Bleed and feed for a water chilled system

[engineers_eyes_and_ears]

I am working on a water chilled hydronic system.

The system is pretty simple and consist of a pump, heat exchanger, pressure drop component and an expansion tank (see attached sketch).
In order to ensure the correct conductivity, the system is equipped with a bleed and feed system consisting of (actuated) valve for filling water and a (actuated) valve for bleeding.

The pump is temperature controlled at the suction side. Leakage detection is very important, so I have added a pressure transmitter at the point of zero pressure change. In case of leakage I should loose pressure in the system and hence detect leakage (The static filling pressure (pump suction side) is 2.8 bar and the pressure after the pump is 14 bar).

Now, I am a bit unsure on how to make the control for the bleed and feed system. More specific, if we assume that the conductivity in the system increases, then we must bleed the system and fill new water in. Hence we can tell the bleed valve to open. However, this would result in a pressure loss which the system would register as a leakage. So that's an issue?

Another method could be to add a level transmitter to the expansion tank so when we bleed, the level falls and then we feed to achieve normal level again. However, the expansion tank cannot be fitted for a level transmitter.

Any ideas?

 

[LittleInch]

If you want to match volume out ot volume in then maybe hook up two piston type pumps or other PD pumps on the same shaft and add a small motor to pump out the same volume going in.

But you are still likely to get transients on pressure during this operation.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[btrueblood]

Assuming you are trying to bleed air from the system, I would advise locating the bleed valve at the highest point in the plumbing.
For filling, characterise the expansion tank's pressure vs. volume and use that information to guide filling. Or (assuming the backside of the expansion tank is an air/nitrogen fill nipple) add a pressure sensor on the air/nitrogen side so you can calculate the volume directly from the charge pressure.

Usually, bleeding is done after installation or maintenance, and done manually/visually (when the bubbles stop coming out, she's full)...

 

[TBP]

This seems like an usual set-up. I've seen automatic blowdown systems on steam boilers and cooling towers, because there's a degree of on-going make-up water. I'm not sure that I've seen a closed-loop system with an automatic bleed system. What would cause the conductivity rise in the system, once it is filled and operating correctly?

 

[LittleInch]

I think the best description would be drain and fill, not bleed and feed....

Also with your expansion tank, you're going to find it difficult to find small leaks as the expansion tank will compensate for the leakage with only a very small pressure drop which will be difficult to see amid the noise of a pumped system.

Other option is to bleed off water down to the min pressure in the expansion tank, pump water in back to some maximum level and then bleed some off, rinse and repeat for as many time as you need to.

You would lose your leak detection system whilst doing this and be limited to the working volume of the expansion tank.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[TBP]

You're not really supposed to double post questions, but I see that you're new here, Eyes & Ears. I think you'll get way more responses if you posted this in the HVAC/R forum.

 

[georgeverghese]

In a closed system like this, the only way to pick up leaks is by loss of level detection at the expansion drum.
Since the drum is gas pressurised, suction and discharge pressure at the recirc. pump will always be maintained even when there is a large leak.
When you have loss in conductivity, activate both drain and fill valves simultaneously through a conductivity gap controller. Fill and drain rates may be adjusted at the field (with local flowmeters) to check flowrates to make fill and drain rates the same. Make some control adjustments to allow for independent operation of either fill or drain valves to enable make up alone or drain alone if there is a long term change in expansion drum level.
Am assuming that there is no mixing of the recirc water with the water at the dead end expansion drum. Where is the conductivity probe located?

 

[crshears]

I concur; bleed-and-feed is eminently suited for systems that are operated without expansion tanks; we used the scheme to control the system pressures in a high-pressure high-temperature water circulating loop used to simulate the physical conditions within nuclear reactors [ no radioactivity ] . . . but with expansion tanks, there is no ready index to extrapolate system leakage, bar using an accumulator with floating "puck" - and those aren't all that common, in my experience.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Rputvin]

A little late to this one, but I deal with low-conductivity and hydronic loops. Typically not together, but the two are not mutually exclusive.

Most of the low-conductivity systems are non-ferrous, or at least as close as possible. Stainless being preferred over copper and yellow metals. Cast/ductile iron and black/steel pipe and fittings leech ions to increase conductivity. You'll want to use the correct corrosion inhibitor as nitrate-based and molybdate-based inhibitors impact conductivity differently (I believe the molybdate is the one you want for LC, but it can be purchased packaged in low-conductivity inhibited glycol as a specific heat transfer fluid). I would not recommend straight city water, but it just depends on how often you want to clean the system and what the costs for the chemical supply and recovery/disposal are.

A lot of the hydronic specialty devices are not available in non-ferrous versions, which is one reason we don't see the two blended very often. Some of the bladder/diaphragm tanks are non-ferrous wetted parts, and the smaller sizes for air separation come in brass (under 2.5")

For hydronic systems I've been adding a pressure switch to the suction header to detect low-fill condition as soon as the system drops below the pre-charge pressure when it's a system without an active makeup source. As mentioned, this is more reactionary for service/maintenance and determining the exact volume change is extremely difficult. But since hydronic systems operate on pressure differential more so than straight system volume, pressure control can be utilized fairly successfully to get pretty close to an exact volume swap, save for any temperature drift that would impact overall volume. But you'll have enough turnover that the exact system volume, which is already +/- you tank acceptance, won't really care that much if you're off by a fraction of that acceptance volume. Worst case you'll need to adjust volume slightly, but high and low-fill triggers and a safety relief are all that's needed to manage it.

There are several ways to accomplish the controls without impacting the operation of the system, but each approach would depend on what you're using for makeup water, what controls hardware is available (relay logic vs PLC), if the system needs to be continuously running, what I/O or sensors are available or can be added, what your control points and operating points look like, etc.

It could also be a pain getting a conductivity sensor for this application. The normal ones packaged for this type of management on a cooling tower, like an Advantage Controls bundle, are low-pressure devices and get installed into clear PVC type piping. The process sensors that carry pressure ratings are orders of magnitude more expensive and are not flow-through sensors. Your service pressure, flow, pipe size will all impact that sensor selection.