A Digital-Twin Enterprise Infrastructure for Plant Asset Monitoring and Predictive Maintenance

With this project, FPoliSolutions will develop an enterprise digital platform for condition monitoring and predictive maintenance augmented with enhanced predictive capabilities by combining digital-twin, machine learning and AI technologies. FPoliSolutions will leverage an existing in-house enterprise digital Asset Data Analytics platform, (ADALIX), which is currently limited to Operating Deflection Shape (ODS) analyses. ADALIX was originally designed to fit a gap in the condition monitoring market. It combines high enterprise scalability with high resolution, physics-based system engineering which is typically only available in small scale single-purpose condition monitoring products. The existing application also leverages INL’s RAVEN tool to handle simulation frameworks and machine learning. In this project the application will be extended to consider other sensor data, well beyond ODS, to provide digital-twin technology that enhances failure predicting capabilities. As an enterprise solution, it will provide a highly collaborative environment within the organization to view, edit, monitor, analyze assets data near-real time and intelligently prioritize maintenance activities. Interpretability, traceability, and overall transparency of the methodology are essential in the nuclear industry (operations specialists, site managers, maintenance engineers and regulators alike). Such quality attributes will require careful consideration in the evolution of the architecture. Key quality attributes of the proposed solution include ease-of-use, scalability (thousands of assets, multiple systems/enterprises), cross-platform compatibility, and user/customer-specific deployment models that meet individual cybersecurity and IT requirements. Effective application program interfaces (APIs) will minimize acquisition costs and provide the ability to manage assets or problems that can be use-case specific. The system will be able to process cascading effects by ‘triaging’ different types of signals and disparate information sets.FPoli has an open dialogue with utilities in this area. In Phase-I of this project, FPoli intends to continue gathering and clarify industry use cases. This step is necessary to optimize the architecture of the product and ensure a strong market fit to the final product. Considering the existing framework, FPoli can achieve an initial minimum viable product (MVP) by the completion of Phase-I during which implementation techniques and approaches will be prototyped and tested. Phase-I work will provide detailed requirements to be considered for Phase-II. The final product will be completed in Phase-II. This will require upgrades and optimization at scale to address specific maintenance problems and programming of new methods into the backend of the platform to solve those problems.