Process optimization is one of the more pleasant and interesting tasks a professional chemical engineer can be called to perform.
For simplicity we can give you our experience with some of the less complex optimization tasks performed. Hopefully one or more of the remarks will help you in your daily optimization efforts. An optimization job usually follows the following path:
- Define the deliverable. Exactly what does the owner want to have in his or her hand when the work is over?
- Define the objective. When the study is over, the modifications made and the bills paid, what will be different? Although we have yet to decline a task at this point, we sometimes find that obtaining definition of the objective to be a very difficult task.
- Collect the data. As part of the data gathering step, along with obtaining the operating data and records and the material use reports, we ask as many people involved in the operation as we can what the problem is, what the solution is, what has been tried, and what were the results. Even though this sounds more like journalism than engineering, we do NOT simply compile answers and publish the collective opinion. We do this because there may be a very diverse set of opinions regarding what the troublesome operation is supposed to do and what it actually does, and those opinions may conflict with the recorded process data. Understanding this conflict has on occasion helped us determine the root problem with an operation.
- Next we ask for the material balance for the existing operation. If one is not available we generate one based on the process data.
- Now that we know what the flow streams are, we ask the direct question "What are the flow steams supposed to be?" The answer to this question may require considerable discussion and agreement by operating forces but it is the key to correcting the problem. In our experience have found a variety of situations such as:
- There was no problem at all. Once everyone understood what was happening, the situation was accepted.
- A tank level control in a receiving tank "would not work". A fluid was being delivered to the tank in regular surges. The level controller manipulated the speed of the output pump. No matter what control loop tuning parameters were tried, the tank level would not hold steady. After several interviews it became clear that the operating and maintenance personnel did not understand the function of the tank and controls. The process down stream of the tank needed a very steady input. Engineering had installed this tank and pump to receive the anticipated surges and deliver a steady stream to the downstream process. The controller was only to modulate the average pump speed to keep the average upstream and downstream flow rates matched. Once the operating and maintenance forces understood how the system worked, operating procedures were revised and this step in the operation ran smoothly.
- "Too much" carrier liquid was being lost at a recovery filter belt. Operating personnel had made several attempts to reduce filter cake output moisture by adding dewatering devices prior to the operation. The material balance quickly revealed that the filter was reducing the moisture from 80% liquid to 8% liquid at any loading rate the process could sustain. Reducing input moisture to 60% moisture or even 50% moisture had no effect on the liquid lost. The losses were directly related to a product rate increase (8%), and the only corrective action was to replace the filter with a much more intensive dewatering devise. The installed filter operation was simply accepted once operating forces understood the situation and the cost of the more effective equipment.
- An engineer changed a tower reflux to enhance overhead quality, which was achieved. This changed flow rates in the overhead by five fold, increased tower pressure drop, and caused large venting losses. Once it was determined that the overhead quality did not need to be increased, the operation was returned to previous conditions.
We repeatedly find that good people doing a good job have a very hard time objectively analyzing their work. We also repeatedly find that the problem that is not clear on the pump curves, the heat exchange calculations, the pipe flow estimates, the vapor liquid diagrams, or vendor cut sheets are almost always clear on the material balance, one of the most powerful tools for the professional chemical engineer.
If you have good people doing a good job and are still in need of process optimization, consider professional chemical engineering.