The Quality Inspection System Designing Of Bevel Gears In Micro-vehicle Main Reducer

The bevel gear set plays an important role in main reducer assembling of automobile, whose manufacturing error and assembling error totally affected the whole quality of main reducer. Before assembling main reducer, we can get rid of some inferior bevel gears by detecting their meshing status. The original bevel gear detecting system assumes manually operated methods, which identifies status of bevel gears by intensity of noise from their meshing. In order to improve its detecting efficiency and accuracy, we refurnish the original detecting system and develop its automotive and intelligent capacity.When the bevel gear set is meshing, vibration signal and noise signal are homologous signal, so the variation of noise features are direct response to the variation of vibration features. The vibration signal caused by bevel gear set meshing includes the malfunction information of bevel gear, thus, the amplitude and frequency of vibration signal are served as the detecting criterion of bevel gear set. The major research work of this dissertation includes the following contents.1. Mechanism of noise from bevel gear set in main reducer, relationship between vibration and noise, error types of manufacturing bevel gears were illustrated in this dissertation.2. Electrical control system was refurnished based on the original machine, and PLC system was added to the new system. The PLC system was programmed, and then the machine can work automatically, the testing results can be displayed.3. Acceleration sensor, charge amplifier, data acquisition card and industrial PC were used to design vibration signal acquisition system. The software system of testing bevel gear set was developed by multithreading technology, based on Windows development platform and C++ builder.4. The vibration signal was processed by using time domain, frequency domain, inverse spectrum, demodulation spectrum, wavelet analysis, etc. According to the characteristics of vibration signal from bevel gear set meshing, the methods of signal feature vector extraction were researched based on based on various time-frequency statistics and the wavelet analysis. 5. Establishing BP neural network model, the effective mixing eigenvector serves as neural network's input vector. Using known samples to train BP neural network, the network have associative memory and pattern matching function to attain the purpose of malfunction detection and diagnosis.In practical application, the refurnished bevel gear set testing system can work effectively and accurately, which has an important significance in industrial manufacturing and testing.