Analysis and Prediction of Noise Pollution of Garden Tractor Using Statistical Methods, Artificial Intelligence and ANFIS

Hearing loss due to harmful sound is one of the most important occupational illnesses. Agricultural mechanization with the use of a variety of machinery causes problems in occupational safety and health for those working in this field. In this study, some of the factors affecting the noise generated by a garden tractor (H.938 Model) were evaluated and predicted. The tests were performed according to international standards of sound such as ISO 5131 and ISO 7216 at different engine speeds, gear positions, road type and microphone distance. The total number of treatment combinations for extracting sound signals released by the garden tractor reached 216 with 72 treatments and three repetitions. The obtained data were analyzed in the form of a factorial test based on a completely random plan. Besides, Artificial Neural Network and ANFIS model was created to predict the sound level in the aforementioned tractor with 216 patterns for training and assessment. The results of the research showed that the effect of the main factors on the level of sound emitted by the garden tractor were significant at probability level of 1%. The results of frequency analysis showed that noise pressure level decreases according to increase in frequency and the sound intensity that reaches the driver's ear is equal to 77.93 dB at the frequency of 4000 Hz. The results of Artificial Neural Network and ANFIS model for forecasting the Sound pressure level of the garden tractor with the highest coefficient of determination (R2) for this research were 99.73 and 99.91, respectively.