Controller over-current fault prediction method and system

Abstract

The invention relates to the technical field of controller fault detection, in particular to a controller overcurrent fault prediction method and system, and the method comprises the steps: collecting the current data of a controller through a current sensor, obtaining various load parameters of the controller, and collecting the environment temperature data in the controller through a temperature sensor; denoising the collected current data, and normalizing the current data, the load parameters and the environment temperature data; constructing a current change feature vector, analyzing the correlation between the load parameter and the current, analyzing the coupling relationship between the environment temperature and the current, and extracting a temperature-current coupling feature; and constructing a controller over-current fault prediction model, inputting data acquired and processed in real time into the model, and predicting the time from the current controller state to the occurrence of the over-current fault. According to the method, the overcurrent fault data features are constructed, and the overcurrent fault prediction model is combined to realize accurate prediction of the overcurrent fault of the controller, so that the prediction reliability is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of controller fault detection, and particularly to a method and system for predicting overcurrent faults of a controller. Background Art

[0002] In various electronic control systems, the controller plays a core role. The overcurrent fault is one of the more common and harmful fault types of the controller. Once it occurs, it may cause the controller components to burn out, the system to shut down, and even lead to more serious safety accidents. The current fault detection methods are mostly for post-diagnosis, and it is difficult to detect potential hazards in time before the overcurrent fault occurs. Although some prediction methods consider multiple types of faults, the prediction accuracy for overcurrent faults is insufficient, and the data characteristics and internal relationships closely related to overcurrent faults are not fully explored. Therefore, there is an urgent practical need to develop a precise and efficient prediction method for the overcurrent faults of the controller. Summary of the Invention

[0003] The present invention aims at the above-mentioned existing technical deficiencies and provides a method and system for predicting overcurrent faults of a controller to achieve comprehensive and precise prediction of overcurrent faults of the controller.

[0004] The present invention is achieved through the following technical solutions:

[0005] A method for predicting overcurrent faults of a controller is provided, and the method includes the following steps:

[0006] Step S10: Collect the current data of the main circuit and each branch circuit of the controller in real time through a current sensor, obtain various parameters of the controller load, and monitor the ambient temperature inside the controller in real time through a temperature sensor;

[0007] Step S20: Use the Kalman filter algorithm to denoise the collected current data, normalize the current data, load parameter data, and ambient temperature data. For the current data, use the zero-mean normalization method to adjust its mean to 0 and the standard deviation to 1. For the load parameter data and ambient temperature data, use the min-max normalization method to map them to the [0, 1] interval for subsequent unified analysis;

[0008] Step S30: Analyze the time-domain characteristics of the current data, construct a current change feature vector, use the Pearson correlation coefficient to analyze the correlation between the load parameters and the current, analyze the coupling relationship between the ambient temperature and the current, and extract the temperature-current coupling feature;

[0009] Step S40: Construct and train a controller overcurrent fault prediction model. Input the data collected and processed in real time into the optimized model. The model outputs the predicted time until an overcurrent fault occurs for the current controller state. When the predicted time is less than the set warning threshold, immediately issue an overcurrent fault warning signal.

[0010] Preferably, step S10 includes:

[0011] Current data acquisition: Use a high-precision current sensor to collect the current data of the main circuit and each branch circuit of the controller in real time. The sampling frequency needs to be reasonably set according to the working frequency of the controller and the current change characteristics, and is set between 1 kHz and 10 kHz to ensure that the subtle fluctuations of the current can be captured.

[0012] Load parameter acquisition: Collect various parameters of the load connected to the controller, such as load resistance, inductance, capacitance, etc. For variable loads, their parameter changes need to be monitored in real time, and these parameters are obtained through intelligent sensors or communication with the load control system.

[0013] Ambient temperature acquisition: Arrange temperature sensors at key positions inside and around the controller to monitor the ambient temperature in real time. The change in ambient temperature will affect the resistance value of the controller components, thereby affecting the current and having a certain correlation with the occurrence of overcurrent faults.

[0014] Preferably, in step S20, the Kalman filtering algorithm is used to denoise the collected current data, including:

[0015] Initialization: Initialize the parameters of the Kalman filter, determine the state vector, including information such as the current value of the current and the rate of change of the current at the current moment, set the initial estimated value of the state vector, and these initial values can be determined based on experience or preliminary analysis of the data. Initialize the covariance matrix, which is used to describe the correlation between the elements of the state vector and the uncertainty of the estimate.

[0016] Prediction stage: Construct a state equation and calculate the state prediction value at the current moment based on the state estimate value at the previous moment. Assume the state equation is X k = AX k-1 + BU k-1 + W k-1 , where X k is the state vector at the current moment, A is the state transition matrix, which describes the change relationship of the state over time, B is the control input matrix (if there is an external control input), U k-1 is the control input at the previous moment, W k-1 is the process noise, which follows a Gaussian distribution. Through this equation, the state vector at the current moment can be predicted where Represents the predicted value of the current - moment state based on the previous - moment information, and calculates the covariance matrix P of the predicted state k|k-1 = AP k-1|k-1 A T + Q, where P k|k-1 is the covariance matrix of the predicted state, P k-1|k-1 is the covariance matrix of the previous - moment state estimate, Q is the covariance matrix of the process noise, which reflects the intensity of the process noise;

[0017] Update stage: According to the real - time current data Z k at the current moment, it is necessary to update the predicted state using the observation equation to obtain the updated state estimate value at the current moment; the observation equation is usually expressed as Z k = HX k + V k , where H is the observation matrix, which describes the relationship between the state vector and the observation data, V k is the observation noise, which also follows a Gaussian distribution, and calculates the Kalman gain K k = P k|k-1 H T (HP k|k-1 H T + R) -1 , where K k is the Kalman gain, which determines the degree of correction of the observation data to the state estimate, R is the covariance matrix of the observation noise, and uses the Kalman gain to update the predicted state to obtain the state estimate value at the current moment, as shown in Equation (1):

[0018]

[0019] where represents the updated state estimate value at the current moment, and updates the covariance matrix P of the state estimate k|k = (I - K k H)P k|k-1 , where I is the identity matrix, and through this formula, the covariance matrix can be adjusted to more accurately reflect the uncertainty of the current state estimate;

[0020] Loop iteration: Take the updated state estimate value and covariance matrix at the current moment as the initial values for the next iteration, repeat the above - mentioned prediction and update steps to process the current - moment current data. As the iteration progresses, the Kalman filter will continuously adjust the state estimate according to the new observation data, gradually approaching the true value of the current signal, thereby effectively removing high - frequency noise interference and retaining the true change trend of the current signal.

[0021] Preferably, in step S20, the step of using the zero-mean normalization method for the current data to adjust its mean to 0 and standard deviation to 1 includes:

[0022] Calculate the mean: Calculate the mean of the controller current data collected.

[0023] Calculate the standard deviation: Calculate the standard deviation of the controller current data collected.

[0024] Normalization processing: Use the zero-mean normalization formula to transform each current data, making the mean of all current data become 0 and the standard deviation become 1. This makes current data with different magnitudes and distributions have a unified scale, facilitating subsequent analysis and comparison, and eliminating the influence of data magnitude differences on the analysis results.

[0025] Preferably, in step S20, the step of using the min-max normalization method for the load parameter data and environmental temperature data to map them to the [0, 1] interval includes:

[0026] Determine the maximum and minimum values of the data: Traverse the collected load parameter data and environmental temperature data to determine the maximum and minimum values in the two sets of data.

[0027] Normalization transformation: After determining the maximum and minimum values, perform normalization calculations on each data according to the min-max normalization formula to map the load parameter data and environmental temperature data to the [0, 1] interval, enabling different types of data to be analyzed on the same scale, facilitating subsequent model training and data processing, and improving the accuracy and efficiency of the analysis.

[0028] Preferably, in step S30, the step of analyzing the time-domain characteristics of the current data and constructing a current change feature vector includes:

[0029] Determine the time window and select data: According to the current characteristics, select a time window from 100 ms to 500 ms, and intercept data segments from the collected current data according to the selected window length.

[0030] Calculate various time-domain characteristic values: According to the current data within the selected time window, calculate the average value of the current, the effective value of the current, and the change rate of the current, and determine the maximum and minimum values of the current. Take these as the time-domain characteristic values of the current data.

[0031] Construct a current change feature vector: Sort the maximum and minimum values, average value, effective value, and change rate of the current obtained in the step of calculating various time-domain characteristic values in the set order to construct a current change feature vector.

[0032] Preferably, in step S30, the Pearson correlation coefficient is used to analyze the correlation between the load parameters and the current, and the coupling relationship between the ambient temperature and the current. The steps of extracting the temperature-current coupling characteristics include:

[0033] Load parameter collection: Collect load parameter and current data at corresponding moments within a period of time, including load resistance, inductance, and capacitance, ensuring that the timestamps of the data are precisely matched and the number of data points is sufficient for analysis;

[0034] Calculate the Pearson correlation coefficient: For each load parameter and current data, use the Pearson correlation coefficient formula to calculate the correlation. The calculation formula is shown in Equation (2):

[0035]

[0036] where r is the Pearson correlation coefficient, x i is the load parameter data, which can be load resistance, inductance, and capacitance, and y i is the current data corresponding to the load parameter data, and are their respective means, and n is the number of data points. When calculating the correlation coefficient between the load resistance and the current, substitute the corresponding data into the formula;

[0037] Correlation judgment and sorting: According to the calculation results of the Pearson correlation coefficient, judge the strength and sign of the correlation. The closer the absolute value of the Pearson correlation coefficient r is to 1, the stronger the correlation. r > 0 indicates a positive correlation, r < 0 indicates a negative correlation, and r = 0 indicates no correlation. Sort the absolute values of the correlation coefficients between all load parameters and the current. The order of sorting represents the degree of influence on the current. The load parameter with a larger absolute value is the parameter that significantly affects the current. For example, if the absolute value of the correlation coefficient of the load resistance is 0.85 and the absolute value of the correlation coefficient of the inductance is 0.3, then the load resistance has a more significant impact on the current. For the load parameters with significant influence, focus on tracking their changes in subsequent analysis, establish a dedicated data recording and analysis process, and monitor the changes of these parameters in real-time or at regular intervals to analyze their influence mechanism and trend on the current;

[0038] Temperature - current coupling feature extraction: According to the variation range of the ambient temperature and actual requirements, divide the temperature intervals. For example, with an interval of 5°C, divide the temperature range into intervals such as [15°C - 20°C], [20°C - 25°C], etc. For each temperature interval, screen out the current data within the corresponding temperature range from the collected temperature and current data. For the current data within each temperature interval, calculate its variation trend. A linear regression method can be used to fit the curve of current changing with time, and obtain the slope of the curve as a quantitative index of the current variation trend. When the slope is positive, it indicates that the current generally shows an upward trend within this temperature interval; when the slope is negative, it shows a downward trend. Taking the temperature intervals as rows and the current data features as columns, construct a temperature - current joint feature matrix to obtain the temperature - current coupling feature.

[0039] Preferably, the steps of constructing and training the controller over - current fault prediction model in step S40 include:

[0040] Construction and training of the controller over - current fault prediction model: Use the support vector machine SVM regression model for over - current fault prediction. Take the current change feature vector, load - current correlation feature, and temperature - current coupling feature after pre - processing and feature extraction as inputs, and use the data before the occurrence of historical over - current faults as training samples to train the SVM regression model so that it can learn the feature change law from the normal state to the over - current fault state;

[0041] Optimization of the controller over - current fault prediction model: Use the particle swarm optimization PSO algorithm to optimize the parameters of the SVM regression model. The PSO algorithm searches for the optimal SVM model parameters, such as the penalty factor C and the kernel function parameter γ, in the parameter space by simulating the foraging behavior of bird flocks to improve the prediction accuracy of the model;

[0042] Real - time prediction and warning: Real - time collect the current data in the controller circuit and perform pre - processing, and input it into the optimized SVM regression model. The model outputs the predicted time from the current state to the occurrence of an over - current fault. When the predicted time of an over - current fault is less than the set warning threshold, for example, the warning threshold is set to 5 minutes, immediately send out a controller over - current fault warning signal to remind the maintenance personnel to take corresponding measures, such as adjusting the load, checking the circuit connection, etc.

[0043] In addition, to achieve the above object, the present invention also proposes a controller over - current fault prediction system, and the controller over - current fault prediction system includes:

[0044] Controller data acquisition module: Used to collect the current data of the main circuit and each branch circuit of the controller in real - time through current sensors, obtain the parameters of each item of the controller load, and monitor the internal ambient temperature of the controller in real - time through temperature sensors;

[0045] Controller Data Denoising and Normalization Module: It is used to denoise the collected current data by using the Kalman filtering algorithm, normalize the current data, load parameter data, and ambient temperature data. For the current data, the zero-mean normalization method is adopted to adjust its mean to 0 and the standard deviation to 1. For the load parameter data and ambient temperature data, the min-max normalization method is adopted to map them to the interval [0, 1] for subsequent unified analysis.

[0046] Feature Extraction and Correlation Analysis Module: It is used to analyze the time-domain features of the current data, construct a current change feature vector, analyze the correlation between the load parameters and the current by using the Pearson correlation coefficient, analyze the coupling relationship between the ambient temperature and the current, and extract the temperature-current coupling features;

[0047] Controller Overcurrent Fault Prediction Model Construction and Overcurrent Fault Prediction Module: It is used to construct and train a controller overcurrent fault prediction model, input the real-time collected and processed data into the optimized model, and the model outputs the predicted time from the current controller state to the occurrence of the overcurrent fault. When the predicted time is less than the set warning threshold, an overcurrent fault warning signal is immediately issued.

[0048] In addition, to achieve the above object, the present invention also proposes a controller overcurrent fault prediction device, which includes: a memory, a processor, and programs such as a controller overcurrent fault prediction algorithm stored on the memory and executable on the processor. The programs such as the controller overcurrent fault prediction algorithm implement the steps of a controller overcurrent fault prediction method as described above.

[0049] In addition, to achieve the above object, the present invention also provides a computer program product, which includes programs such as a controller overcurrent fault prediction algorithm. When the programs such as the controller overcurrent fault prediction algorithm are executed by a processor, they implement a controller overcurrent fault prediction method as described above.

[0050] The advantages and effects of the present invention are:

[0051] A controller overcurrent fault prediction method and system proposed by the present invention, by deeply excavating the data features and internal relationships closely related to the overcurrent fault, combined with the artificial intelligence model algorithm, realizes high-precision prediction of the controller overcurrent fault, greatly improves the accuracy and reliability of the prediction, can issue a warning in advance before the overcurrent fault occurs, and buys sufficient time for maintenance personnel to take measures, effectively avoiding controller damage and system downtime caused by the overcurrent fault, reducing production losses and maintenance costs; in addition, compared with the general fault prediction method, the present invention is specifically designed for the overcurrent fault, and the prediction of the overcurrent fault is more targeted, and can better meet the needs of controller overcurrent fault prevention in practical applications. Brief Description of the Drawings

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the 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 drawings can be obtained based on these drawings.

[0053] Figure 1 It is a flowchart of a method for predicting overcurrent faults of a controller according to the present invention.

[0054] Figure 2 It is a schematic structural diagram of a system for predicting overcurrent faults of a controller according to the present invention.

[0055] Figure 3 It is a schematic block diagram of the structure of an electronic device for predicting overcurrent faults of a controller according to the present invention. Detailed implementation manners

[0056] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0057] As Figure 1 shown, in an embodiment of the present invention, a method for predicting overcurrent faults of a controller includes the following steps:

[0058] Step S10: Real-time collect the current data of the main circuit and each branch circuit of the controller through a current sensor, obtain various parameters of the controller load, and real-time monitor the ambient temperature inside the controller through a temperature sensor.

[0059] Specifically, step S10 includes:

[0060] Current data collection: Use a high-precision current sensor to real-time collect the current data of the main circuit and each branch circuit of the controller. The sampling frequency needs to be reasonably set according to the working frequency of the controller and the current change characteristics, and is set between 1 kHz and 10 kHz to ensure that the subtle fluctuations of the current can be captured;

[0061] Load parameter collection: Collect various parameters of the load connected to the controller, such as load resistance, inductance, capacitance, etc. For variable loads, it is necessary to real-time monitor the parameter changes, and obtain these parameters through intelligent sensors or communicate with the load control system;

[0062] Environmental temperature acquisition: Temperature sensors are arranged at key positions inside and around the controller to monitor the environmental temperature in real time. The change in environmental temperature will affect the resistance value of the controller components, and then affect the current, which is somewhat related to the occurrence of overcurrent faults.

[0063] Step S20: Use the Kalman filter algorithm to denoise the collected current data, normalize the current data, load parameter data, and environmental temperature data. For the current data, use the zero-mean normalization method to adjust its mean to 0 and standard deviation to 1. For the load parameter data and environmental temperature data, use the min-max normalization method to map them to the [0,1] interval for subsequent unified analysis.

[0064] Specifically, in step S20, using the Kalman filter algorithm to denoise the collected current data includes:

[0065] Initialization: Initialize the parameters of the Kalman filter, determine the state vector, including information such as the current current value and the rate of change of the current, set the initial estimated value of the state vector, and these initial values can be determined based on experience or preliminary analysis of the data. Initialize the covariance matrix, which is used to describe the correlation between the elements of the state vector and the uncertainty of the estimate;

[0066] Prediction stage: Construct the state equation, calculate the state prediction value at the current moment based on the state estimate value at the previous moment. Assume the state equation is X k =AX k-1 +BU k-1 +W k-1 , where X k is the state vector at the current moment, A is the state transition matrix, which describes the relationship between the state and time, B is the control input matrix (if there is an external control input), U k-1 is the control input at the previous moment, W k-1 is the process noise, which follows a Gaussian distribution. Through this equation, the state vector at the current moment can be predicted where represents the predicted value of the current state based on the information at the previous moment, and calculate the covariance matrix P k|k-1 =AP k-1|k-1 A T +Q, where P k|k-1 is the covariance matrix of the predicted state, P k-1|k-1 is the covariance matrix of the state estimate at the previous moment, and Q is the covariance matrix of the process noise, which reflects the intensity of the process noise;

[0067] Update stage: According to the real-time current data Z at the current moment k, the observation equation needs to be used to update the predicted state to obtain the updated state estimate at the current moment; the observation equation is usually expressed as Z k =HX k +V k , where H is the observation matrix, which describes the relationship between the state vector and the observation data, and V k is the observation noise, which also follows a Gaussian distribution. Calculate the Kalman gain K k =P k|k-1 H T (HP k|k-1 H T +R) -1 , where K k is the Kalman gain, which determines the degree of correction of the observation data to the state estimate. R is the covariance matrix of the observation noise. Use the Kalman gain to update the predicted state to obtain the state estimate at the current moment, as shown in Equation (1):

[0068]

[0069] where represents the updated state estimate at the current moment. Update the covariance matrix P k|k =(I-K k H)P k|k-1 , where I is the identity matrix. Through this formula, the covariance matrix can be adjusted to more accurately reflect the uncertainty of the current state estimate;

[0070] Loop iteration: Use the updated state estimate and covariance matrix at the current moment as the initial values for the next iteration, repeat the above prediction and update steps, and process the current data at the next moment. As the iteration progresses, the Kalman filter will continuously adjust the state estimate according to the new observation data, gradually approaching the true value of the current signal, thereby effectively removing high-frequency noise interference and retaining the true change trend of the current signal.

[0071] Specifically, the steps of using the zero-mean normalization method for the current data in step S20 to adjust its mean to 0 and standard deviation to 1 include:

[0072] Calculate the mean: Calculate the mean of the collected controller current data;

[0073] Calculate the standard deviation: Calculate the standard deviation of the collected controller current data;

[0074] Normalization processing: Use the zero-mean normalization formula to transform each current data, making the mean of all current data become 0 and the standard deviation become 1. This makes current data of different magnitudes and distributions have a unified scale, facilitating subsequent analysis and comparison, and eliminating the influence of data magnitude differences on the analysis results.

[0075] Specifically, the steps of using the min-max normalization method for the load parameter data and environmental temperature data in step S20 and mapping them to the [0, 1] interval include:

[0076] Determine the maximum and minimum values of the data: Traverse the collected load parameter data and environmental temperature data to determine the maximum and minimum values in these two sets of data;

[0077] Normalization transformation: After determining the maximum and minimum values, perform normalization calculations on each data according to the min-max normalization formula, map the load parameter data and environmental temperature data to the [0, 1] interval, enable different types of data to be analyzed on the same scale, facilitate subsequent model training and data processing, and improve the accuracy and efficiency of the analysis.

[0078] Step S30: Analyze the time-domain characteristics of the current data, construct a current change feature vector, use the Pearson correlation coefficient to analyze the correlation between the load parameters and the current, analyze the coupling relationship between the environmental temperature and the current, and extract the temperature-current coupling features.

[0079] Specifically, the steps of analyzing the time-domain characteristics of the current data and constructing a current change feature vector in step S30 include:

[0080] Determine the time window and select the data: According to the current characteristics, select a time window from 100 ms to 500 ms, and intercept a data segment from the collected current data according to the selected window length. For example, when the acquisition frequency is 10 kHz, a 100-ms time window contains 1000 data points;

[0081] Calculate each time-domain eigenvalue: According to the current data within the selected time window, calculate the average value of the current, the effective value of the current, and the change rate of the current, and determine the maximum and minimum values of the current, and use these as the time-domain eigenvalues of the current data;

[0082] Construct a current change feature vector: Sort the maximum and minimum values of the current, the average value, the effective value, and the change rate obtained in the step of calculating each time-domain eigenvalue in a set order to construct a current change feature vector. For example, arrange them in the order of the current mean value, the maximum and minimum values of the current, the effective value of the current, and the change rate of the current to obtain a feature vector of [current mean value, maximum and minimum values of the current, effective value of the current, change rate of the current], which represents the current change characteristics within this time window and provides key data for the subsequent overcurrent fault prediction model.

[0083] Specifically, in step S30, the Pearson correlation coefficient is used to analyze the correlation between the load parameters and the current, and the coupling relationship between the ambient temperature and the current is analyzed. The steps for extracting the temperature-current coupling characteristics include:

[0084] Load parameter collection: Collect the load parameters and the current data at corresponding moments within a period of time, including the load resistance, inductance, and capacitance, ensuring that the timestamps of the data are precisely matched and the number of data points is sufficient for analysis. For example, collect the load resistance value and the current value simultaneously every 10 milliseconds for 10 minutes, obtaining 6000 sets of data;

[0085] Calculate the Pearson correlation coefficient: For each load parameter and the current data, use the Pearson correlation coefficient formula to calculate the correlation. The calculation formula is shown in Equation (2):

[0086]

[0087] where r is the Pearson correlation coefficient, x i is the load parameter data, which can be the load resistance, inductance, and capacitance, and y i is the current data corresponding to the load parameter data. and are their respective means, and n is the number of data points. When calculating the correlation coefficient between the load resistance and the current, substitute the corresponding data into the formula;

[0088] Correlation judgment and sorting: According to the calculation results of the Pearson correlation coefficient, judge the strength and sign of the correlation. The closer the absolute value of the Pearson correlation coefficient r is to 1, the stronger the correlation. r > 0 indicates a positive correlation, r < 0 indicates a negative correlation, and r = 0 indicates no correlation. Sort the absolute values of the correlation coefficients between all load parameters and the current. The order of sorting represents the degree of influence on the current. The load parameter with a larger absolute value is the parameter that has a significant influence on the current. For example, if the absolute value of the correlation coefficient of the load resistance is 0.85 and the absolute value of the correlation coefficient of the inductance is 0.3, then the load resistance has a more significant influence on the current. For the load parameters with significant influence, focus on tracking their changes in subsequent analysis, establish a dedicated data recording and analysis process, and monitor the changes of these parameters in real-time or at regular intervals to analyze their influence mechanism and trend on the current;

[0089] Temperature - Current Coupling Feature Extraction: According to the variation range of ambient temperature and actual requirements, divide the temperature range into intervals. For example, with an interval of 5°C, the temperature range can be divided into intervals such as [15°C - 20°C], [20°C - 25°C], etc. For each temperature interval, screen out the current data within the corresponding temperature range from the collected temperature and current data. For example, in the interval [15°C - 20°C], select the current values at all time points within this temperature range. For the current data within each temperature interval, calculate its change trend. A linear regression method can be used to fit the curve of current change over time, and the slope of the curve is used as a quantitative indicator of the current change trend. When the slope is positive, it indicates that the current generally shows an upward trend within this temperature interval; when the slope is negative, it shows a downward trend. Taking the temperature interval as rows and the current data features as columns, construct a temperature - current joint feature matrix to obtain the temperature - current coupling feature. For example, the first row corresponds to the interval [15°C - 20°C], containing data such as the current change trend value, average current value, and current fluctuation amplitude within this interval, and the second row corresponds to the interval [20°C - 25°C], and so on. This matrix comprehensively shows the characteristics of the current under different temperature intervals, providing an intuitive data structure for subsequent analysis of the coupling relationship between temperature and current.

[0090] Step S40: Construct and train a controller over - current fault prediction model. Input the real - time collected and processed data into the optimized model. The model outputs the predicted time from the current controller state to the occurrence of an over - current fault. When the predicted time is less than the set warning threshold (such as 5 minutes), immediately send an over - current fault warning signal.

[0091] Specifically, the steps of constructing and training the controller over - current fault prediction model in step S40 include:

[0092] Construction and Training of the Controller Over - Current Fault Prediction Model: Use a support vector machine (SVM) regression model for over - current fault prediction. Take the current change feature vector after pre - processing and feature extraction, the load - current correlation feature, and the temperature - current coupling feature as inputs, and use the data before the occurrence of historical over - current faults as training samples to train the SVM regression model so that it can learn the feature change law from the normal state to the over - current fault state;

[0093] Optimization of the Controller Over - Current Fault Prediction Model: Use the particle swarm optimization (PSO) algorithm to optimize the parameters of the SVM regression model. The PSO algorithm searches for the optimal SVM model parameters, such as the penalty factor C and the kernel function parameter γ, in the parameter space by simulating the foraging behavior of bird flocks to improve the prediction accuracy of the model;

[0094] Real-time prediction and early warning: The current data in the controller circuit is collected in real time and preprocessed, and then input into the optimized SVM regression model. The model outputs the predicted time from the current state to the occurrence of overcurrent fault. When the predicted time of overcurrent fault is less than the set early warning threshold, such as the early warning threshold is set to 5 minutes, an overcurrent fault early warning signal of the controller is immediately sent out to remind the maintenance personnel to take corresponding measures, such as adjusting the load, checking the circuit connection, etc.

[0095] In addition, as Figure 2 shown, in an embodiment of the present invention, a controller overcurrent fault prediction system is proposed. The controller overcurrent fault prediction system includes:

[0096] Controller data acquisition module: It is used to collect the current data of the main circuit and each branch circuit of the controller in real time through a current sensor, obtain various parameters of the controller load, and monitor the internal environment temperature of the controller in real time through a temperature sensor;

[0097] Controller data denoising and normalization module: It is used to denoise the collected current data by using the Kalman filter algorithm, normalize the current data, load parameter data and environment temperature data. For the current data, the zero-mean normalization method is adopted to adjust its mean to 0 and the standard deviation to 1. For the load parameter data and environment temperature data, the minimum-maximum normalization method is adopted to map them to the [0,1] interval for subsequent unified analysis.

[0098] Feature extraction and correlation analysis module: It is used to analyze the time-domain features of the current data, construct a current change feature vector, analyze the correlation between the load parameters and the current by using the Pearson correlation coefficient, analyze the coupling relationship between the environment temperature and the current, and extract the temperature-current coupling features;

[0099] Controller overcurrent fault prediction model construction and overcurrent fault prediction module: It is used to construct and train a controller overcurrent fault prediction model, input the data collected and processed in real time into the optimized model, and the model outputs the predicted time from the current controller state to the occurrence of overcurrent fault. When the predicted time is less than the set early warning threshold (such as 5 minutes), an overcurrent fault early warning signal is immediately sent out.

[0100] The controller overcurrent fault prediction system provided by this application adopts the controller overcurrent fault prediction method in the above embodiment, and can solve the technical problems of low accuracy and weak pertinence of the traditional controller overcurrent fault prediction method. Compared with the prior art, the beneficial effects of the controller overcurrent fault prediction system provided by this application are the same as those of the controller overcurrent fault prediction method provided by the above embodiment, and the other technical features in the controller overcurrent fault prediction system are the same as the features disclosed in the above embodiment method, and will not be elaborated here.

[0101] The present application provides a controller overcurrent fault prediction device, and the controller overcurrent fault prediction device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute a controller overcurrent fault prediction method in Embodiment 1 above.

[0102] As Figure 3 shown, in an embodiment of the present invention, a structural schematic diagram of a controller overcurrent fault prediction device suitable for implementing the embodiments of the present application is shown. A controller overcurrent fault prediction device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions: tablet computers), PMPs (Portable Media Players), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The shown controller overcurrent fault prediction device is only an example and should not impose any limitation on the functions and usage scope of the embodiments of the present application.

[0103] Figure 3A controller overcurrent fault prediction device shown can include a processing system 1001 (such as 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 (ROM: Read Only Memory) 1002 or a program loaded from a storage system 1003 into a random access memory (RAM: Random Access Memory) 1004. In the RAM 1004, various programs and data required for the operation of a controller overcurrent fault prediction device are also stored. The processing system 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input system 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output system 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage system 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication system 1009. The communication system 1009 can allow a controller overcurrent fault prediction device to communicate with other devices wirelessly or wiredly to exchange data. Although a controller overcurrent fault prediction device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the shown systems. More or fewer systems can be alternatively implemented or had.

[0104] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program codes for performing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network through the communication system, or installed from the storage system 1003, or installed from the ROM 1002. When the computer program is executed by the processing system 1001, the above-mentioned functions defined in the methods of the embodiments disclosed in the present application are executed.

[0105] A controller overcurrent fault prediction device provided by the present application adopts a controller overcurrent fault prediction method in the above-mentioned embodiment, and can solve the technical problems of low accuracy and weak pertinence of traditional controller overcurrent fault prediction methods. Compared with the prior art, the beneficial effects of a controller overcurrent fault prediction device provided by the present application are the same as those of a controller overcurrent fault prediction method provided by the above-mentioned embodiment, and other technical features in the controller overcurrent fault prediction device are the same as those disclosed in the method of the previous embodiment, and will not be elaborated here.

[0106] The various parts disclosed in this application can be implemented by hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in a suitable manner in any one or more embodiments or examples.

[0107] This application also provides a computer program product, including a computer program, which when executed by a processor, implements the steps of a method for predicting overcurrent faults of a controller as described above.

[0108] The computer program product provided by this application can solve the technical problems of low accuracy and weak pertinence in the traditional method for predicting overcurrent faults of a controller. Compared with the prior art, the beneficial effects of the computer program product provided by this application are the same as those of the method for predicting overcurrent faults of a controller provided in the above embodiments, and will not be elaborated herein.

[0109] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims

1. A controller overcurrent fault prediction method, characterized in that The method includes the following steps: Step S10: Collect the current data of the main circuit and each branch circuit of the controller in real time through a current sensor, obtain various parameters of the controller load, and monitor the ambient temperature inside the controller in real time through a temperature sensor; Step S20: Denoise the collected current data using the Kalman filter algorithm, normalize the current data, load parameter data, and ambient temperature data. Use the zero-mean normalization method for the current data, and the min-max normalization method for the load parameter data and ambient temperature data; Step S30: Analyze the time-domain characteristics of the current data, construct a current change feature vector, analyze the correlation between the load parameters and the current using the Pearson correlation coefficient, analyze the coupling relationship between the ambient temperature and the current, and extract the temperature-current coupling characteristics; Step S40: Construct and train an overcurrent fault prediction model for the controller, input the data collected and processed in real time into the optimized model, and the model outputs the predicted time from the current controller state to the occurrence of an overcurrent fault. When the predicted time is less than the set warning threshold, immediately issue an overcurrent fault warning signal.

2. The controller overcurrent fault prediction method according to claim 1, wherein In step S20, the Kalman filter algorithm is used to denoise the collected current data, which includes: Initialization: Initialize the parameters of the Kalman filter, determine the state vector, including the current value of the current and the change rate of the current, and set the initial estimated value of the state vector; Prediction stage: Construct a state equation and calculate the state prediction value at the current moment based on the state estimated value at the previous moment; Update stage: Update the predicted state using the observation equation according to the real-time current data at the current moment to obtain the updated state estimated value at the current moment; Loop iteration: Use the updated state estimated value and covariance matrix at the current moment as the initial values for the next iteration, and repeat the above prediction and update steps to process the current data at the next moment.

3. A controller overcurrent fault prediction method according to claim 1, characterized in that, In step S20, the zero-mean normalization method is used for the current data, and the steps to adjust its mean to 0 and standard deviation to 1 include: Calculate the mean: Calculate the mean of the collected controller current data; Calculate the standard deviation: Calculate the standard deviation of the collected controller current data; Normalization process: Use the zero-mean normalization formula to transform each current data, making the mean of all current data become 0 and the standard deviation become 1.

4. A controller overcurrent fault prediction method according to claim 1, characterized in that, In step S20, the min-max normalization method is used for the load parameter data and ambient temperature data, and the steps to map them to the [0,1] interval include: ​ ​ 5. A controller overcurrent fault prediction method according to claim 1, characterized in that, ​ Determine the time window and select data: According to the current characteristics, a time window from 100 ms to 500 ms is selected, and a data segment is intercepted from the collected current data according to the selected window length; Calculate various time-domain characteristic values: According to the current data within the selected time window, calculate the average value of the current, the effective value of the current, and the change rate of the current, and determine the maximum and minimum values of the current. These are used as the time-domain characteristic values of the current data; Construct a current change feature vector: Sort the maximum and minimum values of the current, the average value, the effective value, and the change rate obtained in the step of calculating various time-domain characteristic values in a set order to construct a current change feature vector.

6. A controller overcurrent fault prediction method according to claim 1, characterized in that, In step S30, the Pearson correlation coefficient is used to analyze the correlation between the load parameters and the current, and the coupling relationship between the ambient temperature and the current. The steps for extracting the temperature-current coupling characteristics include: Load parameter collection: Collect load parameters and current data at corresponding moments within a period of time, including load resistance, inductance, and capacitance; Calculate the Pearson correlation coefficient: For each load parameter and current data, use the Pearson correlation coefficient formula to calculate the correlation; Correlation judgment and sorting: According to the calculation results of the Pearson correlation coefficient, judge the strength and sign of the correlation. The closer the absolute value of the Pearson correlation coefficient r is to 1, the stronger the correlation. r > 0 is a positive correlation, r < 0 is a negative correlation, and r = 0 is uncorrelated. Sort the absolute values of the correlation coefficients between all load parameters and the current, and the order of sorting represents the degree of influence on the current; Temperature-current coupling characteristic extraction: According to the change range of the ambient temperature and actual requirements, divide the temperature intervals. For each temperature interval, screen out the current data within the corresponding temperature range from the collected temperature and current data. For the current data within each temperature interval, calculate its change trend. Use the temperature intervals as rows and the current data characteristics as columns to construct a temperature-current joint feature matrix to obtain the temperature-current coupling characteristics.

7. A controller overcurrent fault prediction method according to claim 1, characterized in that In step S40, the steps for constructing and training the controller overcurrent fault prediction model include: Construction and training of the controller overcurrent fault prediction model: Use a support vector machine SVM regression model for overcurrent fault prediction. Take the current change feature vector, the load-current association feature, and the temperature-current coupling feature after preprocessing and feature extraction as inputs, and use the data before the occurrence of historical overcurrent faults as training samples to train the SVM regression model so that it can learn the feature change law from the normal state to the overcurrent fault state; Optimization of the controller overcurrent fault prediction model: Use the particle swarm optimization PSO algorithm to optimize the parameters of the SVM regression model. The PSO algorithm searches for the optimal SVM model parameters in the parameter space by simulating the foraging behavior of bird flocks to improve the prediction accuracy of the model; Real-time prediction and early warning: Collect the current data in the controller circuit in real time and perform preprocessing, then input it into the optimized SVM regression model. The model outputs the predicted time from the current state to the occurrence of overcurrent fault. When the predicted time of overcurrent fault is less than the set early warning threshold, immediately send out an overcurrent fault early warning signal for the controller to remind the maintenance personnel to take corresponding measures.

8. A controller overcurrent fault prediction system, characterized in that, The described overcurrent fault prediction system for a controller includes: Controller data acquisition module: Used to collect the current data of the main circuit and each branch circuit of the controller in real time through a current sensor, obtain various parameters of the controller load, and monitor the internal environmental temperature of the controller in real time through a temperature sensor; Controller data denoising and normalization module: Used to denoise the collected current data using the Kalman filtering algorithm, normalize the current data, load parameter data, and environmental temperature data. For the current data, use the zero-mean normalization method to adjust its mean to 0 and standard deviation to 1. For the load parameter data and environmental temperature data, use the min-max normalization method to map them to the [0,1] interval; Feature extraction and correlation analysis module: Used to analyze the time-domain features of the current data, construct a current change feature vector, analyze the correlation between the load parameters and the current using the Pearson correlation coefficient, analyze the coupling relationship between the environmental temperature and the current, and extract the temperature-current coupling features; Controller overcurrent fault prediction model construction and overcurrent fault prediction module: Used to construct and train a controller overcurrent fault prediction model, input the real-time collected and processed data into the optimized model. The model outputs the predicted time from the current controller state to the occurrence of overcurrent fault. When the predicted time is less than the set early warning threshold, immediately send out an overcurrent fault early warning signal.

9. A controller overcurrent fault prediction device, characterized in that, The described overcurrent fault prediction device for a controller includes: A memory, a processor, and an overcurrent fault prediction program for a controller stored on the memory and executable on the processor. When the overcurrent fault prediction program for a controller is executed by the processor, it implements the overcurrent fault prediction method for a controller according to any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes an overcurrent fault prediction program for a controller. When the overcurrent fault prediction program for a controller is executed by the processor, it implements the overcurrent fault prediction method for a controller according to any one of claims 1 to 7.

Images