Method for predicting health state of battery cell and health state of battery module and electronic equipment

Abstract

The invention provides a method for predicting the health state of a battery cell and the health state of a battery module and electronic equipment. The cell health state prediction method comprises the following steps: obtaining health state SOH time sequence data and external excitation time sequence data of a to-be-detected cell in a preset time period; splitting the SOH time sequence data to obtain trend term time sequence data in the SOH time sequence data; based on a preset ND-ARIMA model in which nonlinear attenuation is introduced, according to the trend term time sequence data, obtaining a trend term prediction value of the to-be-tested cell at the next moment; on the basis of a preset volatility prediction model, according to the external excitation time sequence data, obtaining a fluctuation item prediction value of the to-be-tested cell at the next moment; and according to the trend item prediction value and the fluctuation item prediction value at the next moment, determining the SOH prediction value of the to-be-detected cell at the next moment. According to the scheme, the calculation efficiency is high, the accuracy of cell SOH prediction can be improved, and then the accuracy of cell module SOH prediction is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of batteries, in particular to a method for predicting the state of health of a battery cell and a battery module and an electronic device. BACKGROUND

[0002] Energy storage power stations have high capacity and energy density, and generally establish multi-level power storage units such as modules, cabinets and clusters based on battery cells. The smallest repairable unit is usually a battery module. Therefore, accurately evaluating the state of health of a battery module is crucial for the safe operation, fault diagnosis and prediction, and operation and maintenance efficiency optimization of an energy storage power station. SUMMARY

[0003] To solve the above problems, the present application provides a method for predicting the state of health of a battery cell and a battery module and an electronic device.

[0004] According to a first aspect of the present application, a method for predicting the state of health of a battery cell is provided, comprising:

[0005] obtaining state of health (SOH) time series data and external excitation time series data of a battery cell to be tested within a preset time period;

[0006] splitting the SOH time series data to obtain trend item time series data in the SOH time series data;

[0007] based on a preset ND-ARIMA model introducing non-linear decay, obtaining a trend item prediction value of the battery cell to be tested at the next time point according to the trend item time series data; wherein the ND-ARIMA model introducing non-linear decay is constructed based on the SOH time series data of the battery cell;

[0008] based on a preset volatility prediction model, obtaining a fluctuation item prediction value of the battery cell to be tested at the next time point according to the external excitation time series data; wherein the fluctuation item prediction model has learned the mapping relationship between the external excitation time series data and the fluctuation item at the next time point corresponding thereto;

[0009] determining the SOH prediction value of the battery cell to be tested at the next time point according to the trend item prediction value and the fluctuation item prediction value.

[0010] According to the technical solution of the present application, the non-linear decay trend of the SOH of the battery cell to be tested is predicted by introducing the ND-ARIMA model introducing non-linear decay, and the fluctuations occurring in the SOH decay process of the battery cell to be tested are predicted by the volatility prediction model, and the SOH of the battery cell to be tested is determined based on the predicted decay trend and fluctuations. Not only is the calculation efficiency high, but also the accuracy of the prediction of the SOH of the battery cell can be improved, and the accuracy of the prediction of the SOH of the battery module can be improved.

[0011] According to a second aspect of this application, a method for predicting the health status of a battery module is provided, comprising:

[0012] Based on the cell health status prediction method in the first aspect mentioned above, the predicted SOH value of each cell in the battery module under test at the next moment is determined.

[0013] Based on the predicted SOH value of each cell at the next moment, the predicted SOH value of the battery module under test at the next moment is determined.

[0014] According to the technical solution of this application, the nonlinear decay trend of the SOH of the cell under test is predicted by introducing the nonlinear decay ND-ARIMA model, and the fluctuation prediction model is used to predict the fluctuations that occur during the SOH decay process of the cell under test. Based on the predicted decay trend and fluctuations, the SOH of each cell in the battery module under test is determined. This not only has high computational efficiency, but also improves the accuracy of cell SOH prediction, thereby improving the accuracy of battery module SOH prediction.

[0015] According to a third aspect of this application, a device for predicting the health status of a battery cell is provided, comprising:

[0016] The first acquisition module is used to acquire the health status (SOH) timing data and external excitation timing data of the battery cell under test within a preset time period.

[0017] The splitting module is used to split the SOH time series data and obtain the trend item time series data from the SOH time series data;

[0018] The first prediction module is used to obtain the predicted value of the trend term of the cell under test at the next moment based on the trend term time series data of the preset ND-ARIMA model with nonlinear decay. The ND-ARIMA model with nonlinear decay is constructed based on the cell's SOH time series data.

[0019] The second prediction module is used to obtain the predicted value of the fluctuation term of the battery cell under test at the next moment based on the external excitation time series data and the preset fluctuation prediction model. The fluctuation prediction model has learned the mapping relationship between the external excitation time series data and the corresponding fluctuation term at the next moment.

[0020] The first determining module is used to determine the SOH prediction value of the cell under test at the next moment based on the predicted value of the trend term and the predicted value of the fluctuation term at the next moment.

[0021] According to a fourth aspect of this application, a device for predicting the health status of a battery module is provided, comprising:

[0022] The third prediction module is used to determine the predicted SOH value of each cell in the battery module under test at the next moment based on the cell health state prediction method of the first aspect mentioned above.

[0023] The second determining module is used to determine the predicted SOH value of the battery module under test at the next time step based on the predicted SOH value of each cell at the next time step.

[0024] According to a fifth aspect of this application, an electronic device is provided, comprising: a processor; a memory for storing executable instructions of the processor; wherein the processor is configured to execute the instructions to implement a cell health state prediction method as described in the first aspect above, and / or to implement a battery module health state prediction method as described in the second aspect above.

[0025] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0026] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0027] Figure 1 A flowchart illustrating a method for predicting the health status of a battery cell provided in an embodiment of this application;

[0028] Figure 2 A flowchart illustrating another method for predicting the health status of a battery cell provided in an embodiment of this application;

[0029] Figure 3 A flowchart illustrating yet another method for predicting the health status of a battery cell provided in an embodiment of this application;

[0030] Figure 4 A flowchart illustrating a method for predicting the health status of a battery module provided in an embodiment of this application;

[0031] Figure 5 A flowchart illustrating another method for predicting the health status of a battery module provided in an embodiment of this application;

[0032] Figure 6 A structural block diagram of a cell health status prediction device provided in an embodiment of this application;

[0033] Figure 7 A structural block diagram of a battery module health status prediction device provided in an embodiment of this application;

[0034] Figure 8 This is a structural block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0035] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0036] It should be noted that energy storage power stations have high capacity and energy density, and are generally built with multi-level energy storage units based on battery cells, including modules, cabinets, and clusters. The smallest repairable unit is usually the battery module. Therefore, accurately assessing the state of health (SOH) of battery modules is crucial for the safe operation, fault diagnosis and prediction, and optimization of operation and maintenance efficiency of energy storage power stations. In related technologies, the SOH prediction methods for battery modules mainly include the following two approaches:

[0037] (1) Prediction methods based on deep learning require a large amount of data, which makes it impossible to guarantee the accuracy of the model when there is little historical data. Furthermore, due to the large number of factors affecting the battery's SOH and its own decay trend, the model is often too complex and has low computational efficiency.

[0038] (2) Autoregressive prediction methods are computationally efficient and mainly used to predict SOH decay trends. However, traditional autoregression methods struggle to accurately capture the non-linear SOH decay trend and fluctuations that occur during the decay process. Traditional battery module health state calculations are primarily based on the bucket principle, using the lowest SOH of each cell in the battery module to calibrate the module's health state. While this method is simple in principle and highly efficient, it is limited in scope and cannot capture health state degradation caused by inconsistencies within the module.

[0039] To address the aforementioned issues, this application provides a method and electronic device for predicting the health status of battery cells and battery modules.

[0040] Figure 1 This is a flowchart illustrating a method for predicting the health status of a battery cell, as provided in an embodiment of this application. It should be noted that the method for predicting the health status of a battery cell in this application can be applied to the device for predicting the health status of a battery cell in this application, and this device can be configured in an electronic device. Figure 1 As shown, the method may include the following steps:

[0041] Step 101: Obtain the health status (SOH) timing data and external excitation timing data of the cell under test within a preset time period.

[0042] The cell under test can be any cell in the battery module under test. The preset time period includes the first moment corresponding to the running of the method and a period of time before the first moment, and the length of the preset time period is preset. The health status SOH time series data within the preset time period includes the SOH data at each sampling moment within the preset time period, and the external excitation time series data within the preset time period includes the external excitation data at each sampling moment within the preset time period. It should be noted that the step size between each sampling moment in the health status SOH time series data and the external excitation time series data within the preset time period is consistent with the step size set in the model in steps 103 and 104. The first moment can be the current moment when the method is running, or it can be the moment corresponding to the running of the method after being divided according to the aforementioned step size. For example, if the step size is 1 hour and the running time of the method is 8:20, one possibility is that the first moment is 8:20, and another possibility is that it is divided according to the set step size, predicting the SOH of the cell under test at 0:00, 1:00, 2:00 and so on each day. In this case, the first moment can be 8:00.

[0043] In some embodiments, the SOH data at each sampling time in the SOH time series data can be an index value characterizing the health status of the cell determined based on actual needs. For example, it can be the ratio of the measurable battery capacity to the rated capacity of the cell under test at the corresponding time.

[0044] In some embodiments, the external excitation data at each sampling moment in the external excitation timing data may include the battery module temperature corresponding to the cell under test, the duration of overcharging and over-discharging of the cell under test, the duration of high-rate charging and discharging of the cell under test, etc., wherein overcharging and over-discharging and high-rate charging and discharging can be defined based on actual needs.

[0045] Step 102: Split the SOH time series data to obtain the trend item time series data from the SOH time series data.

[0046] In some embodiments of this application, the SOH time-series data of the battery cell under test can be decomposed using singular spectrum analysis to obtain trend term time-series data and volatility trend data from the SOH time-series data. The specific process of singular spectrum decomposition can be implemented based on singular spectrum decomposition methods in related technologies, and this application does not limit it in this regard.

[0047] Step 103: Based on the preset ND-ARIMA model with nonlinear decay, obtain the predicted value of the trend term of the cell under test at the next moment according to the trend term time series data; wherein, the ND-ARIMA model with nonlinear decay is constructed based on the cell SOH time series data.

[0048] The ARIMA (Autoregressive Integrated Moving Average) model attempts to extract hidden time-series patterns from data through autocorrelation and differencing, and then uses these patterns to predict future data. Specifically: the AR (Autoregressive) component handles the autoregressive part of the time series, considering the influence of past observations on the current value; the I (Internal) component helps non-stationary time series become stationary by eliminating trends and seasonality through first- or second-order equal differencing; and the MA (Moving Average) component handles the moving average part of the time series, considering the impact of past prediction errors on the current value. Combining these three components, the ARIMA model can capture both trend changes and handle temporary, sudden changes, or noisy data. Therefore, the ARIMA model performs well in many time series forecasting problems. However, the ARIMA model struggles to capture the non-linear decay trend of the Solid State of Interest (SOH) and the fluctuations that occur during the decay process. Therefore, this application constructs an ND-ARIMA model that introduces nonlinear decay, uses the ND-ARIMA model to predict the nonlinearly changing SOH decay trend, and uses a volatility prediction model to predict the fluctuations that occur during the SOH decay process, so as to improve the accuracy of cell SOH prediction.

[0049] In some embodiments, the duration of the preset time period is related to the order of the ND-ARIMA model. The duration of the preset time period can be set to a value that satisfies the order of the ND-ARIMA model in order to improve the accuracy of prediction.

[0050] The next moment refers to the moment following the first moment according to the preset step size, that is, the moment that is the preset step size after the first moment.

[0051] It should be noted that the SOH time series data of the battery cells required to build the ND-ARIMA model are the SOH time series data of the same model as the battery cell under test, and the SOH time series data of the battery cells can be time series data covering the entire life cycle of the battery cell SOH, so as to improve the accuracy of the built model.

[0052] In one embodiment, such as Figure 2 As shown, the ND-ARIMA model with nonlinear decay can be pre-constructed through the following steps:

[0053] Step 201: Obtain the full lifecycle timing data of the cell's SOH (State of Health).

[0054] Among them, the full life cycle timing data of the cell SOH is the data of the cell of the same model as the cell under test.

[0055] Step 202: Perform stationarity testing on the full lifecycle time series data of the cell's SOH (State of Health) to determine the order of the first ARIMA model.

[0056] In some embodiments, stationarity detection of the cell's SOH full life-cycle time series data can be performed using ACF (Auto-Correlation Function) and PACF (Partial Auto-Correlation Function), and the order of the first ARIMA model can be determined based on the graphs of ACF and PACF, namely the order p of the autoregressive part, the order q of the moving average part, and the order d of the difference.

[0057] Step 203: Construct the first ARIMA model based on the full life-cycle time-series data of the cell's SOH and the order of the first ARIMA model.

[0058] In some embodiments, the expression of the constructed first ARIMA model is as follows (1):

[0059]

[0060] in, Let be the d-th differenced predicted value of SOH at the t-th sampling point; d is the difference order in the first ARIMA model; C is the offset, which is automatically determined during model execution; p is the order of the autoregressive component in the first ARIMA model; q is the order of the moving average component in the first ARIMA model; r t-i Let be the autocorrelation coefficient of the ith sampling point; Δ d y t-i Let θ be the actual d-th order difference of SOH at the t-th sampling point; t-j Let be the error influence coefficient of the ti-th sampling point; Let be the d-th order difference error of SOH at the tj-th sampling point.

[0061] Step 204: By adding a nonlinear decay coefficient to the first ARIMA model, an ND-ARIMA model with nonlinear decay is obtained.

[0062] In some embodiments, the expression for the ND-ARIMA model with introduced nonlinear decay is shown in equation (2) below:

[0063]

[0064] Among them, K t Let be the nonlinear decay coefficient at the t-th sampling point, containing two hyperparameters, a and b. The hyperparameters of the nonlinear decay coefficient can be obtained by fitting the cell's SOH full life-cycle time-series data using methods such as least squares.

[0065] Step 104: Based on the preset volatility prediction model, obtain the predicted value of the volatility term of the cell under test at the next moment according to the external excitation time series data; wherein, the volatility prediction model has learned the mapping relationship between the external excitation time series data and the corresponding volatility term at the next moment.

[0066] In some embodiments, the external excitation time series data within a preset time period can first be dimensionless processed, and the resulting dimensionless external excitation time series data can be input into a preset volatility prediction model to obtain the predicted volatility value of the cell under test at the next moment. As an example, the dimensionless processing of the external excitation time series data can be based on the following equation (3):

[0067]

[0068] Among them, A i,d The processing result of the i-th sampling point of the external excitation data of dimension d; a i,d The original data of the i-th sampling point of the external stimulus data of dimension d; a max,d and These are the theoretical maximum and minimum values ​​of the external stimulus data for dimension d within a preset time period, respectively.

[0069] In some embodiments, the volatility prediction model can be a Long Short-Term Memory (LSTM) model or other neural network models. In other words, this scheme uses a neural network model to capture the volatility during the SOH decay process.

[0070] In some embodiments, such as Figure 3 As shown, volatility prediction models can be constructed in advance through the following steps:

[0071] Step 301: Obtain SOH timing data samples and external excitation timing data samples of the battery cell within a historical time period.

[0072] It should be noted that the SOH timing data samples and external excitation timing data samples here are data from the same model of battery cell as the battery cell under test, and the parameters included in the SOH timing data samples and external excitation timing data samples are consistent with the parameters included in the SOH timing data and external excitation timing data in the above embodiments.

[0073] Step 302: Split the SOH time series data samples to obtain the fluctuation term time series data samples corresponding to the SOH time series data samples.

[0074] In some embodiments, singular spectrum decomposition can be performed on the SOH time series data samples to obtain the fluctuation term time series data samples corresponding to the SOH time series data samples. The singular spectrum decomposition process can be consistent with the singular spectrum decomposition process in related technologies.

[0075] Step 303: Determine the training samples based on the external stimulus time series data samples and the fluctuation term time series data samples.

[0076] In some embodiments, training samples can be determined based on a preset time series length, using external stimulus time series data samples and volatility term time series data samples. The preset time series length can be consistent with the duration of a preset time period. The training samples include the divided external stimulus time series data samples and their corresponding volatility data for the next time step. As an example, if the external stimulus time series data sample includes external stimulus data 1 at time 1, external stimulus data 2 at time 2, external stimulus data 3 at time 3, external stimulus data 4 at time 4, external stimulus data 5 at time 5, etc., and the fluctuation term time series data sample includes fluctuation term data 1 at time 1, fluctuation term data 2 at time 2, fluctuation term data 3 at time 3, fluctuation term data 4 at time 4, fluctuation term data 5 at time 5, etc., and the preset time series length is 4, then it can be determined that training sample 1 includes external stimulus time series data sample 1, i.e., {external stimulus data 1 at time 1, external stimulus data 2 at time 2, external stimulus data 3 at time 3, external stimulus data 4 at time 4}, and its corresponding fluctuation term data at the next time, i.e., fluctuation term data 5 at time 5; training sample 2 includes external stimulus time series data sample 2, i.e., {external stimulus data 2 at time 2, external stimulus data 3 at time 3, external stimulus data 4 at time 4, external stimulus data 5 at time 5}, and its corresponding fluctuation term data at the next time, i.e., fluctuation term data 6 at time 6, and so on.

[0077] Step 304: Input the training samples into the initial volatility prediction model to train the model and obtain the volatility prediction model.

[0078] Step 105: Determine the SOH prediction value of the cell under test at the next moment based on the predicted value of the trend term and the predicted value of the fluctuation term at the next moment.

[0079] In some embodiments, the process of determining the SOH prediction value of the cell under test at the next time step based on the trend term prediction value and the fluctuation term prediction value at the next time step may include: reconstructing the trend term prediction value and the fluctuation term prediction value at the next time step using singular spectrum analysis to obtain the SOH prediction value of the cell under test at the next time step.

[0080] It should be noted that the cell health status prediction method in this application embodiment can be applied to the battery module health status prediction process. By improving the accuracy of the health status prediction of each cell in the battery module, the accuracy of the battery module health status prediction can be improved.

[0081] According to the cell health state prediction method of the present application embodiment, the nonlinear decay trend of the SOH of the cell under test is predicted by introducing the nonlinear decay ND-ARIMA model, and the fluctuation prediction model is used to predict the fluctuations that occur during the SOH decay process of the cell under test. Based on the predicted decay trend and fluctuations, the SOH of the cell under test is determined. This method not only has high computational efficiency, but also improves the accuracy of cell SOH prediction, thereby improving the accuracy of battery module SOH prediction.

[0082] Next, the cell health status prediction method of the above embodiments will be applied to the battery module health status prediction process.

[0083] Figure 4 This is a flowchart illustrating a method for predicting the health status of a battery module provided in an embodiment of this application. It should be noted that the method for predicting the health status of a battery module in this application can be applied to the battery module health status prediction device in this application, and this device can be configured in an electronic device. Figure 4 As shown, the method may include:

[0084] Step 401: Based on the cell health status prediction method of the above embodiment, determine the SOH prediction value of each cell in the battery module under test at the next moment.

[0085] In other words, using the cell health state prediction method described in the above embodiments, the predicted SOH value of each cell in the battery module under test at the next moment is determined. Since the cells in the battery module are all of the same model, the ND-ARIMA model that introduces nonlinear decay involved in the SOH prediction process of different cells in the battery module can be the same, and the volatility prediction model involved can also be the same.

[0086] Step 402: Determine the predicted SOH value of the battery module under test at the next time step based on the predicted SOH value of each cell at the next time step.

[0087] In some embodiments, step 402 can be implemented as the method in the related art of determining the SOH prediction value of the battery module based on the SOH prediction value of each cell, or other implementation methods proposed by those skilled in the art based on actual needs, which are not limited in this application. As an example, the minimum value of the SOH prediction value of each cell in the battery module at the next moment can be used as the SOH prediction value of the battery module under test at the next moment.

[0088] The battery module health state prediction method according to the embodiments of this application introduces the nonlinear decay trend of the SOH of the cell under test by introducing the nonlinear decay ND-ARIMA model, and uses the volatility prediction model to predict the fluctuations that occur during the SOH decay process of the cell under test. Based on the predicted decay trend and fluctuations, the SOH of each cell in the battery module under test is determined. This method not only has high computational efficiency, but also improves the accuracy of cell SOH prediction, thereby improving the accuracy of battery module SOH prediction.

[0089] Since the consistency of battery cells in a battery module can affect the degradation of the battery module's health status, this application proposes yet another embodiment in order to improve the accuracy of battery module health status prediction.

[0090] Figure 5 A flowchart illustrating another method for predicting the health status of a battery module provided in an embodiment of this application. Figure 5 As shown, the method includes:

[0091] Step 501: Based on the cell health state prediction method of the above embodiment, determine the predicted SOH value of each cell in the battery module under test at the next moment.

[0092] Step 502: Obtain the cell consistency time sequence data of the battery module under test within a preset time period.

[0093] In step 502, the preset time period may be the same as or different from the preset time period in the above embodiments. The duration of the preset time period in this application embodiment is related to the time series duration determined by the model involved in steps 504 and 505 in this application embodiment.

[0094] In some embodiments, the cell consistency time-series data may include cell consistency data of the battery module under test at each sampling time within a preset time period. The cell consistency data may include voltage consistency index data and temperature consistency index data. The voltage consistency index data may be the voltage coefficient of variation, the voltage standard deviation, or the voltage range. The temperature consistency index data may be the temperature coefficient of variation, the temperature standard deviation, or the temperature range. In this embodiment, the voltage coefficient of variation may be used as the voltage consistency index data, and the temperature coefficient of variation may be used as the temperature consistency index data.

[0095] Step 503: Determine the timing data of voltage consistency index and temperature consistency index based on the cell consistency timing data.

[0096] Step 504: Based on the preset neural network model, determine the predicted value of the voltage consistency index of the battery module under test at the next moment according to the voltage consistency index time series data; wherein, the neural network model has learned the mapping relationship between the voltage consistency index time series data and its corresponding voltage consistency index value at the next moment.

[0097] Since the voltage consistency index data has weak periodicity, it can be predicted using a neural network model.

[0098] The neural network model can be a Long Short-Term Memory (LSTM) neural network model, which is trained based on training samples including voltage consistency index time-series data samples and their corresponding voltage consistency index value labels for the next time step. The duration of the time series data samples of the voltage consistency index can be consistent with the duration corresponding to a preset time period.

[0099] In other words, the time series data of the voltage consistency index within a preset time period is input into the trained LSTM neural network model to obtain the predicted value of the voltage consistency index of the battery module under test at the next moment.

[0100] Step 505: Based on the preset second ARIMA model, determine the predicted value of the temperature consistency index of the battery module under test at the next moment according to the time series data of the temperature consistency index; wherein, the second ARIMA model is constructed based on the time series data of the temperature consistency index of the battery module.

[0101] Since the temperature consistency index data changes periodically due to the influence of the external temperature control system, it can be predicted using the second ARIMA model.

[0102] The second ARIMA model can be constructed based on time-series data of temperature consistency indices throughout the entire lifecycle of the battery module's State of Health (SOH). The specific construction process is similar to that of the first ARIMA model and will not be elaborated here. The duration of the preset time period can be determined based on the order specified in the second ARIMA model.

[0103] Step 506: Determine the predicted SOH value of the battery module under test at the next moment based on the predicted values ​​of voltage consistency index, temperature consistency index, and SOH of each cell at the next moment.

[0104] In some embodiments of this application, the implementation process of step 506 may include: determining the cell consistency prediction value of the battery module under test at the next time step based on the predicted value of the voltage consistency index and the predicted value of the temperature consistency index at the next time step; and determining the cell consistency prediction value of the battery module under test at the next time step based on the predicted value of the cell consistency index and the predicted value of the SOH of each cell at the next time step.

[0105] As an example, the process of determining the predicted cell consistency value of the battery module under test at the next time step, based on the predicted values ​​of the voltage and temperature consistency indices, can be achieved by weighted summing of the predicted values ​​of the voltage and temperature consistency indices at the next time step, and then using the weighted sum as the predicted cell consistency value of the battery module under test at the next time step. The weights corresponding to the predicted values ​​of the voltage and temperature consistency indices at the next time step in the weighted calculation process can be determined by expert evaluation.

[0106] As another example, the process of determining the predicted cell consistency value of the battery module under test at the next moment based on the predicted voltage consistency index and the predicted temperature consistency index at the next moment can be achieved based on the following equation (4):

[0107]

[0108] Where CON is the predicted cell consistency value of the battery module under test at the next moment; CON t The predicted value of the temperature consistency index for the next moment; CON U This is the predicted value of the voltage consistency index for the next time step.

[0109] As an example, the process of determining the predicted cell consistency value of the battery module under test at the next time step based on the predicted cell consistency value and the predicted SOH value of each cell can be implemented based on the following equation (5):

[0110]

[0111] Among them, SOH pack is the predicted SOH value of the battery module under test at the next moment; n is the number of cells in the battery module under test;

[0112] This is the predicted cell consistency value for the next time step; SOH i Let SOH be the predicted value of the i-th cell in the battery module under test at the next moment.

[0113] According to the battery module health state prediction method of this application, a neural network model and a second ARIMA model are used to determine the predicted values ​​of voltage consistency index and temperature consistency index of the battery module under test at the next time step based on cell consistency time-series data. These values ​​are then used to determine the predicted cell consistency value of the battery module under test at the next time step. Combining the predicted cell consistency value with the predicted state of health (SOH) value of each cell in the battery module under test at the next time step, the predicted SOH value of the battery module under test is determined. This solution not only improves the accuracy of battery module health state prediction through the aforementioned cell health state prediction method, but also considers the impact of cell consistency on the SOH decay phenomenon of the battery module, thereby further improving the accuracy of battery module health state prediction.

[0114] To achieve the above embodiments, this application also provides a device for predicting the health status of battery cells.

[0115] Figure 6 This is a structural block diagram of a battery cell health status prediction device provided in an embodiment of this application. Figure 6 As shown, the device includes:

[0116] The first acquisition module 601 is used to acquire the health status (SOH) timing data and external excitation timing data of the battery cell under test within a preset time period.

[0117] The splitting module 602 is used to split the SOH time series data and obtain the trend item time series data in the SOH time series data;

[0118] The first prediction module 603 is used to obtain the predicted value of the trend term of the cell under test at the next moment based on the trend term time series data of the preset ND-ARIMA model with nonlinear decay; wherein, the ND-ARIMA model with nonlinear decay is constructed based on the cell SOH time series data.

[0119] The second prediction module 604 is used to obtain the predicted value of the fluctuation term of the battery cell under test at the next moment based on the preset fluctuation prediction model and the external excitation time series data; wherein, the fluctuation prediction model has learned the mapping relationship between the external excitation time series data and the corresponding fluctuation term at the next moment.

[0120] The first determining module 605 is used to determine the SOH prediction value of the cell under test at the next moment based on the trend prediction value and fluctuation prediction value at the next moment.

[0121] In some embodiments, the splitting module 602 is specifically used for:

[0122] The SOH time series data were decomposed using singular spectrum analysis to obtain the trend term time series data.

[0123] In some embodiments, the device further includes a first building module 606, the first building module 606 being configured to:

[0124] Obtain the full lifecycle timing data of the cell's State of Health (SOH);

[0125] Stationarity tests were performed on the time-series data of the cell's SOH throughout its entire life cycle to determine the order of the first ARIMA model;

[0126] Based on the full life cycle time series data of the cell's SOH and the order of the first ARIMA model, the first ARIMA model is constructed.

[0127] By adding a nonlinear decay coefficient to the first ARIMA model, an ND-ARIMA model with nonlinear decay is obtained.

[0128] In some embodiments, the device further includes a second building module 607, the second building module 607 being configured to:

[0129] Acquire SOH timing data samples and external excitation timing data samples of the battery cell within a historical time period;

[0130] The SOH time series data samples are split to obtain the fluctuation term time series data samples corresponding to the SOH time series data samples;

[0131] The training samples are determined based on the time series data samples of external excitation and fluctuation terms;

[0132] The training samples are input into the initial volatility prediction model to train the model and obtain the volatility prediction model.

[0133] In some embodiments, the first determining module 605 is specifically used for:

[0134] Using singular spectrum analysis, the predicted values ​​of the trend term and fluctuation term for the next time step are reconstructed to obtain the predicted value of SOH for the next time step.

[0135] It should be noted that the explanation of the aforementioned method embodiment for predicting the health status of battery cells also applies to the battery cell health status prediction device of this embodiment, and will not be repeated here.

[0136] To achieve the above embodiments, this application also provides a device for predicting the health status of a battery module.

[0137] Figure 7 This is a structural block diagram of a battery module health status prediction device provided in an embodiment of this application. Figure 7 As shown, the device includes:

[0138] The third prediction module 701 is used to determine the predicted SOH value of each cell in the battery module under test at the next moment based on the cell health state prediction method of the first aspect described above.

[0139] The second determining module 702 is used to determine the predicted SOH value of the battery module under test at the next moment based on the predicted SOH value of each cell at the next moment.

[0140] In some embodiments, the device further includes:

[0141] The second acquisition module 703 is used to acquire the cell consistency time sequence data of the battery module under test within a preset time period;

[0142] The third determining module 704 is used to determine the timing data of voltage consistency index and temperature consistency index based on the cell consistency timing data;

[0143] The fourth prediction module 705, based on a preset neural network model, determines the predicted value of the voltage consistency index of the battery module under test at the next moment according to the voltage consistency index time series data; wherein, the neural network model has learned the mapping relationship between the voltage consistency index time series data and its corresponding voltage consistency index value at the next moment.

[0144] The fifth prediction module 706, with a preset second ARIMA model, determines the predicted value of the temperature consistency index of the battery module under test at the next moment based on the time series data of the temperature consistency index; wherein, the second ARIMA model is constructed based on the time series data of the temperature consistency index of the battery module;

[0145] The second determining module 702 is further used for:

[0146] Based on the predicted values ​​of voltage consistency index, temperature consistency index, and SOH of each cell at the next time step, the predicted SOH value of the battery module under test at the next time step is determined.

[0147] In some embodiments, the second determining module 702 is specifically used for:

[0148] Based on the predicted values ​​of voltage consistency index and temperature consistency index at the next time moment, the predicted value of cell consistency of the battery module under test at the next time moment is determined.

[0149] Based on the predicted cell consistency value at the next time step and the predicted SOH value of each cell, the predicted cell consistency value of the battery module under test at the next time step is determined.

[0150] It should be noted that the explanation of the aforementioned method embodiment for predicting the health status of a battery module also applies to the device for predicting the health status of a battery module in this embodiment, and will not be repeated here.

[0151] According to embodiments of this application, this application also provides an electronic device and a readable storage medium.

[0152] Figure 8 This is a structural block diagram of an electronic device provided in an embodiment of this application. The electronic device is used to implement the cell health state prediction method in the above embodiments, and / or to implement the battery module health state prediction method in the above embodiments. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The electronic device can also be a vehicle. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or claimed herein.

[0153] like Figure 8 As shown, the electronic device includes one or more processors 801, a memory 802, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components are interconnected via different buses and can be mounted on a common motherboard or otherwise as required. The processors can process instructions executed within the electronic device, including instructions stored in or on memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In other embodiments, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple electronic devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 8 Take the 801 processor as an example.

[0154] The memory 802 is the non-transitory computer-readable storage medium provided in this application. The memory stores instructions executable by at least one processor to cause the at least one processor to execute the cell health state prediction method and / or the battery module health state prediction method provided in this application. The non-transitory computer-readable storage medium of this application stores computer instructions for causing a computer to execute the cell health state prediction method and / or the battery module health state prediction method provided in this application.

[0155] The memory 802, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the cell health state prediction method and / or the battery module health state prediction method in the embodiments of this application, corresponding to program instructions / modules. The processor 801 executes various server functions and data processing by running the non-transitory software programs, instructions, and modules stored in the memory 802, thereby implementing the cell health state prediction method and / or the battery module health state prediction method in the above method embodiments.

[0156] The memory 802 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function. The data storage area may store data created by the use of the electronic device based on a method for predicting the health status of the battery cells and / or a method for predicting the health status of the battery module. Furthermore, the memory 802 may include high-speed random access memory and may also include non-transient memory, such as at least one disk storage device, flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory 802 may optionally include memory remotely located relative to the processor 801, and these remote memories can be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0157] The electronic device may also include an input device 803 and an output device 804. The processor 801, memory 802, input device 803, and output device 804 can be connected via a bus or other means. Figure 8 Taking the example of a connection between China and Israel via a bus.

[0158] Input device 803 can receive input numerical or character information, and generate key signal inputs related to user settings and function control of the electronic device, such as touch screens, keypads, mice, trackpads, touchpads, joysticks, one or more mouse buttons, trackballs, joysticks, etc. Output device 804 may include display devices, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors). The display device may include, but is not limited to, liquid crystal displays (LCDs), light-emitting diode (LED) displays, and plasma displays. In some embodiments, the display device may be a touch screen.

[0159] Various implementations of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, application-specific integrated circuits (ASICs), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include: implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transferring data and instructions to the storage system, the at least one input device, and the at least one output device.

[0160] These computational programs (also referred to as programs, software, software applications, or code) include machine instructions for a programmable processor and can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. As used herein, the terms “machine-readable medium” and “computer-readable medium” refer to any computer program product, device, and / or apparatus (e.g., disk, optical disk, memory, programmable logic device (PLD)) used to provide machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The term “machine-readable signal” refers to any signal used to provide machine instructions and / or data to a programmable processor.

[0161] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0162] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), the Internet, and blockchain networks.

[0163] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. A server can be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service ecosystem, addressing the shortcomings of traditional physical hosts and VPS (Virtual Private Server, or simply "VPS") services, such as high management difficulty and weak business scalability. Servers can also be servers for distributed systems or servers incorporating blockchain technology.

[0164] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this application can be achieved, and this is not limited herein.

[0165] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for predicting the health status of a battery cell, characterized in that, include: Acquire the State of Health (SOH) timing data and external excitation timing data of the cell under test within a preset time period; The SOH time series data is split to obtain the trend item time series data in the SOH time series data; Based on a pre-defined ND-ARIMA model that incorporates nonlinear decay, the predicted value of the trend term of the cell under test at the next time step is obtained according to the trend term time series data; wherein, the ND-ARIMA model that incorporates nonlinear decay is constructed based on the cell's SOH time series data; Based on a preset volatility prediction model, the predicted value of the volatility term of the battery cell under test at the next moment is obtained according to the external excitation time series data; wherein, the volatility prediction model has learned the mapping relationship between the external excitation time series data and the corresponding volatility term at the next moment. Based on the predicted values ​​of the trend term and the fluctuation term at the next time moment, the predicted SOH value of the cell under test at the next time moment is determined.

2. The method according to claim 1, characterized in that, The step of splitting the SOH time series data to obtain the trend item time series data in the SOH time series data includes: The SOH time series data is decomposed using singular spectrum analysis to obtain the trend term time series data.

3. The method according to claim 1, characterized in that, The ND-ARIMA model that introduces nonlinear decay is constructed in advance through the following steps: Obtain the full lifecycle timing data of the cell's State of Health (SOH); The stationarity of the time-series data of the cell's SOH throughout its entire life cycle is tested to determine the order of the first ARIMA model; Based on the full life-cycle time-series data of the cell's SOH and the order of the first ARIMA model, a first ARIMA model is constructed; By adding a nonlinear decay coefficient to the first ARIMA model, the ND-ARIMA model with introduced nonlinear decay is obtained.

4. The method according to claim 1, characterized in that, The volatility prediction model is constructed in advance through the following steps: Acquire SOH timing data samples and external excitation timing data samples of the battery cell within a historical time period; The SOH time series data samples are split to obtain the fluctuation term time series data samples corresponding to the SOH time series data samples; The training samples are determined based on the external excitation time series data samples and the fluctuation term time series data samples; The training samples are input into the initial volatility prediction model for model training to obtain the volatility prediction model.

5. The method according to claim 2, characterized in that, The step of determining the SOH prediction value of the cell under test at the next time step based on the trend term prediction value and the fluctuation term prediction value at the next time step includes: According to the singular spectrum analysis method, the predicted values ​​of the trend term and fluctuation term for the next time step are reconstructed to obtain the predicted value of SOH for the next time step.

6. A method for predicting the health status of a battery module, characterized in that, include: Based on the cell health state prediction method as described in any one of claims 1-5, the predicted SOH value of each cell in the battery module under test at the next moment is determined; Based on the predicted SOH value of each cell at the next time step, the predicted SOH value of the battery module under test at the next time step is determined.

7. The method according to claim 6, characterized in that, Also includes: Obtain the cell consistency time sequence data of the battery module under test within a preset time period; Based on the cell consistency timing data, determine the voltage consistency index timing data and the temperature consistency index timing data; Based on a preset neural network model, the predicted value of the voltage consistency index of the battery module under test at the next time step is determined according to the voltage consistency index time series data; wherein, the neural network model has learned the mapping relationship between the voltage consistency index time series data and its corresponding voltage consistency index value at the next time step. Based on the preset second ARIMA model, the predicted value of the temperature consistency index of the battery module under test at the next moment is determined according to the time series data of the temperature consistency index; wherein, the second ARIMA model is constructed based on the time series data of the temperature consistency index of the battery module; The step of determining the predicted SOH value of the battery module under test at the next time step based on the predicted SOH value of each cell at the next time step includes: Based on the predicted values ​​of voltage consistency index, temperature consistency index, and SOH of each cell at the next time step, the predicted SOH value of the battery module under test at the next time step is determined.

8. The method according to claim 7, characterized in that, The step of determining the predicted SOH value of the battery module under test at the next moment based on the predicted voltage consistency index, the predicted temperature consistency index, and the predicted SOH value of each cell includes: Based on the predicted values ​​of voltage consistency index and temperature consistency index at the next time moment, the predicted value of cell consistency of the battery module under test at the next time moment is determined. Based on the predicted cell consistency value at the next time step and the predicted SOH value of each cell, the predicted cell consistency value of the battery module under test at the next time step is determined.

9. A device for predicting the health status of a battery cell, characterized in that, include: The first acquisition module is used to acquire the health status (SOH) timing data and external excitation timing data of the battery cell under test within a preset time period. The splitting module is used to split the SOH time series data to obtain the trend item time series data in the SOH time series data; The first prediction module is used to obtain the predicted value of the trend term of the cell under test at the next time step based on the trend term time series data, according to the preset ND-ARIMA model with nonlinear decay. The ND-ARIMA model with nonlinear decay is constructed based on the cell's SOH time series data. The second prediction module is used to obtain the predicted value of the fluctuation term of the battery cell under test at the next time moment based on the external excitation time series data and a preset fluctuation prediction model; wherein the fluctuation prediction model has learned the mapping relationship between the external excitation time series data and the fluctuation term at the corresponding next time moment. The first determining module is used to determine the SOH predicted value of the cell under test at the next time moment based on the predicted value of the trend term and the predicted value of the fluctuation term at the next time moment.

10. An electronic device, characterized in that, include: processor; A memory for storing executable instructions of the processor; wherein the processor is configured to execute the instructions to implement the cell health state prediction method as described in any one of claims 1-5, and / or to implement the battery module health state prediction method as described in any one of claims 6-8.

Images