Control Oscillation 2

[Herman2]

I have two(2) 1000-ton centrifugal chillers and four(4)variable-speed pumps sending 39 deg F chilled water to the primary side of eight(8) plate and frame heat exchangers. My customer circulates water on the other (secondary) side of the heat exchangers. My obligation is to hold 42 degress F on the secondary outlet water.

I also have control valves I can modulate on the chilled water supply to the heat exchangers.

Every control scheme we try oscillates. We get a periodic swing of several degrees (period = 3 minutes) when the thermal load at the heat exchanger is steady.

Any suggestions on how to control this beast? Thanks.

 

[quark]

3⁰F temp. difference is pretty tight between the two streams of the heat exchanger. A generous design could be 5⁰F . Anyhow, in what temp. range the chiller is being operated? What are the flow rates of chilled water and water? Is the secondary side water also recirculating?

Do you have any control valve on secondary side?

The above data will be helpful in analyzing the problem.

Regards,

 

[Herman2]

Thanks for your interest... Here are more details...

We usually run 3 of 4 chillers (in parallel). The outlet temperature is 39 to 40 deg F. The total volumetric flow rate is about 10,000 gpm.

The secondary water flows at a fairly constant rate determined and controlled by our customer. We have data from the secondary side (temperatures and flow rate) but no control of any kind.

The whole system oscillates (primary flow, chiller load, etc.) We need some ideas on how to control this system.

Cheers.

 

[CRG]

A factor that may be causing the problem is your customer. It may not be possible to use controls at your end to fix the problem if your customer’s flow/temperature is oscillating on too steep of a gradient. Carefully look at the rate of temperature and flow gradient in your customer’s return water to see if this is the problem. If the gradient is excessive, correct customer’s oscillating problems before trying to fine tune your controls.

 

[lilliput1]

Tell us the control sequences. Oscillation can be because of controls fighting each other or because the system readout of sensors do not occur often enough to produce smooth action.

 

[Herman2]

Actually, the customer side is quite stable. The flow is steady and the entering water temperature only varies a few degrees. Somehow, we take that situation and make the secondary outlet temperature oscillate.

Control scheme...

We modulate a water valve on our side of the heat exchanger trying to hold the customer's outlet temperature at 42 deg F. We also control speed on the (primary) chilled water pumps. They modulate speed to keep the control valves between 60 and 70% open. There is a very long line from the coolers/pumps to the heat exchangers. It takes over 6 minutes for the water to make a round trip.

 

[ChasBean1]

Here's a two part answer:

1. IF THE HEAT EXCHANGERS ARE YOUR ONLY LOADS - Set your chillers (either two as per your first post or four as per your second) to a 42°F supply temperature. Remove your eight plate and frame heat exchangers, associated control valves, and all extra piping, fittings, and gauges and sell them on Ebay. (Note: I want 50% of the profit on this and 50% of what you save in energy cost over the next year). Control the pumps to maintain a fixed CHW loop differential pressure, supplying the loads directly. Stage the chillers up as necessary to maintain 42°F. Stage them back down based on % load (or amperage).

2. IF YOU HAVE TO MAINTAIN THIS CONFIGURATION - You have a tail wagging the dog thing going on here. You said the pumps "modulate speed to keep the control valves between 60 and 70% open." The control valves in this application need to modulate solely to maintain the secondary side water supply temperature. Trying to maintain them between 60-70% defeats the purpose of having them, and it is therefore the pump speed that is trying to maintain secondary loop temperature, which is a horrific way of trying to control it, especially with a three minute lag between pump modulation and control effect. The three minutes will also become six minutes when the pump speed is half or 1½ minutes when the pump speed is doubled. With this configuration, you need to make sure you constantly have 39°F water near the HX (primary inlet). The use of 3-way control valves at the HX's can help with this. The control valve should respond only to the secondary loop temperature (temp. drops, more primary water enters the HX; temp rises, the 3-way valve bypasses the HX). The CHW pump speed should only control to maintain loop D/P, not valve position or temperature.

 

[gnordo]

I suspect the heat exchangers are a means to seperate the cooling systems primary coolant from that of the 2ndary fluids.

Without a need to maintain seperate fluid standards, the heat exhangers would be a needless capital cost and operting expense. The heat exchanger would add maintenance problems, and also reduce the systems overall efficiency due to increased pumping energy as well as requiring the chillers to operate at a lower discharge temperature.

The need for this isolation while unusual, is not uncommon.
As an example, lets say the 2ndary loops require high purity deionized water that directly cool high powered computers such as Cray computers which rely on direct fluid cooling being circulated through them. By this isolation, the primary system doe not have to be designed to cope/maintain the DI water disipline/materials.

The above is one example of a need to maintain a seperation.

Gnordo

 

[zakh]

As Chas has suggested, you should control the pump speed using a DP sensor...similar to a variable pumping and VAV AC system, while maintaining the 2-way control valves to attain operational cost savings..however, it may not be much as you stated that the entering temperature does not vary much.

Herman...you stated that the pumps are controlled based on the opening of the control valves (60%-70%). Does this 60%-70% map to 0%-100% pump speed?

Could this problem be due to high gain settings on the control loops?

 

[Herman2]

The target valve position range does not map to 0 - 100% pump speed. The customer just thought 60 - 70% was a nice "sweet-spot" for valve operation.

I have implemented several of the group's suggestions and we do not see oscillations. We are now resetting pump speed periodically instead of letting the speed controls chase the same temperature setpoint as the valve controls.

I really liked the idea of selling the heat exchangers on e-bay, but they are required for non-engineering reasons. I never did see much use for a water-to-water heat exchanger at that point in the system.

Thanks...