Modeling the Business Value of a Predictive Maintenance System using Monte Carlo Simulation

Abstract

The automotive industry is undergoing a period of rapid advancement, as original equipment manufacturers race to develop the next generation of electric, autonomous, and connected vehicles. Many manufacturers are investing in prognostics technology, which has made advancements mainly in the aerospace industry over the past couple decades. For vehicle fleet managers who own and operate many vehicles, prognostics and early fault detection can enable predictive maintenance strategies, which can realize cost savings versus corrective or preventative strategies. However, developing the technology required for predictive maintenance can be an expensive undertaking, requiring many parts, months of data collection, and possibly years of engineering effort. It is critical to understand the expected return on investment for developing such a project.

In this paper, we present a framework to model the business value of a predictive maintenance system. The predictive maintenance system is described as the combination of a component being monitored, a network of sensors, a health monitoring algorithm, and a service policy defining the response to those actions. The framework incorporates models of component failure, health monitoring algorithm performance, a policy of actions, and costs associated with those actions. The framework is generic and may be applied to any component where degradation can be modelled by a probability distribution. Monte Carlo simulation is employed to estimate the distribution of repair costs for a particular maintenance strategy, which can then be used to assess the value of a predictive maintenance system.