Reliability prediction method and device, electronic equipment and storage medium
By performing abnormal detection and feature extraction on the historical experimental data of Stirling refrigerators, a reliability prediction model is built, and the problem of long periods of existing detection methods is solved, efficient reliability prediction is achieved, and the detection and detection of the refrigerator are shortened. Production cycle reduces costs.
Patent Information
- Application Number
- CN202411258200.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-09
- Publication Date
- 2025-01-03
AI Technical Summary
The existing Stirling refrigerator performance detection methods, such as sophisticated tests, have a long detection cycle and low efficiency, resulting in extended environmental experiment time and production cycle of the refrigerator, increasing costs.
By obtaining historical experimental data, performing abnormal detection and filling, extracting the characteristics of current and time series data, building a reliability prediction model, and using machine learning algorithms to make reliability prediction based on production data and current experimental data.
It realizes high-precision prediction of the reliability results of the refrigerator, shortens the detection cycle and production cycle, and saves the production and testing costs of the refrigerator.
Smart Images
Figure CN119249078A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of instrument control technology, and in particular to a reliability prediction method, device, electronic device and storage medium. Background Art
[0002] Stirling refrigerators can provide an effective cold source for low-temperature detectors, and their reliability will directly affect the service life of the equipment. Therefore, the long life and high reliability of Stirling refrigerators are the key to their application. Existing Stirling refrigerator performance testing methods include a variety of environmental experiments, such as long-term performance tests, room temperature running-in, temperature cycle tests, and aging tests, which can judge the reliability of Stirling refrigerator operation. However, some of these testing methods, such as aging tests, have a long testing cycle and low testing efficiency, which prolongs the environmental testing time and production cycle of the refrigerator. Summary of the invention
[0003] In order to solve the above technical problems, the embodiments of the present application provide a reliability prediction method, device, electronic device and storage medium.
[0004] In a first aspect, an embodiment of the present application provides a reliability prediction method, the method comprising: Acquire historical experimental data, and form an original data set based on the historical experimental data; Performing anomaly detection and filling on the original data set to obtain a corrected data set; Performing feature extraction on the current data and time series data in the corrected data set to obtain an optimized data set; A reliability prediction model is constructed based on the optimized data set and the preset algorithm, and reliability prediction of the target device is achieved based on the reliability prediction model, wherein the input parameters of the reliability prediction model are the production data and current experimental data of the target device, and the output parameter of the reliability prediction model is the reliability prediction result of the target device.
[0005] In one embodiment, the performing anomaly detection and filling on the original data set includes: Perform anomaly detection on the original data set based on the isolation forest algorithm, remove outliers, and use the locations of the outliers as missing values; The missing values are filled based on the K nearest neighbor algorithm.
[0006] In one embodiment, the feature extraction of the current data in the modified data set includes: Performing Fourier transformation on the current sampling data of the target device in steady state to obtain a current characteristic vector, wherein the current characteristic vector includes a current fundamental wave and each harmonic component; All the current eigenvectors are combined to obtain a current eigenvector matrix.
[0007] In one embodiment, the extracting features from the time series data in the modified data set includes: The time series data is subjected to first-order difference processing to obtain time-derived features.
[0008] In one embodiment, obtaining the optimized data set includes: The optimized data set is obtained based on the current feature vector matrix and the time-derived features.
[0009] In one embodiment, the building of a reliability prediction model based on the optimized data set and a preset algorithm includes: Constructing an initial prediction model; Determine a training set and a test set based on the optimized data set, and train the initial prediction model based on the training set and the test set; The model parameters of the initial prediction model are optimized based on Bayesian optimization to obtain the reliability prediction model.
[0010] In one embodiment, the production data at least includes equipment size data, and the current experimental data at least includes high and low temperature environment experimental current detection data and refrigerator speed.
[0011] In a second aspect, an embodiment of the present application provides a reliability prediction device, the device comprising: An acquisition module, used to acquire historical experimental data and form an original data set based on the historical experimental data; An anomaly detection module, used to perform anomaly detection and filling on the original data set to obtain a corrected data set; A feature extraction module, used for extracting features from the current data and time series data in the modified data set to obtain an optimized data set; A construction module is used to construct a reliability prediction model based on the optimized data set and a preset algorithm, and realize the reliability prediction of the target device based on the reliability prediction model, wherein the input parameters of the reliability prediction model are the production data and current experimental data of the target device, and the output parameter of the reliability prediction model is the reliability prediction result of the target device.
[0012] In a third aspect, an embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and when the computer program is run by the processor, the reliability prediction method provided in the first aspect is executed.
[0013] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program, which executes the reliability prediction method provided in the first aspect when running on a processor.
[0014] In summary, the reliability prediction method provided by the present application includes: acquiring historical experimental data, forming an original data set based on the historical experimental data; performing anomaly detection and filling on the original data set to obtain a corrected data set; performing feature extraction on the current data and time series data in the corrected data set to obtain an optimized data set; constructing a reliability prediction model based on the optimized data set and a preset algorithm, and realizing the reliability prediction of the target device based on the reliability prediction model. The reliability results of the refrigerator can be predicted through easily obtained experimental data with high accuracy, and the reliability of the refrigerator can be effectively judged in advance, thereby shortening the environmental experiment time and production cycle of the refrigerator, thereby greatly saving the production and testing costs of the refrigerator. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present application and should not be regarded as limiting the scope of protection of the present application. In each of the drawings, similar components are numbered similarly.
[0016] Figure 1 A schematic diagram of a reliability prediction method provided in an embodiment of the present application is shown; Figure 2 A schematic flow chart of a sub-step of the reliability prediction method provided in an embodiment of the present application is shown; Figure 3 A schematic flow chart showing another sub-step of the reliability prediction method provided in an embodiment of the present application is shown; Figure 4 A schematic flow chart showing another sub-step of the reliability prediction method provided in an embodiment of the present application is shown; Figure 5 A schematic diagram of the structure of a reliability prediction device provided in an embodiment of the present application is shown; Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.
[0018] The components of the embodiments of the present application generally described and shown in the drawings herein may be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.
[0019] Hereinafter, the terms "including", "having" and their cognates, which may be used in various embodiments of the present application, are intended only to indicate specific features, numbers, steps, operations, elements, components, or combinations of the foregoing items, and should not be understood as first excluding the existence of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing items or adding the possibility of one or more features, numbers, steps, operations, elements, components, or combinations of the foregoing items.
[0020] Furthermore, the terms “first”, “second”, “third”, etc. are merely used for distinguishing descriptions and are not to be understood as indicating or implying relative importance.
[0021] Unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meanings as those generally understood by those skilled in the art to which the various embodiments of the present application belong. The terms (such as those defined in generally used dictionaries) will be interpreted as having the same meanings as the contextual meanings in the relevant technical field and will not be interpreted as having idealized meanings or overly formal meanings unless clearly defined in the various embodiments of the present application.
[0022] As mentioned in the above background technology, the research on the long life and high reliability of Stirling refrigerator is the key technology currently applied. The existing Stirling refrigerator performance detection system mainly judges the operating reliability of Stirling refrigerator through long-term performance testing, room temperature running-in, temperature cycle and aging. However, some of these methods, such as aging test, can detect the reliability of refrigerator operation, but the cycle time for refrigerator reliability detection is too long, and it may also reduce the effective life of the refrigerator. This leads to an increase in the production cycle and cost of the refrigerator.
[0023] Machine learning can be applied to aspects such as data prediction. Therefore, the present application proposes a method for predicting and analyzing a portion of easily accessible data based on a machine learning algorithm to obtain the results of refrigerator reliability. For example, the refrigerator aging results can be obtained by predicting and analyzing the temperature cycle detection data. This can not only reduce the detection cycle of the refrigerator, but also shorten the overall production cycle, which has good economy and timeliness.
[0024] For details, see Figure 1 , this application provides a reliability prediction method, and the specific implementation steps include: Step S110, acquiring historical experimental data, and forming an original data set based on the historical experimental data; The historical experimental data here include various experimental data obtained from various performance tests of the target equipment during historical production and experiments, including data such as normal temperature running-in, temperature cycle test, and aging test, and the above data are integrated to obtain the original data set. In addition, it is also necessary to judge the positive and negative nature of the data samples in the data set through expert experience for subsequent model construction and training.
[0025] The data type of historical experimental data is related to the experimental type used in the historical experiment. For example, if the historical experiment is a temperature cycle test and a aging test, the technical indicators of the temperature cycle test include high temperature, high temperature holding time, drop rate, low temperature, low temperature holding time, rise rate, number of cycles, etc.; the parameters of the aging test include: the current size of the equipment circuit, the harmonic level, etc.
[0026] Step S120, performing anomaly detection and filling on the original data set to obtain a corrected data set; The directly collected data may have problems such as data anomalies or missing data, so the original data set needs to be corrected. The correction in this implementation includes two parts: outlier detection and missing value filling.
[0027] See also Figure 2 In this embodiment, the anomaly detection and filling of the original data set includes: S201, performing anomaly detection on the original data set based on an isolation forest algorithm, removing outliers, and treating the locations of the outliers as missing values.
[0028] S202: Fill in the missing values based on a K-nearest neighbor algorithm.
[0029] The specific anomaly detection steps are: Randomly select n sample points from the collected data set to form a subset, build a tree on the m subsets, randomly select a feature, randomly select a threshold for binary split, and define the probability of anomaly based on the average depth of the m decision trees. See the following formula: in, For sample data in The anomaly index in the iTree trained with the data, is the expected path length of data in all iTrees, is the path length of each tree.
[0030] The path length of each tree The calculation formula can be found in the following formula: in, n is the number of data, To have The height of a binary tree of nodes.
[0031] The specific steps for filling missing values are: The K nearest neighbor algorithm is used to fill the missing values of the missing data, and the standardized Euclidean distance is used. The sampling time scale is , each sampling index value is , then the refrigerator reliability characteristic vector is ,in , is the sampling number, , is the number of sampling indicators.
[0032] For missing data , the missing value points and matching points corresponding to the outlier positions are detected through the anomaly detection step distance.
[0033] The specific missing value filling algorithm can be found in the following formula in, is the complete data set for each feature indicator, is the standard deviation of each dimension, express The distance between the matching point and After calculating the distance from the matching point, select several groups of items with smaller distances to be matched As a reference vector, missing values are obtained by averaging The value of .
[0034] Step S130, performing feature extraction on the current data and time series data in the corrected data set to obtain an optimized data set; See also Figure 3 In this embodiment, the feature extraction of the current data in the corrected data set includes: S301, performing Fourier transform on the current sampling data of the target device in steady state to obtain a current characteristic vector, wherein the current characteristic vector includes a current fundamental wave and each harmonic component; S302: Combine all the current eigenvectors to obtain a current eigenvector matrix.
[0035] According to the start and end time of the circuit transient process (the process from one stable state to another stable state) of the target device, the fast Fourier transform (FFT) is used on the current sampling data in the steady state to obtain the eigenvector composed of the current fundamental wave and each harmonic component. All eigenvectors are combined together to form an eigenvector matrix.
[0036] See also Figure 4 In this embodiment, the feature extraction of the time series data in the modified data set includes: S401, performing first-order difference processing on the time series data to obtain time-derived features.
[0037] In some embodiments, modeling requires the use of stable time series data. A stable time series will not be affected by the time when the series is observed, that is, mathematically speaking, if the mean and variance of a time series remain unchanged and the covariance is independent of time, then the time series is stable. However, since the original time series data set may contain trends and seasonality, the trend may cause the mean of the time series to change over time, and the seasonality may cause the variance of the time series to change over time, thus causing the time series to be unstable. Performing first-order difference processing on the original time series data can remove the influence of trends and seasonality, making unstable time series stable, and obtaining time-derived features that are more conducive to modeling.
[0038] In this embodiment, obtaining the optimized data set includes: obtaining the optimized data set based on the current feature vector matrix and the time-derived feature.
[0039] The original current sampling data and time series data are not suitable for model training or cannot achieve good training results. Therefore, it is necessary to extract features and form a new optimized data set based on the current feature vector matrix and time-derived features obtained by feature extraction.
[0040] Step S140, constructing a reliability prediction model based on the optimized data set and the preset algorithm, and realizing the reliability prediction of the target device based on the reliability prediction model, wherein the input parameters of the reliability prediction model are the production data and current experimental data of the target device, and the output parameters of the reliability prediction model are the reliability prediction results of the target device. Exemplarily, the reliability prediction result may be a quantitative result, such as a score for the reliability of the target device, or a qualitative result, for example, when the reliability score or index is greater than a preset threshold, it indicates that the target device is reliable, otherwise it is unreliable.
[0041] In this embodiment, the reliability prediction model is constructed based on the optimized data set and the preset algorithm, including: Constructing an initial prediction model; determining a training set and a test set based on the optimized data set, and training the initial prediction model based on the training set and the test set; optimizing the model parameters of the initial prediction model based on Bayesian optimization to obtain the reliability prediction model.
[0042] Exemplarily, the method used in constructing the model in this embodiment is the lightGBM (Light Gradient Boosting Machine) algorithm, which is a distributed gradient boosting framework based on the decision tree algorithm. The LightGBM algorithm can reduce the use of memory by data, ensuring that a single machine can use as much data as possible without sacrificing speed; reduce the cost of communication, improve the efficiency of multiple machines in parallel, and achieve linear acceleration in calculations. The LightGBM algorithm is suitable for larger data sets, and it can achieve fast and efficient, low memory usage, high accuracy, and support parallel and large-scale data processing.
[0043] A refrigerator reliability prediction model is constructed, and the current refrigerator production and environmental test related data are used as the input parameters of the model, and whether the refrigerator is qualified is used as the output parameter of the model.
[0044] Exemplarily, the processed optimized data set can be divided into a training set and a test set, and the model is trained through 5-fold cross validation, that is, the optimized data set is divided into 5 subsets, 4 of which are used for training.
[0045] During the training process, the algorithm parameters of LightGBM can also be optimized through Bayesian Optimization. The LightGBM algorithm uses a gradient boosting decision tree. The basic parameters of the LightGBM algorithm include: optimal leaf weight, number of leaves, maximum depth of the tree, feature sampling value, etc. Bayesian optimization constructs a probability model, such as Gaussian process regression, which represents the probability distribution of the objective function. In each iteration, the algorithm selects the next test point that is most likely to provide the greatest benefit. As more data points are added, the model will be continuously updated to guide the search process to gradually approach the optimal solution.
[0046] In this embodiment, the production data at least includes equipment size data, and the current experimental data at least includes high and low temperature environment experimental current detection data and refrigerator speed.
[0047] The data related to refrigerator production and environmental testing may include refrigerator size data, high and low temperature environmental testing current detection data, refrigerator speed data, etc. When making predictions, the above data is input into the trained model for prediction, and the output result is the final result of the environmental experiment. This application can only input easily accessible experimental data (such as temperature cycle test data) to predict the final result of the refrigerator environmental experiment (including aging test data, etc.).
[0048] The reliability prediction method provided in this embodiment forms an original data set based on historical experimental data, performs anomaly detection and filling on the original data set, and extracts features from the current data and time series data to obtain an optimized data set. A reliability prediction model is constructed based on the optimized data set and the lightGBM algorithm, and reliability prediction of the target device is achieved based on the reliability prediction model. The present application can predict the reliability results of the refrigerator through partial production and experimental data with high accuracy, and can effectively determine the reliability of the refrigerator in advance, thereby shortening the environmental test time and production cycle of the refrigerator, avoiding the loss of refrigerant life caused by some tests, and thus greatly saving the production and testing costs of the refrigerator.
[0049] Specifically, Figure 5 As shown, the present application also provides a reliability prediction device 500, comprising: An acquisition module 510 is used to acquire historical experimental data and form an original data set based on the historical experimental data; Anomaly detection module 520, used to perform anomaly detection and filling on the original data set to obtain a corrected data set; A feature extraction module 530 is used to extract features from the current data and time series data in the modified data set to obtain an optimized data set; Construction module 540 is used to construct a reliability prediction model based on the optimized data set and the preset algorithm, and realize the reliability prediction of the target device based on the reliability prediction model, wherein the input parameters of the reliability prediction model are the production data and current experimental data of the target device, and the output parameter of the reliability prediction model is the reliability prediction result of the target device.
[0050] In one embodiment, the anomaly detection module 520 is further used to perform anomaly detection on the original data set based on an isolation forest algorithm, remove outliers, and use the location of the outliers as missing values; The missing values are filled based on the K nearest neighbor algorithm.
[0051] In one embodiment, the feature extraction module 530 is further used to perform Fourier transform on the current sampling data of the target device in steady state to obtain a current feature vector, wherein the current feature vector includes a current fundamental wave and each harmonic component; All the current eigenvectors are combined to obtain a current eigenvector matrix.
[0052] In one embodiment, the feature extraction module 530 is further used to perform first-order difference processing on the time series data to obtain time-derived features.
[0053] In one embodiment, the feature extraction module 530 is further configured to obtain the optimized data set based on the current feature vector matrix and the time-derived feature.
[0054] In one embodiment, the construction module 540 is also used to construct an initial prediction model; determine a training set and a test set based on the optimized data set, and train the initial prediction model based on the training set and the test set; optimize the model parameters of the initial prediction model based on Bayesian optimization to obtain the reliability prediction model.
[0055] The reliability prediction device 500 provided in this embodiment can implement the reliability prediction method provided in Embodiment 1, and will not be described again here to avoid repetition.
[0056] The reliability prediction device provided in this embodiment forms an original data set based on historical experimental data, performs anomaly detection and filling on the original data set, and extracts features from the current data and time series data to obtain an optimized data set. A reliability prediction model is constructed based on the optimized data set and the lightGBM algorithm, and reliability prediction of the target device is achieved based on the reliability prediction model. The present application can predict the reliability results of the refrigerator through partial production and experimental data with high accuracy, and can effectively determine the reliability of the refrigerator in advance, thereby shortening the environmental test time and production cycle of the refrigerator, avoiding the loss of refrigerant life caused by some tests, and thus greatly saving the production and testing costs of the refrigerator.
[0057] In addition, an embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program runs on the processor, the reliability prediction method provided in Embodiment 1 is executed.
[0058] For details, see Figure 6 , the electronic device 600 includes: a transceiver 601, a bus interface and a processor 602, the processor 602 is used to: obtain historical experimental data, and form an original data set based on the historical experimental data; Performing anomaly detection and filling on the original data set to obtain a corrected data set; Performing feature extraction on the current data and time series data in the corrected data set to obtain an optimized data set; A reliability prediction model is constructed based on the optimized data set and the preset algorithm, and reliability prediction of the target device is achieved based on the reliability prediction model, wherein the input parameters of the reliability prediction model are the production data and current experimental data of the target device, and the output parameter of the reliability prediction model is the reliability prediction result of the target device.
[0059] In this embodiment, the processor 602 is further used to perform anomaly detection on the original data set based on the isolation forest algorithm, remove outliers, and use the locations of the outliers as missing values; and fill the missing values based on the K nearest neighbor algorithm.
[0060] In this embodiment, the processor 602 is further used to perform Fourier transform on the current sampling data of the target device in steady state to obtain a current characteristic vector, wherein the current characteristic vector includes a current fundamental wave and each harmonic component; and combine all the current characteristic vectors to obtain a current characteristic vector matrix.
[0061] In this embodiment, the processor 602 is further configured to perform first-order difference processing on the time series data to obtain time-derived features.
[0062] In this embodiment, the processor 602 is further configured to obtain the optimized data set based on the current feature vector matrix and the time-derived feature.
[0063] In this embodiment, the processor 602 is also used to construct an initial prediction model; determine a training set and a test set based on the optimized data set, and train the initial prediction model based on the training set and the test set; optimize the model parameters of the initial prediction model based on Bayesian optimization to obtain the reliability prediction model.
[0064] In the embodiment of the present invention, the electronic device 600 further includes a memory 603. Figure 6In the embodiment, the bus architecture may include any number of interconnected buses and bridges, specifically linking various circuits of one or more processors represented by processor 602 and memory represented by memory 603. The bus architecture may also link various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface. The transceiver 601 may be a plurality of components, i.e., including a transmitter and a receiver, providing a unit for communicating with various other devices on a transmission medium. The processor 602 is responsible for managing the bus architecture and general processing, and the memory 603 may store data used by the processor 602 when performing operations.
[0065] The electronic device provided in this embodiment forms an original data set based on historical experimental data, performs anomaly detection and filling on the original data set, and extracts features from the current data and time series data to obtain an optimized data set. A reliability prediction model is constructed based on the optimized data set and the lightGBM algorithm, and reliability prediction of the target device is achieved based on the reliability prediction model. The present application can predict the reliability results of the refrigerator through partial production and experimental data with high accuracy, and can effectively determine the reliability of the refrigerator in advance, thereby shortening the environmental test time and production cycle of the refrigerator, avoiding the loss of refrigerant life caused by some tests, and thus greatly saving the production and testing costs of the refrigerator.
[0066] The present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the reliability prediction method provided in Example 1 is implemented.
[0067] In this embodiment, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc., and the readable storage medium may be a volatile storage medium or a non-volatile storage medium.
[0068] The computer-readable storage medium provided in this embodiment can implement the reliability prediction method provided in Example 1, and will not be described again here to avoid repetition.
[0069] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or terminal including the element.
[0070] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.
[0071] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.
Claims
1. A reliability prediction method, characterized in that: The method comprises: Acquire historical experimental data, and form an original data set based on the historical experimental data; Performing anomaly detection and filling on the original data set to obtain a corrected data set; Performing feature extraction on the current data and time series data in the corrected data set to obtain an optimized data set; A reliability prediction model is constructed based on the optimized data set and the preset algorithm, and reliability prediction of the target device is achieved based on the reliability prediction model, wherein the input parameters of the reliability prediction model are the production data and current experimental data of the target device, and the output parameter of the reliability prediction model is the reliability prediction result of the target device.
2. The reliability prediction method according to claim 1, characterized in that: The performing anomaly detection and filling on the original data set includes: Perform anomaly detection on the original data set based on the isolation forest algorithm, remove outliers, and use the locations of the outliers as missing values; The missing values are filled based on the K nearest neighbor algorithm.
3. The reliability prediction method according to claim 1, characterized in that: The extracting features of the current data in the corrected data set includes: Performing Fourier transformation on the current sampling data of the target device in steady state to obtain a current characteristic vector, wherein the current characteristic vector includes a current fundamental wave and each harmonic component; All the current eigenvectors are combined to obtain a current eigenvector matrix.
4. The reliability prediction method according to claim 3, characterized in that: The extracting features of the time series data in the corrected data set includes: The time series data is subjected to first-order difference processing to obtain time-derived features.
5. The reliability prediction method according to claim 4, characterized in that: The optimized data set comprises: The optimized data set is obtained based on the current feature vector matrix and the time-derived features.
6. The reliability prediction method according to claim 1, characterized in that: The constructing of a reliability prediction model based on the optimized data set and a preset algorithm includes: Constructing an initial prediction model; Determine a training set and a test set based on the optimized data set, and train the initial prediction model based on the training set and the test set; The model parameters of the initial prediction model are optimized based on Bayesian optimization to obtain the reliability prediction model.
7. The reliability prediction method according to claim 1, characterized in that: The production data at least includes equipment size data, and the current experimental data at least includes high and low temperature environment experimental current detection data and refrigerator speed.
8. A reliability prediction device, characterized in that: The device comprises: An acquisition module, used to acquire historical experimental data and form an original data set based on the historical experimental data; An anomaly detection module, used to perform anomaly detection and filling on the original data set to obtain a corrected data set; A feature extraction module, used for extracting features from the current data and time series data in the modified data set to obtain an optimized data set; A construction module is used to construct a reliability prediction model based on the optimized data set and a preset algorithm, and realize the reliability prediction of the target device based on the reliability prediction model, wherein the input parameters of the reliability prediction model are the production data and current experimental data of the target device, and the output parameter of the reliability prediction model is the reliability prediction result of the target device.
9. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is run by the processor, the reliability prediction method according to any one of claims 1 to 7 is executed.
10. A computer-readable storage medium, characterized in that: The device stores a computer program, which executes the reliability prediction method according to any one of claims 1 to 7 when running on a processor.