Can We Automate Data-Driven Evergreening of API 580 RBI?

This paper presents ways to partially automate risk-based inspection (RBI) evergreening using statistics and computer algorithms. The computer work included in this undertaking uses synthetic data—generated to mimic real inspection results, based on extensive experience with inspection data from several sources. This approach applies analytics and statistical methods to inspection and operational data to predict the future state of equipment based on historical inspection data and operational data. The objective is to fit this approach within the construct of one of the most widely used RBI standards, the American Petroleum Institute Recommended Practice 580, "Risk-Based Inspection" (API 580), to show it is possible to automate evergreening of the RBI and enhancing the risk control process. The applications focus on integrity management of topsides pressure equipment, where the most used standards are lagging far behind what is possible, given the amount of data available and the capacity of modern, computer-driven tools for leveraging the data to update the risk assessment. The result is a new approach that allows the existing RBI for topside pressure equipment to be updated using computerized tools. The observations regarding the application of this approach provide sufficient data to demonstrate the appropriateness and effectiveness of computerized tools, including those that capitalize on feedback loops, to gather information and use it to inform updates to the RBI. While this approach adheres to the API 580 standard, it is important to note that there is still room for subjective judgement in the standard. This approach paves the way for data-driven risk management and has the potential to drive modernization of existing standards so such ways of working are fully incorporated. Demonstrating the effectiveness of this practical automated solution lays the groundwork for updating standards like API 580 to be more data driven and enable the use of computational tools such as statistics, analytics and AI, making them acceptable in executing RBI processes.