WEG Variable Frequency Drive PID Tutorials - Process Variable Options and Optimization

In this video, we will take a quick look at the different kinds of inputs you can use for the process variable, what forward and reverse acting means and how to select the right one, and how to specify the process variable and drive frequency min and max parameters. In all of our examples in this video series we used analog input 1 as the process variable. To use analog input 2, just set parameter 921 to a 2 and populate the analog input 2 parameters. You can also use the difference between the two analog inputs. A common use for that would be to run two processes off of one analog signal, but then use analog input 2 to modify that on the second system. For example, maybe you want the second system to run faster or slower than the first or maybe you just need the ability to tweak system two to keep it in sync with system one. You have a forward acting system if the process variable increases when you increase the drive frequency output. Like in our demos where when we increased the motor speed, the pressure went up. A reverse acting system is common in refrigeration or cooling – where the more the motor speed INCREASES, the more the temperature DECREASES. It’s really important to choose the right one here, because it determines the sign of this input. Get it backwards and PID won’t work. Understanding how to set the process variable min and max along with the drive frequency min and max correctly is the key to getting PID to work well. So it’s worth spending a few minutes to take a close look at that. The starting point with PID is always fundamentally understanding how your system works. In the Quick Start videos, we ran our system across all possible drive speeds and all possible loads to produce these curves. We saw that as the drive speed increased, the pressure increased – which told us we had a forward acting system – and it did it in a roughly linear fashion between these drive speeds. That’s important – the more linear your system is, the better PID will work. Centrifugal pumps always require some minimum speed to operate and in the quick start videos we found for our system that turned out to be around 20 hertz. At the top end, we maxed out our 0 to 15 psi pressure sensor before getting to the full drive speed for three of the loads, so things go non-linear past this point. We want to maintain 6 psi, so we only need to operate the motor between roughly 35 and maybe 42 hertz. But we don’t want to limit PID to that range. We want to give PID as much motor range as we can. Why? So PID can recover quickly. For example, when there is a sudden pressure decrease, PID will want to push the motor really hard to get to 6 psi as fast as possible, but then back the motor speed off as it gets close to the 6 psi setpoint. It’s kind of like when you accelerate your car onto the highway. It’s sure is nice to have horsepower headroom so you can smash the gas pedal like you want to go 120 miles per hour, but then when you get close to 65 miles per hour you back off. It gets you to your desired speed faster. So while we only need to operate in this range, we want to widen that as much as possible to give PID plenty of head room so it can get to its desired setpoint as fast as possible. We know our pump stops working at 20 hertz, so how about we make our bottom end 25 hertz just to give us some margin and put that in parameter 133. The top end is a little harder. If we had chosen a 30-psi sensor then these curves would probably have gone like this and we could have used a larger motor range to give PID more headroom. I intentionally chose an undersized sensor to help you see the kinds of things you need to look out for. So make sure your sensor can maximize the use of your motor. We need to avoid non-linear portions of the curves so we don’t confuse PID, which means we need to limit the max motor speed to somewhere in here. We’ll use 50 hertz as our top end and enter that in parameter 134. It’s important that you understand that when you look at parameter 917 which shows the control output in percent, that this is zero percent and this is 100%. So if parameter 917 says the output is at 50% it means it is half way between 25 and 50 hertz – or whatever you entered for the min and max drive frequency. What do we use for PV max? Normally you only have one curve like if you are simply monitoring the temperature of an oven. But, because we have a different curve for each load, which one do we use? If you just take an average it will get you pretty close and we can tweak it later if we need to. We just grab this guy for PV max and this guy for PV min and enter them in parameters 922 and 923. The point is we are going to hand PID a line that goes from this point to this one and PID will use that line to adjust the motor to react to changes in the process variable to maintain 6 psi. The better that line approximates your actual system the better PID will work. I’ve already entered all four of those into the drive so let’s switch this thing on and try it. I have P set to 2 and I set to 5 for this demo. Looks like to hold 6 psi, PID has the drive running at around 33.4 hertz. Except it’s not quite at 6 psi, which means our line is a little bit off. Well the curves were most spread apart up here so this is where the average is most likely off a bit. So, I simply got to parameter 923 to tweak the PV Max value to bring that up to 6 psi. That’s it – we’re done! I’ll close the valves and sure enough PID tracks it and automatically adjusts the motor speed to get the system back to 6 psi. What if you can’t get access your system to create a nice process variable drive frequency curve like this? Well, there are a couple things you can do that aren’t perfect but will get you by in a pinch. You know your system so you could probably come up with a pretty good guess on a couple data points you can use to extrapolate PV min and max over whatever you know your drives min and max operating frequency should be. You will be surprised at how tolerant PID is of that line you give it as long as you are in the ball park. It won’t give you optimum results, but you’ll be amazed how well it works. As a worst case example of that, the other method is to simply blindly use 0 to 60 Hz and the full range of the sensor, which is 0 to 15 psi for this demo. So instead of this, our line will look like this . Will that work? Let’s find out. I’ll go to parameter 922 – the PV min and push it down to zero. And then to parameter 923 and push it up to my sensor’s max of 15 psi. Then over to parameter 133 to change the min drive frequency to 0 hertz. And parameter 134 to change the max drive frequency to 60 Hz. The drive is now set up to use this line to determine how it is going to change the drive frequency output to react to changes in the process variable. Let’s go to parameter 2 so we can see the drive speed and it looks like PID is holding the drive at around 33.2 or 3 hertz to maintain 5.9 psi … regardless of how many valves I have open or closed. But wait a minute – why isn’t that at 6 psi? Well, it’s because we didn’t use the right points for process variable and drive frequency min and max settings. We could tweak the PV min and max values to make it better like we did a minute ago, OR we can just cheat and go to parameter 911 and artificially raise the setpoint. Looks like if we tell it to go to 6.1, maybe 6.12 psi we will actually get 6.0 psi. Cool. It’s important understand that the only reason this worked so well is my sensor is really well matched to my process. Zero to 15 psi happened to track my process reasonably well across the drive’s 0 to 60 hertz range. Suppose I had a 0 to 100 psi sensor and we used the full range of that sensor for PV min and max. Now our line isn’t close to our process at all, is it? Would that have given us such good results? Well no, it would have been way off. Regardless, even though we didn’t use the actual system curves we’re still able to get a reasonable tuning regardless of load. Pretty cool! What if you did plot your systems performance and you got something odd like this? Well, you just have create a line that best represents whatever you have. PID is pretty tolerant if you can just get close. That should give you a better feel for how to choose the values you need for the min and max process variables and drive frequencies. If you want to learn more then click here to see all of the videos in this series. Click here to subscribe to our YouTube channel so you will be notified when we publish new videos and click here to learn about AutomationDirect’s Free award winning support options.