Research on an optimal parameter prediction method based on digital twin considering forklift multi-factor coupling of human and environment

In order to solve the problem that the conventional multidisciplinary coupling analysis for forklift stability could not be run on embedded system with low power, an optimal parameter prediction method based on support vector regression (SVR) was proposed. The proposed stability mechanism fully considered the complicated multi-factor coupling through the inertial avigation system quantifying the operation habit and surface evenness. A multidisciplinary coupling simulation model of forklift was built with Simulation X software, based on steering and multi-body dynamics. The optimal operational parameters of forklift are aiming at maximizing the operation efficiency under stability condition. PSO algorithm was used to iteratively solve the optimal operational parameters. Further, two acceleration strategies which include excellent solver of CVODE and multicore processor were proposed. On the basis of above parallel optimization, the SVR training set of the optimal parameter was obtained. In experiment, the flight control system was installed to collect data, which were the input of SVR. The experiment showed that the angle variation of the flight control system is basically consistent with the calculation. The method of parallel optimization was 20 times more efficient than Simulation X3.8. The SVR prediction time was only 1.45 s. The SVR mean square error of maximum load capacity was only 151.99 kg 2 and the determination coefficient (R2) reached 0.98. SVR test result showed that it has excellent prediction efficiency and accuracy.