Data recovery method for physical simulation system based on data analysis

Abstract

The present invention relates to the field of data processing technology, and specifically to a data recovery method for a physical simulation system based on data analysis, including: obtaining the data sequence of the physical simulation system in multiple dimensions and the acquisition time corresponding to the data to be recovered; obtaining the updated data segments of two different dimensions in two different time periods; obtaining the cumulative distance matrix of the updated data segments of different dimensions in the same time period in two different time periods and the chain code of the best matching path, obtaining the correlation between each dimension and other dimensions, and the associated dimension of each dimension; obtaining the functional relationship between the data of each dimension and the data of the associated dimension; combining the correlation between each dimension and its associated dimension, obtaining the value of each data to be recovered in each dimension. The present invention aims to solve the problem that part of the data cannot be recovered when recovering the data of a physical simulation system that has failed.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a data recovery method for a physical simulation system based on data analysis. Background Art

[0002] Physical simulation is a simulation environment that uses actual hardware devices to simulate and test and verify the performance, behavior, and fault status of the system. The simulation results obtained in this environment are extremely realistic, but the cost of construction and use is too expensive, the controllability is also poor, and it is difficult to collect data and reproduce faults. Once a fault occurs, it is very likely to damage external devices or even the entire physical environment. Due to the failure of external devices, the physical environment of the entire system changes, which in turn causes system data loss. Therefore, it is necessary to establish better data backup and data recovery methods to ensure that the system simulation data can be fully obtained after a system failure.

[0003] Since the simulation system can generate relatively detailed log records, these logs contain detailed information on key system status information, user operation records and other events. In addition, the system has some backup information. Based on this, the commonly used data recovery method is to restore data through logs, backups and snapshots. However, since there are irregular backups by users and scheduled automatic backups by the system, and most of the log records are about user operations and other related information, system damage may damage some backups, resulting in fragmented data recovery, that is, data of some dimensions at the same acquisition time can be recovered, while data of some dimensions cannot be recovered, resulting in incomplete data in the physical simulation system after data recovery. Summary of the invention

[0004] The present invention provides a data recovery method for a physical simulation system based on data analysis to solve the existing problems.

[0005] The data recovery method for the physical simulation system based on data analysis of the present invention adopts the following technical scheme:

[0006] The present invention proposes a data recovery method for a physical simulation system based on data analysis, the method comprising the following steps:

[0007] Obtaining the data sequences of the physical simulation system in multiple dimensions after the failure and the acquisition time corresponding to the data to be restored;

[0008] The data in the data sequence of each dimension is standardized and segmented to obtain several standard data segments of each dimension, and the standard data segment corresponds to a time period; according to the distribution of the acquisition time of the data to be restored in two different dimensions in two different time periods in the standard data segment, the updated data segments of two different dimensions in two different time periods are obtained; the DTW algorithm is used to obtain the cumulative distance matrix of the updated data segments of different dimensions in the same time period and the chain code of the best matching path; according to the distribution of the cumulative distance in the cumulative distance matrix of the updated data segments of different dimensions in the same time period under two different time periods, and the DTW distance of the chain code of the best matching path, the correlation of the two different dimensions in the two different time periods is obtained, and then the correlation of each dimension with other dimensions is obtained; according to the correlation of each dimension with other dimensions, the associated dimension of each dimension is obtained;

[0009] Based on the data sequence of each dimension and the data sequence of its associated dimensions, the data function relationship between each dimension and each of its associated dimensions is obtained; combined with the data of the associated dimensions corresponding to the collection time of the data to be recovered in each dimension, and the correlation between each dimension and its associated dimensions, the value of each data to be recovered in each dimension is obtained.

[0010] The specific method of obtaining the data sequence of multiple dimensions of the physical simulation system after the failure includes:

[0011] The voltage data sequence, current data sequence, speed data sequence of each gear, vibration frequency data sequence and motor temperature data sequence of the AC / DC reduction motor with fault are obtained from the physical simulation system of the AC / DC reduction motor with fault; the data collection interval of each dimension is seconds, the collection time is sky.

[0012] The data in the data sequence of each dimension is standardized and segmented to obtain a number of standard data segments of each dimension, including the specific method of:

[0013] The values ​​of all data in each dimension of the data sequence are calculated separately. Score standardization, obtaining standard data for each data in the data sequence of each dimension, arranging the standard data for each data in the data sequence of the same dimension according to the order of the data in the data sequence, to obtain a standard sequence of data in each dimension; each standard data in the standard sequence of data in each dimension corresponds to a collection time;

[0014] The collection time is divided into several time periods of equal length, and the data standard sequence of each dimension is divided into several standard data segments according to the divided time periods and the collection time corresponding to the standard data in the data standard sequence of each dimension.

[0015] The method of obtaining updated data segments of two different dimensions in two different time periods according to the distribution of the collection time of the data to be restored of two different dimensions in two different time periods in the standard data segment includes the following specific methods:

[0016] The first Dimensions and The dimension Standard data segment and The order value of the acquisition time corresponding to the data to be restored in the first standard data segment in the standard data segment is recorded as the order value to be removed; Dimensions and The dimension Standard data segment and In the standard data segments, remove the standard data at the collection time corresponding to the order value to be removed, and get the The dimension The update data segment and Update data segment, and The dimension The update data segment and Update data segment.

[0017] The method of obtaining the correlation between two different dimensions in two different time periods based on the distribution of the cumulative distance in the cumulative distance matrix of the updated data segments of different dimensions in the same time period under two different time periods and the DTW distance of the chain code of the best matching path includes:

[0018]

[0019] In the formula, Indicates The first The dimension in The first The relevance of time periods, Indicates Dimensions and The dimension Update data segment and The number of rows or columns contained in the cumulative distance matrix of each updated data segment, Indicates Dimensions and The dimension The cumulative distance matrix of the updated data segment Line The cumulative distance value of the column, Indicates Dimensions and The dimension The cumulative distance matrix of the updated data segment Line The cumulative distance value of the column, Indicates Dimensions and The dimension The chain code of the best matching path of the updated data segment is the same as the Dimensions and The dimension The DTW distance of the chain code of the best matching path of the updated data segment, represents the hyperbolic tangent function, is an exponential function with a natural constant as base, represents the absolute value function.

[0020] The specific method of obtaining the correlation between each dimension and other dimensions includes:

[0021] The first The first The mean of the correlation between the dimensions in any two time periods is denoted as Dimensions and The correlation of dimensions.

[0022] The method of obtaining the associated dimension of each dimension according to the correlation between each dimension and other dimensions includes:

[0023] Preset relevance threshold , let the two dimensions corresponding to the correlation greater than the correlation threshold be mutually correlated dimensions.

[0024] The data function relationship between each dimension and each associated dimension is obtained based on the data sequence of each dimension and the data sequence of the associated dimension, including the specific method of:

[0025] Based on The data series of the dimension The dimension The data sequence of the associated dimension The collection time The data of the first dimension The dimension The data of the associated dimension is used as the The elements corresponding to the acquisition time are mapped to the horizontal axis of the The dimension The data of the associated dimension, the vertical axis is In the two-dimensional coordinate system of the data of the dimension, several data points in the two-dimensional coordinate system are obtained, and several data points in the two-dimensional coordinate system are fitted to obtain the first Dimensions and The dimension The data function relationship of the associated dimensions.

[0026] The method of combining the data of the associated dimension corresponding to the collection time of the data to be restored in each dimension and the correlation between each dimension and its associated dimension to obtain the value of each data to be restored in each dimension includes the following specific methods:

[0027] The first The dimension The collection time corresponding to the data to be restored is recorded as The dimension The recovery time will be The data at the collection time corresponding to the time to be restored is not the data to be restored. The associated dimension of the dimension is denoted as The dimension A comparison dimension on the time to be restored;

[0028] According to The recovery time The dimension The data of the comparison dimension are Dimensions and The dimension The data function relationship of the comparison dimension is obtained The dimension The recovery time The estimated value of the comparison dimension; combined with the The dimension The correlation between the time to be restored and each control dimension is obtained. The dimension The value of the data to be restored.

[0029] The combination The dimension The correlation between the time to be restored and each control dimension is obtained. The dimension The value of the data to be recovered, including the specific method:

[0030]

[0031] In the formula, Indicates The dimension The value of the data to be restored. Indicates The dimension The number of reference dimensions at the time to be restored, Indicates The dimension The recovery time Dimension and its The correlation of the comparison dimensions, Indicates The dimension The recovery time The estimated value of the comparison dimension, represents the weight normalization function.

[0032] The beneficial effect of the technical solution of the present invention is that when restoring data in a physical simulation system that has failed, data of different dimensions can be restored. Score standardization unifies the dimensions of data sequences of different dimensions, thereby improving the quantitative accuracy of the cumulative distance matrix and the best matching path for the data change relationship in the same time period of different dimensions in the subsequent DTW calculation process in the same time period of different dimensions; by analyzing the similarity of the cumulative distance matrix and the best matching path of two different dimensions in two time periods, the correlation of the two dimensions in two time periods is obtained, avoiding the impact of the missing data to be recovered on the quantification of the similarity of the overall change trend of the data, and finding the associated dimensions of each dimension based on the correlation, thereby providing a basis for the estimation and completion of the data to be recovered; by fitting and quantifying the functional relationship between each dimension and the data of its associated dimension, the data of the associated dimension is estimated and the value of the data to be recovered is obtained at the corresponding collection time of the data to be recovered in each dimension. In the comprehensive quantification process, the correlation between the associated dimension and each dimension is considered at the same time, thereby ensuring that the data to be recovered follows the data change trend of the corresponding dimension, while ensuring that the data change trend between each dimension and its associated dimension is similar, ultimately improving the accuracy of data recovery of the physical simulation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 The present invention is a flowchart of the steps of a data recovery method for a physical simulation system based on data analysis. DETAILED DESCRIPTION

[0035] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the data recovery method for the physical simulation system based on data analysis proposed by the present invention, its specific implementation method, structure, characteristics and effects are described in detail as follows in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0036] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0037] The specific scheme of the data recovery method for the physical simulation system based on data analysis provided by the present invention is described in detail below with reference to the accompanying drawings.

[0038] See also Figure 1 , which shows a flow chart of a data recovery method for a physical simulation system based on data analysis provided by an embodiment of the present invention, the method comprising the following steps:

[0039] Step S001: obtaining data sequences in multiple dimensions of a physical simulation system after a failure and acquisition times corresponding to data to be restored.

[0040] This embodiment improves the method of recovering lost data after a physical simulation system fails, thereby obtaining data sequences of various dimensions in the database of the physical simulation system after the failure. This embodiment takes the physical simulation system of AC and DC reduction motors as an example, and other physical simulation systems can be used as examples in other implementations.

[0041] Specifically, when restoring the data in the physical simulation system of the AC / DC reduction motor that has failed, the data of each dimension in the AC / DC reduction motor physical simulation system is restored based on the log records, backup information and snapshot information of the physical simulation system of the AC / DC reduction motor, and the voltage data sequence, current data sequence, speed data sequence of each gear, vibration frequency data sequence and motor temperature data sequence of the AC / DC reduction motor are obtained from the physical simulation system of the AC / DC reduction motor that has failed. The data collection interval of each dimension is seconds, the collection time is Every data in each dimension of the data sequence corresponds to a collection time. The collection interval preset in this implementation is , , this is used as an example for description, and other values ​​may be set in other implementation modes.

[0042] Step S002: standardize and segment the data in the data sequence of each dimension to obtain several standard data segments of each dimension; obtain the updated data segments of the two different dimensions in the two different time periods according to the distribution of the collection time of the data to be restored in the standard data segments in the two different time periods; use the DTW algorithm to obtain the cumulative distance matrix of the updated data segments of different dimensions in the same time period and the chain code of the best matching path; obtain the correlation of the two different dimensions in the two different time periods according to the distribution of the cumulative distance in the cumulative distance matrix of the updated data segments of different dimensions in the same time period under the two different time periods, and the DTW distance of the chain code of the best matching path, and then obtain the correlation of each dimension with other dimensions; obtain the associated dimension of each dimension according to the correlation of each dimension with other dimensions.

[0043] It should be noted that, among the data of multiple dimensions of the physical simulation system, the data of some dimensions are affected by the data of other dimensions, or the data of some dimensions are affected by the data of the same dimension, so that the value of the data of a certain dimension is strongly correlated with the value of the data of some dimensions. Therefore, when restoring the data of a certain dimension, the data of the dimension can be obtained based on the data of other dimensions. Therefore, when restoring data of the physical simulation system, the associated dimensions of each dimension are first obtained.

[0044] It should be further explained that when obtaining the associated dimension of each dimension, since the units and variation ranges of the data of each dimension are different from those of the data of other dimensions, it is too complicated and computationally intensive to directly obtain the associated dimension of each dimension based on the data sequence of each dimension. Therefore, the data of each dimension is standardized to obtain the standard value of each data of each dimension, and then the standard sequence of data of each dimension is obtained to obtain the associated dimension of each dimension.

[0045] It should be further explained that the DTW algorithm is used to calculate the similarity between two sequences. When calculating the similarity between two sequences, it first obtains the cumulative distance matrix and the best matching path of the two sequences. The best matching path of the two sequences reflects the data correspondence between the two sequences. The data of the first dimension If the data of the first dimension have a strong correlation, then The first The data segment of the dimension The similarity between the data segments in the first dimension is The first The data segment of the dimension The similarity between the data segments of the dimensions is relatively close, that is, The next time period The data segment of the dimension The cumulative distance matrix of the data segments of the dimension, the best matching path and the The next time period The data segment of the dimension The cumulative distance matrix and the best matching path of the data segments of the dimensions are relatively similar. Therefore, when obtaining the associated dimension of each dimension, the associated dimension of each dimension is obtained according to the similarity of the cumulative distance matrix and the closest matching path of the data standard segments of the two dimensions in different time periods. Therefore, the data standard sequence of each dimension is segmented to obtain several standard data segments.

[0046] It should be further explained that since the best matching path obtained by the DTW algorithm is a line segment, the similarity between two line segments cannot be calculated by the DTW algorithm. Since the chain code of a line segment reflects the position change relationship of the pixels within the line segment, the similarity between the chain code of each best matching path and the chain code of other best matching paths is used instead of the similarity between each best matching path and other best matching paths.

[0047] Specifically, the values ​​of all data in the data sequence of each dimension are calculated. Score standardization is used to obtain the standard data of each data sequence in each dimension. The standard data of each data sequence in the same dimension is arranged according to the order of the data in the data sequence to obtain the standard data sequence of each dimension. Among them, each standard data in the standard data sequence of each dimension corresponds to a collection time. Among them, the data in a data set are normalized. Score normalization is a known technique and will not be described in detail in this embodiment.

[0048] Further, the acquisition duration is divided into several time periods of equal length. In this embodiment, the acquisition duration is divided into 96 time periods, and the duration of each time period is 1 hour. Other values ​​may be set in other implementations. According to the divided time periods and the acquisition time corresponding to the standard data in the data standard sequence of each dimension, the data standard sequence of each dimension is divided into several standard data segments.

[0049] Furthermore, Dimensions and The dimension Standard data segment and The order value of the acquisition time corresponding to the data to be restored in the first standard data segment in the standard data segment is recorded as the order value to be removed. Dimensions and The dimension Standard data segment and In the standard data segments, remove the standard data at the collection time corresponding to the order value to be removed, and get the The dimension The update data segment and Update data segment, and The dimension The update data segment and Update data segment.

[0050] Further, using the DTW algorithm, we obtain The dimension The update data segment and The dimension The cumulative distance matrix and the best matching path of the updated data segment are obtained. The dimension The update data segment and The dimension The chain code of the best matching path for the updated data segment is obtained using the DTW algorithm. The dimension Update data segment and The dimension The chain code of the best matching path of the updated data segment and the The dimension Update data segment and The dimension The DTW distance of the chain code of the best matching path of the updated data segment. Among them, obtaining the best matching path and the cumulative distance matrix through the DTW algorithm is an existing public technology, obtaining the chain code of a line segment is an existing public technology, and obtaining the DTW distance of two sequences is an existing public technology, which will not be repeated in this embodiment.

[0051] Further, get the The first The dimension in The first The specific calculation formula for the correlation of time periods is as follows:

[0052]

[0053] In the formula, Indicates The first The dimension in The first The relevance of time periods, Indicates Dimensions and The dimension Update data segment and The number of rows or columns contained in the cumulative distance matrix of each updated data segment, Indicates Dimensions and The dimension The cumulative distance matrix of the updated data segment Line The cumulative distance value of the column, Indicates Dimensions and The dimension The cumulative distance matrix of the updated data segment Line The cumulative distance value of the column, Indicates Dimensions and The dimension The chain code of the best matching path of the updated data segment is the same as the Dimensions and The dimension The DTW distance of the chain code of the best matching path of the updated data segment, represents the hyperbolic tangent function, which is used for normalization processing in this embodiment. represents the absolute value function, is an exponential function with a natural constant as the base. The model shows an inverse proportional relationship. As the input of the model, the implementer can set the inverse proportional function according to the actual situation.

[0054] It should be noted that The larger the value of The cumulative distance matrix of the updated data segment is The more similar the cumulative distance values ​​of the same position in the cumulative distance matrix of the updated data segment are, the more similar the cumulative distance matrix is. Dimensions and The dimension in Update data segment and The stronger the correlation is in the time period corresponding to the updated data segment, The larger the value of; The smaller the value of The chain code of the best matching path of the updated data segment is the same as the chain code of the The stronger the similarity of the chain codes of the best matching paths of the updated data segments, the more similar the best matching paths are. The first The data of the dimension is Update data segment and The stronger the correlation is in the time period corresponding to the updated data segment.

[0055] Furthermore, The first The mean of the correlation between the dimensions in any two time periods is denoted as Dimensions and Preset correlation threshold , let the two dimensions corresponding to the correlation greater than the correlation threshold be the correlation dimensions. Among them, the preset correlation threshold , this is used as an example for description, and other values ​​may be set in other implementation modes.

[0056] At this point, the associated dimension of each dimension is obtained.

[0057] Step S003: Based on the data sequence of each dimension and the data sequence of its associated dimension, obtain the data function relationship between each dimension and each of its associated dimensions; combine the data of the associated dimension corresponding to the acquisition time of the data to be recovered in each dimension, and the correlation between each dimension and its associated dimensions, to obtain the value of each data to be recovered in each dimension.

[0058] It should be noted that when two dimensions are related to each other, it means that the data changes of these two dimensions have a certain correlation, that is, if one dimension is collected at the first The value of the data in the dimension is , No. The dimension The value of the data of the associated dimension is ; Then when other collection times are The value of the data in the dimension is , then The dimension The value of the data in the associated dimension is Similar; when other collection times are The dimension The value of the data of the associated dimension is , then The value of the data in each dimension is Therefore, according to The data to be restored in the data sequence of the dimension is collected at the first The data of the associated dimension of the dimension is The data to be recovered in the dimension is recovered. A dimension may have multiple associated dimensions, so according to the The correlation between a dimension and its associated dimension gives The recovery result of the data of the associated dimension of the dimension is a weight, and the The value of the data to be restored in the data sequence of dimensions.

[0059] It should be further explained that, according to The data of the associated dimension of the dimension is When recovering the data to be recovered in the data sequence of dimensions, The data to be restored in the data sequence of the dimension is collected at the first The value of the data of the associated dimension of the dimension may appear for the first time, so it may not be possible to directly The data to be restored in the data sequence of the dimension is collected at the first The value of the data of the associated dimension of the dimension is restored, so The data of the dimension is fitted with the data series of its associated dimension to obtain the The relationship between the data of the dimension and its associated dimension is obtained The value of the data to be restored in the data sequence of dimensions.

[0060] Specifically, based on The data series of the dimension The dimension The data sequence of the associated dimension The collection time The data of the first dimension The dimension The data of the associated dimension is used as the It should be noted that since the data of a certain dimension at the collection time is to be restored, that is, the data of this dimension at the collection time does not exist, the element corresponding to the collection time is incomplete and cannot be mapped to the coordinate system later; all elements corresponding to the collection time are mapped to the horizontal coordinate of the The dimension The data of the associated dimension, the vertical axis is In the two-dimensional coordinate system of the data of the dimension, several data points in the two-dimensional coordinate system are obtained, and the curve fitting is performed on the several data points in the two-dimensional coordinate system by the least squares method to obtain the first Dimensions and The dimension The data function relationship of the associated dimensions. Among them, the least square method is used to perform curve fitting on a number of data points in the two-dimensional coordinate system, which is a well-known technology and will not be described in detail in this embodiment.

[0061] So far, we have obtained Dimensions and The data function relationship of each associated dimension.

[0062] Furthermore, The dimension The collection time corresponding to the data to be restored is recorded as The dimension The recovery time will be The data at the collection time corresponding to the time to be restored is not the data to be restored. The associated dimension of the dimension is denoted as The dimension The comparison dimension of the time to be restored. The recovery time The dimension The data of the comparison dimension are Dimensions and The dimension The data function relationship of the comparison dimension is obtained The dimension The recovery time Estimated value of the comparison dimension.

[0063] Further, get the The dimension The specific calculation formula for the value of the data to be restored is as follows:

[0064]

[0065] In the formula, Indicates The dimension The value of the data to be restored. Indicates The dimension The number of reference dimensions at the time to be restored, Indicates The dimension The recovery time Dimension and its The correlation of the comparison dimensions, Indicates The dimension The recovery time The estimated value of the comparison dimension, Represents the weight normalization function, the normalized object is Dimensions and The collection time The correlation of each comparison dimension with the dimensions.

[0066] At this point, the values ​​of all the data to be restored in the data sequence of each dimension are obtained, and the data in the data sequence after the failure is restored.

[0067] At this point, this embodiment is completed.

[0068] To sum up, in the embodiments of the present invention, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A data recovery method for a physical simulation system based on data analysis, characterized in that: The method comprises the following steps: Obtaining the data sequences of the physical simulation system in multiple dimensions after the failure and the acquisition time corresponding to the data to be restored; The data in the data sequence of each dimension is standardized and segmented to obtain several standard data segments of each dimension, and the standard data segment corresponds to a time period; according to the distribution of the acquisition time of the data to be restored in two different dimensions in two different time periods in the standard data segment, the updated data segments of two different dimensions in two different time periods are obtained; the DTW algorithm is used to obtain the cumulative distance matrix of the updated data segments of different dimensions in the same time period and the chain code of the best matching path; according to the distribution of the cumulative distance in the cumulative distance matrix of the updated data segments of different dimensions in the same time period in two different time periods, and the DTW distance of the chain code of the best matching path, the correlation of the two different dimensions in the two different time periods is obtained, and then the correlation of each dimension with other dimensions is obtained; according to the correlation of each dimension with other dimensions, the associated dimension of each dimension is obtained; Based on the data sequence of each dimension and the data sequence of its associated dimension, the data function relationship between each dimension and each of its associated dimensions is obtained; combining the data of the associated dimension corresponding to the acquisition time of the data to be restored in each dimension, and the correlation between each dimension and its associated dimension, the value of each data to be restored in each dimension is obtained; The method of obtaining updated data segments of two different dimensions in two different time periods according to the distribution of the collection time of the data to be recovered of two different dimensions in two different time periods in the standard data segment includes the following specific methods: The first Dimensions and The dimension Standard data segment and The order value of the acquisition time corresponding to the data to be restored in the first standard data segment in the standard data segment is recorded as the order value to be removed; Dimensions and The dimension Standard data segment and In the standard data segments, remove the standard data at the collection time corresponding to the order value to be removed, and get the The dimension The update data segment and Update data segment, and The dimension The update data segment and Update data segment; The method of obtaining the correlation between two different dimensions in two different time periods based on the distribution of the cumulative distance in the cumulative distance matrix of the updated data segments of different dimensions in the same time period under two different time periods and the DTW distance of the chain code of the best matching path includes: In the formula, Indicates The first The dimension in The first The relevance of time periods, Indicates Dimensions and The dimension Update data segment and The number of rows or columns contained in the cumulative distance matrix of each updated data segment, Indicates Dimensions and The dimension The cumulative distance matrix of the updated data segment Line The cumulative distance value of the column, Indicates Dimensions and The dimension The cumulative distance matrix of the updated data segment Line The cumulative distance value of the column, Indicates Dimensions and The dimension The chain code of the best matching path of the updated data segment is the same as the Dimensions and The dimension The DTW distance of the chain code of the best matching path of the updated data segment, represents the hyperbolic tangent function, is an exponential function with a natural constant as base, represents the absolute value function; The specific method of obtaining the correlation between each dimension and other dimensions includes: The first The first The mean of the correlation between the dimensions in any two time periods is denoted as Dimensions and The correlation of dimensions.

2. The data recovery method for a physical simulation system based on data analysis according to claim 1, characterized in that: The specific method of obtaining the data sequence of multiple dimensions of the physical simulation system after the failure includes: The voltage data sequence, current data sequence, speed data sequence of each gear, vibration frequency data sequence and motor temperature data sequence of the AC / DC reduction motor with fault are obtained from the physical simulation system of the AC / DC reduction motor with fault; the data collection interval of each dimension is seconds, the collection time is sky.

3. The data recovery method for a physical simulation system based on data analysis according to claim 2, characterized in that: The data in the data sequence of each dimension is standardized and segmented to obtain a number of standard data segments of each dimension, including the specific method of: The values ​​of all data in each dimension of the data sequence are calculated separately. Score standardization, obtaining standard data for each data in the data sequence of each dimension, arranging the standard data for each data in the data sequence of the same dimension according to the order of the data in the data sequence, to obtain a standard sequence of data in each dimension; each standard data in the standard sequence of data in each dimension corresponds to a collection time; The collection time is divided into several time periods of equal length, and the data standard sequence of each dimension is divided into several standard data segments according to the divided time periods and the collection time corresponding to the standard data in the data standard sequence of each dimension.

4. The data recovery method for a physical simulation system based on data analysis according to claim 1, characterized in that: The method of obtaining the associated dimension of each dimension according to the correlation between each dimension and other dimensions includes: Preset relevance threshold , let the two dimensions corresponding to the correlation greater than the correlation threshold be mutually correlated dimensions.

5. The data recovery method for a physical simulation system based on data analysis according to claim 1, characterized in that: The data function relationship between each dimension and each associated dimension is obtained based on the data sequence of each dimension and the data sequence of the associated dimension, including the specific method of: Based on The data series of the dimension The dimension The data sequence of the associated dimension The collection time The data of the first dimension The dimension The data of the associated dimension is used as the The elements corresponding to the acquisition time are mapped to the horizontal axis of the The dimension The data of the associated dimension, the vertical axis is In the two-dimensional coordinate system of the data of the dimension, several data points in the two-dimensional coordinate system are obtained, and several data points in the two-dimensional coordinate system are fitted to obtain the first Dimensions and The dimension The data function relationship of the associated dimensions.

6. The data recovery method for a physical simulation system based on data analysis according to claim 1, characterized in that: The method of combining the data of the associated dimension corresponding to the collection time of the data to be restored in each dimension and the correlation between each dimension and its associated dimension to obtain the value of each data to be restored in each dimension includes the following specific methods: The first The dimension The collection time corresponding to the data to be restored is recorded as The dimension The recovery time will be The data at the collection time corresponding to the time to be restored is not the data to be restored. The associated dimension of the dimension is denoted as The dimension A comparison dimension of time to be restored; According to The recovery time The dimension The data of the comparison dimension are Dimensions and The dimension The data function relationship of the comparison dimension is obtained The dimension The recovery time The estimated value of the comparison dimension; combined with The dimension The correlation between the time to be restored and each control dimension is obtained. The dimension The value of the data to be restored.

7. The data recovery method for a physical simulation system based on data analysis according to claim 6, characterized in that: The combination The dimension The correlation between the time to be restored and each control dimension is obtained. The dimension The value of the data to be recovered, including the specific method: In the formula, Indicates The dimension The value of the data to be restored. Indicates The dimension The number of reference dimensions at the time to be restored, Indicates The dimension The recovery time Dimension and its The correlation of the comparison dimensions, Indicates The dimension The recovery time The estimated value of the comparison dimension, represents the weight normalization function.

Images