Pumps and VFDs question

I have a limited knowledge in pumps. What I think I know is, pumps are sized based on the required flow and head pressure. And that pumps have a pump curve which tells you how efficient the pump is based on the flow and head pressure. Right?

Here’s the example for the question - Lets say we have a pump that delivers 100gpm, we have to move 60,000 gallons, so we would run the pump for 1 hour. Now let’s say we have the same pump which is connected to a VFD and the VFD is turned down so that the pump is only delivering 50gpm. So now in order to move the 60,000 gallons the pump will have to run for 2 hours.

Am I correct in thinking that we are using more energy (in the form of electricity) by using a VFD and making the pump run for 2 hours instead of 1 hour? The same amount of work was done, we moved 60,000 gallons, but by running the 100gpm pump at 50gpm we would be lower on the efficiency curve, thereby requiring more power to do the same work.

If I am correct in my thinking, how do I go about determining how much power (money) is wasted by doing something like this?

Tark said:

I have a limited knowledge in pumps. What I think I know is, pumps are sized based on the required flow and head pressure. And that pumps have a pump curve which tells you how efficient the pump is based on the flow and head pressure. Right?

Here’s the example for the question - Lets say we have a pump that delivers 100gpm, we have to move 60,000 gallons, so we would run the pump for 1 hour. Now let’s say we have the same pump which is connected to a VFD and the VFD is turned down so that the pump is only delivering 50gpm. So now in order to move the 60,000 gallons the pump will have to run for 2 hours.

Am I correct in thinking that we are using more energy (in the form of electricity) by using a VFD and making the pump run for 2 hours instead of 1 hour? The same amount of work was done, we moved 60,000 gallons, but by running the 100gpm pump at 50gpm we would be lower on the efficiency curve, thereby requiring more power to do the same work.

If I am correct in my thinking, how do I go about determining how much power (money) is wasted by doing something like this?

Click to expand...

If you're doing the calculations, don't forget to include the watt losses from the VFD (I suspect they will be substantial in this case). If you run the pump at 100% do you still need to use the VFD, or do you have the option of using a motor starter (or contactor) instead of or to bypass the VFD?

The type of pump must also be considered. Centrifugal pumps do not operate very well at lower than desgin speeds. I tried it once and learned my lesson. If you need to limit the flow of a centrifigul pump, I usually install a return line with a control valve on it, or on the discharge side of the pump and clamp the control valve so that it never fully closes.

Well, I simplified the example. I like having a general understanding first so I can ask the correct questions later.

Real situation – We have a pump skid which has 2 pumps each rated at 400gpm. The pumps are vertical turbines. One pump is connected to a starter, the other to a VFD. The pump skid is rated at 800gpm. We have numerous valves connected to the pump skid. Each valve will flow 50gpm. So if I open 16 valves, the pump skid should be pumping at 800gpm. So lets say I have to move 80,000 gallons, if I open 16 valves it should take 1 hour to do so. Now lets say I only open 12 valves, so the pump skid will be pumping 600gpm and will need to run for 1.3 hours to pump 80,000 gallons. The operation of the pumps in this situation would be one pump full speed, the other pump at X speed to make the required flow.

Am I correct in thinking that by running the pump skid at 600gpm that it is going to cost more (electricity) that if the pump skid was run at 800gpm? If so, how would I go about determining how much?

First of all, don't confuse power with energy. It takes a certain amount of energy to lift a certain weight of water to a certain height. (Note: volume x density = weight)

Power is the rate of energy used in a period of time. If you pump that water twice as fast you need twice as much power, but the total energy used is the same.

Electric utilities bill small customers on energy used, and larger customers on both energy used (consumption) and peak power required (demand).

So, if your application is a straight static lift and your are billed for consumption only then the VFD won't save you anything. If the application has a lot of friction head and you are billed for demand as well as consumption, the VFD could save a lot of money.

This is a problem that I deal with regularly, and it takes some analysis to determine the true savings.

Ken, the answer to your problem is to limit the minimum speed of the pump and/or the minimum safe flow. This varies depending on the pump and whether or not the discharge is mostly static or mostly dynamic (friction loss or friction head). You can also cook a pump with a recycle line or a discharge throttling valve. It is the internal recirculation of the fluid in the pump at too low a flow rate that builds up heat and causes the damage.

The pump should be run in the best part of the pump efficiency curve to operate at maximum efficiency.

I work with a lot of pool people and, although I do not know a lot about their business, they do not like running pumps on VSDs if they can help it. They also "throttle" back the pumps by partially closing valves to obtain the desired pressure/flow/efficiency.

I’ve done controls for several closed loop cooling towers and I’ve found that it is generally better to trim the impeller (that means chuck it in a lathe and trim the OD) to achieve the flow rates needed. This is more energy efficient than throttling the pump output, and a VFD on this kind of a system is more trouble than its worth, instead I use soft starters in order to avoid hammer. The pumps are sized as nearly as close as possible before the impeller is trimmed so energy savings from a VFD would be negligible, if at all due to semiconducter heat losses and efficiency losses.

I did the program for a 3 pump municipal system which had two pumps on starters and one on a VFD. I found that the pump on the VFD had to be operated at no less than 80% speed. If speed dropped below 80% the pressure roll off on the pump curve was so steep that the VFD would go hunting. So it saved slightly on operation costs during low usage times, but it turns out that most of the time all three pumps were running. I suspect that the initial cost of the VFD and controller was truly never recaptured in energy savings.

I think I found what I was looking for.

According to a book I have read there are 2 horsepowers in relation to a pump. "Pump input or brake horsepower is the actual horsepower delivered to the pump shaft. Pump output or hydraulic horsepower is the liquid horsepower delivered by the pump."

So to my situtation - If I have a 100gpm pump and slow it down so that it is pumping 50gpm, it's going to cost more to run that pump at 50gpm than at 100gpm since the pump would be running at a lower efficiency on the pump curve. Although, depending on the pump curve, this cost might be a lot or might not be much at all.

Right?

You can't talk about a pump curve without the system curve. Where the two intersect defines the operating point. If you have an application with a high static head (lift) then you may not save much. If you have an application with high friction head, it may.

I've attached a sample analysis for illustrative purposes, but the topic can get quite involved. In one case, if the process only needs 4,000 gpm a constant speed pump with a throttling valve would take almost 200 hp. If you used a VFD it would only take 87 hp. In another case if you needed to pump 80,000 gallons and it didn't matter how long it took, you could pump at full speed for ten hours at 8,000 gallons per minute for a total energy use of 2,300 hp-hrs (1,700 kW-hrs). Or you could pump at 4,000 gpm and reduced speed for 20 hours for a total energy use of 1,700 hp-hrs (1,300 kW-hrs). That's because this system has a lot of friction head.

On the other hand if this system was mostly static lift you would need 2,400 hp-hrs (1,800 kW-hrs) at 4,000 gpm. That's more energy!

Similarly you need to look at the reason for varying the flow. If you have a process that requires variable flow rate, then using a VFD may make sense. If you are just filling a day tank or a stand pipe it may not.

You also need to look at your energy bill. If you are charged for consumption only, then the justification for a VFD is lower. If you are charged for demand as well a VFD might save plenty.
Every application is different.

Replying ASSUMINT these pumps are centrifugal and adhere to the afffinity laws/

FIRST Math little off if pump is rated at 100 gpm and you run it for an hour (at the 100 gpm) you will have pumped 6,000 gallon not 60,000.


AFFINITY LAWS
These apply to centrifugal pumps and squirrel cage blowers.
Rule #1 flow is directly proportional to speed ( 2X speed = 2 X flow)
Rule #2 Head is proportional to speed squared ( 2X speed = 4 X flow)
Rule # 3 Power input is the cube of speed ( 2X speed = 8 X power)

Now if you pump 6,000 gallons and do it in an hour you draw Y watts.
IF you pump the 6,000 gallons and do it in two hours by halving speed the power draw should only be one eighth (not counting some losss of efficiency in BOTH the pump and driver). Of course since you did it for two hours (versus one) I think total power is one fourth ( 1/8 x 2 hours).

On many energy conservation pages I have read the recommendation for flow control is to get rid of throttle and bypass systems and use VFDs.

Throttle does drop power somewhat but it does increase head which is proportional to torque. Bypass is just pure waste. They have been used for years and they do work no argument there but they are wasteful.

IF you throttle do NOT do it too much the pump and most important the seals are cooled by water flow - too much throttle = burned seals = leaks.

Dan Bentler

Dan-


Your assertion above assumes you don't need the pressure you were generating at full speed. That's Tom's point. In you example you not only cut your flow rate in half you reduced your pressure to 1/4 what it was. Now you may not have enough pressure.

Keith

Quite interesting.

I am not a pump/mechanical expert but, as mentioned, several of my clients build swimming pools, life support systems for animals/sea creatures in zoos and the like. I am lead to believe one of these guys designed the aquarium life support system for San Diego Zoo. They are not consultants but water movement and filtration experts.

We have a situation at present where a consultant has suggested to one of his clients that they should consider VSDs on filtration pumps to save on energy usage on the main filtration pumps for an olympic sized pool. They would then be able to open all the valves and run the centrifugal pumps at a lower speed to save power.

The water movement expert has explained, in writing, that the valve(s) have been throttled back to adjust pressure and flow and keep the pump in the optimal curve for the device and that this is normal practice in the industry to achieve optimal performance. The consultant has been asked to put his recommendations/instructions in writing, and reasons why. He came back with the greatest load of rubbish one could imagine and at the end of the day has stated that the contractor installing the equipment will be held responsible for any non performance issues.

You can imagine the response from the water movement expert. All others in the field that have been approached have similarly declined the job. They will probably find some poor local electrician that will see a job in his sights and a buck to be made and at the end of the day he will be crucified for non performance.

Saving power is good but only in certain circumstances. One may suggest that if the client wishes to save power he would be better off putting in power factor correction equipment and leaving the pumps alone.

kamenges said:

Dan-


Your assertion above assumes you don't need the pressure you were generating at full speed. That's Tom's point. In you example you not only cut your flow rate in half you reduced your pressure to 1/4 what it was. Now you may not have enough pressure.
Keith

Click to expand...

I completely agree with you. Was trying to stay with the question regarding energy. Course you are not saving energy if at half speed you have inadequate head to pump - you are wasting energy to drive a "stalled pump"(at shutoff head for that RPM) -- and you will take the chance of burning out the seals and overheat the pump -- not cost effective at all.

Dan

BobB, your situation sounds similar to what we encounter in cooling towers. Once the system is set up, the flow rates aren't adjusted, but we need the proper flows and pressures. We size the pumps as closely as we can, then we achive the final flow rates we need by trimming the pump impellers. You chuck the impeller in a lathe and you trim a timy bit off its OD. Refering to the pump curves will help a bit, but its a trial and error process. But once you have it set then you don't ever have to worry about it. We put a tag on the pump showing the impeller diameter for future reference. Trimming the impeller eleiminates the use of wastefull throtteling or bypass. VFDs are usefull when you need to respond to changing system demands - eg, a municipal pumping system. For pools, cooling towers, and the like they don't gain you much.

I saw a situation once where the VFD salesmam sold a bill of goods with a load of b.s. on an irrigation system where I was asked to come in to make some changes a couple of years after the installation. They had a VFD on a pumping system that was transfering water from a canal to a huge 2 acre storage pond. The idea was to hold the pond at a constant level with speed based on discharge rate, but the entire pupose of the pond was because the canal did not have enough capacity to meet peak demand to distribute enough water to the users. The old way of doing it was more than adequate, with an off level switch and an on level switch a couple of inches below that. On a pond that big a VFD was just plain stupid. I think that salesman laughed all the way to the bank.