Causality analysis using process data.

• 主要作者: Phettongkam Numan.
• 其他作者: School of Chemical and Biomedical Engineering
• 格式: Final Year Project
• 語言: English
• 出版: 2009
• 主題: DRNTU::Engineering::Chemical engineering::Processes and operations
• 在線閱讀: http://hdl.handle.net/10356/15783
• 機構: Nanyang Technological University

Causality analysis methods are initiated by multivariate statistics used in other fields. Due to its strong capability to identify the relationship between the variables and also the prediction, many efforts have been spent to run these methods in the chemical industry field. Nearest neighbor method, with its simplicity to understand and implement, has been applied in this research to analyze the root-cause in the chemical plant. Simulation based on simple relationship between 2 variables has shown that nearest neighbor method is able to analyze the relationship between 2 variables in term of influence. It is able to analyze the root-cause in industrial case study. However, there is still minor error with regard to the identification of the relationship between variable that is not physically feasible in the process. Thus, nearest neighbor method cannot be applied to industrial chemical process data. Further developments by supplementation with other methods should be done to improve the validity.