A method, system and device for detecting faults in an electromechanical brake system

Abstract

The application relates to the technical field of brake control, in particular to a fault detection method, a detection system and a device of an electronic mechanical brake system, the method comprising the following steps: obtaining multiple prediction data of a brake system in a prediction period; analyzing the difference between expected torque corresponding to each prediction data and prediction torque included in the prediction data, obtaining the motor and transmission mechanism abnormal probability of each prediction data, wherein the expected torque is obtained based on the ideal torque after correcting the temperature error, and the ideal torque is calculated based on the prediction current included in the corresponding prediction data; obtaining the signal abnormal index of each prediction data according to the frequency deviation and the duty cycle deviation corresponding to each prediction data; analyzing the motor and transmission mechanism abnormal probability and the signal abnormal index of each prediction data, obtaining the fault probability value of each prediction data; and analyzing the fault probability value of each prediction data, obtaining the fault detection result of the brake system in the prediction period. The application can obtain a more accurate fault detection result.

Description

Technical Field

[0001] This invention relates to the technical field of braking control, and specifically to a fault detection method, detection system and device for an electromechanical braking system. Background Technology

[0002] Electromechanical braking (EMB) is a fully electronically controlled, motor-driven braking system, belonging to the brake-by-wire technology. It eliminates traditional hydraulic or pneumatic brake lines, directly controlling the motor to perform braking actions via electrical signals. It offers advantages such as fast response, high integration, energy efficiency, and environmental friendliness, making it suitable for modern intelligent electric vehicles and autonomous vehicles.

[0003] By detecting faults in the EMB (Electromechanical Braking System), potential braking malfunctions can be identified and addressed promptly, ensuring the long-term reliable operation of the electromechanical braking system. However, in application, it has been found that the accuracy of existing fault detection solutions is poor, easily leading to false alarms. For example, under continuous high-intensity braking conditions, overheating of the braking components can temporarily cause braking failure or a significant decrease in braking performance, which may be misjudged as a fault. Similarly, in environments with strong electromagnetic interference, the transmission of electrical signals used for braking control may be disrupted, resulting in abnormal fluctuations and excessively high braking control delays, which can also be misjudged as faults.

[0004] In other words, existing technologies have poor accuracy in detecting faults in braking systems. Summary of the Invention

[0005] To address the problem of poor accuracy in fault detection of braking systems in existing technologies, the present invention aims to provide a fault detection method, detection system, and device for electromechanical braking systems. The specific technical solution adopted is as follows:

[0006] In a first aspect, one embodiment of the present invention provides a fault detection method for an electromechanical braking system, the method comprising:

[0007] Acquire multiple forecast data points for the braking system within the forecast time period;

[0008] The difference between the expected torque corresponding to each predicted data and the predicted torque included therein is analyzed to obtain the abnormal probability of the motor and transmission mechanism for each predicted data. The expected torque is obtained after correcting for temperature error based on the corresponding ideal torque, and the ideal torque is calculated based on the predicted current included in the corresponding predicted data.

[0009] Based on the frequency deviation and duty cycle deviation corresponding to each predicted data, the signal anomaly index of each predicted data is obtained, wherein the frequency deviation is used to represent the difference between the predicted frequency included in the corresponding predicted data and the corresponding rated frequency, and the duty cycle deviation is used to represent the difference between the predicted duty cycle included in the corresponding predicted data and the corresponding expected duty cycle.

[0010] Analyze the abnormal probability of motor and transmission mechanism and the abnormal signal index of each predicted data to obtain the fault probability value of each predicted data.

[0011] By analyzing the failure probability value of each predicted data point, the failure detection results of the braking system within the predicted time period are obtained.

[0012] In one embodiment, the step of obtaining the expected torque corresponding to each predicted data point includes:

[0013] The ideal torque corresponding to each prediction data point is determined by multiplying the motor torque constant corresponding to the braking system with the predicted current included in each prediction data point.

[0014] Based on the wear data and temperature interference data corresponding to each predicted data, the ideal torque corresponding to each predicted data is corrected to obtain the expected torque corresponding to each predicted data. The wear data is used to represent the material wear accumulated in the braking system, and the temperature interference data is used to represent the torque calculation interference introduced by the braking system due to temperature changes during the predicted time period.

[0015] In one embodiment, the step of obtaining the wear data corresponding to each predicted data includes:

[0016] Acquire multiple historical braking data of the braking system, and the multiple historical braking data correspond one-to-one with multiple braking operations of the braking system;

[0017] The ratio of the first temperature rise rate to the corresponding downforce for each historical braking data is determined as the temperature rise efficiency for each historical braking data. The first temperature rise rate is the average of multiple temperature change rates included in the corresponding historical braking data. The downforce is used to represent the force with which the brake pedal of the braking system is pressed under the corresponding braking operation.

[0018] The sum of the temperature rise efficiency corresponding to each historical braking data point and the temperature rise efficiency corresponding to the next historical braking data point is determined as the temperature rise efficiency of that historical braking data point.

[0019] The ratio of the temperature rise efficiency of each historical braking data point to the braking interval corresponding to that historical braking data point is determined as the wear degree of that historical braking data point, wherein the braking interval is the time interval between the corresponding historical braking data point and the next historical braking data point.

[0020] The wear rate of each historical braking data point is summed to obtain the total wear rate;

[0021] The product of the reciprocal of the total wear rate and the predicted temperature rise efficiency corresponding to each predicted data point is normalized to obtain the wear data corresponding to each predicted data point. The predicted temperature rise efficiency is the ratio of the predicted temperature rise rate to the predicted downforce included in the corresponding predicted data point.

[0022] In one embodiment, the step of obtaining the temperature disturbance data corresponding to each predicted data point includes:

[0023] Each predicted data point is input into a pre-trained random forest model to obtain multiple output values ​​from multiple random trees in the random forest model for the predicted data. The random forest model is used to predict the torque deviation of the motor corresponding to the input data based on the input data.

[0024] The mean of multiple output values ​​corresponding to each predicted data is determined as the first deviation mean of each predicted data, and the absolute value of the mean of multiple temperature deviation values ​​corresponding to multiple predicted data is determined as the second deviation mean, wherein the temperature deviation value is the difference between the predicted temperature included in the corresponding predicted data and the preset reference temperature.

[0025] Based on the temperature influence factor corresponding to the random forest model, the first mean deviation, the second mean deviation, and the temperature difference corresponding to each predicted data, temperature interference data corresponding to each predicted data is obtained. The temperature influence factor represents the degree of influence of temperature indicators on the model output among the various indicators included in the input data of the random forest model. The temperature influence factor is greater than or equal to a preset threshold. The temperature indicators include: temperature rise rate, the time integral corresponding to the difference between temperature and reference temperature, and the product of temperature and corresponding predicted current. The temperature difference is the difference between the mean temperature of the corresponding predicted data and the reference temperature. The mean temperature is the average of multiple predicted temperatures associated with the time segment between the first predicted data and the corresponding predicted data.

[0026] In one embodiment, obtaining the signal anomaly index for each predicted data point based on its corresponding frequency deviation and duty cycle deviation includes:

[0027] Calculate the product of frequency deviation and frequency coefficient for each predicted data point to obtain the frequency difference index for each predicted data point; calculate the product of duty cycle deviation and duty cycle coefficient for each predicted data point to obtain the duty cycle difference index for each predicted data point; wherein the sum of the frequency coefficient and the duty cycle coefficient is 1.

[0028] The sum of the frequency difference index and the duty cycle difference index corresponding to each predicted data point is determined as the signal anomaly index for each predicted data point.

[0029] In one embodiment, before obtaining the signal anomaly index of each predicted data based on the frequency deviation and duty cycle deviation corresponding to each predicted data, the method further includes:

[0030] Acquire multiple historical data from a reference system, wherein the reference system and the braking system are of the same model.

[0031] Based on the aforementioned historical data, multiple frequency deviation samples and multiple duty cycle deviation samples were obtained.

[0032] Correlation analysis was performed on multiple frequency deviation samples and multiple expected torques corresponding to the multiple frequency deviation samples to obtain frequency correlation coefficients; and correlation analysis was performed on multiple duty cycle deviation samples and multiple expected torques corresponding to the multiple duty cycle samples to obtain duty cycle correlation coefficients.

[0033] The frequency correlation coefficient and the duty cycle correlation coefficient are summed to obtain the sum value.

[0034] The frequency coefficient is determined by the ratio of the frequency correlation coefficient to the sum of the coefficients, and the duty cycle coefficient is determined by the ratio of the duty cycle correlation coefficient to the sum of the coefficients.

[0035] In one embodiment, the analysis of the abnormal probability of the motor and transmission mechanism and the signal abnormality index of each predicted data to obtain the fault probability value of each predicted data includes:

[0036] The product of the abnormal probability of the motor and transmission mechanism and the abnormal signal index for each predicted data is determined as the first probability value for each predicted data.

[0037] Based on the abnormal probability of the motor and transmission mechanism and the abnormal signal index of each predicted data, a probability correction factor is obtained for each predicted data. The probability correction factor is used to represent the degree of difference between the corresponding predicted data and multiple fault data of the braking system.

[0038] The ratio of the first probability value of each predicted data point to its corresponding probability correction factor is determined as the second probability value of each predicted data point.

[0039] The second probability value of each predicted data point is normalized to obtain the failure probability value of each predicted data point.

[0040] In one embodiment, the analysis of the fault probability value of each predicted data point to obtain the fault detection result of the braking system within the predicted time period includes:

[0041] Calculate the mean of multiple failure probability values ​​from multiple predicted data to obtain the first mean;

[0042] Based on the fault probability value of each predicted data, the slope value of each predicted data is obtained, wherein the fault slope value is the instantaneous rate of change of the corresponding fault probability value.

[0043] Calculate the mean of multiple slope values ​​for multiple predicted data to obtain a second mean;

[0044] The normalized value of the product of the first mean and the second mean is determined as the target value;

[0045] If the target value is greater than or equal to a preset probability threshold, output a fault detection result indicating that the braking system will fail within the predicted time period.

[0046] If the target value is less than the probability threshold, the output indicates that the braking system will not malfunction within the predicted time period.

[0047] Secondly, another embodiment of the present invention provides a fault detection system for an electromechanical braking system, the system comprising:

[0048] The acquisition module is used to acquire multiple prediction data of the braking system within the prediction time period;

[0049] The overload analysis module is used to analyze the difference between the expected torque corresponding to each predicted data and the predicted torque included therein, and to obtain the abnormal probability of the motor and transmission mechanism for each predicted data. The expected torque is obtained after correcting the temperature error based on the corresponding ideal torque, and the ideal torque is calculated based on the predicted current included in the corresponding predicted data.

[0050] The signal analysis module is used to obtain the signal anomaly index of each predicted data based on the frequency deviation and duty cycle deviation corresponding to each predicted data. The frequency deviation is used to represent the difference between the predicted frequency included in the predicted data and the corresponding rated frequency, and the duty cycle deviation is used to represent the difference between the predicted duty cycle included in the predicted data and the corresponding expected duty cycle.

[0051] The fault probability analysis module is used to analyze the abnormal probability of motor and transmission mechanism and the signal abnormality index of each predicted data to obtain the fault probability value of each predicted data.

[0052] The fault detection module is used to analyze the fault probability value of each predicted data point and obtain the fault detection results of the braking system within the predicted time period.

[0053] Thirdly, another embodiment of the present invention provides a fault detection device for an electromechanical braking system, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method described in the first aspect.

[0054] Fourthly, in another embodiment of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0055] The present invention has the following beneficial effects:

[0056] This invention determines the probability of abnormalities in the predicted motor and transmission mechanism by analyzing the difference between the expected torque corresponding to the predicted data and the predicted torque included therein. Specifically, it analyzes the difference between the actual torque exhibited by the motor under the operating conditions indicated by the predicted data and the corresponding expected torque. After calculating the ideal torque based on the predicted current included in the predicted data, the temperature error included in this ideal torque is corrected to minimize temperature interference, making the calculated expected torque more accurate and thus making the analyzed probability of abnormalities in the motor and transmission mechanism more accurate and reliable. Furthermore, the invention analyzes the difference between the predicted frequency and the rated frequency, as well as the difference between the predicted duty cycle and the expected duty cycle, to analyze whether the electrical signal of the predicted data is abnormal from both frequency and duty cycle perspectives. Anomalies are identified, leading to a signal anomaly index in the predicted data. Then, based on the anomaly probability of the motor and transmission mechanism and the signal anomaly index in the predicted data, the probability value of the motor failing under the operating conditions indicated by the predicted data is comprehensively determined. Finally, the fault detection result of the braking system is output based on the fault probability value. Since the above processing effectively eliminates the adverse effects of temperature interference and electrical signal transmission errors, the accuracy of the fault detection result can be significantly improved. Among these methods, selecting multiple predicted data points for the braking system within the predicted time period is to improve the timeliness of the output fault detection result, so as to identify potential faults before the braking system fails, thereby reminding users to take preventive measures against possible faults in the braking system and reducing the risk of accidents caused by braking system failures. Attached Figure Description

[0057] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0058] Figure 1 This is a flowchart illustrating a fault detection method for an electromechanical braking system according to an embodiment of the present invention.

[0059] Figure 2 A schematic diagram of the structure of a fault detection system for an electromechanical braking system provided in an embodiment of the present invention;

[0060] Figure 3 This is a schematic diagram of the structure of a fault detection device for an electromechanical braking system provided in one embodiment of the present invention. Detailed Implementation

[0061] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a fault detection method, detection system, and apparatus for an electromechanical braking system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0063] The specific scheme of the fault detection method for an electromechanical braking system provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0064] This invention proposes a fault detection method for an electromechanical braking system. Please refer to [link / reference]. Figure 1 The diagram illustrates a schematic flowchart of a fault detection method for an electromechanical braking system according to an embodiment of the present invention. The method includes the following steps:

[0065] Step S1: Obtain multiple prediction data of the braking system within the prediction time period.

[0066] Multiple historical operating data of the braking system within a historical time period (such as the past month or the past year) can be obtained, and a prediction model can be constructed based on the multiple historical operating data. The constructed prediction model can then be used to predict the operating status of the braking system within the prediction time period, thereby obtaining the aforementioned multiple prediction data.

[0067] In one example, the prediction model described above could be an Autoregressive Integrated Moving Average (ARIMA) model.

[0068] The above-mentioned historical operating data corresponds one-to-one with multiple historical operating moments within the historical time period. The historical operating data includes the actual motor torque of the braking system at the corresponding historical operating moment (obtained by a torque sensor), motor speed (collected by an encoder), motor operating current (measured by a current sensor), brake surface temperature (monitored by a temperature sensor), and load pressure (obtained by a pressure sensor).

[0069] In applications, multiple historical operating data can be preprocessed first, and then a prediction model can be built based on the preprocessed historical operating data. The preprocessing includes: removing data with abnormal values ​​(such as temperature data with an instantaneous temperature exceeding 200°C, or torque data that is negative), compensating for short-term missing data (such as filling missing data within 5 seconds by linear interpolation), and normalizing the values ​​(such as using the Min-Max Normalization algorithm for normalization processing).

[0070] The start time of the above-mentioned prediction time period can be the current time or a future time after the current time. The duration of the above-mentioned prediction time period is a preset duration. For example, if the current time is set to 15:15 (24-hour system), the above-mentioned prediction time period can be from 15:15 to 15:16 or from 16:15 to 16:16.

[0071] It should be noted that the above-mentioned multiple prediction data and multiple prediction moments within the prediction time period correspond one-to-one. The prediction data includes the motor torque (i.e., predicted torque), motor speed, motor operating current (i.e., predicted current included in the prediction data), brake surface temperature, the force of the brake pedal being pressed (i.e., predicted downforce), load pressure, etc. of the braking system at the corresponding prediction moment.

[0072] Step S2: Analyze the difference between the expected torque corresponding to each predicted data and the predicted torque it includes, and obtain the abnormal probability of the motor and transmission mechanism for each predicted data.

[0073] The desired torque is obtained by correcting for temperature error based on the corresponding ideal torque, and the ideal torque is calculated based on the predicted current included in the corresponding predicted data.

[0074] In this step, the torque difference between the expected torque (the torque output by the motor under ideal operating conditions) and the actual torque (the torque output by the motor under actual operating conditions at the predicted time, i.e., the predicted torque) at the corresponding predicted time is analyzed to determine the probability of anomalies in the motor at the corresponding predicted time (i.e., the probability of anomalies in the motor and transmission mechanism). The larger the torque difference, the greater the probability of anomalies in the motor at the corresponding predicted time, and vice versa.

[0075] Under ideal operating conditions, the actual output torque of the motor will increase linearly with the increase of the predicted current. Therefore, the ideal torque in this invention can be calculated based on the following formula:

[0076]

[0077] in, This represents the ideal torque of the motor at the corresponding predicted time. The known motor torque constant can be directly obtained from the motor parameters of the braking system. This refers to the predicted current included in the corresponding prediction data.

[0078] In actual operating conditions, due to the existence of temperature error, the motor torque reflected by the predicted current may differ from the ideal torque. Therefore, after calculating the ideal torque, it is necessary to correct the temperature error in the ideal torque to eliminate the calculation error caused by temperature factors, so that the torque difference can accurately reflect the possibility of motor abnormality.

[0079] The aforementioned temperature error may include: material wear accumulated in the brake system during long-term use (such as brake pads gradually thinning during long-term use, resulting in a weakening of braking effect), and resistance changes caused by short-term temperature rise in the brake system during the corresponding braking process (such as the brake disc and friction pads expanding due to heat, leading to an increase in their contact area, which in turn increases the frictional resistance between them, further consuming the energy output of the motor, and causing an increase in the difference between the actual torque of the motor and the expected torque expressed by the current).

[0080] It should be understood that errors caused by material wear and changes in resistance can easily lead to misjudgment problems (i.e., misjudging a motor operating under normal conditions as abnormal). By correcting these errors, the above problems can be effectively avoided, so that the expected torque can accurately reflect the torque that the motor should exhibit under the corresponding predicted current.

[0081] Furthermore, the steps for obtaining the expected torque corresponding to each predicted data point include:

[0082] The ideal torque corresponding to each prediction data point is determined by multiplying the motor torque constant corresponding to the braking system with the predicted current included in each prediction data point.

[0083] Based on the wear data and temperature interference data corresponding to each predicted data, the ideal torque corresponding to each predicted data is corrected to obtain the expected torque corresponding to each predicted data. The wear data is used to represent the material wear accumulated in the braking system, and the temperature interference data is used to represent the torque calculation interference introduced by the braking system due to temperature changes during the predicted time period.

[0084] The wear data mentioned above are used to indicate errors caused by material wear, and the temperature disturbance data mentioned above are used to indicate errors caused by changes in resistance.

[0085] The steps for obtaining the wear data corresponding to each predicted data point include:

[0086] Acquire multiple historical braking data of the braking system, and the multiple historical braking data correspond one-to-one with multiple braking operations of the braking system;

[0087] The ratio of the first temperature rise rate to the corresponding downforce for each historical braking data is determined as the temperature rise efficiency for each historical braking data. The first temperature rise rate is the average of multiple temperature change rates included in the corresponding historical braking data. The downforce is used to represent the force with which the brake pedal of the braking system is pressed under the corresponding braking operation.

[0088] The sum of the temperature rise efficiency corresponding to each historical braking data point and the temperature rise efficiency corresponding to the next historical braking data point is determined as the temperature rise efficiency of that historical braking data point.

[0089] The ratio of the temperature rise efficiency of each historical braking data point to the braking interval corresponding to that historical braking data point is determined as the wear degree of that historical braking data point, wherein the braking interval is the time interval between the corresponding historical braking data point and the next historical braking data point.

[0090] The wear rate of each historical braking data point is summed to obtain the total wear rate;

[0091] The product of the reciprocal of the total wear rate and the predicted temperature rise efficiency corresponding to each predicted data point is normalized to obtain the wear data corresponding to each predicted data point. The predicted temperature rise efficiency is the ratio of the predicted temperature rise rate to the predicted downforce included in the corresponding predicted data point.

[0092] Among them, the multiple braking operations corresponding to multiple historical braking data can be understood as: all braking operations of the braking system before the current moment.

[0093] In one example, a calibrated value (e.g., 100 mm) for the displacement of the brake pedal can be predefined, and the start and end points of a braking operation can be identified based on this calibrated value. The start point is the moment when the displacement of the brake pedal first exceeds 5% of the calibrated value during the gradual increase, and the end point is the moment when the displacement of the brake pedal first falls below 3% of the calibrated value during the gradual decrease.

[0094] In another example, the K-means algorithm can be used to cluster all historical operating data of the braking system to group consecutive sampling time points with similar temperatures into one class, and the elbow method can be used to obtain the optimal K value, thereby obtaining several non-braking process time intervals, and then obtaining multiple braking process time intervals based on the several non-braking process time intervals (the time interval between two adjacent non-braking process time intervals is a braking process time interval), thus obtaining the aforementioned multiple historical braking data.

[0095] As mentioned earlier, one historical braking data point corresponds to one braking operation. Since a braking operation requires a certain amount of time to complete, one historical braking data point corresponds to multiple consecutive historical braking moments. At each historical braking moment, a historical temperature (the surface temperature of the brake) is recorded. By integrating the multiple historical temperatures at multiple historical braking moments corresponding to one historical braking data point over time, multiple temperature change rates corresponding to that historical braking data point can be obtained. By calculating the average of the multiple temperature change rates corresponding to that historical braking data point, the first temperature rise rate corresponding to that historical braking data point can be obtained.

[0096] The aforementioned temperature rise efficiency is used to represent the degree to which the force applied to the brake pedal affects the rate of temperature change in the brake. A higher temperature rise efficiency indicates a faster rate of temperature rise in the brake, resulting in greater resistance and consequently, a greater error due to changes in resistance.

[0097] The predicted temperature rise efficiency corresponding to the predicted data can be: the ratio of the integral value of the predicted temperature over time included in the predicted data to the force of the brake pedal being pressed at the corresponding predicted time.

[0098] The wear rate mentioned above is used to represent the material wear generated by the braking system during the corresponding braking operation. Specifically, the shorter the time interval between the corresponding braking operation and the next braking operation, the greater the material wear. Similarly, the greater the temperature rise efficiency of the corresponding braking process, the greater the material wear.

[0099] For example, the temperature rise efficiency corresponding to the i-th historical braking data (which can be understood as any one of the multiple historical braking data) is defined as follows: The braking interval corresponding to the i-th historical braking data is ( , This is the braking starting point for the (i+1)th historical braking data. (where i is the braking endpoint of the i-th historical braking data), the predicted temperature rise efficiency corresponding to the predicted data at prediction time t in multiple predicted data is: .

[0100] Therefore, the initial wear data corresponding to the predicted time t among the multiple predicted data sets. for:

[0101] In the formula, N represents the total number of historical braking data points.

[0102] The initial wear data corresponding to the predicted data at prediction time t is obtained by using the norm function. By normalizing, we can obtain the wear data corresponding to the predicted time t from multiple predicted data sets. The wear data is dynamically updated as the braking process increases.

[0103] In this process, the material wear of the brake is calculated by temperature rise efficiency, rather than solely by the force applied to the brake pedal. This utilizes the introduced temperature change rate to more accurately reflect the material wear caused by downward pressure during the corresponding braking operation. Furthermore, by statistically analyzing the material wear from multiple past braking processes, a total wear is calculated to fully account for the accumulated wear of the brake during past braking events, making the wear data determined accordingly more accurate and reliable.

[0104] The steps for obtaining the temperature disturbance data corresponding to each predicted data point include:

[0105] Each predicted data point is input into a pre-trained random forest model to obtain multiple output values ​​from multiple random trees in the random forest model for the predicted data. The random forest model is used to predict the torque deviation of the motor corresponding to the input data based on the input data.

[0106] The mean of multiple output values ​​corresponding to each predicted data is determined as the first deviation mean of each predicted data, and the absolute value of the mean of multiple temperature deviation values ​​corresponding to multiple predicted data is determined as the second deviation mean, wherein the temperature deviation value is the difference between the predicted temperature included in the corresponding predicted data and the preset reference temperature.

[0107] Based on the temperature influence factor corresponding to the random forest model, the first mean deviation, the second mean deviation, and the temperature difference corresponding to each predicted data, temperature interference data corresponding to each predicted data is obtained. The temperature influence factor represents the degree of influence of temperature indicators on the model output among the various indicators included in the input data of the random forest model. The temperature influence factor is greater than or equal to a preset threshold. The temperature indicators include: temperature rise rate, the time integral corresponding to the difference between temperature and reference temperature, and the product of temperature and corresponding predicted current. The temperature difference is the difference between the mean temperature of the corresponding predicted data and the reference temperature. The mean temperature is the average of multiple predicted temperatures associated with the time segment between the first predicted data and the corresponding predicted data.

[0108] The process of obtaining multiple random trees in the above random forest model can be as follows:

[0109] Obtain the temperature from each historical braking data point. and current and actual torque At the same time, ensure that the acquisition time of temperature, current and torque data is strictly synchronized (error <10ms) to avoid incorrect association in model learning due to timing misalignment.

[0110] The time-series data of temperature and current are used as basic features, while the temperature difference between adjacent time points (i.e., the aforementioned rate of temperature rise, used to reflect the instantaneous trend of temperature change during historical braking), the time integral of the temperature difference with the reference temperature (used to reflect the cumulative effect of temperature rise caused by the corresponding load during the corresponding historical braking), and the product of current and temperature (used to capture the synergistic effect of current and temperature on torque) are used as derived features.

[0111] Select a reference temperature (e.g., 25°C) and calculate the theoretical torque for each data point. Constructing label vectors That is, actual torque Analysis shows that the deviation from the theoretical value is mainly caused by temperature.

[0112] Next, the dataset (each data in the dataset includes the aforementioned basic features and 3 derived features) is divided into a training set (70%), a validation set (15%), and a test set (15%) in chronological order. Initially, 100 trees are set (meaning 100 random trees will be generated), with a maximum depth of 10 layers and a minimum sample split of 5. The training objective is to minimize the mean squared error (MSE) of the validation set. Multiple random trees are trained to obtain the aforementioned trees. Each random tree can output a predicted label vector Y' based on the input data (including the aforementioned basic features and 3 derived features).

[0113] The temperature influence factor mentioned above can be understood as the sum of the contributions of the three derived features (all temperature features) to the output predicted label vector Y' (which needs to be normalized). If the temperature influence factor is greater than or equal to the preset threshold (e.g., 0.3), it can be considered that temperature will have a significant impact on torque calculation. Therefore, the error caused by temperature needs to be considered in the calculation of the expected torque. Conversely, if the temperature influence factor is less than the preset threshold, it is considered that the influence of temperature on torque calculation is low, and the error caused by temperature can be ignored in the calculation of the expected torque.

[0114] For example, if the temperature influence factor is set to The mean of the first deviation for each predicted data point is Second deviation mean (It should be understood that the mean of the second deviation for different prediction data is the same), then the temperature compensation coefficient for:

[0115]

[0116] The temperature interference data corresponding to each predicted data point is: ,in, This is the difference between the mean temperature of the corresponding predicted data and the reference temperature.

[0117] In this process, based on the introduction of multiple derived features, the influence of temperature on several key indicators with the same torque is fully considered. Based on this, it is analyzed whether temperature will significantly affect the accuracy of the calculation of the expected torque in the current braking system. Under the influence, the output of each random tree and the actual temperature difference reflected by the prediction data are combined to determine more accurate temperature interference data, that is, to determine the torque calculation error introduced by temperature at the corresponding prediction time, so as to ensure the accuracy of the abnormal probability of the motor and transmission mechanism in subsequent analysis.

[0118] In one example, the expected torque corresponding to the predicted data at prediction time t. It can be represented as: in, The predicted data for prediction time t includes the predicted current.

[0119] Step S3: Based on the frequency deviation and duty cycle deviation corresponding to each predicted data, obtain the signal anomaly index of each predicted data.

[0120] The frequency deviation is used to represent the difference between the predicted frequency included in the corresponding predicted data and the corresponding rated frequency, and the duty cycle deviation is used to represent the difference between the predicted duty cycle included in the corresponding predicted data and the corresponding expected duty cycle.

[0121] In this invention, the controller of the braking system transmits braking commands to the motor based on pulse width modulation (PWM) signals to precisely control the deceleration, stopping, or energy recovery process of the motor. By monitoring or predicting the PWM signal, the corresponding frequency and duty cycle can be obtained.

[0122] Further, the step of obtaining the signal anomaly index for each predicted data point based on its corresponding frequency deviation and duty cycle deviation includes:

[0123] Calculate the product of frequency deviation and frequency coefficient for each predicted data point to obtain the frequency difference index for each predicted data point; calculate the product of duty cycle deviation and duty cycle coefficient for each predicted data point to obtain the duty cycle difference index for each predicted data point; wherein the sum of the frequency coefficient and the duty cycle coefficient is 1.

[0124] The sum of the frequency difference index and the duty cycle difference index corresponding to each predicted data point is determined as the signal anomaly index for each predicted data point.

[0125] The frequency deviation corresponding to each predicted data point is the absolute value of the difference between the predicted frequency included in the predicted data and the corresponding rated frequency (e.g., 15kHz). Similarly, the duty cycle deviation corresponding to each predicted data point is the absolute value of the difference between the predicted duty cycle included in the predicted data and the corresponding expected duty cycle (achieved through a pre-built load pressure-duty cycle mapping table, determined by looking up the table based on the predicted load pressure included in the predicted data).

[0126] It should be noted that in the load pressure-duty cycle mapping table, the duty cycle corresponding to each load pressure is the optimal duty cycle for stable torque control of the braking system under the corresponding load pressure (obtained through extensive prior experiments).

[0127] It should be noted that the frequency coefficient and duty cycle coefficient mentioned above are both positive numbers, and the frequency coefficient and duty cycle coefficient can be preset (e.g., both set to 0.5), or the coefficients can be dynamically set by analyzing the degree of influence of frequency and duty cycle on the signal anomaly index. The greater the degree of influence, the larger the coefficient, and vice versa.

[0128] For example, if we set the frequency deviation of the predicted data at prediction time t to be... The corresponding duty cycle deviation is The duty cycle factor is Then the signal anomaly index of the predicted data at prediction time t It can be represented as:

[0129]

[0130] Since PWM signals control the amplitude and on / off time of the motor drive voltage by adjusting the duty cycle and frequency, thereby affecting the motor's output torque, a difference analysis of its frequency and duty cycle is performed to comprehensively determine the PWM signal transmission fault. This avoids the analytical bias that may be caused by a single-dimensional analysis (for example, external disturbances such as electromagnetic interference and power fluctuations usually cause frequency abnormalities, but have a smaller impact on the duty cycle. If only the duty cycle is analyzed, the above abnormalities cannot be effectively identified. Similarly, under conditions such as controller hardware aging and sudden load changes, the duty cycle may become out of control while the frequency remains stable. If only the frequency is analyzed, the above abnormalities also cannot be effectively identified). This makes the determined signal abnormality index more accurate.

[0131] Specifically, before obtaining the signal anomaly index of each predicted data based on the frequency deviation and duty cycle deviation corresponding to each predicted data, the method further includes:

[0132] Acquire multiple historical data from a reference system, wherein the reference system and the braking system are of the same model.

[0133] Based on the aforementioned historical data, multiple frequency deviation samples and multiple duty cycle deviation samples were obtained.

[0134] Correlation analysis was performed on multiple frequency deviation samples and multiple expected torques corresponding to the multiple frequency deviation samples to obtain frequency correlation coefficients; and correlation analysis was performed on multiple duty cycle deviation samples and multiple expected torques corresponding to the multiple duty cycle samples to obtain duty cycle correlation coefficients.

[0135] The frequency correlation coefficient and the duty cycle correlation coefficient are summed to obtain the sum value.

[0136] The frequency coefficient is determined by the ratio of the frequency correlation coefficient to the sum of the coefficients, and the duty cycle coefficient is determined by the ratio of the duty cycle correlation coefficient to the sum of the coefficients.

[0137] The frequency deviation sample / duty cycle deviation sample includes: the motor torque (i.e., predicted torque), motor speed, motor operating current (i.e., the predicted current included in the predicted data) of the braking system at the corresponding historical time, brake surface temperature, the force applied to the brake pedal, load pressure, etc. The process for obtaining the expected torque corresponding to the frequency deviation sample / duty cycle deviation sample can be found in the foregoing explanation and will not be repeated here.

[0138] For example, the process of performing correlation analysis on multiple frequency deviation samples and multiple expected torques corresponding one-to-one with the multiple frequency deviation samples to obtain the frequency correlation coefficient can be as follows:

[0139] Multiple historical frequency values ​​are extracted from multiple frequency deviation samples, and the absolute values ​​of the differences between the multiple historical frequency values ​​and the aforementioned rated frequency are calculated to obtain multiple frequency difference values.

[0140] Multiple frequency differences are ordered to form a frequency deviation sequence according to their corresponding time, and multiple expected torques corresponding to multiple frequency deviation samples are ordered to form a frequency torque sequence according to their corresponding time.

[0141] Calculate the DTW distance between the frequency deviation sequence and the frequency torque sequence, and determine the frequency correlation coefficient by the reciprocal of the calculated DTW distance.

[0142] The method for obtaining the duty cycle correlation coefficient is similar to that for obtaining the frequency correlation coefficient, so it will not be repeated here to avoid repetition.

[0143] Step S4: Analyze the abnormal probability of the motor and transmission mechanism and the abnormal signal index of each predicted data to obtain the fault probability value of each predicted data.

[0144] In this step, the probability of a brake system failure (i.e., the failure probability value) is accurately determined by comprehensively analyzing the possibility of abnormality in the motor and the possibility of abnormality in the corresponding electrical signal during transmission.

[0145] Specifically, the analysis of the abnormal probability of the motor and transmission mechanism and the signal abnormality index of each predicted data point to obtain the fault probability value of each predicted data point includes:

[0146] The product of the abnormal probability of the motor and transmission mechanism and the abnormal signal index for each predicted data is determined as the first probability value for each predicted data.

[0147] Based on the abnormal probability of the motor and transmission mechanism and the abnormal signal index of each predicted data, a probability correction factor is obtained for each predicted data. The probability correction factor is used to represent the degree of difference between the corresponding predicted data and multiple fault data of the braking system.

[0148] The ratio of the first probability value of each predicted data point to its corresponding probability correction factor is determined as the second probability value of each predicted data point.

[0149] The second probability value of each predicted data point is normalized to obtain the failure probability value of each predicted data point.

[0150] For example, the failure probability value of the predicted data at prediction time t. It can be represented as:

[0151] in, This represents the probability correction factor corresponding to the predicted data at prediction time t, where norm represents the normalization operation. This represents the signal anomaly index of the predicted data at prediction time t. This represents the probability of abnormality in the motor and transmission mechanism in the predicted data at prediction time t.

[0152] In one example, after obtaining the aforementioned multiple fault data, the abnormal probability of the motor and transmission mechanism and the abnormal index of the signal can be obtained by referring to the method for obtaining the abnormal probability of the motor and transmission mechanism and the abnormal index of the signal for each fault data. Then, the fault data point corresponding to each fault data is determined in the target rectangular coordinate system. The horizontal axis and vertical axis of the target rectangular coordinate system represent the abnormal probability of the motor and transmission mechanism and the abnormal index of the signal, respectively. By performing linear fitting on multiple fault data points corresponding to multiple fault data, the fault fitting line is determined, and the shortest distance between the predicted data point corresponding to the predicted data in the target rectangular coordinate system and the fault fitting line is calculated. The shortest distance is determined as the probability correction factor corresponding to the predicted data.

[0153] In this step, in addition to longitudinally analyzing the possible motor and signal anomalies reflected in the prediction data, the degree of difference between the situation reflected in the prediction data and multiple historical fault situations will also be compared horizontally. In order to determine a more accurate fault probability value by combining historical experience (reflected by probability correction factors) and actual operating conditions (reflected by the abnormal probability of motor and transmission mechanism and the abnormality index of signal).

[0154] For example, the normalization of the second probability value can be performed based on normalization algorithms such as Min-Max normalization.

[0155] Step S5: Analyze the fault probability value of each predicted data to obtain the fault detection results of the braking system within the predicted time period.

[0156] For example, the mean of multiple failure probability values ​​of multiple predicted data can be calculated, and if the mean is greater than or equal to a preset probability threshold (such as 0.5), a failure detection result indicating that the braking system will fail within the predicted time period can be output.

[0157] When the mean value is less than the probability threshold, the output indicates that the braking system will not malfunction within the predicted time period.

[0158] This invention determines the probability of abnormalities in the predicted motor and transmission mechanism by analyzing the difference between the expected torque corresponding to the predicted data and the predicted torque included therein. Specifically, it analyzes the difference between the actual torque exhibited by the motor under the operating conditions indicated by the predicted data and the corresponding expected torque. After calculating the ideal torque based on the predicted current included in the predicted data, the temperature error included in this ideal torque is corrected to minimize temperature interference, making the calculated expected torque more accurate and thus making the analyzed probability of abnormalities in the motor and transmission mechanism more accurate and reliable. Furthermore, the invention analyzes the difference between the predicted frequency and the rated frequency, as well as the difference between the predicted duty cycle and the expected duty cycle, to analyze whether the electrical signal of the predicted data is abnormal from both frequency and duty cycle perspectives. Anomalies are identified, leading to a signal anomaly index in the predicted data. Then, based on the anomaly probability of the motor and transmission mechanism and the signal anomaly index in the predicted data, the probability value of the motor failing under the operating conditions indicated by the predicted data is comprehensively determined. Finally, the fault detection result of the braking system is output based on the fault probability value. Since the above processing effectively eliminates the adverse effects of temperature interference and electrical signal transmission errors, the accuracy of the fault detection result can be significantly improved. Among these methods, selecting multiple predicted data points for the braking system within the predicted time period is to improve the timeliness of the output fault detection result, so as to identify potential faults before the braking system fails, thereby reminding users to take preventive measures against possible faults in the braking system and reducing the risk of accidents caused by braking system failures.

[0159] In one embodiment, the analysis of the fault probability value of each predicted data point to obtain the fault detection result of the braking system within the predicted time period includes:

[0160] Calculate the mean of multiple failure probability values ​​from multiple predicted data to obtain the first mean;

[0161] Based on the fault probability value of each predicted data, the slope value of each predicted data is obtained, wherein the fault slope value is the instantaneous rate of change of the corresponding fault probability value.

[0162] Calculate the mean of multiple slope values ​​for multiple predicted data to obtain a second mean;

[0163] The normalized value of the product of the first mean and the second mean is determined as the target value;

[0164] If the target value is greater than or equal to a preset probability threshold, output a fault detection result indicating that the braking system will fail within the predicted time period.

[0165] If the target value is less than the probability threshold, the output indicates that the braking system will not malfunction within the predicted time period.

[0166] For example, the slope value of each predicted data point can be obtained by integrating the failure probability value of each predicted data point over time.

[0167] In this embodiment, by introducing a slope value, the probability of failure of the braking system and the rate of change of the probability of failure are considered simultaneously during the prediction period. This adapts to the fact that the braking system will not only generate data indicators with high failure risk in real failure scenarios, but will also experience a process of sudden increase in the probability of failure. This can reduce the detection error caused by the deviation in the calculation of the failure probability value to a certain extent, making the final output failure detection result more accurate and reliable.

[0168] This invention proposes a fault detection system for an electromechanical braking system. Please refer to [link / reference]. Figure 2 The diagram illustrates a structural schematic of a fault detection system 200 for an electromechanical braking system according to an embodiment of the present invention. The system includes:

[0169] The acquisition module 201 is used to acquire multiple prediction data of the braking system within the prediction time period;

[0170] The overload analysis module 202 is used to analyze the difference between the expected torque corresponding to each predicted data and the predicted torque included therein, and to obtain the abnormal probability of the motor and transmission mechanism for each predicted data. The expected torque is obtained after correcting the temperature error based on the corresponding ideal torque, and the ideal torque is calculated based on the predicted current included in the corresponding predicted data.

[0171] The signal analysis module 203 is used to obtain the signal anomaly index of each predicted data according to the frequency deviation and duty cycle deviation corresponding to each predicted data. The frequency deviation is used to represent the difference between the predicted frequency included in the corresponding predicted data and the corresponding rated frequency, and the duty cycle deviation is used to represent the difference between the predicted duty cycle included in the corresponding predicted data and the corresponding expected duty cycle.

[0172] The fault probability analysis module 204 is used to analyze the abnormal probability of motor and transmission mechanism and the abnormal signal index of each predicted data to obtain the fault probability value of each predicted data.

[0173] The fault detection module 205 is used to analyze the fault probability value of each predicted data to obtain the fault detection results of the braking system within the predicted time period.

[0174] The fault detection method for an electromechanical braking system provided in the above embodiments and the fault detection system embodiment for an electromechanical braking system belong to the same concept. The specific implementation process is detailed in the method embodiments and will not be repeated here.

[0175] This invention also provides a fault detection device for an electromechanical braking system. Please refer to [link / reference]. Figure 3 The fault detection device for the electromechanical braking system may include a processor 301, a memory 302, and a program 3021 stored in the memory 302 and capable of running on the processor 301.

[0176] When program 3021 is executed by processor 301, it can achieve the following: Figure 1 Any steps in the corresponding method embodiments and the achievement of the same beneficial effects will not be repeated here.

[0177] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by hardware related to program instructions, and the program can be stored in a readable medium.

[0178] This invention also provides a readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described functions. Figure 1 Any step in the corresponding method embodiment can achieve the same technical effect, and will not be repeated here to avoid repetition.

[0179] The computer-readable storage medium of this invention can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. For example, a computer-readable storage medium can be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0180] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0181] The program code contained on the storage medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0182] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or terminal. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0183] This invention also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned steps to implement a fault detection method for an electromechanical braking system provided in the above embodiments.

[0184] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0185] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A fault detection method for an electromechanical braking system, characterized in that, The method includes: Acquire multiple forecast data points for the braking system within the forecast time period; The difference between the expected torque corresponding to each predicted data and the predicted torque included therein is analyzed to obtain the abnormal probability of the motor and transmission mechanism for each predicted data. The expected torque is obtained after correcting for temperature error based on the corresponding ideal torque, and the ideal torque is calculated based on the predicted current included in the corresponding predicted data. Based on the frequency deviation and duty cycle deviation corresponding to each predicted data, the signal anomaly index of each predicted data is obtained, wherein the frequency deviation is used to represent the difference between the predicted frequency included in the corresponding predicted data and the corresponding rated frequency, and the duty cycle deviation is used to represent the difference between the predicted duty cycle included in the corresponding predicted data and the corresponding expected duty cycle. Analyze the abnormal probability of motor and transmission mechanism and the abnormal signal index of each predicted data to obtain the fault probability value of each predicted data. By analyzing the failure probability value of each predicted data point, the failure detection results of the braking system within the predicted time period are obtained.

2. The fault detection method for an electromechanical braking system according to claim 1, characterized in that, The steps for obtaining the expected torque corresponding to each predicted data point include: The ideal torque corresponding to each prediction data point is determined by multiplying the motor torque constant corresponding to the braking system with the predicted current included in each prediction data point. Based on the wear data and temperature interference data corresponding to each predicted data, the ideal torque corresponding to each predicted data is corrected to obtain the expected torque corresponding to each predicted data. The wear data is used to represent the material wear accumulated in the braking system, and the temperature interference data is used to represent the torque calculation interference introduced by the braking system due to temperature changes during the predicted time period.

3. The fault detection method for an electromechanical braking system according to claim 2, characterized in that, The steps for obtaining the wear data corresponding to each predicted data point include: Acquire multiple historical braking data of the braking system, and the multiple historical braking data correspond one-to-one with multiple braking operations of the braking system; The ratio of the first temperature rise rate to the corresponding downforce for each historical braking data is determined as the temperature rise efficiency for each historical braking data. The first temperature rise rate is the average of multiple temperature change rates included in the corresponding historical braking data. The downforce is used to represent the force with which the brake pedal of the braking system is pressed under the corresponding braking operation. The sum of the temperature rise efficiency corresponding to each historical braking data point and the temperature rise efficiency corresponding to the next historical braking data point is determined as the temperature rise efficiency of that historical braking data point. The ratio of the temperature rise efficiency of each historical braking data point to the braking interval corresponding to that historical braking data point is determined as the wear degree of that historical braking data point, wherein the braking interval is the time interval between the corresponding historical braking data point and the next historical braking data point. The wear rate of each historical braking data point is summed to obtain the total wear rate; The product of the reciprocal of the total wear rate and the predicted temperature rise efficiency corresponding to each predicted data point is normalized to obtain the wear data corresponding to each predicted data point. The predicted temperature rise efficiency is the ratio of the predicted temperature rise rate to the predicted downforce included in the corresponding predicted data point.

4. The fault detection method for an electromechanical braking system according to claim 2, characterized in that, The steps for obtaining temperature disturbance data corresponding to each predicted data point include: Each predicted data point is input into a pre-trained random forest model to obtain multiple output values ​​from multiple random trees in the random forest model for the predicted data. The random forest model is used to predict the torque deviation of the motor corresponding to the input data based on the input data. The mean of multiple output values ​​corresponding to each predicted data is determined as the first deviation mean of each predicted data, and the absolute value of the mean of multiple temperature deviation values ​​corresponding to multiple predicted data is determined as the second deviation mean, wherein the temperature deviation value is the difference between the predicted temperature included in the corresponding predicted data and the preset reference temperature. Based on the temperature influence factor corresponding to the random forest model, the first mean deviation, the second mean deviation, and the temperature difference corresponding to each predicted data, temperature interference data corresponding to each predicted data is obtained. The temperature influence factor represents the degree of influence of temperature indicators on the model output among the various indicators included in the input data of the random forest model. The temperature influence factor is greater than or equal to a preset threshold. The temperature indicators include: temperature rise rate, the time integral corresponding to the difference between temperature and reference temperature, and the product of temperature and corresponding predicted current. The temperature difference is the difference between the mean temperature of the corresponding predicted data and the reference temperature. The mean temperature is the average of multiple predicted temperatures associated with the time segment between the first predicted data and the corresponding predicted data.

5. The fault detection method for an electromechanical braking system according to claim 1, characterized in that, The step of obtaining the signal anomaly index for each predicted data point based on its frequency deviation and duty cycle deviation includes: Calculate the product of frequency deviation and frequency coefficient for each predicted data point to obtain the frequency difference index for each predicted data point; calculate the product of duty cycle deviation and duty cycle coefficient for each predicted data point to obtain the duty cycle difference index for each predicted data point; wherein the sum of the frequency coefficient and the duty cycle coefficient is 1. The sum of the frequency difference index and the duty cycle difference index corresponding to each predicted data point is determined as the signal anomaly index for each predicted data point.

6. The fault detection method for an electromechanical braking system according to claim 5, characterized in that, Before obtaining the signal anomaly index of each predicted data based on the frequency deviation and duty cycle deviation corresponding to each predicted data, the method further includes: Acquire multiple historical data from a reference system, wherein the reference system and the braking system are of the same model. Based on the aforementioned historical data, multiple frequency deviation samples and multiple duty cycle deviation samples were obtained. Correlation analysis was performed on multiple frequency deviation samples and multiple expected torques corresponding to the multiple frequency deviation samples to obtain frequency correlation coefficients; and correlation analysis was performed on multiple duty cycle deviation samples and multiple expected torques corresponding to the multiple duty cycle samples to obtain duty cycle correlation coefficients. The frequency correlation coefficient and the duty cycle correlation coefficient are summed to obtain the sum value. The frequency coefficient is determined by the ratio of the frequency correlation coefficient to the sum of the coefficients, and the duty cycle coefficient is determined by the ratio of the duty cycle correlation coefficient to the sum of the coefficients.

7. The fault detection method for an electromechanical braking system according to claim 1, characterized in that, The analysis of the abnormal probability of the motor and transmission mechanism and the signal abnormality index of each predicted data point yields the fault probability value for each predicted data point, including: The product of the abnormal probability of the motor and transmission mechanism and the abnormal signal index for each predicted data is determined as the first probability value for each predicted data. Based on the abnormal probability of the motor and transmission mechanism and the abnormal signal index of each predicted data, a probability correction factor is obtained for each predicted data. The probability correction factor is used to represent the degree of difference between the corresponding predicted data and multiple fault data of the braking system. The ratio of the first probability value of each predicted data point to its corresponding probability correction factor is determined as the second probability value of each predicted data point. The second probability value of each predicted data point is normalized to obtain the failure probability value of each predicted data point.

8. The fault detection method for an electromechanical braking system according to claim 1, characterized in that, The analysis of the fault probability value of each predicted data point yields the fault detection results of the braking system within the predicted time period, including: Calculate the mean of multiple failure probability values ​​from multiple predicted data to obtain the first mean; Based on the fault probability value of each predicted data, the slope value of each predicted data is obtained, wherein the fault slope value is the instantaneous rate of change of the corresponding fault probability value. Calculate the mean of multiple slope values ​​for multiple predicted data to obtain a second mean; The normalized value of the product of the first mean and the second mean is determined as the target value; If the target value is greater than or equal to a preset probability threshold, output a fault detection result indicating that the braking system will fail within the predicted time period. If the target value is less than the probability threshold, the output indicates that the braking system will not malfunction within the predicted time period.

9. A fault detection system for an electromechanical braking system, characterized in that, The system includes: The acquisition module is used to acquire multiple prediction data of the braking system within the prediction time period; The overload analysis module is used to analyze the difference between the expected torque corresponding to each predicted data and the predicted torque included therein, and to obtain the abnormal probability of the motor and transmission mechanism for each predicted data. The expected torque is obtained after correcting the temperature error based on the corresponding ideal torque, and the ideal torque is calculated based on the predicted current included in the corresponding predicted data. The signal analysis module is used to obtain the signal anomaly index of each predicted data based on the frequency deviation and duty cycle deviation corresponding to each predicted data. The frequency deviation is used to represent the difference between the predicted frequency included in the predicted data and the corresponding rated frequency, and the duty cycle deviation is used to represent the difference between the predicted duty cycle included in the predicted data and the corresponding expected duty cycle. The fault probability analysis module is used to analyze the abnormal probability of motor and transmission mechanism and the signal abnormality index of each predicted data to obtain the fault probability value of each predicted data. The fault detection module is used to analyze the fault probability value of each predicted data point and obtain the fault detection results of the braking system within the predicted time period.

10. A fault detection device for an electromechanical braking system, characterized in that, The apparatus includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the fault detection method for the electromechanical braking system as described in any one of claims 1 to 8.

Images