Heating water in fixed time, need help

I have to resolve one task. There is one water tank, where is needed to heat water in fixed time (for exemple in tank water have 20 degrees have to heat to 70 in 20 minutes). Heating is realized by steam, through analog control valve (4..20mA). Do you have any ideas how to do it? I want to use one controller of simatic S7-300 series.

First thing you have to do It's to be sure that with the valve totally opened you will be able to heat you water in less than 20 minutes. If that is possible, then you can try using an ON/OFF approach or PID.

with totally opened valve the water will heating much faster than 20 min, is needed linear increase of temperature

idell said:

with totally opened valve the water will heating much faster than 20 min, is needed linear increase of temperature

Click to expand...

You can use PID option, but you have to change SP according to your needs, for instance:
SP is 30 degrees first 5 minutes
SP is 40 degrees between 5 and 10 minutes and so on.

thank you widelto for answer, what do you think about my idea: to calculate speed of increase temperature (with cyclic interrupt), and configure one PID loop with PV - speed of increase, SP needed speed (for our exemple ((70-20)/60)*20=16.6 degrees per minute modified (70-20)/20 ), do anyone try this?

yes, you can use 2.5 degrees per minute ( 20 min *2.5 deg/min = 70 deg - 20 deg ), try to use this value to change your SP. It is posible that you have to go faster if you need it. Try changing SP every four or five minutes.

I would (and have) change the SP every minute. This will result in a steadier increase in temperature.

Since temperature will lag the SP, I would hit the endpoint a couple of minutes before the actual time to allow the temperature to settle. The other reason for this is that as temperature increases, there will be less heat transfer (driving force) from the steam.


I have also used a dual ramp of the SP. This is were I would increase the SP by one ratio (slope) for the first 2/3's of the time and another ratio (slope) the last 1/3 of the time. This helped allot on overshoot & settling time. I used two sets of gains when doing this.

See http://www.plctalk.net/qanda/showthread.php?t=5271 for a discussion about ramping.

1. The ramp idea is good but I would increment the ramp just before the PID block is called. This may mean the ramp is being updated every second or so.
2. The PID can follow a ramp without error but the integrator will wind up to do it. When the set point is reached the integrator must be reduced to keep from over shooting. One way to do this is to use a s-curve so there are no abrupt changes in the required output. The other way is to use a feed forward or bias when ramping up.
3. All of this can be calculated fairly accurately if the plant gain, time constant and dead time are known. These can be estimated using the techniques on www.controlguru.com. The bias can also be calculated from the plant parameters.

Thank you men for your advices. Would like to write how was the work done. I used theDave2 ideea about ramp SP. Controler used: S7-1200, CPU1214.
theDave2 - I tried to use PID components from standard library, but unfortunatly i didin't find how to change dinamicaly the parameters of PID, I will write simple P regulator by myself. Below can see the result of my experiment:
Conditions: Time 25 minutes. SPTemperature=120deg, PVTemperature~30. Each 75 seconds SP was increased with ~4deg