An extensive validation of computer simulation frameworks for neural prognostication of tractor tractive efficiency

摘要

To optimize power and energy resources in field operations, there is an inevitable demand for attempts to be conducted regarding determination of performance parameters of tractor-implement. Computer simulation is a potential assistive tool for researchers concerned in this realm. This is the first investigation covering comparative ability of computer simulation frameworks, based on artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS), for neural prognostication of tractor tractive efficiency during tillage operations. In tillage operations, forward speed (2-6 km/h), plowing depth (10-30 cm), and tractor mode (two-wheel drive (2WD) and four-wheel drive (4WD)) were considered as main treatments affecting tractor tractive efficiency. Among several computer simulation frameworks developed in the investigation, the best ANFIS framework yielded coefficient of determination, root mean square error, mean absolute percentage error, and mean of absolute values of simulation residual errors of 0.987, 1.857%, 2.314% and 1.582%, respectively. On the account of obtained statistical parameters, the best ANFIS computer simulation framework was validated as the more distinguished framework than that of the ANN. The ANFIS simulation results revealed that increment of plowing depth and forward speed caused nonlinear decrement of tractor tractive efficiency in each tractor mode. Moreover, physical perception obtained from the integrated ANFIS simulation results indicated that application of the 4WD mode rather than the 2WD mode increased tractor tractive efficiency. Therefore, it can be asserted that the ANFIS simulation results improved state of the art in domain of studying tractor tractive efficiency.