CAJ | 학술논문

针对粒子群优化算法(particle swarm optimization, PSO)在高维空间复杂曲面寻优时易于陷入局部最小值的问题，分组扰动粒子群优化算法(partially-perturbed particle swarm optimization, PPSO)结合问题特征，采用启发式规则，实施参数分组扰动策略，对 PSO 算法进行改进，从而增大了跳出局部极小的可能性。本文主要研究 PPSO 在精馏塔模型参数闭环辨识上的应用，分别针对模型参数可辨识性，参数的不同分组，鲁棒性进行分析验证；并在其他精馏塔模型上进行了相应的验证。仿真实验表明，PPSO 辨识算法比序列近似法等其它辨识算法具有更高的辨识精度，并且具有较强的鲁棒性；在其他精馏塔模型参数辨识上 PPSO 算法也同样取得了很好的辨识精度。实验结果证明了 PPSO 算法在精馏塔模型参数闭环辨识上的可行性和有效性。
침대입자군우화산법(particle swarm optimization, PSO)재고유공간복잡곡면심우시역우함입국부최소치적문제，분조우동입자군우화산법(partially-perturbed particle swarm optimization, PPSO)결합문제특정，채용계발식규칙，실시삼수분조우동책략，대 PSO 산법진행개진，종이증대료도출국부겁소적가능성。본문주요연구 PPSO 재정류탑모형삼수폐배변식상적응용，분별침대모형삼수가변식성，삼수적불동분조，로봉성진행분석험증；병재기타정류탑모형상진행료상응적험증。방진실험표명，PPSO 변식산법비서렬근사법등기타변식산법구유경고적변식정도，병차구유교강적로봉성；재기타정류탑모형삼수변식상 PPSO 산법야동양취득료흔호적변식정도。실험결과증명료 PPSO 산법재정류탑모형삼수폐배변식상적가행성화유효성。
To sovle the problem that the particle swarm optimization algorithm(PSO) is prone to fall into the local minimum especially optimizing in high dimensional space and complex surface, the partially-perturbed particle swarm optimization algorithm(PPSO) uses huristic rules combing with problem characteristics to carry out the parameters group disturbance strategy. Trough improving the PSO, the possibility of avoiding the local minimum is increased in the PPSO. In this paper PPSO is applied for the closed-Loop identification of the parameters of the distillation column model. Different experiments are designed to do verification the identifiability of the model parameters, different groups of parameters and robustness. The PPSO is applied in different distillation column model in this paper. Simulation results show that,PPSO has a higher identification accuracy comparing with other identification methods such as sequential loop closing method, and has strong robustness. PPSO also obtains the good precision in identifying the parameters of other distillation column model. The simulation results demonstrate the the feasibility and effectiveness of the PPSO in parameter identification of distillation column model.