Rotary blower VSD speed control scheme for aeration basins. 1

[silebi66]

Dear professional eng-tips members:

Please bear with me if my question is too trivial for you guys. The variable frequency drive (VFD) of Blowers(2 is on/1 is spare) to be controlled the airflow distribute to the 2 aeration basins and each basin

with a dissolved oxygen (DO) monitor on top.

There are two control options.

1. DO control(fig 1)

By varying the speed of a motor (VSD or VFD), and its driven load, the airflow capacity are adjusted in direct proportion to the DO level. As the flow is proportional to the blower speed, as the DO level in the

basin is reach the set point, the blower speed will be adjusted to lower speed. But, the question bothers me is what is the set point of DO level for the air demand in aeration basin ? And , cause the controllers,

and programming necessary’s cost is high. By DO measuring in each basin and the DO controller adjusts the blower VFD motor set point at its corresponding basin DO controller is quite costly. Please advice me

that this control scheme is appropriate or not.

2. Cascade flow control (fig 2)

The control for positioning the blower inlet valve, and the final challenge is the coordination of multiple blower operation. The method for controlling the blowers is cascade control. There is no DO loop, and the

set point for the airflow controller at each basin are adjusted with flow-paced controller proportional to plant influent flow. DO meter monitored periodically with a portable meter to correlate DO levels with

influent flow. The problem is the in parallel control, whenever the next blower is required to come on line cause operating blowers are out of capacity, the following occurs :

- running blowers are adjusted to minimum airflow by closing the inlet valve.

- The next blower is started and as air demeaned increases, the inlet valves to running blowers are modulated simultaneously to increase air flow, the process is reversed as blowers are dropped off. So, one blower

may have more airflow than the other one blower. The unstable blowers running condition may occurs.

I wonder these two control option description that I wrote is correct or not.

Please kindly correct and comment on these.

Really appreciate for your kindly help.

Thank you very much.

 

[silebi66]

I am sorry to post the fig 2 is not successful at first time.
The following attach is fig 2.

 

[DRWeig]

Hi silebi66,

I've done this exact setup for a large wastewater treatment plant. You don't mention what your process is, so I can't help with DO setpoint. For activated sludge treatment, treatment tanks were maintained at 2.0 ppm, digesters 0.5 ppm.

Scheme to minimize blower energy:

1) Control loop for DO with an air valve or damper at each tank. PID loop to maintain constant DO at setpoint by modulating the air valve or damper.

2) Control loop for system air pressure with blower speed. PID loop to maintain constant air pressure at setpoint by modulating blower speed.

3) Logic and PID loop to decrease air pressure setpoint until at least one air valve is wide open. This insures that the air pressure is at a minimum, and thus energy is at a minimum.

4) If tank(s) with wide open air valves begin to have their DO fall below setpoint, change blower control to modulate pressure setpoint according to DO in the worst tank. This keeps the worst-case valve wide open, allowing minimum air system pressure and thus minimum energy.

5) When demand causes worst-case tank to have its DO fall below setpoint with one blower at full speed, bring on second blower. Modulate them together (both on same output signal) in the same manner as steps 3 and 4 above.

6) Minimize all other fixed pressure drops. Never modulate or close blower inlet or outlet dampers, leave them wide open. Use them for isolation only (close when blower off to prevent backflow).

7) Keep the inlet filters clean!

Implementing the system above reduced energy at one large plant by almost half. The system ran with a fixed pressure setpoint of 8.0 PSI and individual tank control via fixed air valves that were manually adjusted once a day according to lab-analyzed DO. They kept 2 blowers on in winter, 3 in summer at 5000 HP each.

With new control strategy, only 1 blower is needed in winter and 2 in summer. System pressure modulates as above, and it turns out they only needed about 5.5 PSI while running. Of course, we added DO sensors to each tank to enable good control. They still need 8 PSI at startup to overcome static head, but that rarely happens -- municipal wastewater treatment plants don't shut down unless a catastrophe happens. Energy costs reduced from $8 million to about $4.5 million per year. The cost of the controls was paid back with energy savings in less than one month.

Please provide more information about your process, such as purpose of aeration, design system airflow, design system pressure, blower horsepower, depth of air discharge diffusers, and duct sizes. Then wait for others to respond to this thread. My experience is limited to wastewater.

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[bimr]

Regarding "2. Cascade flow control (fig 2)".

This scheme will never work for these reasons among others:

1. It is not possible to modulate air flows with valves at the low pressures that exist in the activated sludge process. You probably need a pressure drop of 3-5 psi across a valve in order to control flow. You don't have that pressure available and that pressure drop is an energy loss.

2. Inlet air dampers are not used on PD blowers, but are used on centrifugal blowers.

3. Air requirements are not proportional to flow, but to BOD loading and mixing.

For aeration control:

1. Minimum setpoint is 2 mg/l for the basic activated sludge process. If you have BNR, there may be other setpoints.

2. You may need multiple blower sizes to improve the range of you aeration system. Most VFD's only have a turndown to approximately 70%.

3. With a PD blower, the backpressure is not set by the blower, it is set by the friction pressure loss and static water head. This is reason that it is close to impossible to precisely modulate the volume out of centrifugal blowers. The blower discharge curve also changes with the pressure that you are blowing against.

4. You need high quality DO probes like the membrane covered ones that Endress+Hauser makes that are resistant to fouling.

5. There is also a minimum air flow requirement that is necessary for adequate mixing that limits the energy savings.

6. Don't try to modulate air flow with valves, it won't work. Valves should be used for on/off control. Control valves are also expensive.

Note that people overstate the energy savings possible with these systems in order to sell them. Little thought is given to how to make them work because the people selling them do not understand the system.

 

[DRWeig]

bimr, most respectfully I have to disagree with a few things:

First a note: I don't sell these things or install them. I'm a controls guy, not a supplier. In this case, my company only specified sequences of control and a mechanical firm did the design. We did, however, get hired to program the controllers and commission the work. I did that myself.

The plant that I worked was an existing system with inlet guide vanes on the blowers and butterfly air valves at the diffuser in each tank (1970 vintage system). It worked OK, but the control for system pressure was not coordinated with the control for DO in each tank. Plus, the DO was only checked daily. When we first arrived at the plant, the pressure setpoint was 8 psi and the most wide-open air valve was at 85 percent closed. Many were more than 95 percent closed. Horrible pressure drop and much energy waste. While they needed high input pressure to overcome hydrostatic head at startup - diffusers were 16 feet deep - only about 60 percent of that pressure was needed when airflow was established. It was a large municipal plant, so the BOD was reasonably constant.

1. After installation of DO transmitters, control of DO via the individual tank air valves was quite effective and stable at 2 ml/l (I goofed when I said ppm above). We didn't change the air flow values much from what they were when we arrived, but they did go down a bit due to better control of DO. We only really messed with system pressure. In a system with 20 aeration tanks running off one aeration header, how else could we control air flow to each tank? The wide-open pressure drop across the valve-diffuser combination was only about 1.5 psid. That's still too much, but when we arrived at the plant it was a shade over 3 psid since the valves were almost closed.

2. We were able to vary the system pressure to keep the worst-case valve wide open, much like we do in a variable-air-volume HVAC system to minimize energy.

3. We were able to control that worst-case tank's DO with the inlet guide vanes once its air valve was out of the way at 100 percent open. That control was not as accurate, but it was within the bounds that the plant operators set. The system pressure was minimized in that way. It was only high enough to meet the demand of the worst case tank.

The proof was in the operation logs. DO values per tank were dead-on 2.0 and very stable. Average annual power demand by the blowers went from 11 MW to a bit over 7 MW if my memory is working right today. Combined with the addition of variable speed to the big lift pumps (4500 hp total) and a bunch of smaller items like efficient lighting and HVAC, the plant's electric bill dropped by $3.5 million year-over-year. Calculations based on metered power to the blowers showed that more than half of that was from automating air pressure control. The system was investigated by some engineers from DOE and awarded the national municipal energy innovator medal that year, since it was really the first integrated VAV system for this application. Keep in mind that this was 30 years ago...

I do agree that VFDs are probably not worth their cost for pressure control in this kind of system. The system pressure can't vary below a limit that's pretty high due to depth of diffuser, and the blowers can be selected to have a high efficiency at the expected pressure and flow.

I also agree that valves are expensive, but in our case they were already there.

You are very right about having DO transmitters that prevent fouling.

Reading back on all I just wrote, I guess my only real disagreement with you is that air valves can control well, and I would look forward to debating it if you think not. At least they did in my experience. I'm long since out of that business.




Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[bimr]

I was not insinuating that you were a supplier of such systems. However, there are plenty of suppliers making claims regarding these systems. Before specifying such a system, it is important to make sure you specify one that will work. It is buyer beware.

I can show you a supplier advertising energy savings with such a system with air valves and centrifugal blowers. Yet if you go to the WWTP, you observe that the control system does not function as it is advertised.

http://www.wwdmag.com/blowers-air/blown-out-wastewater

The backpressure in the aeration system is not set by the blower. The backpressure is set by the friction headloss and static head. As the backpressure increases (by installation of a valve, fouling of diffusers, etc.), the output of a centrifugal blower will decrease. The blower output decreases with increased pressure as the blower operating point moves leftward on the blower operating curve. As you decrease blower output with a VFD, the discharge pressure will also increase. This is too complicated to program in a system with multiple aeration tanks, multiple blowers, etc. Many of the new plants also have different aeration requirements in each tank which adds further complications.

In your own post above, you point out some of the difficulties.

I was actually recommending the PD blower system with VFD:

1. Blower outpout will be directly proportional to blower speed.
A. This eliminates the potential of centrifugal blower surging. Simplifies programming.
B. This eliminates variable centrifugal blower output as output changes with centrifugal blower pressure.
C. Allows multiple blower sizes to be used. Multiple blower sizes are necessary in order to increase the range of operations of the WWtP. (Blowers are not usually all the same size.) It is not possible to use different centrifugal blower sizes because each centrifugal blower will have a different operating curve.

 

[DRWeig]

Good points again, bimr.

Your way sure is more simple, and as long as you have DO control and some way of modulating flow without getting into many control loops, you should be able to minimize blower energy. I was blessed with an existing system to try and optimize, so my choices were limited to control.



Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[silebi66]

Thank you very much for DRWeig and Bimr's generous and kindly help, and really appreciate your help.This waste-water treatment process including anaerobic (UASB)/aerobic aeration pit (activated sludge process),

and clarifier.The wastewater will be cooled to 35C and adjust pH to 6~8 before entering to 4 UASB reactors (upflow sludge blanket reactor,detention time =25 hr, MLSS=16000ppm F/M[food to microorganism ratio]=

0.35kgCOD / kgSS.d, plan to reach COD removing rate=70%, but not sure). Then, wastewater will be transferred into the aerobic basin (MLSS=6000ppm, F/M=2.48kgCOD/kgSS.d,COD removing rate=85%), and then into

clarifier/sand filter process.

Here is the preliminary parameter for this roost type blower:

Blower capacity = 120m3/min

Blower discharge pressure = 2.0 barA

Size of air discharge diffusers= 10"

depth of aerobic aeration basin = 5 m

depth of air discharge diffusers= 0.2kg/cm2

This process will be adopted the roost type blower instead of centrifugal type blower. It is just as you said that centrifugal with a VFD has inherent characteristics of their performance curves –
the variance of the air delivery is not proportional to the speed. And, the head characteristics of the blower change with speed as well.(this is refer from other senior suggestion)


Please correct me anything wrong.

Thank you so much.