19 replies to this topic

[Bill B]

We are looking at replacing a cooling tower. The main option being explored uses CT pumps with TDH = 200 - 250 ft hd. Some of the exchangers fed will be 75 psig. Thought is that due to frictional losses, and to a less extent elevation changes, that the exchangers won't see 75 psig.
My experience is that TDH of pumps is always less than equipment design, so this is new to me.
Anyone have any experience with CT pumps with > equipment design pressure?

[fallah]

Bill B,

Pump shut off pressure should be equal to or lower than downstream equipment's design pressure, otherwise the equipment should be protected against overpressure by a relief valve with proper set pressure.

[latexman]

It's worse than what you describe though. A PSV in liquid service usually needs an inlet pressure < 0.7 x set pressure to reseat. So, if you get lucky with the elevation and frictional losses such that normally the PSV(s) see < set pressure, if any little transient pops the PSV(s) open, they will not reseat until the inlet pressure < 70% of the set pressure. Those heat exchangers may have to have CW valves closed to get them to reseat!

[thorium90]

What i did was to screw on another PSV onto the leaking PSV. haha... Not very permanent solution, but we cant close the CW while the unit is still running...

[kkala]

Pump shut off will be roughly 250*1.2=300 ft H2O g =130 psig at pump level, exchanger will "see" D.P. < 130 psig due to (understood) higher elevation (precise calculation through actual data is needed).
75 psig is understood to be operating pressure of cooling water in the exchangers, can you clarify corresponding design pressure (DP)?
In such cases of used exchangers here, local Mechanical Department has examined possibility of increasing the relevant DP (rerating), just to avoid installation of PSV in liquid service (there can be disposal issues, beside post no 3).
They first look into exchanger mechanical design data to conclude whether this is theoretically possible. If so, a hydraulic test follows.
It is not known to me how corrosion allowance is considered in hydraulic test (stage of hydraulic test not followed by me).
Chances are that eventually the higher DP will be certified, if required increase in DP is small.
It may be worth while investigating this option.

[Atttyub194]

Dear Friends

Good day!

First of all, I would request my friends not to discourage any one comments as such such forums are to share ideas and gain knowledge our selfs.
Friends, I have been involved in the revamping of similar cooling tower system and the question my friend asked can not be answered with carrying out detailed network analysis.At that time we used Fluid flow and sunrise softwares to carry out similar analysis

I would suggest to carry out a detailed pressure survey of complete your cooling water system along with folw measurement to exchangers using utra sonic flow meter , available in market.This will help you is establishing the real bottleneck or you may eliminate the need for replacing cooling tower

Some time your exchangers are already operating at chock , in this case you have to evaluate the cooling tower using software such CTI so that the approach accross the cooling can be increased instead of increasing the flow rate

In most of Ammonia plant , the cooling water header is designed at 100 psig with cooling water pumps at discharge pressure of 60 to 75 psig. On cooling water circuit , you may have seen rupture disks rather than the PSV's which are intended not be cater for over pressurezation of cooling water due to pump issues but to avoid over pressurization due to leakage of high pressure fluid in cooling water circuit. Number of failures have been reported in Ammonia symposium volumes. Additionally, please do consider in design of cooling towers risers of appropriate sizes at the outlet

PSV 's installed on cooling water circuits , especially on exchanger shell are thermal PSV's

In addition please do consdier that cooling water exchangers can not operate at very high differential what I persume from your discussions and may lead to vortex shedding and flow induced vibrations. Therefore such an atempt to increase flow in an existing plant need to be evaluated very care fully


At the end I would simply comment that I have given my recommendation based on your discussions, howver, for specific information lot of date need to be considered, studied

Please feel free for any additional queries , you may have

Best regards and God bless you

[fallah]

It's worse than what you describe though. A PSV in liquid service usually needs an inlet pressure < 0.7 x set pressure to reseat.

Latexman,

In most cases it isn't economical to set/adjust the design pressure of downstream equipment as high as/according to pump shut off pressure, then we have to protect the equipment by a proper safety device. To prevent the problem of long blowdown due to using the relief valve in liquid service, a rupture disk or a proper "Modulating Pilot Operated Safety Valve" can be applied as a credible solution.

[kkala]

Usual local practice (in refineries / fertilizers) is to specify design pressure of downstream equipment equal to pump maximum shut off pressure, even if PSV is installed for other contingencies (e.g. thermal relief on heat exchangers). A liquid discharge from PSV is usually a nuisance (less for water, worse for hydrocarbons). However practices world wide may be different, as I can see.
I have never met rupture disk (RD) combined with PSV (being an applied practice) and even single RDs are seldom used here, probably because they need replacement after "bursting". How can RD + PSV shorten discharge time? Explanation would be appreciated. I understand RD purpose is often to protect from corrosion http://www.eng-tips.....cfm?qid=300956 '> http://www.eng-tips.....cfm?qid=300956 .

[fallah]

kkala,

I did propose modulating POSV because liquid discharge from modulating POSV isn't nuisance and normally having short blowdown, then could be applied in liquid services without usual problem due to applying conventional types.

Indeed i meant applying RD alone not in combination with a PSV. Besides applying in corrosion prevention, a RD alone might be applied against overpressure due to pressure spike because of its fast opening characteristics.

[latexman]

Naser,

By "most cases" are you including non-CW cases, or did you mean most CW cases? The reason I ask is I have rarely seen a lone RD in CW service on a HX. I think the times I have, there was a tube failure scenario, so quick opening was required.

The modulating POSV in CW service is good advice for this case though.

Edited by latexman, 24 January 2013 - 01:07 PM.

[kkala]

Thanks fallah, reaction of RD is indeed fast and could be an option for surge protection. Exchanger tube failure can also use RD (post no 10), due to transients developed (*).
However RD has not been seen here on cooling water network (equipment included). Even PSVs on cooling water or other liquids are avoided as protection against blocked outlet, unless existing (used) equipment of insufficient design pressure are placed in the relevant service.
Of course this is a local "habit" (simplifying operation), not a universal practice.
(*) See http://www.cheresour...ed-to-pump-rate '> http://www.cheresour...ed-to-pump-rate , mainly post no 8.

[latexman]

Naser,

On CW cases, I have to disagree. It is economical to set the design pressure of downstream equipment higher than pump shut off pressure, Of course, our opinions may be influenced by our past experiences in our respective businesses and local conditions. I'll briefly discribe mine.

Most, maybe all, of the CW systems I have been involved with have pressures in the 50-100 psig range. One could specify the MAWP of a HX lower than the CW, but HX fabricators tend to construct their HXs from commercially available materials. A lot of the HXs we use are made from 150 # class pipe and fittings. 150 # pipe and fittings are mass produced by many, many vendors and are very competitive in price. So, if we asked for an MAWP less than 50-100 psig range, we would still get a HX that is good for 150 psig, because the HX vendor can supply that cheaper than a custom fabricated HX with MAWP less than 50-100 psig. We tend to use 5/8-1" tubes, and they do not have to be very thick to be good for 150 psig. Plus, we do not use expensive exotic materials, just CS and 304/304L SS. So, we have found it to be economical to specify a 150 psig MAWP, since in reality that is what most of the HX is going to be good for anyway.

Edited by latexman, 25 January 2013 - 05:14 AM.

[Atttyub194]

Dear all

Good day

First of at , I would appreciate good comments and discussions, which definitely added value
Although the discussion is going on wrong direction so I thought to converge the topic, accordingly following may be considered

• All equipment at Fertilizer complex are designed at a pressure above than cooling water circuit
• All piping at cooling water system can not be designed at pressure rating below 125 of 150 # as no pressure class is available below these pressure rating classes
• All cooling water system need a corrosion allowance of 3mm approx, which if not utilized, eliminates the need of PSV's in the circuit
• Rupture disks are provided on high pressure gas exchanger and these are meant for avoiding failure in case of tube rupture . PSV in the case chatters and damage till the plant is safely shutdown
• PSV provided on demin heater or lube oil coolers cater for thermal conditions only
• In cooling water circuit , you can not balance the flow in each loop, so there are always chances that few exchangers get more cooling water and few get less. This is the main reason that cooling water balance or adjustment is required
• Shut off pressure although matters but most pump manufacturer limit the rise not more than 3-5 % in these cases
• When ever you thought to increase the flow in an existing plant , although you may be operating at pressure below design pressure , even then the flow through few exchangers are more as the flow corresponds to difference is supply and discharge header pressure
• It is not always economical to change even a segment of underground piping or heat exchanger if they are capable to perform duty
• one wise decision is to change fill to get a better approach in cooling tower

Best regards and God Bless you

[kkala]

1. Cooling water data for design use in three local refinery cases is as follows (pressures at grade).
Design pressure, kgf/cm2 g : 8.0 / 12.0 / 7.0
Supply pressure, kgf/cm2 g : 4.5 / 7.0 / 6.0 (evidently minimum)
Return pressure, kgf/cm2 g : 2.5 / 4.0 / 4.0 (evidently minimum)
which more or less complies with data mentioned in post no 12 (by latexman).
2. Post no 13 (by Attyub194) contains useful info on the specific plant; point of para 7 needs clarification, seeing that shutoff pressure is about 1.2×pump head+suction pressure. The latter is higher than 3-5% of discharge pressure in case of cooling tower pump (low suction pressure). Or perhaps meaning is something else? Clarification welcomed.

[fallah]

By "most cases" are you including non-CW cases, or did you mean most CW cases? The reason I ask is I have rarely seen a lone RD in CW service on a HX. I think the times I have, there was a tube failure scenario, so quick opening was required.

Latexman,

I did mean in general not CW cases. Also i am agree with you about using a RD in tube rupture scenario.

Edited by fallah, 25 January 2013 - 02:42 PM.

[latexman]

Okay, I understand the difference now. I was thinking CW only.

[Atttyub194]

Dear Kalkaa

 

Good day!

 

I am not talking of process side of the equipment. It is particular to cooling water system only

 

In  all API pumps, your comment is absolutely valid. However, there is no fun in using Colling water pump on API standards.

 

Gould's and Sulzer are the pumps normally used for Cooling water service with flow capacity of ( 6000 to 10,000 m3/hr per pump) for cooling water service . The manufacturer recommend and can control head by impeller design

 

Could you pleased check the dynamic head of pumps used in refinery , it would be definately in the range of 30 to 45 m with suction at atmosphere or from cooling tower basin.

 

The piping and rest of the circuit with stand as high as pressure as indicated.

 

Best regards,

[kkala]

Good day Attyub194, I understand your post no 17 concerns pump shutoff pressure touched in post No 14. For the refinery pumps specified by us (Process), shut of head = 1.2×differential head at the operating point (not far from best efficiency point). The factor can be a bit higher (say 1.25) and we require clarification from pump suppliers.

Similarly for a cooling water pump located at tower basin (suction pressure close to 0 Barg), shutoff pressure would be expected  1.2×differential pressure at operating point, more or less. Is it so? Or mentioned Goulds / Sulzer pumps (6000-10000 m3/h) have shutoff pressure = 1.03-1.05×differential pressure?  Can it be possible? Cooling water pumps have not been specified (by me), on the other hand recirculation pumps can have peculiar curves <http://aec-central-c.../Pump Tanks.pdf>.

I have not found refinery cooling water pump heads, but data of post no 14 indicate values higher than 45/70/60 m and lower that 80/120/70 m respectively.

[Atttyub194]

Dear kalka

 

Good you

 

Sorry for delayed response. We were in Turnaround

 

Please try to use Sulzer / Goulds online tools. You will get all options for limiting shut off head

However, if you use standard API pump the difference in shut off and normal would be the same as indicated by you

 I would suggest the forum to let come back on the topic and ask our friend Bill B for any further discussion required and whether he wants any additional info and what is the progress

 I would suggest to make the discussion fruit full, we should share the final conculsions after implementation of solution

 

Best regards

[kkala]

Good day Attyub 194.

So big cooling water pumps (Sulzer, Goulds, etc) can have shut off head significantly less than 1.2 x differential pump head, or at least "options" to realize it? What these options could be, probably mechanical recirculation valve? Or the pump curve really results in such a low shutoff head?

Yes, contribution of Bill B would be welcomed.