Artificial intelligence based digital prescriptive maintenance of ships(DiMOS)

With the advent of new emerging and enabling technologies e.g. big data analytics, machine learning, internet of things, intelligent sensors and cloud computing, the approach to maintenance of assets is changing. The advanced augmented approach offers the possibility of performing prescriptive maintenance that offers significant advantages over using traditional (descriptive), preventive, or predictive models individually. Traditional maintenance tends to be reactive-responding to failures in equipment or devices after the fact. This traditional, reactive approach of describing failures after they've occurred is the worst-case scenario for maintenance: reacting to failures in equipment or devices after the fact. Preventive maintenance empowers operators to carry out continuous maintenance.

Prescriptive maintenance goes beyond the realm of descriptive, preventive, and predictive maintenance. Descriptive focuses on what happened in the past. Preventive maintenance is time based. Predictive analytics discovers potential options for the future. Prescriptive maintenance leverages all these approaches and capabilities. The realm of what should happen and the execution of optimized maintenance strategies is precisely the realm of prescriptive maintenance.

The DiMOS project proposes a prescriptive maintenance digital platform for condition monitoring and maintenance planning of ship's structure, engine machinery and auxiliary system by real-time sensor data and AI-based models to prescribe maintenance based on monitored condition and taking into account risk level, maintenance timing and associated cost.

The main application of the DiMOS platform will include (1) Real-time continuous condition monitoring, diagnostic and failure analysis of ships engine machinery, structure and auxiliary systems. (2) Risk-based inspection analysis and provision of critical parts identification of ships components, 3) Detailed prescriptive maintenance of ships components based on condition, time, risk and cost etc.

_The proposed platform will; (1) reduce reliance on experienced and expert inspection engineers to process condition monitoring data and devise a maintenance plan(2) it will reduce interpretation time in devising and implementing maintenance actions reducing maintenance hours by 70% (3) it will automate safety or maintenance operations to the extent where maintenance operations don't require human intervention (4)it will reduce assets unscheduled downtime by 25%, cost by 35% and will improve performance and efficiency of asset (5) it will allow operators to perform cost-effective maintenance on the basis of risk profile of faults detected._

The DiMOS project development is being done based on the collaboration of different partners including Vibtek, ICON, Relmar, KCC, TWI, and BUL.