Immune-inspired self-adaptive collaborative control allocation for multi-level stretching processes

For multi-level stretching processes in large industrial production lines exhibiting complicated non-linear dynamics and interconnected control variables, a novel immune-based self-adaptive collaborative control allocation (ICCA) method is proposed to deal with the tension output fluctuations and some uncertainties. The allocation strategies for coupled stretching ratios at different levels are implemented in the ICCA to achieve the best stable tension performance. The stretching ratio of single level is determined by minimizing the error between the proposed reference model and the actual stretching plant. Compared with distributed control for individual sub-processes, the reversing dynamic programming on each level is introduced which is operated from the last level sequentially. The gain-tuning strategy is directly driven by the optimization result, namely, the self-adaptive allocation algorithm, is just the execution level for operating the solutions of reversing optimization. The optimization and self-adaptive controller are designed to cope with the presence of actuator imperfections and tension tracking fluctuations in the neighboring non-linear process. In addition, the criterion is an important factor in evaluating the control performance of this system, which consists of the objective function. Simulation results show that the ICCA method exhibits better performances than the centralized PI control and the cytokine network-inspired cooperative control in dealing with the desired tension tracking and fluctuations reducing by the control algorithm.