Vibratory Roller Vibration Accelecration Synchronous Detection And Data Processing

Compaction of pavement is an important process in the construction organization. It's proved that compaction could making the pavement with each layer have certain density, can improving the impermeability and strength of the material, reducing the permanent deformation of pavement when it was forced traffic load. Vibratory roller is the main compaction construction machinery, vibratory roller through a centrifugal force which make by the vibratory wheel's eccentric vibrator high speed whirligig to reciprocating shock the pavement, it could change the station of soil particles, so reach the arm of compaction.Because of the traditional method of compaction can't meet the requirements, Vehicle-mounted compaction detector is the main research methods. Tests proved that the process of soil compacting constantly, vibratory roller wheel vibration parameters (vibration displacement, vibration acceleration) and soil compaction are positively correlated, that is, with the increasing degree of soil compaction vibration wheel vibration displacement, vibration acceleration is constantly increasing. So it can fix an acceleration sensor on the vibratory roller wheel, through real-time detecting vibration acceleration to reacting soil compaction.In the process of dealing with the vibration acceleration from the sensor, its main contents are removing interference noise, analyzing singularity of single and making sure the parameters of vibration acceleration. It shows that use of wavelet analysis to deal with this Non-stationary signals has a unique advantage, wavelet analysis could remove the interference noise in high degree, meanwhile, can detect singularity of signals conveniently.Measured data show that the vibration acceleration and soil compaction are positive in the former-medium term of soil compaction, but the linear relationship between them is not very good. Based on BP neural network's ability and characteristics in terms of dealing with nonlinear mapping, it could use of BP neural network to reflect the relationship of them. Simulation and data analyzing showed that, BP neural network is possible.