SHAMOO (Customer) asked a question.
Hi folks! Prototyping a system controller which, among other things, would allow us to quickly and easily assess the % of water in a water tank. We got it to work with one transducer in one test environment, but we added a second transducer and noticed 1. The transducers output different values at identical depths and 2. The solution we currently have has no accommodations for tanks of different sizes, without recoding parts in the PLC.
We think the solution needs some mix of scaling and some sort of calibration method. Our idea is a two-step process done through the CMore HMI where the user would navigate to a config page, hold the transducer in ambient air outside of the tank and hit a button, then lower the transducer into a full tank and hit a button. Then in theory the software has collected the input received at 0% and 100% and that becomes a sort of max & min for the PLC to work off of.
How can we go about doing this? Who's been through this challenge before?
Hi Shamoo, I'm, assuming you're using ultrasonic sensors? If so, setting the points outside of the tank is pretty common. Many times tank geometry is the challenge if the bottom of your tank is not flat (as most aren't) and you may need to accomodate for that depending on what you consider 0% full.
Once you set your min/max values on the sensor, you should be able to just set those values in the Click for 0-100% levels. If you need to handle the geometry, you'll need to do further calculations based on your signal.
Are you using more than one sensor per tank?
We're using torpedo style pressure sensors, two different types. One is a TL-136, I'm not sure what the other one is. We have some tests going on with ultrasonic sensors on Raspberry Pi but unfortunately haven't made much headway with them. I do believe they have some advantages, we just don't have much invested into testing that tech yet.
We're working with vertical water storage tanks so most of them would be flat-bottomed cylinders. We do have a large product line that isn't shaped like that and will require some heavy math to accommodate but we're trying to get the basic tanks handled before getting more technical.
Our test environment has two of the torpedo style sensors in it.
easy way to do it outside the tank is measure the inside height of the tank from bottom to sensor mount then cut a piece of 2" or larger pvc pipe the same height as the inside of the tank, set the sensor on top of the pipe, lay pipe flat on floor against wall or stand on ladder and calibrate for low setting, then mount sensor
That's one of the approaches we've taken. In one approach, we have a math equation that uses the height of the tank to influence the output. I'd just like to make it easier for customers to calibrate their systems, which is why I like the idea of measuring at 0%, measuring at 100%, and the PLC figures out the rest.
Pretty simple if it's just linear. Instead of using constants for raw min/max in your equation use variables.
Have a screen in HMI that displays current raw and scaled EU. Add a button to set current raw as raw min and another for max. In PLC use these to copy the current raw into the variables used for scaling. Can be useful to have numeric entry boxes that display those variables and allows direct editing of you want to quickly change it or adjust it without being able to hit min or max.