سال انتشار: ۱۳۹۱

محل انتشار: دومین کنفرانس بین المللی آکوستیک و ارتعاشات

تعداد صفحات: ۷

نویسنده(ها):

Rahman Abdolahzadeh – Rahman Abdolahzadeh, Dept. Agro-technology of Abouraihan College, University of Te-hran, Tehran, Iran.
Seyed Reza Hassan-Beygi –
Hossein Ahmadian –
Mohammad Aboonajmi –

چکیده:

Harmful machine vibration has been widely reported in the power tillers in on-farm and rural roads, and has become a serious problem due to its negative impacts to human health. There-fore, vibration control and reduction measure or operators protection management is essential to be implemented on these machines. Prerequisite of power tillers vibration control and re-duction is to know about root mean square (RMS) values of vibration emitted from them. A capability to predict the machine vibration with an acceptable accuracy would clearly be very beneficial to machinery manufacturers. This study presents an application of artificial neural networks (ANN) for prediction of a power tiller vibration transmitted to operator body in transportation mode. The vibration acceleration signals for evaluation were obtained in a field experiment using a 13-hp power tiller. Experiments were conducted at five levels of en-gine speed, four levels of transmission gear ratio, three direction levels and two asphalt and dirt rural types of road in transportation conditions at wrist, arm, chest and head positions of operator body. The RMS values of the recorded signals in the time domain were calculated. The results showed that multi layer perceptron networks with a training algorithm of genetic with number of epochs of 1000, population size of 50, and maximum generation of 100 were the best for accurate prediction of RMS values at wrist, arm, chest and head positions of the power tiller operator body. The minimum RMSE for the three-layer perceptron network with a hyperbolic tangential activation function, with four neurons in hidden layer was 0.0946. The correlation coefficient of the network was 0.84 for wrist, 0.94 for arm, 0.99 for chest and 0.98 for head of operator.