Jake was perplexed. The evaporator on his -75°C refrigeration machine indicated signs of fouling on the secondary refrigerant side. The secondary refrigerant was a form of paint stripper designed to keep metal surfaces clean. How could the evaporator possibly be fouled?
[pullquote]
The plant produced a specialty product requiring a separation column operating at -75°C. The new facility that replaced the old one also used the -75°C secondary refrigerant in its columns. Temperatures were drifting up; as a result, the columns were operating at higher-than-design pressures. Capacity wasn’t restricted but there was concern; flow had been increased to the overhead condensers, which helped.
The plant, built during World War II, provided limited system intelligence. The refrigeration system consisted of four cascade refrigeration machines operating on CFCs that were at pressure and temperature limits. With near total vacuum, the system was prone to inward leakage that would introduce non-condensables to the refrigerant. This, in turn, caused high pressures in both condensers. The facility routinely purged the condensers and considered hunting leaks a full-time job.
The easy answer to capacity loss would be the non-condensables, but Jake decided to do a full analysis. Jake set about doing a field performance test. He already had the inlet and outlet brine temperatures, although he was concerned about their accuracy. He lacked interstage temperatures or pressures. Each compressor was a three-stage with equal diameter wheels. Each refrigeration machine had two stages of economizers. Jake was excited to test and analyze the results.
Jake installed test RTDs on the machine. The upper stage required removing the thermometers and installing the RTDs. To install RTDs on the lower stage, Jake had to chip away at the ice ball and then melt the ice with alcohol. Each of the economizer stages included test thermowells. Pressure gauges, limited as they were, were calibrated prior to the test. The compressors had no instrumentation; Jake improvised by chipping the ice down to the case and installing RTDs so they touched the case at each stage entry point. The temperatures were close to what Jake had plotted out on the pressure/enthalpy chart during his pretest analysis.
Jake acquired the original design data for the 40-year old refrigeration system and began his analysis by modeling the system. The manual computations were difficult and required iterations until a solution was reached. With that in hand, Jake began his test fully expecting to find the problem was non-condensables.
During the test, Jake checked the condenser conditions. Instead of showing problems, the temperature of the condensate was very close to the saturation pressure of the refrigerant. No non-condensables!
With manually logged data sheets, Jake analyzed data. A large paper spreadsheet included the averages of readings taken over the one-hour test period. Jake calculated the heat transfer of each heat exchanger. The condenser and the intercooler/condenser showed close-to-design performance but the evaporator showed a much higher temperature difference versus design. Obviously, this was making the compressors work harder to cool the secondary refrigerant.
A novel design for the 1930s, the heat exchanger featured a low charge system. Refrigerant circulated from the bottom of the heat exchanger to a distribution deck in the upper part of the exchanger. From there, the refrigerant dropped to the tubes below and passed over the staggered tubes on its way to the bottom. This was to minimize the submergence penalty of having flooded tubes. Analysis showed high fouling in the tubes.
The only way to find out was to open up the tube side and inspect the tubes. The tubes were carbon steel so iron oxide was a possibility. When the heads were removed, Jake found a very smooth bright red bore on each tube. It wasn’t rust but was thick and acted like fouling.
Checking with plant personnel, he learned that leakage on the secondary refrigerant had been a problem years ago. R&D wanted to test a new substance that used a red pigment to highlight leaks. Unfortunately, the substance had plated out pretty uniformly on the tubes.
Jake had the tubes hydroblasted clean; the refrigeration machine returned to service with a significant improvement in performance. The secondary refrigerant was tested for any residual pigment. Over the next several months, the plant shut down each of the other three refrigeration machines to clean the tubes. Performance went back to design.
So, be thankful you have computers not only to model and analyze test data but also to hook up test rigs for data logging. When you test, keep an open mind and expect the unexpected. Happy energy hunting!
