(76b) A Practical Approach to Advanced Process Control of VCM Plants

Vinyl Chloride Monomer (VCM) is an industrially important chemical whose demand is growing at a fast pace worldwide. VCM is produced by thermal cracking of Ethylene Di-chloride (EDC), which is a fouling service that gradually degrades the performance of cracking furnaces and therefore necessitates periodic shutdown and clean up of cracking furnaces.

Multi-variable Predictive Control (MPC) is a commonly applied technology to processing units to improve the stability, productivity and efficiency of the process. However, MPC cannot be readily applied to VCM plants because some of the key control variables, such as individual furnace EDC conversion, VCM production, coke thickness and tube skin temperature are difficult, if not impossible, to measure.

High fidelity EDC cracking yield model that could accurately predict EDC conversion, VCM production, coke buildup and tube skin temperature in EDC cracking furnaces does exist and could be used in an on-line fashion to provide continuous feedback to MPC. One such approach and its application to Advance Process Control of VCM plants are presented in this paper.