Research On Intergrated Network Based Control In Industrial Processes

In recent years, the increasingly fierce market competition impels large andmedium industrial manufactruing enterprises to implement higer requirements onenergy, efficiency, quality and costs. In oreder to improve production efficiency, productquality, economic efficiency and market competitiveness, as well as achiving energysaving and cost reduction, intergrated control including all the processes in a productionline has been a hot research topic in the process control system community.Concerning the process demonstrated using state space models, a novel two-layernetwork based architecture is proposed, where the down-layer structure is used tocontrol the local plants while the up-layer is used to track the performance indicesassigning to the integrated plants. By adopting an integral and compensation action, it isshown that the evaluation function can track the operation performance indexasymptotically. Under the network environment, both the delay-free and time-delaycases are considered. And to formulate the synthesis conditions in a numericallyefficient way, several different approaches are also presented. It is shown that theresulting derivations are convex or quasi-convex, which are efficiently sovable withnumerical software.When comes to the process described using input-output models, a novelintegrated structure is proposed based on a modified model predictive control method.In networked environment, both forward and backward communication delays areconsidered. Future control sequence is generated to compensate for the backward timedelay. A model predictor is used to compensate the backward time delay, andcontributes to the peformance evaluation fuctions together with the control sequence.The evaluation function can track the operation performance index asymptotically byoptimizing the objective function of the predictive control. And the analytical criteriaare obtained based on Lyapunov steability theory.