Active Suspension Fault Diagnosis Using Inertial-Guided Vibration Denoising
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for fault diagnosis of active suspensions face challenges due to differences between experimental platforms and real vehicles, leading to low diagnosis accuracy and noise interference in collected data.
Innovation Solution
An experimental platform for electro-hydraulic servo active suspensions is developed, which closely replicates real vehicle conditions. This platform includes a gantry frame, wheel suspension apparatus, electro-hydraulic servo actuator, and inertial sensors. A method for fault diagnosis is also introduced, involving noise reduction processing of vibration signals using longitudinal displacement and lateral inclination angle data, and feature extraction for constructing accurate fault feature data sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an existing experimental platform for active suspensions is used, then the platform structure is simple and easy to set up, but the experimental platform greatly differs from a real vehicle, resulting in low diagnosis accuracy rate
Solution Approach 1:
The patent creates a high-fidelity copy of the real vehicle suspension system by replicating its structural configuration, component arrangement, and operational characteristics in the experimental platform. This includes copying the wheel suspension apparatus, electro-hydraulic servo actuator, and sensor placement to match actual vehicle conditions, thereby achieving high measurement precision while maintaining platform usability
Solution Approach 2:
The experimental platform is divided into distinct functional modules including the gantry frame structure, wheel suspension apparatus, electro-hydraulic servo actuator, and sensor systems. Each module can be independently configured and adjusted to replicate specific vehicle conditions, allowing complex vehicle-like behavior to be achieved through modular assembly rather than monolithic complexity
2Reliability
If vibration signals are collected during vehicle motion, then real working conditions are captured, but the obtained vibration signals have relatively strong noise interference from vehicle body motion
Solution Approach 1:
The patent extracts and separates the harmful noise component from the useful vibration signal by using inertial sensors to specifically measure vehicle body motion (longitudinal displacement and lateral inclination). This allows the noise generated by vehicle body motion to be identified and removed from the cylinder body vibration signals, retaining only the fault-related vibration characteristics
Solution Approach 2:
The inertial sensor acts as an intermediary measurement device that captures the vehicle body motion parameters (longitudinal displacement and lateral inclination angle) which serve as noise sources. These intermediary measurements are then used to compensate for and remove the noise influence from the primary vibration sensors, enabling clean fault signal extraction during actual vehicle operation
3Measurement precision
If noise reduction processing is performed using longitudinal displacement and lateral inclination angle data, then diagnosis accuracy is improved, but data processing complexity increases
Solution Approach 1:
The patent replaces complex mechanical noise isolation structures with a computational approach using inertial measurement data. Instead of physically isolating sensors from vehicle motion, the system uses software-based noise reduction that substitutes mechanical complexity with mathematical processing of inertial sensor data to achieve the same noise removal effect
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The experimental platform achieves high accuracy in replicating real vehicle conditions, and the fault diagnosis method significantly improves diagnosis accuracy by reducing noise interference and extracting characteristic fault features, enabling practical application in real vehicle fault diagnosis.
Implementation Method 1
an inertial sensor are sequentially arranged on the bottom plate of the gantry frame from bottom to top
Implementation Method 2
cylinder body vibration sensors, the shock absorbers are arranged at upper parts of the wheels, the hydraulic cylinders are arranged behind the shock absorbers, the cylinder body vibration sensors are arranged on the hydraulic cylinders
Implementation Method 3
an electro-hydraulic servo actuator is arranged on the wheel suspension apparatus, the electro-hydraulic servo actuator includes an oil tank, a hydraulic pump, valve body vibration sensors, servo valves
Implementation Method 4
noise reduction processing is performed on acquired vibration signals by combining with longitudinal displacement and lateral inclination angle data as well as positional relationships between vibration sensors of different hydraulic elements and an inertial sensor
Data Source
AI summary
An experimental platform of the present disclosure is very close to an active suspension apparatus in a real vehicle. Fault feature data acquired on the basis of the experimental platform can well represent actual faults, so that a fault diagnosis model obtained on the basis of training of fault feature data sets has a relatively high practical value. According to the method of the present disclosure, noise reduction processing is performed on acquired vibration signals by combining with longitudinal displacement and lateral inclination angle data as well as positional relationships between vibration sensors of different hydraulic elements and an inertial sensor, and feature extraction is performed on vibration signals after noise reduction processing, so that constructed fault feature data sets are more accurate and more characteristic. The fault feature data sets are gradually expanded according to the actual faults, so that the fault diagnosis model has a certain self-learning capability.