steam flow measurement vs. estimation 2

[andrejsio]

hi all,
I am trying to model the steam consumption of different unit operations in the chemical plant; as a verification, I plan to make the steam consumption measurements for selected unit operations; I want to install the steam flowmeter into the pipeline, but since this is pretty tedious and risky task, I am thinking about calculating the steam flow from the control valve-opening percentage values, which are accesible from the main control system (knowing steam pressure, pipe diameter and estimating the Cv of the valve)
does anybody have practical experience with this? some quantitative information (delta measurement-estimation) would be of great help
thanks,
andrejsio

 

[sshep]

andrejsio

What you are suggesting is technically possible. In practice it won't be a very accurate measure, but is will be fairly reproducible unless there is a problem at the valve (erosion, partial plugged, wet steam, etc). I have a full open valve now only passing 8ton/hr instead of the expected 18ton/hr- without steam measurement I would struggle to know that something was wrong with the valve station.

I see you left downsteam (chest or destination) pressure out of your list. Perhaps in many cases your control valve is running with choked flow, but you can see the calculational challenge of this discontinuity in the flow relation.

As an alternative to measuring the steam flow directly consider an energy balance using data on the process side. Also some cases maybe you can measure condensate flow.

best wishes,
sshep

 

[dcasto]

Here are some thoughts:
You could measure the condensate from the unit in liquid form. A clamp on ultrasonic could work on high pressure steam. Look at Boiler feed water rates. If there are non condensing turbines, look at an orifice meter on the outlet side where the pressure is lower.

 

[andrejsio]

sshep and dcasto thanks a lot for your valuable contributions
I have actually the measurement of the steam consumption for very limited number of unit operations and I am trying to model the theoretical steam (or in general heating utility consumption) based on the process data acquired in defined time interval (temperature, mass inside the unit e.g. reactor) and physicochemical data (heat capacity of the heated substances, evaporation enthalpies of the distilled solvents...)
the first results are showing, that the losses (defined as delta measurement - model) are as high as 20-40% of overall steam consumption; next step was to include the losses to the model and after playing around with complicated equations of heat transfer, I found realtively easy equation, describing the losses very accurately (in this case batch-operated reactor):
losses(i) = a * (theoretical consumption(i))^b
where i is the particular time interval and a and b are fitted coefficients and are probably very dependent on the heating/cooling system type (in my case the half pipe coil with circulating hot water heated with steam injection was used)
I know that it would be convenient to have more generic equation describing the heat transfer phenomena, but this would be too complicated in my case, beacause I would like to make the energy-consumption model for the whole production building (more than 100 unit operations)
so I planned the steam consumption measurements on the subset of the unit operations in order to test and improve the models;
clamp-on measuring device would be the best solution for me, but unfortunately it won't be accurate enough for 5 bar steam which is used (information from manufacturer, if you have some other experiences, please let me know)
another possibility for hot water circulating system would be to install two clamp-on devices - before and after injection of the steam and from the difference calculated the steam consumption...
if you have any ideas and comments, please let me know
cheers,
andrejsio

 

[JimCasey]

It >is< "Possible" to measure flow with a control valve. THe Valtek Starpak does exactly this by measuring P1, P2, and T all inside the valve body, Each Individual valve plug is calibrated for Cv vs travel and that curve is loaded into the electronics, along with the sizing program. Thus it is possible to read and control flow all in one component. It works well but it is not cheap.

You might accomplish your goal with an energy balance, possibly of the process fluid instead of the steam. If you know the flowrate and the Delta-T, of your process fluid, you can look up the specific heat and calculate BTU per hour. Divide by the steam difference in energy at the statepoints entering and leaving the HX (It'll be close to 1000 BTU/lb) and there's your steam flow rate. If the HX is uninsulated, add abot 15% to the steam flow needed.