A parameter identification result processing method and device, and a storage medium

Abstract

The application discloses a parameter identification result processing method and device, and a storage medium. The method comprises the following steps: obtaining an original identification result; the original identification result comprises a plurality of multidimensional data sets identified under a plurality of working conditions, each multidimensional data set comprises a plurality of different electrochemical model parameters, and each electrochemical model parameter corresponds to one dimension; the multidimensional data sets are mapped to a hyperspherical coordinate space and converted into spatial points; the multidimensional spherical coordinates of each spatial point form a sample, and all the samples form a sample set; the angle values of each dimension in the sample set are statistically analyzed to determine a reference point; a target point is selected from the spatial points based on the multidimensional spherical coordinates of the reference point; and the multidimensional data set corresponding to the target point is a target identification result. The electrochemical model parameters with high reliability are obtained by screening the original identification result.

Description

Technical Field

[0001] This application relates to the field of energy storage technology, and in particular to a method and apparatus for processing parameter identification results, and a storage medium. Background Technology

[0002] Electrochemical models of lithium-ion batteries can accurately simulate the charging and discharging process, and are widely used in the field of energy storage technology. An electrochemical model typically consists of multiple partial differential equations involving dozens of physical parameters, such as electrode surface area, electrode thickness, conductivity, and positive and negative electrode ion concentrations. These physical parameters are the electrochemical model parameters. Currently, the specific values ​​of these electrochemical model parameters can be obtained through parameter identification technology.

[0003] Since the internal state of a battery can change at any time, parameter identification must be performed continuously. Each parameter identification will output a large number of identification results. The key to ensuring the reliability of identification is to filter out a large number of identification results. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides a method, apparatus, and storage medium for processing parameter identification results. Specifically, the technical solution of this application is as follows:

[0005] Firstly, this application provides a method for processing parameter identification results, including:

[0006] Obtain the original identification results; the original identification results include multidimensional datasets identified under multiple operating conditions, each multidimensional dataset contains multiple different electrochemical model parameters; and the electrochemical model parameters in different multidimensional datasets correspond one-to-one.

[0007] The multidimensional dataset is mapped to a hyperspherical coordinate space and converted into spatial points; the multidimensional spherical coordinates of each spatial point constitute a sample, and all the samples constitute a sample set; the multidimensional spherical coordinates include a modulus and several angle values;

[0008] Perform statistical analysis on the sample set from different dimensions to determine reference points;

[0009] Based on the multidimensional spherical coordinates of the reference point, a target point is selected from the spatial points; the multidimensional dataset corresponding to the target point is the target identification result.

[0010] In some implementations, the step of performing statistical analysis on the sample set from different dimensions to determine reference points includes:

[0011] The sample set is divided into several sub-sample sets;

[0012] Traverse all the said sub-sample sets, and select target data from the sub-sample sets according to the distribution of the modulus and the angle values ​​in the sub-sample sets to form the multidimensional spherical coordinates of the candidate points;

[0013] Candidate points that meet preset conditions are selected as reference points.

[0014] In some implementations, the preset condition is that the sample repeats the most within a preset sub-sample set;

[0015] Alternatively, the preset condition is repetition within the preset subsample set.

[0016] In some implementations, before mapping the multidimensional dataset to a hypersphere coordinate space and converting it into spatial points, the following steps are included:

[0017] Based on the operating condition dataset corresponding to the operating conditions of each of the multidimensional datasets, determine the upper and lower limits of each electrochemical model parameter in the multidimensional dataset;

[0018] Based on the upper limit and the lower limit, each electrochemical model parameter in the multidimensional dataset is normalized to obtain a normalized value for each electrochemical model parameter in the multidimensional dataset; a new multidimensional dataset is generated based on the normalized values.

[0019] Some implementations also include:

[0020] Obtain the sensitivity value for each of the electrochemical model parameters;

[0021] Sort each of the normalized values ​​in the new multidimensional dataset according to the sensitivity values.

[0022] Secondly, this application provides a processing apparatus for parameter identification results, comprising:

[0023] The first acquisition module is used to acquire the original identification results; the original identification results include multidimensional datasets identified under multiple operating conditions, each multidimensional dataset contains multiple different electrochemical model parameters; and the electrochemical model parameters in different multidimensional datasets correspond one-to-one.

[0024] The transformation module is used to map the multidimensional dataset to a hyperspherical coordinate space, converting it into spatial points; the multidimensional spherical coordinates of each spatial point constitute a sample, and all the samples constitute a sample set; the multidimensional spherical coordinates include a modulus and several angle values;

[0025] The analysis module is used to perform statistical analysis on the sample set in different dimensions to determine reference points;

[0026] The selection module is used to select a target point from the spatial points based on the multidimensional spherical coordinates of the reference point; the multidimensional dataset corresponding to the target point is the target identification result.

[0027] In some implementations, the analysis module is used to divide the sample set into several sub-sample sets; the analysis module is also used to traverse all the sub-sample sets, select target data from the sub-sample sets according to the distribution of the modulus and the angle values ​​in the sub-sample sets, form the multidimensional spherical coordinates of candidate points; and select candidate points that meet preset conditions as the reference points.

[0028] In some implementations, the preset condition is that the sample repeats the most within a preset sub-sample set;

[0029] Alternatively, the preset condition is repetition within the preset subsample set.

[0030] In some implementations, the conversion module includes:

[0031] The determining unit is configured to determine the upper limit and lower limit of each electrochemical model parameter in the multidimensional dataset based on the operating condition dataset corresponding to the operating condition in each multidimensional dataset.

[0032] The processing unit is configured to normalize each electrochemical model parameter in the multidimensional dataset according to the upper limit value and the lower limit value, to obtain a normalized value for each electrochemical model parameter in the multidimensional dataset; and to generate a new multidimensional dataset based on the normalized value.

[0033] Thirdly, this application provides a storage medium storing at least one instruction, which is loaded and executed by a processor to perform the operation performed by the parameter identification result processing method as described in any of the preceding claims.

[0034] Compared with the prior art, the beneficial effects of this application are as follows:

[0035] The identified original electrochemical model parameters are mapped to a hyperspherical coordinate space, and each data row is converted into a spatial point. By performing statistical analysis on the spherical coordinates of the spatial points, a reference point is determined. The target point is then determined from the spatial points based on the reference point. The corresponding original electrochemical model parameters are indexed to the target point, which is the final set of electrochemical model parameters with better reliability. Attached Figure Description

[0036] The accompanying drawings used in the description of the embodiments of this application are briefly introduced below:

[0037] Figure 1This is a flowchart of an embodiment of the parameter identification result processing method of this application;

[0038] Figure 2 This is a schematic diagram showing the distribution of the 24-dimensional hypersphere coordinates corresponding to the sample consisting of 10,000 points.

[0039] Figure 3 It is a comparison between the reproduced voltage of the target point obtained by Ident1 and the original voltage;

[0040] Figure 4 It is a comparison between the reproduced voltage of the target point obtained by Ident2 and the original voltage;

[0041] Figure 5 It is a comparison between the reproduced voltage of the target point obtained by Ident3 and the original voltage;

[0042] Figure 6 This is a structural block diagram of an embodiment of the parameter identification result processing device of this application. Detailed Implementation

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the specific implementation methods of this application will be described below with reference to the accompanying drawings. The drawings and implementation methods described below are merely some embodiments of this application. For those skilled in the art, other drawings and other implementation methods can be obtained based on these drawings and implementation methods without creative effort. Any adjustments and improvements made without departing from the concept of this application are within the protection scope of this application.

[0044] To keep the drawings concise, only the parts relevant to this application are shown schematically in each drawing, and they do not represent the actual structure of the product. In addition, to make the drawings concise and easy to understand, some parts with the same structure or function are only shown schematically in some drawings, and there may actually be more or fewer parts with the same structure or function.

[0045] In this application, unless otherwise expressly specified and limited, ordinal numbers, such as “first”, “second”, etc., are used only to distinguish and describe related objects, and should not be construed as indicating or implying the relative importance or order between related objects; furthermore, they do not represent the number of related objects.

[0046] Electrochemical models of lithium-ion batteries can accurately simulate the charging and discharging process, and are widely used in the field of energy storage technology. An electrochemical model typically consists of multiple partial differential equations involving dozens of physical parameters, such as electrode surface area, electrode thickness, conductivity, and positive and negative electrode ion concentrations. These physical parameters are the electrochemical model parameters. Currently, the specific values ​​of these electrochemical model parameters can be obtained through parameter identification technology. Parameter identification technology can be understood as the process of taking the battery cell's operating data (such as current, voltage, and temperature) as input, and outputting the electrochemical model parameters through a parameter identification algorithm.

[0047] Because chemical reactions occur constantly within the battery cell, its internal state is constantly changing, necessitating continuous parameter identification. For a single battery cell's daily operating data, assuming one identification session yields 200 data rows, representing 200 sets of electrochemical model parameters, with each row containing 24 electrochemical model parameters, the 200 rows would contain 200 * 24 specific data values. Taking the daily data identification of a power plant as an example: operating data is recorded every 10 seconds. Assuming approximately one day's worth of operating data is used (a total of 8442 data entries), the parameter identification algorithm would run for about 40-50 minutes, yielding approximately 100,000 rows of identification results—a massive amount of data.

[0048] Under normal circumstances, the electrochemical parameters corresponding to the operating data of a single battery cell within a single day or a similar timeframe are a stable set of data. However, in the current parameter identification algorithm framework, some parameters exhibit significant fluctuations; for example, values ​​may be too high in some data rows and too low in others, compromising the reliability of the parameter identification results. Therefore, it is necessary to process the original identification results to ultimately determine a set of electrochemical model parameters with better reliability. A common approach is to use the mean method, which calculates the average value of each parameter based on all data rows. Taking 24 electrochemical model parameters as an example, the average value of each parameter is calculated, and these average values ​​are then used as the final electrochemical parameter identification results. However, the mean method averages out the parameter fluctuations, failing to accurately reflect the changing trends of the parameters, resulting in poor reliability.

[0049] Based on this, this application proposes a method for processing parameter identification results. The original electrochemical model parameters obtained by identification are mapped to a hyperspherical coordinate space. Each data row is converted into a spatial point. By performing statistical analysis on the spherical coordinates of the spatial points, a reference point is determined. The target point is determined from the spatial points based on the reference point. The corresponding original electrochemical model parameters are indexed according to the target point, which is the final set of electrochemical model parameters with better reliability.

[0050] The following explanation is provided in conjunction with the accompanying drawings, such as... Figure 1As shown, the method for processing parameter identification results in this embodiment includes the following steps:

[0051] S100 acquires the original identification results.

[0052] Specifically, the electrochemical model includes more than a dozen coupled partial differential equations. Each equation involves several parameters that need to be input. These parameters are the electrochemical model parameters. For the overall electrochemical model, there are often dozens of parameters that need to be input, such as the concentration of positive and negative electrode particles, electrolyte concentration, conductivity, and so on.

[0053] Parameter identification refers to simulating the electrochemical model using parameter identification algorithms based on data from each cell in the energy storage system under different operating conditions, thereby outputting the corresponding electrochemical model parameters for each cell under different operating conditions. A single parameter identification operation yields the electrochemical model parameters for one cell under a specific operating condition, typically referred to as a set of electrochemical model parameters or a row of data. Each row or set of electrochemical model parameters contains specific values ​​for different parameters. It should be noted that different electrochemical models have different numbers and types of parameters. For example, pseudo-two-dimensional (P2D) models or single-particle models with electrolyte dynamics (SPMe) usually have 20-30 electrochemical model parameters, and the specific parameters can be adjusted according to the complexity of the model. For cells belonging to the same energy storage system, the same electrochemical model is applicable.

[0054] For identifying the operating conditions of a single battery cell on a single day, for example, one identification might yield 200 rows of data. If the electrochemical model contains 24 parameters, then each row of data contains the same 24 parameters, although the values ​​of the same parameter may be the same or different in different rows. The original identification result mentioned in the embodiments of this application refers to the collection of these data rows. Taking the daily data identification of a single battery cell in a power station as an example: operating condition data is recorded every 10 seconds. Assuming that one day's operating condition data (a total of 8442 operating condition data records) is used for parameter identification, approximately 100,000 rows of identification results can be obtained. These 100,000 rows of data constitute the original identification result. For all battery cells in a lithium battery energy storage system, electrochemical model parameter identification needs to be performed on a daily data basis.

[0055] Each of the above data rows can be regarded as a dataset. The original identification result is a multidimensional dataset identified under multiple operating conditions. For the same cell, the electrochemical model parameters in different multidimensional datasets correspond one-to-one.

[0056] S200 maps multidimensional datasets to hypersphere coordinate space, converting them into spatial points.

[0057] Specifically, suppose a multidimensional dataset contains electrochemical model parameters x1, x2, ..., x 24 This dataset can be considered as 24-dimensional planar coordinates (x1, x2, ..., x...). 24 By using projection, this multidimensional dataset can be mapped to a hyperspherical coordinate space, converting the planar coordinates into a hyperspherical coordinate representation, i.e., multidimensional spherical coordinates, including 23 angle values ​​(θ1, θ2, ..., θ). 23 ) and a modulus L, denoted as (θ1,θ2,...,θ 23 The specific conversion process is as follows:

[0058]

[0059] Based on the aforementioned multidimensional spherical coordinates, a spatial point can be determined in the hyperspherical coordinate space, and this spatial point corresponds to a multidimensional dataset. Each multidimensional dataset in the original identification result can be converted into a spatial point, and the multidimensional spherical coordinates of each spatial point constitute a sample. All these samples constitute a sample set. It is understood that the embodiments in this application use 24 electrochemical model parameters as an example for illustration. In specific implementations, the actual number of electrochemical model parameters depends on the electrochemical model itself.

[0060] S300 performs statistical analysis on the sample set from different dimensions to determine reference points.

[0061] Specifically, in multidimensional spherical coordinates, both the angle value and the modulus can be regarded as a dimension. Statistical analysis is performed on the data of each dimension to calculate the maximum probability density point of the data in each dimension. All the maximum probability density points are combined to form a new multidimensional spherical coordinate. Based on this coordinate, a new point, that is, a reference point, can be constructed in the above hyperspherical coordinate space.

[0062] With the above (θ1,θ2,...,θ) 23 Taking θ1, θ2, ..., θ as an example, where the modulus L corresponds to one dimension, the modulus L of all samples is extracted for statistical analysis. Based on the distribution of these modulus data (e.g., histogram distribution, Poisson distribution, etc.), the maximum probability density point is calculated. 23 Using the same method, 24 points with the highest probability density will eventually be calculated, thus obtaining the multidimensional spherical coordinates of the reference point.

[0063] The S400 selects target points from spatial points based on the multidimensional spherical coordinates of the reference points; the multidimensional dataset corresponding to the target points is the target identification result.

[0064] Specifically, in order to filter out data that meets reliability requirements from the original identification results, this embodiment uses a reference point as a benchmark to calculate the Euclidean distance between the reference point and each spatial point, and selects the spatial point with the smallest Euclidean distance to the reference point as the target point. The multidimensional spherical coordinates of the target point are obtained through mapping of a multidimensional dataset, so they can be indexed to the specific dataset in the original identification results, thereby achieving the filtering of the original identification results.

[0065] In this embodiment, the identified original electrochemical model parameters are mapped to a hyperspherical coordinate space, and each data row is converted into a spatial point. By performing statistical analysis on the spherical coordinates of the spatial points, a reference point is determined. The target point is then determined from the spatial points based on the reference point. The corresponding original electrochemical model parameters are indexed according to the target point, which is the final set of electrochemical model parameters with better reliability selected.

[0066] Assuming the electrochemical model parameters include A, B, C, D, and E, the original identification results include multidimensional datasets identified under operating conditions 1 through m. The correspondence between operating conditions, multidimensional datasets, and samples (multidimensional spherical coordinates) is shown in Table 1.

[0067] Operating Condition 1 (A1, B1, C1, D1, E1) <![CDATA[Sample 1 (θ 11 , θ 21 , θ 31 , θ 41 , L1)]]> Space point 1 Operating Condition 2 (A2, B2, C2, D2, E2) <![CDATA[Sample 2(θ 12 , θ 22 , θ 32 , θ 42 , L2)]]> Space point 2 …… …… …… …… Working condition m (Am, Bm, Cm, Dm, Em) <![CDATA[sample m(θ 1m , θ 2m , θ 3m , θ 4m , Lm)]]> Space point n

[0068] Table 1. Correspondence between operating conditions, multidimensional datasets, and samples.

[0069] The method for determining the reference point is explained below based on Table 1.

[0070] (1) Divide the sample set into multiple sub-sample sets.

[0071] Specifically, samples 1 through m in the table above constitute the sample set. The m samples are then divided into multiple subsets using an arithmetic progression. Assuming m is 100,000, the subsets are 10,000, 20,000, 30,000, ..., 100,000, representing the number of samples in each subset. Each subsequent subset always contains data from the previous subset. This subdivision automatically increases the sample size and improves reliability. Of course, the subdivision method is not unique; other partitioning intervals or non-arithmetic progression methods can also be used.

[0072] (2) Traverse all subsample sets, and select target data from the subsample sets according to the distribution of modulus and angle values ​​in the subsample sets to form the multidimensional spherical coordinates of candidate points.

[0073] Taking a subset of samples as an example, assuming it contains n samples, we select an appropriate distribution method for statistical analysis based on the distribution of angles θ1, θ2, θ3, θ4 and modulus L. For example, discrete data can use a histogram distribution, and continuous data can use a Poisson distribution. For a sufficiently rich sample, the set of values ​​formed by the same angle or modulus of all points can be regarded as a continuous value, so any distribution method can be used.

[0074] For θ1, it can be determined according to θ 11 ~θ 1m Based on the distribution of the data, calculate the maximum probability density value and denote the angle corresponding to this value as 'a'. Similarly, calculate the maximum probability density values ​​for θ2, θ3, and θ4 respectively, obtaining b, c, and d. The same applies to the modulus L, calculating the maximum probability density value to obtain e. This yields a multidimensional spherical coordinate system (a, b, c, d, e). Based on this spherical coordinate system (i.e., the target data), a new spatial point can be determined in the original hyperspherical coordinate space; this point is the candidate point. Repeat the above steps for each subset of samples to obtain multiple candidate points.

[0075] It should be noted that, in addition to the maximum probability density value, the second-highest probability density value can also be used to select target data. Whether to choose the maximum or second-highest probability density value depends on the actual identification algorithm; different identification algorithms produce different distributions of identification results. Figure 2 As shown, Figure 2 This represents the distribution of values ​​corresponding to 10,000 24-dimensional hypersphere coordinates. The distribution of parameters may exhibit both local peaks and boundary maximum peaks. In such cases, the target data can be selected based on the second maximum probability density value.

[0076] (3) Screen the candidate points and determine the reference points. Specifically, assuming the sample set is divided into N subsets, N candidate points can be obtained through the above steps. Statistical analysis is then performed on the duplicate points and their frequency of repetition among the N candidate points.

[0077] As shown in Table 2, Sample represents the number of samples in each sub-sample set, and Ident1, Ident2, and Ident3 represent the index values ​​of the candidate points in each sub-sample set after the first, second, and third parameter identifications, respectively. For example, the candidate point of sub-sample set 1 in the first identification is the spatial point composed of the 7977th sample out of 10000 samples (arranged in chronological order of data collection time).

[0078]

[0079] Table 2. Index of candidate points in arithmetic progression samples from 10,000 to 100,000.

[0080] Taking the aforementioned 10 sub-sample sets as an example, in one implementation, there exists a candidate point that repeats from the nth (n<10) sub-sample set to the 10th set. For example, the spatial point corresponding to "7977" in Ident1 repeats in sub-sample sets 1 to 10. This means that the sample size is rich enough to make the 24-dimensional electrochemical space corresponding to the battery condition statistically "filled". At this point, further increasing the sample size will not change the maximum probability density value. Therefore, using the candidate point with the most repetitions as the reference point can ensure the stability of the screening results.

[0081] Another scenario involves a candidate point that, while not having the highest repetition rate, has a large sample size in the subset containing it. For example, in Ident2, the spatial point corresponding to "87480" may not have the highest repetition rate across the 10 subsets, but it belongs to subsets 9 and 10, which have large sample sizes. Therefore, the spatial point corresponding to this value can be used as a reference point. In the example above, subsets 9 and 10 are used as the preset subset range; however, this can be flexibly adjusted in practice. Furthermore, assuming multiple repetitive points exist within the preset subset range, the point with the highest repetition rate is selected as the candidate point. To ensure the stability of the reference points, the sample size can be further increased and the number of subsample sets increased. For example, the amount of data in the original identification results can be increased, and the subsample sets can be further divided in an arithmetic progression based on the original subsample sets. Before the data size was increased, there were 100,000 rows of data, which were divided into 10 subsample sets. After the increase, there are 150,000 rows. Based on the 10 subsample sets, 5 more subsample sets (110,000, 120,000, 130,000 to 150,000 points) can be added. In the expanded samples, the maximum probability density value of each dimension is recalculated to confirm whether the value is the same as the probability density value corresponding to the reference point. If they are not the same, the reference point is re-determined by repeating the above steps (1) and (2). The comparison between the reproduced voltage of the reference points obtained in Ident1, Ident2, and Ident3 and the actual voltage is as follows. Figures 3-5 As shown, the MSE (mean squared error loss function) of the voltage is on the order of 9E-4.

[0082] After determining the reference point, the target point is selected from the spatial points based on the multidimensional spherical coordinates of the reference point; the multidimensional dataset corresponding to the target point is the target identification result.

[0083] In one implementation, based on any of the above embodiments, before mapping the cube to a hypersphere coordinate space and converting it into spatial points, the method further includes:

[0084] Based on the operating conditions corresponding to each multidimensional dataset, determine the upper and lower limits of each electrochemical model parameter in the multidimensional dataset.

[0085] Based on the upper and lower limits, the parameters of each electrochemical model in the multidimensional dataset are normalized to obtain the normalized values ​​of each electrochemical model parameter in the multidimensional dataset; a new multidimensional dataset is generated based on the normalized values.

[0086] Specifically, each electrochemical model has a different identification space, which refers to the range defined by the upper and lower limits of the parameters. The parameter identification space varies under different operating conditions, and parameter identification algorithms often update the upper and lower limits of all or some parameters based on the operating condition dataset (e.g., initial SOC values). The upper and lower limits of the parameter identification space for each parameter normalize all points in the resulting sample, thus obtaining a multidimensional dataset with values ​​ranging from 0 to 1, facilitating subsequent data processing.

[0087] It is worth noting that when dealing with larger sample sizes, the normalization range can be set to integers (such as 0-255) to shorten the computation time.

[0088] Furthermore, in one implementation, since different electrochemical model parameters have varying effects on the output voltage (simulated voltage) of the electrochemical model (also known as parameter sensitivity), for example, some electrochemical parameters cause drastic changes in the simulated voltage with slight changes (called high-sensitivity parameters), while others cause drastic changes in the simulated voltage with relatively stable voltage (called low-sensitivity parameters), sensitivity processing can be performed before determining the reference point to ensure that all dimensions of the subsequently transformed hypersphere coordinates have the same sensitivity. Specifically, the parameters are sorted in descending order of their sensitivity values, with parameters of high sensitivity placed first and parameters of low sensitivity placed last. For example, assuming the electrochemical model parameters include four parameters A, B, C, and D, and the sensitivity is C > A > B > D, the parameters in each identified row of data are arranged in the order CABD. For each row of data, during the subsequent hypersphere coordinate transformation, the sensitivity of the parameter values ​​at the beginning is transferred to the cosθ values ​​at the beginning and also to the cosθ values ​​at the end. This ensures that all dimensions of the hypersphere coordinates obtained from the subsequent transformation have maximum sensitivity, thus giving all parameters equal "voting power" in determining the final set of electrochemical parameters, which is equivalent to each parameter having the same weight in the selection process.

[0089] Based on the same technical concept, embodiments of this application provide a processing device for parameter identification results, such as... Figure 6 As shown, it includes a first acquisition module 100, a conversion module 200, an analysis module 300, and a selection module 400, wherein:

[0090] The first acquisition module 100 is used to acquire the original identification results. The original identification results include multidimensional datasets identified under multiple operating conditions. Each multidimensional dataset contains multiple different electrochemical model parameters. The electrochemical model parameters in different multidimensional datasets correspond one-to-one.

[0091] The transformation module 200 is used to map the multidimensional dataset to the hyperspherical coordinate space and convert it into spatial points; the multidimensional spherical coordinates of each spatial point constitute a sample, and all samples constitute a sample set; the multidimensional spherical coordinates include a modulus and several angle values.

[0092] Analysis module 300 is used to perform statistical analysis on the sample set in different dimensions and determine reference points.

[0093] Select module 400, which is used to select target points from spatial points based on multidimensional spherical coordinates of reference points; the multidimensional dataset corresponding to the target points is the target identification result.

[0094] In one implementation, based on the above embodiments, the analysis module 300 is used to divide the sample set into several sub-sample sets; it is also used to traverse all sub-sample sets, select target data from the sub-sample sets according to the distribution of modulus and angle values ​​in the sub-sample sets, and construct the multi-dimensional spherical coordinates of candidate points; and select candidate points that meet preset conditions as reference points. Further, the preset condition is that the points repeat the most within a preset sub-sample set; or, the preset condition is that the points repeat within a preset sub-sample set. The preset sub-sample set range can be a range consisting of all sub-sample sets or a range consisting of some sub-sample sets.

[0095] In one implementation, the above-mentioned conversion module 200 includes:

[0096] The determination unit is used to determine the upper and lower limits of each electrochemical model parameter in the multidimensional dataset based on the operating conditions corresponding to the operating data set of each multidimensional dataset.

[0097] The processing unit is used to normalize each electrochemical model parameter in the multidimensional dataset according to the upper and lower limits, so as to obtain the normalized value of each electrochemical model parameter in the multidimensional dataset; and to generate a new multidimensional dataset based on the normalized value.

[0098] In one implementation, based on any of the foregoing embodiments, the processing apparatus further includes:

[0099] The second acquisition module is used to acquire the sensitivity values ​​of each electrochemical model parameter;

[0100] The sorting module is used to sort each normalized value in the new cube according to the sensitivity value.

[0101] It should be noted that the embodiments of the parameter identification result processing device provided in this application and the embodiments of the parameter identification result processing method provided above are all based on the same inventive concept and can achieve the same technical effect. Therefore, other specific details of the embodiments of the parameter identification result processing device can be referred to the description of the embodiments of the parameter identification result processing method provided above.

[0102] This application also provides a storage medium storing at least one instruction, which is loaded and executed by a processor to implement the operation performed by the processing method corresponding to the above-described parameter identification result. For example, the storage medium may be a read-only memory (ROM), random access memory (RAM), read-only optical disc (CD-ROM), magnetic tape, floppy disk, or optical data storage device. These can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computing device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module.

[0103] This application also provides a storage medium storing at least one instruction, which is loaded and executed by a processor to implement the operation performed in the corresponding embodiment of the capacitance parameter identification method described above. For example, the storage medium can be a read-only memory (ROM), random access memory (RAM), read-only optical disc (CD-ROM), magnetic tape, floppy disk, or optical data storage device. These can be implemented using computer-executable program code, thus allowing them to be stored in a storage device for execution by a computing device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Therefore, this application is not limited to any specific hardware and software combination.

[0104] It should be noted that the above embodiments can be freely combined as needed. The above are only some embodiments of this application. It should be pointed out that for those skilled in the art, several improvements and modifications can be made without departing from the principles of this application, and these improvements and modifications should also be considered within the protection scope of this application.

Claims

1. A method for processing parameter identification results, characterized in that, include: Obtain the original identification results; The original identification results include multidimensional datasets identified under multiple operating conditions, each multidimensional dataset containing multiple different electrochemical model parameters; and the electrochemical model parameters in different multidimensional datasets correspond one-to-one. The multidimensional dataset is mapped to a hyperspherical coordinate space and converted into spatial points; the multidimensional spherical coordinates of each spatial point constitute a sample, and all the samples constitute a sample set; the multidimensional spherical coordinates include a modulus and several angle values; Statistical analysis of the sample set in different dimensions is performed to determine the reference point; the angle value and the magnitude in the multidimensional spherical coordinates are both regarded as a dimension, and statistical analysis is performed on the data of each dimension to calculate the maximum probability density point of the data in each dimension. All the maximum probability density points are combined to form a new multidimensional spherical coordinate, and the reference point is determined in the hyperspherical coordinate space based on the multidimensional spherical coordinate. Based on the multidimensional spherical coordinates of the reference point, a target point is selected from the spatial points; the multidimensional dataset corresponding to the target point is the target identification result.

2. The method for processing parameter identification results according to claim 1, characterized in that, The aforementioned statistical analysis of the sample set across different dimensions to determine reference points includes: The sample set is divided into several sub-sample sets; Traverse all the said sub-sample sets, and select target data from the sub-sample sets according to the distribution of the modulus and the angle values ​​in the sub-sample sets to form the multidimensional spherical coordinates of the candidate points; Candidate points that meet preset conditions are selected as reference points.

3. The method for processing parameter identification results according to claim 2, characterized in that, The preset condition is that the sample is repeated the most times within a preset sub-sample set.

4. The method for processing parameter identification results according to claim 2, characterized in that, The preset condition is repetition within a preset subsample set.

5. The method for processing parameter identification results according to any one of claims 1-4, characterized in that, Before mapping the multidimensional dataset to hypersphere coordinate space and converting it into spatial points, the process includes: Based on the operating condition dataset corresponding to the operating conditions of each of the multidimensional datasets, determine the upper and lower limits of each electrochemical model parameter in the multidimensional dataset; Based on the upper limit and the lower limit, each electrochemical model parameter in the multidimensional dataset is normalized to obtain a normalized value for each electrochemical model parameter in the multidimensional dataset; a new multidimensional dataset is generated based on the normalized values.

6. The method for processing parameter identification results according to claim 5, characterized in that, Also includes: Obtain the sensitivity value for each of the electrochemical model parameters; Sort each of the normalized values ​​in the new multidimensional dataset according to the sensitivity values.

7. A processing device for parameter identification results, characterized in that, include: The first acquisition module is used to acquire the original identification results; The original identification results include multidimensional datasets identified under multiple operating conditions, each multidimensional dataset containing multiple different electrochemical model parameters; and the electrochemical model parameters in different multidimensional datasets correspond one-to-one. The transformation module is used to map the multidimensional dataset to a hyperspherical coordinate space, converting it into spatial points; the multidimensional spherical coordinates of each spatial point constitute a sample, and all the samples constitute a sample set; the multidimensional spherical coordinates include a modulus and several angle values; The analysis module is used to perform statistical analysis on the sample set in different dimensions to determine the reference point; the angle value and the magnitude in the multidimensional spherical coordinates are both regarded as a dimension, and statistical analysis is performed on the data of each dimension to calculate the maximum probability density point of the data in each dimension. All the maximum probability density points are combined to form a new multidimensional spherical coordinate, and the reference point is determined in the hyperspherical coordinate space based on the multidimensional spherical coordinate. The selection module is used to select a target point from the spatial points based on the multidimensional spherical coordinates of the reference point; the multidimensional dataset corresponding to the target point is the target identification result.

8. The processing apparatus for parameter identification results according to claim 7, characterized in that, The analysis module is used to divide the sample set into several sub-sample sets; The analysis module is also used to traverse all the subsample sets, select target data from the subsample sets according to the distribution of the modulus and angle values ​​in the subsample sets, form multidimensional spherical coordinates of candidate points, and select candidate points that meet preset conditions as reference points.

9. The processing apparatus for parameter identification results according to claim 8, characterized in that, The preset condition is that the sample is repeated the most times within a preset sub-sample set.

10. The processing apparatus for parameter identification results according to claim 8, characterized in that, The preset condition is repetition within a preset subsample set.

11. The processing apparatus for parameter identification results according to any one of claims 7-10, characterized in that, The conversion module includes: The determining unit is configured to determine the upper limit and lower limit of each electrochemical model parameter in the multidimensional dataset based on the operating condition dataset corresponding to the operating condition in each multidimensional dataset. The processing unit is configured to normalize each electrochemical model parameter in the multidimensional dataset according to the upper limit value and the lower limit value, to obtain a normalized value for each electrochemical model parameter in the multidimensional dataset; and to generate a new multidimensional dataset based on the normalized value.

12. A storage medium, characterized in that, The storage medium stores at least one instruction, which is loaded and executed by a processor to perform the operation performed by the parameter identification result processing method as described in any one of claims 1 to 6.

Images