CHAPTER 11 - ACOUSTIC FLOW MEASUREMENT

4. System Errors

Error sources for acoustic flowmeters are primarily related to the measurement of the average axial velocity. The main source of errors occurs in the determination of the acoustic path length, L, and the path angle, theta. The error in the velocity measurement is directly proportional to the uncertainty of these two variables. Care must be taken to minimize errors in measuring path length and angles. This care is especially necessary for chordal path meters because the computational procedures require accurate positioning of the acoustic paths. Likewise, errors in the cross-sectional area of the measurement section cause an error in the discharge measurement. This error can be a result of out-of-roundness or shape irregularities caused by temperature, pressure, structural loading, or deposits on the pipe walls. In circular pipe, cross-section dimensional errors can be reduced by averaging diameter measurements made at upstream, midsection, and downstream ends of the measurement section.

Another source of error occurs in measuring the transit times of the acoustic signals and in detecting the acoustic signal in the presence of electrical noise. Signal detection can also be hindered by signal modifications caused by changes in acoustical properties of the liquid which are caused by entrained air, suspended solids, and changes in temperature and pressure. Likewise, transducer fouling by algae or mineral deposits can reduce signal strength.

Secondary flows can create an error in the determination of Vaxial because the calculations are based on flow direction in the axial direction only. Secondary flow is caused by flow disturbances near the measurement section. These disturbances are typically caused by valves, elbows, or transitions. Secondary flow problems can be avoided by careful selection of the measurement section. For small pipes (diameters less than 36 in), 10 pipe diameters of straight pipe upstream and 3 to 5 diameters downstream from the measurement section should be sufficient. For pipe diameters greater than 36 in, 20 to 30 pipe diameters of straight pipe upstream and 3 to 5 diameters downstream may be required to obtain an acceptable velocity profile. If the measurement section must be placed near a bend, secondary flow errors can be reduced by orienting the acoustic paths perpendicular to the plane of the bend and locating the transducer as far downstream as possible. Likewise, another solution is the addition of another acoustic path which crosses the first path. Exact cancellations of secondary flow errors can be accomplished using a cross path configuration. The diametrically reflective path provides a cross path directly. Other variations in velocity profiles, due to Reynolds number effects and pipe wall roughness, can be corrected using a velocity profile correction factor. This correction factor corrects for the difference between the actual velocity profile and the profile assumed in the flowmeter's calculations. In general, deviations in velocity profiles are best accounted for by increasing the number of acoustic paths.