基于免疫-单神经元PID算法的纸浆浓度控制

Pulp Consistency Control Based on Immune-single Neuron Algorithm

中文关键词:  纸浆浓度  免疫机理  单神经元PID算法  增益自调整

Key Words:pulp consistency  immune mechanism  single neuron PID algorithm  gain scheduling control

中文摘要:

      针对单神经元PID算法中的增益K不能自调整引起的动态响应慢的问题,提出了一种将免疫算法与单神经元PID算法相结合的控制算法——免疫-单神元PID算法。依据T细胞免疫机理调节单神经元PID算法中的增益K,使增益K获得自调整功能,以改善单神经元PID算法的动态性能,提高其学习速度。仿真结果表明,该算法可克服纸浆浓度控制过程中存在的多干扰性、时变性、非线性等缺点,能够满足纸浆浓度控制的稳定性、快速性要求。与单神经元PID算法相比,该算法响应速度具有明显的优越性,并具备了单神经元PID算法本身较强的抗干扰能力以及自学习自适应的能力。“THJSK-1”平台上的实时控制也验证了免疫-单神经元PID算法的可行性。

Abstract:

      The gain K of single neuron PID algorithm lacks self adjustment ability and thus the dynamic response time is longer. Aiming at this problem, a new algorithm combining immune mechanism and single neuron method was proposed. T cell immune tuning mechanism could regulate the gain of single neuron PID algorithm. So the gain had the ability of self-adjustment. The system dynamic response performance was proved and the learning rate of single neuron PID algorithm was accelerated. Simulation result showed that the new algorithm could overcome the shortcomings in pulp consistency control which was characterized by large time-lag and time-varying. It also could satisfy the accurate and rapid control requirement of pulp consistency. Compared to single neuron PID algorithm, the new algorithm had the obvious advantage in the system response rate. At the same time, it succeeded the advantages of single neuron PID algorithm which had the ability of strong anti-interference, self learning and self adaptation. The real-time control on THJSK-1 experiment platform indicated this control algorithm was feasible.