Add new comment

Part 1 of 3 Parts   
      Utility engineers usually spend a lot of their time carrying out repetitive administrative tasks. There are estimates that highly trained engineers might spend as much are forty percent of their time on mundane paperwork. The development of artificial intelligence (AI) techniques such as deep learning, natural language processing (NLP) and intelligent automation have made them increasingly valuable for taking over and automating complex and important tasks. It is believed that AI has great potential in reducing repetitive tasks which would allow engineers to spend more time concentrating on engineering work.
     Ontario Power Generation (OPG) is based in Ontario, Canada. It was established in 1999 as a Crown corporation completely owned by the government of Ontario. It has a varied portfolio of more than sixteen gigawatts of capacity which represents about half of the electricity generated in Ontario. This includes almost six gigawatts of nuclear power capacity. They recently demonstrated the potential use of AI during one particular stage of the planning for nuclear unit outages. Such planned maintenance outages often require two to three months per year at OPG nuclear facilities.
      Each nuclear outage requires that twenty thousand to twenty-five thousand separate tasks be scheduled. Most of these tasks are the same as those carried out in previous outages. This task scheduling requires a team of highly trained outage personnel to review procedures. They also manually search for generic tasks on previous outages to fill in the new schedule. The OPG outage planning process currently contains more than forty major milestones that begin more than two years before the outages of any nuclear unit officially begin.
     Industry best practices related to safety, reliability, scope and duration are used as the basis for the outage milestones at OPG. They also have to take into account needs that are specific to a particular nuclear facility. During the execution of these milestones, there are four points at which the schedule for the outage is revised. Each of these revisions is more detailed and comprehensive than the previous revision except the first.
     The Outage AI solution software employed by OPG is aimed at predicting the logical connections for tasks in the outage schedule. It creates the first of the schedules with all of the tasks included. This use of AI in the planning process reduces the need for manual effort while it also ensures that oversight contingency is maintained during the revision process for subsequent schedules. This helps to mitigate both risk and the possible requirement for extensions of the duration of the outage.
     The Outage AI software at OPG is a custom, cloud-hosted application. It integrates smoothly with the existing IT infrastructure at OPG. It leverages AI, deep learning, NLP and intelligent automation. The object is to predict the breakdown of the work structure of up to twenty-five thousand tasks including the previous task and the subsequent task. The Outage AI software automatically schedules these tasks within the upcoming nuclear unit outage Revision ‘B’ schedule.
Please read Part 2