Ultrasonic level transducer to VFD - analog or PID?

Hello,

I've recently installed a ultrasonic level transducer to control a VFD pumping out of a vat. The vat is "slug" filled by other pumps. The purpose of the vat is to act as a balance tank.

I've simply programmed the drive in analog control mode. Vat empty is minimum speed, vat full is max speed. It works fairly good and really smooths out the process, however I am dealing with varying head pressure against the pump. Filtration efficiency and the level in the output tank are placing varying head pressure on the pump, and as such, once pump speed slows to a certain RPM, it will no longer actually pump. Given clean filters and an empty output tank, the vat will empty almost completely at a much lower pump RPM, full output tank and higher filter back pressure requires higher pump RPM to maintain flow.

I've programmed the level transducer (Flowline EchoPod) to read 0 mA at 0" in vat, 20 mA at 46.6" (vat over full).

So, I'm wondering if a PID loop would be a better way to treat this scenario? I've done some trial settings, but can't seem to get it to work like I've envisioned it in my head.

Am I correct that this be a reverse acting (cooling) PID loop?

Would a PID loop be appropriate in this application?

Thanks!

It sounds like you need to use the sleep function. When the vat is close to empty and the pump speed drops below a certain rpm for set length of time, the drive stops running. When vat level reaches a preset height, the drive wakes up and starts the pump again.

Yes, if I did that, I could raise my minimum Hz to always overcome the backpressure, and then drop to 0 at desired level. However, then the pump would be running faster than necessary at most times, and then off a certain amount of time to make up for that. It would be more efficient if the pump ran as slow as possible all day. Which brings me back to desiring a more complicated control process than simply linear analog control.

Honestly, if you study the pump curves, I bet that a higher minimum speed will fit closer to high efficiency than keeping the head too low.

Unless, of course, there is a reason to empty the vat, keeping a reasonable head should minimize short cycling and running dry.

Remember that a motor on a vfd does not see the kind of inrush and heat that full voltage starting does. We have many pumps running on vfds that simply ramp up and down off a single set point (on/off level switch) that perform fine, and noticeably better than when run as full voltage start.

A PID is more for if you wanted the pumps to maintain a certain level in the tank. I suppose you could just use the P term and just have it be a straight proportional control, and set your setpoint to whatever minimum level you want to have in the tank. Don't mess with the I and D terms because (from what I can tell) you don't really care what the tank level actually is, just so long as the pumps are running a speed proportional to the level.

PID won't do a thing to help your problems. As brstilson says, it is only of benefit if you want to maintain a constant level in the tank (vat). It seems to me that the vat is being used to attenuate flow fluctuations and what you really want to accomplish ideally is constant flow. To do that you would need a flow meter and PID or other control to modulate pump speed to maintain constant flow as discharge and suction head varies. You would also need to add limits so the pump would go to full speed if vat level got above a certain point and stop the pump if it got below a certain point.

Most centrifugal pump applications have a minimum speed below which no fluid moves. Most VFDs allow you to set a minimum speed that the pump won't operate below.

If you don't want to add a flow meter you can monitor pump power at the VFD and add a pressure transmitter (they are cheap) and then calculate the flow based on the formula for power as a function of flow, pump efficiency, and pressure. It won't be precise, but based on the above it would be accurate enough for your purposes.

You don't identify the VFD model. Many of them (most of the first tier suppliers anyway) have an extra analog input and some programming capabilities in the drive itself to implement this control. Talk to your drive supplier. Failing that, you can get a small PLC for a few hundred bucks from AutomationDirect.com and many others.

Gene, I've been unable to locate a curve for this pump. I have yet to contact the manufacturer, but one would think it's on their website?

brstilson, that's what I had in mind, a simple control loop, but that's more or less what I've got with an analog control.

Tom, I'm running a DuraPulse drive from Automation Direct. A pressure transducer would be tricky to implement here as it's a sanitary application, making the install more complicated.

I have previously ran this setup on multiple level switches and it works, but creates too much slugging again. If a higher level switch is closed, it would ramp the pump up to a higher rpm, and it would quickly pump down to break that contact again, which was basically a waste of efficiency, as it would have been better off with a slightly higher speed for a longer time period. Which brings you back to more level switches, which ends up being more like a level transducer, with completely variable flow.

Sounds like I'm not doing too bad with the current analog only control setup. If pump head is higher, it simply needs to fill the vat more before higher pump speed overcomes the head pressure. It would sit at this higher level until something major changes to the back pressure, and it would pump down lower and slow the pump down.

Thanks for all the replies.