“What do you mean, the column is upside down?”
Jake and Miles have been colleagues for many years. Both have experience in due diligence and integration during mergers and acquisitions. When transitions were underway, usually some consternation arose when the quick due diligence performed at the start of the process missed key items.
[pullquote]
Miles and Jake had gone to lunch one day and Miles opened up with, “You won’t believe what happened with the new plant, or let me rephrase, you won’t believe what I found.” Jake leaned forward to hear the news.
Miles had kept Jake in the loop regarding an acquisition and the problems with one of the batch distillation systems. In the early phases, Miles learned that instrumentation was minimal and the batch process ran until a gas chromatograph showed the raffinate had reached the desired purity. He understood the operating method and, while he didn’t like it, it did work — as long as the operators knew when to sample.
Miles had modeled columns most of his career and when he saw the column, he first thought that it could do so much more than intended. He asked for a drawing and flowsheet from the previous owner; there were none. At the first offline period, Miles arranged for a junior engineer assigned to the plant to open some of the column’s access ports. They found the column contained the same packing as had been used for years; mirrors determined the location of supports. With those data, they could start modeling the column and hopefully improve its capabilities.
Using the data, Miles built a model. One key element of the more-efficient design was a runaround loop that transferred the heat gain from the overheads condenser to the boil-up heater. However, at higher rates it wasn’t sufficient and because Miles planned on refining different mixes from the original, it wasn’t compatible. So, he designed a system to separate the two operations. Also, he added a full instrumentation package with remote data acquisition and online gas chromatographs during the project stage.
Column Startup
The column started up after the usual project delays. A couple of the junior engineers on Miles’ team stayed at the site to follow the operation and determine corrective actions when problems occurred. All went well and the column was put in to service.
The team noted a significant pressure difference between the top and bottom of the column. They attributed it to possible pressure gauge errors and continued to press forward. Additionally, the pressure/temperature relationships in the column didn’t correspond to the design. Little or no flow went overhead and the overheads condenser didn’t seem to have any load. Maybe the boil-up heater was offline. The team found it working but not at much load. The outlet temperature was at design. At that point, the site engineers called Miles for help.
Miles reviewed the data in his office. He offered a couple of suggestions and then made reservations to travel to the site, scratching his head all the way. On the flight, Miles kept turning the data over and over in his mind. He did a pressure/temperature plot from top to bottom. He laid out the saturation pressure at each point he would expect based on his model. That was strange — the difference in pressure seemed to be a straight line.
By the time he arrived at the site, Miles had a couple of possible solutions. After a calibration check of the temperature sensors and then the pressure sensors, Miles observed the column operating for a few hours. At that point, he said, “Shut it down, empty and clear it. We need to look inside.”
While the shutdown proceeded, Miles sat with the site engineers and explained what he thought was happening. The difference in pressure at each level correlated to the height of that point in the column. Correcting the height for the density of the liquid worked out to be the difference in pressure; essentially the column contained standing liquid.
Removing the column’s top confirmed Miles’ suspicion: the tails section of the column had been installed upside down. The restriction only allowed a small amount of liquid to go overhead, thus filling the column with liquid. Plans were made to bring in a crane for the lift. Once the column was inverted and all the piping reinstalled, it worked as the model predicted.
