Joint early warning method and system for braking risk of van

Abstract

The invention relates to the technical field of braking systems, solves the technical problems that the prior art is easily interfered by the environment, the accuracy of multi-factor analysis is not high, and the early warning accuracy is relatively low, and particularly relates to a combined early warning method and system for the braking risk of a van, and the method comprises the following steps: S1, collecting original data in the van, preprocessing the original data to obtain a dynamic calibration value; s2, a hybrid filtering value is calculated according to the dynamic calibration value, and a feature enhancement value used for achieving feature decoupling is calculated based on the hybrid filtering value, the fault logic can be clearly expressed through causal graph model construction, the fault position can be accurately determined in combination with a Bayesian network model, the fault path can be displayed, and the fault detection accuracy is improved. A clear and accurate analysis basis is provided for the retrospect decision of maintenance personnel, and the AR image display is combined, so that the fault type can be vividly displayed, and the maintenance efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of brake systems, and in particular to a joint early warning method and system for brake risks of passenger and freight vehicles. Background Art

[0002] The joint braking risk warning system is a comprehensive safety system that integrates multiple sensor technologies, data processing algorithms, and communication modules. It monitors the vehicle's driving status, driver's operating behavior, and external environmental information in real time, and uses advanced algorithmic models to fuse and analyze this multi-source, heterogeneous data to predict potential braking risks. Existing technologies typically use a single or limited number of sensors to monitor vehicle and driver information. For example, ABS sensors capture vehicle speed and wheel speed data, or steering wheel angle sensors assist in determining driving intent. Data processing methods often employ simple threshold-based judgment logic, such as triggering a warning when the rate of change of vehicle speed and brake pedal position exceeds a preset value. However, this approach is susceptible to environmental interference, such as inclement weather that affects sensor accuracy, and fails to fully reflect complex driving scenarios. Threshold algorithms lack the ability to analyze multi-factor correlations, resulting in inaccurate fault analysis. This results in low warning accuracy and makes it difficult to meet the complex and ever-changing needs of real-world driving. Summary of the Invention

[0003] In response to the shortcomings of the existing technology, the present invention provides a joint warning method and system for braking risks of passenger and freight vehicles, which solves the technical problems that the existing technology is easily affected by environmental interference, has low accuracy in multi-factor analysis, and has low warning accuracy. It achieves the purpose of avoiding environmental interference, improving accuracy through multi-factor analysis, and greatly improving warning accuracy.

[0004] To solve the above technical problems, the present invention provides the following technical solution: a joint early warning method for braking risks of passenger and freight vehicles, the method comprising the following steps: S1. Collect the original data inside the passenger and freight vehicle, and pre-process the original data to obtain dynamic calibration values ; S2, according to the dynamic calibration value Calculate the mixed filter value , and based on the mixed filter value Calculate feature enhancement values ​​for feature decoupling ; S3, based on feature enhancement value Divide the input set into categories , for the input set Gas path node temperature and gas path node humidity Perform calibration and obtain calibration data set ; S4. Based on the calibration data set Constructed to get thermal coupling value and wet coupling values Digital twin dynamic model , and based on the thermal coupling value and wet coupling values Calculate the pre-feature set used to provide input data for fault diagnosis ; S5. Based on the pre-feature set Construct a brake air circuit fault cause-effect diagram, calculate the judgment probability based on the brake air circuit fault cause-effect diagram, and obtain the primary fault result and secondary abnormal result based on the judgment probability; S6. Calculate the air pump path for judging the result of the pump air failure based on the brake air circuit fault cause and effect diagram and drying drum path ,generate AR animation based on the fault path results.

[0005] Furthermore, in step S1, the specific implementation steps are as follows: S11. Obtain the number N of multiple sensors inside the van and obtain the original data with a dimension of M and a time series length of t. ; S12, based on the highest frequency sampling rate of multiple sensors inside the passenger and freight vehicle Calculate the dynamically adjusted bandwidth value h using the following formula: in, Indicates the response period of the sensor; S13. Based on the original data inside the passenger and freight vehicle Sampling time Calculate the time difference weight matrix of the original data , the calculation formula is: in, Representation and sampling time Corresponding synchronization time; S14, according to the time difference weight matrix Calculate dynamic calibration values , the calculation formula is: in, Indicates environmental pressure The compensation value, Indicates the air pump speed compensation value.

[0006] Furthermore, in step S2, the specific implementation steps are as follows: S21, multiple dynamic calibration values Form a data sequence, build a rectangular window on the data sequence, and calculate the filter weight value within the rectangular window , the calculation formula is: in, represents the standard deviation of the data sequence, k represents the window offset, Represents the window width parameter; S22, according to the filter weight value Computes a mixed filter value for filtering a data sequence , the calculation formula is: in, represents the dynamic calibration value after the window moves k; S23, according to the mixed filter value Computing 4D tensors , the calculation formula is: Among them, S represents the mixed filter value Corresponding voiceprint spectrum characteristics; S24. Convert the four-dimensional tensor Perform tensor decomposition to obtain multiple principal component values and , the calculation formula is: in, and Represent the principal component values ​​in the first r component values, represents the outer product operation, Indicates the decomposition weight corresponding to each component value; S25, based on multiple component values Calculate feature enhancement values ​​for feature decoupling , the calculation formula is: in, represents the Kronecker product.

[0007] Furthermore, in step S3, the specific implementation steps are as follows: S31, feature enhancement value Generate input sets based on the type of original data , the expression is: in, Indicates the key gas path node temperature of the brake system after feature enhancement, The humidity of the key gas circuit nodes of the brake system after feature enhancement, Indicates the moisture content of the drying tube gas path after the characteristics are enhanced, represents the drying drum noise spectrum after feature enhancement, represents the air pump speed after feature enhancement, Indicates environmental pressure after feature enhancement; S32, according to the input set Calculate temperature parameter compensation value and humidity parameter compensation value , the calculation formula is: in, Indicates the rated speed of the air pump. Indicates the ambient standard atmospheric pressure, They represent the coupling compensation coefficients of the air pump speed and ambient pressure to temperature and humidity respectively; S33, repeat step S32 to compensate all temperature parameters and humidity parameter compensation value The calculation is completed and the compensation data set is generated ,in, ; S34, based on temperature parameter compensation value Calculate the temperature standard deviation value for each temperature sensor , the calculation formula is: Where S represents the temperature parameter compensation value the number of Indicates the average temperature of multiple sensors; S35, based on humidity parameter compensation value Calculate the humidity standard deviation value for each humidity sensor , the calculation formula is: Where J represents the humidity parameter compensation value the number of Indicates the average humidity value of multiple sensors; S36, according to the temperature standard deviation Perform drift correction on the temperature sensor; like , then the temperature sensor has data drift, and the temperature sensor needs to be drift corrected; like , the temperature sensor is normal and ends, and the corrected temperature set is obtained ; S37, according to the humidity standard deviation Drift correction for humidity sensor; like , then the humidity sensor has data drift, and the humidity sensor needs to be drift corrected; like , the temperature sensor is normal and ends, and the corrected humidity set is obtained ; S38, set the correct humidity Elements in the drying tube and moisture content in the air path Perform feature fusion to obtain humidity fusion value , the calculation formula is: in, and Respectively represent the fusion weight value; S39, based on the correction temperature set and corrected humidity set Get the calibration dataset .

[0008] Furthermore, in step S4, the specific implementation steps are as follows: S41. Based on the calibration data set Dryer drum noise spectrum Calculate the time domain feature value used to extract time domain features , the calculation formula is: in, Indicates the time domain characteristic peak, N represents the drying drum noise spectrum quantity; S42, based on the calibration data set Dryer drum noise spectrum Calculate frequency domain eigenvalues ​​for extracting frequency domain features , the time domain eigenvalue and frequency domain eigenvalues Synthesize into voiceprint feature vector , the calculation formula is: Wherein, FFT stands for Fast Fourier Transform; S43, according to the humidity fusion value and temperature parameter compensation values Calculate the temperature gradient values ​​separately and humidity gradient values , the calculation formula is: in, and Respectively represent and independent variables; S44, the temperature gradient value and humidity gradient values Composing the gradient matrix ; S45, according to the temperature gradient value and humidity gradient values Building a digital twin dynamic model , the expression is: in, represents the net outflow of fluid heat flux density at the sensor position, D represents the humidity diffusion coefficient, v represents the convection velocity of the fluid, Indicates the heating power of the air pump. Indicates the dehumidification rate of the drying cylinder, represents the thermal coupling value, Indicates the wet coupling value; S46. Obtain historical data of the sensor and calculate the temperature residual threshold of the historical data and humidity residual threshold , the calculation formula is: in, Represent the residual means of temperature and humidity of historical data respectively, represents the confidence coefficient, Represent the residual standard deviation of temperature and humidity of historical data respectively; S47, based on the temperature residual threshold Mark nodes with abnormal temperature; like , then the node corresponding to the temperature data is a temperature abnormality node and is marked; like , then the node corresponding to the temperature data is a normal temperature node and ends; S48, based on the humidity residual threshold Mark the humidity anomaly nodes and merge the temperature anomaly nodes and humidity anomaly nodes into the anomaly node set B; like , then the node corresponding to the humidity data is the humidity abnormal node and is marked; like , then the node corresponding to the humidity data is a normal humidity node and ends; S49, the voiceprint feature vector , gradient matrix , thermal coupling value , wet coupling value Merge with abnormal node set B into pre-feature set .

[0009] Furthermore, in step S5, the specific implementation steps are as follows: S51, based on the pre-feature set The air pump node, drying cylinder node and key air path location node are determined respectively, and a brake air path fault causal graph is constructed based on the air pump node, drying cylinder node and key air path location node. The key air path location node and the air pump node and drying cylinder node are used to construct the propagation path. ; S52. Obtain the cumulative working time of the air pump , based on the accumulated working time Calculate the prior probability of failure of the gas pump node , the calculation formula is: in, and Respectively represent the shape parameters of the air pump; S53. Obtain voiceprint feature vector The time domain eigenvalues ​​in , according to the time domain eigenvalue Calculate the prior probability of tube obstruction at the drying tube node , the calculation formula is: in, Represents the time domain eigenvalue The standard deviation of Represents the time domain eigenvalue The average value of S54. According to the prior probability of failure and the barrier prior probability Calculate the judgment probability, including the post-pump probability of the air pump and the post-drying probability of the drying drum ; S55, according to the probability after pumping and the probability after the tube Determine the fault and generate a primary fault result, wherein the primary fault result includes air pump fault information and drying cylinder fault information; like , then generate air pump fault information; like , then generate drying cylinder fault information; S56, according to the humidity gradient value Calculating humidity gradient threshold , the calculation formula is: Where M represents the humidity gradient value the total number of S57, according to the humidity gradient threshold Determine gas path faults and generate secondary abnormal results; like , then there is a risk of water accumulation in the gas path, a secondary abnormal result is generated and the process ends; like , the gas path is normal and ends.

[0010] Furthermore, the post-pump probability and the probability after the tube The calculation formula is: in, represents the likelihood probability of the air pump speed, represents the likelihood probability of the temperature gradient, represents the joint probability of the air pump speed and temperature gradient, represents the likelihood probability of time domain features, represents the likelihood probability of the humidity gradient, Represents the joint probability of time domain features and humidity gradient.

[0011] Furthermore, in step S6, the specific implementation steps are as follows: S61. Define the air pump node as node A, define the drying cylinder node as node B, define the key air path location node as node C, and calculate the air pump path respectively. and drying drum path ; S62, according to the air pump path and drying drum path Determine the most likely fault path and generate a fault path result, wherein the fault path result includes a pump gas fault result and a cylinder gas fault result; like , then the air pump path For the fault path, generate the pump gas fault result and end; like , then the drying tube path For the fault path, generate the cylinder gas fault result and end; S63. Send the primary fault result and the secondary abnormality result to the early warning center, generate an AR three-dimensional disassembly animation based on the fault path result, and highlight the fault path and faulty components in the early warning system.

[0012] Furthermore, the air pump path and drying drum path The calculation formula is: Where F and G represent the number of AC paths and BC paths, respectively. represents the probability that node C will fail if node A fails, It represents the probability that node C will fail if node B fails.

[0013] The technical solution also provides a system for the above-mentioned passenger and freight vehicle braking risk joint early warning method, the system comprising: The preprocessing module is used to collect the raw data inside the passenger and freight vehicles and preprocess the raw data to obtain dynamic calibration values ; Feature enhancement module for dynamically calibrating Calculate the mixed filter value , and based on the mixed filter value Calculate feature enhancement values ​​for feature decoupling ; Calibration module for feature-based enhancement values Divide the input set into categories , for the input set Gas path node temperature and gas path node humidity Perform calibration and obtain a calibration data set to make the data more accurate ; Digital Twin module for calibrating data sets Constructed to get thermal coupling value and wet coupling values Digital twin dynamic model , and based on the thermal coupling value and wet coupling values Calculate the pre-feature set used to provide reference for fault judgment ; Fault causality module, used to Construct a brake air circuit fault cause-effect diagram, calculate the judgment probability based on the brake air circuit fault cause-effect diagram, and obtain the primary fault result and secondary abnormal result based on the judgment probability; AR generation module, used to calculate the air pump path for judging the results of pump air failure based on the brake air circuit fault cause and effect diagram and drying drum path ,generate AR animation based on the fault path results.

[0014] By means of the above technical solution, the present invention provides a combined warning method and system for braking risk of passenger and freight vehicles, which has at least the following beneficial effects: 1. By calculating the time difference weight matrix of the original data, the present invention can solve the problem of time asynchrony of multi-source sensor data, dynamically compensate for environmental interference, significantly reduce calibration errors, and significantly improve the signal-to-noise ratio. It retains the sensor's fluctuation characteristics while suppressing the high-frequency noise of the microwave sensor, improves feature dimension compression, and retains most of the information, providing structured input for subsequent causal diagrams.

[0015] 2. The present invention uses temperature and humidity sensors in conjunction with microwave humidity detection and system voiceprint analysis to achieve full-dimensional state perception of the gas path, which can significantly increase the detection range in practice. The millisecond-level update of the digital twin model can greatly improve the timeliness of fault prediction. The self-calibration compensation model of the calibration data set can greatly reduce the temperature and humidity measurement errors, thereby improving data accuracy.

[0016] 3. The present invention is constructed through a causal graph model, which can clearly express the fault logic. Combined with the Bayesian network model, it can accurately determine the fault location and display the fault path, providing maintenance personnel with a clear and accurate analysis basis for tracing decisions. Combined with AR image display, it can vividly display the type of fault and improve maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 This is a flow chart of a joint early warning method for braking risks of passenger and freight vehicles according to the present invention; Figure 2 This is a structural block diagram of a passenger and freight vehicle braking risk joint warning system according to the present invention.

[0018] In the figure: 1. Preprocessing module; 2. Feature enhancement module; 3. Calibration module; 4. Digital twin module; 5. Fault cause and effect module; 6. AR generation module. DETAILED DESCRIPTION

[0019] To make the above-mentioned objectives, features, and advantages of the present invention more clearly understood, the present invention is further described below in detail with reference to the accompanying drawings and specific embodiments. This will enable a full understanding of how this application uses technical means to solve technical problems and achieve technical effects, and to implement the invention accordingly.

[0020] Due to the technical problems that the existing technology is easily affected by environmental interference, has low accuracy in multi-factor analysis, and has low early warning accuracy, this embodiment proposes a joint early warning method for passenger and freight vehicle braking risks, such as Figure 1 As shown, it is possible to avoid environmental interference, improve accuracy through multi-factor analysis, and significantly improve the accuracy of early warning. The method includes the following steps: S1. Collect the original data inside the passenger and freight vehicle, and pre-process the original data to obtain dynamic calibration values In the process of acquiring raw data, since the sensors of passenger and freight vehicles are in the external environment, the noise and vibration caused by the vehicle driving can easily make the raw data inaccurate. To solve this problem, the specific methods used are as follows: S11. Obtain the number N of multiple sensors inside the van and obtain the original data with a dimension of M and a time series length of t. ; The original data includes the temperature of the key gas circuit nodes of the brake system and humidity , moisture content of the drying cylinder gas path , Drying drum noise spectrum , air pump speed and environmental pressure ; S12, based on the highest frequency sampling rate of multiple sensors inside the passenger and freight vehicle Calculate the dynamically adjusted bandwidth value h using the following formula: in, Indicates the response period of the sensor; S13. Based on the original data inside the passenger and freight vehicle Sampling time Calculate the time difference weight matrix of the original data , the calculation formula is: in, Representation and sampling time Corresponding synchronization time; the synchronization time and sampling time here will produce deviations, affecting the real-time nature of the data, so it is easy to produce inaccuracies when calculating data. In order to avoid this situation, the time difference weight matrix is ​​calculated based on this deviation. to reduce the size of the deviation.

[0021] S14, according to the time difference weight matrix Calculate dynamic calibration values , the calculation formula is: in, Indicates environmental pressure The compensation value, Indicates the air pump speed The compensation value can be calculated by obtaining the compensation function through offline calibration experiments. Offline calibration experiments are a commonly used method for obtaining compensation functions and will not be described in detail here. By calculating the time difference weight matrix of the original data, the time asynchrony problem of multi-source sensor data can be solved, environmental interference can be dynamically compensated, calibration errors can be greatly reduced, and the signal-to-noise ratio can be greatly improved. The fluctuation characteristics of the sensor are retained while suppressing the high-frequency noise of the microwave sensor. The feature dimension compression is improved while retaining most of the information, providing structured input for the subsequent causal diagram.

[0022] S2, according to the dynamic calibration value Calculate the mixed filter value , and based on the mixed filter value Calculate feature enhancement values ​​for feature decoupling After preprocessing the original data, the characteristics of the data are still not obvious. To solve this problem, this embodiment proposes a more detailed implementation method as follows: S21, multiple dynamic calibration values Form a data sequence, build a rectangular window on the data sequence, and calculate the filter weight value within the rectangular window , the calculation formula is: in, represents the standard deviation of the data sequence, k represents the window offset, Represents the window width parameter; the window width parameter is obtained by rounding the fundamental frequency ratio in the data. The standard deviation and fundamental frequency ratio rounding methods are commonly used data processing methods and are not described in detail here.

[0023] S22, according to the filter weight value Computes a mixed filter value for filtering a data sequence , the calculation formula is: in, represents the dynamic calibration value after the window moves k; S23, according to the mixed filter value Computing 4D tensors , the calculation formula is: Among them, S represents the mixed filter value The corresponding voiceprint spectrum characteristics; the voiceprint spectrum characteristics can be obtained by comparing the drying drum noise spectrum Perform video transformation and acquisition.

[0024] S24. Convert the four-dimensional tensor Perform tensor decomposition to obtain multiple principal component values and , the calculation formula is: in, and Represent the principal component values ​​in the first r component values, represents the outer product operation, Represents the decomposition weight corresponding to each component value; the decomposition weight can be obtained by combining the alternating least squares method with the alternating optimization step. The least squares method and alternating optimization are common methods for obtaining the decomposition weight. r can be obtained by the four-dimensional tensor The variance contribution rate is greater than or equal to 95%. Variance contribution rate is a common data processing method and will not be described in detail here.

[0025] S25, based on multiple component values Calculate feature enhancement values ​​for feature decoupling , the calculation formula is: in, It represents the Kronecker product and eliminates the deviation of multi-source heterogeneous data through a spatiotemporal-physical joint calibration mechanism. At the same time, it takes into account the requirements of nanometer-level sensitivity and rapid response through modal-aware filtering technology. It can also achieve feature decoupling through a tensor fusion architecture, improving the separability of fault features by 32% compared with the traditional PCA method.

[0026] S3, based on feature enhancement value Divide the input set into categories , for the input set Gas path node temperature and gas path node humidity Perform calibration and obtain calibration data set Based on step S2, the data needs to be further calibrated to make the results of subsequent steps more accurate. To solve this problem, the detailed implementation steps are as follows: S31, feature enhancement value Generate input sets based on the type of original data , the expression is: in, Indicates the key gas path node temperature of the brake system after feature enhancement, The humidity of the key gas circuit nodes of the brake system after feature enhancement, Indicates the moisture content of the drying tube gas path after the characteristics are enhanced, represents the drying drum noise spectrum after feature enhancement, represents the air pump speed after feature enhancement, Indicates environmental pressure after feature enhancement; S32, according to the input set Calculate temperature parameter compensation value and humidity parameter compensation value , the calculation formula is: in, Indicates the rated speed of the air pump. Indicates the ambient standard atmospheric pressure, They respectively represent the coupling compensation coefficients of the air pump speed and the ambient pressure to the temperature and humidity; the coupling compensation coefficient can be obtained by the step response method. The step response method applies a step input to the system, measures the output response, and infers the compensation coefficient through the amplitude or phase of the coupling interference. The step response method is a commonly used method for obtaining the coupling compensation coefficient and will not be elaborated here.

[0027] S33, repeat step S32 to compensate all temperature parameters and humidity parameter compensation value The calculation is completed and the compensation data set is generated ,in, ; S34, based on temperature parameter compensation value Calculate the temperature standard deviation value for each temperature sensor , the calculation formula is: Where S represents the temperature parameter compensation value the number of Indicates the average temperature of multiple sensors; S35, based on humidity parameter compensation value Calculate the humidity standard deviation value for each humidity sensor , the calculation formula is: Where J represents the humidity parameter compensation value the number of Indicates the average humidity value of multiple sensors; S36, according to the temperature standard deviation Perform drift correction on the temperature sensor; like , then the temperature sensor has data drift, and the temperature sensor needs to be drift corrected; like , the temperature sensor is normal and ends, and the corrected temperature set is obtained ; S37, according to the humidity standard deviation Drift correction for humidity sensor; like , then the humidity sensor has data drift, and the humidity sensor needs to be drift corrected; like , the temperature sensor is normal and ends, and the corrected humidity set is obtained ; S38, set the correct humidity Elements in the drying tube and moisture content in the air path Perform feature fusion to obtain humidity fusion value , the calculation formula is: in, and They represent fusion weight values ​​respectively; dynamic optimization of fusion weight values ​​through Kalman filtering is a common weight acquisition method, which will not be described here.

[0028] S39, based on the correction temperature set and corrected humidity set Get the calibration dataset ,in, By combining temperature and humidity sensors with microwave humidity detection and system voiceprint analysis, it is possible to cover the full-dimensional state perception of the gas path, which can greatly increase the detection range in practice. The millisecond-level update of the digital twin model can greatly improve the timeliness of fault prediction. The self-calibration compensation model of the calibration data set can greatly reduce the temperature and humidity measurement errors, thereby improving data accuracy.

[0029] S4. Based on the calibration data set Constructed to get thermal coupling value and wet coupling values Digital twin dynamic model , and based on the thermal coupling value and wet coupling values Calculate pre-feature set ; In the obtained calibration data set Due to data volatility and external interference, there are still some abnormal data. To solve this problem, the detailed implementation steps are as follows: S41. Based on the calibration data set Dryer drum noise spectrum Calculate the time domain feature value used to extract time domain features , the calculation formula is: in, Indicates the time domain characteristic peak, N represents the drying drum noise spectrum quantity; S42, based on the calibration data set Dryer drum noise spectrum Calculate frequency domain eigenvalues ​​for extracting frequency domain features , the time domain eigenvalue and frequency domain eigenvalues Synthesize into voiceprint feature vector , the calculation formula is: Among them, FFT stands for Fast Fourier Transform; Fast Fourier Transform is an efficient algorithm used to convert time domain signals into frequency domain signals to obtain their frequency domain eigenvalues. Fast Fourier Transform is a common method for obtaining frequency domain features and will not be described in detail here.

[0030] S43, according to the humidity fusion value and temperature parameter compensation values Calculate the temperature gradient values ​​separately and humidity gradient values , the calculation formula is: in, and Respectively represent and independent variables; S44, the temperature gradient value and humidity gradient values Composing the gradient matrix ,in, ; S45, according to the temperature gradient value and humidity gradient values Building a digital twin dynamic model , the expression is: in, represents the net outflow of fluid heat flux density at the sensor position, D represents the humidity diffusion coefficient, v represents the convection velocity of the fluid, Indicates the heating power of the air pump. Indicates the dehumidification rate of the drying cylinder, represents the thermal coupling value, It represents the wet coupling value; the humidity diffusion coefficient, the convection velocity of the fluid, the heating power of the air pump, and the dehumidification rate of the drying cylinder can all be obtained through sensors. The boundary conditions are set according to the air inlet and outlet of the air path. In practice, the rated value of the temperature and humidity at the air inlet is generally set as the boundary condition, and the boundary condition at the air outlet is that the partial derivative of the ratio of the temperature and humidity to the normal vector of the air outlet boundary is equal to 0.

[0031] S46. Obtain historical data of the sensor and calculate the temperature residual threshold of the historical data and humidity residual threshold , the calculation formula is: in, Represent the residual means of temperature and humidity of historical data respectively, represents the confidence coefficient, Represents the residual standard deviation of temperature and humidity of historical data respectively; in practice, the confidence coefficient is generally set to The numbers between correspond to the 95% to 99.7% confidence interval. The calculation of the residual mean and residual standard deviation is a commonly used calculation method and will not be described in detail here.

[0032] S47, based on the temperature residual threshold Mark nodes with abnormal temperature; like , then the node corresponding to the temperature data is a temperature abnormality node and is marked; like , then the node corresponding to the temperature data is a normal temperature node and ends; S48, based on the humidity residual threshold Mark the humidity anomaly nodes and merge the temperature anomaly nodes and humidity anomaly nodes into the anomaly node set B; like , then the node corresponding to the humidity data is the humidity abnormal node and is marked; like , then the node corresponding to the humidity data is a normal humidity node and ends; S49, the voiceprint feature vector , gradient matrix , thermal coupling value , wet coupling value Merge with abnormal node set B into pre-feature set ,in, By combining temperature and humidity sensors with microwave humidity detection and system voiceprint analysis, it is possible to cover the full-dimensional state perception of the gas path, which can greatly increase the detection range in practice. The millisecond-level update of the digital twin model can greatly improve the timeliness of fault prediction. The self-calibration compensation model of the calibration data set can greatly reduce the temperature and humidity measurement errors, thereby improving data accuracy.

[0033] S5. Based on the pre-feature set Construct a brake air circuit fault causal diagram, calculate the judgment probability based on the brake air circuit fault causal diagram, and obtain the primary fault result and secondary abnormal result based on the judgment probability, among which the judgment probability includes the post-pump probability , post-tube probability and humidity gradient threshold ; In the pre-feature set On this basis, it is necessary to quickly find the fault type and path through the brake air circuit fault cause-effect diagram. To solve this problem, this embodiment proposes a more detailed implementation method as follows: S51, increase the air pump speed Air pump heating power As the air pump node of the brake air circuit fault causal graph, the voiceprint feature vector The frequency domain eigenvalues ​​in and dehumidification rate of the drying drum As the drying tube node of the brake air circuit fault causal graph, the pre-feature set The gradient matrix in , thermal coupling value and wet coupling values As the key gas path location node, a brake gas path fault causal graph is constructed based on the air pump node, the drying cylinder node and the key gas path location node. The connection between the nodes is the propagation path. ; S52. Obtain the cumulative working time of the air pump , based on the accumulated working time Calculate the prior probability of failure of the gas pump node , the calculation formula is: in, and They respectively represent the shape parameters of the air pump; the shape parameters of the air pump can be obtained by staff through direct measurement. The direct measurement method is a common method for obtaining shape parameters and will not be described here.

[0034] S53. Obtain voiceprint feature vector The time domain eigenvalues ​​in , according to the time domain eigenvalue Calculate the prior probability of tube obstruction at the drying tube node , the calculation formula is: in, Represents the time domain eigenvalue The standard deviation of Represents the time domain eigenvalue The average value of S54. According to the prior probability of failure and the barrier prior probability Calculate the post-pump probability of the air pump separately and the post-drying probability of the drying drum , the calculation formula is: in, represents the likelihood probability of the air pump speed, represents the likelihood probability of the temperature gradient, represents the joint probability of the air pump speed and temperature gradient, represents the likelihood probability of time domain features, represents the likelihood probability of the humidity gradient, Represents the joint probability of time domain features and humidity gradient; likelihood probability and joint probability are common methods for calculating posterior probability, which will not be described here.

[0035] S55, according to the probability after pumping and the probability after the tube Determine the fault and generate a primary fault result, wherein the primary fault result includes air pump fault information and drying cylinder fault information; like , then generate air pump fault information; like , then generate drying cylinder fault information; S56, according to the humidity gradient value Calculating humidity gradient threshold , the calculation formula is: Where M represents the humidity gradient value the total number of S57, according to the humidity gradient threshold Determine gas path faults and generate secondary abnormal results; like , then there is a risk of water accumulation in the gas path, a secondary abnormal result is generated and the process ends; like , the gas path is normal and ends. Through the construction of the causal graph model, the fault logic can be clearly expressed. Combined with the Bayesian network model, the fault location can be accurately determined, and the fault path can be displayed, providing maintenance personnel with a clear and accurate analysis basis for tracing decisions. Combined with AR image display, it can vividly display the type of fault and improve maintenance efficiency.

[0036] S6. Calculate the air pump path based on the brake air circuit fault cause-effect diagram and drying drum path , according to the air pump path and drying drum path The pump failure result is obtained and an AR 3D disassembly animation is generated. Based on step S5, after finding the possible faulty device, it is necessary to determine the fault path and generate a vivid image animation. To solve this problem, the specific implementation method of this embodiment is as follows: S61. Define the air pump node as node A, define the drying cylinder node as node B, define the key air path location node as node C, and calculate the air pump path respectively. and drying drum path , the calculation formula is: Where F and G represent the number of AC paths and BC paths, respectively. represents the probability that node C will fail if node A fails, Indicates the probability that node C will also fail if node B fails. Since the path formed in the brake system is generally from the air pump to the air circuit or from the drying cylinder to the air circuit, it is necessary to know the two paths and compare them for AR display.

[0037] S62, according to the air pump path and drying drum path Determine the most likely fault path and generate a fault path result, wherein the fault path result includes a pump gas fault result and a cylinder gas fault result; like , then the air pump path For the fault path, generate the pump gas fault result and end; like , then the drying tube path For the fault path, generate the cylinder gas fault result and end; S63. Generate AR 3D disassembly animation based on the fault path results, highlight the fault path and faulty components in the early warning system, and build a causal graph model to clearly express the fault logic. Combined with the Bayesian network model, it can accurately determine the fault location and display the fault path, providing maintenance personnel with a clear and accurate analysis basis for tracing decisions. Combined with AR image display, it can vividly display the type of fault, thereby improving maintenance efficiency.

[0038] Since the existing technology is susceptible to environmental interference, the accuracy of multi-factor analysis is not high, and the warning accuracy is low, this embodiment also proposes a passenger and freight vehicle braking risk joint warning system, which can avoid environmental interference, improve accuracy through multi-factor analysis, and greatly improve the warning accuracy, such as Figure 2 As shown, the monitoring system includes a preprocessing module 1, a feature enhancement module 2, a calibration module 3, a digital twin module 4, a fault cause and effect module 5 and an AR generation module 6.

[0039] Preprocessing module 1 is used to collect the original data inside the passenger and freight vehicle, and preprocess the original data to obtain dynamic calibration values ; Feature enhancement module 2, used to adjust the dynamic calibration value Calculate the mixed filter value , and based on the mixed filter value Calculate feature enhancement values ​​for feature decoupling ; Calibration module 3, for enhancing the value based on the feature Divide the input set into categories , for the input set Gas path node temperature and gas path node humidity Perform calibration and obtain a calibration data set to make the data more accurate ; Digital twin module 4, used to Constructed to get thermal coupling value and wet coupling values Digital twin dynamic model , and based on the thermal coupling value and wet coupling values Calculate the pre-feature set used to provide reference for fault judgment ; Fault causality module 5, used to Construct a brake air circuit fault causal graph, calculate the judgment probability based on the brake air circuit fault causal graph, and obtain the primary fault result and secondary abnormal result based on the judgment probability; AR generation module 6 is used to calculate the air pump path for judging the pump air fault result based on the brake air circuit fault causal graph and drying drum path ,generate AR animation based on the fault path results.

[0040] Those skilled in the art will appreciate that all or part of the steps in the above-mentioned embodiment methods can be accomplished by instructing the relevant hardware through a program. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0041] Each embodiment in this specification is described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the same or similar parts between the embodiments. For the above embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For relevant parts, refer to the partial description of the method embodiments.

[0042] The above embodiments provide a detailed introduction to the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A joint early warning method for braking risk of passenger and freight vehicles, characterized in that: The method comprises the following steps: S1. Collect the original data inside the passenger and freight vehicle, and pre-process the original data to obtain dynamic calibration values ; S2, according to the dynamic calibration value Calculate the mixed filter value , and based on the mixed filter value Calculate feature enhancement values ​​for feature decoupling ; S3, based on feature enhancement value Divide the input set into categories , for the input set Gas path node temperature and gas path node humidity Perform calibration and obtain calibration data set ; S4. Based on the calibration data set Constructed to get thermal coupling value and wet coupling values Digital twin dynamic model , and based on the thermal coupling value and wet coupling values Calculate the pre-feature set used to provide input data for fault diagnosis ; S5. Based on the pre-feature set Construct a brake air circuit fault cause-effect diagram, calculate the judgment probability based on the brake air circuit fault cause-effect diagram, and obtain the primary fault result and secondary abnormal result based on the judgment probability; S6. Calculate the air pump path for judging the result of the pump air failure based on the brake air circuit fault cause and effect diagram and drying drum path ,generate AR animation based on the fault path results.

2. The joint early warning method according to claim 1, characterized in that: In step S1, the specific implementation steps are as follows: S11. Obtain the number N of multiple sensors inside the van and obtain the original data with a dimension of M and a time series length of t. ; S12, based on the highest frequency sampling rate of multiple sensors inside the passenger and freight vehicle Calculate the dynamically adjusted bandwidth value h using the following formula: in, Indicates the response period of the sensor; S13. Based on the original data inside the passenger and freight vehicle Sampling time Calculate the time difference weight matrix of the original data , the calculation formula is: in, Representation and sampling time Corresponding synchronization time; S14, according to the time difference weight matrix Calculate dynamic calibration values , the calculation formula is: in, Indicates environmental pressure The compensation value, Indicates the air pump speed compensation value.

3. The joint early warning method according to claim 1, characterized in that: In step S2, the specific implementation steps are as follows: S21, multiple dynamic calibration values Form a data sequence, build a rectangular window on the data sequence, and calculate the filter weight value within the rectangular window , the calculation formula is: in, represents the standard deviation of the data sequence, k represents the window offset, Represents the window width parameter; S22, according to the filter weight value Computes a mixed filter value for filtering a data sequence , the calculation formula is: in, represents the dynamic calibration value after the window moves k; S23, according to the mixed filter value Computing 4D tensors , the calculation formula is: Among them, S represents the mixed filter value Corresponding voiceprint spectrum characteristics; S24. Convert the four-dimensional tensor Perform tensor decomposition to obtain multiple principal component values and , the calculation formula is: in, Represent the principal component values ​​in the first r component values, represents the outer product operation, Indicates the decomposition weight corresponding to each component value; S25, based on multiple component values and Calculate feature enhancement values ​​for feature decoupling , the calculation formula is: in, represents the Kronecker product.

4. The joint early warning method according to claim 1, characterized in that: In step S3, the specific implementation steps are as follows: S31, feature enhancement value Generate input sets based on the type of original data , the expression is: in, Indicates the key gas path node temperature of the brake system after feature enhancement, The humidity of the key gas circuit nodes of the brake system after feature enhancement, Indicates the moisture content of the drying tube gas path after the characteristics are enhanced, represents the drying drum noise spectrum after feature enhancement, Indicates the air pump speed after feature enhancement, Indicates environmental pressure after feature enhancement; S32, according to the input set Calculate temperature parameter compensation value and humidity parameter compensation value , the calculation formula is: in, Indicates the rated speed of the air pump. Indicates the ambient standard atmospheric pressure, They represent the coupling compensation coefficients of the air pump speed and ambient pressure to temperature and humidity respectively; S33, repeat step S32 to compensate all temperature parameters and humidity parameter compensation value The calculation is completed and the compensation data set is generated ,in, ; S34, based on temperature parameter compensation value Calculate the temperature standard deviation value for each temperature sensor , the calculation formula is: Where S represents the temperature parameter compensation value the number of Indicates the average temperature of multiple sensors; S35, based on humidity parameter compensation value Calculate the humidity standard deviation value for each humidity sensor , the calculation formula is: Where J represents the humidity parameter compensation value the number of Indicates the average humidity of multiple sensors; S36, according to the temperature standard deviation Perform drift correction on the temperature sensor; like , then the temperature sensor has data drift, and the temperature sensor needs to be drift corrected; like , the temperature sensor is normal and ends, and the corrected temperature set is obtained ; S37, according to the humidity standard deviation Drift correction for humidity sensor; like , then the humidity sensor has data drift, and the humidity sensor needs to be drift corrected; like , the temperature sensor is normal and ends, and the corrected humidity set is obtained ; S38, set the correct humidity Elements in the drying tube and moisture content in the air path Perform feature fusion to obtain humidity fusion value , the calculation formula is: in, and Respectively represent the fusion weight value; S39, based on the correction temperature set and corrected humidity set Get the calibration dataset .

5. The joint early warning method according to claim 1, characterized in that: In step S4, the specific implementation steps are as follows: S41. Based on the calibration data set Dryer drum noise spectrum Calculate the time domain feature value used to extract time domain features , the calculation formula is: in, Indicates the time domain characteristic peak, N represents the drying drum noise spectrum quantity; S42, based on the calibration data set Dryer drum noise spectrum Calculate the frequency domain eigenvalues ​​used to extract frequency domain features , the time domain eigenvalue and frequency domain eigenvalues Synthesize into voiceprint feature vector , the calculation formula is: Wherein, FFT stands for Fast Fourier Transform; S43, according to the humidity fusion value and temperature parameter compensation values Calculate the temperature gradient values ​​separately and humidity gradient values , the calculation formula is: in, and Respectively represent and independent variables; S44, the temperature gradient value and humidity gradient values Composing the gradient matrix ; S45, according to the temperature gradient value and humidity gradient values Building a digital twin dynamic model , the expression is: in, represents the net outflow of fluid heat flux density at the sensor position, D represents the humidity diffusion coefficient, v represents the convection velocity of the fluid, Indicates the heating power of the air pump. Indicates the dehumidification rate of the drying cylinder, represents the thermal coupling value, Indicates the wet coupling value; S46. Obtain historical data of the sensor and calculate the temperature residual threshold of the historical data and humidity residual threshold , the calculation formula is: in, Represent the residual means of temperature and humidity of historical data respectively, represents the confidence coefficient, Represent the residual standard deviation of temperature and humidity of historical data respectively; S47, based on the temperature residual threshold Mark nodes with abnormal temperature; like , then the node corresponding to the temperature data is a temperature abnormality node and is marked; like , then the node corresponding to the temperature data is a normal temperature node and ends; S48, based on the humidity residual threshold Mark the humidity anomaly nodes and merge the temperature anomaly nodes and humidity anomaly nodes into the anomaly node set B; like , then the node corresponding to the humidity data is the humidity abnormal node and is marked; like , then the node corresponding to the humidity data is a normal humidity node and ends; S49, the voiceprint feature vector , gradient matrix , thermal coupling value , wet coupling value Merge with abnormal node set B into pre-feature set .

6. The joint early warning method according to claim 1, characterized in that: In step S5, the specific implementation steps are as follows: S51, based on the pre-feature set The air pump node, drying cylinder node and key air path location node are determined respectively, and a brake air path fault causal graph is constructed based on the air pump node, drying cylinder node and key air path location node. The key air path location node and the air pump node and drying cylinder node are used to construct a propagation path. ; S52. Obtain the cumulative working time of the air pump , based on the accumulated working time Calculate the prior probability of failure of the gas pump node , the calculation formula is: in, and Respectively represent the shape parameters of the air pump; S53. Obtain voiceprint feature vector The time domain eigenvalues ​​in , according to the time domain eigenvalue Calculate the prior probability of tube obstruction at the drying tube node , the calculation formula is: in, Represents the time domain eigenvalue The standard deviation of Represents the time domain eigenvalue The average value of S54. According to the prior probability of failure and the barrier prior probability Calculate the judgment probability, including the post-pump probability of the air pump and the post-drying probability of the drying drum ; S55, according to the probability after pumping and the probability after the tube Determine the fault and generate a primary fault result, wherein the primary fault result includes air pump fault information and drying cylinder fault information; like , then generate air pump fault information; like , then generate drying cylinder fault information; S56, according to the humidity gradient value Calculating humidity gradient threshold , the calculation formula is: Where M represents the humidity gradient value the total number of S57, according to the humidity gradient threshold Determine gas path faults and generate secondary abnormal results; like , then there is a risk of water accumulation in the gas path, a secondary abnormal result is generated and the process ends; like , the gas path is normal and ends.

7. The joint early warning method according to claim 6, characterized in that: The post-pump probability and the probability after the tube The calculation formula is: in, represents the likelihood probability of the air pump speed, represents the likelihood probability of the temperature gradient, represents the joint probability of the air pump speed and temperature gradient, represents the likelihood probability of time domain features, represents the likelihood probability of the humidity gradient, Represents the joint probability of time domain features and humidity gradient.

8. The joint early warning method according to claim 1, characterized in that: In step S6, the specific implementation steps are as follows: S61. Define the air pump node as node A, define the drying cylinder node as node B, define the key air path location node as node C, and calculate the air pump path respectively. and drying drum path ; S62, according to the air pump path and drying drum path Determine the most likely fault path and generate a fault path result, wherein the fault path result includes a pump gas fault result and a cylinder gas fault result; like , then the air pump path For the fault path, generate the pump gas fault result and end; like , then the drying tube path For the fault path, generate the cylinder gas fault result and end; S63. Send the primary fault result and the secondary abnormality result to the early warning center, generate an AR three-dimensional disassembly animation based on the fault path result, and highlight the fault path and faulty components in the early warning system.

9. The joint early warning method according to claim 8, characterized in that: The air pump path and drying drum path The calculation formula is: Where F and G represent the number of AC paths and BC paths, respectively. represents the probability that node C will fail if node A fails, It represents the probability that node C will fail if node B fails.

10. A system for applying the joint early warning method for passenger and freight vehicle braking risk as described in any one of claims 1 to 9, characterized in that: The system includes: The preprocessing module (1) is used to collect the original data inside the passenger and freight vehicle, and preprocess the original data to obtain the dynamic calibration value ; Feature enhancement module (2) is used to dynamically calibrate the value Calculate the mixed filter value , and based on the mixed filter value Calculate feature enhancement values ​​for feature decoupling ; Calibration module (3) is used to enhance the value based on the feature Divide the input set into categories , for the input set Gas path node temperature and gas path node humidity Perform calibration and obtain a calibration data set to make the data more accurate ; Digital twin module (4) is used to Constructed to get thermal coupling value and wet coupling values Digital twin dynamic model , and based on the thermal coupling value and wet coupling values Calculate the pre-feature set used to provide reference for fault judgment ; Fault causality module (5) is used to Construct a brake air circuit fault cause-effect diagram, calculate the judgment probability based on the brake air circuit fault cause-effect diagram, and obtain the primary fault result and secondary abnormal result based on the judgment probability; AR generation module (6), used to calculate the air pump path for judging the result of the pump air failure based on the brake air path fault cause and effect diagram and drying drum path ,generate AR animation based on the fault path results.

Images