Simatic 545 Pid

maintpro

Member
Join Date
May 2003
Location
Osceola, IN
Posts
28
I am having problems with the PID operations of my SIMATIC 545 PLC. We use the controller in a wax melting operation. The wax kettles are heated by steam and the PID is adjusting the valve position. The valve position range is -25% to 100%.
The problem is that when the PV is the same or greater than the SP, the the valves are occasionally still open 50% or sometimes greater (I say occasionally because it is not every time). This leads to overheating of the kettles and burning of the wax.
If we turn the steam off (this takes the loop out of auto) and then back on again, the output will correct itself. We have had this system in place for a number of years now and according to the operators, it has never worked properly to begin with.

My loop perameters are as follows:
POS/VEL PID ALGORITHM: POS
LOOP VFLAG ADDRESS: C702
SAMPLE RATE: 2.00 SEC
PROCESS VARIABLE ADDRESS: WX177
PV RANGE LOW: 0
HIGH: 300
PV IS BIPOLAR: NO
SQUARE ROOT OF PV: NO
20% OFFSET ON PV: YES
RAMP/SOAK PROGRAMMED: NO
RAMP/SOAK FOR SP: NO
ALARM DEADBAND: 0
MONITOR LOW-LOW/HI-HI: NO
MONITOR LOW/HI: NO
PV ALARMS LOW-LOW: 0
LOW: 0
HIGH: 100
HIGH-HIGH: 100
REMOTE SETPOINT: NONE
CLAMP SP LIMITS LOW: O
HIGH: 300
LOOP GAIN: 50
RESET: 35
RATE: 0
FREEZE BIAS: NO
DERIVATIVE GAIN LIMITING: NO
LIMITING COEFFICIENT: 10
SPECIAL CALCULATION ON: NONE
SPECIAL FUNCTION: NONE
LOCK SETPOINT: NO
LOCK AUTO/MANUAL: NO
LOCK CASCADE: NO
ERROR OPERATION: NONE
REVERSE ACTING: NO
MONIOR DEVIATION: NO
DEVIATION ALARM YELLOW: 100
ORANGE: 100
MONITOR RATE OF CHANGE: NO
RATE OF CHANGE ALARM: 0
MONITOR BROKEN TRANSMITTER: NO

This is the first time that I have had to deal with a PID system and am uncertain if these perameters, especially the gain, are correct.
I am also wondering if the problems are due to fluctuations in the process variable and if I should freeze this value for loop use by moving it into a V-register.

I have monitored the program and found that temperature values received by the plc are correct.

Any help you can provide would be greatly apprecieated.
 
You say that the control valve is open when PV=setpoint.
Is the same PID loop used for cooling also? Do you have a split range controller? Can you put the loop in manual and set the output to zero, what is the valve position? I think the I/P you are using and the control valve may not be set up for the same control air pressure ranges. Also your PID gain may be sort of aggressive for a temperature loop.
The first thing to check is if the PLC can fully open and close the control valve. If it can then you need to tune your loop.
 
The PID loop is used for heating only. I can manually set the outputs to any setting and the valve will track to there. Closed= -25% and 0psi on the valve and the valve is closed. 0%= approx 6psi and the valve is slightly open. Full open= 100% and the valve is fully open.
The PLC is capable of fully opening and closing the valves but is responding way to slowly in doing so. I say that the valves are still open when PV=SP and they are. However, the loop eventually corrects itself and will adjust the outputs accordingly but sometimes not until the kettles have become too hot. This time frame varys from a couple of minutes to over an hour.
 
Had same problems yesterday. Is it possible that since I am using the direct analog input and that the reading is not stable that this is delaying my loop reaction.
 
I don't know your process, but it sounds like you are having an issue with integral-term windup. That is, a large integral error is built up during the heating phase, and it will take some time for the integral error to go away.

I'd start by seeing what effect modifying the RESET (integral) term has. You might also find out that you need some derivative (RATE), which would be used to limit the rate of change of the temperature.
 
Hi

The integral wind-up issue referred to by rdrast is controlled in TI terms by the setting of the "Freeze Bias" parameter.

Whenever the loop output saturates (i.e. reaches fully on or fully off) you can't increase (or decrease, respectively) the output any further. As a result, any further integral calculation all gets added in to the loop bias term. Eventually when things begin to come back within the controllable zone of the loop, there is all this bias which needs to be reduced by successive integral calculations again. And this can all take time. The "Freeze Bias" option tells the loop controller what to do when the loop is saturated. If you have it as "No", then it just keeps adding to the bias as I have described. If you have it as "Yes", then the bias term is frozen at whatever value it had when the loop output saturated. No further change is made to the bias until it comes back in to the control zone again. Significantly, if rdrast's diagnosis is correct, yours is set as "No".

I'd rather change something like this (it's an either/or change and quick to change back) rather than try changing one tuning term without taking care of the others as well. Whether the tuning terms are correct or not is impossible to tell without being there. Don't start getting suspicious of a high figure or a low figure just because it's high or low. Having said that, if the system has never worked, it may just be worth starting from a clean page and trying to re-commission the loop from scratch. Throw away any preconceptions about what the 'right' numbers are, and get the process dynamics to tell you how they need to be controlled to you satisfaction. If the art of loop tuning itself is the problem, then there have been many, many threads in this forum on that very subject. Some excellent advice has been written over the years - I've always found Peter Nachtwey's contributions on PID to be first class. Search for him and PID.

Regards

Ken
 

Similar Topics

What software and cable are used to program a Simatic 545? We have 3 on winders and need to make a small program change for a new product line...
Replies
5
Views
1,790
Hi - I am looking for contractors in the western US experienced in TI / Simatic 545 programming. This would be small project type work -...
Replies
2
Views
1,910
Hi: I have a CTI-2552-A card which is "isolated rtd input". The card replaced a 505-7038. The temperature values are reading too high, on all...
Replies
0
Views
1,290
I had a problem on a plc I am not versed on. a 545 could not get it into run and could not connect... we replaced the card dumped the program in...
Replies
2
Views
3,103
I am currently trying to convert a cti2500 program to a 545-1106 and everytime I transfer online I get a failed to write I/O configuration...
Replies
5
Views
1,965
Back
Top Bottom