This presentation explores the innovative capabilities of LLM augmented with Reliability Centered Maintenance that leverages Retrieval-Augmented Generation (RAG) to provide intelligent responses and actions based on a comprehensive array of asset documents, including battery test reports, work orders, and asset manuals for informed decisions and asset information search. Advanced custom Optical Character Recognition (OCR) methods have been implemented to extract data from particularly challenging tables, ensuring high accuracy in information retrieval. The data ingestion process is efficiently managed on Databricks, allowing for seamless integration and processing of large datasets. A custom frontend developed by Microsoft facilitates user interaction and testing. Utilizing Microsoft Promptflow, the chatbot adeptly handles user queries while formatting responses to deliver concise, context-rich information. This cutting-edge solution enhances operational efficiency and supports informed decision-making within the organization.
Download “Unlocking Asset Data: RAG-Enhanced LLM for Efficient Asset Management” Asset-Health-GenAI-NYPA-June-2024.pdf – Downloaded 4 times – 1.85 MB