Application Research Of Multivariable Statistal Process Control

After long time running, there are some changes in any producing system, which will influence the quality of products unavoidably. Besides strict supervision and reliable design, the setting up of a sound process supervision system can ensure and improve the quality of products. Multivariable Statistical Process Control (MSPC) is a modeling method only depending on the process past data for this purpose. In company with the advance of automation level and the increase of computer memory capacity in working site, the collection and storage of process data lay foundation for MSPC.The method to apply the statistical analysis theory in supervision of quality of product in manufacture process is Statistical Process Control (SPC), which is the effectual tool to get eligible product and the foundation for supervising process performance and fault diagnosis. Single Variable Statistical Process Control (SVSPC) and MSPC are both SPC, the shortage of SVSPC is that it only notices the value of one variable at some moment, not suiting to analyze the multivariable process data with interrelation among them. Founding on the space mapping, MSPC can deal with multi-dimension data, solving a set of problems appearing in practice. But there are some problems in MSPC.From the application of MSPC, this dissertation investigates the application of MSPC and its shortage, gives settlement and tries to explain the means of PC, making MSPC more practical. The main research work includes as following:1. Firstly, we introduce the concept and technique of SPC (Statistical Process Control), summarize the application and development of SPC and sum up it's characteristic.2. Principal Component Analysis (PCA) and its technique are introduced, the and the essential of PCA is analyzed. The meaning and means to interpret the practice means of PC are brought forward. Basing on theory above, PCR (Principal Component Recursion) and PLS (Partial Least Squares) are presented. For batchprocess, Multi-way SPC is the appropriate technique, and Glide Windows Multi-way SPC is put forward.3. Compared with the SVSPC and its problem and shortage, the main statistical variables and technique of MSPC are emphasized. Multi-dimension data is transformed to the analysis of statistical variables themselves basing on these statistical variables4. Basing on the model of synthesizing of Phosphonic Acid Dimethyl, we realize the MSPC on this process through simulation.5. Glide Windows Multi-way SPC is applied hi the model of batch cooling crystallizer, and it works well.Finally, the whole thesis is summarized, and some future research areas are highlighted.