Multi-source data fusion method for nonlinear feature modeling

Abstract

The invention discloses a nonlinear feature modeling-oriented multi-source data fusion method. The method comprises the following steps of: reading simulation data and experimental test data of the same experimental working condition point from a database; obtaining experimental test data features, and obtaining simulation test data features; aligning the simulation data according to a total time length relationship between the simulation data and the experimental test data, and dividing the aligned simulation data based on a sampling interval of a sensor to obtain a plurality of data sets; operating data in the data set, calculating a maximum value, a weighted average value and a median, constructing a data sequence, and generating a data matrix containing different features; and combining the data matrixes containing different characteristics, performing system identification on the data matrixes, and selecting an optimal result as a data fusion result. The invention further discloses a multi-source data fusion device for nonlinear feature modeling, corresponding equipment and a storage medium. According to the invention, the accuracy and reliability of data fusion are obviously improved.

Description

Technical Field

[0001] This application relates to the field of data fusion technology, and more specifically, to a multi-source data fusion method for non-linear feature modeling. Background Art

[0002] In today's data-driven era, data fusion, as a key technology, has become the core means to improve system performance and decision-making quality. With the rapid development of information technology, big data and artificial intelligence technologies have gradually become important supports in all walks of life. However, in the face of a large amount of multi-source heterogeneous data, how to effectively extract representative non-linear features has become an urgent problem to be solved. Traditional linear feature extraction methods are often powerless in the face of complex non-linear problems. The current technical challenges mainly focus on the following aspects: First, how to handle data fusion problems in several cases such as different types, different sampling frequencies, and soft (simulation data) / hard (sensor data) heterogeneous data fusion; second, how to ensure a high degree of restoration of the non-linear dynamic performance of the model when the data is incomplete or defective.

[0003] Although existing data fusion methods can handle multi-source data to a certain extent, there are still obvious deficiencies when dealing with complex non-linear features. For example, when traditional methods handle the asynchronous problem of simulation data and sensor data, they often cannot effectively align the data, resulting in the inability to construct a data matrix, which in turn affects the application of advanced algorithms. In addition, existing technologies are difficult to ensure the performance of non-linear feature extraction and model construction when the data is incomplete or defective, and cannot accurately restore the non-linear dynamic characteristics of the system. These problems limit the application of data fusion technology in complex data environments, and there is an urgent need for a new method to overcome these challenges. Summary of the Invention

[0004] In view of at least one defect or improvement requirement of the existing technology, the present invention provides a multi-source data fusion method for non-linear feature modeling, which can solve at least one of the problems existing in the above background art.

[0005] To achieve the above object, according to the first aspect of the present invention, there is provided a multi-source data fusion method for non-linear feature modeling, the method comprising:

[0006] Read simulation data and test data of the same test condition point from the database;

[0007] Obtain the characteristics of the test data, including the total time length LT, sampling interval, test start time, test end time, and spectrum, and obtain the characteristics of the simulation test data, including the total time length LS, minimum simulation interval, maximum simulation interval, maximum data value, maximum data point position, minimum data value, and minimum data point position;

[0008] According to the relationship between the total time lengths of the simulation data and the test data, supplement or truncate the simulation data to achieve data alignment. Divide the aligned simulation data based on the sampling interval of the sensor to obtain multiple data sets in units of time blocks;

[0009] Operate on the data in the multiple data sets in units of time blocks, calculate the maximum value, weighted average value, and median, construct a data sequence, and generate a data matrix containing different features;

[0010] Merge the data matrices containing different features, perform system identification on the data matrix, and select the optimal result as the data fusion result.

[0011] Furthermore, for the multi-source data fusion method for non-linear feature modeling described above, the supplementing or truncating of the simulation data to achieve data alignment specifically includes:

[0012] When the total time length of the simulation data is equal to the total time length of the test data, divide the simulation data according to the sampling interval of the sensing data;

[0013] When the total time length of the simulation data is less than the total time length of the test data, fill the simulation data set in the form of supplementing simulation tests;

[0014] When the total time length of the simulation data is greater than the total time length of the test data, truncate the redundant simulation data.

[0015] Furthermore, for the multi-source data fusion method for non-linear feature modeling described above, after filling the simulation data set in the form of supplementing simulation tests or truncating the redundant simulation data, divide the simulation data according to the sampling interval of the sensing data.

[0016] Furthermore, for the multi-source data fusion method for non-linear feature modeling described above, the operation on the data in the multiple data sets in units of time blocks specifically includes:

[0017] If the number of data points in the time block is greater than or equal to 1, calculate the maximum value, weighted average value, and median, and construct a data sequence;

[0018] If the number of data points in the time block is equal to 0, select the last two data points of the previous step and the first data point of the next step for extrapolation calculation and interpolation calculation, and calculate the average value of the interpolation and extrapolation as the data filling value of this data point.

[0019] Furthermore, for the multi-source data fusion method for non-linear feature modeling described above, the data matrix includes a maximum value filled data matrix, a weighted average value filled data matrix, and a median filled data matrix.

[0020] Furthermore, before reading the simulation data and test data of the same test operating condition point from the database in the above multi-source data fusion method for non-linear feature modeling, the method further includes:

[0021] In the database, tests are respectively carried out on the test bench and the simulation platform according to the experimental design. For the same test operating condition point, sensor data is collected on the test bench as test data, and non-easily measurable data and unmeasurable data are additionally obtained on the simulation platform as simulation data. The simulation data and test data of the same test operating condition point collected are stored in the format of time-data.

[0022] According to the second aspect of the present invention, there is also provided a multi-source data fusion device for non-linear feature modeling, which includes:

[0023] A data reading module, configured to read the simulation data and test data of the same test operating condition point from the database;

[0024] A feature acquisition module, configured to acquire the features of the test data, including the total time length LT, the sampling interval, the test start time, the test end time, and the frequency spectrum, and acquire the features of the simulation test data, including the total time length LS, the minimum simulation interval, the maximum simulation interval, the maximum data value, the maximum data point position, the minimum data value, and the minimum data point position;

[0025] A data partitioning module, configured to supplement or truncate the simulation data according to the relationship between the total time lengths of the simulation data and the test data to achieve data alignment, and partition the aligned simulation data based on the sampling interval of the sensor to obtain multiple data sets in units of time blocks;

[0026] A matrix construction module, configured to operate on the data in the multiple data sets in units of time blocks, calculate the maximum value, the weighted average value, and the median, construct a data sequence, and generate a data matrix including different features;

[0027] A result acquisition module, configured to merge and generate a data matrix including different features, perform system identification on the data matrix, and select the optimal result as the data fusion result.

[0028] According to the third aspect of the present invention, there is also provided a multi-source data fusion device for non-linear feature modeling, which includes at least one processing unit and at least one storage unit. Wherein, the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit is enabled to execute the steps of the method described in any one of the above.

[0029] According to the fourth aspect of the present invention, there is also provided a storage medium storing a computer program executable by a multi-source data fusion device for non-linear feature modeling. When the computer program runs on the multi-source data fusion device for non-linear feature modeling, the multi-source data fusion device for non-linear feature modeling is caused to execute the steps of the method described in any one of the above.

[0030] According to the fifth aspect of the present invention, there is also provided a computer program product including a computer program, characterized in that when the computer program is executed by a processor, it implements the steps of the multi-source data fusion method for non-linear feature modeling as described in any one of the above claims.

[0031] Generally speaking, compared with the prior art by the above technical solution conceived by the present invention, the following beneficial effects can be achieved:

[0032] A multi-source data fusion method for non-linear feature modeling provided by the present invention ensures the synchronization of data in time and space through data alignment, supplementation or truncation operations. At the same time, a variety of features are used to construct a data matrix, which reflects the data feature information from different angles, significantly improving the accuracy and reliability of data fusion. In addition, the optimal data matrix is selected through system identification as the data fusion result, further ensuring the reliability of the data and the prediction accuracy of the model, and providing high-quality data support for subsequent non-linear feature modeling. Finally, this method can effectively handle the fusion problem of multi-source heterogeneous data, improve the modeling and analysis capabilities of complex systems, and significantly enhance the overall performance and decision-making quality of the system. Description of the Drawings

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0034] Figure 1 It is a schematic flowchart of a multi-source data fusion method for non-linear feature modeling provided by an embodiment of the present application. Detailed Embodiments

[0035] In order to make the objectives, technical solutions and advantages of the present invention clearer, the following will further elaborate on the present invention in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0036] In the description, claims and the above - mentioned drawings of this application, the terms "first", "second", "third", etc. are used to distinguish different objects rather than to describe a specific order. In addition, the terms "comprise" and "have" and any variations thereof are intended to cover non - exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0037] Figure 1 As shown in the flowchart of a multi - source data fusion method for non - linear feature modeling provided by an embodiment of this application, Figure 1 As shown, a multi - source data fusion method for non - linear feature modeling provided by an embodiment of this application includes the following steps:

[0038] S1 Read the simulation data and test data of the same test condition point from the database;

[0039] S2 Obtain the characteristics of the test data, including the total time length LT, sampling interval, test start time, test end time and spectrum, and obtain the characteristics of the simulation test data, including the total time length LS, minimum simulation interval, maximum simulation interval, maximum data value, maximum data point position, minimum data value and minimum data point position;

[0040] S3 According to the relationship between the total time lengths of the simulation data and the test data, supplement or truncate the simulation data to achieve data alignment, and divide the aligned simulation data based on the sampling interval of the sensor to obtain multiple data sets in units of time blocks;

[0041] S4 Operate on the data in the multiple data sets in units of time blocks, calculate the maximum value, weighted average value and median, construct a data sequence, and generate a data matrix containing different characteristics;

[0042] S5 Merge the data matrices containing different characteristics, perform system identification on the data matrix, and select the optimal result as the data fusion result.

[0043] Specifically, read the simulation data and test data of the same test condition point from the database. These data are respectively from the test platform (actual operation data collected by sensors) and the simulation platform (generated data that is difficult or impossible to measure). The data is stored in the time - data format for subsequent processing. This data reading method ensures the integrity and consistency of the data, laying a foundation for subsequent data processing and fusion.

[0044] Obtain the characteristics of the experimental test data, including the total time length LT, sampling interval, test start time, test end time, and spectrum, etc. At the same time, obtain the characteristics of the simulation test data, including the total time length LS, minimum simulation interval, maximum simulation interval, maximum data value, maximum data point position, minimum data value, and minimum data point position, etc. These characteristics are used for subsequent data alignment and processing. By extracting data characteristics in detail, the characteristics of the data can be comprehensively understood.

[0045] According to the relationship between the total time lengths of the simulation data and the experimental test data, supplement or truncate the simulation data to achieve data alignment:

[0046] If the total time length of the simulation data is equal to the total time length of the experimental test data, directly divide the simulation data according to the sampling interval of the sensor data, and divide the simulation data into multiple data sets with time blocks as units. If the total time length of the simulation data is less than the total time length of the experimental test data, fill the simulation data set by supplementing the simulation experiment to align the length of the simulation data with the length of the experimental test data. If the total time length of the simulation data is greater than the total time length of the experimental test data, truncate the redundant simulation data and retain the valid values of the simulation data to ensure data alignment.

[0047] Data alignment is a key step in multi-source data fusion. Through the above methods, the asynchronous problem between the simulation data and the experimental test data can be effectively solved, ensuring the synchronization of data in time and space.

[0048] Operate on the data in multiple data sets with time blocks as units. For the data in each time block, calculate parameters such as the maximum value, weighted average value, and median, and construct three groups of data sequences, namely the maximum value sequence, weighted average value sequence, and median sequence. Then, generate data matrices containing different characteristics according to these data sequences, such as the maximum value filling data matrix, weighted average value filling data matrix, and median filling data matrix. These data matrices can reflect the characteristic information of the data from different angles and provide rich data support for subsequent system identification. By extracting multiple characteristics and constructing data matrices, the non-linear characteristics in the data can be effectively captured, improving the quality and accuracy of data fusion.

[0049] Merge the generated data matrices containing different features and perform system identification on these data matrices. The purpose of system identification is to select the optimal result as the data fusion result by analyzing the prediction accuracy of the model. Specifically, by constructing a system dynamic model, parameter estimation and model verification are carried out using input and output data, and finally the data matrix with the highest prediction accuracy is selected as the data fusion result. This process not only ensures the reliability of the data, but also provides high-quality data support for subsequent non-linear feature modeling. System identification is the final link of data fusion, and through model verification and optimization, the overall performance and reliability of the system can be effectively improved.

[0050] A multi-source data fusion method for non-linear feature modeling provided by an embodiment of the present application ensures the synchronization of data in time and space through data alignment, supplementation or truncation operations. At the same time, multiple features are used to construct data matrices, which reflect data feature information from different angles, significantly improving the accuracy and reliability of data fusion. In addition, the optimal data matrix is selected as the data fusion result through system identification, further ensuring the reliability of the data and the prediction accuracy of the model, and providing high-quality data support for subsequent non-linear feature modeling. Finally, this method can effectively handle the fusion problem of multi-source heterogeneous data, improve the modeling and analysis capabilities of complex systems, and significantly enhance the overall performance and decision-making quality of the system.

[0051] Optionally, for the multi-source data fusion method for non-linear feature modeling provided by an embodiment of the present application, the supplementation or truncation of the simulation data to achieve data alignment specifically includes:

[0052] When the total time length of the simulation data is equal to the total time length of the test data, the simulation data is divided according to the sampling interval of the sensing data;

[0053] When the total time length of the simulation data is less than the total time length of the test data, the simulation data set is filled by supplementing the simulation test;

[0054] When the total time length of the simulation data is greater than the total time length of the test data, the redundant simulation data is truncated.

[0055] Optionally, for the multi-source data fusion method for non-linear feature modeling provided by an embodiment of the present application, after filling the simulation data set by supplementing the simulation test or truncating the redundant simulation data, the simulation data is divided according to the sampling interval of the sensing data.

[0056] Specifically, according to the relationship between the total time lengths of the simulation data and the test data, the simulation data is supplemented or truncated to achieve data alignment:

[0057] When the total length of the simulation data time is less than the total length of the test data time, the simulation data set is filled by supplementing the simulation test to make the total length of the simulation data time equal to the total length of the test data time. The supplemented simulation data should maintain the same characteristics and distribution as the original simulation data to ensure data consistency and reliability; when the total length of the simulation data time is greater than the total length of the test data time, the redundant simulation data is truncated, and the part exceeding the total length of the test data time is removed starting from the end of the simulation data until the total length of the simulation data time is equal to the total length of the test data time. During the truncation process, the valid part of the simulation data should be ensured to be retained to avoid losing important information.

[0058] After completing the data alignment operation, the simulation data is divided according to the sampling interval of the sensor data. Obtain the sampling interval of the test data, which is the time interval for the sensor to collect data during the test. According to the sampling interval, the aligned simulation data is divided into multiple time blocks. The length of each time block is equal to the sampling interval. The specific operation is to divide the time axis of the simulation data to ensure that the data within each time block is aligned with the test data in time. For the simulation data within each time block, further processing is performed to extract feature values, such as calculating parameters such as the maximum value, weighted average value, and median, for subsequent data matrix construction and system identification.

[0059] Optionally, for the multi-source data fusion method for non-linear feature modeling provided in the embodiments of the present application, the data in multiple data sets with time blocks as units is operated on, specifically including:

[0060] If the number of data points in the time block is greater than or equal to 1, calculate the maximum value, weighted average value, median, and construct a data sequence;

[0061] If the number of data points in the time block is equal to 0, select the last two data points in the previous step and the first data point in the next step for extrapolation calculation and interpolation calculation, and calculate the average value of the interpolation and extrapolation as the data filling value for this data point.

[0062] Specifically, after aligning the simulation data with the test data, the simulation data is divided into multiple time blocks according to the sampling interval of the sensor data. The length of each time block is equal to the sampling interval, and the number of data points within each time block may be different, specifically depending on the sampling frequency and data integrity of the data.

[0063] Process the data within each time block. If the number of data points within a time block is greater than or equal to 1, calculate the features of the data within that time block. Calculate the maximum value of all data points within that time block. Calculate the weighted average according to the weights of the data points. The weights can be determined based on the reliability, importance, or other factors of the data points. Calculate the median of all data points within that time block. Construct the calculated maximum value, weighted average, and median into three groups of data sequences respectively for subsequent data matrix construction and system identification.

[0064] If the number of data points within a time block is 0, it means there are no data points within that time block, and data filling is required. Select the last two data points of the previous step (i.e., the previous time block) and the first data point of the next step (i.e., the next time block). According to the selected reference data points, calculate the extrapolation value of that time block using the extrapolation method. The extrapolation method can select a suitable mathematical model according to the characteristics of the data, such as linear extrapolation, polynomial extrapolation, etc. Similarly, according to the selected reference data points, calculate the interpolation value of that time block using the interpolation method. The interpolation method can also select a suitable mathematical model according to the characteristics of the data, such as linear interpolation, spline interpolation, etc. Take the average of the calculated extrapolation value and interpolation value as the data filling value for that time block. This filling method can effectively balance the smoothness and information retention of the data, ensuring the integrity and reliability of the data.

[0065] Optionally, in the multi-source data fusion method for non-linear feature modeling provided by the embodiments of the present application, the data matrix includes a maximum value filling data matrix, a weighted average filling data matrix, and a median filling data matrix.

[0066] Specifically, construct the following three data matrices:

[0067] Maximum value filling data matrix: Arrange the maximum value of each time block in chronological order to form a data matrix. Each row of this matrix represents the maximum value of a time block, and each column represents the maximum value sequence of different time blocks.

[0068] Weighted average filling data matrix: Arrange the weighted average of each time block in chronological order to form a data matrix. Each row of this matrix represents the weighted average of a time block, and each column represents the weighted average sequence of different time blocks.

[0069] Median filling data matrix: Arrange the median of each time block in chronological order to form a data matrix. Each row of this matrix represents the median of a time block, and each column represents the median sequence of different time blocks.

[0070] Optionally, before reading the simulation data and test data of the same test operating condition point from the database, the multi-source data fusion method for non-linear feature modeling provided by the embodiments of the present application further includes:

[0071] In the database, tests are carried out on the test bench and the simulation platform respectively according to the experimental design. For the same test operating condition point, sensor data is collected on the test bench as test data, and non-easily measurable data and unmeasurable data are supplemented and obtained on the simulation platform as simulation data. The simulation data and test data of the same test operating condition point collected are stored in the format of time-data.

[0072] Specifically, according to the test requirements, test operating condition points are designed. These operating condition points should cover various typical states of the system operation to comprehensively reflect the dynamic characteristics of the system.

[0073] Tests are carried out on the test bench (actual test platform) and the simulation platform respectively. The test bench test is mainly used to collect sensor data, while the simulation platform is used to supplement and obtain non-easily measurable data and unmeasurable data.

[0074] Actual tests are carried out on the test bench, and test data is collected through sensors. These data include various measurable parameters of the system, such as temperature, pressure, flow rate, etc. The collected data should have high precision and high reliability to ensure the accuracy of subsequent analysis. Simulation tests are carried out on the simulation platform to generate simulation data. The simulation data is used to supplement the parameters that are difficult to measure or cannot be measured in the test bench test, such as internal state variables, outputs of complex physical processes, etc. The simulation data should be based on reliable physical models and algorithms to ensure its consistency with the actual system.

[0075] The collected test data and simulation data are sorted in the format of time-data. The time-data format means that the data is stored in the form of a time series, and each data point corresponds to a specific timestamp. This format is convenient for subsequent data alignment and processing. Ensure that the timestamps of the test data and simulation data have the same precision and time reference so that data alignment can be accurately carried out in subsequent steps.

[0076] The sorted test data and simulation data are stored in the database. The database should have efficient data storage and retrieval capabilities so that the required data can be quickly read in subsequent data processing steps. In the database, independent data records are created for each test operating condition point to ensure the integrity and traceability of the data. Each record should contain information such as the identifier of the test operating condition point, the test time, sensor data, and simulation data.

[0077] The embodiments of the present application also provide a multi-source data fusion device for non-linear feature modeling, including:

[0078] A data reading module, configured to read simulation data and test data of the same test operating condition point from a database;

[0079] A feature acquisition module, configured to acquire test data features, including the total time length LT, sampling interval, test start time, test end time, and spectrum, and acquire simulation test data features, including the total time length LS, minimum simulation interval, maximum simulation interval, maximum data value, maximum data point position, minimum data value, and minimum data point position;

[0080] A data partitioning module, configured to supplement or truncate the simulation data according to the relationship between the total time lengths of the simulation data and the test data to achieve data alignment, and partition the aligned simulation data based on the sampling interval of the sensor to obtain multiple data sets in units of time blocks;

[0081] A matrix construction module, configured to operate on the data in multiple data sets in units of time blocks, calculate the maximum value, weighted average value, and median, construct a data sequence, and generate a data matrix including different features;

[0082] A result acquisition module, configured to merge and generate a data matrix including different features, perform system identification on the data matrix, and select the optimal result as the data fusion result.

[0083] The present application also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented. Among them, the computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical discs, DVDs, CD-ROMs, micro drives, and magneto-optical discs, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nano-systems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0084] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps may be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present application.

[0085] In the above embodiments, the descriptions of the various embodiments have their own focuses. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0086] In several embodiments provided by this application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling, direct coupling, or communication connection between each other can be through some service interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical or other form.

[0087] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0088] In addition, each functional unit in various embodiments of this application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0089] If the above integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. And the aforementioned memory includes: USB flash drive, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk, or optical disc, etc., which can store program codes.

[0090] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program. This program can be stored in a computer-readable memory. The memory can include: flash drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disc, etc.

[0091] The above are only exemplary embodiments of the present disclosure, and the scope of the present disclosure cannot be limited thereby. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure still fall within the scope covered by the present disclosure. Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the specification and practicing the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not described in the present disclosure. The specification and embodiments are only illustrative, and the scope and spirit of the present disclosure are defined by the claims.

[0092] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the various technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0093] It is easy for those skilled in the art to understand that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-source data fusion method for non-linear feature modeling, characterized in that It includes the following steps: Read the simulation data and test data of the same test condition point from the database; Obtain the characteristics of the test data, including the total time length LT, sampling interval, test start time, test end time and spectrum, and obtain the characteristics of the simulation test data, including the total time length LS, minimum simulation interval, maximum simulation interval, maximum data value, maximum data point position, minimum data value and minimum data point position; According to the relationship between the total time lengths of the simulation data and the test data, supplement or truncate the simulation data to achieve data alignment, and divide the aligned simulation data based on the sampling interval of the sensor to obtain multiple data sets in units of time blocks; Operate on the data in the multiple data sets in units of time blocks, calculate the maximum value, weighted average value and median, construct a data sequence, and generate a data matrix containing different characteristics; Merge the data matrices containing different characteristics, perform system identification on the data matrix, and select the optimal result as the data fusion result.

2. The multi-source data fusion method for non-linear feature modeling according to claim 1, wherein The supplementing or truncating of the simulation data to achieve data alignment specifically includes: When the total time length of the simulation data is equal to the total time length of the test data, divide the simulation data according to the sampling interval of the sensing data; When the total time length of the simulation data is less than the total time length of the test data, fill the simulation data set by supplementing the simulation test; When the total time length of the simulation data is greater than the total time length of the test data, truncate the redundant simulation data.

3. The multi-source data fusion method for non-linear feature modeling according to claim 2, wherein After filling the simulation data set by supplementing the simulation test or truncating the redundant simulation data, divide the simulation data according to the sampling interval of the sensing data.

4. The multi-source data fusion method for non-linear feature modeling according to claim 1, wherein The operation on the data in the multiple data sets in units of time blocks specifically includes: If the number of data points in the time block is greater than or equal to 1, calculate the maximum value, weighted average value, median, and construct a data sequence; If the number of data points in the time block is equal to 0, select the last two data points of the previous step and the first data point of the next step for extrapolation calculation and interpolation calculation, and calculate the average value of the interpolation and extrapolation as the data filling value of this data point.

5. The multi-source data fusion method for non-linear feature modeling according to claim 1, characterized in that The data matrix includes a maximum value filling data matrix, a weighted average value filling data matrix, and a median filling data matrix.

6. The multi-source data fusion method for non-linear feature modeling according to claim 1, wherein Before reading the simulation data and test data of the same test condition point from the database, it further includes: In the database, conduct tests on the test bench and the simulation platform respectively according to the test design. For the same test condition point, collect sensor data on the test bench as the test data, and supplement and obtain non-easily measurable data and unmeasurable data on the simulation platform as the simulation data, and store the simulation data and test data of the same test condition point collected in the time-data format.

7. A multi-source data fusion device for non-linear feature modeling, characterized in that, It includes: A data reading module for reading the simulation data and test data of the same test condition point from the database; A feature acquisition module, configured to acquire features of experimental test data, including the total time length LT, sampling interval, test start time, test end time, and spectrum, and acquire features of simulation test data, including the total time length LS, minimum simulation interval, maximum simulation interval, maximum data value, maximum data point position, minimum data value, and minimum data point position; A data division module, configured to supplement or truncate the simulation data according to the relationship between the total time lengths of the simulation data and the experimental test data to achieve data alignment, and divide the aligned simulation data based on the sampling interval of the sensor to obtain multiple data sets in units of time blocks; A matrix construction module, configured to operate on the data in multiple data sets in units of time blocks, calculate the maximum value, weighted average value, and median, construct a data sequence, and generate a data matrix containing different features; A result acquisition module, configured to merge and generate a data matrix containing different features, perform system identification on the data matrix, and select the optimal result as the data fusion result.

8. A multi-source data fusion device for non-linear feature modeling, characterized in that, It includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes the steps of the method according to any one of claims 1 to 6.

9. A storage medium, characterized in that, It stores a computer program executable by a multi-source data fusion device for non-linear feature modeling. When the computer program runs on the multi-source data fusion device for non-linear feature modeling, the multi-source data fusion device for non-linear feature modeling executes the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the multi-source data fusion method for non-linear feature modeling according to any one of claims 1 to 6.

Images