6 replies to this topic

[steven12]

Hi guys. I am currently doing a little hands on project involving a feasibility study. We have a chiller that has currently got extra capacity and we are trying to provide the cooling water to several mills, rather than just one mill it is connected to at the moment.(so right now alot of cooling capacity is not being used).

First I have to make sure the pump can provide the flowrate of cooling water to all mills. I assumed a safe figure for total flowrate. (if it were to provide for 5 mills instead of one). The problem is the technical data supplied on the pump is not very detailed. It had

pump(1.1kw)
Pump flow (at max pressure) 3000(L/hr) / 500(kpa)
Pump flow ( at max flowrate) 14000 (L/hr) / 250 (kpa)

I have calculated the possible pressure drop from all the heat exchangers and pipes and fittings, which comes to about 300kpa, and there is no height elevation. And I need a flowrate of around 9000L. How do I know the pump on the chiller is adequate?

Another problem I have is, there is no flowmeter from the chiller side, so i have no idea what he flowrate of the cooling water it is providing now, is it somehow possble to calculate this?

---------------
On the chilling side,the chiller gave 40000W at 10C supply. I was trying to convert this to ton of cooling. I was wondering if i did this correctly.

40000* (60min*60s) = 144000000(J/hr)
=136600 (btu/hr)
= 11.4 (ton)
(*12000btu/hr = 1 ton)

the mill that the chiller is connected to now has 2gpm cooling requirement, so i used the equation

Q(cooling req)= 2 gpm * 500 * dT(3.74k)
= 3740 (btu/hr)
= 0.311 (ton)

spare capacity = 11.07 (ton)

I dont know if that is correct. can someone confirm this please.

[Art Montemayor]

Steven:

Welcome to our Forums. I don’t understand how you are handling an industrial project (however trivial) if you are a student. Nevertheless, I tried to make heads or tails out of your explanations and equations and, like most students, you don’t communicate that well. Let me explain:

You need a centrifugal pump performance curve (although you failed to say so, I have to assume you are dealing with a centrifugal pump) in order to calculate if your pump is going to be able to do the job you need. There is no alternative. Having just two points of the performance curve doesn’t cut it. The way to work the problem is that you calculate and draw what is called a “System Curve” for your piping system. This defines the system on your performance curve graph and it intersects the pump’s performance curve. Where the two curves intersect is where your pump will operate. And that’s the answer. Therefore, if you don’t have a performance curve, you can’t generate the correct answer. All you can do is guess. Try to obtain a real performance curve for your pump from the manufacturer. At this point, I have to also assume that you know what I am talking about. I don’t want to talk over your head, but I don’t know your profile or your level of technical preparation or training. But I have to assume something.

You can back-calculate the refrigeration capacity of your chiller the way I think you have tried to do it, but you have to tell us what you did. Engineers don’t just do calculations and let others try to guess what logic or algorithm was used. We can’t read your mind. You have to tell us what it is that you are calculating and where and how you got your data for input. For example, what do you mean by “the chiller gave 40,000 W at 10 oC”? How did you get the power rating – did you read it on the nameplate? All your calculations are just power conversions.

Next, how do you employ the sensible heat transfer equation, Q = W Cp (T2-T1)? What do you use for the temperature difference? Are you assuming that you would be cooling the 2 gpm of water flow? If so, what is your temperature range?

If your heat load in the mill is 0.3 tons of refrigeration and you have a chiller that can put out a NET 11.4 tons of refrigeration then yes, you have 11 tons of refrigeration as extra capacity. But what that means is you are going to have to control (turn-down) that capacity – and that is going to be a BIG problema. Not knowing the type or manufacture of chiller you have, I can’t even guess how to turn-down the unit. But the magnitude of the turn-down is huge – in my opinion – and doesn’t seem practical. You need more cooling load for the chiller to run continuously and efficiently.

I hope these comments help you out. If you need help in performing a System Curve, state so, and we can help you on that. If you really did accurate and correct pressure drop calculations using the Darcy Equation, then you should be able to do a System Curve.

[steven12]

Art Montemayor

Thank you for the reply, I apologise for not being clear in the beginning. I'm currently a student doing some vacational work experience in a company. They have hand me some small tasks and I am just like I said, doing a feasbility of whether it is possible to use the extra cooling capacity the chiller has, and if thats the case, would the pump have enough power to pump the new flowrate.and of couse I'm supervised onsite. I have just completed second year, and this is the first time I applying my limited knowledge of chemical engineering.

To answer your question on the equipment. The chiller is connected to only 1 mill at the moment (Drais Megavantis ACS-101). The chiller is produced by an australian company. Aqua cooler R2000. There are 5 more mills and 2 tanks that we are thinking of connecting to.

Yes, I do understand to work out what the flowrate is without a flowmeter, I need a pump characteristic curve for the pump, and construct the system curve by working out dynamic head as flowrate increases. where it intersects on the pump curve is where it should operating. Now I dont have the pump curve, all the technical data provided me was the 2 points. And I dont know if the valve for controlling flowrate is fully open or partially open (assuming it has one).
----------
Side note

I was wondering if it is possible just to draw a general curve around these 2 points. since i know what most pump curve looks like.

Pump flow (at max pressure) 3000(L/hr) / 500(kpa)
Pump flow ( at max flowrate) 14000 (L/hr) / 250 (kpa)

(I'm trying to get in contact with company but I havent got a reply yet).


-------

Now back to my problem, there is no flow meter anywhere on pipings, between the chiller and the mill. So like you said I have to get the pump data and make my best guess.

But here is something I dont understand.

The technical data provided for the mill(megavantis)requires 2gpm of cooling water supply. But on the utility requirement page. It suggesting a cooling water flowrate of 2500L/hr. now 2gpm converts to about 450L/hr. So this doesn’t add up, I can only assume 2500L/hr is what I should be working with.

Another utility requirement states:
Pressure differential min.(3bars) . ( if no opening cooling water outlet but counter pressure exists after the mill in a closed cooling water circuit) . I have no idea what that means.
-------------------------

I have another problem with this chiller/mill setup. On the entry and return pipe at the Chiller. there are 2 pressure gauges. So the pressure difference is how much head is lost by pumping the water through pipe and the mill.

I recorded the pressure, the outgoing pressure is 270 kPa,and returning pressure is 10 kpa. so there is 260kpa is lost.. Now I have calculated the pressure loss for the piping, which is miniscule at the moment, there is not height elevation. So I must assume the bulk of the pressure loss is lost due to heat exchange in the mill?(not sure if this has something to do with the pressure differential).But the pressure seems abit too high, doesnt it?

If this is the case , and using a flowrate of 2500L/hr. (2500 L/hr =requirement for the mill)

Power required will be

P = ( 260000/ (9.8*1000) ) * ( 2.5) / (368 *0.7)
=0.25 (kw)

Of pumping power right, is that right?

But that is the trend , with the way things go, I seriously doubt there will be enough pumping for 5 mills (WAB dyno-mills) and at 9000L/hr


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

On the chilling side.

The 40000W/ 10C is the cooling capacity, I assume it means it can cool 40000W if we want to produce 10C of supply water from the return water. On the data sheet, it listed cooling capacity at 5C.10C,15C and20C. The higher the supply temperature, the higher the cooling capacity.

There is no temperature gauge as well, so I manually tested the pipe surface temperature at the miller on the entry/outlet using a special device. There was 3.74k of temperature difference. (I took the temp readings on few different spots and averaged them).

So if I assume for the moment the flowrate is (2500L/hr)(this just an assumption, the actual flowrate will probably be a higher) . A simple m.cp.dt gives me 10.8 kW of cooling is required. Is this correct?

So 40 kW -10.8kW = 30 kW
------------------------


What I normally do to calculate system curve, I usually just assume velocity of 1m/s. use that to multiply by the pipe cross-sectional area to get the flowrate. Find the Reynolds number and find the corresponding friction factor, then add in equivalent diameters for bends and fittings. Then use (8f(L/d)pu2)/2 to find the dynamic head. Then use 2m/s velocity and so on. Then plot it on the pump curve.

Now I don’t know whether to include the pressure differential for the mill which is 3 bars, as the static head . As you can see I observed a pressure loss of 260kpa which shouldn’t happen if the pressure is loss due to just the pipings. The bulk of that pressure must be due to heat exchange or the equipment. If heat exchanger pressure loss remains constant regardless of what the flowrate is, then is should be static head. But if it increases with the flowrate, then its part of the dynamic head? If that’s the case , how do you adjust it accordingly with the flowrate?

[Art Montemayor]

Steven:

Thank you for the clear and detailed response. The length of your response is not important. What is important is that you have explained your case and given your basis for your calculations. This enables all of us to analyze and reply to your query with much more accuracy and value.

Before going any further, now that I suspect that you have had at least a preliminary course in Fluid Mechanics – or an introduction to the subject – you should get a copy of Crane's Technical Paper #410. This is the professional, basic document that defines the engineer’s needs and tools for employing Fluid Mechanics in the design of process and utility piping. You find yourself in the lucky position of being out in industry with a hands-on challenge in front of you and a sponsoring mentor supervising you. Use your supervisor to guide you to a copy of the Crane Technical Paper (hopefully, free) from the company library or from a valve or fittings vendor. Often, vendors give free, bound copies to regular customers or to friends. It is an opportunity you cannot afford to miss out on. I am attaching a copy of my workbook where I have presented the examples of the problems resolved in the Crane booklet. Study them and you will find the answer to a lot of fluid flow problems. Get the book and study it and you will understand everything else regarding fluid flow – prior to taking the other 2 years of related engineering fluid courses in your future. Now to your questions / comments:

1. I don’t believe that you can correctly draw a pump performance curve with only two points identified. And I challenge the source of those two points. The correct thing to do in an industrial, real-life situation is to refer right back to the manufacturer and obtain the correct, applicable performance curve for the pump in question.
2. First you said the mill requires 2 gpm (454 lph); now it seems it could be 2,500 lph (11 gpm) – which sounds more reasonable. Whatever it is, it should be accurately established before going any further. Again, you do this by referring back to the manufacturer. A telephone or email should suffice. While you are at it, also request the mill manufacturer to give you the expected pressure drop of cooling water going through the mill’s cooling coil (or bundle). It is important to get this figure in order to proceed with your calculations.
3. You say you are recording a pressure drop loss across the chiller’s water circulation loop of 260 kpa (38 psi). this is relatively high, but I can’t say that it is abnormal. You haven’t furnished a piping layout or isometric sketch. This could be easily done on a spreadsheet, together with the hydraulic calculations. Again, without the mill’s certified, expected pressure drop you can’t really attribute any certainty to your pressure drop expectations on the loop. A phone call or email is called for here, as I stated earlier.
4. The cooling capacity of the chiller is normally rated in tons of refrigeration – which, as you know, relates to the heat flow of a “ton” of refrigeration as 12,000 Btu/hr. However, what you are not explaining is what is the cooling range of the circulated water in the chiller. In other words, you must fix the inlet water temperature to the chiller as well as the outlet temperature. Knowing the two, you use the sensible heat transfer equation (Q = W Cp (T2 – T1) ) to resolve the circulation rate required to furnish the cooling demand. Stated another way, the same equation gives you the amount of water that can be circulated (W) in order to furnish the rated cooling capacity (11 tons of refrigeration) while cooling the same water within a given temperature range.

I am surprised that you don’t know the Darcy-Weisbach equation or its derivation. Any Fluid Mechanics course has to deal with this basic relationship – whether you call it Fanning, Colebrook, or any other name. The basic equation is the basis of the Crane Tech Paper #410 and that is why it is so important to obtain and study that booklet. Without an accurate and complete understanding of the fluid flow and mechanics involved, you cannot be expected to generate a credible System Curve. You have described the generation of the System Curve correctly. I fail to understand why you don’t know of Darcy-Weisbach. In any event, call it what you may, you must perform the correct fluid calculations – and to do that you need the correct data on the expected pressure drops across the equipment. These you obtain from the manufacturer or your supervisor – mentor.

I hope this helps you out.

 Crane_Fluid_Flow_Problems.zip   659.79KB   219 downloads

[steven12]

Quick Update,


I have obtained the pump curve from the chiller company, its a P7 pump and i have include the pump curve in the excel file. I have also included the technical data for the chiller into the excel file as well.(I put the scanned image in different tabs).

I know the discharge pressure from the chiller pump is 270kpa, or nearly 3 bars. So I drew a horozontal line cross the P7 pump curve on chart. Straight down it points to 2.7 L/s. using the Delta T
I measured before for the Drais Mill(ACS101) inlet and outlet of cooling water, which is 3.74K

so 2.7*3.74*4.18= 42.2 (kW) of cooling.
=12 (ton)

This is nearly all the cooling capacity at 10 deg supply, which is 40000W, if thats the case , the chiller is barely able to handle the cooling cooling!!!!

I'll post an update on this latter.

Attached Files

•  data.xls   225.5KB   148 downloads

[Art Montemayor]

Steven:

Please refer to the attached workbook that I made comments and notes in.

Your pump curve is not an orthodox one. That's OK. However, the manufacturer must state outright the suction head he/she is assuming when they use the pump discharge pressure (as you assume) as the value for the ordinate.

Pump performance curves are normally drawn with the TDH (Total Developed Head) as the ordinate value. The TDH is normally the difference between the Discharge Head and Suction Head (in feet of liquid).

Therefore, unless you know the Suction Head (which is not the case on the graph) you can't use the pump's discharge pressure as your "TDH".

Besides that, I get 3.0 l/sec on the graph.

You've done the correct, engineering footwork by reverting back to the manufacturers. That's what I recommended in my last post. Keep it up - go back to the pump manufacturer and ask for the true head definition as depicted on their performance curve. See how important Fluid Mechanics is?

Good Luck.
 Steven_Chiller.xls   261.5KB   214 downloads

[steven12]

Art:



Thanks for replying, From your response,and i will try to work through those examples.

I assume the 270kpa I recorded at the pump outlet was a gauge discharge head (Hgd) but not Total discharge Head? and to calculate total discharge head and and total suction head I would need the velocity of the water at corresponding locations? Since Total Discharge Head= Gauge discharge Head +velocity head.


*The problem is I dont have any flowrate readings.(If thats true

Or did I understand that wrong, and that pressure gauge indicates the Total Discharge Head (which include velocity component), and all I need is the suction head gauge reading, then I can proceed with

TDH = H(d)- H(S)

*Side Note
(One more thing I forgot to mention, the manufacturer stated a by-pass valve is used to control the flowrate. From what i understand , the bypass valve directs some of the discharge flowrate back into the entry, and recirculates it. There is a region on the pump curve (0(L/s) to 1 (L/s)) , which is grey and states that within that region, mandatory bypass valve is to be used.
Does that mean that even though the pump can operate at 0 to 1 (L/s), it should never be operated within this range because it is undesirable, and instead pump should operate above 1(L/s) , and a bypass(recirculation) should be utilised to achieve below 1 (l/s) )

I'm go back and ask the manufacturer if there is a flow meter somewhere inside the chiller, would make my life so much easier.I think i will have alot of difficulties developing my own system curve, because first I dont know how to account for the heat exchanger as flowrate increases( the cooling water actually split up into 3 streams in hoses in the mill, going into different parts, then rejoin again at the exit),secondly I dont know if the bypass valve is fully open or partially open.(and if it is partially open , how do you account that in the darcy equation?). So If I know suction gauge pressure and discharge pressure, I can find TDH and find the flowrate. but then that will only give me flowrate for that single mill circuit, I cannot predict the flowrate for muti-mill circuit......(so i still need to develop my system curve).

(hold on, if i know what flowrate is required, which I do (8700L/hr), I can simply calculate the velocity, find the pressure loss in terms of head, Now if that head using that velocity is UNDER the pump curve, it means it has extra head available which simply means YES it can provide the flowrate I want. Is my logic correct? I know extra head would convert to faster flowrate, but that doesnt concern me cos I can simply use the bypass valve)

Another thing I think I have to find out whether the flow of water is in closed circuit loop, or does the return water go into some kind of reservoir, cooled and redrawn from the pump back to circulation again, are you familiar with the way chillers are usually designed? I remember my supervisor mentioned something about a sump which explains why the pressure on return pipe was so low. I have to clarify that.

Another problem Im facing is estimating the heat outputed by two large grinding tanks, these things are like huge, 10000L each, they are used to grind certain solvent/primer? I'm not sure, the thing is they have no temperature gauge anywhere on the inlet and outlet of the cooling water, and manual method of temperture detection with my heat sensor equipement dont work cos the piping is too thick to detect any temperature difference. Is there someway i can estimate the heat output, assuming i know the RPM, impeller size, and solvent viscosity etc? (this is more trivial than my first 2 problems which are finding the pump flowrate and cooling capacity, for now im putting aside these 2 at the moment , cos I seriously doubt that chiller can cool these 2 tanks).

Update: I called up the installer guy for the chiller, there is no flowmeter inside the chiller, and appararently the chiller is not a closed circuit, the return water goes into a 500L tank, where the water is cooled, and redrawn again back into circulation. He suggested I take a pressure detection equipment and measure the pressure at the inlet and outlet of the mill to find the pressure difference. Then again he doesnt know about how much the bypass valve is bypassing the discharge water back into the tank and by how much. Im thinkinng the easist way would to disconnect the return pipe to the chiller and just find out manually how much water is flowing back.
* If the water is redrawn from a tank after it has been chilled, assuming i know what the dimension are, cant I just work out the suction head by finding the height of the tank? assume the pipe lies on the bottom of the tank? that way I know what my total head developed is by (TDH=discharge head- suction head)
*how does that bypass valve affect my system curve in the scheme of things?
*I have also been suggested acquiring a magnetic flowmeter. Im pretty sure the cooling is not deionised. I have to rent it somewhere though.

I have uploaded a copy of the chiller/mill schematic.

Attached Files

•  chiller_schematic.pdf   823.45KB   121 downloads