Process Fault Diagnosis Using Artificial Intelligence and Partial Process Knowledge
碩士 === 國立清華大學 === 化學工程學系 === 87 === On-line data acquisition system will record mass operating data during the continuous chemical process running. Much of process knowledge was hidden under these data. When a process fault occur, the operator may not notice the fault but the computer has...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
1999
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/41482139502916272352 |
| Summary: | 碩士 === 國立清華大學 === 化學工程學系 === 87 === On-line data acquisition system will record mass operating data during the continuous chemical process running. Much of process knowledge was hidden under these data. When a process fault occur, the operator may not notice the fault but the computer has already record the fault on the storage media. Process knowledge can be explored by analyzing of these routine data. However, the analysis of routine data always relies on skill or experience of process engineer. Now, the application of artificial intelligence let it possible to analyze mass process routine data by computers. Our research is about using the method of machine learning, which is called ID3 algorithm to generate rule from routine data and use as rule base for fault diagnosis. Here we simulate CPC xylene distillate process in order to prove the practicability of using inductive classification algorithm ID3 on chemical process.
|
|---|