What is the difference between uncertainty and accuracy in flowmeters?

Feb. 2, 2016
Béla Lipták says the there is no such thing as a perfect detector, therefore we never know what "true value" is.

This column is moderated by Béla Lipták, automation and safety consultant and editor of the Instrument and Automation Engineers’ Handbook (IAEH). If you have an automation-related question for this column, write to [email protected].

Question

What is the difference between uncertainty and accuracy in flowmeters? Why do custody transfer ultrasonic flowmeter manufacturers state uncertainty value instead of accuracy?

A. Rahimi
[email protected]

Answer

Accuracy (or more precisely, "inaccuracy" or error) can be defined as the closeness of the result of a measurement to the true value of the measurand. Unfortunately, we never know what that "true value" is, because there is no such thing as a perfect detector. Uncertainty should reflect this, by using the term uncertainty as the sum of the inaccuracies of the instrument and of the reference. Unfortunately, some vendors play "specmanship" with these terms, and the worst ones don't even state the basis of their numbers—they don't state if the percentage they give is based on the full scale (FS) or the actual reading (AR) of the measurement.

As far as the performance of ultrasonic flowmeters, it depends on the nature of the process fluid (medium) and on the meter's “linkage.” The medium is the acoustically conductive process fluid through which the sound is traveling. Linkage is based on the integrity of the bond between adjacent mediums (transducer and pipe wall, pipe wall and lining if present, lining and process fluid etc.). From this, it can be seen that wetted transducer installations will exhibit superior results than "clamp-on" ones.

Now, if we disregard the reference error, measurement uncertainty is composed of two types of error: precision error and systematic error. The precision or random error is a function of the quality of vendor's hardware, which hopefully was determined in the carefully controlled environment of a flow lab. This is the error usually listed in the manufacturer's specifications. It is, however, impossible to replicate the carefully controlled conditions of the flow lab in the field. This introduces the second type of error, the systematic error. This second error component encompasses all factors that are related to the meter's actual installation. Some of these include: the pipe's outside surface, the pipe's inside surface, the bonding of the liner if liners are used, coating or buildup in the pipe, pipe eccentricity, roundness or ovality, pipe material, environmental factors and flow profile (Reynolds number), to name some of the more common ones. This systematic error can add up to 5% FS to the total uncertainty of the measurement.

So how good are ultrasonic flow meters? That depends on three design features:

  1. Wetted or clamp-on?
  2. Time of flight (TOF) or Doppler?
  3. Rangeability required

The accuracy of a wetted, TOF unit is about 1% FS. The reading accuracy at the actual flow is very much a function of the ratio between maximum and minimum flow to be detected (rangeability). Remember, at minimum flow, the Reynolds number must still exceed 10,000. So at 50% flow, the accuracy of the above meter is 2% AR (actual reading) and at 20% flow, it is 5% AR. If Doppler designs are used, the error can double, and when clamp-on units are used it can double again.

Béla Lipták
[email protected]