To clarify, what I am particularly interested in is how to select the appropriate calibration range for the lower range transmitter.
I expect to calibrate the higher range transmitter to go from zero to the flowmeter full scale differential pressure. The lower range transmitter will then be calibrated from zero to an upper range value that is some fraction of the flowmeter full scale dp. The control system continuously reads both and selects the lower range transmitter when the flow is less than some value at or near its upper range. When flow is higher than that value, the system selects the higher range transmitter. When flow is above this transfer point, the lower range transmitter will be pegged upscale, but the reading will be ignored. That's how I think it works. Looking for confirmation of this method and help establishing the lower range transmitter calibration range.
Thanks.
And BTW: That is really old technology that goes back to the Foxboro Pneumatic transmitter days.
Modern instruments such as magnetic flowmeters for liquids or vortex-shedding meters for gases or low viscosity liquids have vastly greater rangeability than orifice plates. Stacking DP transmitters on orifice plates I really wouldn't trust a range greater than 10:1, while mags or vortex meters reach 100:1.
I agree with JimCasey - and the stacked dp transmitter concept still works. I would use at least five and perhaps higher as the range multiplication factor for the highly accurate smart dp transmitters.
Your post lacked specifics such as the line size and fluid phase etc. I like the turndown with Coreolis meters for smaller line sizes.
You are going to need a flow computer to do all this. I suggest you look at their literature first.
The flow computers I've used have up to three stacked DP's. I've only used 2 myself. We would put a 0 to 50" and a 0 to 250". The accuracy range is now 1" to 250" which is about 16 to 1 turndown. Adding a 3rd DP is 64 to 1.
The problem lies in the switch point and dead band between the transmitters. We had one station that spent most of the time running at 48" to 52". We had poor mass balance (over 1.5%) error. We switched to a 0 to 150" single tranmitter and got less than 1% closure on the system.
Kevin,
I guess you already have the flow element, and you are just trying to get the best rangeability from it? So replacing it with another type of flowmeter may not be practical if you cans shut down the line.
You have not given us any flow rates, pipe sizes, pressures or temperatures, so it is a little difficult to give a complete answer. But A word of warning. Flowmeter manufacturers claim turndown figures that are completely unuseable. For example vortex meters (which are velocity devices) are designes with a meter max of up to 130m/sec. They may claim 100:1 turndown, but on compressed air it is undesireable to to have high pressure drop so your ideal max velocity sould be 12m/sec. This would leave you with a useable turndown 12/(130/100) =9.2:1 No better than 2 DP transmitters! And their answer to this may be to reduce the pipe size and create a pressure drop that your compressor will burn fuel trying to overcome.
To give a reasonable answer to your question, Lets answer it in three parts.
Firstly you should split the range of the transmitters based on the turndown of flow that you require and not the turndown of DP. Orifice plates etc have no moving parts, so the square law will follow all of the way down to 0% (according to ISO5167). So the problem lies in accurately measuring the DP producds.
Secondly the transmitters you select for the upper and lower dp measurement should have different measureing cells, because the lower reange unit needs to have greater accuracy,sensitivity and stability.
Thirdly, the chageover point is normally the point that causes the problems. If you use a switching point, then you will get a step change in the signal. Some flow computers avoid this by phasing one signal in and the other out within a change over zone. The easyest way to achieve this is to scale the input signals to flow units before the changeover. Set up your phase zone and then use a weighted average of the two signals within the zone.
If you have a PLC or DCS this will be possible with some maths. If not you may want to go down the flow computer route.
I am thinking that stacking DP transmitters is ok. And I have done it for gases and liquids. And I agree for the most part with the 3 points but I would like to add that you have to watchout for density variation in the flow (liquid vs gas, and P & T variations), so you might want to stack multivariable transmitters instead of DP transmitters if you are measuring density variation plus Rd variations. Lower flows with lower Rd's deviate from square root relationship and multivariable transmitters do RG compensatation, like ISO5167.
