The Evolution of Data Management to Empower Integrity Management Decisions: A Case Study of an Enterprise Implementation

Abstract

BP’s Natural Gas Liquids business unit (NGLBU) has conducted integrity investigation and mitigation activities on its pipelines and has been following this best practice for numerous years. In recent times, NGLBU’s data management initiatives focused on establishing an enterprise Geographic Information System (GIS) coupled tightly with a derivative of the Pipeline Open Data Standard (PODS) data model. During successful implementation of the GIS, an analysis identified gaps in existing data management processes for pipeline integrity information. Consequently, the business unit adopted Baseline Technology’s Pipeline Information Control System (PICS) and its modules to support the pipeline integrity decision-making process on its 9000km of pipeline. The PICS implementation leverages the existing GIS implementation while addressing a number of unresolved data management and integration issues, including: • Integration of inline inspection with excavation results; • Migration of above ground surveys to a common repository; • Integration of multiple inline inspections; • Facilitation of corrosion growth modeling; • Structured process for prioritization of remediation; • Structured process for integration of inline inspections with risk parameters; • Defined data collection, storage, and integration standards. Data management solutions based solely on a GIS require pipeline surveys without explicit positional information to be converted into a common linear reference system (typically chainage or stationing) such that disparate data sets may be overlaid and compared. This conversion, or spatial normalization, process is where much of the data management effort is spent and is often prone to error introduction. Even when small errors are introduced, the normalization process is often performed such that it is not auditable. If the underlying spatial errors are not reported, addressed, and understood, the value of the data integration and any subsequent analysis of the combined data set is questionable.