Electric motor cooling control system having fault diagnosis function
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-03-18
- Publication Date
- 2026-08-13
Smart Images
Figure CN2025083189_13082026_PF_FP_ABST
Abstract
Description
A motor cooling control system with fault diagnosis function Technical Field
[0001] This application relates to the field of motor cooling control, and in particular to a motor cooling control system with fault diagnosis function. Background Technology
[0002] Motors generate a significant amount of heat during operation. If cooling is not implemented promptly, the motor temperature will overheat, affecting its performance and lifespan. Therefore, a proper cooling control method is an indispensable part of motor design. Water cooling is a commonly used motor cooling method. It involves installing water cooling pipes inside or outside the motor, using circulating water to absorb and remove the heat generated by the motor. Water cooling offers excellent heat dissipation and is suitable for high-power, high-heat-generating motors. However, water cooling systems are relatively complex and prone to failure. Failure to address these faults in advance can lead to serious consequences such as motor overheating.
[0003] In the existing technology, motor cooling systems are usually inspected manually on a regular basis. This method is inefficient and cannot detect faults before they occur or in their early stages, let alone predict and handle them. Summary of the Invention
[0004] In order to solve the technical problem of fault prediction in the existing motor cooling control system, this application provides a motor cooling control system with fault diagnosis function.
[0005] This application is achieved through the following technical solution:
[0006] A motor cooling control system with fault diagnosis function includes: a water pump, a water pump controller, pipelines, several vibration sensors, a flow sensor, a differential pressure sensor, a temperature sensor, and a fault diagnosis module.
[0007] The vibration sensor is installed on the water pump and the pipeline; the flow sensor and differential pressure sensor are installed in the pipeline; the temperature sensor is installed on the cooled motor; the water pump controller controls the water pump.
[0008] The fault diagnosis module predicts faults in the motor cooling control system through the following steps:
[0009] S1: Acquire signals to obtain the current, speed, and vibration signals of the water pump, the temperature signal of the motor, and the vibration signal, flow rate, and pressure difference of the pipeline;
[0010] S2: Process the acquired signals and extract features;
[0011] S3: Classify the state of the motor cooling control system based on K-means clustering algorithm combined with LSTM model;
[0012] S4: Determine the status of the motor cooling control system; if the current performance of the water pump in the motor cooling control system is degraded, but there is no water pump failure, pipe blockage or pipe leakage, proceed to step S5; otherwise, proceed to step S6.
[0013] S5: Calculate the pump degradation level and make adaptive adjustments; and
[0014] S6: Predict the failure probability of the system based on the competing failure model and provide early warning.
[0015] Furthermore, the processing of the acquired signals includes blind source separation of the vibration signals of the water pump and the pipeline based on the independent component analysis method.
[0016] Furthermore, the classification results of the motor cooling control system status include normal, water pump performance degradation, water pump failure, pipe blockage, and pipe leakage; among them, water pump failure, pipe blockage, and pipe leakage are respectively divided into three levels of severity: minor, moderate, and severe.
[0017] Furthermore, step S3 also includes:
[0018] S31: Cluster the sample data based on the K-means clustering algorithm;
[0019] S32: Determine training and validation samples based on the clustering results;
[0020] S33: Build and train the LSTM model; and
[0021] S34: Based on the trained LSTM model, the input data is classified to obtain the classification results of the state of the motor cooling control system.
[0022] Furthermore, determining the training and validation samples based on the clustering results includes performing mixed sampling on the clustering results to determine the training and validation samples, wherein the mixed sampling includes stratified sampling and random sampling.
[0023] Stratified sampling obtains sampling results from N clustered categorical samples and discrete samples, as shown in the following expression:
[0024] Among them, D i Let r be the sample set of the i-th cluster center among N cluster classification samples. i L represents the sampling ratio, and L is the discrete sample set.
[0025] Furthermore, r i Determined in the following manner:
[0026] Where, k i DB is the adjustment coefficient. i Let q0 be the Davis-Bolding index of the i-th cluster, q0 be the set Davis-Bolding index threshold, and r0 be a set constant.
[0027] Furthermore, step S5 includes:
[0028] S51: Calculate the degree of degradation of the water pump;
[0029] S52: When the degradation degree of the water pump reaches a certain threshold, the control parameters of the water pump are calculated based on the adaptive control algorithm and sent to the water pump controller to realize adaptive adjustment control.
[0030] Furthermore, the calculation of the pump's degradation degree includes:
[0031] Obtain the time series of the pump's current, pressure, and flow rate; predict the pump's current, pressure, and flow rate based on the autoregressive integral moving average model with periodic correlation; calculate the pump's degradation degree based on the prediction results of the time series.
[0032] Furthermore, the control of the adaptive adjustment includes:
[0033] An adaptive adjustment model is constructed based on an FNN network to obtain the degradation degree of the water pump, which is used as the input parameter of the adaptive adjustment model. The degradation degree is then input into the adaptive adjustment model to obtain the control parameters of the water pump output by the adaptive adjustment model.
[0034] Furthermore, step S6 includes: S61: Calculate the reliability R(t) of the motor cooling control system: R(t) = R1(t) + R2(t)
[0035] Where R1(t) is the reliability of the motor cooling control system under time t when only the water pump deteriorates and fails without any sudden failure factors, and R2(t) is the reliability of the motor cooling control system under time t when sudden failure factors are present.
[0036] S62: Calculate and predict the failure probability of the motor cooling control system:
[0037] Where f(t) is the system density function, expressed as follows:
[0038] Compared with the prior art, the beneficial effects of this application are as follows:
[0039] This application provides early warning for faults such as pipe blockage, pipe leakage, pump degradation, and pump failure, and features adaptive compensation. It automatically adjusts control parameters to maintain optimal system performance and improve overall reliability when some components degrade. To enhance prediction accuracy, this application uses independent component analysis to perform blind source separation of pump and pipe vibration signals. Furthermore, it employs a clustering algorithm to perform cluster sampling of the sample data, ensuring the comprehensiveness and accuracy of the training data through mixed sampling and reducing sample redundancy. Additionally, in predicting pump degradation, a periodically correlated autoregressive integral moving average model is used, taking into account the periodic changes in motor temperature, thus improving prediction accuracy. Attached Figure Description
[0040] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:
[0041] Figure 1 is a schematic diagram of the fault prediction process of the motor cooling control system according to an embodiment of this application;
[0042] Figure 2 is a schematic diagram of the method for classifying the state of a motor cooling control system based on the K-means clustering algorithm combined with the LSTM model according to an embodiment of this application. Detailed Implementation
[0043] The embodiments of this application will now be described in detail with reference to the accompanying drawings.
[0044] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0045] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The illustrations only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the shape, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.
[0046] A motor cooling control system with fault diagnosis function includes a water pump, a water pump controller, pipes, several vibration sensors, a flow sensor, a differential pressure sensor, a temperature sensor, and a fault diagnosis module.
[0047] The vibration sensor is installed on the water pump and the pipeline; the flow sensor and the differential pressure sensor are installed in the pipeline; the temperature sensor is installed on the cooled motor; the water pump controller controls the water pump.
[0048] As shown in Figure 1, the fault diagnosis module predicts faults in the motor cooling control system through the following steps:
[0049] S1: Acquire signals, including: current, speed and vibration signals of the water pump, temperature signal of the motor, vibration signal of the pipeline, flow rate and pressure difference;
[0050] The vibration signal of the water pump is acquired by a vibration sensor installed on the water pump; the vibration signal of the pipeline is acquired by a vibration sensor installed on the pipeline; the temperature signal of the motor refers to the temperature of the motor being cooled;
[0051] S2: Process the acquired signals and extract features, including:
[0052] S21: Signal preprocessing, including filtering, denoising and normalizing the acquired signal;
[0053] S22: Perform blind source separation on the vibration signals of the water pump and the pipeline;
[0054] Since the vibrations of the water pump and the pipeline affect each other, the vibration signals acquired by the vibration sensors installed on the water pump and the pipeline are the result of their mutual influence. To achieve more accurate subsequent signal analysis and processing, it is necessary to separate the two signals. This application uses the independent component analysis method to restore the observed vibration signals received from the two vibration sensors into independent vibration source signals, including the following steps:
[0055] a. Preprocess the acquired vibration sensor signals, including centering and whitening.
[0056] b. The preprocessed vibration sensor signal is pre-separated by the convolutional ICA algorithm to obtain pre-separated signals y1 and y2, which correspond to the water pump vibration source signal and the pipeline vibration source signal, respectively.
[0057] c. Normalize y1(n) and y2(n), then perform short-time Fourier transform to the time-frequency domain to obtain Y1 and Y2. Estimate the binarized mask and perform cepstral smoothing.
[0058] d. Enhance Y1 and Y2 using the mask values after cepstral smoothing.
[0059] e. Using inverse short-time Fourier transform to... Converted into pure, isolated vibration source signals That is, the vibration signal of the water pump and the vibration signal of the pipeline after processing.
[0060] S23: Feature extraction, including time-domain and frequency-domain feature extraction of pump current, pump vibration signal, and pipeline vibration signal; and time-domain feature extraction of pump speed, motor temperature signal, pipeline flow rate, and differential pressure signal.
[0061] S3: The state of the motor cooling control system is classified based on the K-means clustering algorithm combined with the LSTM model, as shown in Figure 2. The classification results include normal, water pump performance degradation, water pump failure, pipe blockage, and pipe leakage. Among them, water pump failure, pipe blockage, and pipe leakage are respectively divided into three levels of severity: minor, moderate, and severe.
[0062] S31: Cluster the sample data based on the K-means clustering algorithm;
[0063] In a preferred embodiment, clustering includes the following steps:
[0064] a. Import the sample dataset into the list; the sample dataset includes the acquired signals and extracted features;
[0065] b. For a given dataset, construct a set containing N cluster centers. Using the improved k-means algorithm based on the minimax principle, initialize the cluster centers, and then perform k-means clustering on each category of data. Finally, the N optimal cluster centers are obtained.
[0066] c. Calculate the Euclidean distance between two sample vectors;
[0067] d. Create a matrix to store the cluster assignment results for each point, including the cluster index value and error. Calculate the distance from the current sample to each cluster center, find the nearest cluster center, and assign the sample to the nearest cluster.
[0068] e. After all samples have been assigned, recalculate the N cluster centers;
[0069] f. Compare the N cluster centers obtained in the previous calculation. If the cluster centers have changed, proceed to step c.
[0070] Otherwise, proceed to step g;
[0071] g. Stop and output the clustering results when the cluster centers do not change;
[0072] h. The clustering results are merged, and duplicate results are removed to obtain the final set.
[0073] S32: Determine training and validation samples based on the clustering results;
[0074] This includes performing mixed sampling on the clustering results to determine training and validation samples. The clustering results include N clustered classification samples and discrete samples. The mixed sampling includes stratified sampling and random sampling. The proportion of training and validation samples is set to a% obtained by stratified sampling and b% obtained by random sampling, and a% + b% = 1.
[0075] In a preferred embodiment, sampling results are obtained from N clustered classification samples and discrete samples through stratified sampling, as expressed below:
[0076] Among them, D i Let r be the sample set of the i-th cluster center among N cluster classification samples. i L represents the sampling ratio, and L is the discrete sample set.
[0077] Optional, r i Determined in the following manner:
[0078] Where, k i DB is an adjustment coefficient, determined based on actual conditions. i Let q0 be the Davis-Bolding index of the i-th cluster, q0 be the set Davis-Bolding index threshold, and r0 be a set constant.
[0079] The above sampling method determines the sampling parameters based on the performance indicators of clustering, achieving a more comprehensive sampling method and reducing the redundancy of the sampling data.
[0080] S33: Build and train an LSTM model, including:
[0081] The sample dataset and classification categories are encapsulated to construct a training set, establish a classification model, and determine the loss function. The classification model is then trained using the training set, and the change in the loss function value with the number of iterations is observed until the classification accuracy requirements are met.
[0082] S34: Based on the trained LSTM model, the input data is classified to obtain the classification results of the motor cooling control system status. The classification results include normal, water pump performance degradation, water pump failure, pipe blockage, and pipe leakage. Among them, water pump failure, pipe blockage, and pipe leakage are respectively divided into three levels of severity: minor, moderate, and severe.
[0083] S4: Motor cooling control system status judgment; if the current motor cooling control system water pump performance is degraded, but there is no water pump failure, pipe blockage or pipe leakage, then proceed to step S5; otherwise proceed to S6.
[0084] S5: Calculate the pump degradation degree and make adaptive adjustments;
[0085] S51: Calculate the pump degradation degree, specifically including the following steps:
[0086] S511: Obtain the time series of water pump current, pressure, and flow rate;
[0087] S512: A time series prediction of water pump current, pressure, and flow rate based on an autoregressive integral moving average model with periodic correlation is performed. Since the temperature change of the motor is periodic, the control parameters and flow rate of the cooling system's water pump also change periodically. Therefore, this application uses an autoregressive integral moving average model with periodic correlation to predict the time series of parameters. The model is as follows:
[0088] In the formula: y t ε is the time series value at time t; t Let c be the white noise sequence value at time t; c is the constant term; B is the delay operator; B s To make y t Shifting backwards s times in time, i.e., B s y t =y t -s, and θ q (B) represents the p-order autoregressive polynomial and the q-order moving average polynomial, respectively; φ P (B s ) and m Q (B s The numbers (p, d, q, s, P, D, Q) represent the P-order periodic autoregressive polynomial and the Q-order periodic moving average polynomial, respectively. The model contains seven parameters: p, d, q, s, P, D, and Q. The process of determining these parameters is called model order determination. p, d, and q represent the orders of the autoregressive, differencing, and moving average polynomials, respectively, while P, D, and Q represent the orders of the periodic autoregressive, differencing, and moving average polynomials with a period of s.
[0089] The model building and prediction process includes: 1) Data processing: Based on the time series plot of the original data, observe whether the sequence is stationary. If it is non-stationary, perform a difference transformation to make it stationary; 2) Model parameter identification: Identify the model type through the autocorrelation coefficient and preliminarily determine the order values of p, d, q and P, D, Q; 3) Model identification and model diagnosis: Select the most suitable model through the Akaike information criterion; 4) Model validation: Validate the optimal model using the validation set sample data; 5) Model prediction: Use the optimal model to predict the time series of pump current, pressure, and flow.
[0090] S513: Calculate the pump degradation degree based on the time series prediction results;
[0091] Optionally, the pump degradation degree can be calculated based on a supervised learning classifier.
[0092] S52: When the pump degradation reaches a certain threshold, the pump control parameters are calculated based on the adaptive control algorithm and sent to the pump controller to achieve adaptive adjustment control. The specific steps are as follows:
[0093] S521: Constructing an Adaptive Adjustment Model
[0094] The model is constructed using an FNN network, which consists of five layers: an input layer, a fuzzification layer, a rule inference layer, a synthesis layer, a defuzzification layer, and an output layer. The network membership function is expressed as follows:
[0095] Where m is the center point, a is the width, and b is the slope.
[0096] S522: Obtain the pump degradation degree as an input parameter for the adaptive adjustment model;
[0097] S523: Input the degradation prediction results into the adaptive adjustment model to obtain the pump motor control parameters output by the adaptive adjustment model, including the modulation ratio, modulation degree, duty cycle, phase offset, and overmodulation of SVPWM.
[0098] S6: Predict system failure probability based on a competing failure model and provide early warning; including the following steps:
[0099] S61: Calculate the system reliability R(t) = R1(t) + R2(t)
[0100] Where R1(t) is the system reliability under time t with only pump degradation failure and no sudden failure factors, and R2(t) is the system reliability under time t with sudden failure factors.
[0101] S62: Calculate the predicted system failure probability:
[0102] Where f(t) is the system density function, expressed as follows:
[0103] In this embodiment, early warning of faults such as pipe blockage, pipe leakage, pump degradation, and pump failure is achieved, and adaptive compensation function is provided. When the performance of some components degrades, control parameters can be automatically adjusted to maintain the system's optimal operating state and improve overall reliability. To ensure the accuracy of the prediction results, this application uses independent component analysis to perform blind source separation of the pump vibration signal and the pipe vibration signal, and uses a clustering algorithm to perform cluster sampling of the sample data. Combined with mixed sampling, this ensures the comprehensiveness and accuracy of the training sample data and reduces sample redundancy. Furthermore, in the prediction of pump degradation, a periodically correlated autoregressive integral moving average model is used to consider the periodic changes in motor temperature, improving the accuracy of the prediction.
[0104] The embodiments described above are merely preferred embodiments of this application and are not intended to limit the scope of this application. Any modifications and improvements made by those skilled in the art to the technical solutions of this application without departing from the spirit of this application should fall within the protection scope defined by the claims of this application.
Claims
1. A motor cooling control system with fault diagnosis function, characterized in that, include: Water pump, water pump controller, pipeline, several vibration sensors, flow sensors, differential pressure sensors, temperature sensors, and fault diagnosis module; The vibration sensor is installed on the water pump and the pipeline; the flow sensor and the differential pressure sensor are installed in the pipeline; the temperature sensor is installed on the cooled motor; the water pump controller controls the water pump; The fault diagnosis module predicts faults in the motor cooling control system through the following steps: S1: Acquiring signals, including: acquiring the current, speed and vibration signals of the water pump, the temperature signal of the motor, and the vibration signal, flow rate and pressure difference of the pipeline; S2: Process the acquired signals and extract features; S3: Classify the state of the motor cooling control system based on the K-means clustering algorithm combined with the LSTM model; S4: Determine the status of the motor cooling control system; if the current performance of the water pump in the motor cooling control system is degraded, but there is no water pump failure, pipe blockage or pipe leakage, proceed to step S5; otherwise, proceed to step S6. S5: Calculate the degradation degree of the water pump and make adaptive adjustments; and S6: Predict the failure probability of the motor cooling control system based on the competitive failure model and issue an early warning.
2. The motor cooling control system with fault diagnosis function according to claim 1, characterized in that, The processing of the acquired signals includes performing blind source separation on the vibration signals of the water pump and the pipeline based on the independent component analysis method.
3. The motor cooling control system with fault diagnosis function according to claim 1, characterized in that, The classification results of the motor cooling control system status include normal, water pump performance degradation, water pump failure, pipe blockage, and pipe leakage; among them, water pump failure, pipe blockage, and pipe leakage are respectively divided into three levels of severity: minor, moderate, and severe.
4. The motor cooling control system with fault diagnosis function according to claim 3, characterized in that, Step S3 includes: S31: Cluster the sample data based on the K-means clustering algorithm; S32: Determine training and validation samples based on the clustering results; S33: Build and train the LSTM model; and S34: Classify the input data based on the trained LSTM model to obtain the classification result of the state of the motor cooling control system.
5. The motor cooling control system with fault diagnosis function according to claim 4, characterized in that, The step of determining training and validation samples based on clustering results includes performing mixed sampling on the clustering results to determine the training and validation samples, wherein the mixed sampling includes stratified sampling and random sampling; The sampling result is obtained from N clustered and discrete samples through stratified sampling, as expressed below: Among them, D i Let r be the sample set of the i-th cluster center among the N cluster classification samples. i L represents the sampling ratio, and L is the discrete sample set.
6. The motor cooling control system with fault diagnosis function according to claim 5, characterized in that, r i Determined in the following manner: Where, k i DB is the adjustment coefficient. i Let q0 be the Davis-Bolding index of the i-th cluster, q0 be the set Davis-Bolding index threshold, and r0 be a set constant.
7. The motor cooling control system with fault diagnosis function according to claim 1, characterized in that, Step S5 includes: S51: Calculate the degree of degradation of the water pump; and S52: When the degradation degree of the water pump reaches a certain threshold, the control parameters of the water pump are calculated based on the adaptive control algorithm, and the control parameters of the water pump are sent to the water pump controller to realize the adaptive adjustment control.
8. The motor cooling control system with fault diagnosis function according to claim 7, characterized in that, The calculation of the degradation degree of the water pump includes: The time series of the current, pressure, and flow rate of the water pump are obtained; the time series of the current, pressure, and flow rate of the water pump are predicted based on the autoregressive integral moving average model of periodic correlation; and the degradation degree of the water pump is calculated based on the prediction results of the time series.
9. The motor cooling control system with fault diagnosis function according to claim 7, characterized in that, The control of the adaptive adjustment includes: An adaptive adjustment model is constructed based on an FNN network to obtain the degradation degree of the water pump, which is used as the input parameter of the adaptive adjustment model. The degradation degree is then input into the adaptive adjustment model to obtain the control parameters of the water pump output by the adaptive adjustment model.
10. The motor cooling control system with fault diagnosis function according to claim 1, characterized in that, Step S6 includes: S61: Calculate the reliability R(t) of the motor cooling control system: R(t) = R1(t) + R2(t) Wherein, R1(t) is the reliability of the motor cooling control system under time t when only the water pump deteriorates and fails without any sudden failure factors, and R2(t) is the reliability of the motor cooling control system under time t when sudden failure factors are present. S62: Calculate and predict the failure probability of the motor cooling control system: Where f(t) is the system density function, expressed as follows: