AI/ML Maintenance On-Prem Platform

OUSD (R&E) CRITICAL TECHNOLOGY AREA(S): Trusted AI and Autonomy; Advanced Computing and Software; Human-Machine Interfaces OBJECTIVE: A stand-alone, scalable asset management software providing interactive local logic memory (LLM) knowledge base capability and AI/ML module for guided troubleshooting and maintenance of critical IPE (Industrial Plant Equipment). DESCRIPTION: A real-time, dynamic, and interactive knowledge base capability for efficient technical data management and interaction is sought in this initiative. The capability developed here should provide for custom database generation allowing the underlying software to process and vectorize content using an AI/ML algorithms. The transformation of static textual data into a dynamic LLM database here should yield guided troubleshooting 'wizard' capability to decrease maintenance touch-time for critical/complex DIPE (Depot Industrial Plant Equipment). The software in this instance should also be able to link textual part data with 3D images for the purpose of streamlining procurement capability. PHASE I: As this is a Direct-to-Phase-II (D2P2) topic, no Phase I awards will be made as a result of this topic. To qualify for this D2P2 topic, the Government expects the applicant(s) to demonstrate feasibility by means of a prior Phase I-type effort that does not constitute work undertaken as part of a prior or ongoing SBIR/STTR funding agreement. Applicant(s) will demonstrate the ability to automatically analyze commercial/technical manuals for IPE assets (CNC/Robotics) to identify specific maintenance & operational procedures in response to user inquiries. They must demonstrate the accuracy and completeness of findings to appropriately answer inquiries. PHASE II: The objectives of this D2P2 is varied. Development of a stand-alone maintenance platform version allowing software to run on CPU/tablets with limited connectivity in industrial facilities is desired. Also, development of the capability for user to create multiple instances or databases to be processed through AI algorithm for LLM interactive capability is sought. Further, development and enablement of capture of custom SOP (standard operating procedures) for procedures not defined in existing COTs manuals or database which can then be uploaded and vectorized through AI/ML algorithm is being pursued. Still further, development, maturation, and integration of LLM AI modules into software providing an interactive 'chatbot' experience for guided troubleshooting capability of complex/critical IPE is needed. Additionally, development of a capability to link textual part data with 3D images of IPE subcomponents is sought after. Lastly, development of the capability to provide all associated materials and tools required for maintenance/operational procedures derived from user inquiries is needed PHASE III DUAL USE APPLICATIONS: In the event the D2P2 is successful, WR-ALC leadership is committed to providing support to commercialize this capability. Further development will refine analysis to increase accuracy and reliability of maintenance procedure predictions/inquiries. REFERENCES: William Z. Bernstein and David Lechevalier Volume 124, Article No. 124011 (2019) https://doi.org/10.6028/jres.124.011 Journal of Research of National Institute of Standards and Technology - A Reference Schema for the Unit Manufacturing Process Information Model. Izabela Rojek, Ma gorzata Jasiulewicz-Kaczmarek, Mariusz Piechowski and Dariusz Miko ajewski, Appl. Sci. 2023, 13(8), 4971; https://doi.org/10.3390/app13084971 - An Artificial Intelligence Approach for Improving Maintenance to Supervise Machine Failures and Support Their Repair. Mohammed Misbahuddin, Abul Kashem Mohammed Azad, Veysel Demir College, College of Engineering and Engineering Technology, Northern Illinois University, DeKalb, USA. https://doi.org/10.4236/ait.2023.134008 - Machine-to-Machine Collaboration Utilizing Internet of Things and Machine Learning; KEYWORDS: AI/ML; LLM; Guided troubleshooting; interactive knowledge base