Highly purified water(HPW) distribution system

[quark]

Dear fellows!

I would appreciate any comments on the scheme I am proposing for the subject matter. There are many issues with the entire distribution system but, at the moment, we are attacking the main problems.

The System
There is the 5kL main storage tank and a pump(on second floor) with cold standby arrangement. The pump discharge splits into two streams (this is the first problem). The discharge piping is of 50OD upto 24.88 meters, i.e upto the branch split. The first distribution line goes to the ground floor and returns back to the tank. There are 7 user points with zero dead leg valves. The tube size is 38OD and the ZDL valves are of 38OD size. The total length of this loop is 216 meters and there are 70 bends.

The second loop runs to the first floor and returns back to the tank. The length is 269 meters, 73 bends and 11 user points with ZDL valves.

There is no specific schedule as to the usage of water at various user points.

Requirement
The user points are mostly for agitated vessels and there are few wash points. The tanks are fitted with spray balls that require about 1 to 2 barg pressure. The return line velocity must be around 0.5 m/s. This corresponds to about 1.75cu.mtr/hr. The consumption is 3cu.mtr/hr/loop, at an average. The first user point is about 50 meters from the pump discharge and the last point is about 220 meters.

Proposed Scheme

As the system has many combinations of usage, I have considered designing the system with two boundary conditions. First one is that, entire 3cu.mtr/hr will be consumed at the first user point of each loop and this gives me the lowest pressure drop in the system and thus head requirement. Second one is that, entire 3cu.mtr/hr will be consumed at the last point of each loop and this gives me the highest pressure drop in the system. Now the selected pump may operate anywhere between these two points.

Each line will be fitted with a back pressure regulating valve(before the return line tank entry) and this takes care of spray ball operation apart from complying the regulatory guidelines.

I am planning to install a pressure transmitter just before one of the BPRVs (the loop with maximum pressure drop) which in turn controls the speed of the pump. The main aim of going for a variable speed system is not to save energy but to provide return line flow in all conditions and also to ensure effective cleaning of the tanks.

I welcome any better method.

PS: The best condition gives me a total head of 20m (including that of BPRV) and 50m in worst condition. I have used methods of both Hooper and Darby for PD calculation and they concur to a great extent.

Regards,

 

[BigInch]

Well, without doing a simulation on this, I'd probably start like this,

Set the minimum pressures you need at both branch outlets to the tank when each branch is flowing maximum with maximum head drop. Install a pressure transmitter at both branch outlets to the tank.

Set the pump rpm to vary to and program its controller so as to compare and maintain the highest pressure of either of those two outlet pressure setpoints at any given time.

You could even turn down the pressure requirements at both branches during the night shift or something, if you wanted to.

http://virtualpipeline.spaces.msn.com

 

[BigInch]

BTW Your PT is redundant for a VSD, since you need'nt control the rpm from that value, as that point (tank inlet pressure) can very easily be made a constant pressure point (minimum tank inlet pressure) with value also equal to the discharge pressure from the long loop flow at last station case and the system curve calculated from there.

Without a variable feedback to pump rpm, this sys should have a constant speed pump discharge head that yields the required flowing pressure for the case with flow at the first station of the short loop. Calculate upstream from some minimum tank inlet pressure to get whatever pressure you need at the branch for the worse case flow condition (presumedly the highest required head at the branch, presumedly first station flow on the long loop).

http://virtualpipeline.spaces.msn.com

 

[HEC]

It sounds like you have either a purified water system, a water for injection system, or a sterile water system, for each case you need to maintain minimum flow velocity in the system to ensure purity is not compromised. With this as the stated aim you would need to monitor the return flows in both return lines to maintain this minimum. The flow meters would be used to control the pump speed and teh back pressure valves utilised to control pressure to the sprayballs. If the spray ball valves are automated the back pressure control can be active only when required and fully open to reduce pump power consumption when the system is "idle".

Mark Hutton

 

[quark]

First of all thanks for the replies.

There is some problem with keeping the pressure constant. The line is with TC couplings and I don't want to keep the line at higher pressures because I hate to attend the leakage on top of a false ceiling. I have worst experiences when the pressure goes beyond 5 barg. Secondly, this is not a constant volume flow requirement system. 3cu.mtr/hr is the maximum flowrate and only minimum flow requirement is to be maintained in the return lines (i.e no user requirement).

So, the option will be to activate an automatic back pressure valve by a pressure signal, if I have to maintain constant header pressure. But I want to grab that little energy and save it and rather go for a manual BPRV that will be set at minimum flow requirement + spray ball working pressure.

Flowmeters can be an option but installing two flow meters will be much costlier business as my only option is straight tube coriolis meters (vortexes were used in the past but are fast becoming obsolete due to reservations of regulatory authorities)

As the manufacturer suggested to use the pump on trial basis, I will have some data before finalizing the system. I will keep you posted.

 

[EdStainless]

Double check and make sure that you are not too far off of BEP at these different conditions.

What on earth do you have in the system that could leak at 75psi? These systems should handle three or four times that.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[BigInch]

Holes?

http://virtualpipeline.spaces.msn.com

 

[quark]

The fluid in the pipeline is highly purified water (approximately equals cold WFI). It is not that water starts jetting from all points when the pressure goes up. When it starts leaking, you will know it only when the water pools on the false ceiling and starts dripping from light and HVAC boxes or on the walls. By this time you have enough damage.

I am not knowledgeable about the triclover clamps that can withstand 15 to 20bar pressure. Can you give me some links or details about these?

PS: The system also needs weekly steam sanitization.

 

[pleckner]

@quark, I have experience in the bio-pharm area of process design. The tri-clover fittings should be good for up to 150 psig. Although there are clients of ours that want to maintain high velocities in the piping, studies have shown that flows equivalent to Reynolds numbers around 30,000 should be adequate to keep high purity water clean. On the surface, I don't see any particular problem with what you propose just as long as you can maintain that minimum 30,000 Reynolds number at all times.

 

[quark]

pleckner,

Thanks for the suggestion. I came across a paper in Pharmaceutical Engineering, about 8 years back (roughly), that deals with pipe velocity and corresponding shear forces that can take care of biofilms on the pipe internals. Somehow I lost it. Do you have any references to this or any other relevant material?

Regards,

 

[EdStainless]

Oh, you are worried about 'slop' clamps leaking. If your flanges, clamps and gaskets are all from the same supplier then you should be OK. If you play mix and match then you could be in trouble. Not everyone uses the same angles, thicknesses, or gasket dimensions. Unless you have your own internal dimensional standard I would be careful.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[pleckner]

Pardon my ignorance but what is a "slop" clamp?

A Tri-clover clamp is a pretty standard way of connecting two pipes to each other, or a pipe to equipment, fitting or valve, in the Bio-Pharmecutical industry. It replaces flanges. The gaskets are basically O-rings. The connections are such that no material can easily get trapped between them, creating a spot for the production of microbes, and are more easily and quickly removable. There is no problem with interchanging brands of gasket material, only the type of gasket material used.

@quark: The criteria of 30,000 Reynolds number was given to me by a colleague. I'll have to see in my bag of information (at home) what I have to and can offer.

 

[quark]

Ed,

The dimensional specifications for these fittings are well defined and manufacturers generally meet these requirements.

pleckner,

I got the references. Let me know if you want to have a look into them.

 

[HEC]

quark
I have had some experience in Pharma, a few years old now, tempted into the resources sector! I'm interested to hear that vortex meters are being frowned upon by authorities, why is this. Particularly if the system is being steam sterilised weekly. Ultrasonic flow measurement may be a viable alternative, though high temperature sensors will be required.
To maintain the back pressure, why not use a sprayball on each return pipe into the storage tank. If the minimum flow is matched to the required system back pressure the system will be self balancing.
In both food and pharma I have used minimum line velocity of 1.5m/sec rather than Re number.
Regards

Mark Hutton

 

[quark]

Mark,

The reason is as silly as the shedder bar protruding inside. (PS:I, long back, stopped arguing with pharmacists and regulatory guys).

You can't validate sterilization of pipeline so we call it as sanitization (or sanitation, at some places).

You are right that Reynolds Number may mislead you, particularly, at higher pipe sizes. 1.5m/s is the general guideline and we take it to be atleast 1m/s in the return line after the last user point.

However, the references I cited above gives the following velocities at 25⁰C.

1/2" - 0.5fps
3/4" - 1.3fps
1" - 1.9fps
1.5" - 2.9fps
2" - 4fps
2.5" - 5.1fps
3" - 6.2fps
4" - 8.2fps
6" - 12.5fps.

These are for standard sanitary OD tubing and the velocity is based upon shear stress greater than 130N/m² which ensures minimum adhesion of biofilms.

There are spray balls in the storage tank but you can't ensure a constant back pressure with them. At the most they will not rotate or may not spray properly if the flow is lowered. BPRVs are spring loaded and thus take care of return line pressure.

 

[EdStainless]

quark and pleck,
I have a small collection of flanges, clamps and gaskets on my desk from four or five sources. If I play mix an match with them some combinations will not seal. Some leave significant voids at the joint. And some are so tight that the gasket protrusion tears the gasket.
I know that there are standards (EHEDG and BPE), but many examples that I have seen don't conform with the nominal requirements close enough to yield the desired results.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[pleckner]

@quark:

I would be interested in seeing what your references say.

Thanks

 

[kenvlach]

quark,
Don't know your requirements, but the bio-pharmacological & semiconductor processing industries are pretty rigorous and both use a lot of electropolished 316L tubing & fitting. Easier to keep clean & re-passivate, re-sterilize, etc., than non-EPed material.
“STAINLESS STEEL TUBING IN THE BIOTECHNOLOGY INDUSTRY” gives many standards.

http://www.asepco.com/docs/PDF_files/Stainless_tubing.pdf

Valex in Ventura, California is a big supplier of EP-ed SS. Some details:

http://www.valex.com/html/2_con.asp?page=specs

The Semiconductor Equipment and Materials International (SEMI) people have some (expensive) standards of possible interest:

http://downloads.semi.org/pubs/SEMI...878f08055d4a230c88256a16007e9fb2!OpenDocument

To help set up in-situ cleaning & passivation processes:

http://www.astropak.com/o_high_purity_cleaning.php

 

[quark]

Ken,

Thanks. The first link is very good and gives me consolidated information in one article. So far, I have it in bits and pieces.