Font Size: a A A

The Study Of Batch Process Monitoring And Fault Diagnosis Based On The Theory Of Multivariate Statistic And Its Applications

Posted on:2004-12-18Degree:MasterType:Thesis
Country:ChinaCandidate:Q LiuFull Text:PDF
GTID:2121360092981265Subject:Chemical Engineering
Abstract/Summary:PDF Full Text Request
Batch process control, fault detection and diagnosis and the theory of multivariate statistics are studied in this thesis. The main purpose of this thesis is to put forward the approach of batch process monitoring and fault diagnosis based on the theory of multivariate statistic and its applications. Some satisfying results have been made by applying those methods to batch polymerization process control and its fault diagnosis.(1) The development of batch process control is reviewed. Classification of process data, batch process control and its application are introduced. Some newly research areas related to batch process control are also briefly presented.(2) The Principal Component Analysis (PCA) and Partial Least Square Regression (PLSR) methods are applied to batch process monitoring and fault diagnosis, and the two types of multivariate statistics~SPE (Squares Prediction Error), Hotelling T2 梐re systematically defined and analyzed.(3) Multi-way PCA (MPCA)--the method of batch process monitoring and fault diagnosis-is discussed systematically. The method to deal with three-way batch process data by MPCA is introduced. Satisfying results are obtained by applied MPCA to batch process monitoring and fault diagnosis and a new method to exclude outliers is proposed.(4) To fill up missing data MPCA assumes that the future deviations from the mean trajectories will remain for the rest of the batch duration at their current values at time interval k . Comparing to it, a method assumed that the future observations are in perfect accordance with their mean trajectories is proposed.ABSTRACT(5) Comparing to MPCA a new way to handle three-way batch process data is proposed and a method of batch process monitoring and fault diagnosis-AMPCA-is developed. Good results are obtained by applying AMPCA to batch polymerization process monitoring and its fault diagnosis.Finally, a summary of the thesis is given and some conclusions are made through the results of the research. Further research interests in the future are prospected.
Keywords/Search Tags:Multivariate
PDF Full Text Request
Related items