Condition-Based Calibration of Flow Meters Through Data-Driven Modelling

Case Study

Case Study

Sector Challenge

By implementing a condition-based calibration (CBC) schedule, end-users, in theory, have the ability to reduce costs through a more dynamic operating pattern. This is achieved by the continuous monitoring of key digital diagnostic and standard instrument outputs, which are then compared to known baseline conditions. In addition, with enough historical data, CBC systems can be used to predict device calibration drift over time and correct for this.
However, in reality the performance of a given meter will vary between manufacturers, as will the availability and relevance of the digital ‘diagnostic’ data. There is also a great number of physical factors to account for. For example, in a hydrocarbon production operation there will most certainly be pipe bends and valves as well as pressure and temperature effects, all of which are known to influence the performance of flow meters to different extents depending on type and manufacturer.

How Can TÜV SÜD National Engineering Laboratory Help?

Condition-Based Calibration InfographicBy partnering with a Coriolis flow meter manufacturer and through the use of our flow laboratories, we have conducted experimentation designed to intentionally induce errors in a Coriolis meter’s flow measurement value. By logging all the available digital variables from the meter transmitter throughout testing, a detailed picture of how each individual variable responds as a result of the error state was obtained. This data was then used to develop a data-driven model, capable of assessing the hidden correlations between the variables under each scenario and therefore alert an end-user to the fact that not only was there a measurement error but identify the cause of said error. This information was then communicated through simple correlation diagrams such as those above which match the prediction to reality.

Sector Impact

By partnering with a meter manufacturer, the models developed by the Digital Metrology team have the experience of both parties built in. By utilising these techniques, end-users are empowered with a greater understanding and confidence in meter performance without the need to be a Coriolis flow meter expert and by extension the need to be adept at interpreting multiple screens of individual diagnostic trends. We can create tools for users that provide simple end-user functionality, such as the ability to identify error states quickly, that in turn significantly reduce plant operator financial exposure while increasing overall operational efficiency.

Next Steps

Site Selector