A Real "Old Wives Tail"

A friend of mine hails from the West Coast of South Africa (just north of Cape Town) and recounted this story to me a while ago. Being close to the sea, fresh fish was always a big part of their menu, and for years he observed his wife trimming 2 inches of prime flesh off the tail while preparing the fish to put in the oven. One day he thought to ask why the strange preparation ritual, and was told that this was the way her mother had taught her, who had been taught by her mother et cetera. This was a tradition in their family going back to her great grandmother and she had never thought to question it. On further investigation it seems that her great grandfather had been a fisherman, and would always return home from a trip with a large fresh fish. Unfortunately the largest dish they had was 16 in wide so she had to trim the tail to fit it in the oven. This physical limitation had evolved over time into a strict requirement ingrained into the cook.

This got me thinking of a practice I see in many of the plants I visit - throttling the pump to limit the amps drawn by the motor. As in the case of the fish story above, the original reason for throttling the pump has often disappeared in the midst of time. Peering through our looking glass, perhaps some of the following reasons could have applied :

What is the problem with throttling anyway. Anyone who has attended a course on pumping system efficiency will know that the specific energy of a system increases when valves are throttled. This is because the additional friction loss through the valve adds additional head to the pump duty, thus consuming more power in order to pump the same amount of fluid. This is clearly shown in the diagram below, where the amount of useful energy (the energy being used to provide actual work) is only a small proportion of the total energy being consumed by the pump.

Fig 1. A typical system where a valve is used to control the pump flow rate. As the valve is closed, friction loss though the valve creates additional head, forcing the pump left on its curve and reducing the flow supplied by the pump. The amount of wasted energy in this system is represented by the size of the red block.

In cases where the pump is being throttled to reduce the flow rate to a downstream process (pumps are over sized and deliver more flow than the system requires), throttling the pump can be seen as a practical, though inefficient, solution. However when the system demands a higher flow rate than the pump is currently delivering, as in most cooling water systems, and yet we still throttle the pump, then this is just throwing money away.

A case study : Throttled pumps at a steel mill

Steel mills produce a huge amount of heat, and the cooling water systems are often some of the biggest consumers of power in the plant. One such system investigated consisted of 8 pumps in parallel, of which 6 pumps were running, feeding cold water through the shells of the continuous arc furnace. During the plant walk we noted that 3 of the 6 pumps were throttled 50%, while one valve was almost 75% closed. According to the operators the reason for the throttling was to keep the amps drawn below a certain value, but they did not know what value this was, and the amps could only be measured in the motor control centre. They just knew to close the valves to the positions as stated above.

The max rated current for the 3.3 KV motor read off the nameplate was 60.5 amps, but when we measured the actual amps drawn in the MCC (motor control centre), they were all around 56A.

We sequentially opened each of the valves and checked that the amps drawn were less than the max rated amps for the motor (they were well within this), and ended up being able to shut down one pump when all the valves were 100% open, and still achieve the same flow rate. Running 5 pumps instead of 6 led to savings of 258 kW of energy, translating to 2.1 GW.hrs of reduced energy cost every year, for zero cost of implementation.

Why were the pumps being throttled? One possible reason was that the operator was told to keep the discharge valve set to 50% closed, because at some point in time this corresponded to maximum amps on the motor. But as we know things change, and if they are not measuring the amps, then how do they know that at 50% closed, the amps are at their maximum.

Even if there was a valid reason for limiting the amps or throttling the valve, does that reason still apply today, or do we continue to throw away 2 inches of prime fish because this is what we have always done.

Fig 2. Pumps in parallel with the discharge valves throttled, to ensure that amps drawn are kept below max rated amps. The author has found numerous instance where the amps drawn is far less than max rated amps for the motor, and it is possible to fully open the valve and still not overload the motor. Fig 3. Some ammeters have a red needle that can be preset to the maximum amps (as above), or drawn on with a red pen. If you observe that the amps drawn is very close to the maximum, there is a good chance that this pump is being throttled. Make sure this value is the maximum rated amps for the motor.

All of the above assumes that the pumps are being throttled for no valid reason. What happens in cases where the motor is drawing more amps than its maximum. Surely this is a case when we are justified to throttle the pump. Wrong! There are other ways to fix the situation, and none of them involve wasting energy through throttling. However this is a topic for the next article when we will discuss energy efficient ways to reduce pump output, such as trimming the impeller, reducing the pump speed or changing the pump.