Equipment fault positioning and early warning method and system based on aging degree analysis

Abstract

The invention discloses an equipment fault positioning early warning method and system based on aging degree analysis, and relates to the technical field of fault monitoring, and the method comprises the steps: obtaining the historical operation data of equipment, carrying out the preprocessing of the historical operation data of the equipment, obtaining the standardized data, and carrying out the feature extraction of the standardized data, and obtaining the operation feature data of the equipment; constructing an equipment aging model based on the digital twin model and the implicit semi-Markov model in combination with the equipment operation feature data; acquiring equipment operation characteristic data in real time, monitoring and updating an equipment aging state by using the equipment aging model, and performing fault prediction based on the equipment aging state; and based on a fault prediction result, combining the obtained equipment operation characteristic data, extracting fault information, and carrying out fault positioning and early warning according to the fault information. According to the invention, through the equipment aging model, the equipment aging state is monitored and updated, potential faults can be identified in advance, the problem root can be accurately positioned, and an effective maintenance plan is formulated.

Description

Technical Field

[0001] The present invention relates to the technical field of fault monitoring, and more specifically, to a method and system for equipment fault location and early warning based on aging degree analysis. Background Art

[0002] As an important part of the urban public transportation system, subway trains undertake the important task of transporting a large number of passengers efficiently and safely; with the influence of the service life of subway trains and environmental factors, train equipment such as traction motors and braking equipment gradually ages, resulting in performance degradation and increased fault risks; if only relying on traditional threshold alarms without aging degree analysis, it is difficult to accurately predict the decline of equipment, and it is easy to miss the best maintenance opportunity, leading to an increased risk of sudden failures; for traction motors and braking equipment with long-term service, without the support of aging degree data, relying solely on the method of traditional threshold alarms cannot effectively identify early fault signs, easily miss the best maintenance opportunity, and increase the risk of sudden failures.

[0003] In addition, warning mechanisms that do not consider equipment aging factors often rely on simple monitoring of the current state, which may lead to frequent false alarms or missed alarms, affecting the safe and stable operation of equipment; this method is difficult to capture the physical changes and performance degradation of equipment over time, lacks an understanding of the wear accumulation within the equipment life cycle, and is prone to ignoring key factors related to aging, resulting in inaccurate fault location and inability to fundamentally solve the problem; at the same time, without the support of aging degree analysis, the formulation of maintenance plans may be more blind, increasing unnecessary maintenance costs, and may lead to service interruptions due to failure to detect potential faults in a timely manner, affecting the normal travel of passengers.

[0004] In response to the problems in the related art, no effective solutions have been proposed yet. Summary of the Invention

[0005] In response to the problems in the related art, the present invention proposes a method and system for equipment fault location and early warning based on aging degree analysis to overcome the above technical problems existing in the existing related technologies.

[0006] To this end, the specific technical solutions adopted by the present invention are as follows:

[0007] According to one aspect of the present invention, there is provided a method for equipment fault location and early warning based on aging degree analysis, the method comprising:

[0008] S1. Obtain historical operation data of the equipment, preprocess the historical operation data of the equipment to obtain standardized data, and extract features from the standardized data to obtain equipment operation feature data;

[0009] S2. Based on the digital twin model and the implicit semi - Markov model, combined with the device operation characteristic data, construct a device aging model;

[0010] S3. Obtain the device operation characteristic data in real - time, use the device aging model to monitor and update the device aging state, and conduct fault prediction based on the device aging state;

[0011] S4. Based on the fault prediction results, combined with the obtained device operation characteristic data, extract fault information, and conduct fault location warning according to the fault information.

[0012] Furthermore, obtain the device historical operation data, pre - process the device historical operation data to obtain standardized data, and conduct feature extraction on the standardized data to obtain device operation characteristic data, including:

[0013] S11. Obtain the device historical operation data, conduct noise reduction processing on the device historical operation data, and conduct standardization processing on the noise - reduced data to obtain standardized data;

[0014] S12. According to the standardized data, extract the time - domain, frequency - domain, and time - frequency - domain feature data of the device, and conduct feature integration on the time - domain, frequency - domain, and time - frequency - domain feature data of the device to obtain a device feature dataset;

[0015] S13. Use the Pearson correlation coefficient to conduct aging degree correlation analysis on the device feature data, and remove redundant features based on the correlation analysis results to obtain device operation characteristic data;

[0016] Among them, the calculation formula of the Pearson correlation coefficient is:

[0017]

[0018] In the formula, Z k represents the Pearson correlation coefficient of the k - th device feature in the device feature dataset; t represents device t; a represents the total number of devices; T k,t represents the value of the k - th device feature in device t; represents the mean value of the k - th device feature of all devices; L t represents the aging degree of device t; represents the mean value of the aging degrees of all devices.

[0019] Furthermore, according to the standardized data, extract the time - domain, frequency - domain, and time - frequency - domain feature data of the device, and conduct feature integration on the time - domain, frequency - domain, and time - frequency - domain feature data of the device to obtain device feature data, including:

[0020] S121. According to the standardized data, calculate the statistical features of the device operation state to obtain the device time - domain feature data;

[0021] S122. Use Fourier transform, combined with the standardized data, to calculate the frequency characteristics of the device operation state and obtain the frequency-domain characteristic data;

[0022] S123. Use wavelet transform, combined with the standardized data, to calculate the wavelet energy and wavelet entropy of the device operation state and obtain the time-frequency domain characteristic data;

[0023] S124. Integrate the time-domain, frequency-domain, and time-frequency domain characteristic data of the device to obtain the device characteristic dataset.

[0024] Furthermore, the calculation formula of Fourier transform is:

[0025]

[0026] In the formula, X b represents the amplitude of the b-th frequency component; x n represents the standardized data at the n-th time point; n represents the n-th time point; N is the total number of time points; b represents the index of the frequency component; i represents the imaginary unit;

[0027] The calculation formula of wavelet transform is:

[0028]

[0029] In the formula, W j,k represents the wavelet coefficient; ψ j,k (n) represents the wavelet basis function.

[0030] Furthermore, based on the digital twin model and the implicit semi-Markov model, combined with the device operation characteristic data, constructing the device aging model includes:

[0031] S21. Based on the digital twin model, construct the device operation state model, and combined with the device operation characteristic data, obtain the device operation state data;

[0032] S22. Discretize the device operation state data to obtain the discretized state data, and set the device state in the implicit semi-Markov model according to the device state index;

[0033] S23. Use the implicit semi-Markov model, combined with the discretized state data, to construct the device aging model.

[0034] Furthermore, based on the digital twin model, construct the device operation state model, and combined with the device operation characteristic data, obtaining the device operation state data includes:

[0035] S211. Use the subway train equipment operation knowledge base to construct a physical model of equipment operation, and use machine learning algorithms to construct a model of equipment operation rules in combination with equipment operation characteristic data;

[0036] S212. Use the weighted average method to linearly fuse the physical model of equipment operation and the model of equipment operation rules to obtain an initial equipment operation state model;

[0037] S213. Verify the equipment operation state model through simulation, and optimize the parameters of the equipment operation state model according to the verification results to obtain the equipment operation state model;

[0038] S214. Based on the equipment operation state model, combine the equipment operation characteristic data to obtain the equipment operation state data.

[0039] Furthermore, use the implicit semi-Markov model to construct an equipment aging model in combination with discretized state data, including:

[0040] S231. Based on the implicit semi-Markov model, combine the discretized state data to analyze the equipment state transition frequency, and use the maximum likelihood estimation to obtain the state transition probability;

[0041] S232. Use the distribution function to combine the discretized state data to obtain the equipment state duration, and construct an initial equipment aging model according to the state transition probability;

[0042] S233. Optimize the parameters of the initial equipment aging model through cross-validation to obtain the equipment aging model.

[0043] Furthermore, obtain the equipment operation characteristic data in real time, use the equipment aging model to monitor and update the equipment aging state, and conduct fault prediction based on the equipment aging state, including:

[0044] S31. Obtain the equipment operation characteristic data in real time, use the equipment aging model to monitor and update the equipment aging state;

[0045] S32. Based on the equipment aging state, evaluate the health status of the equipment, and analyze the aging trend and fault risk of the equipment according to the health status of the equipment;

[0046] S33. According to the aging trend and fault risk of the equipment, conduct equipment fault prediction to obtain the fault prediction result.

[0047] Furthermore, the fault information includes: fault location information, aging degree information, and fault time prediction information.

[0048] According to another aspect of the present invention, a device fault location and early warning system based on aging degree analysis is provided. The device fault location and early warning system based on aging degree analysis includes: a device operation feature extraction module, a device aging model construction module, a device fault prediction module, and a device fault location and early warning module;

[0049] The device operation feature extraction module is used to obtain the historical operation data of the device, preprocess the historical operation data of the device to obtain standardized data, and extract features from the standardized data to obtain device operation feature data;

[0050] The device aging model construction module is used to construct a device aging model based on the digital twin model and the implicit semi-Markov model in combination with the device operation feature data;

[0051] The device fault prediction module is used to obtain the device operation feature data in real time, use the device aging model to monitor and update the device aging state, and perform fault prediction based on the device aging state;

[0052] The device fault location and early warning module is used to extract fault information based on the fault prediction result in combination with the obtained device operation feature data, and perform fault location and early warning according to the fault information.

[0053] The beneficial effects of the present invention are as follows:

[0054] 1. Through the device aging model, the present invention monitors and updates the device aging state, uses the device aging state for fault prediction, ensures that potential faults can be identified in advance and the root cause of the problem can be accurately located, and formulates an effective maintenance plan; thereby avoiding service interruption caused by potential faults, seizing the best maintenance opportunity, optimizing the maintenance cost, extending the service life of the device, and at the same time, it can more accurately evaluate the device state and its development trend, optimize the maintenance strategy, and ensure the safe and efficient operation of the subway train.

[0055] 2. Through the digital twin model and the implicit semi-Markov model, the present invention accurately masters the device operation state data, state transition probability, and device state duration, so as to be able to capture the physical changes and performance degradation of the device over time, accurately grasp the wear accumulation within the device life cycle, predict the aging trend of the device, give early warning to the device about to fail, reduce service interruption caused by sudden faults, and thus ensure that the subway train can maintain the best operation state for a long time. Description of the Drawings

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required in the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0057] Figure 1 is a flowchart of a device fault location and warning method based on aging degree analysis according to an embodiment of the present invention;

[0058] Figure 2 is a principle block diagram of a device fault location and warning system based on aging degree analysis according to an embodiment of the present invention.

[0059] In the figure:

[0060] 1. Device operation feature extraction module; 2. Device aging model construction module; 3. Device fault prediction module; 4. Device fault location and warning module. Detailed implementation manners

[0061] To further illustrate the embodiments, the present invention provides accompanying drawings. These accompanying drawings are part of the disclosure of the present invention. They are mainly used to illustrate the embodiments and can be combined with the relevant descriptions in the specification to explain the operation principle of the embodiments. With reference to these contents, those of ordinary skill in the art should be able to understand other possible implementation manners and the advantages of the present invention.

[0062] According to an embodiment of the present invention, a device fault location and warning method and system based on aging degree analysis are provided.

[0063] Now, the present invention will be further described in combination with the accompanying drawings and specific implementation manners. As Figure 1 shown, according to an embodiment of the present invention, a device fault location and warning method based on aging degree analysis is provided. The method includes:

[0064] S1. Obtain the historical operation data of the device, preprocess the historical operation data of the device to obtain standardized data, and perform feature extraction on the standardized data to obtain device operation feature data.

[0065] It should be noted that obtaining the historical operation data of the device specifically means obtaining information such as the historical operation temperature, vibration, pressure, current, operation time of the device, and device load.

[0066] Specifically, obtaining the historical operation data of the device, preprocessing the historical operation data of the device to obtain standardized data, and performing feature extraction on the standardized data to obtain device operation feature data includes:

[0067] S11. Obtain the historical operation data of the device, perform noise reduction processing on the historical operation data of the device, and perform standardization processing on the denoised data to obtain standardized data.

[0068] It should be noted that the noise reduction processing includes removing missing values, outliers or noise data; for missing values, interpolation or mean filling is adopted; for outliers, detection and correction are performed through the 3-sigma criterion.

[0069] It should be noted that the standardization processing is to normalize the collected data to make it meet the unified dimension and range.

[0070] S12. According to the standardized data, extract the time domain, frequency domain and time-frequency domain feature data of the device, and perform feature integration on the time domain, frequency domain and time-frequency domain feature data of the device to obtain the device feature dataset.

[0071] Specifically, according to the standardized data, extracting the time domain, frequency domain and time-frequency domain feature data of the device, and performing feature integration on the time domain, frequency domain and time-frequency domain feature data of the device to obtain the device feature data includes:

[0072] S121. According to the standardized data, calculate the statistical features of the device operation state to obtain the time domain feature data of the device;

[0073] It should be noted that the statistical features of the device operation state include mean, variance, standard deviation, peak value, valley value, kurtosis and skewness, etc.

[0074] S122. Use the Fourier transform, combined with the standardized data, to calculate the frequency features of the device operation state to obtain the frequency domain feature data.

[0075] It should be noted that using the Fourier transform, combined with the standardized data, to calculate the frequency features of the device operation state specifically means performing the Fourier transform on the standardized data to obtain the spectrogram, and calculating the main frequency, spectral entropy, average frequency and root mean square frequency, etc.

[0076] S123. Use the wavelet transform, combined with the standardized data, to calculate the wavelet energy and wavelet entropy of the device operation state to obtain the time-frequency domain feature data.

[0077] S124. Integrate the time domain, frequency domain and time-frequency domain feature data of the device to obtain the device feature dataset.

[0078] Specifically, the calculation formula of the Fourier transform is:

[0079]

[0080] In the formula, X b represents the amplitude of the b-th frequency component; x nThe standardized data representing the nth time point; n represents the nth time point; N is the total number of time points; b represents the index of the frequency component; i represents the imaginary unit;

[0081] The calculation formula of wavelet transform is:

[0082]

[0083] In the formula, W j,k represents the wavelet coefficient; ψ j,k (n) represents the wavelet basis function.

[0084] It should be noted that the frequency characteristics of the operating state of the computing device include the main frequency, spectral entropy, average frequency, and root mean square frequency calculated based on X b The calculated main frequency, spectral entropy, average frequency, and root mean square frequency.

[0085] It should be noted that the calculation formula of wavelet energy is:

[0086]

[0087] In the formula, E represents the wavelet energy; j represents the scale index; k represents the translation index; J represents the maximum value of the scale index; K represents the maximum value of the translation index.

[0088] It should be noted that the calculation formula of wavelet entropy is:

[0089]

[0090] In the formula, T represents the wavelet entropy; p j,k represents the normalized wavelet energy probability distribution.

[0091] Among them, the calculation formula of the normalized wavelet energy probability distribution is:

[0092]

[0093] It should be noted that integrating the time domain, frequency domain, and time-frequency domain characteristic data of the device into a characteristic data set specifically means unifying the data formats of the time domain, frequency domain, and time-frequency domain characteristic data of the device to form a characteristic set.

[0094] S13. Use the Pearson correlation coefficient to perform an aging degree correlation analysis on the device characteristic data, and remove redundant characteristics based on the correlation analysis results to obtain the device operation characteristic data;

[0095] Among them, the calculation formula of the Pearson correlation coefficient is:

[0096]

[0097] In the formula, Z kIt represents the Pearson correlation coefficient of the k-th device feature in the device feature dataset; t represents device t; a represents the total number of devices; T k,t It represents the value of the k-th device feature in device t; It represents the mean value of the k-th device feature of all devices; L t It represents the aging degree of device t; It represents the mean value of the aging degrees of all devices.

[0098] Among them, the specific value of the aging degree of device t is as follows: when device t is in a healthy state, the value of the aging degree is zero; when device t is in a mildly aged state, the value of the aging degree is one; when device t is in a moderately aged state, the value of the aging degree is two; when device t is in a severely aged state, the value of the aging degree is three; when device t is in a faulty state, the value of the aging degree is four.

[0099] It should be noted that based on the results of the correlation analysis for redundant feature removal, by setting the threshold of the Pearson correlation coefficient, when the Pearson correlation coefficient is above 0.6, the higher the correlation between the device feature data and the aging degree; then select two device features with high correlation and use the Pearson correlation coefficient to calculate the correlation between the two device features. When the threshold is 0.8 or 0.9, it indicates that the correlation between the two device features is high, and one of the related features needs to be removed.

[0100] S2. Based on the digital twin model and the implicit semi-Markov model, combined with the device operation feature data, construct a device aging model.

[0101] Specifically, based on the digital twin model and the implicit semi-Markov model, combined with the device operation feature data, constructing a device aging model includes:

[0102] S21. Based on the digital twin model, construct a device operation state model, and combined with the device operation feature data, obtain the device operation state data.

[0103] Specifically, based on the digital twin model, construct a device operation state model, and combined with the device operation feature data, obtaining the device operation state data includes:

[0104] S211. Utilize the subway train device operation knowledge base to construct a device operation physical model, and use machine learning algorithms, combined with the device operation feature data, to construct a device operation law model.

[0105] It should be noted that the subway train equipment operation knowledge base includes the physical characteristics of subway train equipment (such as motor torque, load torque, etc.), thermodynamic equations, mechanical equations, and electrical equations. Specifically, constructing the physical model of equipment operation is based on the performance indicators of the equipment, such as power output, efficiency, etc.; the physical parameters of the equipment, such as the resistance, inductance, and thermal resistance of the motor, etc.; and the operating conditions of the equipment, such as load, speed, ambient temperature, etc. First, select the physical model, substitute the physical characteristics of the equipment into the thermodynamic equation to describe the heat transfer and heat dissipation of the equipment during operation, and obtain Physical Model 1; substitute the physical parameters of the equipment into the electrical equation to describe the electrical characteristics of the equipment, and obtain Physical Model 2; then substitute the performance indicators of the equipment, the operating conditions of the equipment, and the physical characteristics of the equipment to select the corresponding data into the mechanical equation to describe the mechanical characteristics of the equipment, and obtain Physical Model 3. Then, construct the numerical simulation of the physical model, initialize the input parameters of the model, including performance indicators, physical parameters, and operating conditions; use the mathematical form of the physical model for numerical simulation to calculate the state of the equipment under different operating conditions; record the simulation results, including the operating state and performance indicators of the equipment; use the actual operating data to verify the accuracy of the physical model, compare the results predicted by the model with the actual measured data, and evaluate the error of the model; adjust the model parameters according to the verification results to optimize the performance of the model; for example, adjust the value of the thermal resistance or resistance to improve the accuracy of the model; construct a physical model that can accurately describe the operating state of the equipment. This physical model of equipment operation can not only simulate the behavior of the equipment under different operating conditions, but also provide a basis for subsequent equipment operation law models and aging analysis.

[0106] It should be noted that by using the linear regression model and combining the equipment operation characteristic data, the coefficients of the linear regression model can represent the influence degree of each characteristic on the equipment aging degree; by analyzing the magnitude of the coefficients, the characteristics that have the greatest influence on equipment aging can be identified. Specifically, constructing the equipment operation law model is to divide the equipment operation characteristic data into a training set and a test set, usually in a ratio of 80% training set and 20% test set, and use the training set to input into the linear regression model to construct the equipment operation law model for predicting the continuous value of the equipment aging degree.

[0107] S212. Use the weighted average method to linearly fuse the physical model of equipment operation and the equipment operation law model to obtain the initial equipment operation state model.

[0108] It should be noted that according to the importance of the description of the equipment operation state by the physical model of equipment operation and the equipment operation law model, set the weights corresponding to the physical model of equipment operation and the equipment operation law model, and perform weighted summation of the weights and the output results of the corresponding physical model of equipment operation and the output results of the equipment operation law model to obtain the total output result, and then obtain the initial equipment operation state model.

[0109] S213. Simulate and verify the device operation status model, and optimize the parameters of the device operation status model according to the verification results to obtain the device operation status model.

[0110] It should be noted that simulating and verifying the device operation status model means comparing the output results of the device operation status model with the actual operation status, and calculating and judging the error between the output results of the device operation status model and the actual operation status.

[0111] S214. Based on the device operation status model and combined with the device operation characteristic data, obtain the device operation status data.

[0112] S22. Discretize the device operation status data to obtain discretized status data, and set the device status in the implicit semi-Markov model according to the status indicators of the device.

[0113] It should be noted that the device status in the implicit semi-Markov model specifically includes the healthy state, the mild aging state, the moderate aging state, the severe aging state, and the fault state.

[0114] Among them, the healthy state means that the device operates normally without signs of aging; the mild aging state means that the device performance begins to decline but does not affect normal operation; the moderate aging state means that the device performance further declines and potential faults may occur; the severe aging state means that the device performance significantly declines and approaches the fault critical point; the fault state means that the device cannot operate normally and needs to be shut down for maintenance.

[0115] S23. Use the implicit semi-Markov model and combine with the discretized status data to construct a device aging model.

[0116] Specifically, using the implicit semi-Markov model and combining with the discretized status data to construct a device aging model includes:

[0117] S231. Based on the implicit semi-Markov model and combined with the discretized status data, analyze the device state transition frequency and obtain the state transition probability using the maximum likelihood estimation.

[0118] S232. Use the distribution function and combine with the discretized status data to obtain the device state duration, and construct an initial device aging model according to the state transition probability.

[0119] S233. Through cross-validation, optimize the parameters of the initial device aging model to obtain the device aging model.

[0120] It should be noted that the device operation status data specifically includes the healthy state, mild aging state, moderate aging state, severe aging state, and fault state according to the device status. The data thresholds for dividing the device status are used to divide the device operation status data according to the actual device status data thresholds, and discretized status data is obtained.

[0121] It should be noted that based on the implicit semi-Markov model, combined with the discretized status data, the device status transition frequency is analyzed, and the state transition probability is obtained by using the maximum likelihood estimation. Specifically, the number of transitions between different states of the device is counted to obtain the state transition matrix, and the maximum likelihood estimation is used to calculate the state transition probability matrix; and the duration of the device status is continuously recorded.

[0122] S3. Real-time obtain the device operation characteristic data, use the device aging model to monitor and update the device aging state, and perform fault prediction based on the device aging state.

[0123] Specifically, real-time obtaining the device operation characteristic data, using the device aging model to monitor and update the device aging state, and performing fault prediction based on the device aging state includes:

[0124] S31. Real-time obtain the device operation characteristic data, use the device aging model to monitor and update the device aging state.

[0125] It should be noted that real-time obtaining the device operation characteristic data is specifically to collect the device operation data in real time through sensors, preprocess the device historical operation data to obtain standardized data, and perform feature extraction on the standardized data to obtain the device operation characteristic data.

[0126] S32. Based on the device aging state, evaluate the health status of the device, and analyze the aging trend and fault risk of the device according to the health status of the device;

[0127] S33. According to the aging trend and fault risk of the device, perform device fault prediction to obtain the fault prediction result.

[0128] It should be noted that using the equipment aging model to monitor and update the equipment aging status specifically means obtaining the current equipment status, the equipment duration under the current equipment status, the equipment running time under the current equipment status, and the state transition frequency under the current equipment status through the equipment aging model. The remaining time to jump to the next equipment status is obtained by subtracting the equipment running time under the current equipment status from the equipment duration under the current equipment status. Then, the remaining time of the current equipment status is obtained by multiplying the state transition frequency under the current equipment status by the remaining time to jump to the next equipment status. The data obtained from the equipment aging model is used to overwrite the previously obtained equipment status information to achieve the update of the equipment aging status. Based on the current equipment status, that is, if the current equipment status is a healthy state, a mild aging state, or a moderate aging state, the equipment is judged to be healthy; if the current equipment status is a severe aging state or a fault state, the equipment is judged to be unhealthy; and the aging trend of the equipment is judged by combining the remaining time of the current equipment status and the equipment health condition. If the equipment is healthy, the aging trend is judged through the current equipment status and the remaining time of the current equipment status, that is, if it is in a healthy state and the remaining time of the current equipment status is greater than half of the equipment duration under the current equipment status, the aging trend of the equipment is non-aging and the aging speed is slow; if it is in a mild aging state and the remaining time of the current equipment status is greater than half of the equipment duration under the current equipment status, the aging trend of the equipment is mild aging and the aging speed is medium; if it is in a moderate aging state and the remaining time of the current equipment status is greater than half of the equipment duration under the current equipment status, the aging trend of the equipment is moderate aging and the aging speed is fast; if it is an unhealthy equipment and the remaining time of the current equipment status is greater than half of the equipment duration under the current equipment status, the aging trend of the equipment is severe aging and the aging speed is the fastest. And according to the equipment duration under the current equipment status of the unhealthy equipment, a reminder is given that there is a risk of failure, and the fault prediction result is the equipment status and the equipment duration under the current equipment status.

[0129] S4. Based on the fault prediction result, combined with the obtained equipment operation characteristic data, extract fault information, and conduct fault location warning according to the fault information.

[0130] It should be noted that the aging degree information is obtained according to the equipment status of the fault prediction result, the equipment status and the equipment duration under the current equipment status are used for fault time prediction to obtain the fault time prediction information, the fault location information of the faulty equipment is obtained through the obtained equipment operation characteristic data, the fault location is carried out according to the fault location information, the fault warning time point is determined according to the aging degree information and the fault time prediction information, and the fault location warning is carried out in combination with the warning equipment.

[0131] Specifically, the fault information includes: fault location information, aging degree information, and fault time prediction information.

[0132] Such as Figure 2As shown in the figure, according to another embodiment of the present invention, a device fault location and early warning system based on aging degree analysis is provided. The device fault location and early warning system based on aging degree analysis includes: a device operation feature extraction module 1, a device aging model construction module 2, a device fault prediction module 3, and a device fault location and early warning module 4;

[0133] The device operation feature extraction module 1 is used to obtain the historical operation data of the device, preprocess the historical operation data of the device to obtain standardized data, and extract features from the standardized data to obtain device operation feature data;

[0134] The device aging model construction module 2 is used to construct a device aging model based on the digital twin model and the implicit semi-Markov model in combination with the device operation feature data;

[0135] The device fault prediction module 3 is used to obtain the device operation feature data in real time, monitor and update the device aging state by using the device aging model, and perform fault prediction based on the device aging state;

[0136] The device fault location and early warning module 4 is used to extract fault information based on the fault prediction result in combination with the obtained device operation feature data, and perform fault location and early warning according to the fault information.

[0137] In summary, by means of the above technical solutions of the present invention, the present invention monitors and updates the device aging state through the device aging model, uses the device aging state for fault prediction, ensures that potential faults can be identified in advance and the root cause of the problem can be accurately located, and formulates an effective maintenance plan; thus avoiding service interruption caused by potential faults, seizing the best maintenance opportunity, optimizing the maintenance cost, extending the service life of the device, and at the same time can more accurately evaluate the device state and its development trend, optimize the maintenance strategy, and ensure the safe and efficient operation of the subway train; through the digital twin model and the implicit semi-Markov model, accurately master the device operation state data, state transition probability, and device state duration, so as to be able to capture the physical changes and performance decline of the device over time, accurately grasp the wear accumulation within the device life cycle, predict the aging trend of the device, give early warning to the device about to fail, reduce service interruption caused by sudden faults, and thus ensure that the subway train can maintain the best operation state for a long time.

[0138] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for equipment fault location and early warning based on aging degree analysis, characterized in that The method includes: S1. Obtain the historical operation data of the device, preprocess the historical operation data of the device to obtain standardized data, and extract features from the standardized data to obtain device operation feature data; S2. Based on the digital twin model and the implicit semi-Markov model, combine the device operation feature data to construct a device aging model; S3. Obtain the device operation feature data in real time, use the device aging model to monitor and update the device aging state, and perform fault prediction based on the device aging state; S4. Based on the fault prediction result, combine the obtained device operation feature data, extract fault information, and perform fault location warning according to the fault information.

2. The device fault location and early warning method based on aging degree analysis according to claim 1, characterized in that, The obtaining the historical operation data of the device, preprocessing the historical operation data of the device to obtain standardized data, and extracting features from the standardized data to obtain device operation feature data includes: S11. Obtain the historical operation data of the device, perform noise reduction processing on the historical operation data of the device, and perform standardized processing on the noise-reduced data to obtain standardized data; S12. According to the standardized data, extract the time domain, frequency domain, and time-frequency domain feature data of the device, and perform feature integration on the time domain, frequency domain, and time-frequency domain feature data of the device to obtain a device feature dataset; S13. Use the Pearson correlation coefficient to perform correlation analysis of the aging degree on the device feature dataset, and remove redundant features based on the correlation analysis result to obtain device operation feature data; Among them, the calculation formula of the Pearson correlation coefficient is: Where Z k represents the Pearson correlation coefficient of the k-th device feature in the device feature dataset; t represents device t; a represents the total number of devices; T k,t represents the value of the k-th device feature in device t; represents the mean value of the k-th device feature of all devices; L t represents the aging degree of device t; represents the mean value of the aging degrees of all devices.

3. The device fault location and warning method based on aging degree analysis according to claim 2, characterized in that The extracting the time domain, frequency domain, and time-frequency domain feature data of the device according to the standardized data, and performing feature integration on the time domain, frequency domain, and time-frequency domain feature data of the device to obtain a device feature dataset includes: S121. According to the standardized data, calculate the statistical features of the device operation state to obtain the time domain feature data of the device; S122. Use the Fourier transform, combine with the standardized data, calculate the frequency features of the device operation state to obtain the frequency domain feature data; S123. Use the wavelet transform, combine with the standardized data, calculate the wavelet energy and wavelet entropy of the device operation state to obtain the time-frequency domain feature data; S124. Integrate the time domain, frequency domain, and time-frequency domain feature data of the device to obtain a device feature dataset.

4. The device fault location and early warning method based on aging degree analysis according to claim 3, characterized in that, The calculation formula of the Fourier transform is: where X b represents the amplitude of the b-th frequency component; x n represents the normalized data at the n-th time point; n represents the n-th time point; N is the total number of time points; b represents the index of the frequency component; i represents the imaginary unit; The calculation formula of the wavelet transform is: Where, W j,k represents the wavelet coefficient; ψ j,k (n) represents the wavelet basis function.

5. A method for device fault location and early warning based on aging degree analysis according to claim 1, characterized in that, The constructing a device aging model based on the digital twin model and the implicit semi-Markov model, and combining the device operation feature data includes: S21. Based on the digital twin model, construct a device operation state model, and combine with the device operation feature data to obtain device operation state data; S22. Discretize the device operation state data to obtain discretized state data, and set the device state in the implicit semi-Markov model according to the state index of the device; S23. Use the implicit semi-Markov model, combine with the discretized state data, to construct a device aging model.

6. The method for device fault location and early warning based on aging degree analysis according to claim 5, wherein, The constructing a device operation state model based on the digital twin model, and combining with the device operation feature data to obtain device operation state data includes: S211. Use the subway train equipment operation knowledge base to construct a physical model of equipment operation, and use machine learning algorithms to construct a model of equipment operation rules in combination with equipment operation characteristic data. S212. Use the weighted average method to linearly fuse the physical model of equipment operation and the model of equipment operation rules to obtain an initial equipment operation state model. S213. Conduct simulation verification on the equipment operation state model, and optimize the parameters of the equipment operation state model according to the verification results to obtain the equipment operation state model. S214. Based on the equipment operation state model, combine with the equipment operation characteristic data to obtain equipment operation state data.

7. A method for equipment fault location and early warning based on aging degree analysis according to claim 5, characterized in that, The construction of the equipment aging model by using the implicit semi-Markov model in combination with discretized state data includes: S231. Based on the implicit semi-Markov model, combine with the discretized state data to analyze the equipment state transition frequency, and use the maximum likelihood estimation to obtain the state transition probability. S232. Use the distribution function to combine with the discretized state data to obtain the equipment state duration, and construct an initial equipment aging model according to the state transition probability. S233. Through cross-validation, optimize the parameters of the initial equipment aging model to obtain the equipment aging model.

8. A method for equipment fault location and early warning based on aging degree analysis according to claim 1, characterized in that, The real-time acquisition of equipment operation characteristic data, the use of the equipment aging model to monitor and update the equipment aging state, and the fault prediction based on the equipment aging state include: S31. Real-time acquire equipment operation characteristic data, use the equipment aging model to monitor and update the equipment aging state. S32. Based on the equipment aging state, evaluate the health status of the equipment, and analyze the aging trend and fault risk of the equipment according to the health status of the equipment. S33. According to the aging trend and fault risk of the equipment, conduct equipment fault prediction to obtain the fault prediction result.

9. A method for device fault location and warning based on aging degree analysis according to claim 1, characterized in that, The fault information includes: fault location information, aging degree information, and fault time prediction information.

10. An equipment fault location and early warning system based on aging degree analysis, which is used to implement the equipment fault location and early warning method based on aging degree analysis described in any one of claims 1-9, and is characterized in that, The equipment fault location and warning system based on aging degree analysis includes: an equipment operation characteristic extraction module, an equipment aging model construction module, an equipment fault prediction module, and an equipment fault location and warning module. The equipment operation characteristic extraction module is used to acquire the equipment historical operation data, preprocess the equipment historical operation data to obtain standardized data, and extract features from the standardized data to obtain equipment operation characteristic data. The equipment aging model construction module is used to construct an equipment aging model based on the digital twin model and the implicit semi-Markov model in combination with the equipment operation characteristic data. The equipment fault prediction module is used to real-time acquire equipment operation characteristic data, use the equipment aging model to monitor and update the equipment aging state, and conduct fault prediction based on the equipment aging state. The equipment fault location and warning module is used to extract fault information based on the fault prediction result in combination with the acquired equipment operation characteristic data, and conduct fault location and warning according to the fault information.

Images