As a common part of mechanical equipment to transmit power and motion, gearbox has been widely used in modern mechanical equipment, such as metal cutting machine tools, aviation, electric power system, agricultural machinery, metallurgical machinery and so on, due to its advantages of fixed transmission ratio, compact structure and high transmission accuracy.
However, due to the complex structure and bad working environment of the gearbox, the gearbox is prone to failure in mechanical equipment. Especially in today’s rapid development of computer and electronic technology, mechanical equipment continues to develop in the direction of large-scale, high-speed, automation and intelligence, and the failure and failure of gearbox cause more and more losses to production and society.
Due to the high speed, heavy load, extreme working temperature, pollution and other harsh conditions, the gearbox is vulnerable to all kinds of damage. The unexpected failure of gearbox will destroy the whole mechanical system, cause huge economic losses, even catastrophic failure. Therefore, the research of gearbox condition monitoring and fault diagnosis technology is of great significance to reduce and avoid the whole set of equipment stop work caused by component failure, the transformation from post maintenance, regular maintenance to condition based maintenance, and the extension of gearbox service time and service life
Righteousness. At the same time, it is also of great significance in ensuring the personal safety of staff, improving the quality of maintenance, reducing spare parts reserves and reducing unnecessary economic losses.
In the past decades, people have made great efforts in developing various fault diagnosis methods by using feature extraction and classification methods. Generally speaking, gearbox fault feature extraction method can be summarized into three stages: data acquisition, fault feature extraction and fault pattern recognition. In these three stages, the gearbox feature extraction is the key, which directly determines the final classification results.
Entropy is an index to detect the dynamic change of time series, which has been widely used in the fault diagnosis of rotating machinery. The vibration signal of healthy gearbox is unstable and entropy value is large; The instability and entropy of vibration signal of fault gearbox are small. Firstly, the feature extraction algorithm based on multi-scale fuzzy entropy (MFE) is used to extract the fault features of gearbox; Then the fault features are input into the k-nearest neighbor (KNN) classifier to automatically identify the health status of the gearbox. The gear fault simulation experiment is carried out on the gear box test-bed, and the experimental scheme is designed. Through the analysis of the experimental results, the effectiveness of the vibration diagnosis method in the gear box fault diagnosis is verified. The effectiveness of the mfe-knn method is verified by comparing with the traditional diagnosis methods, and the quantitative diagnosis of different types of gear faults is realized.
A fault diagnosis method based on multi-scale fuzzy entropy and KNN method is proposed. This method can extract fault information from signals with strong background noise. The experimental results show that the proposed method is superior to multi-scale sample entropy and traditional time and frequency domain methods, and five different fault types are successfully identified by KNN classifier, which provides a new idea for gear fault diagnosis.
