System Curve of a Throttled Valve 2

[ericdayo26]

Hi there fellow engineers,

I have a question with regards to a system with a throttled valve.
First let me explain the system,

A centrifugal pump supplies stock (3.5% consistency) to the top of Chest B, from Chest A (both chest are open to the atmosphere).
The Levels on both chest are maintained on a certain level, so we can assume that NPSHa does not really change.
The Pump's nameplate reads; 36.6m, 209m3/hr, 1780rpm. The motor is rated 50Hp, 575volts.
A control valve, 41% open, restricts the flow to chest B, so that it is maintained at 190m3/hr during stable operations, but is sometimes fully opened to during certain situations when Chest A's level are near maximum. The flow when the Control Valve is fully open is 270m3/hr.
I have computed for the Friction head of the piping with the Elevation head, and got 31 meters, which is very near when plotted against the Pumps curve, so i guess my system curve is correct. I have attached my plotted curve.
The pump was evaluated because we saw potential power savings on the pump, because it is normally open at 41% only. And also the pump is an old model, so the boss wanted to replace it with a new one that will have available spare parts.
But as I evaluate the system, i think that the pump is not really wasting energy on the throttled valve, because it is still operating near BEP at a 41% open valve. Perhaps it was designed that way to adjust with varying flow requirement. Plus the fact that it not really a giving problems to the maintenance people.
Anyway, what i dont get is how the system curve works? Because normally, i would suggest an option to replace the pump with a smaller one that is rated at 190m3/hr and 32 meters of head. But based on what i've read on articles about the system curve, it is suggesting that I only need 20 meters of head. Am I right with 190m3/hr x 32 meters? or do i only need 20m?
I'm also suggesting to replace the pump with a variable speed drive, which i think is the best option for this system.
Can anyone confirm help me out? thanks

 

[LittleInch]

Eric,

Well done for a well structured query and attached graphs and data. From your description and pump curve, the only error I can see is the first system curve (the steeper one, which interescts the pump curve at 190 m3/hr. From your description I don't think the system curve changes. What happens is that the control valve at 41% open is represented by a vertical line which goes from the pump curve down to meet the system curve at 190 m3/hr. Hence you have a head loss accross the control valve of about 17m (~1.7 bar). This pressure loss is an energy loss which is why, although the pump is running efficiently, you are not using that fluid energy efficiently and essntially getting rid of half of it as heat into the fluid caused by the pressure drop accross the control valve.

However, you also have another duty point, which is your max flow case (even if this is only occasionally) and therefore your pump still needs to be able to do 270 m3hr at 31m, so if you replaced this pump with a 190 x 20m pump, that is all it would do and hence chest A would perhaps overflow??

If you search VSD pumps on this forum this will give you lots of info and there are various thoughts about this. In my opinion, if the majority of your time is spent at 190m3 at 20m head then yes, a VSD would seem to be worth it, but they are not a free lunch and have extra costs for the VSD equipment, new motor probably, and sometimes power supply issues. Only you can do a full study to see if the power savings are worth it bearing in mind that the VSD equipment itself uses power. I'm sure pthers will guide you well on this subject. You could also consider two pumps, one for the normal flow and an additional pump in parallel or series to cope with the max flow case.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[LittleInch]

This presentation has been posted on other VFD posts by Stannier I think so I take no credit for it, but is very useful.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[TenPenny]

The valve is part of the system, so you don't have one system curve, you have to have system curves that represent the range of opening of the valve. Your two system curves are perfect for describing the two scenarios that you give.

To answer one of your questions, yes, you are 'wasting' energy, by virtue of whatever pressure drop occurs across the valve. This is energy you have put into the stock, and are now simply dumping away. However, that said, what are your options to control the flow rate without using a valve? Your option is likely to use a pump that is sized so that it can reach your maximum flow condition at whatever head is required(without the valve pressure drop), while also being able to reach your minimum flow condition and control it using a vfd.

But you'll have to examine it carefully to make sure that you can achieve all of the operating scenarios. As well, typically, on anything less than 100 hp stock pump, you'll be unlikely to save any money using a vfd in a system where the control valves and infrastructure are already in place.

 

[zdas04]

Stainer's PowerPoint is an excellent representation of the power side of the equation and he does a great job of describing the considerations, but on the use side there can be (and often are) overriding considerations. For example, on an oil-flooded screw compressor skid I specified 4 VFD's:
[ul]
[li]One driving the compressor motor itself. I find that if I can vary the rpm on a screw compressor (instead of opening the slide/turn valve, throttling the suction more or less, or throttling the discharge more or less) I can maintain a very constant pressure upstream of the skid which is what gas wells need, the other alternatives are not as effective at this[/li]
[li]VFD on the fan for the Fin-Fan Cooler. The one crucial parameter on a flooded screw compressor is the temperature out of the screw. If you get that right, then everything else is just sweet. In my temperature control scheme I slow and/or stop the fan on the cooler first[/li]
[li]VFD on the glycol pump. The oil is cooled in a plate and frame heat exchanger with glycol on the cooling side going to the fin-fan cooler. If the fan speed going to minimum still provides too low an oil temp then after I've stopped it I slow (and eventually stop) the glycol pump second[/li]
[li]VFD on oil pump. If the temperature is still too low, I can slow the oil pump to the minimum[/li]
[/ul]

This scheme is in no way power-efficient, but it is absolutely the most process-efficient screw compressor I've ever investigated. This control scheme operates with the slide valve fully shut much of the time, the suction controller fully open all the time, and the recycle fully shut all the time with a constant suction pressure with flow rate swinging nearly an order of magnitude. It also maintains oil temperature out of the screw within 1°F with flow swings and ambient temperature swings from -40°F to 105°F and gas flow rates from 200 MSCF/day to 1.8 MMSCF/day. No mechanical devices could come close to that performance. Nothing but a series of VFD driven motors. I made that decision with full understanding of the issues that Stainer points out in his PowerPoint.

When deciding on a VFD we often focus on (often imaginary) power savings, but most of us are not in the business of burning amps, we are facilitating a process, and need to take into account the process needs along with evaluating the equipment selection. A 10-15% improvement in the process can easily swamp the 2-3% increase in power or the 4-5% increase in maintenance, or the significant increase in capital.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[LittleInch]

Zdas04 - I'm with you on this and I like using VFDs to get the most out of a system and also like the ability to create controlled conditions which are not always justifiable by power loss alone (though I hate wasting energy). Given that this system is realtively small, (approx 35kW) the energy saving might not be huge, but without knowing the running time per year or cost / avaialbility of electricity where the OP is it is difficult to judge.

Ten penny, I can see your point, but I beleive that there is only one system curve here, the thing that makes the two curves meet is the pressure drop across the control valve which I normally show as a vertical line joining the two at whatever flow or head you want to achieve. I think this is a better way of thinking about it rather than having different system curves, but maybe that's just me.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[BigInch]

David, you seem to have posted against the wrong thread. Looks like some good info that needs to be moved over there.

Way to Go TENPENNY! Star. I agree that there is LITTLE to NO benefit to using a VFD and many more disadvantages. Personally I think there is NEGATIVE NET BENEFIT. Search this forum VFD to see the BIG LIST of Disadvantages and headaches.

A VFD on this system can be evaluated by knowing the power consumption at 190 m3/h and the power consumption at 270 m3/h. I figure it as 44 HP at 270 m3/h and 37 HP at 190 m3/h. The ABSOLUTE maximum savings we are looking at here is therefore 7 HP difference between the two flowrates.
That power savings will be obtained by using either a VFD or a Control Valve.

CV or VFD?
The head required using a contol valve is 39 meters. A VFD needs about 33 meters. Difference is 6 meters (20 ft) at 190 m3/hr, the maximum power saved is 1.86 cfs * 20 ft * 62.4 lbs/ft3 / 550 = 4.2 HP or 3 kW
1 hour operating with at a power cost of 0.15 /h = 0.47 $/hr maximum savings.

Maybe if you operate at 190 m3/hr for 2000 hours this year, you might make back the cost of a cheap VFD.
But then you probably have to trash the old motor and buy a new motor suitable for VFD use.
IF you're lucky and have no power supply problems induced by the VFD and can afford a new VFD rated motor, maybe you'll finally start to get a payback after 4 or 5 years or so.

I've done a quick VFD analysis for you. A bit rough, as I had to make up a representative efficiency plot and control valve effect, but you get the picture, right?










Independent events are seldomly independent.

 

[Artisi]

As a Paper stock pump, the flows and heads are probably quite variable throughout any given period as is the absorbed power input due to any varying stock density and air content within the stock - so chasing a few HP savings is probably spinning wheels. You may be better off looking into smoothing the demand and dumping excess flow into stock chest A elsewhere if possible.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[BigInch]

Right. Tanks (stock chest) allow the most efficient operation overall for any hydraulic system, because if enough tankage can be installed, the pump can always operate at BEP flow. Not even a perfect VFD can do better than that. Tanks do cost something though, so including one still might not produce the optimum total cost solution.

zdas, sorry, I just noticed the reference to stanier's thread.

Independent events are seldomly independent.

 

[LittleInch]

Big inch. Whilst I can see that a vfd may not be appropriate in all cases, I think you've got some figures wrong. At 190m3/hr, the power input to the pump is about 26kw/36hp. However the system only needs 20m head. Therefore energy lost across the cv is about 11kW. If you had a vfd, then some of this energy just thrown away would not be expended. Let's say 3kw for the vfd drive inefficiency and say 2 kW if the pump efficiency is a bit lower still makes 6 kW saving. Tho op says 190 is the normal flowrate. If it runs all day, that's 21$ /day. Only the OP knows how much his pump runs per day /year.

He also says the pump is old and they want to replace it. At the very least Eric, make sure the motor you buy is capable of vfd drive. It doesn't cost much more, but will allow a retrofit later on.

For sure, vfd is not a pain free option or suitable for many applications and this ones probably on the limit, but can still make sense.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[BigInch]

Could be. It was late and I didn't check it. But anyway, it does show it's not likely to save much unless he runs a great deal of time at the 190 rate, in which case he probably should take another look at reconfiguration. Equally, a small, or slightly larger tank, might be the best solution. Only the %Time run at reduced rate vs Total Time will answer that question.

Independent events are seldomly independent.

 

[ericdayo26]

Ok you guys, thank you for giving your time on this.
Little inch, I agree with you with regards to your explanation of the system curve. So this means that even though i calculated the Total Dynamic Head to be 32M, I only need 20 meters of head with a VSD to get 190m3/hr?

Additional info to this problem is the Actual power consumption at 41% open Control Valve to be 29.7kW. I've counter checked this data using this formula:
(52.78LPS * 38M)/(102*0.74*0.93) = 28.2kW
where: 0.74 is pump's eff, and 0.93 is motor eff, 190m3/hr = 52.78LPS, 102 is just a conversion factor..

I got a slightly lower Power consumption value of 28.2kW versus the actual load which is 29.7kW so I assumed that perhaps the consistency of the stock, 3.5%, has something to do with it. Since stock is still considered slurry i assumed its not to far from water values.

Using the same formula for the 100% open control valve scenario;
(75LPS * 32M)/(102*0.74*0.93) = 33.74kW
where: 270m3/hr = 75LPS

To defend my option to install a smaller pump that will run for the normal operations, I am guessing what head to use, will it be 20m?(that is in reference with the system curve and drawing a vertical line down that intersects at 190m3/hr)
Or should i use my computed Total Dynamic head, which is about 32m?

As we can see, substituting a higher value of head on the formula above will get a higher power rate.

Also, as to BigInch's inquiry, we (in the plant) agreed to use 297days/year for the 41% opening and 33days/year for the 100% opening, for a total of 330 days per year of operation. Thanks BigInch for your help, it made me rethink the option for the VFD, and also your power consumption computation, though i might need to tweek it a bit now.

Again to LittleInch, thanks for the link on Stainer's presentation, but i can't seem to download it. Is is possible for you to repost it? or send it to my email: ericdayo26@gmail.com. I would really appreciate it.

I'm really confused with the system curve topic because from what i know, when designing pump and piping systems, you have to compute for the Total Head first (i.e, Friction head + Elevation head + Pressure head) in reference to the desired flow. I got aprox. 32m for this problem, but the system curve is saying i only need 20m. So, which is it?

Thanks everyone, in advance
Kind regards,
Eric

 

[BigInch]

That's the problem I had. I couldn't be sure at what flowrate you computed that system curve head at, or if that was including some %Closed of the valve. You should compute the head loss at 190 m3/h and another head loss at 270 m3/h, and some in-between points might be good too.. Then connect all the dots to get the whole curve. In my approximation, I assumed your 32 meters head loss was at 270 m3/hr, which gave about 23 meters of head loss head required at 190 m3/h. Not including any valve effect.

The contribution to the system curve head loss of a valve should actually be computed based on dH = K * dP^2, where dP is the head loss across the valve, although it can be kind of fudged like I did there. You should also use a curve (depending on type of valve) to get the K value for each percent open too.

Independent events are seldomly independent.

 

[LittleInch]

Eric,

Glad we helped (sort of). 20m @ 190m3 comes from your system curve. I would always add a bit, but you really need to check the pressure immeadiatly downstream of your control valve and then calcualte head from that to be sure. From your total figure, I get the friction loss + elevation difference (chest B liquid level minus the chest A liquid level, but not sure what you mean by "Pressure head". Give us the numbers you used and then we might be able to see where your error is, but taking a reading u/s and d/s the control valve will give you hard data to compare it against.

If you get a smaller fixed speed unit ( 20m head) then 190 m3/hr is all that it will pump, you won't get 270 m3/hr out of it.

If you go the VSD route the the VSD pump data sheet needs to have your two main duty points, which as far as I can see are 190m3/hr @ 21m and 270 m3/hr @ 32m for max flow. The pump vendor will then provide a pump that can do both with different speeds. Make sure you note that the 190 case is the one to be the most efficient. If you're going to run this automatically, the VSD controller also needs to be able to accept an input from a level transmitter to control the speed automatically. If it's manual then a speed control button is all you need to increase and decrease flow.

Re posted the ppt link, but the earlier one opened ok for me.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Artisi]

For me, in a paper mill application such as this one under discussion, the expense of fitting and maintaining a VFD for the possible chance that if "might" be needed 33 days a years is ridiculous. All it is doing is adding a complication into a very simple installation to save how much power over the years operation - I'll leave that to you to calculate.

One small point for your own clarification when pumping paper stock, once a certain velocity thru the pipe is reached the friction loss drops to below that of water for the same velocity /flowrate. Check on the ITT Goulds pump website - somewhere in there is detail on paper stock / friction losses with discusses this point.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[LittleInch]

Eric,

As noted before, when doing any electrical saving calualtion, remember that the VFD unit has some significant heat losses which need to be added in. These are 5-10% of the power into the motor - you'll get this data from the VFD supplier.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[ericdayo26]

LittleInch, I think what your describing is the efficiency of the VFD? If that is it, then yes, actually I do multiply an additional 90% eff for the VFD, when calculating for the Power. Say, Pump eff * Motor eff * VFD eff. BTW, I got to download the file this time, thanks again. Also, about the Total Head equation I stated, I didn't actually add any pressure head because the chest are not pressurised, they are atmospheric. There isn't really any pressure gauge installed before and after the Control Valve, so when I calculated the Total head of 32m and it matched the pump curve at 270m3/hr, I assumed that my L/D value for the Control valve at 100% open was correct. Yes, a smaller pump will indeed be a fixed speed motor of about 40hp. When there is a need for the flow to increased, I suggest a parallel pump be intalled. But after doing a Life Cycle cost analysis on this option, i found that the payback period was too long and there isn't much savings per year.

Artisi, I think I understand what your saying, so I would like to rephrase what I said previously about the current system. We kinda just assumed the 41% open scenario because in actual operation, the valve openning changes from 30 to 50%. During our load survey, the opening was about 40% percent with a flow of 190m3/hr, it got a load of 29.7kW. For 30% opening, the flow is 170m3/hr, I did not get to have the load checked with this scenario. But as you can see, the flow still varies during normal flow depending on the level of the chest A and B, that fluctuates also. The current pump also has a purallel pump intalled which they also use during those very desperate situations. In that scenario when the two pumps are running, they regulate the Control Valve opening to 60%, that gives a flow of 320m3/hr.

BigInch, with regards to my calculation of Total Head, I only calculated based on a flow of 190m3/hr, since they don't actual target a flow of 270m3/hr. It just happens that, that is their maximum flowrate at 100% opening, whenever they need hurry transferring stock from Chest A to B. I use the equivalent length method, which for me is easier to do. Although, as far as I know, this method can be inaccurate with varying valve opening. If you share with me a tutorial of a different method of computing for Head, I will appreciate that as well.

Kind regards
Eric

 

[BigInch]

I think your pump has relatively the same efficiency at both flowrates, or not very much different, and they are both relatively close to max efficiency. It would help if you showed that curve. There is probably very little wasted efficiency with the BEP at a flowrate somewhere around the middle of (190+270)/2, which is probably what you've got now. I wouldn't change the pump characteristics at all, unless you were to have maybe a much more skewed time ratio between the two flows, 75-25, or 80-20. Then it might start to make some small amount of difference. Equal %time for each flow would best be served by a BEP at mid way.

The VFD will probably have 90 up to 95% efficiency at that power loading. It doesn't start to drop until you have a 50% or lower power load. Same wtih the motor, but a little faster drop, something like 75% efficiency at 50% load, linear to zero from there.

Calculate the head loss of the piping system using D'arcy equation. Calculate the head loss of the valve by obtaining it's coefficient, K from typical values, or from the manufacturer, preferably for all %Open Positions, then the head loss for the valve in the system depending on the percent open position and differential pressure on the valve at any given time. That depends on the flowrate. Add the two together to get the head loss of system including the valve. It is an iterative solution, typically for flowrate. Assume a flowrate, calculate pump differential head, system loss, valve loss and see if it matches actual pressure drop. If not, increase, or decrease the assumed flowrate as required.

Independent events are seldomly independent.

 

[ericdayo26]

BigInch, yes the flow is very much within the range of the pump curve's BEP. I'm going to try and post the pump's curve later.
I have another question though..
why do they say that you will save more power if you operate the centrifugal pump near BEP?
Say a centrifugal pump with a control valve runs near BEP at 100% open. When the valve is throttled, say to a maximum of 30%, the duty point moves up to left of the curve, but it then relatively gets a decrease in power needed if you plot a vertical line down to the power curve. So therefore, as you throttle a valve, the load on the motor decreases also right?

 

[BigInch]

Post the efficiency curve if you can. The curve demonstrates the pump's efficiency at all flowrates.
You will not save more power. BEP means Best Efficiency Point. If you operate there, you will not save any power, but you will be using the power you do use there most efficiently. In other words, the (Power_consumed at BEP flowrate) divided by (mass_of_fluid_pumped) is the highest of any point within the pump's range. Simply put, you are using the energy most efficiently at BEP. IN GENERAL, and not considering the system curve, looking only at the pump curve,

at lower flowrates, even though the pump will discharge at a higher head, hydraulic power demand will be less, due to the lower mass flows.
At higher flowrates, the high mass flow increases the hydraulic power demand, even though the discharge head is lower then when at BEP.
Since total power demand (brake horsepower) is calculated by dividing hydraulic power by efficiency those general rules can vary a little when considering efficiency at each flowrate, but it doesn't vary by much, unless your efficiency curve is very unusual.
Have a look at how I calculated Brake Horsepower on that spreadsheet. Brake Horsepower = Hydraulic horsepower / eff

From the looks of your pump head curve, I just think the efficiency curve will show relatively good efficiencies across the entire flowrate range you are talking about. There may be a difference in efficiencies of 2-4% from your two target flowrates and BEP flowrate, which is not bad at all. That indicates a relatively good pump selection was made to optimize efficiency completely across your entire desired range of flowrates. Unless your efficiency curve shows a "spike" at BEP, and drops rapidly on either side of BEP, I would think the pump type is good. In fact, very good.

Independent events are seldomly independent.