Service life prediction method and device based on optimized support vector machine, equipment and medium

Abstract

The invention discloses a service life prediction method and device based on an optimized support vector machine, equipment and a medium, and relates to the technical field of prediction and health management, and the method comprises the steps: obtaining the information data of target equipment, carrying out the feature extraction of the information data, obtaining the feature data, processing the feature data according to a particle swarm optimization algorithm, and obtaining the service life of the target equipment; and obtaining the prediction information data, and obtaining the remaining service life of the target equipment according to the prediction information data. And through integrated data processing and an optimization algorithm, the accuracy and the efficiency of predicting the residual life of the equipment are improved.

Description

Technical Field

[0001] The present invention relates to the field of prediction and health management technology, and in particular to a life prediction method, device, equipment and medium based on an optimized support vector machine. Background Art

[0002] In the current field of information management and equipment health management, especially in the context of widespread application of Internet of Things technology, effective life prediction and health management of key equipment and systems based on optimized support vector machines has become an important part of ensuring industrial production safety and improving operational efficiency. With the rapid development of the industrial Internet, equipment maintenance has shifted from passive response to failures to active preventive maintenance, which has put forward higher requirements for real-time status monitoring, fault prediction and life assessment of equipment. As one of the core technologies of equipment health management, the life prediction algorithm based on optimized support vector machines is crucial for optimizing maintenance strategies, reducing operation and maintenance costs, and extending the service life of equipment.

[0003] Prediction and health management mainly revolve around the assessment and health management of remaining useful life (RUL), among which the life prediction method based on optimized support vector machine based on machine learning occupies an important position. Machine learning algorithms, including supervised learning, unsupervised learning and reinforcement learning, have been widely used in life prediction based on optimized support vector machine. Among them, representatives of supervised learning such as support vector machine (SVM) and artificial neural network (ANN), as well as clustering algorithms and dimensionality reduction algorithms of unsupervised learning, are trying to extract information from a large amount of condition monitoring data to predict the remaining useful life of equipment. In particular, the life prediction algorithm based on optimized support vector machine based on improved support vector machine selects the features most related to the fault through correlation analysis, and uses particle swarm optimization (PSO) to adjust the support vector machine parameters, which improves the generalization ability of prediction and the online application effect.

[0004] In some cases, the machine learning-based optimized support vector machine-based life prediction algorithm has poor generalization ability and is difficult to adapt to various complex industrial application scenarios, especially for online applications of real-time data, which leads to challenges in actual deployment. When processing actual data, it is limited by data quality, quantity and labeling status. Especially in the process of data preprocessing, feature selection and model training, there are problems such as sensitivity to noise, feature redundancy or missing, which affect the prediction accuracy. Although the existing machine learning-based optimized support vector machine-based life prediction algorithm has made progress in the field of health management-based optimized support vector machine-based life prediction, there is an urgent need for more advanced, efficient and good generalization prediction algorithms. Summary of the invention

[0005] The main purpose of this application is to provide a life prediction method, device, equipment and medium based on an optimized support vector machine, aiming to solve the technical problem of how to improve the efficiency and accuracy of equipment life prediction.

[0006] To achieve the above objectives, the present application proposes a life prediction method based on an optimized support vector machine, the method comprising:

[0007] Get information data of the target device;

[0008] Extracting features from the information data to obtain feature data;

[0009] Processing the characteristic data according to the life prediction model to obtain prediction information data;

[0010] The remaining service life of the target device is obtained according to the prediction information data.

[0011] Optionally, before the step of processing the characteristic data according to the life prediction model to obtain prediction information data, the step includes:

[0012] Dividing the feature data into a training set, a validation set and a test set;

[0013] Obtain training set feature information and establish an initial lifespan prediction model;

[0014] The initial life prediction model is trained according to the feature information of the training set using a particle swarm optimization algorithm to obtain a preset life prediction model.

[0015] Optionally, the step of training the initial life prediction model according to the training set feature information using a particle swarm optimization algorithm to obtain a preset life prediction model includes:

[0016] Obtain the particle swarm in the training set feature information;

[0017] Initializing the positions and velocities of particles in the particle swarm and calculating the fitness of the particles;

[0018] Set the number of iterations and calculate the historical optimal value of the particle;

[0019] By comparing the fitness of the particles in each iteration, a global optimal value of the particle swarm is obtained;

[0020] According to the historical optimal value and the global optimal value, the position and speed of the particle are updated;

[0021] Determine whether the position and speed of the particle meet a preset range, and if not, perform boundary condition processing on the particle;

[0022] If the number of iterations is reached or the error is less than the preset error, the preset life prediction model is output.

[0023] Optionally, after the step of training the initial life prediction model according to the training set feature information using a particle swarm optimization algorithm to obtain a preset life prediction model, the method further comprises:

[0024] Evaluating the preset life prediction model to obtain an evaluation result;

[0025] If the evaluation result does not meet the prediction requirement, retrain the preset life prediction model until it meets the prediction requirement;

[0026] If the evaluation result meets the prediction requirements, the preset life prediction model is used as the final life prediction model, and the remaining service life is analyzed.

[0027] Optionally, the step of acquiring information data of the target device includes:

[0028] Get the original information data of the target device;

[0029] Performing denoising on the original information data according to wavelet transform to obtain denoised information data;

[0030] The information data after the noise removal is subjected to time-frequency analysis and normalization processing to obtain the information data of the target device.

[0031] Optionally, the step of extracting features from the information data to obtain feature data includes:

[0032] Decomposing the information data at multiple levels to obtain signals in multiple frequency bands;

[0033] Performing Fourier transform on the signals in the multiple frequency bands to obtain frequency domain signals;

[0034] Performing sliding window analysis on the frequency domain signal to obtain distribution characteristics;

[0035] The characteristic data is obtained by selecting according to the distribution characteristics.

[0036] Optionally, before the step of obtaining the remaining useful life according to the prediction information data, the following steps are included:

[0037] Obtain sample information data;

[0038] Establishing a linear risk model based on the information sample data;

[0039] The step of obtaining the remaining service life according to the prediction information data further includes:

[0040] Obtaining risk probability according to the predicted information data through the linear risk model;

[0041] Comparing the risk probability with the preset damage probability, obtaining the survival probability;

[0042] The remaining service life is calculated based on the survival probability.

[0043] In addition, to achieve the above purpose, the present application also proposes a life prediction device based on an optimized support vector machine, the life prediction device based on an optimized support vector machine comprising:

[0044] An acquisition module is used to acquire information data of a target device;

[0045] A feature extraction module, used to extract features from the information data to obtain feature data;

[0046] An acquisition module, used for processing the characteristic data according to the life prediction model to obtain prediction information data;

[0047] The result module is used to obtain the remaining service life of the target device according to the prediction information data.

[0048] In addition, to achieve the above-mentioned purpose, the present application also proposes a medium, which is a computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the life prediction method based on the optimized support vector machine as described above are implemented.

[0049] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the life prediction method based on the optimized support vector machine as described above.

[0050] This application obtains information data of the target device, extracts features from the information data, obtains feature data, processes the feature data according to the particle swarm optimization algorithm, obtains predicted information data, and obtains the remaining service life of the target device based on the predicted information data. The accuracy and efficiency of the prediction of the remaining service life of the equipment are improved through integrated data processing and optimization algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0052] Figure 1 This is a flow chart of the first embodiment of the life prediction method based on optimized support vector machine of this application;

[0053] Figure 2 Vibration and temperature signal diagram of the collected samples of the first embodiment of the life prediction method based on the optimized support vector machine of this application;

[0054] Figure 3 This is a schematic diagram of the wavelet transform processing result of the first embodiment of the life prediction method based on the optimized support vector machine of this application;

[0055] Figure 4 This is a standard deviation prediction result diagram of the first embodiment of the life prediction method based on the optimized support vector machine of this application;

[0056] Figure 5 This is a diagram showing the root amplitude prediction result of the first embodiment of the life prediction method based on the optimized support vector machine of the present application;

[0057] Figure 6 This is a diagram of the root mean square prediction results of the first embodiment of the life prediction method based on the optimized support vector machine of this application;

[0058] Figure 7 This is a flow chart of a second embodiment of a life prediction method based on an optimized support vector machine of the present application;

[0059] Figure 8 This is a risk probability assessment result diagram of the second embodiment of the life prediction method based on the optimized support vector machine of this application;

[0060] Fig. 9 This is a graph showing the survival probability assessment result of the second embodiment of the life prediction method based on the optimized support vector machine of the present application;

[0061] Fig.10 This is a comparison diagram of the predicted absolute relative errors of the second embodiment of the life prediction method based on the optimized support vector machine of this application;

[0062] Fig.11 This is a schematic diagram of the module structure of a life prediction device based on an optimized support vector machine according to an embodiment of the present application;

[0063] Fig.12 A schematic diagram of the device structure of the hardware operating environment involved in the life prediction method based on the optimized support vector machine in the embodiment of the present application;

[0064] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0065] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0066] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0067] The main solution of the embodiment of the present application is: obtaining information data of the target device, performing feature extraction on the information data to obtain feature data, processing the feature data according to a particle swarm optimization algorithm to obtain predicted information data, and obtaining the remaining service life of the target device based on the predicted information data.

[0068] Based on this, the embodiment of the present application provides a life prediction method based on an optimized support vector machine, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the life prediction method based on optimized support vector machine of this application.

[0069] In this embodiment, the life prediction method based on optimized support vector machine includes steps S10 to S40:

[0070] Step S10, obtaining information data of the target device.

[0071] It should be noted that obtaining information data of the target equipment covers a wide range of monitoring technologies and data collection methods. First, sensor technology, such as vibration sensors, temperature sensors, pressure sensors, etc., is used to comprehensively monitor the operating status of the equipment. These sensors are precisely arranged in key parts of the equipment to continuously collect data at a high sampling rate to ensure the real-time and integrity of the data. The panoramic view of the test bench includes rotating parts, decay generating parts and measuring parts. The rotating parts are driven by a motor with a power of 250W and a speed of 1500rpm; the decay generating part is the core part of the entire test bench. A load exciter generates a dynamic load exceeding the maximum bearing of the bearing, that is, 4000N, during the operation of the bearing. The vibration sensors and temperature sensors installed on the test bench complete the task of collecting data. Figure 2 The vibration and temperature signal diagram of the collected samples shown in the figure is based on the data monitored under the condition of bearing load 5000N. A total of 1111040 vibration signal samples and 42717 temperature signal samples were monitored during the whole process. The sampling frequencies of vibration and temperature signals were 25.6kHz and 0.1Hz respectively. In order to avoid damage to the test bench, the data monitoring behavior was terminated when the vibration amplitude of the bearing exceeded 20g.

[0072] Furthermore, before obtaining the information data of the target device, the information data must undergo a strict preprocessing stage, including denoising, filtering, signal smoothing and other operations to eliminate environmental interference and accidental errors and ensure data quality. For example, the use of wavelet transform or frequency domain filtering technology can effectively separate the noise component in the signal and highlight the true reflection of the device status. In addition, the data must also be standardized, such as normalization or standardization, to ensure that the data collected by different sensors are comparable on a unified scale, which is convenient for subsequent analysis. Figure 3 The schematic diagram of the wavelet transform processing result shown in the figure first obtains the original information data of the target device, then performs denoising on the original information data according to the wavelet transform to obtain the denoised information data, performs time-frequency analysis and normalization on the denoised information data, and finally obtains the information data of the target device.

[0073] Step S20, extracting features from the information data to obtain feature data.

[0074] It should be noted that feature extraction is to convert pre-processed data into meaningful characteristic parameters that are closely related to the health status of the equipment. This step usually involves statistical analysis, spectrum analysis, time domain analysis and other technologies to extract key indicators that can reflect equipment wear, fatigue, and performance degradation from a large amount of raw data. For example, the root mean square value, kurtosis, peak factor, etc. of the vibration signal are extracted, which can indicate the vibration intensity, irregularity and potential failure trend of the equipment.

[0075] Furthermore, the information data is decomposed at multiple levels to obtain signals in multiple frequency bands. By decomposing the complex original signal into multiple frequency bands, it helps to reveal the impact of different frequency components on the change of device status. This process usually uses technologies such as wavelet transform or empirical mode decomposition (EMD). The former can provide signal details and overviews at different scales, while the latter can adaptively decompose the signal into a series of intrinsic mode functions (IMFs), each of which corresponds to a specific frequency band, making the frequency structure of the signal clearer.

[0076] Furthermore, the multiple frequency band signals are subjected to Fourier transform to obtain frequency domain signals. The decomposed multiple frequency band signals are subjected to Fourier transform, which is an important step to convert the time domain signal into the frequency domain signal. Frequency domain analysis can reveal the energy distribution of the signal and help identify the characteristic frequencies of the equipment operation status, such as harmonics, sidebands or subharmonics that are unique to the fault. Through spectrum analysis, fault signs that are not easy to directly observe in the time domain can be found, providing more abundant fault mode information for the prediction model.

[0077] Furthermore, the frequency domain signal is subjected to sliding window analysis to obtain distribution characteristics. Based on the frequency domain signal, further implementation of sliding window analysis is a dynamic analysis strategy that gradually analyzes the frequency domain characteristics of each time period by moving a fixed-length window on the signal sequence. This analysis can capture the changing trend of the signal over time and extract features that reflect the dynamic evolution of the device status, such as frequency drift and spectrum slope changes. The selection of the sliding window size is usually based on the signal characteristics and prediction objectives. It should be able to capture the instantaneous changes of the signal while avoiding noise amplification caused by over-refinement.

[0078] Furthermore, feature data is obtained based on the distribution feature selection. The distribution features extracted based on the above steps contain multi-dimensional information about the health status of the equipment, such as the peak position of the spectrum, the spectrum density distribution, the proportional relationship of the frequency components, etc. Through a carefully designed feature selection algorithm, such as correlation analysis, mutual information or principal component analysis (PCA), the time dimension features closely related to equipment failure are selected from a large number of distribution features, such as the peak-to-peak value (x r ), mean(x m ), peak value (x p ), root amplitude (x ra ), RMS(x rms ), standard deviation (x std ), skewness(x ske ), kurtosis (x k ), peak index (x c ), margin index (x ma ), Waveform Indicator (x sh a) and impulse indicators (x i ). The first five features can reflect the amplitude and energy of the signal per unit time, and the last seven features can represent the time series distribution of the signal per unit time. The specific mathematical expression of the features is:

[0079] x r =max(x(n))-min(x(n))

[0080]

[0081] x p =maxx(n)

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091] Where x(n) is a signal sequence with a length of N, and N represents the number of samples. The most relevant feature is selected by comparing the correlation coefficient table of the 12 features with the operating status of the bearing.

[0092] Table 1 Operation status correlation coefficient table

[0093]

[0094] By comparing the correlation coefficients, it can be clearly observed that the three characteristics of standard deviation, root mean square and root amplitude are most closely related to the operating status. Therefore, in this embodiment, standard deviation, root mean square and root amplitude are selected as inputs of the life prediction model.

[0095] Step S30, processing the characteristic data according to the life prediction model to obtain prediction information data.

[0096] It is important to understand that the feature data is divided into training set, validation set and test set, and the feature information of the training set is obtained, and the initial life prediction model is established. The support vector machine (SVM) model is very suitable for processing small samples, nonlinear classification and regression problems due to its excellent generalization ability. The initial life prediction model is established as the SVM model.

[0097] Furthermore, the initial life prediction model is trained according to the particle swarm optimization algorithm based on the characteristic information of the training set to obtain the preset life prediction model. In order to optimize the performance of the SVM model, the particle swarm optimization algorithm is introduced to automatically adjust the model parameters. PSO guides particles (i.e. parameter configuration) to find the optimal solution in the search space by simulating the foraging behavior of bird flocks. Each particle represents a possible combination of SVM parameters, and by continuously iteratively updating the position and speed, it eventually converges to the global optimal solution, thereby improving the prediction accuracy of the model.

[0098] Furthermore, the particle swarm in the feature information of the training set is obtained, the position and velocity of the particles in the particle swarm are initialized, and the fitness of the particles is calculated. First, a group of particles are randomly initialized, each particle represents a candidate solution for feature data processing, that is, a set of feature weights or model parameters. The positions (solutions) and velocities (update directions and rates of solutions) of these particles are defined in a multidimensional space, corresponding to various possibilities for feature data processing. The position of each particle, that is, the feature data processing strategy it represents, is evaluated. In the life prediction scenario, this usually means applying the processing strategy to the feature data and measuring its effectiveness through a predefined evaluation criterion (such as minimizing prediction error, maximizing model accuracy, etc.). This step is the basis of optimization and determines the "fitness" of the particles.

[0099] Furthermore, the number of iterations is set, the historical optimal value of the particle is calculated, and then the global optimal value of the particle swarm is obtained by comparing the fitness of the particle in each iteration. The position and velocity of the particle are updated according to the historical optimal value and the global optimal value. Each particle updates its position and velocity according to its own historical optimal position (individual optimality) and the global optimal value of the particle swarm. This mechanism encourages particles to approach a better solution while maintaining the diversity of the search and avoiding premature convergence.

[0100] Furthermore, it is determined whether the position and speed of the particle meet the preset range. If not, the boundary conditions of the particle are processed. If the number of iterations is reached or the error is less than the preset error, the preset life prediction model is output and repeated. The evaluation and update steps are performed until the preset stop condition is met (such as reaching a certain number of iterations, minimal improvement in fitness, etc.). As the iterations proceed, the particle swarm gradually converges on a better solution and obtains a particle swarm optimization support vector machine model, namely the PSVM model, thereby generating accurate prediction information data. The standard deviation, root mean square and root amplitude are input into the optimized life prediction model to obtain the prediction results. Figure 4 The standard deviation prediction result graph shown, Figure 5 The root amplitude prediction results shown in the figure and Figure 6 The RMS prediction results are shown in Figure 2.

[0101] Step S40, obtaining the remaining service life of the target device according to the prediction information data.

[0102] It should be noted that the prediction information data is obtained through the optimized SVM algorithm model. These data reflect the trend of equipment status changes in the future and are the basis for evaluating the remaining life.

[0103] Once these prediction data are obtained, the next step is to input them into the linear risk model. In this embodiment, the Weibull linear risk model is used, which evaluates the performance degradation rate and risk rate of the equipment by considering multiple key characteristic parameters of the equipment, such as standard deviation, root mean square and root amplitude, as well as possible temperature effects. Model parameters such as λ, η, β1, β2 and β3 are determined through an optimization process to ensure that the model can accurately reflect the complex relationship between the equipment status and the remaining life. The prediction results are usually presented in the form of risk rate and survival probability, which reveal the potential failure probability and the possibility of continuous operation of the equipment at different time points. For example, a sudden increase in risk rate may indicate that the equipment is about to enter an unstable state. Set a threshold, such as a survival probability of 0.3. When the risk rate predicted by the model reaches this threshold, the expected failure time point of the equipment can be determined. In order to verify the accuracy of the prediction results, they are compared with actual cases and the prediction error is calculated. By comparing the prediction results with other algorithms such as basic SVM and BP neural network, the effectiveness and advantages of the proposed algorithm can be further confirmed. If the prediction accuracy is high, it means that the model can better predict the remaining service life of the equipment, which is of significant value in planning maintenance in advance, reducing unexpected downtime, and optimizing resource allocation.

[0104] This embodiment provides a life prediction method based on an optimized support vector machine, which obtains information data of a target device, extracts features from the information data, obtains feature data, processes the feature data according to a particle swarm optimization algorithm, obtains predicted information data, and obtains the remaining service life of the target device based on the predicted information data. The accuracy and efficiency of the prediction of the remaining service life of the device are improved through integrated data processing and optimization algorithms.

[0105] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can refer to the above introduction, and will not be repeated later. Figure 7 The life prediction method based on optimized support vector machine step S40 also includes steps S401 to S403:

[0106] Step S401, obtaining the risk probability through a linear risk model according to the prediction information data.

[0107] It should be noted that before predicting the remaining service life of the equipment, sample information data is obtained, and a linear risk model is established based on the information sample data. The particle swarm optimization support vector machine model (PSVM) is used to predict the future trend of the standard deviation, root mean square and root amplitude of the relevant features. PSVM enhances the generalization ability and prediction performance of the model by optimizing parameter selection. In this embodiment, 4340 sample points were collected, but only 3520 sample points were used to evaluate and predict the operating performance and remaining service life of the bearing. Therefore, after selecting three characteristic parameters, the 3520 sample data corresponding to each of them are input into the support vector machine regression model to predict the subsequent characteristic trend. On this basis, the Weber linear risk model is constructed by predicting the characteristic trend data obtained. The Weber linear risk model is a statistical analysis tool that can evaluate the performance status and remaining life of the equipment according to the changing trend of the input characteristic parameters. Model parameters such as λ, η, β, etc. are determined by optimization algorithms to best fit the actual situation. Specifically, the Weber linear risk model can be written as:

[0108]

[0109] where z 1 (t) is the standard deviation, z 2 (t) is the root mean square, z 3 (t) is the root amplitude, and the model parameters λ = 1.0366, η = 0.6837, β1 = 6.7870, β2 = 3.5752 and β3 = -13.2698.

[0110] The relevant features are then input into the established Weibull linear risk model to obtain the risk probability, which quantifies the possibility of equipment failure at a certain point in the future. The calculation of risk probability is based on the current health status of the equipment, the trend of characteristic parameters, and the statistical laws of historical data.

[0111] Step S402, comparing the risk probability with the preset damage probability to obtain the survival probability.

[0112] It should be understood that if the risk probability exceeds the preset damage probability, which is 0.3 in this embodiment, when the predicted risk probability reaches 0.3, it is considered that the device has reached the end of its expected service life. Figure 8 The risk probability assessment results shown in the figure and Fig. 9 The survival probability evaluation result diagram shown is based on Figure 8 It can be seen that the risk rate is close to 0 at the first 3020 sample points. At the 4210th sample point, the risk rate increases sharply due to the growth of the three eigenvalues. The larger the risk rate, the worse the stability. Fig. 9 The corresponding survival probability can be observed.

[0113] Step S403, obtaining the remaining useful life according to the survival probability calculation.

[0114] It should be noted that when calculating the remaining service life, based on the critical value of the set survival probability, for example, when the predicted survival probability drops to 0.3 for the first time, it is considered that the equipment is about to enter a high-risk period. Using the time series of equipment status data, combined with the changes in survival probability predicted by the model, it is possible to determine the expected healthy operation time of the equipment from the current moment, that is, the remaining service life. In the prediction sequence, the risk rate of the 4349th sample point reaches 0.3, which means that after the 3520th sample point, the bearing survives for 829s ((4349-3520)*1=829), while the actual remaining life is 820s. It can be seen that the accuracy of prediction using the PSVM model is 98.9%. Fig.10 From the comparison chart of absolute relative errors of predictions shown in the figure, it can be clearly seen that compared with the basic SVM model and the BP neural network model, the PSVM model has smaller prediction errors for samples 1#, 2#, 5#, 6#, 9#, and 10#.

[0115] This embodiment obtains the risk probability through a linear risk model based on the predicted information data, obtains the survival probability by comparing the risk probability with the preset damage probability, and obtains the remaining service life based on the calculation of the survival probability. Through detailed risk assessment and survival probability analysis, the accuracy of predicting the remaining service life of the equipment is improved.

[0116] This application also provides a life prediction device based on optimized support vector machine, please refer to Fig.11 , the device comprises:

[0117] The acquisition module 10 is used to acquire information data of the target device.

[0118] The feature extraction module 20 is used to extract features from information data to obtain feature data.

[0119] The acquisition module 10 is used to process the characteristic data according to the life prediction model to obtain prediction information data.

[0120] The result module 30 is used to obtain the remaining service life of the target device according to the prediction information data.

[0121] The life prediction device based on the optimized support vector machine provided by the present application adopts the life prediction method based on the optimized support vector machine in the above embodiment, which can solve the technical problem of how to improve the efficiency and accuracy of equipment life prediction. Compared with the prior art, the beneficial effects of the life prediction device based on the optimized support vector machine provided by the present application are the same as the beneficial effects of the life prediction method based on the optimized support vector machine provided by the above embodiment, and the other technical features in the life prediction device based on the optimized support vector machine are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0122] In one embodiment, the acquisition module 10 is also used to divide the feature data into a training set, a validation set and a test set; obtain the feature information of the training set and establish an initial life prediction model; train the initial life prediction model according to the particle swarm optimization algorithm based on the feature information of the training set to obtain a preset life prediction model.

[0123] In one embodiment, the acquisition module 10 is also used to obtain the particle swarm in the training set feature information; initialize the position and velocity of the particles in the particle swarm and calculate the fitness of the particles; set the number of iterations and calculate the historical optimal value of the particles; obtain the global optimal value of the particle swarm by comparing the fitness of the particles in each iteration; update the position and velocity of the particles according to the historical optimal value and the global optimal value; determine whether the position and velocity of the particles meet the preset range, if not, perform boundary condition processing on the particles; if the number of iterations is reached or the error is less than the preset error, output the preset life prediction model.

[0124] In one embodiment, the acquisition module 10 is also used to evaluate the preset life prediction model to obtain an evaluation result; if the evaluation result does not meet the prediction requirements, the preset life prediction model is retrained until the prediction requirements are met; if the evaluation result meets the prediction requirements, the preset life prediction model is used as the final life prediction model, and the remaining service life is analyzed.

[0125] In one embodiment, the acquisition module 10 is also used to acquire original information data of the target device; perform denoising on the original information data according to wavelet transform to obtain denoised information data; perform time-frequency analysis and normalization on the denoised information data to obtain information data of the target device.

[0126] In one embodiment, the feature extraction module 20 is also used to perform multi-level decomposition on the information data to obtain signals in multiple frequency bands; perform Fourier transform on the multiple frequency band signals to obtain frequency domain signals; perform sliding window analysis on the frequency domain signals to obtain distribution characteristics; and select feature data based on the distribution characteristics.

[0127] In one embodiment, the result module 30 is also used to obtain sample information data; establish a linear risk model based on the information sample data; obtain the risk probability through the linear risk model based on the predicted information data; obtain the survival probability by comparing the risk probability with the preset damage probability; and obtain the remaining service life based on the survival probability calculation.

[0128] The present application provides a life prediction device based on an optimized support vector machine, and the life prediction device based on an optimized support vector machine includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the life prediction method based on the optimized support vector machine in the above-mentioned embodiment one.

[0129] Reference below Fig.12 , which shows a schematic diagram of the structure of a life prediction device based on an optimized support vector machine suitable for implementing the embodiment of the present application. The life prediction device based on an optimized support vector machine in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig.12 The life prediction device based on the optimized support vector machine shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0130] like Fig.12As shown, the life prediction device based on the optimized support vector machine may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM: Read Only Memory) 1002 or the program loaded from the storage device 1003 to the random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the life prediction device based on the optimized support vector machine are also stored. The processing device 1001, ROM1002 and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the life prediction device based on the optimized support vector machine to communicate wirelessly or wired with other devices to exchange data. Although the figure shows a life prediction device based on the optimized support vector machine with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have instead.

[0131] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0132] The life prediction device based on the optimized support vector machine provided by the present application adopts the life prediction method based on the optimized support vector machine in the above embodiment, which can solve the technical problem of how to improve the efficiency and accuracy of equipment life prediction. Compared with the prior art, the beneficial effects of the life prediction device based on the optimized support vector machine provided by the present application are the same as the beneficial effects of the life prediction method based on the optimized support vector machine provided by the above embodiment, and the other technical features in the life prediction device based on the optimized support vector machine are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0133] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0134] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0135] The present application provides a computer-readable medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the life prediction method based on the optimized support vector machine in the above-mentioned embodiment.

[0136] The computer-readable medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0137] The computer-readable medium may be included in the life prediction device based on the optimized support vector machine; or may exist independently without being assembled into the life prediction device based on the optimized support vector machine.

[0138] The computer-readable medium carries one or more programs. When the one or more programs are executed by the life prediction device based on the optimized support vector machine, the life prediction device based on the optimized support vector machine can be written in one or more programming languages ​​or a combination thereof to write computer program codes for performing the operations of the present application. The programming languages ​​include object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed completely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0139] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0140] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0141] The readable medium provided by the present application is a computer-readable medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned life prediction method based on the optimized support vector machine, and can solve the technical problem of how to improve the efficiency and accuracy of equipment life prediction. Compared with the prior art, the beneficial effects of the computer-readable medium provided by the present application are the same as the beneficial effects of the life prediction method based on the optimized support vector machine provided by the above-mentioned embodiment, and will not be repeated here.

[0142] The present application also provides a computer program product, including a computer program, which implements the steps of the life prediction method based on the optimized support vector machine as described above when executed by a processor.

[0143] The computer program product provided by this application can solve the technical problem of how to improve the efficiency and accuracy of equipment life prediction. Compared with the prior art, the beneficial effects of the computer program product provided by this application are the same as the beneficial effects of the life prediction method based on the optimized support vector machine provided in the above embodiment, which will not be repeated here.

[0144] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A life prediction method based on optimized support vector machine, characterized in that: The method comprises: Get information data of the target device; Extracting features from the information data to obtain feature data; Processing the characteristic data according to the life prediction model to obtain prediction information data; The remaining service life of the target device is obtained according to the prediction information data.

2. The method according to claim 1, characterized in that Before the step of processing the characteristic data according to the life prediction model to obtain prediction information data, the method includes: Dividing the feature data into a training set, a validation set and a test set; Obtain training set feature information and establish an initial lifespan prediction model; The initial life prediction model is trained according to the feature information of the training set using a particle swarm optimization algorithm to obtain a preset life prediction model.

3. The method according to claim 2, characterized in that The step of training the initial life prediction model according to the feature information of the training set using a particle swarm optimization algorithm to obtain a preset life prediction model includes: Obtain the particle swarm in the training set feature information; Initializing the positions and velocities of particles in the particle swarm and calculating the fitness of the particles; Set the number of iterations and calculate the historical optimal value of the particle; By comparing the fitness of the particles in each iteration, a global optimal value of the particle swarm is obtained; According to the historical optimal value and the global optimal value, the position and speed of the particle are updated; Determine whether the position and speed of the particle meet a preset range, and if not, perform boundary condition processing on the particle; If the number of iterations is reached or the error is less than the preset error, the preset life prediction model is output.

4. The method according to claim 2, characterized in that After the step of training the initial life prediction model according to the training set feature information using a particle swarm optimization algorithm to obtain a preset life prediction model, the method includes: Evaluating the preset life prediction model to obtain an evaluation result; If the evaluation result does not meet the prediction requirement, retrain the preset life prediction model until it meets the prediction requirement; If the evaluation result meets the prediction requirements, the preset life prediction model is used as the final life prediction model, and the remaining service life is analyzed.

5. The method according to claim 1, characterized in that The step of obtaining information data of the target device includes: Get the original information data of the target device; Performing denoising on the original information data according to wavelet transform to obtain denoised information data; The information data after the noise removal is subjected to time-frequency analysis and normalization processing to obtain the information data of the target device.

6. The method according to claim 1, characterized in that The step of extracting features from the information data to obtain feature data comprises: Decomposing the information data at multiple levels to obtain signals in multiple frequency bands; Performing Fourier transform on the signals in the multiple frequency bands to obtain frequency domain signals; Performing sliding window analysis on the frequency domain signal to obtain distribution characteristics; The characteristic data is obtained by selecting according to the distribution characteristics.

7. The method according to any one of claims 1 to 6, characterized in that: Before the step of obtaining the remaining useful life according to the prediction information data, the method includes: Obtain sample information data; Establishing a linear risk model based on the information sample data; The step of obtaining the remaining service life according to the prediction information data further includes: Obtaining risk probability according to the predicted information data through the linear risk model; Comparing the risk probability with the preset damage probability, obtaining the survival probability; The remaining service life is calculated based on the survival probability.

8. A life prediction device based on optimized support vector machine, characterized in that: The device comprises: An acquisition module is used to acquire information data of a target device; A feature extraction module, used to extract features from the information data to obtain feature data; An acquisition module, used for processing the characteristic data according to the life prediction model to obtain prediction information data; The result module is used to obtain the remaining service life of the target device according to the prediction information data.

9. A life prediction device based on optimized support vector machine, characterized in that: The device includes: a memory, a processor, and a life prediction program based on an optimized support vector machine stored in the memory and executable on the processor, wherein the life prediction program based on the optimized support vector machine is configured to implement the steps of the life prediction method based on the optimized support vector machine as described in any one of claims 1 to 7.

10. A medium, characterized in that The medium stores a life prediction program based on an optimized support vector machine, and when the life prediction program based on the optimized support vector machine is executed by a processor, the steps of the life prediction method based on the optimized support vector machine as described in any one of claims 1 to 7 are implemented.

Images