CAJ | 학술논문

针对多输入多输出（MIM0）热工过程的非线性、强耦合、变工况及参数时变等特点，提出了一种基于系统输入输出数据和模糊自适应竞争聚类的模型辨识新方法．该方法首先依据系统的各个典型运行工况，使用模糊自适应竞争聚类对输入输出数据进行聚类划分，并对T—S模糊模型进行结构辨识，以确定系统的模型结构和参数；然后采用最小二乘递推算法对模型后件参数进行辨识，同时对结构辨识参数进行精确修正．将所提出的模型辨识方法用于锅炉一汽轮机非线性系统的模型辨识，仿真结果验证了该方法的有效性．
침대다수입다수출（MIM0）열공과정적비선성、강우합、변공황급삼수시변등특점，제출료일충기우계통수입수출수거화모호자괄응경쟁취류적모형변식신방법．해방법수선의거계통적각개전형운행공황，사용모호자괄응경쟁취류대수입수출수거진행취류화분，병대T—S모호모형진행결구변식，이학정계통적모형결구화삼수；연후채용최소이승체추산법대모형후건삼수진행변식，동시대결구변식삼수진행정학수정．장소제출적모형변식방법용우과로일기륜궤비선성계통적모형변식，방진결과험증료해방법적유효성．
Considering the nonlinear, strong coupling, parameters time-varying characteristics and variable work conditions of multi input multi output(MIMO) thermal process, a new model identification method is proposed based on input-output data of the system and fuzzy adaptive competitive clustering, with which the input-output data are firstly clustered according to typical work conditions, while the T-S model structure identified so as to determine the antecedent parameters of the system; then, the consequent parameters of the fuzzy rules are identified using recursive least squares, while the antecedent parameters modified exactly. Application results on the nonlinear model of a boiler-turbine unit prove the model identification method to be effective.