Genetic programming for prediction a power tiller vibration transmitted to driver during pulling a trailer at rural asphalt road

Power tillers are the main sources of power supply in small and medium size farms. These machines generate high levels of vibrations due to lack of suspension system. Harmful ma-chine vibration has been widely reported and has become a serious problem due to its nega-tive impacts to human health. A capability to predict the occurrence of machine vibration with an acceptable accuracy would, clearly, be beneficial to machinery manufacturers. The vibration acceleration signals were obtained in a field experiment using a 13-hp power tiller. Experiments were conducted at five levels of engine speed (1400, 1600, 1800, 2000 and 2200 rpm), four levels of transmission gear ratio (2 heavy, 2 light, 3 heavy and 3 light), four posi-tion of accelerometer (wrist, arm, chest and head) and in the lateral, longitudinal and vertical directions in transportation conditions for the asphalt rural road. In scope of this study, we have introduced a data driven evolutionary algorithm (i.e. Genetic Programming) for predict-ing the machine vibration values. A detailed analysis of the results, obtained from the GP models, shows that the vibration acceleration increases by an increasing in engine speed at all gear ratios and, it achieves the highest value at vertical direction in all tests. The allowable exposure time, also, was higher in heavy gear ratio and lower engine speed. The acceptable performance of the new GP model in present study, in terms of the statistical criteria, proves the accurate predicting capability of the new developed model for predicting the machine vi-bration values.

Keywords:

Genetic programming , power tiller , prediction , transportation mode