Meter Factor Selection using a Multinomial Figure of Merit Algorithm
Proceedings Publication Date
Flow meters are used in pipelines to measure the rate of flow of the fluid. There are certain properties that affect these measurements, including temperature, pressure, gravity, flow rate and mechanical factors of the meter itself. Some of these properties can be measured and used to adjust the flow meter results using mathematical calculations. To calibrate a flow meter more accurately a process of “meter proving” is used. Meter Proving uses a fixed volumetric length of pipeline that can be compared to the meter flow measurement. The ratio of fixed volume to measured volume is called a meter correction factor. This meter correction factor is used in conjunction with measured fluid properties to account for mechanical errors that cannot be directly determined. In an ideal pipeline, meters can be recalibrated using “proving” whenever flow conditions or fluid properties change. Practically, there are times when the volume of liquid is too small to gather enough information to create a new meter factor. There are also economic limitations where a meter prover is shared between meters such that it cannot be run as often as desired. In these cases a historical meter factor can be used to calibrate the meter. General Electric has developed a novel algorithm to select the best historical meter factor using a multinomial figure of merit. This algorithm first uses a selection process that eliminates any inappropriate meter factors, and then uses a figure of meter calculation to rank the remaining acceptable meter factors that are suitable for current conditions. By using this process, the accuracy of pipeline metering is improved, which can have large economic benefits.