摘要
通过对铅锌烧结过程的机理分析,确定了透气性预测具有模型不确定性以及输入变量的不确定性等特点。将基于满意聚类的Takagi-Sugeno建模方法和机理分析方法结合起来解决这些问题,并采用带有梯度加速的混合粒子群算法进行辨识,来解决运算量大和可能出现不可辨识的情况。仿真结果表明,所提出的方法能较好地克服预测过程中出现的不确定性因素,能较准确地预测透气性,具有一定的鲁棒性。
According to the analysis of mechanism of sintering process,the prediction of permeability is verified to having the character with uncertainty in the prediction model and the input parameters.The Takagi-Sugeno method combined with mechanism analysis is employed to resolving the uncertainty.In order to avoid a great deal of computation and non-identifiable problem possibly presented in the traditional identification method,the hybrid particle swarm algorithm with gradient acceleration is applied to identify the parameters of Takagi-Sugeno.The simulation results show that the proposed method can overcome the uncertainty successfully,predict the permeability accurately,and have the robust character to a certain degree.
出处
《湖南工业大学学报》
2007年第6期56-59,共4页
Journal of Hunan University of Technology
基金
国家杰出青年科学基金资助项目(60425310)
关键词
透气性
T-S模糊预测
混合粒子群算法
permeability
T-S fuzzy prediction
hybrid particle swarm algorithm
