Future AI-based Maintenance

Purpose and goal:

In the project, a framework for predictive maintenance and planning under uncertainty in industry will be developed. It includes development of a platform for prediction models, new methods for forecasting component lifetime, methods for dynamic and robust planning, and simulation tools to evaluate and compare different advanced maintenance strategies.

The project aims to reduce business risks by switching to more efficient and AI-based maintenance strategies and thereby increasing resource efficiency and competitiveness in Swedish industry. Expected results and effects:

A new modular platform for using predictive maintenance and planning under uncertainty in industry will be developed. A new and improved prediction methods for component lifetime forecasting, based on machine learning and causal modeling, especially focusing on uncertainty management will be developed. New methods for dynamic and robust planning under uncertainty in industry as well as a simulation tool for decision support that enables evaluation and comparison of different advanced maintenance strategies, with regard to risk and cost.

Approach and implementation: To achieve the goal of the project there are four main areas to address, with corresponding work packages: - a common probabilistic representation of lifetime predictions, costs, and risks, i.e. the “glue” of the system - methods for prediction of remaining useful life and its uncertainty, suitable for a range of components

- dynamic maintenance planning under uncertainty, taking production into consideration - a simulation tool for evaluation of maintenance strategies In addition, there is a work package for project management and knowledge dissemination.