σ3, where σ3 < σ2 < σ1. Using the grouped pressure drop rate sequence as input and remaining lifetime as output, the loss function is minimized using gradient descent. If the MSE of all groups is less than or equal to the threshold MSE0, it indicates that all trained models have reached the fitting accuracy. Test the trained model with several sets of new mid-term battery data: Input a set of voltage drop rate sequences and compare the deviation between the model's predicted remaining lifetime and the measured value. If the deviation of a certain fluctuation group is greater than ΔL, adjust the σ value or the number of hidden layer neurons h of that group and retrain the model until the deviation is less than or equal to ΔL, where ΔL is the preset upper limit of lifetime deviation.

9. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 1, characterized in that, The specific steps for constructing the grey prediction GM(1,1) model are as follows: Collect data on the instantaneous increase in internal resistance and the range of SOC fluctuation at the end of the target battery period, and remove samples with an instantaneous increase in internal resistance < 0 and a range of SOC fluctuation < ΔSOC0, where ΔSOC0 is the minimum fluctuation range allowed by the charging and discharging strategy. Construct a non-negative original sequence X(0)={x(0)(1),x(0)(2),...,x(0)(n)} of instantaneous internal resistance increments, where x(0)(i) is the instantaneous internal resistance increment of the i-th record; The original sequence X(0) is summed once to generate the sequence X(1)(k)=sum(X(0)(i)), where sum() is the summation function and i∈[1,n]. Establish the gray differential equation: X(0)(k)+a×X(1)(k)=b, where a is the development coefficient and b is the gray action quantity; Solve for the parameter vector [a, b] using the least squares method. T The prediction equation for the cumulative sequence is obtained as follows: ; The predicted value of the original sequence is obtained by cumulative subtraction: That is, the predicted value of the instantaneous increment of internal resistance in the (k+1)th iteration; The cumulative internal resistance value Rpred is obtained by successively accumulating the predicted instantaneous increment of internal resistance. The number of cycles when Rpred first ≥ rth is found, and the difference between Rpred and the current number of cycles is calculated to obtain the preliminary remaining lifetime L. At the same time, Rpred / rth is calculated. If (Rpred / rth)≤ɳ, then L does not need to be corrected; if ɳ<(Rpred / rth)≤1, then the remaining lifespan needs to be corrected to obtain L×ε; if (Rpred / rth)>1, the battery needs to be stopped immediately and L is set to 0; where ɳ is the threshold and ε is the correction coefficient.

10. The method for improving and optimizing the operating efficiency of energy storage power stations based on big data analysis according to claim 1, characterized in that, The specific operational procedure for determining the stage based on distance is as follows: Extract the cluster centers of clusters Q1, Q2, and Q3, denoted as C1 (early stage), C2 (middle stage), and C3 (late stage). Each center contains three features: SOH (saturation loss), denoted as f1; capacity decay acceleration, denoted as f2; and internal resistance growth rate, denoted as f3, forming a center feature matrix. ； The stage discrimination variance of each feature among the three centers is calculated using the following formula: ,in, Ci(fj) is the mean of the feature across the three centers, and Ci(fj) is the specific feature value of the i-th cluster center on the j-th feature dimension, where j∈[1,3]. Calculate weights The weight matrix W = diag(wf1, wf2, wf3) is obtained. The real-time feature vector Y = [Y(f1), Y(f2), Y(f3)] is collected by BMS, where Y(f1) is the SOH after Kalman filtering correction, Y(f2) is the average capacity decay acceleration of the most recent 100 cycles, and Y(f3) is the average internal resistance growth rate of the day. First, calculate the covariance matrix of the central features of the three stages. Construct a weighted covariance matrix Calculate the weighted Mahalanobis distance between the real-time feature vector Y and each center, using the following formula: We obtain D(Y,C1), D(Y,C2), and D(Y,C3) respectively, where, Let be the inverse of the weighted covariance matrix, and let Ci be the cluster center of the i-th cluster; Select the minimum value among the three distances and take the corresponding center stage as the candidate stage. Calculate the difference ΔDis between the minimum distance and the second smallest distance. If ΔDis > ΔDth, the candidate stage is the final stage. If ΔDis ≤ ΔDth, the stage at the previous moment is maintained. ΔDth is a preset threshold.

Images