基于PSO和ESNs的马铃薯贮藏库温度预测控制  

Potato Storage Temperature Prediction Control Based on PSO and ESNs

作　　者：刘俊霞[1,2] 熊新荣[1] 胡兵[1] 

机构地区：[1]新疆工程学院电气与信息工程系,乌鲁木齐830023 [2]新疆大学信息科学与工程学院,乌鲁木齐830046

出　　处：《系统仿真学报》2016年第7期1701-1705,共5页Journal of System Simulation

基　　金：新疆维吾尔自治区高校科研计划资助重点项目(XJEDU2012I43);新疆维吾尔自治区高校科研计划青年教师科研启动基金(XJEDU2014S074)

摘　　要：针对马铃薯存储现状及其加工企业原材料和人力资源的严重浪费问题,结合新疆大罗素农业科技开发有限公司马铃薯储藏库温度数据的非线性、时变性等特点,采用常规控制方法控制效果不理想。提出了新的温度预测控制方法,它是用粒子群优化算法滚动优化输入控制量并得到实际输出值,用ESNs对储藏库内温度进行预测,利用预测输出和实际输出的偏差来对系统反馈较正。仿真结果表明:提出的温度预测控制方法是有效的,不仅在控制效果上优于LS-SVM预测控制,还有良好自适应性以及对扰动信号有较好的鲁棒性。According to the current situation of potato storage and serious waste of raw materials and human resources of potato processing enterprises, combined with temperature data characteristics of potato storage from Xinjiang Great Russell Agricultural Science and Technology Development LTD., and control effect of the conventional control methods which is not ideal, a new temperaturepredictive control method was proposed, which rolled to optimize the input control variable by using particle swarm optimization algorithm and get the actual output values, use ESNs to forecast the temperature in the storage, the system was corrected by deviation between forecasting output and actual output. The simulation results show that the proposed temperature control method is effective, not only the control effect is better than that of the LS-SVM predictive control, but also has good adaptability and robustness to disturbance signals.

关 键 词：粒子群优化算法 回声状态网络 温度预测控制 马铃薯储藏库 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]