In the aviation industry, the hydraulic system performance directly affects the safety of aircraft and passenger's life, and the hydraulic pump is a hydraulic system power source, so the hydraulic pump condition monitoring and fault diagnosis is particularly important. Hydraulic pump bearing failure is one common failure mode, due to bearing failure caused by additional vibration relative to the natural vibration of hydraulic pump is weak, making it difficult to fault information decouple from the signal. So far, the pump bearing fault diagnosis is still a lack of very effective method. In this paper, pour in the frequency domain and frequency domain feature extraction, feature extraction difficult to solve bearing problems and solutions with integrated multi-BP network fault diagnosis and identification and robustness issues.
1, the hydraulic pump bearing fault feature extraction
for mechanical systems, where failure would cause the system is the additional vibration. Vibration signal is a dynamic signal, which contains the information-rich; it is suitable for fault diagnosis. However, if additional vibration signals due to the inherent signal interference or the interference of the fault signal is large and flooded, then how to extract useful signal from the vibration signal becomes very critical.
According to teratology theory, when the bearing surface of the inner flow, the outer ring raceway and the emergence of an injury on the roller, smooth the surface of raceway destruction, damage whenever the roller roll point, will produce a vibration. Assumed bearing parts of a rigid body, without considering the impact of contact deformation of the roller along the raceway for the pure rolling.
Hilbert transform for signal analysis in time domain signal seeking the envelope in order to achieve a smooth power spectrum to highlight the fault information. Define the signal: the optimal profile. Real cestrum envelope model is obtained from the sensor signal cestrum analysis, and its cestrum to extract the signal envelope, which highlights the dual nature of failure information for small signal to noise ratio provides fault feature extraction basis.
2, integrated fault diagnosis of BP network theory
neural network structure is characterized by the area of problem solving decisions. As the complexity of fault diagnosis system, the neural network fault diagnostic system design will be the organization of large-scale neural networks and learning problems. To reduce the complexity of the work, to reduce the network learning time, this will be a collection of troubleshooting knowledge into several logically separate subset, each subset and then broken down into several sub-set of rules, then a subset of rules to organize the network. Each rule is a logical subset of independent sub-network mapping, the link between a subset of the rules, right through the sub-network system matrix. Independent of each sub-network by BP learning algorithm for learning and training, respectively. As the decomposed sub-networks much smaller than the original network and the problem of localization, and thus greatly reduce the training time. BP n in the aviation industry, the hydraulic system performance directly affects the safety of aircraft and passenger's life, and the hydraulic pump is a hydraulic system power source, so the hydraulic pump condition monitoring and fault diagnosis is particularly important. Hydraulic pump bearing failure is one common failure mode, due to bearing failure caused by additional vibration relative to the natural vibration of hydraulic pump is weak, making it difficult to fault information decouple from the signal. So far, the pump bearing fault diagnosis is still a lack of very effective method. In this paper, pour in the frequency domain and frequency domain feature extraction, feature extraction difficult to solve bearing problems and solutions with integrated multi-BP network fault diagnosis and identification and robustness issues.



Other News:
Bearing with solid standard Superagriseal seal
Hydraulic pump bearing fault diagnosis method research network
Bearing failure analysis
Industrial FAG bearing damage and judgment
Substandard products bearing rings remedial measures
The function of the bearing parts, structure features and technical requirements
SKF bearing is available judgment
Spherical Bearings Series Introduction