In one of the gas filtering and metering stations, a butterfly valve (B.V) skid (two runs) is installed upstream ultrasonic-based (USM) metering skid (two streams) such that the outlet isolation ball valve of the B.V skid goes to the same header to which the inlet isolation ball valve of the USM is connected.

The B.V is equipped with smart Siemens positioner and Biffi pilot.

The B.V is required to function as flow limiting tool.

The station PLC (installed in the control room) compares flow rate set values (manually daily-entered) with USM actual flow measurements and sends electrical analog (4/20 mA error signal) to the smart positioner which drives the Biffi pilot and finally this pilot control the B.V movement (opening percentage).

The problem is as follows:

As long as the PLC 'sees' the actual flow (measured by USM) is greater than the flow rate values set points, it forces the B.V to close till reaches 15% opening. Then every thing went into chaos/disorder.

At increasing load demand from power plant, the butterfly valve is strongly throttling the flow, the buffer size between the gas station and the power plant is small (few meters pipeline), the pressure downstream the B.V starts to decay and the DP across the B.V starts to increase, noise / vibration generated in the P/L and the USM paths starts to be lost and of course the error signal of the PLC sent to the B.V is erratic and the USM could give erroneous reading of flow rates 3 times higher than the station max. flow. This situation will continue till we isolate the Butterfly operation.

My Question is :

What is exactly the cause of this problem ?
What is this butterfly doing ?
Could any one give me problem analysis ?
Any suggestions to fix the problem?

 

I'm not a valve expert so all I can do is to encourage you to speak to a valve expert from Fisher. From what I understand about valves is that butterfly valves are not suitable for control, particularly near closed which is were you appear to be facing problems. Problem is, you cannot easily replace a butterfly valve with a ball or globe valve because the face-to-face distance is too different. However, Fisher has a new solution to this common problem called a "control disc". It has the same face-to-face dimension as a butterfly valve, but is shaped in such a way as to provide better low-flow control performance. Speak to one of their application engineers to see if that could be a solution to your problem.

Cheers,
Jonas

 

We are still waiting for a sketch, your description is too hard to understand but it sounds as though your flow meter is too close to the valve.

I also wonder if you would be better of with some sort of cascade control with pressure as the primary loop.

Roy

 

Ayman,

My experience using butterfly valves to control flow is limited to water applications, but my first thought was along the lines of something that Jonas mentioned above... that flow-control is difficult when the butterfly valve is near to being fully-closed.

So in those cases, we use pipe reducers on both sides of the butterfly to ensure that the butterfly valve is not over-sized for the application. This allows your average operating point to be "more open".

For example, a 6" pipeline flow-control application may work best with a 2" butterfly valve. Using a full 6" butterfly may result in cavitation which would shorten the life of the butterfly valve and also cause excessive noise and vibration.

Hope this helps!

 

> We are still waiting for a sketch, your description is too hard to understand but it sounds as though your flow meter is too close to the valve. <
>
> I also wonder if you would be better of with some sort of cascade control with pressure as the primary loop. <

Sorry for not being able to give a sketch.

The USM lines and butterfly valves are parallel to each other. In some cases, the butterfly skid (2 runs with I/O isolation valves) is installed upstream the USM and the USM skid (2 runs also with I/O isolation valves) is installed after the butterfly skid. In some other worst cases, the situation is reversed and the USM skid is installed downstream the butterfly valve skid.

Any how, the butterfly runs go into (end to) outlet header. From the same header, the butterfly runs emerged (starts). It is single common header.

I hope it is clear now enough, if not please tell me

Ayman

 

Ayman Kamha wrote :

Thanks for all those people who shared their opinions for that problem with me.

I just want to add something that might help the experts analyze that problem.

We made another trial with the Butterfly valve. We kept manually throttling by giving different descending opening percentages from 100% to nearly about 20%.

A very strange thing happened. The flow rate remained as it is about 65,000 SM3/H although we kept gradually manually-decreasing the Butterfly valve opening percentages from 100% to 20%.

To our surprise, the differential pressure drop across the Butterfly valve kept growing from nearly 0.4 bar till 10 Bar at 21% opening.

The flow was not controlled at all.

The Butterfly valve did not even start to "take action" and by "take action" I mean the growing up in DP from except maybe at 50% opening percentage.

From 100% opening till 50% opening DP was less than or equal 1 bar.

This means simply that the Butterfly valve does not exist till its opening percentage reaches about 50%, then it starts throttling and when it does it creates large DP across it and leaves the flow uncontrolled !!

HELP, Any explanation ??? Could any one "ANALYZE THAT" ??

 

Ayman,

First, I have no explaination for your lack of flow change except for you may want to check your flow meter! That's just not right.

However, The controlability that you are discribing is fairly typical for butterfly valves. If you look at the characteristic curve for a butterfly valve, you will see that it looks like an "S". Typically, you want to control in the range of 50 to 30 degrees open if at all possible. Worst case, you want to control up to 70 and down to 20 degrees open (really pushing it).

The lack of control response on the top end is due to the disk design. The instability on the bottom end is also due to the valve design and/or actuator sizing.

What to do? - Options
1. If you can give me all of your process data and valve/actuator information, I can resize your valve and see what I come up with. I can then make a recommendation. I would recommend that you check all of your process conditions to confirm they are what you are seeing in the real world.

2. Check your supply pressure to your actuator. If okay - Increase the actuator size or, if the actuator is double acting, increase the supply pressure. This could help with your stability. Changing the actuator could get expensive, so I would resize the valve before I did that.

3. adjust the positioner gain so that the signal response is slower. This may tune it up enough to control.

4. Combination of 2 and 3.

I hope this helps...

Cheers,
Charles

 

Raw guess on my part.

The ultrasonic flow meters are probably transit time meters, not doppler technology. The ultrasonic flow meter is down stream of the butterfly valve.

The nature of the butterfly control action is such that it cavitates at some point along its curve. Cavitation produces bubbles, which implode. But the concurrent turbulent action produces other aeration bubbles which do not implode. Those bubbles drive the transit time meter nuts.

Good transit time measurement assumes liquid, low aeration and near zero suspended solids, whereas Doppler relies on bubbles and suspended solids.

The result is that the flow meter's measurements in this 'bubble' state are not valid for transit time measurement technology.

I suspect that the readings are false readings due to the ultrasonics's close location downstream of the butterfly valve with its modulating performance.