Flow Hunting Problem

What is flowing?
What rate is it flowing at?
What shape of pipe work is it flowing through?
What is the temperature (you say40, then you say high temperature... these don't tally).
What range of flow are you trying to achieve?
What is the nominal position if the valve . E.G is it working nearly fully open, fully closed, in the middle.

You say the flow in hunting... over what range?
And over what timescale/frequency? That may tell you a lot.

When you say varying the input by 1%... do you mean 1 degree of angular position? Or if you mean 1%.... 1% of what?... my shoe size?

Simply adding a long straight pipe run filled with smaller pipes to linearise the flow may help.
How do you know the hunting is in the actual flow and not a function of the flow sensors or their position, installation,software etc.
I'm sure someone who knows more about this stuff will ask some more pertinent questions, but these few are a good start.
Whew... nuff for now... time for tea and toast.
Del

(There is no backlash error. Statements like that worry me)

This sounds whacky to me. A 0.1% delta gives 30 TPH variation?? What is 30 TPH?

Tons per hour? Throughput per hour? Tadpoles per hour?

You must have one gigantic flow if 0.1% is 30 TPH?

Whacky, whacky, whacky by cracky!

Fisher.

Thanks a lot for the responses.

Well, answers to queries raised are as below:-

1. Range of Flow is 200 ton/hr.

2. The liquid is DTA-dowtherm-A (Therminol-VP), boiling point of 257 deg C, freezing point of 12 deg C.

3. Normally we need to maintain the flow of 140 TPH with a low-low set point of 111 TPH.

4. Temperature is of the order of 400 deg C.

5. The flow is through a horizontal pipe of 6" dia.

6. The installed valve is Fisher's V-300 (rotary type with 90 deg of rotation) with actuator type 1052-40 and CV of 1100.

7. The original positioner was DVC6020.

8. The valve generally opens around 36 to 37 % to get the flow.

Problem faced after installation of new positioner:-

1. The valve is kept in manual mode and the flow is around 150 TPH at opening of 36.3%.

2. When we reduce the opening to 36.2%, (linear output from DCS in terms of 4-20 ma, which is converted into angular movement at positioner side) the flow after sometimes drops to 125 TPH. This abnormality is what I am looking forward for your suggestions.

3. The valve is physically checked and position marked by marker and noticed that there is no abrupt physical movement in valve, so why does the flow (which is measured by two FT's) get varied.

4. Just for information, I get the same flow at 36% opening, which I use to get with Fisher positioner. But when I vary 0.1% in new case, there is large variation in flow, which is next to impossible but is being observed.

5. I need suggestions what can be the probable reasons of such a behavior.

<...boiling point of 257 deg C...Temperature is of the order of 400 deg C...>

One of them could be that the liquid is boiling as it goes through the valve.

<...carried out an in-house modification and installed a smart Metso make 9102 positioner (on the front side with a gap of 30 mm from the body)...>

Was this modification approved by Fisher? If not, why not?

<...The valve is tuned and put to service....>

How odd. Normally, automatic control loops are tuned, and yet this loop is in manual. How abstruse.

How did the telephone call with Fisher go? Do tell.

How does the flow measuring device respond to boiling liquid passing through it? What sort of flowmeter is it?

If the flow low-low is so low, then why is the fluctuation of any immediate concern?

...and why is anyone twiddling with the valve position in any case?

I expect the downstream pressure to matter in this case. You appear to be far above the boiling point for this Dowtherm, and maybe you have misapplied the fluid to the situation. Make sure the valve positioner output 4-20 mA signal matches what is going on with the valve. You should also install a microphone and record various sounds that take place, and thereby gain some insight.

Dowtherm-A has a vapor pressure of approximately 10 bars (Engineer's Toolbox data) at 400 C. It sounds to me like a session at the chalk board is in order.

first problem: insufficient definition of the exact nature of what is the process involved, and the magnitude. 30 ton/hr (is that metric) = ~500 L/m, or is that English ton/hr = ~1000 L/m (assuming water density) and that is the peak-peak variance off the control point? Have you considered this could be the wrong type valve for your application? If the system pressure to the valve is high (300#), and downstream pressure is low, then perhaps your control position is not really 35% open, but is only 35% rotation, and the seat is not well cleared. IF turbulent conditions exist at the valve orifice, this could set up vibrations which could "chatter" the valve position.

Second problem: valve controller fails due to local heat. Here's your sign. If the controller is no good for the environment, then there's your problem. Change the environment, and the controller.

Something does not look right.

You say boiling point is 275 DEgc and normal temp is 400 Degc .

If the flow is showing a change of around 10% (140-125)/140 when you change the signal by 0.1 % something is not correct about your flow measurement.

If you are using an orifice plate you may be double extracting the square root .

Check the change in signal directly by connecting a manometer when you move the valve slightly.

While one might expect a constant flow when a positioner is in manual mode and the valve position is fixed, that's true only
- if all other flow conditions remain constant, and
- if you're not operating on a edge of the flash point at which liquid vaporizes to gas

PWSlack: <...boiling point of 257 deg C...Temperature is of the order of 400 deg C...>
One of them could be that the liquid is boiling as it goes through the valve.
James Stewart: Dowtherm-A has a vapor pressure of approximately 10 bars (Engineer's Toolbox data) at 400 C

I think these suggestions warrant serious consideration. I suspect that the temperature is higher than the operating pressure (or the operating pressure is lower than the temperature) needed to remain below the flash point.

Dowtherm's spec sheet gives the vapor pressure at 400°C as 10.62bara.

- What is the line pressure at the valve and at the flow meter?

- Are you recording the temperature and pressure so that you can track those with respect to the liquid's flash point?

- Is there some addtional load downstream that can come on-line which would reduce the line pressure which would drop the line pressure below flash point?

- Is there a diverter valve downstream that can alter the load conditions/line pressure?

Did this temp control loop have a history of working properly, or are you debugging a newly commissioned loop?

Certainly all of these questions are pertinent. It makes a huge difference to flow transmitters if the flow is single liquid phase, or two phase vapor and liquid.