Using Augmented Intelligence to Automate Subsea Inspection Data Acquisition, Processing, Analysis, Reporting and Access

Augmented Intelligence (AI2) involves fusing Analyst Intuition with Artificial Intelligence to deliver an optimised combination of human-machine decision support. AI2 is being incorporated by i-Tech Services / Leidos into the physical inspection of offshore Oil, Gas, and Renewables assets, delivering valuable data driven insights that contribute to greater efficiency, enhanced condition monitoring, improved asset integrity and asset life extension. The deployment of vehicular and diver assets to obtain such inspection data, with associated support vessels, remains a major cost challenge for Operators. We believe the industry needs to approach this challenge from two key directions. Firstly, through the application of autonomous systems for data acquisition and delivery, reducing vessel reliance, and secondly through automating the acquisition and processing of data and maximising the insight provided by the data. This paper will examine the use of Augmented Intelligence to optimise the Subsea Inspection data workflow as a key use case, to demonstrate the principles. The historic paradigm consists of a fragmented evolving approach, with insufficient consideration and design across all the sensors, processing analytical engines and data visualisation. The approach being adopted is to closely link all aspects of the data workflow, within the context of delivering the data and beyond in terms of harvesting additional insight and value. To achieve the optimum workflow a number of developmental initiatives are being knitted into a modular platform, each element providing standalone value but the sum of the parts generates the most significant value and cost reduction. The elements being combined are automatic data quality control at acquisition source and through the full workflow, automated processing, machine vision for object recognition and reporting and machine learning to optimise the system intelligence. All of these are designed to augment the expertise of the analyst / user, detecting change to learnt parameters, by using real time data and critically by referencing large historical data sets and as-built data. The outputs from a system holistic approach will be improved data acquisition with more efficient high quality right first time data reporting. In addition layers of analytics, with smart, intuitive data access and retrieval will optimise delivery of key information within large data sets, together with maximising value and insight.