THIS MONTH'S PUZZLER
An old bank of three Freon chillers provides cooling throughout our plant. The units are manually controlled. Operators frequently turn on too many units in preparation for anticipated heat loads, causing trips in the chillers that delay production because of upward temperature swings. As part of a corporate-sponsored energy campaign, funds are available for the chillers. Where would you start? How can we get more out of these old units?
CONSIDER AN ICE BANK
It is most likely that this problem is reported from a batch processing plant where wild swings of refrigeration load are experienced. I can suggest two solutions to this problem: 1) A programmable logic controller (PLC) should be used to control the Freon compressor bank. Depending on the load, as sensed by return chilled-water temperature or return Freon pressure, the compressors should start and stop automatically. Use an automatic algorithm in the PLC to rotate the sequence operation so that the load is shared among the compressors. 2) If the system works by circulating chilled water cooled by Freon, an ice bank system can be considered. This will even-out the load on chillers. The Freon will produce ice on the coils in the ice bank when the plant load is low. When the load is high, the return chilled water will be warmer and will rapidly melt the ice, thus suddenly releasing a lot of "cold." This will have a beneficial load-leveling effect. This is a method used in breweries where ammonia is used as the refrigerant to cool circulating chilled water.
Either or both of the above solutions can be used depending on the nature and severity of the problem or availability of funds.
Sebastian Thomas, engineer
Sai Dem America, Houston
DO THE ANALYSIS
There could be a number of issues: aging chillers, poor control and lack of operator training. You may need more of them as a standby train or 50%-on/50%-off concept if not implemented already. Indiscriminate loading of the chillers is proof of a need for improved operator training for a better understanding of the system usage and capabilities. First take some time to analyze the system loads and make automation improvements. This is important for a successful revamp your people are planning.
Sameer Chawla, lead process engineer
Worley Parson, Kuala Lumpur, Malaysia
START WITH PROCESS CONTROL
As a minimum, sequence the units based on temperature and pressure. Control the demand for cooling by temperature and raise the pressure as needed in response to increase in flow to meet the temperature. It's a simple cascade. (Using this approach with a cooling tower, I was able to cut electrical costs by 10%.) A step beyond this would be a load-following PLC, separate but in communication with the plant PLC.
But, a larger issue is at stake: production integrity. As the figure shows (via the large diamonds), the Freon chiller has a separate controller from the PLC running the plant. This is typical of old equipment. Either the controls are proprietary or the project engineer was lazy. The plant needs these key inputs. So, it's no surprise the chiller crashes unexpectedly given the secondary controller is responsible for preventing freeze damage.
I have seen this before with chillers, furnaces and heat tracing. Proprietary PLCs don't talk to the distributed control system (DCS). Here, there isn't even an alarm on the Freon temperature loop. It appears a low-pressure alarm on the brine flow to the evaporator is the only alert the board operator would receive; there is always a lag in the temperature, so the temperature indicators on the brine to and from the evaporator won't respond until the chiller has been down for several minutes. When systems fail, the board operator is flying blind. These controls must be paralleled off so status is reported at the PLC or DCS.