Aeration blower D.O application

Hello,

Does anyone have an alternative idea to correlate the speed of a VFD on an aeration blower to a desired Dissolved Oxygen setpoint. Currently I am using a controller based PID control, but this has proved to be quite hard to tune and over/undershoots the setpoint continually resulting in undesired life on the blower motor and cyclical D.O.

Any recommendations/ideas would be greatly appreciated.

Thanks,

Ben

Because the lag time for the PV is so long between changes in the CV, I've found PID control as implemented in PLCs to be a poor choice for controlling D.O..

A company I used to work for had some very cool and very proprietary control logic for D.O for anywhere from 1 to 16 blowers in cascading, lead lag, priority, KWh Time-of-Day etc control that would be overkill for your application but the basics are the same. You could roll your own slow PI control - look at your error after a length of time (adjustable, say.. 6 minutes at first or whatever your lag time is) and adjust your CV output proportionally. Use a ramp - the larger the error, the greater the output change (keeping in mind a minimum output). Rinse repeat. Add a dead band to prevent constant CV changes and you should be good to go.

what is your blower type, do you just have 1 in 1 tank ? and what is your operating parameters.
The remaining o2 in water is far from being linear against blower motor speed so you are right...
The reaction time for the do is very slow for a pid loop too, i use only my own integral part to make it easier...or G / D would need to be reduced a lot...
I delay the scan to every 5 to 10 minutes to let the water react before making a correction...otherwise, the speed will just bounce around the place until something break.

bjh said:

Because the lag time for the PV is so long between changes in the CV, I've found PID control as implemented in PLCs to be a poor choice for controlling D.O..

A company I used to work for had some very cool and very proprietary control logic for D.O for anywhere from 1 to 16 blowers in cascading, lead lag, priority, KWh Time-of-Day etc control that would be overkill for your application but the basics are the same. You could roll your own slow PI control - look at your error after a length of time (adjustable, say.. 6 minutes at first or whatever your lag time is) and adjust your CV output proportionally. Use a ramp - the larger the error, the greater the output change (keeping in mind a minimum output). Rinse repeat. Add a dead band to prevent constant CV changes and you should be good to go.

Click to expand...

Exactly, lol i was writting my post when you respond!

I'll put in a plug for my book: http://www.wiley.com/WileyCDA/WileyTitle/productCd-1118389980.html

I was never happy with PI or PID control for DO, and developed a floating control algorithm that worked quite well. There are a few vendors that have adapted my technique and would potentially be system suppliers:
https://www.necontrols.com/
https://www.howden.com/Resources/Documents/B-Controls_8.25x11.pdf
http://www.wastewater.com/symphony

Jeff,

Thanks for the follow-up. As well as your links Tom. It is a centrifugal blower with VFD speed control. There are 4 blowers. One in each channel of a carousel. Outer,inner,outer, inner. Typically, in cooler months, the inners are sufficient to maintain D.O. Previously, they only had low and high speed options. Trying to expand on that. They can manually enter the speed hertz, with a minimum run speed.

I guess I'll just have to keep playing with the PID. I was able to get it to track fairly dead on with compare instructions regarding the SP and PV then stepping the PID output speed at a lower increment. But obviously as that hones in the SP, you get to a point where it will not get back to Max speed.

But all the info is good. I appreciate it. I haven't tuned with the deadband much yet. I saw that as the speed reference not changing with a range of PV fluctuation. Which by the time the momentum of the D.O. starts going,it's too late.

You can improve your control, regardless of algorithm, by changing the controlled variable to air flow rate instead of speed.

The oxygen transfer rate to the process is a complex multi-variable function, but essentially the controllable parameter is mass air flow rate. The air flow rate provided by the blowers is also a complex multi-variable function, with ambient temperature and discharge pressure having significant impacts. The important fact is that the air flow from a centrifugal blower at a given speed can vary dramatically.

I suggest you use air flow to the process as your primary controlled variable, and employ a second control loop to modulate blower speed to provide the required air flow rate.

by the way be carefull centrifugal blowers have a minimum flow to keep to avoid destroying them by flow surge. if you reach that point, there will be flow pulse that can break pipes and the blower internal assembly...

Tom Jenkins said:

I suggest you use air flow to the process as your primary controlled variable, and employ a second control loop to modulate blower speed to provide the required air flow rate.

Click to expand...

Tom, do you mean "use mass air flow to the process" as opposed to volumetric air flow (not compensated for volumetric changes due to temperature or line pressure)?

danw said:

Tom, do you mean "use mass air flow to the process" as opposed to volumetric air flow (not compensated for volumetric changes due to temperature or line pressure)?

Click to expand...

I'm sure that is what he means.

For DO could you do a feedforward calculation based on the DO reading of the pond and volume of the pond to come up with a targeted mass flow rate of air?
The mass flow rate target could be the output of the out cascaded loop with an inner loop that is actually controlling the the blower to match the mass flow.

That's just an idea, I do not know anything about DO control in waste water.

danw said:

Tom, do you mean "use mass air flow to the process" as opposed to volumetric air flow (not compensated for volumetric changes due to temperature or line pressure)?

Click to expand...

Mass air flow rate is best if it is available. There are thermal mass air flow meters that measure this directly, or you can use temperature and pressure correction on venturis or other volumetric devices.

However, because this is a feedback loop, you can get satisfactory results by using volumetric flow rate instead. Your control logic will just keep tweaking the air flow from the blower volmuetric until you reach the desired DO.

Blowers are volumetric flow devices. Jeff mentioned surge - minimum air flow. This is a volumetric flow based parameter. If you don't have a blower manufacturer's surge protection panel then you need to include the minimum flow rate in your logic. You can get by with minimum speed, but it has to be set conservatively to avoid catastrophic blower failure.