Example 3 Diagnosis of Operating Problems in a Batch Polymer Reactor

Comprehensive Example 3 Diagnosis of Operating Problems in a Batch Polymer Reactor 

This example illustrates how numeric-symbolic interpreters are combined with symbolic-symbolic interpretation for root cause diagnosis. It 

-statistics over time for A = 1 to S3 minutes for reactor data. 

Consider again a batch polymerization process where the process is characterized by the sequential execution of a number of steps that take place in the two reactors. These are steps such as initial reactor charge, titration, reaction initiation, polymerization, and transfer. Because much of the critical product quality information is available only at the end of a batch cycle, the data interpretation system has been designed for diagnosis at the end of a cycle. At the end of a particular run, the data are analyzed and the identification of any problems is translated into corrective actions that are implemented for the next cycle. The interpretations of interest include root causes having to do with process problems (e.g., contamination or transfer problems), equipment malfunctions (e.g., valve problems or instrument failures), and step execution problems (e.g., titration too fast or too much catalyst added). The output dimension of the process is large with more than 300 possible root causes. Additional detail on the diagnostic system can be found in Sravana (1994). 

The overall objective of the system is to map from three types of numeric input process data into, generally, one to three root causes out of the possible 300. The data available include numeric information from sensors, product-specific numeric information such as molecular weight and area under peak from gel permeation chromatography (GPC) analysis of the product, and additional information from the GPC in the form of variances in expected shapes of traces. The plant also uses univariate statistical methods for data analysis of numeric product information. 

---