Pipeline operators and researchers face persistent challenges in detecting, analyzing, and learning from pipeline leak events due to fragmented data, inconsistent schemas, and limited mechanisms for secure data sharing. Operational time series, physical pipeline attributes, fluid properties, and contextual incident details are often stored in disconnected formats that are difficult to integrate, reproduce, or reuse for simulation, analysis, and research. These challenges are compounded by industry concerns around data privacy, security, and loss of control when sharing sensitive opera-tional information,...
Pipeline operators and researchers face persistent challenges in detecting, analyzing, and learning from pipeline leak events due to fragmented data, inconsistent schemas, and limited mechanisms for secure data sharing. Operational time series, physical pipeline attributes, fluid properties, and contextual incident details are often stored in disconnected formats that are difficult to integrate, reproduce, or reuse for simulation, analysis, and research. These challenges are compounded by industry concerns around data privacy, security, and loss of control when sharing sensitive opera-tional information, limiting collaboration, and slowing progress in leak detection methodologies.
This document presents the design and implementation of a comprehensive Leak Detection Data-base Repository Framework intended to standardize how pipeline system and incident data are archived, managed, and shared. The framework defines a normalized relational data model encom-passing core pipeline entities including companies, pipelines, nodes, externals, pipes, profiles, valves and valve curves, pumps, sensors, fluids and fluid properties, time series, data quality indi-cators, and leak events with clearly defined primary and foreign key relationships. A unified mod-eling language (UML) entity‑relationship diagram is provided to visually convey the system ar-chitecture and interdependencies that support leak detection analysis and pipeline simulation, in-cluding real‑time transient models.
The primary audience for this document includes leak detection stakeholders and researchers re-sponsible for capturing and managing pipeline operational and incident data. Secondary audiences include academic researchers, leak detection algorithm developers, software vendors, and industry consortia members seeking standardized, high‑quality datasets for modeling, validation, and inno-vation in leak detection and pipeline monitoring. In daily operations, practitioners can use this framework to consistently archive leak events along with the full physical, operational, and con-textual description of the pipeline at the time of the incident.
This zipped deliverable contains the following items:
1. PR764-243904-R01 Leak Detection Database Repository Framework: Final Report. 2. PR764-243904-E01 APIMLX Pipeline Data API Documentation.html: Document with docu-mented code, example workflow of doing ETL on data, and creating database. 3. PR764-243904-E02 SQL Coder DB Integration.html: Document example of local LLM use of gptPipeline. 4. PR764-243904-E03 Stored Procedure.html: Document with stored procedure to query all data for specific company and pipeline.