CAJ | 학술논문

针对阳离子聚合反应器的温度分布建模与控制问题,提出了一种基于B样条神经网络的广义PI控制方法.首先采用B样条复合网络建立分布函数的动态和静态模型,并基于该模型,将分布函数的跟踪问题等效为动态权值向量的时间域跟踪问题.最后给出一种新型的广义PI控制方法,实现对给定温度分布的跟踪控制.同时,为了更好地抑制未知干扰、参数摄动以及模型不匹配等问题,模型权值状态、模型输出与实测温度分布所对应的权值误差都被引入到反馈控制回路,因此能够大大增强系统的鲁棒性与抗干扰能力.仿真结果表明该方法的可行性.
침대양리자취합반응기적온도분포건모여공제문제,제출료일충기우B양조신경망락적엄의PI공제방법.수선채용B양조복합망락건립분포함수적동태화정태모형,병기우해모형,장분포함수적근종문제등효위동태권치향량적시간역근종문제.최후급출일충신형적엄의PI공제방법,실현대급정온도분포적근종공제.동시,위료경호지억제미지간우、삼수섭동이급모형불필배등문제,모형권치상태、모형수출여실측온도분포소대응적권치오차도피인입도반궤공제회로,인차능구대대증강계통적로봉성여항간우능력.방진결과표명해방법적가행성.
Model of temperature distribution in a tubular polymerization reaction is developed using a B-spline neural network,in which both dynamic and static network are applied to resolve the modeling of distribution function from a high dimensional data set.Based on this dynamic network model,a new-type generalized PI control algorithm has been studied.Then a control problem for distributed system is reduced to a tracking problem of nonlinear dynamic weights,which separates the time and the space effectively.In order to restrain unknown disturbances and parameter perturbation,not only the weights state of the network model are turn into feedback,but also the output error vector between the model and the real process is introduced at a certain percentage.This provides a feedback channel for the control,and therefore the robustness and anti-disturbance performance is largely enhanced.Simulation results demonstrate the effectiveness of the proposed method.