CSV or VFD 4

[gare]

There are times where a VFD seems appropriate but I then encountered a Cycle Stop Valve on a system not functioning correctly. Has anyone had experience with using these to maintain constant pressure with varying demands on an irrigation system? I thought it was just to stop hammering, but the manufacturer states it does much more, and this seems true, I am not sure of their reliability though.

 

[itsmoked]

CSVs verse VFDs have been debated at heated length here.

CSVs waste energy but are robust.

They waste energy but can help a pump stay in the middle of its curve.

They waste energy but are less expensive to purchase and install.

They waste energy but they don't have potential electrical noise issues, nor life time issues.


There is one manufacturer who is rabid about pushing them as the ONLY VALID solution. That showed a dramatic ability to ignore, or paint over facts, to draw that conclusion exclusively.


VFDs allow you to do things you can't with just a CSV. But if those things have little value in your particular application then they might point away from VFDs.

CSV essentially do their job by raising back pressure to the pump. This costs more energy given the same amount of flow with less restriction that you'd get with a VFD.

Essentially if the only thing you want to do is keep a constant delivery for a short time the CSV is probably a good way to go. If you instead pump long periods and the CSV has a significant head penalty then you should look at a VFD.

Many sites have energy calculators that let you mess with various scenarios, using a VFD to see if one might make any kind of economic sense over a CSV.



Keith Cress
kcress -

http://www.flaminsystems.com

 

[BigInch]

Keith,
PLEASE. Any valve wastes energy, just as any VFD wastes energy (remember at least 4 or 5 percent just due to VFD inefficiency, at lower flowrates MUCH more due to even lower motor and VFD inefficiencies at partial loads). Your statement shows about as much unwarrented favoritism for VFDs as the Cycle Stop Valve guy had for valves. Which option wastes LESS energy is the question that should be answered.

Gare,
As I said above, the CRITICAL question to be answered when designing all systems is which option will waste LESS energy when considering all the flow and head requirements of the system and the manner in which the system will be operated throught its entire lifetime. For a system that has little static head variation, but will be operated only at one flowrate, or within a narrow flowrate and head margin, it will pretty much ALWAYS be more efficient WITHOUT a VFD sapping off energy. In fact, you might not even need a valve. Same goes for a system that has high static head requirements at all flowrates.

For a system that has head requirements that vary with the square of speed and wide ranges of flowrates, you could consider a vfd, provided that there IS a wide range of required flowrates. For a system that has high static head requirements at most flowrates, VFDs are not even an option.

I've spent most of the past year analyzing many cases of VFD vs VALVE or NO VALVE and can say with a very high degree of authority, VFDs are not the panacia many think they are. In fact, as a rule of thumb in many systems, work well only between 50% to 80% of what would be the BEP flowrate and speed of a single speed pump installation, given that the variable head they produce at reduced rpm still allows those partial flowrates in the system. That "work well" criteria includes a necessity to operate in that 50% to 80% range for a significant portion of the systems lifetime. Many times it is a very close call, in which case maintenance concerns for VFD related symptoms of bearing and power quality problems often tip the balance to favor not using a vfd. The whole point being here is that, if you fall outside some very small margins of my rules of thumb, a general question about VFD vs Valve cannot be answered without a very carefull and sometimes a very detailed analysis of your specific system and how it will be operated.

A perfect example of what kind of system it is and how it will be operated is to consider an irrigation system, which often have high static head requirements if the pump is lifting from a well, so VFD may not even be an option. Secondly, and most importantly, in an irrigation system, WHY do you even need variable flow? Why not just water the plants using one flowrate and turn off the pump when they get enough, or switch to a different field? That's how I handle my grass. Do you need to water one lawn at 2 gpm and the other at 5 gpm? You could water lawn 1 at 1 gpm for 2 hours and the other at 1 gpm for 5 hours. Other options might be to build a tank, pump at one flowrate until the tank is full, then turn the pump off and trickle flow from the tank. Why use a VFD to get trickle flow? IMO, irrigation is perfectly suited for single speed operation, no VFD no VAVLE.

IMO, in irrigation systems, cycle stop valves are best used for hammer prevention and adjusting head and flow, but to a lesser estent. Stopping hammers will also stop-short a lot of maintenance problems with broken pipe joints and initiating pinhole leaks. Their biggest advantage in an irrigation system is they can do that without using a VFD, which I believe are basically not necessary in such a system. If you have an industrial process where you have to feed glue, or some other fluid, from the same spout at many flowrates go VFD. If you have a water demand that varies alot, but you can't turn on or off 1 to 4 pumps, get a VFD, but only if your required head also varies accordingly. If head remains relatively the same, use a valve, or don't use a valve if you don't need it. VFD, use them if you need them, if you don't... don't.


**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[gare]

I apologize for asking a question already heavily debated and researched, I need to search better before asking. However, that both of you responded is appreciated. I asked the question in here because researching on the internet yields (for me) a lot of great information on both sides, disabling my ability to make a choice with confidence. I suppose this is what people must put up with when they do not just go out and higher an engineer the first time around, or the second for that matter. I am an installer, and really just the electrical portion.

The entire system has such varying zones that flow rate is dramatically different. I have not looked at each zone and done specific calculations, but there are small drip systems and multiple large rotary heads, as well as the pool-fill, two fountain fills, and four moat fills, and who knows how many hose bibs. It probably should have been designed with more than one pump. The pump sits in a pit/below grade vault that is filled with well water, property drain water, and/or city water. The water in the pit is dedicated to irrigation.

On large zones the motor cycled every 45 seconds and the starter at times bounced up to eight noticeable times before seating. I am assuming (I hate that word) that voltage drop on start-ups, especially as the motor gets hotter, was causing this.

The system ran smooth after start-up, but it only took about 20 seconds for the pump to reach its pressure shut off point, and that on large demand zones. At first we though of just adding more tank capacity, but the room also houses a large generator, 4 pumps (none of which have to do with this, they are moats and fountains), transfer switch, etc., so space is cramped. If I had known what the CSV should be doing, we would have focused on that. We could have overlapped stopped and start times on zones to keep the pump running throughout the irrigation cycle. We even thought of just putting a pressure relief in the sump (I am not sure what to call what the submersible sits in) and letting excess water dump back in during irrigation.

We are putting in the VFD because we bought it (a rather embarrassing statement). We were going to pull it out if I got hammered in this forum, but I do feel it is an OK choice, if not the best. I would prefer not putting it in, as it introduces another level of technical stuff that most of us in residential irrigation may rarely, if ever, see.

These happen to be rather large estates. They should pay for up-front design and then make sure competent people do any adds or mods to the system. The sad case is that many people look at the bottom line and merely ask, why can so-and-so do it for X and you are 2X? We will hire X. Worse, after the initial design and installation, the gardener adds zones, repairs heads with something they found at OSH and can be installed in any 1/2" or 3/4" FA, etc.

Today we pull the CSV out and put the VFD online. We had never dealt with a CSV until this job, perhaps it has failed, or needs adjusting, I do not know.

Another issue is the ease of maintenance, as referenced somewhere above. I am concerned over what the VFD may do to the motor over time as well as the life expectancy of the VFD itself.

The landscaping is most likely in the 7 figure arena We would like to see longevity in the irrigation system, but they were tripping out the motor every month and getting billed $400 per call out by the installer, plus replacing a motor now and then, I am not sure of frequency. Maybe twice in the past 5 years?

Bottom-line, keeping the landscape green. BigInch's statement <b>"That "work well" criteria includes a necessity to operate in that 50% to 80% range for a significant portion of the systems lifetime. Many times it is a very close call, in which case maintenance concerns for VFD related symptoms of bearing and power quality problems often tip the balance to favor not using a vfd. The whole point being here is that, if you fall outside some very small margins of my rules of thumb, a general question about VFD vs Valve cannot be answered without a very careful and sometimes a very detailed analysis of your specific system and how it will be operated"</b> is a concern now.

I have to say, I hope it works well. Otherwise, we eat it and do something else. This <b>"without a very careful and sometimes a very detailed analysis of your specific system and how it will be operated"</b> would probably cost them as much as the VFD. Such a drag.

 

[BigInch]

Sounds like a crap original design that a VFD has no chance of fixing. Without solving the 45 second problem, your vfd will just keep running constantly at very low efficiency using far too much electricity.

Its taking the pipe 45 seconds to pressure up initially and with some pipe expansion in a big system, and some pressure reflections after the pipe pressurizes in some areas, pipe overexpands and then sends some of the pressure back to the pump pressure sensor tripping the pump off until the pressure is relatively stabilized for the current demand flowrate. I feel the cycle stop valve can solve that problem, if it is adjusted correctly. Get somebody out there who understands how that valve should work. I would suspect that the pressure sensor on and off trips should be adjusted for a wider margin.

What I would suggest is that you definitely do not install a VFD ... yet. Somebody should have a look at balancing the zones using the same flowrates to all zones, but with timers to vary the time on/off to each zone by opening and closing (in turn) valve actuators at the inlet to each zone. The fountain draws and miscellaneous hoses etc. should not be enough of a constant water load to affect anything near the design capacity point. It might also be beneficial to install a smaller topping-off jockey pump for those loads, if per chance they turn out to be a significant demand draw that lowers pressure too much.

From what you've said so far, the VFD won't solve any problem you've mentioned, except perhaps slowing down the effect of pump startup, something the cycle stop valve should be perfectly able to handle. In lieu of a VFD, if you do decide to go that general route, may be a soft starter for that pump.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[gare]

Thank you BigInch. There are three entities now involved, not including the homeowner or their manager. Landscaper, property maintenance, us(me).

I agree that your approach is the optimum. However (drum roll) due to the sequence of events we must do things backward, make it work "OK" right now and look at long term proper solutions after.

I have sent the image of the installation to the CSV manufacturer to show the location of the existing CSV and described the problem. I do think the CSV SHOULD work, but it is apparent that it has not functioned properly, and this for some extended time. If it is just due to improper adjustment, shame on those before me and ME.

Short-term: The tripping/cycling needs to be stopped. We want the system at a constant pressure. CSV or VFD? If this VFD just does this for now, we are winners in the hearts of the owner and manager. What a sad approach to proper functioning.

Long-term: Multiple smaller pumps?, balanced zones?, ???. Whatever the owner will allow for cost-wise. All of the input from here leads to a more comprehensive report/proposal to the owner. Sadly, it is rarely do the absolute right thing the first time, whether due to ignorance, money, immediate need, or whatever.

I will post the final resolution.

 

[BigInch]

The easiest way to get constant pressure is with a single speed pump. As long as you keep the flow within the pump's discharge capacity, you can have a constant pressure. That's exactly what you are attempting to do with the pressure sensor. Or at least keep it within a narrow range of pressures. That should work, provided the start-stop pressure range can be "tuned" properly to the system response characteristics, ie not too much pressure run down for reasonable demands.

A VFD will not make holding pressure at flows greater than the pump capacity any easier. In fact a VFD installed on the same pump will run at slightly less pressure at full speed than a constant speed pump for any given flow. To get the same head as a constant speed pump a pump used with a VFD should be selected with a curve about 5% higher than the constant speed one.

The only real difference in the two alternatives, is that the vfd changes speed and flow to try to match your set pressure, whereas a constant speed pump tries to hold the same flow while it turns on/off to try to match the set pressure. If it can hold the pressure you want without a valve, so much the better. If you have too much pressure, a valve is required to reduce the pressure. That also reduces the flow somewhat too, depending on the valve's characteristics.

A VFD may help the pressure pulsing, but only if it slows the pump's speed changes down. If its only happening on startup and the cycle stop valve won't work, a soft starter would be the way to go.

The ace card for VFD use is a real necessity to have variable demands on a limited number of pumps, 1 or 2, no high heads at low flows and no alternatives towads making the flows more constant in nature. In that situation they can work well.

I think the first solution in this case should be to try adjusting the valve correctly, especially since the valve is already there, and trying to better balance any variable demand loads if that is possible, but if you've no choice, you've no choice.

Good luck.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[ScottyUK]

Hi BigInch,

I'm not going to get in to the endless valves v's VFDs argument but "In fact a VFD installed on the same pump will run at slightly less pressure at full speed than a constant speed pump for any given flow." is a very strange statement. The motor just turns the pump. If the motor is turning at the same speed then whether it is being supplied by a VFD or direct from mains is totally irelelevant because the only variable - speed - is identical.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[itsmoked]

In fact a VFD installed on the same pump will run at slightly less pressure at full speed than a constant speed pump for any given flow. To get the same head as a constant speed pump a pump used with a VFD should be selected with a curve about 5% higher than the constant speed one.

This is not true. You would be implying that the turning of the motor's shaft would be somehow altered by virtue of the motor being fed by a VFD. That is not the case. Your installation might draw 3% more power(at full speed) but the motor is not going to spin 'differently'.


This installation with:

The entire system has such varying zones that flow rate is dramatically different. I have not looked at each zone and done specific calculations, but there are small drip systems and multiple large rotary heads, as well as the pool-fill, two fountain fills, and four moat fills, and who knows how many hose bibs. It probably should have been designed with more than one pump. The pump sits in a pit/below grade vault that is filled with well water, property drain water, and/or city water. The water in the pit is dedicated to irrigation.

And a single pump trying to cover these large variations in flow does sound to me like a VFD candidate.

Add to this the motor start up issues mentioned and the ability to smoothly bring up pressure into a use that has a large variety of connections also points in that direction.(for me)

I could easily see this application living in the 50% to 80% realm most of the time.

Trying to synchronize all the various loads of this system into some well averaged flow is probably an unreasonable expectation. You are always going to have something uncontrollable happening. You can balance the irrigation all you want and someone is going to show up and decide to fill a moat or start washing a car, or running 3 or 4 hose bibs into temporary watering tools.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[BigInch]

Oh ya, put-um up. That's not the typical design procedure. In pipe work its probably backwards from the electrical procedure. In pipe work you usually know the required head and flow and must chose from pumps from a catalogue with curves defined by manufacturer's test speeds, not from curves defined at VFD speeds. Maybe in electrical you start from curves with VFD speeds, ... or something ???.

Therefore,

If you need a pump for a system that requires 64.6 feet of head at a flowrate of 260 gpm, as in this example by PIPE-FLO,

http://kb.eng-software.com/questions/117/Modeling+a+Pump+with+a+VFD+to+Maintain+a+Fixed+Pressure

and you will use a VFD, you can see that you must select a pump (in this case with a rated fixed speed of 1780 rpm) with the head curve way up about 87 feet, much higher than the 64.6 ft, such that the fixed speed head can be scaled down to the vfd rpm of 1630 and achive the same head of 64.6 feet. So when selecting the pump BASED ON A REQUIRED FLOWRATE AND A REQUIRED HEAD for use with a vfd fitted pump, you must chose from pumps with some head curve defined by the manufacturer's test speed , say (1780 rpm) and the pump you chose on that basis must have a 10 inch impeller curve that is quite a bit higher head at that same flowrate, so it can be scaled down to the VFD's rpm. With no VFD you could select that pump with a 9.25 inch impeller and get the same head. See what I mean? A very common rookie mistake is to decide to use VFD and then select that pump with a 9.25 inch impeller just 'cause its close to the operating point; forgetting that the 9.25 inch impeller curve is developed at the test speed of 1780 rpm. If using a fixed speed pump, you could select the 9.25 in impeller.



**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[gare]

Hi Everyone. And Thank you itsmoked, BigInch and ScottyUK. Hey itsmoked, you helped me a ton on the 12kv ungrounded primary as well, back in March or April. I had signed up as garetoo and have no idea how to get into that now, I no longer have the email and I forgot the password.

Well, being an electrician, I rarely design anything , I install something already designed. I also whine about some designs I install.

So, when I was asked to look at why this motor was tripping the sensor out (senses phase, overcurrent, undercurrent, and voltage - opens the control circuit and illuminates a red LED that no one can see unless they bypass the latching of the cabinet and look at it - no one had done that) and I heard the starter hammering, and saw the pressure shooting up and down I was thinking, "how do I approach this electrically?" and I thought VFD without any thought of pump curves, etc. I admit I have a narrow perspective at times. I needed the work, so I said the problem could be fixed with the VFD and they approved the job.

I imagine when the job was first done that it was engineered properly, with the CSV (now we know it was manufactured in 2001). I also know things change and people do not call the engineer when they want a bubbler system on the roses. The gardener, like me with the VFD, says, "I can do that for only $8.95" and he does.

I think controllers should get huge labels "DO NOT ADD ZONES REQUIRING LESS THAN XXX or MORE THAN XXX GPM, MAY DAMAGE ENTIRE SYSTEM". That may scare most hackers away.

Before I went to the site this morning, I shot an email off to the manufacturer of the CSV and he called me back, that was very nice. I learned more about the valve in question. I decided we would try messing with the valve to see if the system would work, but we wanted to look at all of it first, so we pulled it out. We found that one piece had come loose and there were bits of diaphragm and rust in it (the galvanized nipples had a lot of buildup in them).

We had no parts to make any repairs, so we kept it out, made up a nipple to take its place, put the VFD on line and it works sorta great. There is some fluctuation, like 2 to 3 pounds rapidly, when the pressure hits the lowest point and the pump starts picking up speed. I am guessing this may be a partial product of the expansion and contraction BigInch speaks of, but all-in-all it maintains pressure within 5 PSI regardless of the zone we put on. We did turn on two large zones and the pump could not keep up.

I worry about the motor, and the manager wants assurance the motor is not going to burn up. I could not commit to that though I assured her the hammering of the starter and constant stop-starts was harder on it than the VFD, I hope this is true. I also told here there would be no more tripping as before. I did see the VFD requires Class T fast-acting fuses, right now the original TR slow blows are on it.

From a maintenance perspective, I think I would rather just replace the few parts in a CSV and deal with a NEMA starter than troubleshoot a VFD.

What is odd is that this CSV has probably been out of commission for some time and no one looked at it, they just kept resetting and adjusting the sensitivity of the monitor to inhibit tripping.

Other than needing to get some adapters that allow for Class T fast acting fuses into the fused disconnect now in place, I believe we will leave the VFD running the motor, but in the future I want to try the CSV. We are going to rebuild this one and wait for the opportunity.

Thank you

 

[BigInch]

Even a vfd cannot make the pump supply a demand that exceeds what the pump can supply. Effectively, it is a pump runout condition where excessive power can be consumned and the equipment is possibly at risk of burning up unless the motor, the vfd and the power supply cables are all rated for that full power load when it is at runout conditions.



**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[BigInch]

I would like some EE comments about this statement from ASPE Handbook, "Pumps and Pump Systems" by Chan and Meckler
which apparently verifies by example the method I mentioned above.

Maximum variable speed in rpm is about 90% of synchronous speed. Since basic pump curves are plotted in terms of synchronous speeds, the true selection point must be converted to a synchronous speed selection point. This is accomplished as follows:

For a system requirement of 300 gpm at 132 ft,

Synchronous speed selection: gpm = 300/0.9 = 334 gpm
Ft Head = 132/0.9^2 = 132/0.81 = 163 ft head.

If I were to select a constant speed pump, I would simply select a pump with a BEP of 300 gpm and 132 ft head, not
a pump of 334 and 163 ft. which in certain borderline ASME flange rated pressures might make a considerable difference in the pump pressure spec and the resulting cost, perhaps requiring a ASME 900# rating for a VFD pump, rather than a 600# rating in what could be a worse case scenario, where it might be plausible to disconnect the VFD (perhaps in an emergency parts availability situation) and reconnect to run at full synchronous speed, even if only temporary. (admittedly extreme) Or perhaps in another situation, where to get that extra head and flow curve, an entirely different pump might have to be selected from the catalogue.

(I said 10% head in my comment above, but that should have been a 10% higher flow, even more of a head increase.)

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[BigInch]

And if that statement above is true, retrofitting a VFD on an existing pump would preclude ever pumping at the previous maximum flow, given the system curve remained the same. So, with a vfd on a retrofit, a lower pressure output is assured, more than likely reducing the previous maximum flow too. Like I say, variable flows for a high percentage of op time between 50 to 85% of required BEP is pretty much a necessity to justify a vfd.

Some of the work I've done last year involved studying normally distributed flowrates with different standard deviations about a mean flowrate, that mean also used as the basis for selection of the optimum pipeline diameter. What I discovered, especially when pumping a lower value products is the extreme importance of maintaining a flowrate as near as possible to the given design flowrate in order for the investment in pipe and pump to payback time. I honestly thought, or didn't think, that the lower flowrates, those in the range of 50 to 80% where VFDs pay back best, wouldn't hurt the payback time of the project as much as they did. Believe me, you definitely want to keep flowrates as close as possible to the pipeline optimum design flowrate. Flowrates lower than 75%, I'm sure you can imagine, drastically affect payback time of a typical project.


**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[itsmoked]

Big; I totally agree with your assessments in most the cases you are presenting. But. I think these small non-steady state or even quasi-steady state situations like the OPs really wreck a lot of the logical designed system expectations.

The system is there! It is installed. It sort of works. It would work better if it was either less chaotic or designed correctly or differently in the first place. Now it's add something to get it to work a little better. Energy considerations are taking the back seat.

gare; If you think the CSV is busted and can be fixed, you should give it a shot.

Big; As for your quote question.. Async speed is, of course, sub synchronous speed. 3,600rpm > 3450rpm(typically) which represents about 4% below sync - not 10%.

If that statement is supposed to be referring to a VFD,(no mention), it makes no sense to me. A VFD motor does not run one bit slower than a DOL motor. The statement is incorrect. A VFD runs its motor at the same 60Hz that DOL provides.

On the contrary, if the motor has some service factor you can actually run the motor up to the synchronous speed and the pump to its synchronous rating.


Keith Cress
kcress -

http://www.flaminsystems.com

 

[BigInch]

That would be with no load and consequently no slip. It must refer to max rpm when loaded, which I inferred from this bit,

Almost all commercially used variable speed drives operate in a "slip" manner. Design slip is at 10% for fluid drives and SCR interrupted voltage units.

This book is getting on in years, maybe there are photon and warp drives today.

Keith,
Can we go to email?; as with this topic we might drift even farther away from the thread.

Gare,
Yes, even Keith agreed... try the csv first.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[BigInch]

Probably should have thought about this first, but here's one last thought.

If its a CSV with a strainer, be sure that its clean.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[ScottyUK]

BigInch,

That quote contains a number of statements which are either misleading or just plain wrong.

Unless the motor is a synchronous motor then synchronous speed is irrelevant to the DOL / VFD discussion. I doubt you'll find a synchronous machine on any pump requiring less than a megawatt, and not on many below about 3MW. Induction motors reign in the sub-megawatt range, and they always run at a slightly sub-synchronous speed, usually achieving about 97% of sync speed.

Some VFDs can compensate for the slip inherent in any induction motor and raise the drive output frequency slightly above the nominal 50 or 60Hz to bring the actual speed to (say) 3000rpm or 3600rpm, so as far as the pump is concerned it could be being driven by a synchronous machine and not an induction machine. The pump output with a slip-compensated VFD will be slightly higher than with the same pump and motor running DOL.

After all that... I agree that a VFD doesn't suit every application and every pump design.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[BigInch]

Thanks Scotty.

Lately I've been putting together what is now a rather large spreadsheet which attempts to consolidate the entire pipeline system optimization process, including whether it is more advantageous to use a VFD or not. Due to the large number of variables that can have a significant affect on such problems, my ultimate goal is to see if I can narrow down the most plausible solution set by solving using genetic algorithms.

The classic procedure for optimizing a pipeline design is to first balance a cost of pipeline diameter versus cost of power consumed ... for a given flowrate, product and pipeline length and its typically left at that. To give you an idea of what I am attempting, it is to solve without even defining those fundamentals, other than a given pipeline length. Although I am only considering one product at this time. I soon realized that in a "corporate scenario" pipeline system optimization, its not only minimizing power consumption; maximizing profit from products delivered is the ultimate goal (which diminishes the valve-vfd question considerably for high value products). To further complicate this, we might not even have one known flowrate, but rather some kind of normal distribution of demands at city A, B and C, etc., compounded by the fact that those demands can vary by day, month and/or year and can change considerably over the lifetime of such a pipeline system. The first question, what should be the (initial) design flowrate, becomes of vital importance and indeed does not have a trivial answer. I say initial flowrate as the construction of such a system can be staged over many years by adding pump stations and pipe loops as required to increase capacity to a 50 year ultimate design flowrate. I reverted to a Monte Carlo solution where, even for simple problems, I discovered the extreme effect that various demand distributions have on the ultimate profit that can be made from each "optimized" design flowrate and diameter. So much so, that even designing for a normally distributed flowrate is questinable. Which is presumedly the precise reason that liquid product distribution pipelines generally wind up running around pretty much a very narrow band near one flowrate, probably also why the classic design procedure of choosing a design flowrate works out OK, at least until the pipeline needs to be looped. As I say, it was surprizing how close to that flowrate that operations need to be maintained to allow the pipeline to pay back its invested capital in a reasonable time period. Consequently, unless a pump is being started or stopped, you don't find many running at lows of even 75% capacity. Another item with product pipelines is that they are typically operated at constant head, allowing natural flowrates to be assumed for the combination of all product batches travelling down the line, rather than by trying to put them on flow or pressure control with vfds. Over any given time of pumping multiple product batches, "more product has been found to be delivered" using that method than any other. The key word being "found", indicates to me that the complex pipeline optimization problem is difficult to solve analytically, but it keeps the neurons firing. I think GA could have potential here, but perhaps only if the lifetime design is broked down into possible staged developments and considered separately, otherwise the answer may become unintelligable. What I have discovered is that in such a problem, the vfd vs valve question isn't of much interest. Its far more important to always flow at the optimum flowrate, or rip up the pipe.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[gare]

ALL - We should have fixed the CSV first, but I was ignorant of what it was. I thought it only stopped the hammering, I did not research it up front as I should have because I only looked at it as to what I could do electrically - VFD

ITS - the CSV is broken, we are going to fix it and we are going to install it in the future.

BI - It does have a strainer, the manufacturer explained to us that this CSV probably should not be where it is and they recommended a different one. But we are leaving the CSV out of this installation for now. We had to get the system up, the manager is not going to pay us to work on optimizing anything, we already ate 12 hours of labor, but it was worth the learning curve.

ALL - The bad thing on many residential systems, seldom does anyone have any original design criteria, the systems have been modified to the nth degree with ZERO documentation, the homeowners are sick of spending a ton of money, most technicians do not know everything they (me,we) should know.

Algorithms are great for original design and if this were held to (do not know how to show the quote):

QUOTE BI -
The first question, what should be the (initial) design flowrate, becomes of vital importance and indeed does not have a trivial answer. I say initial flowrate as the construction of such a system can be staged over many years by adding pump stations and pipe loops as required to increase capacity to a 50 year ultimate design flowrate.
END QUOTE

I can see this being done on a refinery where they have a gazillion dollars and plan expansion. This does not negate the fact it should ALWAYS be done, but up-front cots are always a variable in residential. She wants the $200,000 mantelpiece from a french castle, so forget about potential changes to irrigation and hire the guy who will do it for X versus XX.

The existing pump on the VFD maintains flow on any zone. The irrigation service group is MAYBE going to look at doubling up some of the smaller zones run times (not combining via pipe) only to get the watering done earlier. No one is planning on pulling the pump up to see what make/module, getting pump curves, doing any calcs. Do I think they should, yes. Someone has to pay for that though. Perhaps when the motor fries.

I too think this thread is complete and again I thank everyone. It always blows me away when I see how something should be done up front with respect to the engineering of a system. Hence eng-tips.com! Where people like me can ask a group of engineers a question.