Device failure time prediction method, device, apparatus, storage medium and computer program product

Abstract

The application relates to the technical field of fault time prediction, and discloses a device fault time prediction method, device, equipment, storage medium and computer program product, which comprises the following steps: generating an initial prediction parameter set according to monitoring data and component feature information of a target device; determining a prediction accuracy index of each parameter combination in the initial prediction parameter set; selecting an optimal parameter combination from the initial prediction parameter set based on the prediction accuracy index; and inputting the optimal parameter combination into a preset time sequence model to output a fault time prediction result of the target device. Since the optimal parameter combination of the current state of the device to be predicted in the time sequence model is determined through the monitoring data and the component feature information of the target device, accurate prediction of the device fault time is realized.

Description

Technical Field

[0001] This application relates to the field of failure time prediction technology, and in particular to a method, apparatus, device, storage medium and computer program product for predicting equipment failure time. Background Technology

[0002] Traditional power plant equipment failure prediction methods are usually based on time series analysis, which is suitable for univariate prediction, but requires the assumption of data stationarity, and trend separation is required when predicting multiple factors, resulting in high computational complexity and limited accuracy. Summary of the Invention

[0003] The main objective of this application is to provide a method, apparatus, device, storage medium, and computer program product for predicting equipment failure time, aiming to solve the technical problems of existing failure time prediction methods, which require trend separation when predicting multiple factors, have high computational complexity, and limited accuracy.

[0004] To achieve the above objectives, this application proposes a method for predicting equipment failure time, the method comprising:

[0005] An initial set of prediction parameters is generated based on the monitoring data and component characteristic information of the target equipment;

[0006] Determine the prediction accuracy index for each parameter combination in the initial prediction parameter set;

[0007] The optimal parameter combination is selected from the initial prediction parameter set based on the prediction accuracy index;

[0008] The optimal parameter combination is input into a preset time series model, and the failure time prediction result of the target device is output.

[0009] Optionally, the step of generating an initial prediction parameter set based on the monitoring data and component feature information of the target device includes:

[0010] Based on the monitoring data of the target equipment, multi-dimensional fault feature encoding is performed on the component feature information to generate an initial feature set;

[0011] Construct an initial prediction parameter set for the target device based on the initial feature set;

[0012] Optionally, the initial feature set includes a subset of sensor failure features, a subset of mechanical wear features, and a subset of abnormal vibration features;

[0013] The step of constructing the initial prediction parameter set of the target device based on the initial feature set includes:

[0014] A weighted recombination of a subset of sensor failure features is performed to generate a failure feature sequence containing the installation location information of the target device.

[0015] A time-domain degradation analysis was performed on a subset of mechanical wear characteristics to generate a wear trend sequence;

[0016] Multi-scale entropy feature extraction is performed on a subset of abnormal vibration features to generate a vibration pattern sequence that includes vibration energy distribution and abrupt change features;

[0017] An initial prediction parameter set is generated based on the failure feature sequence, the wear trend sequence, and the vibration mode sequence.

[0018] Optionally, the step of weighted reorganization of the subset of sensor failure features to generate a failure feature sequence containing the installation location information of the target device includes:

[0019] The sensor failure feature subset is divided according to the functional area of ​​the target device, and the monitoring data of different functional areas are cross-validated to obtain the verification results.

[0020] Based on the verification results, determine the data confidence weights of the sensor failure feature subsets in each functional area;

[0021] The sensor failure feature subset is dynamically weighted and recombined based on the data confidence weights to generate a failure feature sequence.

[0022] Optionally, the step of determining the prediction accuracy index of each parameter combination in the initial prediction parameter set includes:

[0023] The prediction accuracy quantification value of each parameter combination is determined by a preset evaluation function;

[0024] Time constraints are set based on the maintenance cycle of the target equipment;

[0025] The prediction accuracy quantization value is adjusted according to the time constraint to obtain the prediction accuracy index of each parameter combination.

[0026] Optionally, after the step of inputting the optimal parameter combination into a preset time series model and outputting the failure time prediction result of the target device, the method further includes:

[0027] Calculate the error volatility of the failure time prediction result over a continuous prediction period;

[0028] The deviation between the predicted time point and the actual failure point in the failure time prediction results is statistically analyzed.

[0029] The error volatility and interval deviation are normalized to generate adjustment correction coefficients, and the time series model is adjusted based on the adjustment correction coefficients.

[0030] Furthermore, to achieve the above objectives, this application also proposes an equipment failure time prediction device, which includes:

[0031] The parameter acquisition module is used to generate an initial set of predicted parameters based on the monitoring data and component feature information of the target device.

[0032] The indicator determination module is used to determine the prediction accuracy index of each parameter combination in the initial prediction parameter set;

[0033] The combination selection module is used to select the optimal parameter combination from the initial prediction parameter set based on the prediction accuracy index.

[0034] The time prediction module is used to input the optimal parameter combination into a preset time series model and output the failure time prediction result of the target device.

[0035] In addition, to achieve the above objectives, this application also proposes an equipment failure time prediction device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the equipment failure time prediction method as described above.

[0036] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the equipment failure time prediction method described above.

[0037] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the device failure time prediction method described above.

[0038] This application discloses a method for generating an initial prediction parameter set based on monitoring data and component feature information of a target device; determining the prediction accuracy index of each parameter combination in the initial prediction parameter set; selecting the optimal parameter combination from the initial prediction parameter set based on the prediction accuracy index; inputting the optimal parameter combination into a preset time series model; and outputting the failure time prediction result of the target device. Because this application determines the optimal parameter combination for the current state of the device to be predicted in the time series model using monitoring data and component feature information of the target device, it achieves accurate prediction of device failure time. Attached Figure Description

[0039] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0040] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart illustrating the first embodiment of the equipment failure time prediction method of this application;

[0042] Figure 2 This is a flowchart illustrating the second embodiment of the equipment failure time prediction method of this application;

[0043] Figure 3 This is a flowchart illustrating the third embodiment of the equipment failure time prediction method of this application;

[0044] Figure 4 This is a schematic diagram of the module structure of the equipment failure time prediction device according to an embodiment of this application;

[0045] Figure 5 This is a schematic diagram of the hardware operating environment involved in the device failure time prediction method in the embodiments of this application.

[0046] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0047] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0048] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0049] The main solution of this application embodiment is: to generate an initial prediction parameter set based on the monitoring data and component feature information of the target device; to determine the prediction accuracy index of each parameter combination in the initial prediction parameter set; to select the optimal parameter combination from the initial prediction parameter set based on the prediction accuracy index; to input the optimal parameter combination into a preset time series model and output the failure time prediction result of the target device.

[0050] Current equipment failure prediction technologies generally employ single-dimensional data analysis, such as vibration spectrum or temperature threshold monitoring, which suffers from insufficient ability to fuse multi-source heterogeneous data. Traditional methods, using LSTM networks to process time-series data, fail to effectively integrate the spatial topology of the equipment, resulting in a false correlation misjudgment rate of up to 25% for sensor failure characteristics. Furthermore, using fixed-weight fusion of multiple parameters leads to significant fluctuations in prediction errors during sudden changes in operating conditions. Existing equipment failure time prediction technologies mainly suffer from the following bottlenecks: 1. Mechanical wear analysis is largely limited to time-domain statistics, lacking joint modeling of time-frequency domain degradation trends, resulting in excessive missed detection rates for early bearing pitting corrosion; 2. Abnormal vibration detection relies on fixed-band energy analysis, leading to long response delays for impact failures; 3. The static parameter optimization process cannot adapt to feature drift caused by equipment aging, requiring weekly manual model calibration, increasing maintenance costs. In addition, traditional 3D visualization methods only display the current state, failing to dynamically extrapolate fault propagation paths, resulting in a lack of predictability in maintenance decisions.

[0051] This application provides an intelligent prediction system based on multi-dimensional coupling of space, time, and operating conditions. By structurally processing monitoring data and component characteristics, it establishes a complete technical closed loop from parameter generation to prediction output, meeting the real-time requirements of industrial sites.

[0052] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, intelligent decision-making, and program execution functions, such as a power plant intelligent management platform, or an electronic device capable of performing the above functions. The following description uses a power plant equipment detection system as an example to illustrate this embodiment and the subsequent embodiments.

[0053] Based on this, embodiments of this application provide a method for predicting equipment failure time, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the equipment failure time prediction method of this application.

[0054] In this embodiment, the equipment failure time prediction method includes:

[0055] Step S10: Generate an initial prediction parameter set based on the monitoring data and component feature information of the target equipment.

[0056] It should be noted that the target equipment refers to the specific equipment object for which failure time prediction needs to be performed. In industrial scenarios, the target equipment can be a specific rotating machine (such as a steam turbine or compressor), production line equipment (such as a CNC machine tool), or critical infrastructure (such as power transmission and transformation equipment), etc.

[0057] The monitoring data of the target equipment refers to the real-time operating status data of the equipment collected by various sensors, such as temperature, pressure, and vibration frequency; the component characteristic information refers to the inherent attributes and characteristics of each component of the target equipment, such as the component's material, design parameters, and service life. The initial prediction parameter set consists of a series of parameters extracted, transformed, and combined from the monitoring data and component characteristic information. These parameters will serve as the basic input for subsequent failure time prediction.

[0058] Understandably, the initial prediction parameter set contains feature information related to equipment failure extracted from multiple dimensions, which can more comprehensively and accurately reflect the operating status of the equipment.

[0059] It should be understood that during data acquisition, attention must be paid to the sensor's installation location and calibration to ensure data accuracy. For example, accelerometers should be installed on critical vibration points of the equipment and calibrated regularly. Furthermore, the acquired data should undergo preprocessing such as noise reduction and filtering.

[0060] In one example, the monitoring data of the power plant turbine includes bearing vibration signals, lubricating oil temperature, rotor speed, etc. Through multi-dimensional fault feature coding, features such as kurtosis (reflecting impact faults) and spectral entropy (measuring signal complexity) can be extracted from the bearing vibration signals, and features such as temperature change rate can be extracted from the lubricating oil temperature data to form an initial feature set.

[0061] Step S20: Determine the prediction accuracy index of each parameter combination in the initial prediction parameter set.

[0062] It should be noted that prediction accuracy indicators are quantitative metrics used to measure the difference between the predicted result and the actual value, such as root mean square error and mean absolute percentage error.

[0063] Understandably, when determining the prediction accuracy index, the impact of different parameter combinations on the failure time prediction results needs to be considered. The prediction accuracy index can be a combination of multiple evaluation indicators, such as root mean square error, mean absolute error, and prediction accuracy. In actual implementation, each parameter combination can be quantitatively evaluated using a preset evaluation function to obtain a quantified value of the prediction accuracy for each parameter combination.

[0064] Furthermore, in order to adjust the quantized values ​​according to time constraints to obtain a prediction accuracy index, and to ensure that the index can more accurately reflect the prediction performance of the parameter combination within the actual equipment maintenance cycle, thereby improving the accuracy of selecting the optimal parameter combination, step S20 may include:

[0065] The prediction accuracy quantification value of each parameter combination is determined by a preset evaluation function; a time constraint condition is set by the maintenance cycle of the target equipment; the prediction accuracy quantification value is adjusted according to the time constraint condition to obtain the prediction accuracy index of each parameter combination.

[0066] It should be noted that the preset evaluation function is a pre-defined mathematical function that quantifies the accuracy of different parameter combinations in predicting equipment failure time. The design of the preset evaluation function considers the difference between the predicted result and the actual failure time, such as common mean square error and mean absolute error. The quantified prediction accuracy value is a specific numerical value obtained after calculating each parameter combination using the preset evaluation function, demonstrating the accuracy of that parameter combination in predicting equipment failure time. The maintenance cycle is the time interval between the completion of one maintenance and the start of the next maintenance for the target equipment. The time constraint is a limitation on the prediction time range set according to the equipment maintenance plan or the degree of failure impact, to avoid conflicts between the predicted failure time and the maintenance schedule.

[0067] Understandably, time constraints can be dynamically adjusted based on the importance of the equipment and the consequences of failure. For example, the prediction time constraints for critical equipment (such as boilers) can be more stringent.

[0068] Step S30: Select the optimal parameter combination from the initial prediction parameter set based on the prediction accuracy index.

[0069] Understandably, when selecting the optimal parameter combination, both prediction accuracy and practical application requirements need to be considered. For example, if the consequences of a device failure are very serious, the parameter combination with the highest prediction accuracy may be chosen first, even if its computational complexity is high; if the device needs to perform real-time fault prediction, a parameter combination with higher computational efficiency can be selected while also ensuring a certain level of prediction accuracy.

[0070] In some embodiments of this application, intelligent optimization algorithms such as genetic algorithms and particle swarm optimization algorithms can be used to search for the optimal parameter combination from the initial set of predicted parameters. These algorithms can efficiently find the optimal solution in a complex parameter space, improving the accuracy and efficiency of parameter selection.

[0071] In one example, for fault prediction of power plant generators, a genetic algorithm is used to select the optimal combination from an initial set of prediction parameters including vibration, temperature, and pressure characteristics. First, the parameter combinations are represented as chromosomes, such as binary codes. Then, chromosomes with high fitness are selected based on prediction accuracy. Hints suggest generating new chromosomes through crossover operations and random mutation to maintain the diversity of parameter combinations. After several generations of evolution, a set of parameter combinations is obtained whose prediction accuracy reaches its optimal level within the maintenance cycle.

[0072] Step S40: Input the optimal parameter combination into a preset time series model and output the failure time prediction result of the target device.

[0073] It should be understood that the preset time series model can be of various types, such as ARIMA, LSTM, and Prophet models. These models can analyze and predict time series data, inferring future trends based on patterns in historical data.

[0074] It is understood that, in some embodiments of this application, the model can be periodically updated and optimized to improve its prediction accuracy and adaptability. For example, when new monitoring data is acquired, the model can be retrained, or the model parameters can be adjusted based on the new data.

[0075] In this embodiment, an initial prediction parameter set is generated based on the monitoring data and component feature information of the target device; the prediction accuracy index of each parameter combination in the initial prediction parameter set is determined; the optimal parameter combination is selected from the initial prediction parameter set based on the prediction accuracy index; the optimal parameter combination is input into a preset time series model, and the failure time prediction result of the target device is output. Because this application determines the optimal parameter combination of the current state of the device to be predicted in the time series model using the monitoring data and component feature information of the target device, accurate prediction of device failure time is achieved.

[0076] Reference Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the equipment failure time prediction method of this application. Based on the first embodiment described above, a second embodiment of the equipment failure time prediction method of this application is proposed.

[0077] In the second embodiment, step S10 includes:

[0078] Step S101: Based on the monitoring data of the target equipment, multi-dimensional fault feature encoding is performed on the component feature information to generate an initial feature set.

[0079] It should be noted that multi-dimensional fault feature coding refers to analyzing and encoding component feature information from multiple different perspectives to extract more comprehensive and representative fault features. For example, vibration signals can be analyzed from multiple dimensions such as the time domain, frequency domain, and time-frequency domain to extract various features such as mean, variance, kurtosis, and spectral entropy.

[0080] In some embodiments of this application, deep learning techniques can be used for multi-dimensional fault feature encoding. For example, a convolutional neural network (CNN) can be used to extract features from vibration signals. CNN can automatically learn deep features in the data, improving the efficiency and accuracy of feature extraction.

[0081] In one example, for boiler equipment in a power plant, a CNN is used to extract features from its vibration signals. Automatic learning can reflect the deep features of faults such as furnace coking and tube rupture. First, the vibration signal and other data are converted into a format suitable for CNN input (such as a two-dimensional matrix). Then, local features are extracted through convolutional kernels. The extracted features are mapped to the fault feature space in a fully connected layer. Finally, features from different dimensions (such as vibration, temperature, and pressure) are concatenated or weighted and fused to form a comprehensive fault feature set, resulting in an initial feature set.

[0082] Step S102: Construct the initial prediction parameter set for the target device based on the initial feature set.

[0083] Understandably, when constructing the initial prediction parameter set, it is necessary to reasonably combine and filter the various features in the initial feature set to form a parameter set that can effectively reflect the equipment fault state. For example, sensor failure features, mechanical wear features, and abnormal vibration features can be combined to form a comprehensive parameter set.

[0084] It should be understood that when there are many parameters in the initial feature set, dimensionality reduction techniques such as principal component analysis (PCA) and linear discriminant analysis (LDA) can be used to process the initial feature set, reduce the dimensionality of the features, and improve the computational efficiency and prediction accuracy of the subsequent model.

[0085] Furthermore, in order to extract spatial topology, temporal degradation, and frequency domain abrupt change features through categorized feature processing, and to synthesize these features to generate an initial prediction parameter set, a more comprehensive and accurate description of the equipment's fault characteristics can be achieved. Step S102 may include:

[0086] A weighted recombination of the sensor failure feature subset generates a failure feature sequence containing the installation location information of the target equipment; a time-domain degradation analysis is performed on the mechanical wear feature subset to generate a wear trend sequence; multi-scale entropy feature extraction is performed on the abnormal vibration feature subset to generate a vibration mode sequence containing vibration energy distribution and abrupt change features; an initial prediction parameter set is generated based on the failure feature sequence, the wear trend sequence, and the vibration mode sequence.

[0087] It should be noted that the sensor failure feature subset is anomaly features extracted from the raw signals collected by sensors such as vibration and temperature sensors, including peak values, kurtosis, and energy distribution, reflecting the sensor's own health status (e.g., decreased sensitivity, noise interference). Installation location information represents the topological relationship of the equipment's physical structure, such as the location of bearings at key support points in the rotor system, used to distinguish reliability differences in data from different areas. The mechanical wear feature subset is a collection of time-domain signal features reflecting the physical wear of equipment components, such as kurtosis change rate and the proportion of defect frequency energy. Time-domain degradation analysis uses statistical models to fit the trend of wear rate changes over time. The wear trend sequence is a continuous curve with time on the horizontal axis and wear rate on the vertical axis, used to predict remaining service life. The abnormal vibration feature subset includes non-stationary signal features such as impact pulses (SPI) and high-frequency resonant band energy. Multi-scale entropy calculates signal complexity at different time scales, measuring abrupt changes in the degradation process. The vibration mode sequence characterizes the abrupt change patterns of vibration signals in the time-frequency domain.

[0088] In one example, when predicting bearing failures in a power plant turbine unit, the equipment is first divided into multiple functional areas (such as the bearing housing area and the gearbox area), with independent sensors deployed in each area. Correlation analysis is then performed on data from multiple areas under the same operating condition, and the confidence weights for each area's data are calculated.

[0089]

[0090] Where, σ i and σ j The data variance of the sensors is divided into regions i and j. W is the position sensitivity coefficient, which can be preset according to the equipment manual, such as setting the weight of the bearing housing region to 0.6. The formula for weighting the subset of sensor failure characteristics is as follows:

[0091]

[0092] Among them, S i F represents the standardized regional feature values. miss This is the recombined failure characteristic sequence. By combining sensor data from various locations, false alarms caused by single sensor failures (such as sensor noise interference in a certain area) can be eliminated, improving the reliability of the characteristics. For example, a turbine bearing housing sensor may falsely alarm due to vibration attenuation, but the sensor data in the gearbox area may be normal. After weighted recombining, outliers are removed, preventing false alarms.

[0093] When performing time-domain degradation analysis on a subset of mechanical wear features, the kurtosis rate of change (KR) of the vibration signal is first calculated:

[0094]

[0095] Among them, K tK0 represents the current kurtosis, and K0 represents the initial kurtosis value.

[0096] Next, the energy percentage of the bearing inner ring defect frequency (BPFI) is extracted:

[0097]

[0098] E BPFI Based on the characteristic frequency of the bearing component, P(f) represents the energy distribution density of the vibration signal at frequency f. The integral of vibrational energy centered at BPFI and within the frequency range of BPFI±5 represents the total energy within this specific frequency interval.

[0099] The LSTM model trained using historical data will combine KR and E. BPFI As input, the output wear rate prediction value R(t) is:

[0100] R(t) = a·e bt +c

[0101] Based on a subset of mechanical wear characteristics, early signs of minor wear (such as BPFI energy percentage increasing from 5% to 15%) can be detected, thus preventing sudden failures.

[0102] When analyzing a subset of abnormal vibration features, signal preprocessing is required. First, the vibration signal is mean-reduced and normalized. Then, wavelet transform is used to decompose the signal into multiple scale components, and the sample entropy (SE) is calculated for each scale component.

[0103]

[0104] Where m is the embedding dimension, used to construct the vector dimension of the phase space, r is the tolerance threshold, defining the criterion for judging data similarity, and N is the data length of the time series. C m (r) and C m+1 (r) represents the proportion of the total number of sample pairs where the distance between two sequences is less than or equal to r when the embedding dimensions are m and m+1, respectively.

[0105] Finally, the entropy values ​​at each scale are weighted and combined to generate a vibration mode sequence:

[0106]

[0107] Where, α k The scale weights are determined using the mutual information method. n is the number of scale components.

[0108] After obtaining the failure feature sequence, wear trend sequence, and vibration pattern sequence, a comprehensive input vector combining multi-source features is used as the initial prediction parameter set for the time series model output. Alternatively, grid search and genetic algorithms can be used to select the optimal feature combination to reduce the risk of model overfitting.

[0109] First, align the three subsequences by time and concatenate them into a high-dimensional feature vector X:

[0110] X = [F] miss (t),R(t),V(t)]

[0111] Next, define the candidate parameter space (e.g., the number of LSTM layers {2, 3}, and the learning rate {0.001, 0.01}). Evaluate the model performance L using the cross-entropy loss function:

[0112]

[0113] Among them, y i This refers to the actual downtime. These are predicted values.

[0114] Furthermore, in order to determine the data confidence weights based on the verification results, and to ensure that the data from different functional areas receive reasonable weight allocation during the reorganization process, the differences in sensor data from different functional areas can be fully considered, thereby improving the quality of the failure feature sequence. The step of weighted reorganization of the sensor failure feature subset to generate a failure feature sequence containing the installation location information of the target device may include:

[0115] The sensor failure feature subsets are divided according to the functional areas of the target device, and the monitoring data of different functional areas are cross-validated to obtain the validation results. Based on the validation results, the data confidence weights of the sensor failure feature subsets in each functional area are determined. The sensor failure feature subsets are dynamically weighted and recombined according to the data confidence weights to generate a failure feature sequence.

[0116] Understandably, sensor data is categorized into different subsets based on the functional structure of the target equipment (such as mechanical modules or system units). When performing feature cross-validation on monitoring data from different functional areas, analysis can be conducted by setting validation rules. For example, during normal equipment operation, temperature changes in the target area should match the parameter change trends in related areas; or the spatiotemporal correlation of sensor data from different areas can be compared. For instance, when the temperature of a turbine bearing rises, it's possible to check whether the corresponding vibration sensor data shows synchronous anomalies.

[0117] It should be understood that when quantifying the reliability of data in each region based on cross-validation results, regions with high reliability are assigned high weights, and vice versa. By weighting and combining the subsets of sensor failure features in each region based on confidence weights, a sequence that comprehensively reflects the overall failure state of the equipment can be generated.

[0118] In this embodiment, multi-dimensional fault feature encoding is performed on component feature information based on monitoring data of the target device to generate an initial feature set; an initial prediction parameter set for the target device is then constructed based on the initial feature set. By generating an initial feature set through multi-dimensional fault feature encoding of component feature information based on monitoring data, the fault features of the device can be more comprehensively and deeply mined, improving the richness and effectiveness of the features.

[0119] Reference Figure 3 , Figure 3 This is a flowchart illustrating the third embodiment of the equipment failure time prediction method of this application. Based on the second embodiment described above, a third embodiment of the equipment failure time prediction method of this application is proposed.

[0120] In the third embodiment, after step S40, the method further includes:

[0121] Step S501: Calculate the error fluctuation rate of the fault time prediction result within the continuous prediction period.

[0122] It should be noted that error volatility is the degree of fluctuation in prediction error over a continuous prediction period, and it can be used to assess the stability of model predictions. Error volatility reflects the fluctuation of prediction results over a period of time, and calculating error volatility can help understand the stability of model predictions.

[0123] Understandably, one can calculate the degree of deviation between the error at each prediction point in a continuous prediction period and the average error, and then express the error volatility in the form of standard deviation.

[0124] In one example, error volatility is expressed in the form of standard deviation, which is a mathematical measure of how much the error deviates from the average error. The formula is as follows:

[0125]

[0126] Among them, e t Let be the prediction error at time point t. Let N be the average error, and N be the total number of time points.

[0127] For each predicted time point t, the predicted failure time needs to be calculated. Compared with the actual failure time y i The error is obtained as e t and standard deviation σ.

[0128] Calculate the average error e within each window t The average of the standard deviations of all windows is then used as the total error volatility (EV).

[0129]

[0130] Where K is the total number of windows.

[0131] Step S502: Calculate the deviation between the predicted time point and the actual fault point in the fault time prediction result.

[0132] It should be noted that the interval deviation refers to the time difference between the predicted time point and the actual failure point. Statistical interval deviation can be used to evaluate the accuracy and timeliness of the model's prediction. For example, if the predicted time point is earlier than the actual failure point, it means that the model can provide an early warning; if the predicted time point is later than the actual failure point, it means that the model's prediction has a delay.

[0133] Understandably, when calculating the deviation of the statistical interval, the accuracy of the timestamp of the actual fault point can be ensured through equipment logs or manual verification. For cases where the fault is predicted in advance, further analysis is needed to determine whether the lead time is sufficient for maintenance operations.

[0134] Step S503: Normalize the error volatility and interval deviation to generate adjustment correction coefficients, and adjust the time series model based on the adjustment correction coefficients.

[0135] It should be noted that normalization is performed to convert error volatility and interval deviations of different dimensions into comparable values ​​for comprehensive evaluation. After generating adjustment and correction coefficients, the parameters of the time series model can be adjusted based on these coefficients, such as adjusting the model's weights and learning rate, so that the model can better adapt to the actual operating conditions of the equipment.

[0136] Understandably, after normalizing the error volatility and interval deviation, weights can be set based on historical experience to calculate a comprehensive correction coefficient. If the comprehensive correction coefficient is greater than the preset value, it indicates that the model is unstable and has a large bias. In this case, grid search should be used to re-optimize the LSTM hyperparameters, such as the learning rate and the number of hidden layer nodes. If the comprehensive correction coefficient is less than the preset value, the model is overfitting. Noisy data or synthetic data can be added to expand the training set.

[0137] In this embodiment, the error volatility of the fault time prediction result over a continuous prediction period is calculated; the deviation between the predicted time point and the actual fault point in the fault time prediction result is statistically analyzed; the error volatility and the deviation are normalized to generate adjustment correction coefficients, and the time series model is adjusted based on these adjustment correction coefficients. Normalizing the error volatility and the deviation to generate adjustment correction coefficients, and then adjusting the time series model accordingly, allows the model to continuously learn and adapt to the actual operating conditions of the equipment, improving the model's predictive performance and adaptability, thereby maintaining high prediction accuracy over long-term prediction processes.

[0138] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the equipment failure time prediction method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0139] This application also provides a device for predicting equipment failure time; please refer to... Figure 4 The equipment failure time prediction device includes:

[0140] The parameter acquisition module 10 is used to generate an initial prediction parameter set based on the monitoring data and component feature information of the target device;

[0141] The indicator determination module 20 is used to determine the prediction accuracy index of each parameter combination in the initial prediction parameter set;

[0142] The combination selection module 30 is used to select the optimal parameter combination from the initial prediction parameter set based on the prediction accuracy index.

[0143] The time prediction module 40 is used to input the optimal parameter combination into a preset time series model and output the failure time prediction result of the target device.

[0144] The equipment failure time prediction device provided in this application, employing the equipment failure time prediction method described in the above embodiments, can solve the technical problems of existing failure time prediction methods, which require trend separation in multi-factor prediction, resulting in high computational complexity and limited accuracy. Compared with the prior art, the beneficial effects of the equipment failure time prediction device provided in this application are the same as those of the equipment failure time prediction method provided in the above embodiments, and other technical features in the equipment failure time prediction device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0145] This application provides a device for predicting device failure time, which 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, which are executed by the at least one processor to enable the at least one processor to perform the device failure time prediction method in the first embodiment described above.

[0146] The following is for reference. Figure 5 The diagram illustrates a structural schematic suitable for implementing a device failure time prediction device according to embodiments of this application. The device failure time prediction device in embodiments of this 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 Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 5 The device failure time prediction device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0147] like Figure 5 As shown, the device failure time prediction device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the device failure time prediction device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the device failure time prediction device to communicate wirelessly or wiredly with other devices to exchange data. While the figure shows device failure time prediction devices with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0148] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0149] The equipment failure time prediction device provided in this application, employing the equipment failure time prediction method described in the above embodiments, can solve the technical problems of existing failure time prediction methods, which require trend separation in multi-factor prediction, resulting in high computational complexity and limited accuracy. Compared with the prior art, the beneficial effects of the equipment failure time prediction device provided in this application are the same as those of the equipment failure time prediction method provided in the above embodiments, and other technical features of this equipment failure time prediction device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0150] It should be understood that the various parts disclosed in this application can be implemented using 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 suitable manner in one or more embodiments or examples.

[0151] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0152] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the device failure time prediction method in the above embodiments.

[0153] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0154] The aforementioned computer-readable storage medium may be included in the device failure time prediction device; or it may exist independently and not assembled into the device failure time prediction device.

[0155] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the device failure time prediction device, cause the device failure time prediction device to perform the device failure time prediction method described above.

[0156] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0157] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0158] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0159] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described equipment failure time prediction method. This solves the technical problems of existing failure time prediction methods, which require trend separation in multi-factor prediction, resulting in high computational complexity and limited accuracy. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the equipment failure time prediction method provided in the above embodiments, and will not be repeated here.

[0160] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the device failure time prediction method described above.

[0161] The computer program product provided in this application can solve the technical problems of existing failure time prediction methods, which require trend separation in multi-factor prediction, resulting in high computational complexity and limited accuracy. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the equipment failure time prediction method provided in the above embodiments, and will not be repeated here.

[0162] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for predicting equipment failure time, characterized in that, The equipment failure time prediction method includes: An initial prediction parameter set is generated based on the monitoring data and component feature information of the target equipment, wherein the target equipment is a bearing of a rotating machine in a power plant, the monitoring data includes vibration signals, the component feature information includes the frequency of defects in the inner ring of the bearing, and the initial prediction parameter set includes time-domain features and frequency-domain features extracted from the vibration signals. Determine the prediction accuracy index for each parameter combination in the initial prediction parameter set; The optimal parameter combination is selected from the initial prediction parameter set based on the prediction accuracy index; The optimal parameter combination is input into a preset time series model, and the failure time prediction result of the target device is output. The step of generating an initial prediction parameter set based on the monitoring data and component feature information of the target device includes: Based on the monitoring data of the target equipment, multi-dimensional fault feature encoding is performed on the component feature information to generate an initial feature set; Construct an initial prediction parameter set for the target device based on the initial feature set; The initial feature set includes a subset of sensor failure features, a subset of mechanical wear features, and a subset of abnormal vibration features; The step of constructing the initial prediction parameter set of the target device based on the initial feature set includes: The confidence weights are determined based on the data variance of the sensor data in each functional area of ​​the target device. The standardized regional feature values ​​in the sensor failure feature subset are then weighted and recombined based on the confidence weights to generate a failure feature sequence containing the installation location information of the target device. The sensor failure feature subset includes abnormal features extracted from the original signals collected by vibration sensors or temperature sensors. The kurtosis rate of the vibration signal of the target equipment and the corresponding bearing inner ring defect frequency energy ratio are calculated based on the mechanical wear feature subset. The kurtosis rate and the bearing inner ring defect frequency energy ratio are then input into the trained long short-term memory network model to output the wear trend sequence. Multi-scale entropy feature extraction is performed on a subset of abnormal vibration features to generate a vibration pattern sequence that includes vibration energy distribution and abrupt change features; An initial prediction parameter set is generated based on the failure feature sequence, the wear trend sequence, and the vibration mode sequence.

2. The equipment failure time prediction method as described in claim 1, characterized in that, The step of weighting and recombining the standardized regional feature values ​​in the subset of sensor failure features according to the confidence weights to generate a failure feature sequence containing the installation location information of the target equipment includes: The sensor failure feature subset is divided according to the functional area of ​​the target device, and the monitoring data of different functional areas are cross-validated to obtain the verification results. Based on the verification results, determine the data confidence weights of the sensor failure feature subsets in each functional area; The sensor failure feature subset is dynamically weighted and recombined based on the data confidence weights to generate a failure feature sequence.

3. The equipment failure time prediction method as described in claim 1, characterized in that, The step of determining the prediction accuracy index of each parameter combination in the initial prediction parameter set includes: The prediction accuracy quantification value of each parameter combination is determined by a preset evaluation function; Time constraints are set based on the maintenance cycle of the target equipment; The prediction accuracy quantization value is adjusted according to the time constraint to obtain the prediction accuracy index of each parameter combination.

4. The equipment failure time prediction method as described in any one of claims 1 to 3, characterized in that, After the step of inputting the optimal parameter combination into a preset time series model and outputting the failure time prediction result of the target device, the method further includes: Calculate the error volatility of the failure time prediction result over a continuous prediction period; The deviation between the predicted time point and the actual failure point in the failure time prediction results is statistically analyzed. The error volatility and interval deviation are normalized to generate adjustment correction coefficients, and the time series model is adjusted based on the adjustment correction coefficients.

5. A device for predicting equipment failure time, characterized in that, The device includes: The parameter acquisition module is used to generate an initial prediction parameter set based on the monitoring data and component feature information of the target equipment. The target equipment is a bearing of a rotating machine in a power plant. The monitoring data includes vibration signals, the component feature information includes the bearing inner ring defect frequency, and the initial prediction parameter set includes time-domain features and frequency-domain features extracted from the vibration signals. The indicator determination module is used to determine the prediction accuracy index of each parameter combination in the initial prediction parameter set; The combination selection module is used to select the optimal parameter combination from the initial prediction parameter set based on the prediction accuracy index. The time prediction module is used to input the optimal parameter combination into a preset time series model and output the failure time prediction result of the target device. The parameter acquisition module is also used to encode the component feature information in multiple dimensions based on the monitoring data of the target device to generate an initial feature set; and to construct an initial prediction parameter set for the target device based on the initial feature set; the initial feature set includes a subset of sensor failure features, a subset of mechanical wear features, and a subset of abnormal vibration features; The parameter acquisition module is further configured to determine the corresponding confidence weights based on the data variance of sensor data in each functional area of ​​the target device, and to perform weighted recombination of the standardized regional feature values ​​in the sensor failure feature subset based on the confidence weights to generate a failure feature sequence containing the installation location information of the target device. The sensor failure feature subset includes abnormal features extracted from the original signals collected by vibration sensors or temperature sensors. Based on the mechanical wear feature subset, the module calculates the kurtosis rate of the vibration signal of the target device and the corresponding bearing inner ring defect frequency energy ratio, and inputs the kurtosis rate and the bearing inner ring defect frequency energy ratio into the trained long short-term memory network model to output a wear trend sequence. The module performs multi-scale entropy feature extraction on the abnormal vibration feature subset to generate a vibration mode sequence containing vibration energy distribution and abrupt change features. Based on the failure feature sequence, the wear trend sequence, and the vibration mode sequence, the module generates an initial prediction parameter set.

6. A device for predicting equipment failure time, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the device failure time prediction method as described in any one of claims 1 to 4.

7. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the equipment failure time prediction method as described in any one of claims 1 to 4.

8. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the device failure time prediction method as described in any one of claims 1 to 4.

Images