Data fault detection method, electronic equipment, storage medium and program product

Abstract

The invention discloses a data fault detection method, electronic equipment, a storage medium and a program product. The method comprises the following steps: acquiring a cache data set generated by a plurality of isomorphic devices in a preset time period; feature extraction is carried out on the multiple pieces of cache data contained in the different cache data sets, and feature data sequences corresponding to the different isomorphic devices are obtained; based on the feature data sequence, reference feature parameters corresponding to the plurality of isomorphic devices are determined, and the reference feature parameters are used for representing feature parameters of the plurality of cache data when the plurality of isomorphic devices operate normally; and performing fault detection on the different isomorphic devices based on the reference feature parameters and the feature data sequence to obtain fault detection results of the different isomorphic devices. According to the invention, the technical problem of low accuracy of fault detection of the isomorphic equipment of which historical data is difficult to obtain in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of fault detection, and in particular to a data fault detection method, electronic equipment, storage medium and program product. Background Art

[0002] At present, for some homogeneous equipment without historical data records, most fault diagnosis methods that rely on long-term historical data face significant limitations. Some industrial enterprises, especially when the equipment is not the main production link, often do not retain detailed equipment operation history data due to cost considerations or storage resource limitations. In this case, traditional methods of fault detection for homogeneous equipment, such as expert systems, fault tree analysis, statistical models, neural network models, etc., may not be able to guarantee the accuracy of fault detection results during real-time fault detection due to lack of necessary knowledge updates and lack of training data.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] Embodiments of the present invention provide a data fault detection method, an electronic device, a storage medium, and a program product to at least solve the technical problem in the related art of low accuracy in fault detection for homogeneous devices for which historical data is difficult to obtain.

[0005] According to one aspect of an embodiment of the present invention, a data fault detection method is provided, including: obtaining a cache data set generated by multiple homogeneous devices in a preset time period; performing feature extraction on multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different homogeneous devices; based on the feature data sequence, determining baseline feature parameters corresponding to the multiple homogeneous devices, wherein the baseline feature parameters are used to characterize the feature parameters of the multiple cache data when the multiple homogeneous devices are operating normally; performing fault detection on different homogeneous devices based on the baseline feature parameters and the feature data sequence to obtain fault detection results for different homogeneous devices.

[0006] Furthermore, the characteristic data sequence is used to characterize the data change trend of multiple cache data; feature extraction is performed on multiple cache data contained in different cache data sets to obtain characteristic data sequences corresponding to different isomorphic devices, including: smoothing filtering is performed on multiple cache data to obtain filtered data; based on the filtered data, a data change curve corresponding to the isomorphic device is constructed; feature extraction is performed on the data change curve to obtain the trend turning point corresponding to the data change curve; based on the trend turning point, a characteristic data sequence corresponding to the isomorphic device is constructed.

[0007] Furthermore, the baseline characteristic parameters are used to characterize the data change trend of multiple cached data under normal circumstances; based on the characteristic data sequence, the baseline characteristic parameters corresponding to multiple isomorphic devices are determined, including: respectively performing correlation detection on the characteristic data sequences corresponding to different isomorphic devices and the characteristic data sequences corresponding to other isomorphic devices in the multiple isomorphic devices to obtain correlation detection results corresponding to different isomorphic devices; based on the correlation detection results, evaluating the characteristic data sequences corresponding to different isomorphic devices to obtain sequence evaluation results; based on the sequence evaluation results corresponding to different isomorphic devices, determining the baseline characteristic parameters from the characteristic data sequences corresponding to different isomorphic devices.

[0008] Furthermore, the fault detection result is used to characterize whether there is an abnormality in the change trend of multiple cached data; fault detection is performed on different isomorphic devices based on baseline characteristic parameters and characteristic data sequences to obtain fault detection results of different isomorphic devices, including: matching the characteristic data sequence with the baseline characteristic parameters to obtain the degree of deviation between the characteristic data sequence and the baseline characteristic parameters; in response to the degree of deviation being greater than a preset threshold, determining that the fault detection result is whether there is an abnormality in the change trend of multiple cached data; in response to the degree of deviation being less than or equal to a preset threshold, determining that the fault detection result is that the change trend of multiple cached data is normal.

[0009] Furthermore, feature extraction is performed on multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different isomorphic devices, including: based on a preset time window, determining multiple window data from multiple cache data, and aggregating the multiple window data corresponding to multiple isomorphic devices to obtain multiple summary data; obtaining the number of devices of the multiple isomorphic devices; in response to the number of devices being less than or equal to a preset value, sorting the multiple summary data to obtain a data sorting result, and constructing a feature data sequence based on the Euclidean distance between adjacent summary data in the data sorting result; in response to the number of devices being greater than a preset value, constructing a feature data sequence based on the multiple summary data.

[0010] Furthermore, based on the characteristic data sequence, the benchmark characteristic parameters corresponding to the multiple isomorphic devices are determined, including: in response to the number of devices being less than or equal to a preset value, the benchmark characteristic parameters are obtained based on the average value of the Euclidean distance contained in the Euclidean distance sequence; in response to the number of devices being greater than a preset value, the data average value is obtained based on the average value of multiple summarized data, and the data standard deviation is obtained based on the standard deviation of the multiple summarized data, and the benchmark characteristic parameters are constructed based on the data average value and the data standard deviation.

[0011] Furthermore, the fault detection result is used to characterize whether there is an abnormality in the values ​​of multiple cache data corresponding to different isomorphic devices; fault detection is performed on different isomorphic devices based on baseline feature parameters and feature data sequences to obtain fault detection results of different isomorphic devices, including: in response to the number of devices being less than or equal to a preset value, clustering multiple summary data based on the average Euclidean distance to obtain a data clustering set, and determining the fault detection result based on the number of data clustering sets and the number of devices; in response to the number of devices being greater than a preset value, matching multiple summary data with a numerical interval to obtain a numerical matching result, and determining the fault detection result based on the numerical matching result.

[0012] According to one aspect of an embodiment of the present invention, a data fault detection device is also provided, including: a set acquisition module, used to obtain a cache data set generated by multiple homogeneous devices in a preset time period; a feature extraction module, used to extract features of multiple cache data contained in different cache data sets, and obtain feature data sequences corresponding to different homogeneous devices; a parameter determination module, used to determine baseline feature parameters corresponding to multiple homogeneous devices based on the feature data sequence, wherein the baseline feature parameters are used to characterize the feature parameters of multiple cache data when the multiple homogeneous devices are operating normally; a fault detection module, used to perform fault detection on different homogeneous devices based on the baseline feature parameters and the feature data sequence, and obtain fault detection results for different homogeneous devices.

[0013] Furthermore, the feature data sequence is used to characterize the data change trend of multiple cache data; the feature extraction module is also used to: perform smoothing filtering on the multiple cache data to obtain filtered data; based on the filtered data, construct a data change curve corresponding to the isomorphic device; perform feature extraction on the data change curve to obtain the trend turning point corresponding to the data change curve; based on the trend turning point, construct a feature data sequence corresponding to the isomorphic device.

[0014] Furthermore, the baseline characteristic parameters are used to characterize the data change trend of multiple cached data under normal circumstances; the parameter determination module is also used to: perform correlation detection on the characteristic data sequences corresponding to different isomorphic devices and the characteristic data sequences corresponding to other isomorphic devices in the multiple isomorphic devices, and obtain the correlation detection results corresponding to the different isomorphic devices; based on the correlation detection results, evaluate the characteristic data sequences corresponding to the different isomorphic devices to obtain sequence evaluation results; based on the sequence evaluation results corresponding to the different isomorphic devices, determine the baseline characteristic parameters from the characteristic data sequences corresponding to the different isomorphic devices.

[0015] Furthermore, the fault detection result is used to characterize whether there is an abnormality in the changing trend of multiple cached data; the fault detection module is also used to: match the characteristic data sequence with the benchmark characteristic parameters to obtain the degree of deviation between the characteristic data sequence and the benchmark characteristic parameters; in response to the degree of deviation being greater than a preset threshold, determine that the fault detection result is whether there is an abnormality in the changing trend of multiple cached data; in response to the degree of deviation being less than or equal to the preset threshold, determine that the fault detection result is that the changing trend of multiple cached data is normal.

[0016] Furthermore, the feature extraction module is also used to: determine multiple window data from multiple cache data based on a preset time window, and summarize the multiple window data corresponding to multiple isomorphic devices to obtain multiple summary data; obtain the device quantity of multiple isomorphic devices; in response to the device quantity being less than or equal to a preset value, sort the multiple summary data to obtain a data sorting result, and construct a feature data sequence based on the Euclidean distance between adjacent summary data in the data sorting result; in response to the device quantity being greater than a preset value, construct a feature data sequence based on the multiple summary data.

[0017] Furthermore, the parameter determination module is also used to: in response to the number of devices being less than or equal to a preset value, obtain a benchmark feature parameter based on the average value of the Euclidean distance contained in the Euclidean distance sequence; in response to the number of devices being greater than a preset value, obtain a data average value based on the average value of multiple summarized data, and obtain a data standard deviation based on the standard deviation of multiple summarized data, and construct a benchmark feature parameter based on the data average value and the data standard deviation.

[0018] Furthermore, the fault detection result is used to characterize whether there is an abnormality in the values ​​of multiple cache data corresponding to different isomorphic devices; the fault detection module is also used to: in response to the number of devices being less than or equal to a preset value, clustering multiple summary data based on the average value of the Euclidean distance to obtain a data clustering set, and determining the fault detection result based on the number of data clustering sets and the number of devices; in response to the number of devices being greater than a preset value, matching multiple summary data with a numerical interval to obtain a numerical matching result, and determining the fault detection result based on the numerical matching result.

[0019] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the method in each embodiment of the present invention is executed when the program is running.

[0020] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0021] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0022] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0023] According to another aspect of the embodiments of the present invention, a computer program is further provided. When the computer program is executed by a processor, the methods in the embodiments of the present invention are implemented.

[0024] In an embodiment of the present invention, a method is adopted in which a cache data set generated by multiple isomorphic devices in a preset time period is obtained; features are extracted from multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different isomorphic devices; based on the feature data sequences, baseline feature parameters corresponding to multiple isomorphic devices are determined; based on the baseline feature parameters and the feature data sequences, fault detection is performed on different isomorphic devices to obtain fault detection results for different isomorphic devices. By determining the baseline feature parameters that will be generated by multiple isomorphic devices in normal operation through multiple cache data generated by different isomorphic data in a short period of time, and then using the baseline feature parameters to perform fault detection on different isomorphic devices, it is possible to reduce the degree of dependence of the fault detection process on long-term historical data while achieving real-time and accurate fault detection of isomorphic devices, thereby solving the technical problem of low accuracy in fault detection of isomorphic devices for which historical data is difficult to obtain in the related art. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0026] Figure 1 is a flow chart of a data fault detection method according to an embodiment of the present invention;

[0027] Figure 2 It is a structural block diagram of a data fault detection device according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] According to an embodiment of the present invention, a method embodiment of fault detection is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0031] Figure 1 FIG. 1 is a flow chart of a data fault detection method according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0032] Step S102: obtaining a cache data set generated by multiple homogeneous devices in a preset time period.

[0033] In an optional solution of this embodiment, considering that when a large number of homogeneous devices are currently being detected for faults, it is usually necessary to use the historical data generated by these homogeneous devices in a long historical period, and then judge whether these homogeneous devices have faults based on the acquired historical data through methods such as expert systems, fault trees, large statistical models, neural network models, etc., but due to the existence of a series of influencing factors such as high data confidentiality, poor quality of historical data, and changes in the application environment, when these methods are used to actually detect faults for multiple homogeneous devices, detection errors may occur due to excessive reliance on long-term historical data. Therefore, in order to avoid the poor quality of the acquired historical data, or even the inability to acquire historical data, resulting in the inability to accurately detect faults for multiple homogeneous devices, the fault detection system can record and acquire the cache data generated by multiple homogeneous devices in a short period of time in real time, and determine the standard data that these homogeneous devices may generate when running normally according to the cache data corresponding to different homogeneous devices through data simulation, neural network model prediction, etc., and then match the determined standard data with the acquired cache data, so as to judge whether the corresponding homogeneous device has faults according to the matching results.

[0034] Based on this, when performing fault detection on multiple homogeneous devices, the detection system can first obtain the cache data set generated by these multiple homogeneous devices in a short period of time and flowing into the above-mentioned preset time period, and use the multiple cache data contained in the cache data set to perform fault detection on the corresponding different homogeneous devices.

[0035] Step S104 , extracting features from the plurality of cache data included in different cache data sets to obtain feature data sequences corresponding to different isomorphic devices.

[0036] In an optional scheme of the present embodiment, in order to determine the standard data generated by these multiple isomorphic devices when operating under normal circumstances, the detection system may first perform feature extraction on the multiple cache data contained in different cache data sets to determine the key features used to analyze the multiple cache data, such as the data change trend, the correlation and similarity between different data, etc., so as to obtain a feature data sequence that can reflect the changes in the above-mentioned multiple cache data, so that the detection system can capture and analyze the changes in the data generated by different isomorphic devices within a preset time period and the changes in the operating status of different isomorphic devices through the feature data sequence, and then indirectly determine whether the different isomorphic devices are currently operating normally.

[0037] Step S106: determining the reference characteristic parameters corresponding to the plurality of isomorphic devices based on the characteristic data sequence.

[0038] The benchmark characteristic parameters are used to characterize characteristic parameters of multiple cache data when multiple homogeneous devices operate normally.

[0039] In an optional solution of the present embodiment, in order to improve the accuracy of fault detection for different isomorphic devices, the detection system may analyze the characteristic data contained in the characteristic data sequence after determining the characteristic data sequence corresponding to the different isomorphic devices, for example, analyzing the change trend of the data contained in the characteristic data sequence to determine the standard change trend corresponding to the multiple cache data when multiple isomorphic devices generate multiple cache data under normal circumstances, or analyzing the size of the data contained in the characteristic data sequence to determine the standard data size corresponding to the multiple cache data when multiple isomorphic devices generate multiple cache data under normal circumstances, thereby determining the standard parameters generated when the multiple isomorphic devices operate normally, that is, the above-mentioned baseline characteristic parameters. Considering that different isomorphic devices have the same design, structure, function and operating principle, the above-mentioned baseline characteristic parameters corresponding to the multiple isomorphic devices may also be the same.

[0040] Step S108, performing fault detection on different homogeneous devices based on the reference characteristic parameters and the characteristic data sequence to obtain fault detection results of different homogeneous devices.

[0041] In an optional solution of the present embodiment, after determining the reference characteristic parameters corresponding to the multiple isomorphic devices, the detection system can perform fault detection on the different isomorphic devices according to the reference characteristic parameters and the characteristic data sequences currently presented by the different isomorphic devices, so as to obtain the fault detection results corresponding to the different isomorphic devices. For example, assuming that the reference characteristic parameters determined above include the standard change trend and the standard data size, the detection system can match the data change trend of the multiple cache data generated by the different isomorphic devices in the above characteristic data sequence with the standard change trend to determine whether the change trend of the multiple cache data currently generated by the different isomorphic devices is normal, or the detection system can match the data size of the multiple cache data generated by the different isomorphic devices in the above characteristic data sequence with the standard data size to determine whether the current data size of the multiple cache data is normal. Correspondingly, if it is detected that the current change trend of the multiple cache data is abnormal, or the current data size of the multiple cache data is abnormal, the detection system can determine that the isomorphic devices corresponding to the multiple cache data are faulty, otherwise it can be determined that the isomorphic devices are normal.

[0042] In an embodiment of the present invention, a method is adopted in which a cache data set generated by multiple isomorphic devices in a preset time period is obtained; features are extracted from multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different isomorphic devices; based on the feature data sequences, baseline feature parameters corresponding to multiple isomorphic devices are determined; based on the baseline feature parameters and the feature data sequences, fault detection is performed on different isomorphic devices to obtain fault detection results for different isomorphic devices. By determining the baseline feature parameters that will be generated by multiple isomorphic devices in normal operation through multiple cache data generated by different isomorphic data in a short period of time, and then using the baseline feature parameters to perform fault detection on different isomorphic devices, it is possible to reduce the degree of dependence of the fault detection process on long-term historical data while achieving real-time and accurate fault detection of isomorphic devices, thereby solving the technical problem of low accuracy in fault detection of isomorphic devices for which historical data is difficult to obtain in the related art.

[0043] Furthermore, the characteristic data sequence is used to characterize the data change trend of multiple cache data; feature extraction is performed on multiple cache data contained in different cache data sets to obtain characteristic data sequences corresponding to different isomorphic devices, including: smoothing filtering is performed on multiple cache data to obtain filtered data; based on the filtered data, a data change curve corresponding to the isomorphic device is constructed; feature extraction is performed on the data change curve to obtain the trend turning point corresponding to the data change curve; based on the trend turning point, a characteristic data sequence corresponding to the isomorphic device is constructed.

[0044] In an optional scheme of the present embodiment, as mentioned above, in order to accurately determine whether the multiple cache data currently generated by different isomorphic devices are abnormal, so as to determine whether the corresponding isomorphic devices fail, the detection system can construct feature data sequences corresponding to different isomorphic devices based on the data change trends of the multiple cache data corresponding to different isomorphic devices. Based on this, in order to accurately obtain the above-mentioned feature data sequence, the detection system can first perform smoothing filtering on the multiple cache data to obtain the corresponding filtered data, and then construct the data change curve of the corresponding isomorphic device in a preset time period based on the filtered data and the recording time corresponding to the different data. Feature extraction is performed on the constructed data change curve to obtain key points that can reflect the trend change of the data change curve, that is, the above-mentioned trend turning points. Finally, according to the recording time corresponding to the trend turning points, the detection system can construct the corresponding feature data sequence.

[0045] Among them, in order to ensure the accuracy of the extracted trend turning point, multiple Savitzky-Golay smoothing filters can be used, and the signal can be smoothed by a local polynomial fitting method. The signal is polynomially fitted in a sliding window (usually a window of odd length) to estimate the local trend in the window. The Savitzky-Golay filter can effectively eliminate high-frequency noise in the signal while retaining the low-frequency components of the signal, so it is often used for preprocessing and noise reduction. By performing multiple Savitzky-Golay smoothing filters, the detection system can remove high-frequency noise in the data, which can significantly improve the signal-to-noise ratio of the data, making it easier to effectively analyze the trend characteristics in the data and filter multiple cache data corresponding to different isomorphic devices.

[0046] Furthermore, the baseline characteristic parameters are used to characterize the data change trend of multiple cached data under normal circumstances; based on the characteristic data sequence, the baseline characteristic parameters corresponding to multiple isomorphic devices are determined, including: respectively performing correlation detection on the characteristic data sequences corresponding to different isomorphic devices and the characteristic data sequences corresponding to other isomorphic devices in the multiple isomorphic devices to obtain correlation detection results corresponding to different isomorphic devices; based on the correlation detection results, evaluating the characteristic data sequences corresponding to different isomorphic devices to obtain sequence evaluation results; based on the sequence evaluation results corresponding to different isomorphic devices, determining the baseline characteristic parameters from the characteristic data sequences corresponding to different isomorphic devices.

[0047] In an optional scheme of the present embodiment, in order to accurately determine whether there is an abnormality in the change trend of multiple cache data corresponding to different isomorphic devices, after determining the feature data sequences corresponding to different isomorphic devices, the detection system can detect the correlation between different feature data sequences. For example, the detection system can select a target isomorphic device from multiple isomorphic devices in a polling, random selection, etc. manner, and perform a correlation detection on the first feature data sequence corresponding to the target isomorphic device and the second feature data sequences corresponding to other isomorphic devices. For example, the position, turning direction, turning strength and other parameters of the multiple trend turning points contained in the first feature data sequence are detected for similarity with the position, turning direction, turning strength and other parameters of the multiple trend turning points contained in different second feature data sequences to obtain the correlation between the first feature data sequence and the different second feature data sequences, thereby obtaining the correlation between the target isomorphic device and other isomorphic devices, and summarizing the correlations corresponding to other isomorphic devices to obtain the correlation detection result corresponding to the target isomorphic device. According to the above process, the characteristic data sequences corresponding to different isomorphic devices are respectively subjected to correlation detection with the characteristic data sequences corresponding to other isomorphic devices in the multiple isomorphic devices, and the correlation detection results corresponding to the different isomorphic devices can be obtained. After obtaining the correlation detection results corresponding to the different isomorphic devices, the detection system can evaluate the characteristic data sequences corresponding to the different isomorphic devices according to the correlation detection results, so as to obtain the sequence evaluation results corresponding to the different characteristic data sequences. Finally, according to the sequence evaluation results obtained by the evaluation, the detection system can select the target characteristic sequence from the different characteristic data sequences as the above-mentioned benchmark characteristic parameters to reflect the change trend of the cache data generated by the different isomorphic devices during normal operation.

[0048] For example, taking the number of votes as the sequence evaluation result, if the correlation detection result corresponding to the isomorphic device A is that there are currently 5 feature data sequences whose correlation with the feature data sequence A is greater than the preset correlation threshold, then it can be determined that the number of votes for the feature data sequence A is 5 votes; if the correlation detection result corresponding to the isomorphic device B is that there are currently 4 feature data sequences whose correlation with the feature data sequence B is greater than the preset correlation threshold, then it can be determined that the number of votes for the feature data sequence B is 4 votes; if the correlation detection result corresponding to the isomorphic device C is that there are currently 3 feature data sequences whose correlation with the feature data sequence C is greater than the preset correlation threshold, then it can be determined that the number of votes for the feature data sequence C is 3 votes. Based on this, it can be determined that, among the feature data sequences A, B, and C, the change trend reflected by the feature data sequence A can better reflect the change trend of the cache data generated by most isomorphic devices during operation. Based on this, the detection system can use the data change trend reflected by the feature data sequence A as the change trend of the cache data generated by different isomorphic devices under normal circumstances, and correspondingly use the feature data sequence A as the above-mentioned benchmark feature parameter.

[0049] It should be noted that the above process and parameters for determining the sequence evaluation results corresponding to different feature data sequences are only exemplary. In addition, weighted summation, sequence evaluation model evaluation and other methods can be used to evaluate different feature data sequences. Users can choose according to actual conditions, which will not be elaborated here.

[0050] Furthermore, the fault detection result is used to characterize whether there is an abnormality in the change trend of multiple cached data; fault detection is performed on different isomorphic devices based on baseline characteristic parameters and characteristic data sequences to obtain fault detection results of different isomorphic devices, including: matching the characteristic data sequence with the baseline characteristic parameters to obtain the degree of deviation between the characteristic data sequence and the baseline characteristic parameters; in response to the degree of deviation being greater than a preset threshold, determining that the fault detection result is whether there is an abnormality in the change trend of multiple cached data; in response to the degree of deviation being less than or equal to a preset threshold, determining that the fault detection result is that the change trend of multiple cached data is normal.

[0051] In an optional scheme of the present embodiment, after determining the baseline characteristic parameters corresponding to multiple passing devices, in order to detect whether there is an abnormality in the changing trend of the cache data currently generated by different isomorphic devices, the detection system may match the determined baseline characteristic parameters with the characteristic data sequences corresponding to different isomorphic devices to determine the degree of deviation between the different characteristic data sequences and the baseline characteristic parameters. For example, the corresponding degree of deviation is determined based on the degree of matching of a series of parameters such as the turning direction and turning strength of the trend turning points with the same position in different characteristic sequences and the baseline characteristic parameters. Correspondingly, if the determined degree of deviation is greater than a preset threshold, it can be determined that the corresponding fault detection result is that there is an abnormality in the changing trend of the cache data currently generated by the isomorphic device; if the determined degree of deviation is less than or equal to the preset threshold, it can be determined that the corresponding fault detection result is that the changing trend of the cache data currently generated by the isomorphic device is normal.

[0052] Furthermore, feature extraction is performed on multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different isomorphic devices, including: based on a preset time window, determining multiple window data from multiple cache data, and aggregating the multiple window data corresponding to multiple isomorphic devices to obtain multiple summary data; obtaining the number of devices of the multiple isomorphic devices; in response to the number of devices being less than or equal to a preset value, sorting the multiple summary data to obtain a data sorting result, and constructing a feature data sequence based on the Euclidean distance between adjacent summary data in the data sorting result; in response to the number of devices being greater than a preset value, constructing a feature data sequence based on the multiple summary data.

[0053] In an optional solution of this embodiment, in addition to performing fault detection on different isomorphic devices according to the change trend of the cache data currently generated by different isomorphic devices, the detection system can also perform fault detection on different isomorphic devices according to the data size of the cache data currently generated by different isomorphic devices. Based on this, in order to reflect the data size of the cache data generated by different isomorphic devices under normal circumstances, the detection system can first use a preset time window, with the current moment as the end time, to determine multiple window data from multiple cache data, and summarize the multiple window data corresponding to different isomorphic devices, thereby obtaining the above-mentioned multiple summary data. Considering that the greater the number of multiple isomorphic devices, the closer the distribution of the data means of these multiple summary data will be to the normal distribution, and the corresponding assumption of the normal distribution using the sigma analysis method will be more reasonable. Therefore, when using the data size of the cache data currently generated by different isomorphic devices to perform fault detection on different isomorphic devices, the detection system can also obtain the current number of devices of the multiple isomorphic devices. If the number of devices is less than or equal to a preset value, such as 30, the multiple summary data can be sorted in order from large to small or from small to small to obtain the above-mentioned data sorting result, and then the corresponding Euclidean distance sequence is constructed according to the Euclidean distance between two adjacent summary data. At this time, the detection system can determine the constructed Euclidean distance sequence as the above-mentioned feature data sequence; if the number of devices is greater than the preset value, such as 30, the detection system can directly construct the corresponding feature data sequence based on the multiple summary data.

[0054] Furthermore, based on the characteristic data sequence, the benchmark characteristic parameters corresponding to the multiple isomorphic devices are determined, including: in response to the number of devices being less than or equal to a preset value, the benchmark characteristic parameters are obtained based on the average value of the Euclidean distance contained in the Euclidean distance sequence; in response to the number of devices being greater than a preset value, the data average value is obtained based on the average value of multiple summarized data, and the data standard deviation is obtained based on the standard deviation of the multiple summarized data, and the benchmark characteristic parameters are constructed based on the data average value and the data standard deviation.

[0055] In an optional scheme of the present embodiment, corresponding to determining the characteristic data sequence, if the number of devices is less than or equal to a preset value, the detection system can determine the average value of the Euclidean distance contained in the Euclidean distance sequence, and determine the obtained average value as the above-mentioned benchmark characteristic parameter; if the number of devices is greater than the preset value, the detection system can obtain the corresponding data average value based on the average values ​​of multiple summarized data, and at the same time obtain the corresponding data standard deviation based on the standard deviations of multiple summarized data, and finally construct the corresponding benchmark characteristic parameter based on the data average value and the data standard deviation. At this time, the benchmark characteristic parameter can be the numerical range of the data size of the cache data generated by different isomorphic devices during normal operation, and the minimum value of the corresponding data interval can be the difference between the data mean and three times the data standard deviation, and the maximum value of the data interval can be the sum of the data mean and three times the data standard deviation.

[0056] Furthermore, the fault detection result is used to characterize whether there is an abnormality in the values ​​of multiple cache data corresponding to different isomorphic devices; fault detection is performed on different isomorphic devices based on baseline feature parameters and feature data sequences to obtain fault detection results of different isomorphic devices, including: in response to the number of devices being less than or equal to a preset value, clustering multiple summary data based on the average Euclidean distance to obtain a data clustering set, and determining the fault detection result based on the number of data clustering sets and the number of devices; in response to the number of devices being greater than a preset value, matching multiple summary data with a numerical interval to obtain a numerical matching result, and determining the fault detection result based on the numerical matching result.

[0057] In an optional scheme of the present embodiment, in order to accurately determine whether the data size of the cache data currently generated by different isomorphic devices is normal, if the number of devices is less than or equal to a preset value, the detection system can perform neighbor clustering on multiple summary data according to the determined average value of the Euclidean distance to obtain the above-mentioned data clustering set. If the number of data sets is 1, it can be considered that the above-mentioned multiple summary data are in one cluster, and then it can be determined that the data size of the cache data is normal; if the number of data sets is the same as the number of devices, that is, the window data of different isomorphic devices are in different clusters, and the data has the characteristics of uniform divergence, then it can also be determined that the data size of the cache data is normal; if the number of data sets is between 1 and the number of devices, and the number of points of the cluster centers corresponding to different clusters is greater than the ratio of the number of devices to the number of data sets, then it can also be determined that the data size of the cache data corresponding to the cluster is normal, otherwise, it can be determined that the data size of the cache data is abnormal. If the number of devices is greater than the preset value, the detection system can directly determine whether the multiple cache data corresponding to different homogeneous devices are within the numerical range. If so, it can be determined that the data size of the cache data currently generated by the corresponding homogeneous device is normal. If not, it can be determined that the data size of the cache data currently generated by the corresponding homogeneous device is abnormal.

[0058] According to an embodiment of the present invention, a device embodiment for fault detection is provided. It should be noted that the device can be used to execute the above-mentioned data fault detection method. Figure 2 is a structural block diagram of a data fault detection device according to an embodiment of the present application. Figure 2 As shown, the apparatus may include: a set acquisition module 202 , a feature extraction module 204 , a parameter determination module 206 and a fault detection module 208 .

[0059] Among them, the set acquisition module 202 is used to obtain the cache data set generated by multiple homogeneous devices in a preset time period; the feature extraction module 204 is used to extract features from multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different homogeneous devices; the parameter determination module 206 is used to determine the baseline feature parameters corresponding to multiple homogeneous devices based on the feature data sequence, wherein the baseline feature parameters are used to characterize the feature parameters of multiple cache data when multiple homogeneous devices operate normally; the fault detection module 208 is used to perform fault detection on different homogeneous devices based on the baseline feature parameters and the feature data sequence to obtain fault detection results for different homogeneous devices.

[0060] Furthermore, the feature data sequence is used to characterize the data change trend of multiple cache data; the feature extraction module is also used to: perform smoothing filtering on the multiple cache data to obtain filtered data; based on the filtered data, construct a data change curve corresponding to the isomorphic device; perform feature extraction on the data change curve to obtain the trend turning point corresponding to the data change curve; based on the trend turning point, construct a feature data sequence corresponding to the isomorphic device.

[0061] Furthermore, the baseline characteristic parameters are used to characterize the data change trend of multiple cached data under normal circumstances; the parameter determination module is also used to: perform correlation detection on the characteristic data sequences corresponding to different isomorphic devices and the characteristic data sequences corresponding to other isomorphic devices in the multiple isomorphic devices, and obtain the correlation detection results corresponding to the different isomorphic devices; based on the correlation detection results, evaluate the characteristic data sequences corresponding to the different isomorphic devices to obtain sequence evaluation results; based on the sequence evaluation results corresponding to the different isomorphic devices, determine the baseline characteristic parameters from the characteristic data sequences corresponding to the different isomorphic devices.

[0062] Furthermore, the fault detection result is used to characterize whether there is an abnormality in the changing trend of multiple cached data; the fault detection module is also used to: match the characteristic data sequence with the benchmark characteristic parameters to obtain the degree of deviation between the characteristic data sequence and the benchmark characteristic parameters; in response to the degree of deviation being greater than a preset threshold, determine that the fault detection result is whether there is an abnormality in the changing trend of multiple cached data; in response to the degree of deviation being less than or equal to the preset threshold, determine that the fault detection result is that the changing trend of multiple cached data is normal.

[0063] Furthermore, the feature extraction module is also used to: determine multiple window data from multiple cache data based on a preset time window, and summarize the multiple window data corresponding to multiple isomorphic devices to obtain multiple summary data; obtain the device quantity of multiple isomorphic devices; in response to the device quantity being less than or equal to a preset value, sort the multiple summary data to obtain a data sorting result, and construct a feature data sequence based on the Euclidean distance between adjacent summary data in the data sorting result; in response to the device quantity being greater than a preset value, construct a feature data sequence based on the multiple summary data.

[0064] Furthermore, the parameter determination module is also used to: in response to the number of devices being less than or equal to a preset value, obtain a benchmark feature parameter based on the average value of the Euclidean distance contained in the Euclidean distance sequence; in response to the number of devices being greater than a preset value, obtain a data average value based on the average value of multiple summarized data, and obtain a data standard deviation based on the standard deviation of multiple summarized data, and construct a benchmark feature parameter based on the data average value and the data standard deviation.

[0065] Furthermore, the fault detection result is used to characterize whether there is an abnormality in the values ​​of multiple cache data corresponding to different isomorphic devices; the fault detection module is also used to: in response to the number of devices being less than or equal to a preset value, clustering multiple summary data based on the average value of the Euclidean distance to obtain a data clustering set, and determining the fault detection result based on the number of data clustering sets and the number of devices; in response to the number of devices being greater than a preset value, matching multiple summary data with a numerical interval to obtain a numerical matching result, and determining the fault detection result based on the numerical matching result.

[0066] An embodiment of the present application further provides an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of the present invention when running.

[0067] An embodiment of the present application further provides a computer-readable storage medium, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0068] An embodiment of the present application further provides a computer program product, including a computer program, which implements the methods in various embodiments of the present invention when executed by a processor.

[0069] An embodiment of the present application further provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium is used to store a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0070] The embodiments of the present application further provide a computer program, which implements the methods in the above-mentioned embodiments of the present invention when executed by a processor.

[0071] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0072] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. 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 mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0073] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0074] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0075] If the 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 storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0076] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A data fault detection method, characterized in that: include: Obtain cache data sets generated by multiple homogeneous devices in a preset time period; Perform feature extraction on multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different isomorphic devices; Determine, based on the characteristic data sequence, benchmark characteristic parameters corresponding to the multiple homogeneous devices, wherein the benchmark characteristic parameters are used to characterize characteristic parameters of the multiple cached data when the multiple homogeneous devices operate normally; Fault detection is performed on different isomorphic devices based on the reference characteristic parameters and the characteristic data sequence to obtain fault detection results of different isomorphic devices.

2. The method according to claim 1, characterized in that: The characteristic data sequence is used to characterize the data change trend of the plurality of cached data; Feature extraction is performed on multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different homogeneous devices, including: Performing smoothing filtering on the plurality of cached data to obtain filtered data; Based on the filtered data, construct a data change curve corresponding to the homogeneous device; Extracting features from the data change curve to obtain a trend turning point corresponding to the data change curve; Based on the trend turning point, the characteristic data sequence corresponding to the isomorphic device is constructed.

3. The method according to claim 2, characterized in that The reference characteristic parameter is used to characterize the data change trend of the plurality of cached data under normal circumstances; Determining, based on the characteristic data sequence, reference characteristic parameters corresponding to the plurality of homogeneous devices, includes: Respectively performing correlation detection on the characteristic data sequences corresponding to different isomorphic devices and the characteristic data sequences corresponding to other isomorphic devices among the multiple isomorphic devices to obtain correlation detection results corresponding to different isomorphic devices; Based on the correlation detection result, evaluating the characteristic data sequences corresponding to different isomorphic devices to obtain a sequence evaluation result; Based on the sequence evaluation results corresponding to different isomorphic devices, the reference feature parameters are determined from the feature data sequences corresponding to different isomorphic devices.

4. The method according to claim 3, characterized in that The fault detection result is used to characterize whether there is an abnormality in the change trend of the plurality of cache data; Performing fault detection on different homogeneous devices based on the reference characteristic parameters and the characteristic data sequence to obtain fault detection results of different homogeneous devices includes: Matching the characteristic data sequence with the reference characteristic parameter to obtain a degree of deviation between the characteristic data sequence and the reference characteristic parameter; In response to the deviation being greater than a preset threshold, determining whether the fault detection result is abnormal in the change trend of the plurality of cache data; In response to the deviation degree being less than or equal to the preset threshold, it is determined that the fault detection result is that the change trend of the plurality of cache data is normal.

5. The method according to claim 1, characterized in that Feature extraction is performed on multiple cache data contained in different cache data sets to obtain feature data sequences corresponding to different homogeneous devices, including: Based on a preset time window, determine a plurality of window data from the plurality of cache data, and aggregate the plurality of window data corresponding to the plurality of homogeneous devices to obtain a plurality of aggregate data; Obtain the number of devices of the multiple homogeneous devices; In response to the number of devices being less than or equal to a preset value, sorting the plurality of summary data to obtain a data sorting result, and constructing the feature data sequence based on the Euclidean distance between adjacent summary data in the data sorting result; In response to the number of devices being greater than the preset value, the characteristic data sequence is constructed based on the plurality of summary data.

6. The method according to claim 5, characterized in that Determining, based on the characteristic data sequence, reference characteristic parameters corresponding to the plurality of homogeneous devices, includes: In response to the number of devices being less than or equal to a preset value, obtaining the reference feature parameter based on an average value of the Euclidean distances included in the Euclidean distance sequence; In response to the number of devices being greater than the preset value, a data average is obtained based on an average of the multiple summarized data, and a data standard deviation is obtained based on a standard deviation of the multiple summarized data, and the benchmark feature parameter is constructed based on the data average and the data standard deviation.

7. The method according to claim 6, characterized in that The fault detection result is used to characterize whether there is an abnormality in the values ​​of the plurality of cache data corresponding to different homogeneous devices; Performing fault detection on different homogeneous devices based on the reference characteristic parameters and the characteristic data sequence to obtain fault detection results of different homogeneous devices includes: In response to the number of devices being less than or equal to a preset value, clustering the plurality of summary data based on the average value of the Euclidean distance to obtain a data cluster set, and determining the fault detection result based on the number of the data cluster sets and the number of devices; In response to the number of devices being greater than the preset value, the plurality of summary data are matched with the numerical value interval to obtain a numerical value matching result, and the fault detection result is determined based on the numerical value matching result.

8. An electronic device, characterized in that: include: A memory storing an executable program; A processor, configured to run the program, wherein the program executes the method according to any one of claims 1 to 7 when running.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is executed, the device where the storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images