Pre-loader

Sensor Quality-Check and Value Correction

  • Client:

    National Oil Company
  • Field:

    NCS Oil Field

Highly accurate timeseries of production data provided by the BPT Twins - used as the essential input layer for Oil Company digital ecosystem

In brief:

For this specific field, initially the quality-check of production data was done with the traditional approach using the standard data validation embedded in the plant historical database. This included sensor-by-sensor considerations, mainly Out-Of-Range and Rate-Changes, as well as 1-Out-Of-2 or 2-Out-Of-3 sensor considerations (duplicate or triple instrumentation for critical measurements). This is by all means a relatively simple method of quality-checking sensor values provided by the field instrumentation, however still used across the oil & gas industry.

While the traditional data validation is based on a local sensor-by-sensor approach, the sensor quality-check and corrector function in the heart of BPT Digital Production Twins is holistically driven. For this specific field, the BPT Twins using all the production related sensor data from the subsea production system and topside processing facility, approx. 250 and 2500 sensor values respectively, as input received from the plant historian (OSI PI).

The BPT patented reconciliation solver provides quality assured, accurate and consistent data. As part of the solver, one of the industry-standard 1st principle process simulator models covers the entire oil & gas production facility with high-fidelity thermodynamics throughout. The output of the solution is complete consistent datasets of best-fit sensor values. The solution identifies faulty or bad sensors and provide replacement values.

The complete dataset is automatically updated frequently. The BPT Twins dataset is used as the essential input layer for the corporate designed and operated digital ecosystem consisting of a wide range of digital applications (built on machine learning, artificial intelligence, advanced control approaches, etc.) instead of using the raw data from the field instruments directly.

The operator and BPT have spent considerable effort to benchmark the quality of the corrected sensor values with the conclusion that accuracy is significantly improved and could be used with high degree of confidence.

The trend plot above is an example from the field solution showing a temperature sensor output (blue points and trend curve) versus the quality-checked and corrected output from BPT Twins (orange curve). As could be seen, the sensor output value is unreliable in more than 50% of the time, while the twin provided replacement values has been proved to match field performance with high accuracy.

Benefits and key findings:
  • BPT Twins using all the production related sensor data from subsea production system and topside processing facility as input in a superb holistic approach compared to traditional data validation in plant historian
  • The output of the BPT Twins is complete consistent datasets of best-fit sensor values, automatically updated frequently
  • BPT Twins detects faulty or bad sensors and provide replacement values
  • The BPT Twins generated dataset is used as the essential input layer for Oil Company corporate-wide digital ecosystem, feeding a wide range of digital applications with reliable and high-quality input data

Back to References Gallery