中文摘要：

对多变量耦合且存在时变性的焦炉集气管压力系统，提出了一种基于支持向量机结合自适应PID的控制方法．采用支持向量机逆系统的方法来进行解耦控制，使得MIMO（多输入多输出）的集气管压力系统解耦成相互独立的SISO（单输入单输出）伪线性子系统．对于解耦后的SISO（单输入单输出）系统采用单神经元自适应控制算法，实时在线调整PID参数．仿真结果表明该控制策略实现了集气管压力系统的动态解耦控制，迅速跟踪变化，提高了系统的快速调节能力和稳态精度，增强了系统的鲁棒性，可以保证焦炉集气管压力稳定在现场工艺要求的范围之内．

英文摘要：

The control method based on support vector machine combined with adaptive PID is proposed for the multi variable coupling and the existence of the coke oven gas collector pressure system. The support vector machine （SVM） inverse system method is used to decouple the control system.The MIMO＇s （ Multi input and Multi output） manifold pressure system is decoupled into independent SISO （Single Input and Single Output） pseudo linear subsystems.For the decoupled SISO （Single Input and Single Output）system, the single neuron adaptive control algorithm is used to adjust the PID parameters online in real time.The simulation results show that the control strategy can realize the dynamic decoupling control of the gas collector pressure system, which can improve the system＇s ability of fast adjustment and steady state accuracy, and enhance the robustness of the system. It can ensure that the coke oven gas collector pressure is stable in the range of field process requirements.