Daikin Industries and Hitachi Ltd have begun a new trial to commercialise an AI agent that supports equipment failure diagnostics in factories.
Known as the AI Agent for Equipment Failure Diagnostics, it is part of a broader project which includes the creation of a factory with zero equipment breakdowns.
This is a factory that never stops and is part of a next generation production model that includes the digitalisation of brazing work and the digitalisation of fluorochemical reaction processes.
But back to the trial at hand, which is taking place at the Sakai Plant-Rinkai Factory (Sakai City, Osaka Prefecture) where Daikin manufactures commercial air conditioning equipment. The aim of this specific trial is to commercialise an AI agent that supports equipment failure diagnostics in factories.
This initiative takes the OT knowledge cultivated by Daikin and Hitachi at their respective worksites over many years and unites it with Hitachi’s advanced IT.
When a maintenance technician armed with a tablet device discovers a failure in a pump or valve while inspecting production equipment, the "AI Agent for Equipment Failure Diagnostics" will indicate the cause and needed corrective actions to the technician.
The system works by first converting the factory equipment drawings that Daikin has accumulated, such as various production and utility equipment, into knowledge graphs that the generative AI can read.
Then, the generative AI learns the OT data, which includes knowledge graphs, maintenance records, and OT skills, as well as Hitachi's unique equipment failure cause analysis process which is based on STAMP.
This achieves the failure diagnostics that is on par with or exceeds the capabilities of Daikin's standard maintenance technicians. A preliminary demonstration test confirmed that the "AI Agent for Equipment Failure Diagnostics" could identify the cause of equipment failure and recommend corrective actions within 10 seconds at an accuracy of over 90 per cent.
Daikin and Hitachi plan to complete the trial operation by September 2025 before beginning practical application of the AI agent.
By developing it to Daikin's manufacturing bases both in Japan and overseas, the company intends to share its tacit knowledge of equipment maintenance as organisational knowledge in order to ensure global quality, the transfer of skills, and improved productivity of frontline workers.
For its part, Hitachi will adopt the "AI Agent for Equipment Failure Diagnostics" as a Lumada solution throughout the manufacturing industry and provide applications combining OT knowledge and IT into fields other than equipment maintenance.
Hitachi will also help frontline workers improve work efficiency by including services such as the "AI Agent Development, Operation, and Environment Provision Service, which rapidly develops and generates AI agents that utilise OT knowledge.