Flow Measurement Using Load Cells


Thread Starter

Juan Pardo

I need to measure the flow that is coming out of a vessel using only a load cell that measures the weight of the tank. The objective of this measurement is to control a proportional valve that is controlling the output of the tank, which flows by gravity through a pipe below the tank.

Anyone has experience on this? Any thoughts? Thanks!

If you want to write directly, here is my email
[email protected]
That method (loss of weight, flow measurement) is not often used in automatic control.

It requires a sampled data control algorithm with a calculated process variable, although you might be able to do it with integral only control. In any event, the response time for stable control is very slow.

The level of control would be comparable to manual control of the valve based on the rate of weight loss.

Good luck
I agree with "d". This is not the way to go. Rather install a clamp on ultra-sonic flow meter on your outlet pipe and use a local hardware flow controller to control your control valve if you don't have a PLC or supervisory system close by. Speak to your supplier on which sensors to use on the USFM since depending on your product one type might be more suitable than others. If you have something like explosive emulsion, take your time in selecting the right sensors since ultra-slow and thick liquid measurements are very difficult to do.

Once you have a good reliable flow meter going, the PID flow controller will control the FCV as per your set point and will open and close to try and keep to your set point. The reaction time of your FCV can then be adjusted to suit your process. Start with a fast tuning set and work your way down until you have a smooth operation.
You didn't say what the process material was so I assumed it was liquid. Working with solids and gravel in such an application is very different from what I have described above. With solids life becomes a nightmare for instrument technicians and engineers. Liquids are easy to measure and control so for your same I hope your process is some form of liquid.
I used that method with good results on a small pilot autoclave.
The slurry flow was too low to measure with a mag flowmeter.

From memory I used a 6 minute sample time.
There is a gravitic density meter that measures 3000 samples per second with a 0.33 second response time. It's 99.5% accurate, non-nuclear, and in-line.

More information here: www.directdensity.com

Hope this helps..


The use of mass PV's is simple. I assume you are asking about actually installing the load cells. Every support for the tank, or the frame, can have a separate load cell which are summed. If you have a lot of motion, taking the mean and re-suming is a simple approach- high accuracy requires modelling (does inertia on one corner subtract from the other support corners or not?).

A very heavy tank can ignore piping (mass doesnt need to be accurate, just proportional to the flow), mid range can be modelled, light tanks would need flexible piping (the easiest in all cases).

As far as the signal, it's likely to be noisy, so smooth it as much as possible relative to the response time you need. Your PV and SP would be in kg/m or similar, the slope of the mass readings.