Automated performance assessments on gas metering stations predict upcoming meter failures and save cost by avoiding mismeasurements, optimizing recalibration schedules, and avoiding meter downtime.

SHARE:

Gas meters in general, and ultrasonic gas meters in particular, can and need to be monitored to ensure performance. Mismeasurements require intensive manual corrections and can lead to increased grid imbalance and loss of revenues.

Monitoring each gas meter individually can be laborious: a manual job for most operators today. Moreover, operating a portfolio of meters from different vendors requires running several different programs to monitor each meter type individually. Only after inspecting every single meter, can an operator identify those that are facing performance issues.

Automated meter population performance assessment and alerts

DNV has launched the MECADA platform (Metering & Calibration Data Analytics), which monitors and assesses the performance of an entire meter population of Ultrasonic Gas meters from their diagnostic data. Monitoring of Gas Chromatographs and Pressure/Temperature sensors are included as well, and with that, the online comparison of measured vs. calculated speed of sound as a powerful diagnostic.

MECADA is a single platform where all metering station data is combined into a single dashboard. When issues arise, the operator is alerted and the particular meter is flagged on the dashboard. If there are no issues, the operator can check the dashboard and focus on their day-to-day duties.

Benefits overview:

• Advanced visualization tools for the entire meter population, with the possibility to zoom-in on individual meters. This offers a single interface to analyse the performance of all meter brands/types and their auxiliary equipment (GC, P, T).
• Metering dashboard with alert functionality, saving time by visualizing the state of an entire meter population in one single view.
• Share a custody transfer meter between the buying and selling party, solving the issue of transparency for the meter owner and the issue of auditability for the other operator.

Planned features of future releases:

• The ability to temporarily share a meter on the platform with the manufacturer, to facilitate easy troubleshooting by service engineers.
• A benchmarking option to anonymously compare meter performance with other operators, to gauge whether performance should be improved.
• Live prediction of meter error (%) during field service as a final goal, by using a top-down data-driven Machine Learning model.

MECADA models will use meter performance data from all participating operators to train machine learning algorithms on DNV’s VERACITY open industry platform. By using big data from multiple operators, the model’s performance can become better than any individual operator can achieve on their own. The model performance keeps improving as more data comes in.

Source: Simulated dataset for demonstration purposes