A/C PID Control?

Hi, I've hard-wired an 9,000BTU window A/C unit so that it can be operated via my PLC. So what I do is turn ON the compressor and fan using a couple of relays when I need to bring the temperature down.

I have a Vaisala HMT100 temperature/humidity sensor hooked up to an Allen Bradley MicroLogix 1400 PLC with a 0-20mA analog module.

It seems to me that a PID is mostly designed to control a process that is affected by a variable device (i.e. valve, heating element, fan, etc) where opening/intensity/level/speed/etc can be adjusted, so as to achieve an almost perfect SP.

So since the compressor is either turned OFF or ON at any given time, does it still make sense to us a PID? Should I just use logic to compare the SP to manually keep the compressor ON/OFF using ladder logic. So for example if the SP is 25 degC, should I just turn the compressor OFF when the temperature lowers to 25.2 and turn back ON when temperature reaches 25.8, for example.

I'm still learning about how to tune PIDs, so please bear with me if what I'm asking doesn't make sense and it's obvious to an experienced PLC programmer to always use PIDs, even in situations like this.

Some information: The SP=25 degC and it takes about 25 minutes to bring a temperature of 30 degC down to the setpoint (as shown in below image).



Any thoughts?

Thanks in advance!

SergioB said:

Hi, I've hard-wired an 9,000BTU window A/C unit so that it can be operated via my PLC. So what I do is turn ON the compressor and fan using a couple of relays when I need to bring the temperature down.

I have a Vaisala HMT100 temperature/humidity sensor hooked up to an Allen Bradley MicroLogix 1400 PLC with a 0-20mA analog module.

It seems to me that a PID is mostly designed to control a process that is affected by a variable device (i.e. valve, heating element, fan, etc) where opening/intensity/level/speed/etc can be adjusted, so as to achieve an almost perfect SP.

So since the compressor is either turned OFF or ON at any given time, does it still make sense to us a PID? Should I just use logic to compare the SP to manually keep the compressor ON/OFF using ladder logic. So for example if the SP is 25 degC, should I just turn the compressor OFF when the temperature lowers to 25.2 and turn back ON when temperature reaches 25.8, for example.

I'm still learning about how to tune PIDs, so please bear with me if what I'm asking doesn't make sense and it's obvious to an experienced PLC programmer to always use PIDs, even in situations like this.

Some information: The SP=25 degC and it takes about 25 minutes to bring a temperature of 30 degC down to the setpoint (as shown in below image).

Any thoughts?

Thanks in advance!

Click to expand...

Broadly speaking, PID control is only useful if you have a proportional output and field device. ON/OFF makes the most sense in this application.

JHarbin said:

Broadly speaking, PID control is only useful if you have a proportional output and field device. ON/OFF makes the most sense in this application.

Click to expand...

I agree with the above, also the deadband you have chosen seems to small to me ( .6 degress) you don't want to cycle your AC too much.

Or did you mean 24.8 to 25.2, even worse.

Mickey said:

I agree with the above, also the deadband you have chosen seems to small to me ( .6 degress) you don't want to cycle your AC too much.

Or did you mean 24.8 to 25.2, even worse.

Click to expand...

I agree about not wanting to cycle too much, but have you noticed that today's Electronic Thermostats control at less than +/- 1 ºF (and he is using ºC)?

To get around the deadband/ rapid cycling issue Mickey mentioned that is hard on the compressor, put an anti-restart timer in the logic so that it prevents the unit from restarting until it's been off for XX minutes (15?)

It's all about tradeoffs, a little more swing is more desirable than a failed compressor.

JHarbin said:

I agree about not wanting to cycle too much, but have you noticed that today's Electronic Thermostats control at less than +/- 1 ºF (and he is using ºC)?

Click to expand...

Didn't know that, my house is 28 years old and the thermostat I have has a deadband a whole bigger then 1 degree.

In this application, there needs to be a minimum off time for the compressor - once the compressor is cycled off it needs to stay off long enough for the high pressure side to bleed off, otherwise the compressor will not be able to start. I agree that simple on/off is probably best for this project.

...assuming this window unit A/C is built like any other I've seen.

edit - you guys are fast, 2 posts while I was composing that one!

Mickey said:

Didn't know that, my house is 28 years old and the thermostat I have has a deadband a whole bigger then 1 degree.

Click to expand...

My house is a little older than that, and my old electronic one had a bigger deadband too, but when I got a new Heat Pump, they put in a new Electronic Thermostat it was less than +/- 1 ºF. I was surprised so I started investigating and they all were programmed that way.

Ok, understood. It's nice to get some confirmation as to what I already suspected! Sure, I can relax the control so it uses +/- 2º C. The control is not that strict and it can fluctuate easily 2º C or 3º C.

Great idea about the anti-restart timer! I hadn't thought of doing that!

You guys are awesome, thanks for the quick reply!!!!

Rapid cycling of an A/C compressor will ruin the compressor.

As was mentioned, when the compressor turns off the high pressure side needs time to bleed down or it will stall when you restart it. Not only does the compressor require a minimum off time, it also requires a minimum on time. While the compressor is off refrigerant disolved in the compressor oil boils off, turning the oil into foam. When the compressor is started the oil foam is blown downstream, and the compressor has to run long enough to return that oil back to the compressor. When it is short cycled the compressor runs out of oil, with all the oil pumped out to the evaporator and condenser and none in the crankcase. Many a PLC programmer has trashed a compressor because he didn't understand the nuances of refrigeration.

If you really want to do capacity control then pump out the refrigerant and install a hot gas bypass valve and then draw a vacuum on the system and then put the refrigerant back in, and then time proportion the hot gas bypass valve. (This will add significantly to your electric bill because your compressor will run a lot longer, but you'll have tighter capacity and temperature control). Or you can replace the compressor with one of Copeland's scrolls that has a variable engagement scroll that can be controlled by an analog signal. Here again, your compressor will run longer.

If you're only experimenting and don't really care about the A/C unit then a fun project would be to try a blended control similar to a hot/cold deck used in large buildings. You could run the A/C at full speed and use a variable speed fan to draw in outside air and blend it to achieve a desired continuous temperature airflow.

Cool technical info. It's good to learn new things!

Thanks a lot!
sb

I was just doing some checking and I see that Copeland now has a compressor out that is suitable for use with a VFD. Its about freaking time. The A/C compressor market is long overdue. Now maybe that segment of the HVAC industry and the refrigeration industry can finally start making the energy efficiency gains that the heating segment has made.

TConnolly said:

I was just doing some checking and I see that Copeland now has a compressor out that is suitable for use with a VFD. Its about freaking time. The A/C compressor market is long overdue. Now maybe that segment of the HVAC industry and the refrigeration industry can finally start making the energy efficiency gains that the heating segment has made.

Click to expand...

Water Cooled and Air Cooled Chillers have had VFD options for some time.

You're right. The big centrifugal compressors used on industrial chillers have been variable speed for a long time. I should have been more clear - VFD compatibility for hermetic compressors like you would find in the home HVAC market, commercial installations, and small chillers. That is the market that I meant is long overdue for adopting VFDs, even though VFDs and modulated burners started making their way into home furnaces some time ago.

Just a note; On/Off Control does not exclude the use of a PID. There are instances (electric heater with an SSR) where On/Off Control is quite suitable with a PID. You would have a cycle time, which is controlled by your PID output percentage. So if your cycle time is 10 seconds, and your output is 75%, you would turn the output on for 7.5s and off for 2.5s

This is dependent on what you are controlling however, and if it can handle they cycling. It won't work in this compressor situation, but just something to keep in mind. PID is not exclusive to analog.