INEOS Chlor is one of the major chlor-alkali producers in Europe and a global leader in chlorine derivatives. Our Runcorn site in the northwest of the U.K. has two large chlorine plants: the original J Unit that uses mercury cells and a more-recent membrane unit. Both yield hydrogen as a byproduct. This gas can be sold or used as a fuel for onsite boilers that generate power.
Building the Membrane Chlorine Plant (MCP) involved major restructuring of the site hydrogen system; as a result there was a period when not all the hydrogen from the new plant could be used in the boilers and was vented. When hydrogen is vented in this way, additional natural gas has to be purchased to replace it, which could have increased the annual fuel bill by several million pounds.
Continuous improvement of our manufacturing processes has enabled the site to achieve "best in class" cost and environmental performance. As part of this improvement program we wanted to minimize vented hydrogen and maximize the value of this resource at both plants.
This wasn't as simple as just redirecting all available hydrogen to the boilers. Their burners are very sensitive to hydrogen pressure — wide variations can lead to boiler trips. Such variations can stem from startup or shutdown of a hydrogen compressor supplying external customers, shutdown of a chlorine production stream, or a sudden change in output from such a stream as may occur when an individual electrolyzer is switched off. The MCP process itself also is sensitive to boiler startups and trips, which can cause back-pressures that can damage the membrane.
The task therefore was to increase the hydrogen to the boiler while simultaneously protecting both the boiler and plant from pressure disturbances.
Minimizing Pressure Variations
Events that create disturbances in the system are inevitable. So, we needed to develop an advanced control scheme that could minimize resulting pressure variations. Our approach was a mixture of feedback or proportional-integral-derivative control and predetermined valve movements to keep the differential pressure within defined bounds.
To implement this control scheme, we took advantage of the DeltaV automation system already controlling the boiler plant and both chlorine plants. It's flexible and powerful and offers advanced control capabilities ideal for our approach.
First we built a mathematical model of the process so we could run dynamic simulations of the entire system. We tested several control scheme designs with the dynamic simulator. This enabled us to fully understand if each design was accurate and whether it would work.
In particular, this testing process revealed the need to completely isolate a hydrogen stream if a boiler tripped at a significant load. To cope with such a trip, the export valve closes rapidly and the vent valve opens to a load-dependent position that allows it to vent all current hydrogen production.
We selected the control scheme that promised the best results; our process engineer verified it would meet both production and safety requirements. We then configured the solution in the control system — a task made easier by the DeltaV system's wide range of available function blocks and flexibility to write custom code as needed. You can do pretty much whatever you want so long as you have the imagination to take advantage of its capabilities.