10 replies to this topic

[smuk]

Hi,

I have a problem of finding out flow through a cooling water network, comprising of piping and heat exchangers. The size, length and fittings for the piping network is known. Also number of tube passes, size of tubes, number of tubes per pass, size of inlet/outlet nozzles are known for the heat exchangers. For a given cooling water supply and return pressures, the problem is to find the cooling water flow through each and every branch of the cooling water network. Can this problem be solved with Aspen HYSIS?

Thanks

smuk

[breizh]

Hi smuk ,
I would advise to buy or borrow a portable flow meter like ultrasonic doppler meter to do the mapping of your network .It's worth to have it .

Breizh

[Zauberberg]

The approach you are looking for is something that is done during design stage. For practical purposes, in an operating plant, you'll never get your simulation outputs even close to the real values. I know, I have tried many times.

Take an excellent advice given by Breizh and obtain a non-intrusive flowmeter. That will certainly provide you with more accurate information regarding the actual cooling water flow. These devices are fairly simple and relatively cheap, and definitely fit for purpose.

[smuk]

Hi,

Thanks for the replies. The cooling water network in the plant is known to have restrictions to flow, mainly due to undersized piping. The approach suggested is excellent for finding out restrictions in flow due to fouling of the water side of the coolers. The task at hand is to suggest increase in line sizes considering the entire network. So a simulation approach is being considered.

Thanks,

smuk

[djack77494]

smuk,
You need a simulation package specifically designed to handle complex hydraulic networks. Many years ago I used such a program known as "Fathom". I know it is still available and think it may be the answer to your problem. I would not attempt this type of problem with Hysys (but that doesn't mean it could not be successfully done).

Edited by djack77494, 23 August 2010 - 06:06 PM.

[Erwin APRIANDI]

since this is for pipenetworking, tryout pipenet from sunrise sys
this is a hydraulics simulation tools which normally used
for the fire water system ring, which quire similar to your cooling water syetem

for this application, still you have to know the flow first then simulated
this tool will only calculate the hydraulics of fluid in the system
but it will not recommend the flow, as the flow of cooling water
is also depend on the consumer, how much it is required to cool

the software is quite cheap comparing to other
and steady state is enough for your need

[smuk]

Thanks for the help.However, can you tell if Pipenet will also handle heat exchanger pressure drops.

Thanks

smuk

[Erwin APRIANDI]

Thanks for the help.However, can you tell if Pipenet will also handle heat exchanger pressure drops.

Thanks

smuk

smuk, you can simulate the heat exchanger as a valve and put the maximum pressure drop of the heat exchanger.
I'm not sure whether in the new version they have added a heat echanger. but in case os steady state, what you need to consider from the heat exchanger is only the pressure drop

[Padmakar Katre]

smuk,
You need a simulation package specifically designed to handle complex hydraulic networks. Many years ago I used such a program known as "Fathom". I know it is still available and think it may be the answer to your problem. I would not attempt this type of problem with Hysys (but that doesn't mean it could not be successfully done).

Dear DJack,
Indeed an useful and reliable tool by Applied Flow Technology. I happened in past to model the cooling water network with AFT-Fathom which was really helpful.

[Shehryar]

Going through a complete Cooling Water Circuit is difficult, but not impossible, depending upon the number of Water Cooled Exchangers u have. You can use Aspen Hysys Hydraulics for your network, but for this you must have each exchanger heat duty, CWS & CWR temperatures (will help in adjusting CW flow through that Exchanger) , You must also have isometrics with you, to have exact take off point of CW supply for each exchanger. By adding multiple Sub-FlowSheets in Aspen Hysys with Each Exchanger it can be done. It will take some of your time, as you might have to go through your plant data, but I believe one can do it this way...

[kkala]

Having worked in a fertilizer plant (1975-81), I was surprised at the difference between design flows and actual flows, even in the same pipe of unchanged diameter. Extending this a bit (no simulation at that time), I agree to Zauberberg's opinion that a simulation alone will most probably not give results precise enough. E.g. pipe roughness can increase up to 10 times in long term; scale effect on ΔP cannot be precisely estimated; flow division into two identical branch pipes can be of different rate depending on status of each branch line and its valves. If you are far from the plant, it is hard to have "experimental" data; but if you are in the plant (it seems so), you can combine actual plant data with simulation. That is to try "interpreting" actual data through simulation by changing its parameters. Some data may be wrong, needing "reconciliation", which can be indicated by simulation.
Pressures around exchangers are usually available, as well as at a few pipelines. Flows in lines without flow meters are not easily measured (e.g through diverting them to tanks, sumps, barrels of known volume, while using a chronometer); pump curve & suction / discharge pressure do not seem to give results precise enough (pump can be run down, or curve rather flat).
If I knew "non intrusive flow meters" at that time (1975-81), I would try to have one bought for the fertilizer plant (on condition that price was then reasonable). Operating principle of such flow meters can be seen e.g. at http://www.rshydro.co.uk/index2.shtmlor http://www.flowmeter...c_02051301.html; there seems to be two types, "doppler" or "travel on flight". Disadvantage of doppler type is the requirement of 100 ppm of suspended solids (or gas) to act as reflectors to ultra sound waves sent, look athttp://www.omega.com...flowmeters.html. Cooling water does not normally contain them. Nevertheless these non invasive flow meters (mentioned by breiz and Zauberberg) could make a way out, even if injection of a "reflectors" solution into cooling water is needed during measurements. Of course getting such a flow meter (after investigation & order) may need some time.
After above, following could be a rough suggestion for the task.
1. Take a representative "photograph" of the situation, by measuring flows and pressures over whole cooling water piping and exchangers (as much detail as possible). This may need more than one trial for the case to be representative (steady state and normal conditions).
2. Insert measured data into the simulation software. Try to explain deviations between measured and simulation results; modify simulation parameters accordingly with an intent to nullify deviations. Such parameters can concern line roughness, scale (but ΔP software known to me does not "understand" it), a control valve or a common valve, a filter online, etc. Changed simulation parameters had better be withing expected range, something that could occur but not be detected. Not expected values could be given to few simulation parameters, only in case that measured results cannot be obtained otherwise (keep an eye for reconciliation of them later).
3. When deviations get small enough (meaning that actual and simulated case is close to each other), check velocity and frictional pressure drop (ΔP per unit length) according to applicable practices. Max allowable flow, velocity and pressure drop for lateral and main cooling water new pipe can be also found in C. R. Branan's "Pocket Guide to Chemical Engineering", Table 1-3, Gulf Publishing Co, 1999. Replace piping of too small a diameter, restricting the flow. Heat exchangers and control valves should take majority of frictional ΔP, not pipes.
4. After these modifications, run updated simulation and recheck remaining piping (there will have been some redistribution of flows). Parameters for not affected piping have to be same as per step 2. Final updated simulation can be a prediction of the new situation, as good as it can be.

Edited by kkala, 01 May 2011 - 07:27 AM.