PID toolkit suitable for this process??

[stu22]

Hi, i'm considering purchasing the PID toolkit to control a process that must have a constant temperature. Basically i use a furnace to heat up salt to a molten temperature of 800 degree's celsius. The salt must be kept at 800 degrees C for long periods of time e.g. 24 hours. I have a thermocouple in the salt, and this is where the temperature is monitored. A furnace keeps the salt constantly heated, but i need a way of keeping the salt temperature constant. Would a PID with a feedback loop do this??

Thanks Stuart

[johnfr]

I have used the PID tool kit for numerous temperature control applications. While my applications were used to control between -30C to +65C, I don't see why it would not work for your application. I am able to control the temperature of a disc drive under test to within +/-.5C.  The thermal couple should work good as a feed back loop. That is what I am using. Keep in mind that most thermal couples for your temperature range are only accurate to +/-5C. You may want to have a 2 channel (2 thermocouples) system with a redundant sensor in case the first one fails. As I recall, the PID controller can sense an open thermocouple and shut down should it detect a failed sensor.

The PID tool kit has an auto tune feature that works fair.  Generally, you need to have some initial understanding of what kind of time delays occur between when heat is applied and the temperature starts to change. I would assume you have significant mass of material along with some fairly high power heaters. Try to establish what the time delay is for step change in temperature is. You will then want to seed the "I" parameter of the PID controller with that time.  When using the auto tuner make sure the system is supervised throughout the auto tune process. With the kinds of power you are dealing with you don't want it going out of control.

Good Luck.

[AnalogKid2DigitalMan]

Stuart:

Is using LabVIEW a requirement for this application? I would suggest an off-the-shelf PID process controller, which is a pure hardware approach and safer (and would get you up and running much faster).

If you do use LabVIEW, make sure you account for open thermocouple conditions which could lead to thermal runaway. Also, what if the PC locks up?

With either approach I would also suggest implemententing a separate over-temperature safety relay in case the relay or SSR applying power to the furnace fails in a shorted state.

I routinely deal with furnaces and heaters that run in the 350-1800 C range, safety is paramount in these applications.

-AK2DM

[Andrew McL]

Hi Stuart,

Thank you for contacting National Instruments. PID would be a good way of implementing this type of process control. I've attached a link here to an article that gives an overview of PID and specifically talks a little about some of the PID functions contained within the LabVIEW Toolkit. 

As AK2DM correctly points out, doing process control with a PC is risky in that a PC can lock up. It would be worth considering either using a Real-Time PXI or cRIO platform to give the determinism and reliability you will need. I would recommend getting in touch with the National Instruments Sales Team if you are interested in either of these options so that they can advise you better on what system you need.

If you have any further questions, please let me know and I shall try and answer these.

Many thanks,