Distillation: Do You Understand Partial Pressures?

Non-azeotropic mixtures can cause problems in ancillary equipment.

As we work on various refining trains and columns, we oftentimes take for granted how and where they work. Partial pressures are one of the basic facts in separations. However, what happens when a partial pressure causes poor performance in ancillary equipment? Many of us have performed binary separation column calculations and compared the saturation curves for two components to try to optimize how we design and run our columns. A third component or non-condensable gas can wreak havoc with both types of calculations.

Without warning, the familiar sound of a full-blown surge echoed through the warehouse.

Jake’s Story

Jake had been asked to assist with analyzing a new refrigerant to solve a short-term supply issue with a chlorofluorocarbon (CFC) that was to be phased out. On the surface, the refrigerant looked good. The three-part mixture promised similar performance to one of the primary three refrigerants used in centrifugal chillers. In the process, Jake learned to two new terms, non-azeotropic and glide. He understood azeotropic, i.e., two substances when combined acting as a single gas. The plant used the technique for columns with extraordinary levels of water as a way to finesse partial separation. It also had been used with several previous refrigerants.

Non-azeotropic substances, as he learned, due to their different saturation curves could serve to simulate performance of a single refrigerant. This would work as long as there was a constant flow with no residence place that could cause the gases to separate. The initial analysis looked good. Based on the refrigerant holding together, Jake’s calculations indicated it would be a suitable replacement.

He also learned that for non-azeotropic mixtures, at any given temperature, the liquid has the mixture composition at the bubble point while the vapor has the composition at the dew point — and that “glide” reflects how the gap between the boiling and dew points changes with temperature and represents the composition shift across the saturation dome.

Jake began calculating based on the evaporator side that acted like a reboiler with one composition coming in and another composition exiting in the gas phase. That meant the low boilers in the mixture were vaporized and the higher boilers were only partially boiled off so most would remain in the evaporator. Jake moved through several iterations and concluded the mixture eventually would not be compressible by the centrifugal compressor and the unit would surge and probably trip. This lead to the decision to limit the use of slug-flow-type systems such as direct-expansion refrigeration where no liquid gas residence interface could be established.

However, Jake’s good friend Gavin challenged Jake’s theory and set up a test at a vendor shop. The shop hooked a refrigeration unit to a test loop and filled the unit with the refrigerant mixture. Gavin invited Jake to witness the test; he was present from the start. After checking everything, Gavin gave the operator the go-ahead to start the unit. Jake held his breath. The compressor started with the usual snarl as it passed through the surge line and on up to operating conditions. The unit held and started to work down the evaporator pressure. Then, almost without warning, the familiar sound of a full-blown surge echoed through the warehouse. Several similar rounds followed as the compressor worked to regain the head required to condense the mixture. It then shut down on a low evaporator pressure trip. Gavin looked disheartened.

Jake asked, “Well, what did we learn from this?”

Gavin quipped, “Not to challenge Jake’s theories.”

Jake replied, “No, you had the guts to do that and now we have confirmation!”

So, what happened? The mixture contained the major high-pressure gas, HCFC22. As the compressor lowered the evaporator pressure, the HCFC22 preferentially boiled off along with small quantities of the other two gases; the compressor drew away the gases and pushed them into the condenser. Unfortunately, the condenser needed the sum of the partial pressures and resultant temperature to start condensing the vapor. Because the compressor didn’t have the capability to compress the HCFC22 to its saturation pressure, flow was reduced and the compressor crossed the surge line, setting up the instability. As a result, the compressor tripped.

There are many situations where you can use partial pressures to investigate system faults or design systems for energy optimization. Brush up on your understanding of the partial pressure calculations and begin optimizing or troubleshooting. Happy energy hunting!

Earl M. Clark, PE, – Engineering Manager, Global Energy Services. Clark retired from DuPont after a career of 39 years and 11 months and joined Hudson’s Global Energy Systems Group as Engineering Manager. During his over 43 years in the industry, he has worked in nearly all aspects of the energy field; building, operating and troubleshooting energy facilities for DuPont. He began his energy career with Duke Power and Clemson University during the energy crisis in the 1970s.

Active in both, the American Society of Mechanical Engineers and the American Society of Heating, Ventilating, Refrigerating, and Air-Conditioning Engineers (ASHRAE), Clark was chairman of ASHRAE's task group on Halocarbon Emissions and served on the committee that created ASHRAE SPG3 - Guideline for Reducing Halocarbon Emissions. He has written numerous papers on CFC alternatives and retrofitting CFC chillers. He was awarded a U.S. patent on a method for reducing emissions from refrigeration equipment. He has served as technical resource for several others.