Fault early warning method and device for air spring, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a fault early warning method and device for an air spring, electronic equipment and a storage medium. The method comprises the steps that driving state data, air spring control data and driving condition data of a vehicle are obtained; feature extraction is carried out on the driving state data and the air spring control data, multiple pieces of feature data are obtained, and the multiple pieces of feature data comprise time domain feature data, frequency domain feature data, vehicle asymmetry feature data, attitude stability feature data and control signal feature data; according to the driving condition data, determining a weight parameter corresponding to each piece of feature data; based on the weight parameter corresponding to each piece of feature data, performing weighted fusion on the plurality of pieces of feature data to obtain fused feature data; and determining the health state and the fault type of the air spring according to the fused feature data. According to the invention, misrecognition can be effectively reduced, and the accuracy of health state and fault type determination is improved.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, and in particular to a fault warning method, device, electronic device, and storage medium for air springs. Background Technology

[0002] As a core component of automotive air suspension systems, air springs provide elastic support through compressed air. They can dynamically adjust suspension stiffness and damping according to vehicle load, road conditions, and driving needs, thereby improving ride comfort, handling stability, and passability.

[0003] During long-term operation, air springs are prone to problems such as rubber aging, fabric fatigue, seal failure, or airbag leakage, leading to a gradual decline in system performance. In severe cases, this can even cause safety issues such as suspension collapse and loss of vehicle control. Because air spring failures often exhibit both gradual and sudden characteristics, they are difficult to detect in time through routine inspections before failure occurs. Therefore, life assessment and fault warning for air springs are particularly important.

[0004] In response, related technologies utilize digital twin models to simulate the operation of air springs using finite element analysis to estimate their remaining lifespan. However, this method suffers from high computational complexity, high costs in building and maintaining digital twin models, and difficulty in meeting the needs for real-time early warning. Summary of the Invention

[0005] This application provides a fault early warning method, device, electronic device, and storage medium for air springs, which can effectively reduce false identification and improve the accuracy of health status and fault type determination.

[0006] In a first aspect, embodiments of this application provide a fault warning method for air springs. The method includes: acquiring vehicle driving state data, air spring control data, and driving condition data; extracting features from the driving state data and air spring control data respectively to obtain multiple feature data, including time-domain feature data, frequency-domain feature data, vehicle asymmetry feature data, attitude stability feature data, and control signal feature data; determining weight parameters corresponding to each feature data based on the driving condition data; performing weighted fusion of the multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data; and determining the health status and fault type of the air spring based on the fused feature data.

[0007] In one possible implementation, determining the weight parameters corresponding to each feature data based on driving condition data includes: adjusting the initial weight parameters corresponding to each feature data based on driving condition data to obtain target weight parameters corresponding to each feature data; and weighting and fusing multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data includes: weighting and fusing multiple feature data based on the target weight parameters corresponding to each feature data to obtain fused feature data.

[0008] In one possible implementation, the aforementioned driving condition data includes vehicle speed, steering wheel angle, and road surface type. The adjustment of the initial weight parameters corresponding to each feature data based on the driving condition data to obtain target weight parameters for each feature data includes: when the vehicle speed is within a first preset range, the steering wheel angle is less than or equal to a first preset angle, and the road surface type is a flat road surface, the initial weight parameters corresponding to each feature data remain unchanged; when the vehicle speed is within a second preset range, and / or the steering wheel angle is greater than the first preset angle, and / or the road surface type is not a flat road surface, the initial weight parameters corresponding to each feature data are adjusted to obtain target weight parameters for each feature data; wherein the lower limit of the second preset range is greater than the upper limit of the first preset range.

[0009] In one possible implementation, adjusting the initial weight parameters corresponding to each feature data to obtain the target weight parameters for each feature data includes: when the vehicle speed is within a second preset range, the steering wheel angle is less than or equal to a first preset angle, and the road surface type is a flat road surface, increasing the initial weight parameters corresponding to the time-domain feature data, frequency-domain feature data, and attitude stability feature data, and decreasing the initial weight parameter corresponding to the control signal feature data to obtain the target weight parameters for each feature data; when the road surface type is not a flat road surface, increasing the initial weight parameters corresponding to the time-domain feature data, frequency-domain feature data, and vehicle asymmetry feature data, and decreasing the initial weight parameters corresponding to the attitude stability feature data and control signal feature data to obtain the target weight parameters for each feature data; when the steering wheel angle is greater than the first preset angle, increasing the initial weight parameters corresponding to the vehicle asymmetry feature data and attitude stability feature data, and decreasing the initial weight parameter corresponding to the control signal feature data to obtain the target weight parameters for each feature data.

[0010] In one possible implementation, the aforementioned driving state data includes the wheel speed and wheel acceleration of each wheel, and also includes the left suspension travel, right suspension travel, vehicle pitch rate, and roll rate. Feature extraction is performed on the driving state data and air spring control data to obtain multiple feature data, including: frequency domain feature extraction on the driving state data to obtain frequency domain feature data; time domain feature extraction on the driving state data to obtain time domain feature data; determining vehicle asymmetry feature data based on the left suspension travel, right suspension travel, wheel speed, and wheel acceleration of each wheel; determining attitude stability feature data based on pitch rate and roll rate; and extracting features from the air spring control data to obtain control signal feature data.

[0011] In one possible implementation, determining the weight parameters corresponding to each feature data based on driving condition data includes: determining the weight parameters corresponding to each feature data based on the target fault warning model and driving condition data; and weighting and fusing multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data, and determining the health status and fault type of the air spring based on the fused feature data, including: using the target fault warning model, weighting and fusing multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data, and determining the health status and fault type of the air spring based on the fused feature data.

[0012] In one possible implementation, the method further includes: acquiring multiple sets of training sample data, each set of training sample data including sample driving state data, sample air spring control data, sample driving condition data, real health status, and real fault type; using the real health status and real fault type as training labels, and employing multiple sets of training samples, training the initial fault warning model to obtain a target fault warning model, the target fault warning model being used to predict the health status and fault type of the air spring based on the vehicle's driving state data, air spring control data, and driving condition data.

[0013] Secondly, embodiments of this application provide a fault warning device for an air spring, comprising:

[0014] The data acquisition module is used to acquire vehicle driving status data, air spring control data, and driving condition data;

[0015] The feature extraction module is used to extract features from driving state data and air spring control data respectively, and obtain multiple feature data, including time domain feature data, frequency domain feature data, vehicle asymmetry feature data, attitude stability feature data and control signal feature data;

[0016] The weight allocation module is used to determine the weight parameters corresponding to each feature data based on driving condition data.

[0017] The feature fusion module is used to perform weighted fusion of multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data;

[0018] The fault determination module is used to determine the health status and fault type of the air spring based on the fused feature data.

[0019] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0020] The memory stores the instructions that the computer executes;

[0021] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0023] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0024] In this embodiment, during vehicle operation, driving status data, air spring control data, and driving condition data are acquired. Feature extraction is performed on the driving status data and air spring control data to obtain feature data across multiple dimensions, including time-domain features, frequency-domain features, vehicle asymmetry features, attitude stability features, and control signal features. Based on the driving condition data, weight parameters are determined for each feature data, and multiple feature data are weighted and fused. The fused feature data is then used to determine the health status and fault type of the air spring. This multi-dimensional feature data, encompassing time-domain, frequency-domain, vehicle asymmetry, attitude stability, and control signal features, can distinguish between normal operating conditions (such as high-speed cornering) and sudden faults (such as air leakage), effectively reducing false identification and improving the accuracy of health status and fault type determination. Furthermore, early fault warnings prevent sudden failures such as suspension collapse, ensuring driving safety. Additionally, it reduces onboard computing power requirements, enables model sharing via cloud interconnection, and avoids the high-cost modeling of digital twin technology. Attached Figure Description

[0025] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0026] Figure 1 This is a schematic diagram illustrating an application scenario of a fault warning method for an air spring according to an embodiment of this application;

[0027] Figure 2 A flowchart illustrating the fault warning method for an air spring provided in this application embodiment;

[0028] Figure 3 A schematic diagram of the architecture of a target fault early warning model provided in an embodiment of this application;

[0029] Figure 4 A schematic diagram of the structure of an air spring fault warning device provided in an embodiment of this application;

[0030] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0031] The embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described below do not represent all embodiments consistent with this application. They are merely examples of systems and methods consistent with some aspects of this application as detailed in the claims.

[0032] It should be noted that the brief descriptions of terms in this application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning.

[0033] The terms "first," "second," "third," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar or related objects or entities, and do not necessarily imply a specific order or sequence, unless otherwise specified. It should be understood that such terms are interchangeable where appropriate.

[0034] The terms “comprising” and “having”, and any variations thereof, are intended to cover but not exclude inclusion, for example, a product or device that includes a range of components is not necessarily limited to all of the components that are clearly listed, but may include other components that are not clearly listed or that are inherent to such product or device.

[0035] The term "module" refers to any known or subsequently developed hardware, software, firmware, artificial intelligence, fuzzy logic, or combination of hardware and / or software code that is capable of performing the functions associated with that element.

[0036] As a core component of automotive air suspension systems, air springs provide elastic support through compressed air. They can dynamically adjust suspension stiffness and damping according to vehicle load, road conditions, and driving needs, thereby improving ride comfort, handling stability, and passability.

[0037] During long-term operation, air springs are prone to problems such as rubber aging, fabric fatigue, seal failure, or airbag leakage, leading to a gradual decline in system performance. In severe cases, this can even cause safety issues such as suspension collapse and loss of vehicle control. Because air spring failures often exhibit both gradual and sudden characteristics, they are difficult to detect in time through routine inspections before failure occurs. Therefore, life assessment and fault warning for air springs are particularly important.

[0038] In response, related technologies involve installing dedicated sensors on various components of the air spring to directly collect its physical state parameters, which are then used for lifespan assessment and fault prediction. However, this method requires additional sensors, resulting in high costs and limited scalability.

[0039] In some examples, digital twin models are used to simulate the operation of air springs through finite element analysis to estimate their remaining lifespan. However, this method has high computational complexity, high costs for building and maintaining digital twin models, and difficulty in meeting the needs of real-time early warning.

[0040] In view of this, this application provides a fault early warning method for air springs. During vehicle operation, it acquires vehicle driving status data, air spring control data, and driving condition data. Features are extracted from the driving status data and air spring control data to obtain feature data across multiple dimensions, including time-domain features, frequency-domain features, vehicle asymmetry features, attitude stability features, and control signal features. Based on the driving condition data, weight parameters are determined for each feature data, and multiple feature data are weighted and fused. The fused feature data is then used to determine the health status and fault type of the air spring. In this way, by using feature data across multiple dimensions such as time-domain features, frequency-domain features, vehicle asymmetry features, attitude stability features, and control signal features, it is possible to distinguish between normal operating conditions (such as high-speed cornering) and sudden faults (such as air leakage), effectively reducing false identification and improving the accuracy of health status and fault type determination. Furthermore, early fault warnings prevent sudden failures such as suspension collapse, ensuring driving safety. Additionally, it reduces the need for onboard computing power, enables model sharing via cloud interconnection, and avoids the high-cost modeling of digital twin technology.

[0041] Before introducing the air spring fault warning method provided in the embodiments of this application, the application scenarios of the air spring fault warning method will be explained first.

[0042] Figure 1 This is a schematic diagram illustrating an application scenario of a fault warning method for an air spring according to an embodiment of this application.

[0043] The method provided in this application embodiment can be used to detect the health status of air springs in a vehicle suspension system, and to determine the fault type of the air spring when an abnormal state is detected.

[0044] Specifically, such as Figure 1 As shown, during the driving process of vehicle 10, driving status data, air spring control signals (hereinafter referred to as air spring control data), and driving condition data of vehicle 10 are acquired. Among them, driving status data can be collected by sensors arranged on vehicle 10, such as longitudinal, lateral, and vertical acceleration data (i.e., acceleration data in the x, y, and z directions) collected by the data acquisition unit through inertial sensors, as well as pitch angular velocity, roll angular velocity (i.e., roll angular velocity), and yaw angular velocity data (i.e., yaw angular velocity), wheel speed and wheel acceleration collected by sensors on each wheel, and suspension travel (i.e., the extension of the air spring) collected by suspension travel sensors.

[0045] Air spring control data can be obtained by parsing the control signals from the air spring controller via a communication bus (such as a CAN bus). In this embodiment, the air spring may include an air bladder, an inflation valve, and an exhaust valve. Accordingly, the air spring control data may include target altitude, target pressure, actual altitude, actual pressure, inflation valve data, and exhaust valve data.

[0046] Driving condition data can characterize how the current driving condition of a vehicle affects driving status data. Driving condition data may include, for example, the vehicle's speed, steering wheel angle, and road surface type.

[0047] After acquiring driving status data, air spring control data, and driving condition data, the fault assessment system 20 can detect the health status and fault type of the air spring based on the driving status data, air spring control data, and driving condition data.

[0048] In some examples, the fault assessment system 20 extracts features from driving status data and air spring control data to obtain multiple feature data. Based on the driving condition data, it determines the weight parameters corresponding to each feature data, performs weighted fusion of multiple feature data, and determines the health status and fault type of the air spring based on the obtained fused feature data.

[0049] In some examples, a target fault warning model (i.e., an air suspension fault detection model) can be deployed in the fault assessment system 20. In practical applications, the acquired driving status data, air spring control data, and driving condition data are input into the target fault warning model, which outputs the health status and fault type of the air spring.

[0050] For example, the target fault warning model can be deployed on vehicle 10, that is, the fault assessment system 20 can be deployed on vehicle 10. In this scenario, during the driving process of vehicle 10, driving status data, air spring control data and driving condition data of vehicle 10 are acquired, and the target fault warning model is used to process the driving status data, air spring control data and driving condition data, and output the health status and fault type of the air spring.

[0051] For example, the target fault warning model can be deployed on a server (i.e., the cloud), meaning the fault assessment system 20 can be deployed on a server. In this scenario, during the driving of vehicle 10, driving status data, air spring control data, and driving condition data of vehicle 10 are acquired and sent to the server. The server receives the driving status data, air spring control data, and driving condition data, and uses the target fault warning model to process these data to determine the health status and fault type of the air springs. After determining the health status and fault type of the air springs, the server sends the health status and fault type of the air springs to vehicle 10, causing vehicle 10 to output corresponding prompt information.

[0052] The fault warning method for air springs provided in this application embodiment will be described in detail below with reference to the accompanying drawings and application scenarios.

[0053] Figure 2 This is a flowchart illustrating a fault warning method for an air spring provided in an embodiment of this application. Figure 2 As shown, the fault warning method for this air spring may include the following steps:

[0054] S201 acquires vehicle driving status data, air spring control data, and driving condition data.

[0055] The driving status data is used to characterize the vehicle's current driving state. For example, the driving status data may include acceleration and angular velocity data collected by inertial sensors. Acceleration data may include the vehicle's longitudinal acceleration, lateral acceleration (or side acceleration), and vertical acceleration. Angular velocity data may include the vehicle's pitch, roll, and yaw rates. The acceleration and angular velocity data can reflect the vehicle's linear motion in the x, y, and z directions and its rotational motion about the three axes.

[0056] Driving status data can also include wheel speeds and wheel accelerations of each wheel. Specifically, this can be collected by sensors on each wheel. The wheel speeds can be used to calculate the vehicle's current actual speed and wheel speed difference. Wheel acceleration can include the vehicle's longitudinal wheel acceleration and lateral wheel acceleration, reflecting the road surface excitation experienced by the wheels.

[0057] Driving status data may also include suspension travel data collected by suspension travel sensors. Suspension travel refers to the actual extension of the air spring in the suspension system, reflecting the compression and extension state of the air spring.

[0058] Air spring control data can include the target height, target pressure, actual height, and actual pressure of the air spring. The target height and target pressure are desired values. The target height refers to the set desired vehicle body height, which can be calculated based on the target extension of the air spring. It can be understood that the target height is related to the vehicle's current load, driving mode, and road conditions. The target extension of the air spring can be set based on the current load, driving mode, and road conditions to obtain the target height. The target pressure is the desired pressure required to control the vehicle body height to reach the desired vehicle body height (i.e., the target height).

[0059] Actual height and actual pressure are actual measured values. Actual height refers to the vehicle's current actual height, that is, the actual height of the vehicle body relative to the ground. Actual height can be calculated based on the actual extension of the air spring. The actual extension of the air spring is the suspension travel measured by the suspension travel sensor. Actual pressure corresponds to actual height; actual pressure refers to the actual pressure value inside the air spring measured by the pressure sensor.

[0060] The air spring control data may also include inflation valve data and deflation valve data. The air spring may include an airbag, an inflation valve, and a deflation valve. In response to an inflation command sent by the controller, the inflation valve opens, allowing compressed air to enter the airbag, causing the airbag to extend and thus raising the vehicle height or increasing suspension stiffness. In response to a deflation command, the deflation valve opens, allowing compressed air to escape from the airbag, causing the airbag to compress and thus lowering the vehicle height or decreasing suspension stiffness. The inflation valve data may include inflation valve status and inflation valve commands, with the inflation valve status including open and closed states, and the inflation valve commands including open and closed commands. Similarly, the deflation valve data may include deflation valve status and deflation valve commands, with the deflation valve status including open and closed states, and the deflation valve commands including open and closed commands.

[0061] Air spring control data may also include the operating current (i.e., output current) of the air spring controller, which is the current value that drives the solenoid valve (such as an inflation valve or an exhaust valve) or other actuators, to reflect the workload of the air spring controller.

[0062] Air spring control data can also include stiffness and damping estimates. Specifically, these estimates can be determined based on the air spring's current actual pressure, actual height, and operating temperature. The stiffness and damping estimates reflect the equivalent stiffness and damping characteristics of the air spring.

[0063] Driving condition data reflects the current driving conditions of a vehicle. For example, driving condition data may include the vehicle's current speed and steering wheel angle. Driving condition data may also include road surface type, such as flat road surface.

[0064] For example, driving condition data may also include vehicle model information.

[0065] For example, driving condition data may also include ambient temperature and ambient humidity. Ambient temperature can be acquired by an external temperature sensor, reflecting the ambient temperature of the environment in which the vehicle is located. Ambient humidity can be acquired by an external humidity sensor, reflecting the ambient humidity of the environment in which the vehicle is located.

[0066] In some embodiments, after acquiring driving status data, air spring control data, and driving condition data, the driving status data, air spring control data, and driving condition data are cleaned and normalized.

[0067] S202 extracts features from driving state data and air spring control data respectively, and obtains multiple feature data, including frequency domain feature data, time domain feature data, vehicle asymmetry feature data, attitude stability feature data and control signal feature data.

[0068] In this embodiment, the acquired data (i.e., driving status data, air spring control data, and driving condition data) are all continuous time-series data. Therefore, according to a preset sliding window, each data is segmented to obtain multiple time windows, each time window from t1 to t2. L We have data from L time steps. For each time window, we extract features from the data within that time window to obtain the corresponding feature data.

[0069] For example, the preset window length of the sliding window can be 2-10 seconds, with a step size of 1 second.

[0070] The following sections provide illustrative explanations of the extraction of frequency domain feature data, time domain feature data, vehicle asymmetry feature data, attitude stability feature data, and control signal feature data.

[0071] 1. Frequency domain characteristic data

[0072] Frequency domain characteristic data is used to characterize the distribution characteristics of vibration energy in the vertical direction of a vehicle, that is, to characterize the distribution characteristics of vibration energy in different frequency bands caused by changes in the damping performance of air springs.

[0073] In some examples, for each type of driving state data, frequency domain features are extracted from the driving state data to obtain the frequency domain features corresponding to that driving state data, and the frequency domain features corresponding to each driving state data are used as frequency domain feature data.

[0074] For example, for any type of driving state data, a preset sliding window is used to segment the driving state data, resulting in multiple time windows. For each time window, a Fourier transform is performed on the driving state data within that time window to obtain a spectral signal. According to a preset strategy, the spectral signal is divided into multiple frequency bands, and the energy value within each frequency band is calculated as the frequency domain energy feature of the driving state data. Additionally, the standard deviation of the energy value within each frequency band is calculated as the energy distribution fluctuation feature of the driving state data. The frequency domain energy feature and the energy distribution fluctuation feature are then used as the frequency domain features of the driving state data.

[0075] In practical applications, following the above process, frequency domain features are extracted from acceleration data (such as vertical acceleration), angular velocity data (such as pitch and roll angular velocities), suspension travel, and wheel speed and wheel acceleration of each wheel, respectively, to obtain the frequency domain features of acceleration data, angular velocity data, suspension travel, wheel speed, and wheel acceleration of each wheel, thus obtaining the frequency domain feature data.

[0076] 2. Time-domain feature data

[0077] Temporal feature data is used to characterize the changes in vibration amplitude and impact intensity of vehicles in the vertical direction.

[0078] In some examples, for each type of driving state data, time-domain features are extracted from the driving state data to obtain the time-domain features corresponding to that driving state data, and the time-domain features corresponding to each driving state data are used as time-domain feature data.

[0079] For example, for any type of driving state data, a preset sliding window is used to segment the driving state data, resulting in multiple time windows. For each time window, the standard deviation, maximum value, minimum value, and peak value of the driving state data within that time window are calculated to obtain the time-domain characteristics of the driving state data. The standard deviation of the driving state data reflects the dispersion and vibration intensity of the signal within the window. The maximum and minimum values ​​of the driving state data are used to capture positive and negative extreme impacts. The peak value of the driving state data, i.e., the maximum absolute value of the signal, is used to capture bidirectional extreme impacts.

[0080] Specifically, following the above process, time-domain features are extracted from acceleration data (such as vertical acceleration) and angular velocity data (such as pitch and roll angular velocity) to obtain the time-domain features of acceleration data and angular velocity data, thus obtaining time-domain feature data.

[0081] 3. Vehicle asymmetric feature data

[0082] Vehicle asymmetric feature data is used to characterize the difference in motion state between the left and right sides of the vehicle, that is, to reflect the difference in motion state between the left and right sides of the vehicle caused by air spring failure.

[0083] In some examples, the driving status data includes wheel speeds and wheel accelerations of each wheel, as well as left and right suspension travel data collected by suspension travel sensors on both sides. Based on this, vehicle asymmetry characteristic data are determined.

[0084] For example, a preset sliding window is used to segment the wheel speed and wheel acceleration of each wheel, resulting in multiple time windows. For each time window, the root mean square (RMS) of the wheel speed of the left wheel and the right wheel within that time window are calculated. The difference between the RMS of the left and right wheel speeds is used to determine wheel speed asymmetry characteristics. Furthermore, for each time window, the maximum and minimum values ​​of the wheel speeds of the left and right wheels are calculated, and the maximum range of wheel speeds is determined based on these values.

[0085] For example, a preset sliding window is used to segment the left and right suspension travel, resulting in multiple time windows. For each time window, the root mean square (RMS) of the left and right suspension travel within that time window is calculated. The difference between the RMS of the left and right suspension travel is used to determine the suspension travel asymmetry characteristics. Furthermore, for each time window, the maximum and minimum values ​​of the left and right suspension travel are calculated, and the maximum range of suspension travel is determined based on these values.

[0086] Wheel speed asymmetry, suspension travel asymmetry, maximum wheel speed range, and maximum suspension travel range are used as vehicle asymmetry characteristic data.

[0087] 4. Attitude stability feature data

[0088] Attitude stability characteristic data is used to characterize the stability of the vehicle's pitch and roll attitude during driving. When the performance of the air springs deteriorates (such as air leakage or decreased stiffness), the suspension's ability to support the vehicle's attitude decreases, leading to an increase in the steady-state deviation of pitch and roll attitude and a wider range of dynamic fluctuations.

[0089] In some examples, attitude stability characteristic data are determined based on pitch and roll angular velocities.

[0090] For example, a preset sliding window is used to segment the pitch and roll angular velocities, resulting in multiple time windows. For each time window, the average, maximum, and minimum values ​​of the pitch angular velocity within that time window are calculated. Based on the maximum and minimum values ​​of the pitch angular velocity, the fluctuation range of the pitch angular velocity is determined.

[0091] Furthermore, the average, maximum, and minimum values ​​of the roll angular velocity within this time window are calculated, and the fluctuation range of the roll angular velocity is determined based on the maximum and minimum values. The average pitch angular velocity, the fluctuation range of the pitch angular velocity, the average roll angular velocity, and the fluctuation range of the roll angular velocity are used as attitude stability characteristics.

[0092] The average values ​​of pitch and roll angular velocities are used to reflect the steady-state deviation of the vehicle body posture. When there is air leakage in one side of the air spring or abnormal vehicle load distribution, the average values ​​will deviate from the normal range. The fluctuation range of pitch and roll angular velocities is used to reflect the dynamic stability of the vehicle body posture. When the damping performance of the air springs decreases, the fluctuation range will increase significantly.

[0093] 5. Control signal characteristic data

[0094] Control signal characteristic data are used to characterize the regulating behavior and execution efficiency of the air spring control system.

[0095] In some examples, feature extraction is performed on the air spring control data to obtain control signal feature data.

[0096] For example, a preset sliding window is used to segment the control data of each air spring (such as target height and actual height) to obtain multiple time windows. For each time window, the average height difference and the maximum height difference between the target height and the actual height are calculated. The average height difference reflects the steady-state tracking error of the control system, while the maximum height difference reflects the transient tracking deviation. When the air spring leaks air or its stiffness decreases, the controller has difficulty maintaining the target height, and the height difference will increase significantly.

[0097] A preset sliding window is used to segment the control data of each air spring (such as inflation valve data and deflation valve data) into multiple time windows. For each time window, the maximum inflation duration, inflation frequency, maximum deflation duration, and deflation frequency are calculated. The maximum inflation duration and maximum deflation duration are the longest duration of a single inflation or deflation event within the window; these values ​​increase when system efficiency decreases. The inflation frequency and deflation frequency are the number of inflation or deflation events per unit time; these values ​​increase when the controller needs to make frequent adjustments. Furthermore, the control command frequency is determined based on the inflation valve commands and deflation valve commands within each time window. This determines the frequency at which the controller issues adjustment commands. When the air spring performance deteriorates, the controller needs to issue adjustment commands more frequently to maintain the target state.

[0098] S203, based on driving condition data, determine the weight parameters corresponding to each feature data.

[0099] The driving condition data includes vehicle speed, steering wheel angle, and road surface type. Weighting parameters for each feature data point can be determined based on the vehicle's current speed range, steering wheel angle, and road surface type.

[0100] In some examples, when the vehicle speed is within a first preset range, the steering wheel angle is less than or equal to a first preset angle, and the road surface type is a flat road surface, the initial weight parameters corresponding to each feature data remain unchanged.

[0101] The first preset range is used to determine whether the vehicle's current speed is too low. The first preset angle is used to determine whether the vehicle's current steering wheel angle is too small.

[0102] When the vehicle speed is within the first preset range, the steering wheel angle is less than or equal to the first preset angle, and the road surface type is flat, it indicates that the current vehicle speed is low, the steering wheel angle is small, and the road surface is relatively flat. At this time, the vehicle's driving state is relatively stable. The various feature data determined in the previous step can reflect the actual situation of the air springs in the suspension system. Therefore, the initial weight parameters corresponding to each feature data can be used to perform weighted fusion processing on multiple feature data. Then, combined with subsequent steps, the fused feature data can be used to determine the health status and fault type of the air springs.

[0103] For example, the initial weight parameters corresponding to each feature data can be preset according to the actual situation. Optionally, the initial weight parameters corresponding to each feature data can be the same or different.

[0104] For example, the initial weights of all types of feature data can be set to 1.0. Alternatively, based on experiments or experience, the initial weight parameters for time-domain feature data can be set to 0.8, for frequency-domain feature data to 0.8, for vehicle asymmetry feature data to 1.0, for attitude stability feature data to 1.0, and for control signal feature data to 1.0.

[0105] In some examples, when the vehicle speed is within a second preset range, and / or the steering wheel angle is greater than a first preset angle, and / or the road surface type is not a flat road surface type, the initial weight parameters corresponding to each feature data are adjusted to obtain the target weight parameters corresponding to each feature data.

[0106] The lower limit of the second preset range is greater than the upper limit of the first preset range. The second preset range is used to determine whether the vehicle's current speed is too high.

[0107] When the vehicle speed is within the second preset range, and / or the steering wheel angle is greater than the first preset angle, and / or the road surface type is not flat, it indicates that the current vehicle speed is high, and / or the steering wheel angle is large, and / or the road surface is bumpy. At this time, the vehicle's driving state is unstable, which will affect the various feature data determined in the previous step. Therefore, the initial weight parameters corresponding to each feature data are adjusted. Based on the adjusted target weight parameters, multiple feature data are weighted and fused. The resulting fused feature data can more accurately reflect the health status of the air spring. In combination with subsequent steps, the health status and fault type of the air spring can be determined, which can improve the accuracy of the assessment.

[0108] For example, when the vehicle speed is within the second preset range, the steering wheel angle is less than or equal to the first preset angle, and the road surface type is a flat road surface, the initial weight parameters corresponding to the time domain feature data, frequency domain feature data, and attitude stability feature data are increased, while the initial weight parameters corresponding to the control signal feature data are decreased, to obtain the target weight parameters corresponding to each feature data.

[0109] In other words, when the current vehicle speed is high, the steering wheel angle is small, and the road surface is relatively flat, the initial weight parameters corresponding to the time domain feature data, frequency domain feature data, and attitude stability feature data are increased, while the initial weight parameters corresponding to the control signal feature data are decreased, thus obtaining the target weight parameters corresponding to each feature data.

[0110] In this embodiment, when the vehicle speed is within the second preset range, the steering wheel angle is less than or equal to the first preset angle, and the road surface type is a flat road surface (i.e., high-speed flat straight driving condition), increasing the weights of time-domain feature data, frequency-domain feature data, and attitude stability feature data can enhance the sensitivity to the dynamic response of the suspension system and detect early signs of air spring performance degradation earlier when driving at high speeds. At the same time, reducing the weights of control signal feature data can avoid misjudging normal high-speed adjustment behavior as a fault, thereby improving the accuracy and timeliness of the warning.

[0111] For example, when the road surface type is not a flat road surface type, the initial weight parameters corresponding to the time domain feature data, frequency domain feature data, and vehicle asymmetry feature data are increased, while the initial weight parameters corresponding to the attitude stability feature data and control signal feature data are decreased, so as to obtain the target weight parameters corresponding to each feature data.

[0112] In this embodiment, when the road surface type is not a flat road surface (i.e., non-flat road surface condition), increasing the weights of time-domain feature data, frequency-domain feature data, and vehicle asymmetry feature data can enhance the ability to detect single-sided air spring faults; at the same time, reducing the weights of attitude stability feature data and control signal feature data can avoid misjudging normal attitude fluctuations and adjustment behaviors caused by road bumps as faults, effectively reducing the false alarm rate.

[0113] For example, when the steering wheel angle is greater than a first preset angle, the initial weight parameters corresponding to the vehicle asymmetry feature data and attitude stability feature data are increased, and the initial weight parameters corresponding to the control signal feature data are decreased, so as to obtain the target weight parameters corresponding to each feature data.

[0114] In this embodiment, when the steering wheel angle is greater than the first preset angle (i.e., large-angle steering condition), the weights of vehicle asymmetry feature data and attitude stability feature data are increased, which can enhance the sensitivity to abnormal body roll during turning and the ability to detect unilateral faults; at the same time, the weights of control signal feature data are reduced, which can avoid misjudging the adjustment behavior during normal steering as a fault and achieve accurate early warning.

[0115] For example, the weight parameters of each feature data under different driving conditions are shown in Table 1 below.

[0116]

[0117] In Table 1 above, the weights The weights are the weight parameters corresponding to the time-domain feature data. The weights are the weight parameters corresponding to the frequency domain feature data. The weights are the weight parameters corresponding to the vehicle asymmetry feature data. The weights are the weight parameters corresponding to the attitude stability feature data. These are the control signal feature data. In case 1, the weight parameters corresponding to each feature data are the initial weight parameters.

[0118] S204, based on the weight parameters corresponding to each feature data, performs weighted fusion of multiple feature data to obtain fused feature data.

[0119] For example, the fused feature data can be determined according to the following formula (1).

[0120] (1)

[0121] in, This indicates fused feature data; Represents time-domain feature data, This represents the weight parameters corresponding to the time-domain feature data; Represents frequency domain characteristic data, This represents the weight parameters corresponding to the frequency domain feature data; This represents the asymmetric characteristic data of the vehicle. The weight parameters represent the asymmetric feature data of the vehicle. This represents attitude stability feature data. The weight parameters represent the attitude stability feature data. This represents the characteristic data of the control signal. This represents the weight parameters corresponding to the characteristic data of the control signal.

[0122] When a certain type of feature data contains multiple feature terms (for example, time-domain feature data includes multiple feature values ​​such as standard deviation, maximum value, minimum value, peak value, etc.), the weight parameter corresponding to this type of feature is a scalar, and the same weight coefficient is applied to all feature terms within this type of feature, which is formula (2):

[0123] (2)

[0124] in, , , , These are the individual feature items in the feature data.

[0125] Optionally, weighted fusion can also employ a feature concatenation method, where various feature data are multiplied by their corresponding weight parameters and then concatenated along the feature dimension to form a fused feature vector; or a non-linear fusion method, where various feature data are adaptively fused through a neural network layer. Those skilled in the art can choose an appropriate fusion method based on the actual application scenario.

[0126] By using the weighted fusion method described above, the contribution of various features can be dynamically adjusted according to different driving conditions, so that the fused feature data can more accurately reflect the true health status of the air spring, providing high-quality input for subsequent fault diagnosis, thereby reducing the false alarm rate and improving the accuracy of fault detection.

[0127] S205, based on the fused feature data, determine the health status and fault type of the air spring.

[0128] For example, the health status of an air spring can include a normal state and an abnormal state. Fault types of an air spring can include: air bladder leakage, decreased stiffness, valve sticking, controller malfunction, air path blockage, sensor failure, etc.

[0129] In some embodiments, multiple extracted feature data can be input into a pre-trained target fault warning model. Based on driving condition data, the weight parameters corresponding to each feature data are determined. Based on the weight parameters corresponding to each feature data, multiple feature data are weighted and fused to obtain fused feature data. Then, based on the fused feature data, the health status and fault type of the air spring are determined.

[0130] Specifically, the aforementioned driving status data, air spring control data, and driving condition data are all continuous time-series data. In practical applications, after acquiring the driving status data, air spring control data, and driving condition data, each data point is segmented according to a preset sliding window to obtain multiple time windows. Each time window contains data from t1 to t2. LData is collected over L time steps. For each time window, the raw data and extracted feature data within that window are concatenated to form a two-dimensional matrix X. This two-dimensional matrix is ​​then used as input to a pre-trained target fault early warning model, which outputs the health status and fault type of the air spring.

[0131] For example, a target fault early warning model can be constructed using supervised learning, unsupervised learning, or semi-supervised learning methods.

[0132] For example, such as Figure 3 As shown, taking a Transformer-based deep learning model as an example, the target fault warning model includes an input embedding layer, a position encoding layer, an encoder layer, a sequence convergence layer, and an output layer. The encoder layer consists of multiple identical encoders stacked together (e.g., 3 encoders), and each encoder contains a multi-head self-attention mechanism (e.g., 6 heads).

[0133] The input embedding layer receives a two-dimensional matrix X, composed of concatenated raw data and feature data. A linear transformation is performed on X to obtain the transformed matrix E1. The position encoding layer receives the transformed matrix E1 and adds the position-encoded cosine vector P to E1 to obtain E2, which is then input into the encoder layer. The outputs of all encoder heads are concatenated and projected back to the original dimension through a linear layer, followed by a feedforward neural network and residual connections. Finally, layer normalization is performed to generate matrix Z. The sequence pooling layer pools the matrix Z output by the last encoder through average pooling to generate vector C. The output layer inputs vector C into a multilayer perceptron and outputs information representing the current health state and fault type (i.e., fault direction) of the air spring, i.e., outputting a health value and fault type label, represented as (a, b).

[0134] Where a is the health value of the air spring (i.e., the overall health value), with a value range of [0, 100]. The larger the value, the healthier the air spring is; b is a multi-dimensional vector, where each component represents the confidence level of a certain fault type.

[0135] The health status of the air spring is determined by comparing its health value with a preset health threshold, i.e., whether the air spring is in an abnormal state. Furthermore, the type of air spring failure is determined by comparing each component with a preset fault threshold.

[0136] For example, a preset health threshold is set to 60 and a preset fault threshold is set to 0.6. When the health value 'a' of the air spring is less than the preset health threshold, the air spring is determined to be in an abnormal state; when any component of the multidimensional vector 'b' is greater than the preset fault threshold, a fault type corresponding to that component is determined to exist. Optionally, multiple thresholds can be set to achieve graded early warning, for example, a first health threshold and a second health threshold can be set, corresponding to minor abnormalities and serious faults respectively, triggering different levels of early warning measures.

[0137] In this embodiment, the judgment result is used as the system output to trigger the vehicle's in-vehicle alarm and remind the driver to repair it in time; at the same time, the fault information is recorded and stored in the cloud server, and a signal is output to the vehicle motion control system according to the severity of the fault. If necessary, the vehicle speed or suspension height adjustment function is restricted to prevent the fault from worsening.

[0138] In some embodiments, driving status data, air spring control data, and driving condition data (i.e., raw data) can be directly input into a pre-trained target fault warning model, which will automatically complete feature extraction, weight allocation, and fault determination, and output the health status and fault type of the air spring.

[0139] In some examples, multiple sets of training sample data are obtained. Each set of training sample data includes sample driving status data, sample air spring control data, sample driving condition data, real health status, and real fault type. Using the real health status and real fault type as training labels, the initial fault warning model is trained using multiple sets of training samples to obtain the target fault warning model. The target fault warning model is used to predict the health status and fault type of the air spring based on the vehicle's driving status data, air spring control data, and driving condition data.

[0140] The above-mentioned method based on the target fault early warning model realizes end-to-end processing from raw data input to fault judgment. There is no need to manually design feature extraction rules and weight adjustment rules. The model automatically learns the temporal dependency relationship and the adaptive capability of the working condition, which can accurately identify the health status and fault type of the air spring, effectively improving the accuracy and real-time performance of fault early warning.

[0141] In this embodiment, during vehicle operation, driving state data, air spring control data, and driving condition data are acquired. Feature extraction is performed on the driving state data and air spring control data to obtain feature data across multiple dimensions, including time-domain features, frequency-domain features, vehicle asymmetry features, attitude stability features, and control signal features, thus constructing a comprehensive representation of the air spring's operating state. Based on this, weight parameters corresponding to each feature data are dynamically determined according to the driving condition data, and multiple feature data are weighted and fused based on these weight parameters. This allows the model to adaptively adjust the contribution of each feature in fault determination according to the current driving scenario (such as high-speed cornering, bumpy roads, etc.), effectively distinguishing between dynamic responses under normal operating conditions and abnormal performance caused by real faults, significantly reducing the false recognition rate, and improving the accuracy of health status and fault type determination.

[0142] Furthermore, the fused feature data is input into the fault diagnosis model, which outputs the health status and specific fault type of the air spring. Early fault warnings can prevent sudden failures such as suspension collapse and ensure driving safety.

[0143] Furthermore, the model used in this application embodiment can be deployed in the cloud, and parameters can be continuously optimized through big data to adapt to different vehicle models and environmental changes. At the same time, the lightweight model design reduces the demand for on-board computing power, and cloud interconnection enables model sharing, effectively avoiding the high cost of modeling and computing power in digital twin technology.

[0144] This application also provides a fault warning device for air springs. For example... Figure 4 As shown, the air spring fault warning device 400 includes a data acquisition module 401, a feature extraction module 402, a weight allocation module 403, a feature fusion module 404, and a fault determination module 405. The data acquisition module 401 acquires vehicle driving status data, air spring control data, and driving condition data. The feature extraction module 402 extracts features from the driving status data and air spring control data respectively, obtaining multiple feature data, including time-domain feature data, frequency-domain feature data, vehicle asymmetry feature data, attitude stability feature data, and control signal feature data. The weight allocation module 403 determines the weight parameters corresponding to each feature data based on the driving condition data. The feature fusion module 404 performs weighted fusion of multiple feature data based on the weight parameters corresponding to each feature data, obtaining fused feature data. The fault determination module 405 determines the health status and fault type of the air spring based on the fused feature data.

[0145] In some embodiments, the weight allocation module 403 is specifically used to adjust the initial weight parameters corresponding to each feature data according to the driving condition data to obtain the target weight parameters corresponding to each feature data. The feature fusion module 404 is specifically used to perform weighted fusion of multiple feature data based on the target weight parameters corresponding to each feature data to obtain fused feature data.

[0146] In some embodiments, driving condition data includes vehicle speed, steering wheel angle, and road surface type. The weight allocation module 403 is specifically configured to: maintain the initial weight parameters corresponding to each feature data unchanged when the vehicle speed is within a first preset range, the steering wheel angle is less than or equal to a first preset angle, and the road surface type is a flat road surface; and adjust the initial weight parameters corresponding to each feature data to obtain target weight parameters corresponding to each feature data when the vehicle speed is within a second preset range, and / or the steering wheel angle is greater than the first preset angle, and / or the road surface type is not a flat road surface; wherein the lower limit of the second preset range is greater than the upper limit of the first preset range.

[0147] In some embodiments, the weight allocation module 403 is specifically configured to: when the vehicle speed is within a second preset range, the steering wheel angle is less than or equal to a first preset angle, and the road surface type is a flat road surface, increase the initial weight parameters corresponding to the time-domain feature data, frequency-domain feature data, and attitude stability feature data, and decrease the initial weight parameters corresponding to the control signal feature data to obtain target weight parameters corresponding to each feature data; when the road surface type is not a flat road surface, increase the initial weight parameters corresponding to the time-domain feature data, frequency-domain feature data, and vehicle asymmetry feature data, and decrease the initial weight parameters corresponding to the attitude stability feature data and control signal feature data to obtain target weight parameters corresponding to each feature data; when the steering wheel angle is greater than the first preset angle, increase the initial weight parameters corresponding to the vehicle asymmetry feature data and attitude stability feature data, and decrease the initial weight parameters corresponding to the control signal feature data to obtain target weight parameters corresponding to each feature data.

[0148] In some embodiments, the driving state data includes the wheel speed and wheel acceleration of each wheel, and also includes the left suspension travel, right suspension travel, vehicle pitch rate, and roll rate. The feature extraction module 402 is specifically used for: performing frequency domain feature extraction on the driving state data to obtain frequency domain feature data; performing time domain feature extraction on the driving state data to obtain time domain feature data; determining vehicle asymmetry feature data based on the left suspension travel, right suspension travel, wheel speed, and wheel acceleration of each wheel; determining attitude stability feature data based on the pitch rate and roll rate; and extracting features from the air spring control data to obtain control signal feature data.

[0149] In some embodiments, the weight allocation module 403 is specifically used to determine the weight parameters corresponding to each feature data based on the target fault warning model and driving condition data. The feature fusion module 404 is specifically used to use the target fault warning model to perform weighted fusion of multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data, and to determine the health status and fault type of the air spring based on the fused feature data.

[0150] In some embodiments, the air spring fault warning device 400 further includes a training module. The training module is used to acquire multiple sets of training sample data, each set including sample driving state data, sample air spring control data, sample driving condition data, actual health status, and actual fault type. Using the actual health status and actual fault type as training labels, the initial fault warning model is trained using multiple sets of training samples to obtain a target fault warning model. The target fault warning model is used to predict the health status and fault type of the air spring based on the vehicle's driving state data, air spring control data, and driving condition data.

[0151] The electronic device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0152] Figure 5 A schematic diagram of the structure of the electronic device provided in this application. Figure 5 As shown, the electronic device 50 provided in this embodiment includes at least one processor 501 and a memory 502. Optionally, the electronic device 50 further includes a communication component 503. The processor 501, memory 502, and communication component 503 are connected via a bus 504.

[0153] In a specific implementation, at least one processor 501 executes computer execution instructions stored in memory 502, causing at least one processor 501 to perform the above-described method.

[0154] The specific implementation process of processor 501 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0155] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0156] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0157] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0158] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0159] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0160] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0161] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0162] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0163] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0164] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0165] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0166] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0167] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A fault early warning method for an air spring, characterized in that, The method includes: Acquire vehicle driving status data, air spring control data, and driving condition data; Feature extraction is performed on the driving state data and the air spring control data respectively to obtain multiple feature data, including time domain feature data, frequency domain feature data, vehicle asymmetry feature data, attitude stability feature data and control signal feature data; Based on the driving condition data, determine the weight parameters corresponding to each feature data; Based on the weight parameters corresponding to each feature data, the multiple feature data are weighted and fused to obtain fused feature data; Based on the fused feature data, the health status and fault type of the air spring are determined.

2. The method according to claim 1, characterized in that, The step of determining the weight parameters corresponding to each feature data based on the driving condition data includes: Based on the driving condition data, the initial weight parameters corresponding to each feature data are adjusted to obtain the target weight parameters corresponding to each feature data. The weighted fusion of the multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data includes: Based on the target weight parameters corresponding to each feature data, the multiple feature data are weighted and fused to obtain fused feature data.

3. The method according to claim 2, characterized in that, The driving condition data includes vehicle speed, steering wheel angle, and road surface type; the adjustment of the initial weight parameters corresponding to each feature data based on the driving condition data to obtain the target weight parameters corresponding to each feature data includes: When the vehicle speed is within a first preset range, the steering wheel angle is less than or equal to a first preset angle, and the road surface type is a flat road surface, the initial weight parameters corresponding to each feature data remain unchanged. When the vehicle speed is within the second preset range, and / or the steering wheel angle is greater than the first preset angle, and / or the road surface type does not belong to the flat road surface type, the initial weight parameters corresponding to each feature data are adjusted to obtain the target weight parameters corresponding to each feature data; wherein, the lower limit of the second preset range is greater than the upper limit of the first preset range.

4. The method according to claim 3, characterized in that, The step of adjusting the initial weight parameters corresponding to each feature data to obtain the target weight parameters corresponding to each feature data includes: When the vehicle speed is within the second preset range, the steering wheel angle is less than or equal to the first preset angle, and the road surface type is the flat road surface type, the initial weight parameters corresponding to the time domain feature data, the frequency domain feature data, and the attitude stability feature data are increased, and the initial weight parameter corresponding to the control signal feature data is decreased, so as to obtain the target weight parameters corresponding to each feature data. If the road surface type does not belong to the flat road surface type, the initial weight parameters corresponding to the time domain feature data, the frequency domain feature data, and the vehicle asymmetry feature data are increased, and the initial weight parameters corresponding to the attitude stability feature data and the control signal feature data are decreased, so as to obtain the target weight parameters corresponding to each feature data. When the steering wheel angle is greater than the first preset angle, the initial weight parameters corresponding to the vehicle asymmetry feature data and the attitude stability feature data are increased, and the initial weight parameters corresponding to the control signal feature data are decreased, so as to obtain the target weight parameters corresponding to each feature data.

5. The method according to claim 1, characterized in that, The driving status data includes the wheel speed and wheel acceleration of each wheel, and also includes the left suspension travel, right suspension travel, vehicle pitch rate, and roll rate. Feature extraction is performed on the driving status data and the air spring control data to obtain multiple feature data, including: Frequency domain features are extracted from the driving state data to obtain the frequency domain feature data; The driving state data is subjected to time-domain feature extraction to obtain the time-domain feature data; The vehicle asymmetry characteristic data are determined based on the left suspension travel, the right suspension travel, the wheel speed and wheel acceleration of each wheel; The attitude stability characteristic data are determined based on the pitch angular velocity and the roll angular velocity. Feature extraction is performed on the air spring control data to obtain the control signal feature data.

6. The method according to claim 1, characterized in that, The step of determining the weight parameters corresponding to each feature data based on the driving condition data includes: Based on the target fault warning model and the driving condition data, determine the weight parameters corresponding to each feature data. The step of weighting and fusing the multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data, and determining the health status and fault type of the air spring based on the fused feature data, includes: Using the target fault early warning model, the multiple feature data are weighted and fused based on the weight parameters corresponding to each feature data to obtain fused feature data. Based on the fused feature data, the health status and fault type of the air spring are determined.

7. The method according to any one of claims 1-6, characterized in that, The method further includes: Acquire multiple sets of training sample data. Each set of training sample data includes sample driving status data, sample air spring control data, sample driving condition data, real health status, and real fault type. Using the actual health status and the actual fault type as training labels, and employing the multiple sets of training samples, the initial fault warning model is trained to obtain the target fault warning model. The target fault warning model is used to predict the health status and fault type of the air spring based on the vehicle's driving status data, air spring control data, and driving condition data.

8. A fault warning device for an air spring, characterized in that, include: The data acquisition module is used to acquire vehicle driving status data, air spring control data, and driving condition data; The feature extraction module is used to extract features from the driving state data and the air spring control data respectively to obtain multiple feature data, including time domain feature data, frequency domain feature data, vehicle asymmetry feature data, attitude stability feature data and control signal feature data; The weight allocation module is used to determine the weight parameters corresponding to each feature data based on the driving condition data. The feature fusion module is used to perform weighted fusion of the multiple feature data based on the weight parameters corresponding to each feature data to obtain fused feature data; The fault determination module is used to determine the health status and fault type of the air spring based on the fused feature data.

9. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-7.

Images