基于可拓控制的燃料电池焊接机器人多能源动力系统多模型切换控制研究

燃料电池具有比能高、环保等优点，将其引入焊接机器人供电系统，有望解决无缆焊接机器人发展中的能源动力限制问题。将燃料电池和小容量的辅助供电装置共同组成机器人的多能源动力系统，需要解决燃料电池动态响应的时滞性和能量双向流动的难题。同时，此动力系统是一个多输入多输出、强耦合的非线性系统，保证动力系统高效平稳运行也是其中的难点与关键。本项目提出基于可拓控制的多模型切换控制方法，首先基于模糊聚类方法建立燃料电池、辅助供电装置等子系统的混合神经网络模型，在此基础上分别建立动力系统负载平稳、负载突增、能量回馈等多个运行状态下各典型工作区的多能源动力系统的能量流模型；然后为各能量流模型设计底层模型预测控制策略；最后设计上层可拓控制器，通过对底层控制策略进行自适应切换优化能量流模型，从而实现系统的高效稳定运行。

For the characteristics of high energy density and high efficient protection for the environment, the fuel cell can be employed to the welding robot power supply for solving the driving energy restrict the development of the no-cable welding robot. The hybrid power system of the robots, which is constituted by the fuel cell and the auxiliary power supply device with small capacity, needs to solve the problems of time delay and bi-direction power flow. The power system is a kind of multiple input-multiple output nonlinear system with strong coupling. Therefore, it is difficult improve the efficient steady operation of the power system. This project presents a multiple model extension control method. Firstly, the hybrid neural network subsystem model is built based on the fuzzy clustering theory. Then energy flow models of multiple energy sources power system are established in the states of the typical work area, such as load stability, load increases, load anticlimax, energy feedback and other operational state. The predictive control strategy of the underlying model is designed for energy flow models. Finally, in order to realize the efficient and stable operation of the system, the upper extension controller is designed which can optimize the energy flow models based on the control strategy of adaptive switch.