Performance monitoring method and device, computer equipment, storage medium and program product
By applying machine learning-based performance monitoring methods in transformers and analyzing transformer state data using initial and optimized machine learning models, the problem of inaccurate monitoring of traditional DGA methods is solved, and higher monitoring accuracy and early fault recognition capabilities are achieved.
Patent Information
- Application Number
- CN202411582469.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-07
- Publication Date
- 2025-03-18
AI Technical Summary
The traditional dissolved gas analysis (DGA) method in transformer oil has problems with inaccurate monitoring, making it difficult to accurately identify the type and severity of potential faults inside the transformer.
Through the performance monitoring method based on machine learning, the initial machine learning model is used to train the status monitoring data of the target device, generate an initial prediction model, and optimize the model through incremental learning to generate a second prediction model. Then, the performance monitoring results of the equipment are determined based on the prediction results of the two models.
It improves the accuracy of monitoring the performance of the target equipment, can more accurately determine whether the performance of the equipment has changed during adjacent time periods, and enhances the early identification of internal faults of the transformer.
Smart Images
Figure CN119646529A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of transformers, and in particular to a performance monitoring method, device, computer equipment, storage medium and program product. Background Art
[0002] Dissolved Gas Analysis (DGA) in transformer oil is an important technology used to evaluate transformer internal faults and insulation conditions. During transformer operation, due to electrical and thermal stress, insulating oil and solid insulating materials will decompose and produce a variety of gases, such as hydrogen (H2), methane (CH4), ethylene (C4H4), acetylene (C2H2), etc. DGA can detect and analyze the concentration and composition of these dissolved gases to identify the potential fault type and severity of the transformer at an early stage.
[0003] However, the traditional DGA performance monitoring method has the problem of inaccurate monitoring. Summary of the invention
[0004] Based on this, it is necessary to provide a performance monitoring method, device, computer equipment, storage medium and program product that can improve monitoring accuracy in response to the above technical problems.
[0005] In a first aspect, the present application provides a performance monitoring method, comprising:
[0006] The initial machine learning model is trained according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device;
[0007] According to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block are obtained; the first data block and the second data block are monitoring data of adjacent preset time periods of the target device;
[0008] A performance monitoring result of the target device is determined according to the first prediction result and the second prediction result.
[0009] In one of the embodiments, obtaining the first prediction result and the second prediction result corresponding to the second data block according to the first prediction model and the second data block in the data to be predicted includes:
[0010] Inputting the second data block into the first prediction model to obtain the first prediction result;
[0011] Training the first prediction model according to the second data block to obtain a second prediction model;
[0012] The second data block is input into the second prediction model to obtain the second prediction result.
[0013] In one embodiment, the training of the first prediction model according to the second data block to obtain the second prediction model includes:
[0014] The monitoring data and label data in the second data block are used as training samples, and incremental learning training is performed on the first prediction model to obtain the second prediction model.
[0015] In one embodiment, determining the performance monitoring result of the target device according to the first prediction result and the second prediction result includes:
[0016] Determining a similarity between the first prediction result and the second prediction result;
[0017] The performance monitoring result of the target device is determined according to the similarity and a preset adaptive threshold.
[0018] In one embodiment, determining the performance monitoring result of the target device according to the similarity and a preset adaptive threshold includes:
[0019] If the similarity is greater than the adaptive threshold, determining that the performance monitoring result is that performance drift occurs on the target device in the time period corresponding to the second data block;
[0020] If the similarity is not greater than the adaptive threshold, it is determined that the performance monitoring result is that no performance drift occurs to the target device in the time period corresponding to the second data block.
[0021] In one embodiment, the first prediction result and the second prediction result are vectors of the same length, and determining the similarity between the first prediction result and the second prediction result includes:
[0022] Determining a product value of the transpose of the first prediction result and the second prediction result;
[0023] The similarity between the first prediction result and the second prediction result is determined according to the product value, the absolute value of the first prediction result and the absolute value of the second prediction result.
[0024] In a second aspect, the present application also provides a performance monitoring device, comprising:
[0025] A training model is used to train an initial machine learning model according to a first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is status monitoring data of a target device;
[0026] A first determination module, configured to obtain a first prediction result and a second prediction result corresponding to a second data block in the data to be predicted according to the first prediction model and the second data block; the first data block and the second data block are monitoring data of adjacent preset time periods of the target device;
[0027] The second determination module is used to determine the performance monitoring result of the target device according to the first prediction result and the second prediction result.
[0028] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:
[0029] The initial machine learning model is trained according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device;
[0030] According to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block are obtained; the first data block and the second data block are monitoring data of adjacent preset time periods of the target device;
[0031] A performance monitoring result of the target device is determined according to the first prediction result and the second prediction result.
[0032] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:
[0033] The initial machine learning model is trained according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device;
[0034] According to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block are obtained; the first data block and the second data block are monitoring data of adjacent preset time periods of the target device;
[0035] A performance monitoring result of the target device is determined according to the first prediction result and the second prediction result.
[0036] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:
[0037] The initial machine learning model is trained according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device;
[0038] According to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block are obtained; the first data block and the second data block are monitoring data of adjacent preset time periods of the target device;
[0039] A performance monitoring result of the target device is determined according to the first prediction result and the second prediction result.
[0040] The above-mentioned performance monitoring method, device, computer equipment, storage medium and program product train the initial machine learning model according to the first data block in the data to be predicted to obtain the first prediction model; the data to be predicted is the status monitoring data of the target device; according to the first prediction model and the second data block in the data to be predicted, the first prediction result and the second prediction result corresponding to the second data block are obtained; the first data block and the second data block are the monitoring data of the target device in adjacent preset time periods; the performance monitoring result of the target device is determined according to the first prediction result and the second prediction result. By judging the first prediction result and the second prediction result corresponding to the data blocks in adjacent preset time periods, it is determined whether the performance of the target device has changed in the adjacent time periods, thereby improving the accuracy of performance monitoring of the target device, and in each monitoring process, the initial machine learning model is trained according to the first data block, so that the first prediction model has a higher matching degree with each monitoring process, thereby further improving the accuracy of performance monitoring of the target device. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0042] Figure 1 An application environment diagram of a performance monitoring method in an embodiment;
[0043] Figure 2 A schematic diagram of a performance monitoring method in one embodiment;
[0044] Figure 3 is a flow chart of a performance monitoring method in another embodiment;
[0045] Figure 4is a flow chart of a performance monitoring method in another embodiment;
[0046] Figure 5 is a flow chart of a performance monitoring method in another embodiment;
[0047] Figure 6 is a flow chart of a performance monitoring method in another embodiment;
[0048] Figure 7 is a flow chart of a performance monitoring method in another embodiment;
[0049] Figure 8 is a structural block diagram of a performance monitoring device in one embodiment;
[0050] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0051] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0052] The performance monitoring method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The server 104 can obtain the data to be predicted from the terminal 103 in advance, so that the server 104 can monitor the performance of the target device according to the data to be predicted. Among them, the terminal 102 can be but not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.
[0053] In one embodiment, Figure 2 As shown, a performance monitoring method is provided, which is applied to Figure 1 The following is an example of a server in the example, including:
[0054] S201, training an initial machine learning model according to a first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is status monitoring data of a target device.
[0055] The data to be predicted may be Dissolved Gas Analysis (DGA) data of the target device. The DGA data may include the contents of hydrogen, methane, acetylene, ethylene, and ethane. Therefore, the DGA data at each moment is a multi-dimensional vector.
[0056] In the embodiment of the present application, a DGA system may be installed on the terminal, so that the acquisition component corresponding to the DGA system is called to collect the target device's data to be predicted, and the target device's data to be predicted is sent to the server; or, a DGA system may be installed in the server, so that the acquisition component corresponding to the DGA system is called to collect the target device's data to be predicted. Optionally, the terminal or server may call the DGA component to collect DGA data once every first preset time interval. For example, when the first preset time interval is 1 minute, the terminal or server collects DGA data once every 1 minute.
[0057] Furthermore, the server divides the data to be predicted into multiple data blocks according to the second preset time length, and marks the state type corresponding to each data in the data block. Optionally, the state type may include normal, low-energy discharge, medium-energy discharge, high-energy discharge, low-temperature overheating, medium-temperature overheating, and high-temperature overheating; the server may mark the state type corresponding to each data in the data block according to the user's input data, or the server may determine the state type corresponding to each data in the data block according to the content of each gas. Optionally, if the second preset time length is 1 hour, the server merges the data of each hour into a data block. , where i is the sequence number of the data block.
[0058] In the embodiment of the present application, the first data block in the data to be predicted can be any data block in the data to be predicted, or, in the first round of monitoring, the first data block is the first data in the data to be predicted, and in the nth round of monitoring, the first data block is the nth data in the data to be predicted. The first data block is used as sample data to perform supervised training on the initial machine learning model to obtain a first prediction model.
[0059] Optionally, multiple sample subsets are randomly extracted from the first data block with replacement by a bootstrap method, and each sample subset is used to train a decision tree. Suppose the first data block contains N samples. Each time N samples are extracted, some samples may be repeatedly selected, while others may not be selected. Furthermore, when partitioning at each node, a feature subset is randomly selected instead of using all features. Suppose there are M features, and m features are randomly selected for partitioning each time. For classification problems, usually For each sample subset and feature subset, a decision tree is constructed. No pruning is required during the tree construction process, and the tree is allowed to grow to its maximum depth.
[0060] S202, obtaining a first prediction result and a second prediction result corresponding to the second data block according to the first prediction model and the second data block in the data to be predicted; the first data block and the second data block are monitoring data of the target device in adjacent preset time periods.
[0061] The second data block is the next data block adjacent to the first data block, that is, the first data block and the second data block are monitoring data of a preset time period adjacent to the target device.
[0062] In the embodiment of the present application, according to the first prediction model and the second data block, the first prediction result corresponding to the second data block is obtained, and according to the first prediction model and the second data block, the second prediction result corresponding to the second data block is obtained. Optionally, the second data block can be input into the first prediction model to obtain the first prediction result and the second prediction result; or, the second data block can be input into the first prediction model to obtain the first prediction result, and the first prediction model is optimized, and the second database is input into the optimized first prediction module to obtain the second prediction result.
[0063] S203: Determine a performance monitoring result of the target device according to the first prediction result and the second prediction result.
[0064] The performance monitoring result may indicate whether the target device has an abnormality within the target time period, for example, performance drift. Optionally, the target time period may be the monitoring time period corresponding to the second data block, or the target time period may be the monitoring time period corresponding to the first data block and the second data block.
[0065] In the embodiment of the present application, the first prediction result and the second prediction result are compared to determine whether the target device has performance drift during the monitoring time corresponding to the second data block. Optionally, if the first prediction result is the same as the second prediction result, the performance monitoring result of the target device is determined to be that the performance has not drifted; if the first prediction result is different from the second prediction result, the performance monitoring result of the target device is determined to be that the performance has drifted.
[0066] Optionally, after determining the performance monitoring result of the target device based on the first prediction result and the second prediction result, the second data block can be used as a new first data block, and the next data block adjacent to the second data block can be used as a new second data block, and the above-mentioned step of training the initial machine learning model based on the first data block in the data to be predicted to obtain the first prediction model is returned; the data to be predicted is the status monitoring data of the target device, until the performance monitoring results corresponding to all time periods are determined.
[0067] In the above performance monitoring method, the initial machine learning model is trained according to the first data block in the data to be predicted to obtain the first prediction model; the data to be predicted is the status monitoring data of the target device; according to the first prediction model and the second data block in the data to be predicted, the first prediction result and the second prediction result corresponding to the second data block are obtained; the first data block and the second data block are the monitoring data of the target device in adjacent preset time periods; the performance monitoring result of the target device is determined according to the first prediction result and the second prediction result. By judging the first prediction result and the second prediction result corresponding to the data blocks in adjacent preset time periods, it is determined whether the performance of the target device has changed in the adjacent time periods, thereby improving the accuracy of performance monitoring of the target device, and in each monitoring process, the initial machine learning model is trained according to the first data block, so that the first prediction model has a higher matching degree with each monitoring process, thereby further improving the accuracy of performance monitoring of the target device.
[0068] In one implementation, an implementation of the above S202 is provided, such as Figure 3 As shown, the above “obtaining a first prediction result and a second prediction result corresponding to the second data block according to the first prediction model and the second data block in the data to be predicted” includes:
[0069] S301, input the second data block into the first prediction model to obtain a first prediction result.
[0070] Optionally, the first prediction model can be expressed as , for each input data , and the judgment result is , is a vector of length 7, each element of which is the confidence level of the corresponding category (“normal”, “low energy discharge”, “medium energy discharge”, “high energy discharge”, “low temperature overheating”, “medium temperature overheating”, “high temperature overheating”).
[0071] In the embodiment of the present application, the second data block can be expressed as , input the second data block into the first prediction model to obtain the first prediction result .
[0072] S302: Train the first prediction model according to the second data block to obtain a second prediction model.
[0073] In the embodiment of the present application, the second data block is used as sample data to further train the first prediction model to obtain an updated first prediction model, and the updated first prediction model is used as the second prediction model. .
[0074] Optionally, the monitoring data and label data in the second data block are used as training samples to perform incremental learning training on the first prediction model to obtain a second prediction model.
[0075] Among them, incremental learning means that a learning system can continuously learn new knowledge from new samples and save most of the previously learned knowledge. Incremental learning does not need to save historical data, thereby reducing storage space occupancy. It can also use historical training results to significantly reduce subsequent training time.
[0076] In an embodiment of the present application, the second data block is labeled data, so the monitoring data and label data in the second data block are used as training samples to perform supervised training on the first prediction model to obtain the second prediction model.
[0077] S303: Input the second data block into the second prediction model to obtain a second prediction result.
[0078] In the embodiment of the present application, the second data block Input to the second prediction model , and get the second prediction result .
[0079] In the above-mentioned application embodiment, the second prediction model is trained using the second data block to obtain a second prediction model that better matches the second data block, thereby monitoring the time period corresponding to the second data block based on the prediction results of the first prediction model and the second prediction model on the second database, thereby improving the accuracy of the monitoring results.
[0080] In one embodiment, an implementation of the above S203 is provided, such as Figure 4 As shown, the above “determining the performance monitoring result of the target device according to the first prediction result and the second prediction result” includes:
[0081] S401, determining the similarity between the first prediction result and the second prediction result.
[0082] In an embodiment of the present application, the first prediction result and the second prediction result are both vectors including multiple elements. The first prediction result can be used as the first vector, and the second prediction result can be used as the second vector. The similarity between the first prediction result and the second prediction result can be determined based on the first vector and the second vector.
[0083] As an optional implementation, the number of identical vector elements in the first vector and the number of vector elements in the second vector is determined, and the ratio of the identical number to the number of elements in each vector is used as the similarity between the first prediction result and the second prediction result.
[0084] As another optional implementation, a first absolute value of the first vector and a second absolute value of the second vector are determined, so that a ratio of the first vector to the second vector is used as the similarity between the first prediction result and the second prediction result.
[0085] S402: Determine the performance monitoring result of the target device according to the similarity and a preset adaptive threshold.
[0086] In the embodiment of the present application, for the same number of drift samples, the degree of difference of the updated model will vary with the number of trained samples. The more training samples there are, the smaller the impact of new samples on the model. Therefore, an adaptive threshold must be used to intercept multiple data blocks using a sliding window and calculate their mean. and standard deviation , and take a fixed coefficient , adaptive threshold It can be expressed as: .
[0087] In the embodiment of the present application, the similarity is compared with the size of the preset adaptive threshold, so as to determine the performance monitoring result of the target device according to the comparison result. Optionally, if the similarity is greater than the adaptive threshold, the performance monitoring result is determined to be that the performance of the target device is abnormal; if the similarity is not greater than the adaptive threshold, the performance monitoring result is determined to be that the performance of the target device is normal.
[0088] In the above application embodiment, the adaptive threshold is determined by multiple data blocks, thereby improving the accuracy of the adaptive threshold. Further, the prediction result is determined based on the similarity and the adaptive threshold, thereby improving the accuracy of the prediction result.
[0089] In the embodiment of the present application, an implementation method of the above S402 is provided, such as Figure 5 As shown, the above “determining the performance monitoring result of the target device according to the similarity and the preset adaptive threshold value” includes:
[0090] S501: If the similarity is greater than the adaptive threshold, determine that the performance monitoring result is that the target device has a performance drift in the time period corresponding to the second data block.
[0091] In the embodiment of the present application, if the similarity is greater than the adaptive threshold, the difference between the first prediction result and the second prediction result is large, and the performance monitoring result is determined to be that the target device has performance drift in the time period corresponding to the second data block.
[0092] It should be noted that when the similarity between the detection results before and after the update is greater than the threshold, it means that the performance of the target device has drifted. In order to continue to detect the newly input data, it is necessary to reinitialize .
[0093] S502: If the similarity is not greater than the adaptive threshold, determine that the performance monitoring result is that the target device has no performance drift in the time period corresponding to the second data block.
[0094] In an embodiment of the present application, if the similarity is not greater than the adaptive threshold, the difference between the first prediction result and the second prediction result is small, and the performance monitoring result is determined to be that the target device has no performance drift in the time period corresponding to the second data block.
[0095] In the above application embodiment, whether performance drift occurs in the target device during the time period corresponding to the second data block is determined based on similarity and adaptive thresholds. The determination method is simple, accurate and error-prone, thereby improving the accuracy of the monitoring results.
[0096] In one embodiment, an implementation of the above S401 is provided, such as Figure 6 As shown, the first prediction result and the second prediction result are vectors of the same length, and the above “determining the similarity between the first prediction result and the second prediction result” includes:
[0097] S601, determine a product value of the transpose of the first prediction result and the second prediction result.
[0098] In the embodiment of the present application, the first prediction result is , the second prediction result is , then the product of the transpose of the first prediction result and the second prediction result is .
[0099] S602: Determine the similarity between the first prediction result and the second prediction result according to the product value, the absolute value of the first prediction result and the absolute value of the second prediction result.
[0100] In the embodiment of the present application, the similarity between the first prediction result and the second prediction result is determined according to the product value, the absolute value of the first prediction result, the absolute value of the second prediction result and a preset relationship.
[0101] Optionally, the absolute value of the first prediction result can be expressed as , the absolute value of the second prediction result can be expressed as , further, according to the preset relationship, the similarity between the first prediction result and the second prediction result can be expressed as:
[0102]
[0103] In the above application embodiment, the similarity between the first prediction result and the second prediction result is determined according to a preset relationship, which improves the accuracy of the similarity without introducing other parameters, thereby improving the determination efficiency and accuracy.
[0104] In one embodiment, a complete performance monitoring method is provided, such as Figure 7 As shown, the method includes:
[0105] S1, training an initial machine learning model according to a first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is status monitoring data of a target device.
[0106] S2, inputting the second data block into the first prediction model to obtain a first prediction result; the first data block and the second data block are monitoring data of the target device in adjacent preset time periods.
[0107] S3, using the monitoring data and label data in the second data block as training samples, performing incremental learning training on the first prediction model to obtain a second prediction model.
[0108] S4, inputting the second data block into the second prediction model to obtain a second prediction result.
[0109] S5, determining a product value of the transpose of the first prediction result and the second prediction result.
[0110] S6, determining the similarity between the first prediction result and the second prediction result according to the product value, the absolute value of the first prediction result and the absolute value of the second prediction result.
[0111] S7, if the similarity is greater than the adaptive threshold, determine that the performance monitoring result is that the target device has performance drift in the time period corresponding to the second data block; if the similarity is not greater than the adaptive threshold, determine that the performance monitoring result is that the target device has no performance drift in the time period corresponding to the second data block.
[0112] In the above performance monitoring method, the initial machine learning model is trained according to the first data block in the data to be predicted to obtain the first prediction model; the data to be predicted is the status monitoring data of the target device; according to the first prediction model and the second data block in the data to be predicted, the first prediction result and the second prediction result corresponding to the second data block are obtained; the first data block and the second data block are the monitoring data of the target device in adjacent preset time periods; the performance monitoring result of the target device is determined according to the first prediction result and the second prediction result. By judging the first prediction result and the second prediction result corresponding to the data blocks in adjacent preset time periods, it is determined whether the performance of the target device has changed in the adjacent time periods, thereby improving the accuracy of performance monitoring of the target device, and in each monitoring process, the initial machine learning model is trained according to the first data block, so that the first prediction model has a higher matching degree with each monitoring process, thereby further improving the accuracy of performance monitoring of the target device.
[0113] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0114] Based on the same inventive concept, the embodiment of the present application also provides a performance monitoring device for implementing the above-mentioned performance monitoring method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more performance monitoring device embodiments provided below can refer to the limitations on the performance monitoring method above, and will not be repeated here.
[0115] In one embodiment, Figure 8 As shown, a performance monitoring device is provided, comprising: a training module 10, a first determination module 11 and a second determination module 12, wherein:
[0116] The training model 10 is used to train the initial machine learning model according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device.
[0117] The first determination module 11 is used to obtain a first prediction result and a second prediction result corresponding to the second data block according to the first prediction model and the second data block in the data to be predicted; the first data block and the second data block are monitoring data of adjacent preset time periods of the target device.
[0118] The second determination module 12 is used to determine the performance monitoring result of the target device according to the first prediction result and the second prediction result.
[0119] In one embodiment, the first determination module 11 includes: a first prediction unit, a training unit, and a second prediction unit, wherein:
[0120] The first prediction unit is used to input the second data block into the first prediction model to obtain a first prediction result.
[0121] The training unit is used to train the first prediction model according to the second data block to obtain a second prediction model.
[0122] The second prediction unit is used to input the second data block into the second prediction model to obtain a second prediction result.
[0123] In one embodiment, the second prediction unit is specifically configured to use the monitoring data and label data in the second data block as training samples, perform incremental learning training on the first prediction model, and obtain the second prediction model.
[0124] In one embodiment, the second determination module 12 includes: a first determination unit and a second determination unit, wherein:
[0125] A first determining unit, configured to determine a similarity between the first prediction result and the second prediction result;
[0126] The second determining unit is used to determine the performance monitoring result of the target device according to the similarity and a preset adaptive threshold.
[0127] In one embodiment, the above-mentioned second determination unit is specifically used to determine that the performance monitoring result is that the target device has performance drift in the time period corresponding to the second data block if the similarity is greater than the adaptive threshold; if the similarity is not greater than the adaptive threshold, determine that the performance monitoring result is that the target device has not experienced performance drift in the time period corresponding to the second data block.
[0128] In one embodiment, the first determination unit is specifically used to determine the product value of the transpose of the first prediction result and the second prediction result; and determine the similarity between the first prediction result and the second prediction result based on the product value, the absolute value of the first prediction result and the absolute value of the second prediction result.
[0129] Each module in the above performance monitoring device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.
[0130] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig. 9As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store performance monitoring data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a performance monitoring method is implemented.
[0131] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0132] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:
[0133] The initial machine learning model is trained according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device;
[0134] According to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block are obtained; the first data block and the second data block are monitoring data of the target device in adjacent preset time periods;
[0135] A performance monitoring result of the target device is determined according to the first prediction result and the second prediction result.
[0136] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0137] Inputting the second data block into the first prediction model to obtain a first prediction result;
[0138] Training the first prediction model according to the second data block to obtain a second prediction model;
[0139] The second data block is input into the second prediction model to obtain a second prediction result.
[0140] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0141] The monitoring data and label data in the second data block are used as training samples, and incremental learning training is performed on the first prediction model to obtain a second prediction model.
[0142] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0143] Determining a similarity between the first prediction result and the second prediction result;
[0144] The performance monitoring result of the target device is determined based on the similarity and the preset adaptive threshold.
[0145] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0146] If the similarity is greater than the adaptive threshold, it is determined that the performance monitoring result is that the target device has a performance drift in the time period corresponding to the second data block;
[0147] If the similarity is not greater than the adaptive threshold, it is determined that the performance monitoring result is that the target device has no performance drift in the time period corresponding to the second data block.
[0148] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:
[0149] Determining a product value of the transpose of the first prediction result and the second prediction result;
[0150] The similarity between the first prediction result and the second prediction result is determined according to the product value, the absolute value of the first prediction result and the absolute value of the second prediction result.
[0151] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
[0152] The initial machine learning model is trained according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device;
[0153] According to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block are obtained; the first data block and the second data block are monitoring data of the target device in adjacent preset time periods;
[0154] A performance monitoring result of the target device is determined according to the first prediction result and the second prediction result.
[0155] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0156] Inputting the second data block into the first prediction model to obtain a first prediction result;
[0157] Training the first prediction model according to the second data block to obtain a second prediction model;
[0158] The second data block is input into the second prediction model to obtain a second prediction result.
[0159] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0160] The monitoring data and label data in the second data block are used as training samples, and incremental learning training is performed on the first prediction model to obtain a second prediction model.
[0161] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0162] Determining a similarity between the first prediction result and the second prediction result;
[0163] The performance monitoring result of the target device is determined based on the similarity and the preset adaptive threshold.
[0164] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0165] If the similarity is greater than the adaptive threshold, it is determined that the performance monitoring result is that the target device has a performance drift in the time period corresponding to the second data block;
[0166] If the similarity is not greater than the adaptive threshold, it is determined that the performance monitoring result is that the target device has no performance drift in the time period corresponding to the second data block.
[0167] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0168] Determining a product value of the transpose of the first prediction result and the second prediction result;
[0169] The similarity between the first prediction result and the second prediction result is determined according to the product value, the absolute value of the first prediction result and the absolute value of the second prediction result.
[0170] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:
[0171] The initial machine learning model is trained according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device;
[0172] According to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block are obtained; the first data block and the second data block are monitoring data of the target device in adjacent preset time periods;
[0173] A performance monitoring result of the target device is determined according to the first prediction result and the second prediction result.
[0174] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0175] Inputting the second data block into the first prediction model to obtain a first prediction result;
[0176] Training the first prediction model according to the second data block to obtain a second prediction model;
[0177] The second data block is input into the second prediction model to obtain a second prediction result.
[0178] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0179] The monitoring data and label data in the second data block are used as training samples, and incremental learning training is performed on the first prediction model to obtain a second prediction model.
[0180] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0181] Determining a similarity between the first prediction result and the second prediction result;
[0182] The performance monitoring result of the target device is determined based on the similarity and the preset adaptive threshold.
[0183] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0184] If the similarity is greater than the adaptive threshold, it is determined that the performance monitoring result is that the target device has a performance drift in the time period corresponding to the second data block;
[0185] If the similarity is not greater than the adaptive threshold, it is determined that the performance monitoring result is that the target device has no performance drift in the time period corresponding to the second data block.
[0186] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0187] Determining a product value of the transpose of the first prediction result and the second prediction result;
[0188] The similarity between the first prediction result and the second prediction result is determined according to the product value, the absolute value of the first prediction result and the absolute value of the second prediction result.
[0189] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.
[0190] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0191] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A performance monitoring method, characterized in that: The method comprises: The initial machine learning model is trained according to the first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is the status monitoring data of the target device; According to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block are obtained; the first data block and the second data block are monitoring data of adjacent preset time periods of the target device; A performance monitoring result of the target device is determined according to the first prediction result and the second prediction result.
2. The method according to claim 1, characterized in that The obtaining, according to the first prediction model and the second data block in the data to be predicted, a first prediction result and a second prediction result corresponding to the second data block includes: Inputting the second data block into the first prediction model to obtain the first prediction result; Training the first prediction model according to the second data block to obtain a second prediction model; The second data block is input into the second prediction model to obtain the second prediction result.
3. The method according to claim 2, characterized in that The step of training the first prediction model according to the second data block to obtain a second prediction model includes: The monitoring data and label data in the second data block are used as training samples, and incremental learning training is performed on the first prediction model to obtain the second prediction model.
4. The method according to any one of claims 1 to 3, characterized in that: Determining the performance monitoring result of the target device according to the first prediction result and the second prediction result includes: Determining a similarity between the first prediction result and the second prediction result; The performance monitoring result of the target device is determined according to the similarity and a preset adaptive threshold.
5. The method according to claim 4, characterized in that Determining the performance monitoring result of the target device according to the similarity and a preset adaptive threshold includes: If the similarity is greater than the adaptive threshold, determining that the performance monitoring result is that the target device has a performance drift in the time period corresponding to the second data block; If the similarity is not greater than the adaptive threshold, it is determined that the performance monitoring result is that no performance drift occurs to the target device in the time period corresponding to the second data block.
6. The method according to claim 4, characterized in that The first prediction result and the second prediction result are vectors of the same length, and determining the similarity between the first prediction result and the second prediction result includes: Determining a product value of the transpose of the first prediction result and the second prediction result; The similarity between the first prediction result and the second prediction result is determined according to the product value, the absolute value of the first prediction result and the absolute value of the second prediction result.
7. A performance monitoring device, characterized in that: The device comprises: A training model is used to train an initial machine learning model according to a first data block in the data to be predicted to obtain a first prediction model; the data to be predicted is status monitoring data of a target device; A first determination module, configured to obtain a first prediction result and a second prediction result corresponding to a second data block in the data to be predicted according to the first prediction model and the second data block; the first data block and the second data block are monitoring data of adjacent preset time periods of the target device; The second determination module is used to determine the performance monitoring result of the target device according to the first prediction result and the second prediction result.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.