Nova Transportadora do Sudeste S.A. (NTS) is a TSO in the Southeast Region of Brazil, responsible for a 2,060 km network with a capacity of 70 MMm3/d. This network connects the most industrialized region of the country to the main offshore production fields of the Brazilian Pre-Salt and operating all assets remotely.

To enhance its operations, NTS's Operations Department launched the Logistics and Operations Analytical Data (LOAD) program, featuring 12 operational optimization projects. The first one, LOAD-METAGUS, has been conceived and carried out with the main objective of clearly identify the potential and critical offenders of the measurements of all metering systems that directly affect volume of gas to be billed, including Unaccounted for Gas (UAG) balance and errors in Operational Losses monitoring. The program also positively impacts the correct accounting of the volume of atmospheric emissions, in convergency with ESG requirements.

To achieve this objective, 6,722 main operational variables tags were evaluated, the balancing behaviour of 12 gas pipelines balances, measurement systems of 47 delivery points and 5 supply points, as well as 16 types of data categories were statistically studied. Thousands of field events recorded in the Control Room's logbook management system were also evaluated.

In this way, a self-refined algorithmic model based on machine learning was created to explain and stratify the UAG by pipeline and indicate possible operational losses to be purged from the hypothetical UAG (due to the impact on ESG). The application of this algorithm made it possible to cross-reference all data from the field measurement systems with the records in the Control Room’s logbooks, in addition to generating guidance for maintenance actions related to the meters by pipeline.

This paper aims to show the project results, discussing the gains and performing a critical analysis of the results obtained up to this stage.