November 17, 2020
On November 17, 2020, the Ministry of Economy, Trade and Industry (METI), in collaboration with the Fire and Disaster Management Agency (FDMA) and the Ministry of Health, Labour and Welfare (MHLW), compiled Guidelines for helping plant operators to determine the reliability of specific AI for use in petroleum and chemical plants and a “Collection of Case Examples of Leading Companies Introducing AI into Plants” that explains key points in successful introduction of AI into plants by leading companies and their achievements. These Guidelines and Collection are expected to facilitate the safe introduction of AI into the maintenance and operation of petroleum and chemical plants and to help people engaged in plant safety to resolve existing and future challenges, through developing measures for overcoming labor shortages and for reducing the number of accidents.
Utilization of AI may become a way to dramatically sophisticate the safety of petroleum and chemical plants. Appropriate utilization of AI will help plant operators to analyze hundreds of pieces of real-time sensor data and discover small signs of abnormality at plants or to see images of pipes and judge their corrosion based on uniform criteria, etc.
Despite these advantages, the number of companies which have introduced AI into their plants is currently limited because of the following factors.
As plant operators are rigorously required to ensure safety in plant operation, they are required to appropriately assess the effectiveness and safety of using such AI; however, currently there is no official methodology for determining such performance for AI.
Some plant operators hesitate to introduce AI into their plants since the returns on investment in introducing an AI are as yet unknown and they may face significant challenges in the process of such introduction.
Measures in addressing Factor 1: Guidelines on Assessment of AI Reliability in the Field of Plant Safety
The Guidelines present approaches to appropriate management of the reliability of AI, i.e., securing the quality of AI expected to improve the safety and productivity of plants, exclusively for the field of plant safety. Plant operators’ utilization of the Guidelines is expected to bring about the following effects.
- Plant owner companies will be able to improve the safety and productivity of plants by making use of highly reliable AI and also smoothly fulfill accountability to internal and external stakeholders by demonstrating the reliability of AI.
- For vendor companies, if the AI that they develop satisfies the requirements stipulated in the Guidelines, the reliability of their AI will be much easier to explain to potential customers.
- Both plant owner companies and AI vendor companies will be able to appropriately define the terms and conditions that are necessary for the development and use of the AI.
Measures in Addressing Factor 2: Collection of Case Examples of Leading Companies Introducing AI into Plants
The Guidelines present measures for addressing challenges in relation to AI reliability that plant operators may face “in the process of introducing AI.” Meanwhile, the Collection of Case Examples provides information related to both “determining investments in AI” and “developing AI,” which are the steps that are taken prior to the measures in the Guidelines.* Specifically, the case examples of leading companies explain [i] effects of AI and [ii] methods for solving typical challenges in introducing AI, e.g., shortages of human resources with expertise in AI and difficulties in setting goals.
*Note: Appropriate assessment of the reliability of AI requires companies to start such assessment in the early stages of the project for introducing AI. Moreover, the case examples in the Collection do not comprehensively explain methods of assessing the reliability of AI. Accordingly, METI recommends that plant operators refer to the Guidelines from the early stages of such projects in order to ensure safety in introducing AI.
METI has been promoting smart safety in petroleum and chemical plants, aiming to provide solutions to social and infrastructural issues, such as labor shortages and highly aged facilities, by making use of new technologies. In July 2020, it jointly formulated the Action Plan for Smart Industrial Safety in the Field of High Pressure Gas Industrial Safety with companies that are leading the field. The Action Plan presented the formulation of the Guidelines and the Collection of Case Examples as efforts by the public sector. It organized the actions taken by the private sector in introducing AI in “Section 4. Sophistication of Decision Making.”
The Guidelines and the Collection of Case Examples will be published on the website operated by the Liaison Council of Three Ministries on Disaster Prevention of Petroleum Complexes and Other Facilities.