To VFD or not to VFD: Centrifugal pump flow control
ฝัง
- เผยแพร่เมื่อ 22 พ.ค. 2024
- This video looks at where a Variable Frequency Drive (VFD) will provide good flow control for a centrifugal pump. We look at steep vs flat pump curves, and friction vs static head dominated hydraulic systems. This is a thinking exercise at the whiteboard, not a detailed mathematical analysis.
- วิทยาศาสตร์และเทคโนโลยี
![ใจเสาะบนเกาะนาคี - อ๊ะอาย สกุณาพร เสียงวิหค [Official MV] | ค่ายเพลงเป็นหนึ่ง](http://i.ytimg.com/vi/jbcmGQkfyOg/mqdefault.jpg)
I love how pleased you are with yourself for this joke. Thank you, it makes the fluid dynamics go by easier.
I promise not quit my day job. Thanks.
0:00 - 0:06 this is why we get a bad wrap when people think of STEM professionals🤦🏾♂️
Thank you for spending the time to create and share this content awareness 🤙🏾
Awesome video Kevin, really hard to find content and specific when it comes to real world engineering.
Awesome Explanation. Thanks
Thank you sir, amazing simple and effective communication.
Wow Kevin Fantastic presentation
Thank you.
Good explaination. 😊
Thank you. Nice and Simple.
Terrific. Thanks for this.
Thanks with good explanation of vfd , thank you very much sir
Thank you
Kevin, thanks for the "Clear as clean water" explanation!
I would be happy to know if, in your experience, you have some idea on what percentage of the total of pump applications, the VFD is not favourable? Thanks in advance!
Of the applications where a VFD is used with a centrifugal pump, I think about 25% are not great. The strangest application I have seen is a VFD to modulate flow rate, plus a throttling valve to control discharge pressure. No surprise that MANUAL operation was common practice here.
@@kevindormaconsulting4763 Kevin, I do agree with your estimate of about 25% of applications are not favorable to install a VFD on pumps, but on your comment about having a valve throttling the discharge, it is not an absurd idea, because of how some of those 25% systems curves do interact with a variable speed pump, and also it needs to be taken into account what kind of dynamic pump is being used... not all rotodynamic pumps (the correct name of so called "centrifugal" pumps) are started with its discharge valve closed or just cracked open. It has to correspond with the "Specific Speed" or "Ns" value the pump has. Thus, in some cases, a discharge valve can be indeed useful in Variable Speed controls. In all cases, it depends on the shape of the pump's curve against the Hydraulic System curve against the power curve of the specific pump used on the system, which mainly depends on the pump "Specific Speed". An intelligent use of both VFD and Throttling Valve can be advantageous under some combinations of pumps and their systems. A good inderstanding of the system interactions is through the use of a 3D graph, plotting both system and pump curves and superimposing the effect of the valve position or Delta P. Best Regards.
Alfredo, I fully agree with your observations. The application that I saw was FAR less sophisticated than your experience. The pump had a very flat curve (low specific speed), and speed was varied to maintain constant flow. And there was a backpressure control valve to hold the pump discharge pressure at a fixed value. The intent of the control valve was to "ensure the pump always operated on the curve." A pump with a flat curve, operating with a fixed discharge pressure, is difficult to operate. The Operations Teams found the answer: Open the minimum flow recycle enough to move the flow rate well past the maximum allowable flow rate and into the steep part of the pump curve. Seal failures were common. The flat pump curve and perfectly flat system curve also had TERRIBLE flow control. The flow rate was oscillatory, and this had reliability impacts on the upstream vessel (frequent level excursions) and the downstream process operation (poor performance). Sadly, none of this evidence was accepted.
@@kevindormaconsulting4763 A pleasure reading your posts. Good information to spare. One thing I say in my pump courses, is that "the Rotodynamic pump is the simplest... but only mechanically, because it has the far from simple hydraulics".
Very nice tutorials. Sir I have a question: does the system curve changes if I have branches in the downstream pipe?
The system curve does not change if there are branches in the downstream piping. The system curve only changes if there is a resistance that changes (someone adjusts a valve). System curves can be difficult to describe for piping with several downstream branches because the destination pressures may be different for each destination. The system curve tells us the supply pressure that is needed to push into the destination pressure (at zero flow, the supply pressure matches the destination pressure).
@@kevindormaconsulting4763 thanks for your answer sir. I have another one: what about positive displacement pumps like an screw one? can I use VFD in that case? since the curve is only FLOW rate constant?
@@rucaskeeper666 a VFD is a good application for a positive displacement pump. Flow rate is proportional to pump speed. The
Hi Kevin, I hope you upload more videos.
I have a question, I have a motor whose nameplate indicates 3500rpm. Can I install a VFD and push the motor to higer rpm than 3500?
Can you recomend some books about this topic?
Are there any restrictions (apart from the explanations that you give in the video) related to motors and VFD?
CAn VFD be implemented on any motor?
The nameplate will be the rpm that the motor will run at rated voltage and frequency. The higher the frequency at the top end, the lower the torque. Depending on the application, this can be an issue in itself. There will be diminishing returns here and if you need to run faster than 60hz, then you're most likely going to need a different motor than the one you're intending to use. VFD's can not be put on just any motor as well. I'm sure you can find most of this information to answer your questions with a quick google search.
I have asked that question a few times. The answer I was given is You might be able to go a little faster than synchronous, but not a lot faster. Higher speed draws more power and may damage the motor. I don’t know if a VFD can be applied to any motor. Good question.
Congrats for the video. It is good to see someone sharing knowledge. Sir another question: by using a VFD and changing the speed of the motor there is no posibility of problems related to speed? I was reading about sincronous and pump curve in terms of torque and speed and it seems this speed should have a calculations in order to match the curve torque vs speed in the pump
Good question. If the system curve is flat then the motor requires constant torque and variable speed. The amount of heat generated by the motor is proportional with motor speed. But a fan cooled motor may lot be able to provide enough cooling at reduced speed. I have not calculated the amount of cooling at lower speed. I suggest only using VFD where there is steep system curve and no danger of overheating.
Kevin I would like to ask you a more fundamental question: why the flowrate of the increases with the increase of the pump speed?
That is a good question. I have never seen a derivation of the affinity laws. I need some time to think about this.
Thank you so much for your reply! Together with this question, I try to give an answer to another similar question: The pump creates/increases pressure or flow?
Hi sir what do u mean by large gain @6:54? are u referring to the power?
This is a large gain in static head. A large increase in pressure or elevation.
Hello, I have a question. I want to propose a control strategy for an HVAC closed-loop system, where it works at partial load. We need to reduce the electrical consumption energy using a VFD. The system must have a constant pressure in the mainline to ensure enough flow to the farthest point or load. I already checked the pump curves and the pump can reduce its flow working above the minimum flow permitted. The minimum flow is 50 gpm and in the control strategy, we can reduce until 300 gpm. There won't be any trouble if the pump works above the minimum flow value.?
Did you really think this guy was going to answer your question? He's not here to do free engineering work for you.
@@Impedancenetwork I already checked by myself, and I solved the issue without his help or yours. Unfortunately for some people like you, every thing is about profits and benefits.
@@Rafaverdadhai bro..i got same question in my mind..did u able to introduce the vsd without sacrifice the head pressure?
Kevin
I need a clarification about the valve controlled recirculation loop installed in a centrifugal pump.
When DISCHARGE
The pump delivers fluid to two destinations. The main destination is the process. If the process takes less fluid than the MCF, then the pump will have seal problems if it is operated for a long time. The minimum flow loop opens another path and ensures that the total flow rate through the pump is at the MCF. In this case, we need a flow meter on the total discharge from the pump (the main process flow plus the recirculation flow). Did this answer your question?
@@kevindormaconsulting4763
Your reply is convincing and satisfactory for the case where the system requires lesser fluid recirculation line enables the balancing of pump fixed discharge. My company is installing a recirculation line to a desalter crude Feed pump where the throttling of feed is not essential and it remains constant, what could be the need for the recirculation line towards balancing? Kindly comment plz
@@sabeerali5610 Dear Sabeer, it happens that one of the expertise areas we developed in my engineering company, was precisely the intell unigent selection of the best type and size of pump for feeding the Crude Oil Dehydrators or Desalters at our State owned petroleum company ("PEMEX").
In that service, the use of a pump is required mainly because of the Desalter needing a certain pressure margin above the oil-gas equilibrium pressure, owing to the gas tendency to be released from the crude oil, and thus creating small bubbles that interfere with the water separation from the oil... That happens because the gas-liquid equilibrium corresponds to the last oil- gas Separator before the Desalter, and the fact that the Desalter or Dehydrator has SOME pressure drop resulting from the internals, like the crude oil distributor and its piping. Therefore, all Desalter vendors establish the requirement of supplying and maintaining at least 20 PSI above the gas-liquid equilibrium, to avoid the gas fraction to separate, form bubbles and interfere with the water droplets separating and falling to the bottom of the Desalter. Thus, the need to install pumps before the Desalter in order to rise and keep the pressure inside the top if the vessel at least 20 PSI or 1.4 bar above the pressure at the last separator before the Desalter.
BUT, almost any pump will unfortunatelly produce the highly undesirable water droplet shearing, due to the turbulence and mechanical shearing of the water droplets... Even Positive Displacement pumps shear the water droplets enough to reduce the separation at the Desalters. Remember that separation of those droplets is governed by the Stokes Law, which says that halving the mean droplet size, will require FOUR TIMES for the same separation, and reducing the droplet size to one third, would necessitate a NINE TIMES larger vessel for the same separation (because the droplet diameter is squared). Therefore, it is very important to keep the water droplet size as large as possible, which dictates that the pump must have the lowest droplet shearing possible. As this subject is not easy or trivial, I suggest you to contact me at amarquezclaussen at hotmail dot com , for an ample discussion on these matters. For the moment, allow me to tell you that many standard pumps used before Desalters, are totally inadequate in regards to the harm they impart to the oil-water emulsion, and that makes the separation quite more difficult because care is not applied to pump type selection or its best sizing and nothing is done in order to have the pump operate at the lowest damaging droplet shear on its curve, and remember that any valve placed before the Desalter will also cause undesirable droplet shearing. A different engineered approach, that takes into account the droplet shearing, will reduce the damage and promote a much better separation efficiency from the Dehydrators or Desalters.
Additionally, using the conventional "mixing valve" to mix the desalting water stream with the crude oil to be desalted, results in lower Desalting, reduced separation, and a larger water use. There, a more intelligent approach is in order too. Myself, I have dedicated many years at studying and improving the Electrostatic Dehydrator or Desalter operation, with measurably good results, and that demands a different pump selection and sizing criteria, that is NOT the usual criteria followed by pump vendors and manufacturers. Go and ask any commercial pump vendor, and require the "Emulsification Parameter" of their pump when pumping a Oil-Water mixture to a Desalter... I'm quite shure he won't be able to answer, because 99% of pump designers, have never taken the consideration of the undesirable mixing and emulsifying effect of their pumps, having only dealt with single liquid phase fluid handling, but pumping a crude oil-water emulsion is quite another matter, and too little has been done globally in order to keep the water droplet size distribution without having little droplet shearing. I wait for your contact soon. Sincerely, Alfredo M. Claussen, 42 years experience at the Mexican Institute for Petroleum in Mexico City.
@@alfredomarquez9777 Awesome explanation😀
Kindly give me sometime to understand your context in full and I shall definitely get back to you vide your email stated Sir.
You read my mind clearly and wrote a wonderful and convincing reply.... My sincere thanks and hearted greet for your efforts Sir👍
@@kevindormaconsulting4763 Sure I shall revert to you with the typical sketch of recirculation project soon
how would this apply to residential well water systems? Good idea or not?
A well water system is dominated by static head, so the system curve is flat. For a VFD to work, the pump curve needs to be very steep. I have well pumps with steep curves, but only for large industrial systems. A small pump for a residential well is likely to have a flat curve. No, I don't think a VFD is appropriate for this application.
@@kevindormaconsulting4763 interesting, okay thanks.
@@kevindormaconsulting4763 Hmm ... my local well pump service company is suggesting a new pump with VFD for my 195 ft well, which currently has a 45 year old, 1.5 hp, 240v submersible pump. This system is no doubt dominated by static head as the above-ground pipeline length is about 100 ft with 1.25" diameter pipe. I assume the well service company will choose a compatible pump that has a steep pump curve, but can I be sure? My use is typical for a home (two residents) with little or no landscape irrigation. I'm having doubts that I need a VFD rather than a conventional pressure tank system, and worried about the problems you mentioned in the video. What is your advice? Also, will I need 3-phase, 240v power at the well to use a VFD? Or, does the VFD take 1-phase input and convert to 3-phase? Thanks!
@@GKALLIO interesting application. If this is a multistage pump then it might have a steep curve. Ask the service company to provide the pump curve and check if it is steep. Ask them how they will test their application to ensure that the pump can provide low and high flow while maintaining constant water pressure at the surface. The long run of pipe May provide enough resistance to make this really suitable for a VFD, but only if the pressure sensor is located at the far end. Your gut feel may be worth following.
for the scenario at 9:20 cant the VFD operate at say 88% and still be within the system curve?
In this case, the pump would only be able to run between say 88 and 92%. It would be difficult to control the flow rate. A control valve gives better reliability in this case.
How to start up a varible speed pump
VFDs are not a cure all. For large HP applications, the VFD can be the size of a truck trailer and very expensive. VFD's can also play havoc with plant controls if the wiring between the controller and the motor is not properly isolated away from other control wiring or properly grounded. Also the motor selected, especially with high HP motors must be specifically designed for VFD control. This means retrofitting a pump with a VFD need to be fully investigated with the motor manufacture. As with most things, the devil is in the details. VFD applications need the combined effort the vendor, an experienced mechanical design engineer and an electrical engineer who has experience in controls and motor design. VFDs work great when done right.
You are completely right Bob: The many VFD observed failures have the Devil in the details. Details like improper wiring (VFD cables have special construction and must be terminated in the correct way). Many times there is a definite limitation on their Lenght and Electrical Impedance, Filtering and Harmonic control, their routing and undesirable coupling, proper grounding etc. But that does not necessarily mean that it can't be done with excellent results, just that ALL those "details" need to be correct. As time progresses, VFDs are getting more rugged and reliable every day. The use of more electrically rugged power semiconductors on the modern VFDs is making them more usable and reliable, but their installation still needs to be perfect.
I can't understand
t=1:30 -> Can you really say definitively that a control valve uses less power than a variable speed drive? There's a small loss in the VFD and clearly a loss in the CV (depending on how you have sized the CV).
t=3:21 Typically a system curve is the pump differential pressure, not the inlet pressure. Inlet pressure is a guage or absolute pressure. Saying that the pump inlet pressure will (must) rise as flow increases is either confusing or ignores some other assumptions that must be true. In fact the source pressure in your example is probably not changing and the incrased flow rate is causing greater frictional losses, so the inlet pressure is dropping with increased flow rate. Pump head is a differential pressure, so to plot inlet pressure on a pump head curve is a little confusing. However, plotting system differential pressure on the curve does make sense.
t= 7:43 So, that's a good point. Thanks for pointing that out.
A VFD should (almost always) consume less power than fixed speed pump and control valve. This is because the control valve must take 10-20% of the total pressure drop. I did not clarify where the inlet is. This is the pipe inlet (pump discharge). I assume pump suction P constant. System curve is the pipe, starting at the fixed outlet pressure. Pump curve is the pump discharge, which is the pump DP plus suction pressure.
@@kevindormaconsulting4763 I think there's also a case for using a control valve if the situation calls for a process to be isolated from the pump, and using a valve is necessary for that.
Where there are multiple destinations, it's possible only one of the control valves would be eliminated with the use of a VFD. You still may save some energy with a VFD though.
NEVER, wherever possible particularly on single phase pumps. .
In principal...great. in practice...do not go there. My advice.
@@christophertownley6734 I was being generous with the suitable applications for a VFD and a centrifugal pump. I have seen laughable applications of a VFD with a centrifugal pump, and people still insisted the seal failures were not caused by running way outside the curve (despite the data). Flat pump curve / steep system or steep pump curve / flat system will require a lot of analysis to ensure the pump is reliable and operable. Steep / steep is a great fit, but I have not seen many examples of this. What did you mean by single phase pumps? Single phase electrical?