(73c) Perform Better Planning and Scheduling of VCM Furnace Decoke-Shutdowns with Enhanced Capability of Load Balancing, through Predictive Modeling

Cracking of EDC in furnaces produces coke as a byproduct, this coke layers retards heat transfer and increases tube metal temperature and pressure drop across the coil. The accumulation of deposits on the surface of the coil necessitates periodic removal of this coke by steam-air decoking. Thus after a few months of operation, furnace shut down is taken for decoking of the furnace. In cracking furnaces, it is difficult to estimate the remaining run-length of the furnaces and is generally determined intuitively with the help of process parameters trends, furnace coil pressure drop and TMT. This makes it difficult to plan and schedule decoke shutdowns for these furnaces. Any unplanned decoke-shutdown, will ultimately lead to productivity losses. In addition, absence of coke monitoring, load balancing is performed intuitively compounding the issue of unplanned decokes.

For a VCM unit in USA, a forecasting model was developed to predict the furnace run-length based on critical operating parameters. This model is trained on historical data and is continuously fed with live data from the plant. The model is a dynamically auto-tuned model making it receptive to changes in plant operating conditions. Therefore, in case of any sudden increase in slope of pressure drop, it will immediately alert the operator to take actions to control the coking rate.

This model is supplemented with a deviation tracker which will also help the operator identify the coil and tubes where coking rate is higher, so that appropriate actions can be taken by adjusting the burners near that tube.

To make the overall application more utilizable in decision making, a “What-if” feature was built to help predict the furnace run-length at different feed rates. For example, it can indicate that the furnace run-length will increase/decrease by 10 days if the feed is decreased/increased by 5% today. This will help in better production planning of the furnaces and will help in aligning the furnace end-of run dates with the planned decoking dates.

Application Utilization:

1. Monitor furnace coking rates and implement operational changes to control the coking rates and achieve desired run length.
2. Perform “What-if” provision to predict future furnace performance at different feed rates, for better planning and scheduling of furnace operation and decoke cycles.
3. Perform informed load balancing of the furnaces, with combined insights from above evaluations.

Application Benefits:

• The model has been made online in the discussed unit few months ago, and is performing as expected.
• Quantitively, potential benefits of implementing this model in the unit are estimated to be ~ $0.5 million/annum.
• Qualitatively, augmenting the intelligence of field/panel engineers with such a tool will help in pro-active and continuous improvement of furnace operation leading to better asset utilization without requiring any major CAPEX with enhanced reliability and financial benefits.