Control question

[Vidmas30]

Good day,

 We have a heating system (closed loop) with boiler and circulation pump with frequency converter (on return). The task is to maintain constant t-re (on return) while keeping pressure in users desirable limits. Maybe someone had similar task and could share his oppinion about control method?  Thank you in advance!

[Walkerok]

The level of inforamtion given is not enough to give an actual answer.  Are you controlling between responding fast enough when someone creates a momentary open loop demand (faucet or shower) and the rest of the time where it is a closed loop system and you just want to circulate water so that hot water is avaialble at the end point quickly when required.  If this is the case then a PID control setup will probably never work as you want because you will always have long delay periods between usages and there is no consumption where the compounding I and/or D values will get you into trouble.  You can force your system to not allow I or D to contribute more than say 20% influence over the PID and that will make it better, but it is still very hard to control. 

 

If I have guessed correctly in your application then I would sugest to create a signal following step type routine instead of a PID.  While this is more up front work and I am sure that someone out there can make a much more elegant answer this will work.  You will want to put some time delay on each step so that electrical noise does not cause your VFD to shimmy shimmy shake eternally. (say 2 seconds for a pressure to be held before taking action)  For example:

 

1) Usable VFD speed range is from 20 - 60Hz so your effective range in 4-20ma is 8.8 - 20ma or 2252 - 4095 output from your PLC to your VFD.

2) you want to hold 80PSI on the line

 

If your set point is 80 PSI, you will make a routine that drives your analog output to your VFD based on striclt how far away you are from that pressure point.

 

 

 

set a range from 78 - 82 PSI to control your VFD

 

PSI     Output

82       2252

81.9    2298

81.8    2344

81.7    2390

.

.

.

.

.

.

.

.

78.2   4003

78.1   4049

78      4095

 

 

Like I said before there are more elegant ways but this should do what you need.

 

Good luck

Keith

[Vidmas30]

Thank you, Keith,

 It is closed loop ( see attached drawing ) heating system. Now, as our customer wanted, we control pump (with VFD) to keep constant pressure on return (user chooses to keep constant set manually or variable setpoint according outside temperature). It is not problem with pressure, system works fine. But now customer want to keep constant temperature on return while keeping pressure in desirable limits (set manually i.e. MIN - 3bar MAX - 5bar). Still can't figure it out how to do it better..

[Walkerok]

Since you have two sets of simultaneous contraints the problem becomes one of when one set of conditions is violated which condition is most important pressure or temperature.  I can easily see a situation where you will not be able to provide a fast enough flow to keep the temperature stable. 

 

I have one question though.  If this is a closed loop hydronic heating application what pressure are you making since the water is not actually being consumed you are simple recirculating the water in an infinite loop.  It seems the question is how fast do you have to circulate the water to get the heat transfer you want at the end point in at the same time keep the delta T on your boiler to a required amoount (many times 20 degrees so that you do not create condensation on the tubes of the boilers).  The faster you move the water through the heating system the more stable your temperature will be, but at the same time the faster you move the water through the more erosion you create on bends, fittings, and valves.  It is genrally acceptable to say that you move the fluid through the papes at a rate that returns fluid back to the boiler at 20 degrees cooler than the outgoing water of the boiler andf that if you attempt to close this delta T to less than 20 degrees then you are generally approaching the max flow rate through the pipes you should allow.

 

Hydronic heating is a vastly complicated issue for what on the surface looks like a fairly simple problem.

 

If you only have one point of heat load then it should be possible to make a flow rate that gives the best chance of being stable at the location that you care about.  If you input the temperature data for the load point then when temperature drops you speed up flow, if the temperature is too hot slow down flow.

 

I think I am still not understanding what you are exactly doing and I am given not so helpful suggestions.

 

Sorry

Keith

[Simon]

It looks to me like this:

 

* you have two variables, temperature and pressure.

* I assume the boiler is on/off control, not proportional?  The boiler control is potentially a 3rd variable.

* in the first case you controlled pressure to a setpoint, and allowed the temperature to do whatever.

* in the second case you want to control temperature and have some constraints on pressure.

 

So in the second case you may be looking at a cascade control loop.  To do that, leave your existing pressure loop unchanged.  Then create a second PID loop for temperature.  The CV of the temperature loop becomes the SP to the pressure loop.  So if temp drops the Temp loop issues a command to the pressure loop for more pressure (flow).  You will need to set the constraints of 3-5bar as min and max limits in your pressure controller.  You will also need to suspend any integral action in the Temp loop if the pressure loop hits either the min or max limit.  Otherwise you will get integrator windup in the Temp loop.  You should also set the update time/sample rate on the Temp loop to be 10 times slower than the Pressure loop.

 

You may be able to approximate some of this, for example by using a single PID controller with temperature as the PV and pressure/flow as the CV.  But without knowing your system I can't way whether this will be stable or not.

 

I hope these comments are helpful, I can't go much further into the details, as I don't do this type of control everyday (it was something I did in a previous "life"), and don't have the time.

[Vidmas30]

Many thanks, Walkerok and Simon,

  I had some thoughts about pressure control (because it's main variable) and correcting it in some limits according to temperature. Simon's comments really helped to sort things Thank you!