نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشگاه علوم کشاورزی و منابع طبیعی ساری- دانشکده مهندسی زراعی- گروه مهندسی مکانیک بیوسیستم
2 دانشیار گروه مهندسی مکانیک بیوسیستم، دانشکده مهندسی زراعی، دانشگاه علوم کشاورزی و منابع طبیعی ساری، ساری، ایران
3 استادیار گروه مهندسی مکانیک بیوسیستم، دانشکده مهندسی زراعی، دانشگاه علوم کشاورزی و منابع طبیعی ساری، ساری، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
The interaction of wheel and soil has received much attention from the researchers of this sector due to its effects on energy consumption and soil characteristics, especially in the field of agriculture. In this research, the amount of wheel indentation in the soil and the rolling resistance of the moving wheel, which are the determining parameters in the wheel-soil interaction, were measured using a wheel tester in the laboratory soil bin. Tests in 2 different levels of speed (0.386 and 0.879 km/h), 3 different levels of tire pressure (18, 25, and 32 lb/in2) and 3 different levels of vertical load on wheel (150, 300, and 450 kg) were carried out in the form of a randomized complete block design (RCBD) in 3 repetitions and a total of 54 treatments. Then, using adaptive neuro-fuzzy inference system and multivariable regression model, the amount of indentation and rolling resistance were predicted. In order to evaluate these models, correlation coefficient and mean square error were used. The results showed that inflation pressure, vertical load, and forward speed of the moving wheel have significant effects on the wheel indentation in the soil (P<0.01). They do not have correlation coefficient in predicting the amount of indentation and rolling resistance by adaptive neuro-fuzzy inference system models was equal to 0.99 and 0.69, respectively, which was much higher than the correlation coefficient in regression models (0.87 and 0.40, respectively). Also, the mean square error in adaptive neuro-fuzzy inference system models regarding the indentation and rolling resistance of the wheel was 0.0231 mm2 and 0.0101 kgf2, respectively, which is much lower than the mean square error in the regression models (0.864 mm2 and 0.918 kgf2, respectively). Therefore, adaptive neuro-fuzzy inference system models have higher accuracy and less error than regression models.
کلیدواژهها [English]