基于改进蜘蛛群集算法的木薯收获机块根拔起速度优化

针对挖拔式木薯收获机由于较难获得块根拔起速度控制系统最优化控制参数,造成块根拔起速度控制精度较低,木薯块根收获拔断损失率较大的问题,开展拔起机构块根拔起速度控制系统控制参数优化的算法研究。该文以模糊PI作为块根拔起速度控制算法,采用多领域的动力学仿真技术,构建木薯收获机块根拔起机构控制系统的联合仿真模型,以较优块根拔起速度模型为控制目标,块根拔起阻力为条件,开展结合局部搜索算子的蜘蛛群集算法的模糊PI控制参数的优化研究,且进行了田间木薯块根拔起试验。结果表明,结合局部搜索的蜘蛛群集算法比蜘蛛群集算法具有较快的收敛速度和较高的搜索精度,模糊PI控制参数的优化结果,控制参数K_p为0.841,K_i为0.203 9,在优化的控制参数条件下,块根拔起速度能较好跟随较优块根拔起速度,木薯块根的垂直拔起速度与块根拔起较优速度的平均相对误差为4.5%,滑块位移与理论值的平均相对误差为3.7%。

Optimization of tuber lifting velocity of cassava harvester based on improved spider clustering algorithm

Abstract: When the cassava tubers are lifted up by the dig-pull cassava tuber harvester, the harvester has low power consumption and a high adaptability to the soil. However, the control precision of the lifting velocity of cassava tuber is low, thus, the broken and loss rate of the cassava tubers are larger in the cassava tuber harvesting. And the control precision of the control system of the tuber lifting velocity mainly depends on the quality of control system''s control parameters. Whether the optimal control parameters could be obtained by the optimization algorithm of the control system''s control parameters determines the quality of the parameters. Therefore, the optimization algorithm of the control parameters of the lifting velocity control system of the cassava tuber lifting mechanism is studied using the advanced method and technology which has important significance to improve the control precision of the cassava tuber lifting velocity and the harvesting quality of the cassava tubers. The broken and loss rate of the cassava tubers are larger in the cassava tuber harvesting when the control precision of the tuber lifting velocity of the dig-pull cassava harvester is low. Firstly, the co-simulation model of the control system of the tuber lifting mechanism of the dig-pull type cassava tuber harvester was established. The fuzzy PI algorithm was used as the control algorithm of the mechanically optimal tuber lifting velocity of the tuber lifting mechanism. The multi-domain dynamics simulation technology was also used in the co-simulation model. The mechanically optimal lifting velocity model of the cassava tuber was obtained using the cassava tuber lifting tests of the experienced farmers and the optimized velocity model of manually pulling tubers as well as the numerical simulation tests. The mechanically optimal lifting velocity model of the cassava tuber was used as the control target, and meanwhile, the constant cassava tuber lifting force, the cassava tuber lifting force in the soft soil as well as the cassava tuber lifting force in the hard soil, respectively, were used as the condition. The study of the spider clustering algorithm combined with the local search operator was carried out. Then, using a combination of local search operator and spider cluster algorithm, the control parameters of the cassava tuber lifting mechanism system were optimized by iterative optimization. In addition, the common test function was used to verify the convergence and search accuracy of the spider cluster algorithm combined with local search operator. Finally, the cassava tuber lifting test verification was carried out in the field. The error analysis of the verification test was carried out by means of the mean error and the maximum error. The results show that the spider cluster algorithm combined with local search operator which can avoid getting into the local optimal solution in the iterative process, has faster convergence speed and higher search accuracy than the spider clustering algorithm. The spider clustering algorithm combined with the local search operator is suitable for solving the extremum problems of high-dimensional complex function. The optimization result of the Fuzzy PI control system''s control parameters: Kp and Ki are 0.841 and 0.203 9, respectively. Using the optimized control parameter, the actual lifting speeds of the cassava tuber can follow mechanically optimal lifting velocity model. And the dynamic performance is great. The average relative error between the actual vertical lifting speed of the cassava tuber and mechanically optimal lifting velocity of the cassava tuber is 4.5%. The maximum error is 6.7%. The average relative error between actual slide displacement and the theoretical value is 3.7%. The maximum error is 5.4%. The spider clustering algorithm combined with local search operator can be used in controlling the cassava tuber lifting process of the dig-pull type cassava tuber harvester. The control precision of the tuber lifting velocity has high precision.