Experimental research and prediction of screw dosing based on least squares support vector machine
CUI Shoujuan1, ZHANG Xiliang1, XU Yunfeng2, SU Qiang1, SUN Xiang1, ZHANG Yu1
1.School of Mechanical Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, China; 2.Key Laboratory of Modern Agricultural Equipment and Technology, Ministry of Education, Jiangsu University, Zhenjiang, Jiangsu 212013, China
The feeding precision of screw feeder is determined by various factors, such as the rotational speed, filling rate and so on. It is usually difficult to calculate the accurate feeding quantity through a precise mathematical model. In this paper, we predicted the feeding quantity through machine learning based on three factors. The most important factor was the rotational angle, the other two factors were rotational speed and filling rate. The least squares support vector machine(LS-SVM)was used to learn the nonlinear relationship between the feeding quantity and these factors. To achieve the best performance, the parameters of LS-SVM were tuned through the cross-validation. A real feeding experiment was carried out to verify our method. The experimental results demonstrate that the prediction values of LS-SVM are consistent with the real values. Moreover, the precision of our LS-SVM is better than the estimation from theoretical calculation and BPNN. The prediction error of our method is only ±0.02 if the filling rate is close to 1. In summary, our method has high precision and can be applied to the control of screw feeder.
CUI Shou-Juan-,ZHANG Xi-Liang-,XU Yun-Feng- et al. Experimental research and prediction of screw dosing based on least squares support vector machine[J]. Journal of Drainage and Irrigation Machinery Engin, 2016, 34(12): 1099-1104.
LONG Wen, LIANG Ximing, LONG Zhuqiang. O2 content in flue gas of boilers predictive control based on hybrid PSO and LSSVM[J]. Journal of Central South University(science and technology), 2012, 43(3): 980-985.(in Chinese)
WAN Changjiaing, JIANG Song. Application of compu-ter aided design(CAD)for designing components in rice whitener[J]. Food and feed industry. 1997(8):10-12.(in Chinese)
VAPNIK V,CHAPELLE O.Bounds on error expectation for support vector machines[J].Neural computation,2000,12(9): 2013-2036.
SCHOLKOPF B, SMOLA A J. Learning with kernels[M]. Cambridge, MA: MIT Press, 2001: 222-226.