Differential Flow totalizer

Hi Folks,

I have searched the Internet high and low for an answer to this question. If it has already been asked please accept my apologies.

I have a flow totalizing system that works like this:

Coriolis flow meters measure the flow feeding into and returning from a closed loop water system. When an operator opens one of the feed points the PLC integrates the difference in feed flow vs return flow. This is then used as the total flow over a given time.( I hope I made this clear).

My questions are:

Have you seen this approach before.

Could you help me explain what would cause a 30 % error in totalized flow. The only variable that has changed is the flow rate of the loop. When I last tested the errors were negligible. They are now huge +25-30%

I can't put up the ladder logic right now because I am not at work. If anyone is interested in helping me I will upload it tomorrow if need be.

Thanks
John

I'm a little foggy.

"Closed loop" infers to me that 'what goes in' is the same as 'what comes out'. In other words, no diversion flow paths within the loop. But not in your case?

Why wouldn't your inlet flow = outlet flow (averaged over time)?

Yet someone saw fit to install two coreolis meters, meaning that measurement of inlet must have been considered to be different than outlet flow; otherwise why not just monitor a single flow rate, at either the inlet or at the outlet?

I would think that integrating the difference between inlet flow and outlet flow would yield zero:
(Total closed loop inlet volume) minus (total closed loop outlet volume) = zero.
or total (inlet flow - outlet flow) = zero, since the integral of zero is zero.

>When I last tested the errors were negligible.

What did you test? Flow rate on each coriolis, independently? What are some of the values you got?

What shows 30% deviation? instantaneous flow rates or the totalized value?

What is calculating the total, the coreolis meters or the PLC?

The system is a USP water supply system that feeds six feed points. It runs at a constant but changeable flow rate. The Coriolis meters measure flow to the six point loop and flow returning from them.

If one (or more) of the delivery point valves is opened the return flow is then less than the feed flow. This difference is integrated.

So, if the static flow rate is 30 GPM and a feed point is opened and 10 GPM goes to the feed point 20 GPM goes back to the return meter. If the valve is held open for 10 minutes than 100 Gallons of water was delivered.

Of course only one feed point can be open at any one time or the results are indeterminate.

I tested the accuracy of the totalizer. I filled one of the reactors with 25 gallons (Based on the totalizer) and weighted it. In Feb of last year it was spot on. I repeated this test several times.

Friday I tested it again and it was off by 27%.

The only thing that changed was the flow rate of the loop. It was in the low 30 GPM range and now it is in the high 30 GPM range.
The totalizer is what is showing the error. I believe the flow-meters are fine.

The total is being calculated by the PLC.


John

jpocreva said:

The only thing that changed was the flow rate of the loop. It was in the low 30 GPM range and now it is in the high 30 GPM range.John

Click to expand...

How are the flow rates being read into the PLC? Analog Input? Maybe by raising the flow rate from low 30s to high 30s, you are out of the scaled range for the analog signal. You could also be above the range of the flowmeters. There could also be an issue with the PLC calculation for the totalizers.

What kind of PLC?

How is the integration being performed? (i.e. what is the specific instruction/method being used?)

In my experience, if you are simply totalizing discrete amounts over time (summing flow_rate_per_minute every minute), that can lead to not-insignificant errors.

"The totalizer is what is showing the error. I believe the flow-meters are fine.

The total is being calculated by the PLC."

Either u r using a Totalizer at Field or must be done at PLC Control Panel .
PLC Ladder Logic uses Compare blocks through which they get Interlocked with process drives and Flow integrator Display (Math Blocks )are used as to show the Total Flow.

The approach may be there(not new) , however the problem faced can be rectified with schematic diagram.

Flow Totalizer solved

For the historical value I will post the solution to this one.

Thanks to all for the replies.

The flow rate is via 4-20mA. The system is running right at the upper range of the AI. So, when a valve is opened the flow increases out of range and viola we are getting a higher flow rate then we think. This causes the system to overcharge.

Thanks again for all of the help.

John