| The prediction of key parameters of chemical processes provides important guidance and early warning for safe,smooth and economic operations of chemical processes.Taking into account of the inherent complexity and nonlinearity of chemical processes,especially large time delays among process variables caused by integrations of pipelines and vessels,it is a great technical challenge to establish an effective method for predicting key parameters of chemical processes.The fuzzy cognitive map is a new method for soft computing capable of knowledge representations and reasoning.Motivated by these observations,the thesis proposes a time-delay estimation based fuzzy cognitive map for predictions of chemical process key parameters.The research contents and achievements of the paper are presented as follows.1.An analysis of advantages of fuzzy cognitive maps in knowledge representations is carried out along with studies on structures and transfer functions of traditional fuzzy cognitive maps.In response to the shortcoming of temporal information treatments,a fuzzy cognitive map based on time delay estimations is proposed.Through the improvement of map structures and temporal information treatments,the established fuzzy cognitive map is more in line with the actual process operations,accurately describing the relationship among process parameters.2.Considering the time delay involved among chemical process variables and the fuzziness of relationship between key process parameters and related variables,a time delay fuzzy cognitive maps based prediction method for key parameters of chemical processes is proposed.Empirical knowledge and gray correlation analysis are employed to design the structure of fuzzy cognitive maps and cross-correlation functions are used to estimate overall time delays.Finally,intelligent optimization algorithms are used to learning model parameters in building the time-delay estimation based fuzzy cognitive maps.The effectiveness of the proposed method is verified by comparisons with other prediction methods.3.The proposed method is applied to an industrial coal gasification drum process.Process data of plant are having been collected to predict the key parameter,drum level.Satisfactory results have been achieved,showing that the proposed method is feasible for practical engineering applications. |
