Integrated maintenance and mission planning using remaining useful life information

Abstract

The modern world requires high reliability and availability with minimum ownership cost for complex industrial systems (high-value assets). Maintenance and mission planning are two major interrelated tasks affecting availability and ownership cost. Both tasks play critical roles in cost savings and effective utilization of the assets, and cannot be performed without taking each other into consideration. Maintenance schedule may make an asset unavailable or too risky to use for a mission. Mission type and duration affect the health of the system, which affects the maintenance schedule. This article presents a mathematical formulation for integrated maintenance and mission planning for a fleet of high-value assets, using their current and forecast health information. An illustrative example for a fleet of unmanned aerial vehicles is demonstrated and evolutionary-based solutions are presented.

A Comprehensive Health Indicator Integrated by the Dynamic Risk Profile from Condition Monitoring Data and the Function of Financial Losses

“…The COF is assessed by loss function using Equations (9) and (10). By solving Equation (11) using the least squares method, with financial loss L = 0.5•EML [20] and shutdown threshold d 2 calculated in Section 3.2.1, the shape parameter γ can be obtained as γ = 4.1975.…”

“…This result was in accordance with the model designed in Section 2.1.2. The is calculated using Equations (9) and (10). By solving Equation (11) using least squares method, with financial loss = 0.5 • [20], the shape parameter can be obtained.…”

“…After 2 October, the health status of the compressor became worse as the MD value kept increasing. The COF is calculated using Equations (9) and (10). By solving Equation (11) using least squares method, with financial loss L = 0.5•EML [20], the shape parameter γ can be obtained.…”

“…Industry today seeks maintenance solutions based on real-time health monitoring of assets [10,11]. Instead of collecting data from similar systems, the condition monitoring data give information on the individual degradation process of the target system, providing an opportunity to update the risk factor before actual failure occurs.…”

Towards Developing a Novel Framework for Practical PHM: a Sequential Decision Problem solved by Reinforcement Learning and Artificial Neural Networks

“…), which in turn depend on the Operation and Maintenance (O&M) decisions taken in time for optimal equipment usage. When predicting the RUL, these future conditions of equipment usage are generally assumed constant (e.g., (Camci et al, 2018;Van Horenbeek & Pintelon, 2013;Sankararaman, Ling, Shantz, & Mahadevan, 2011;Cadini, Zio, & Avram, 2009;Cadini, Sbarufatti, Corbetta, & Giglio, 2017;Leser et al, 2017)) or behaving according to some known exogenous stochastic process (e.g., (Ling & Mahadevan, 2012;Sankararaman, 2015;Ding, Tian, Zhao, & Xu, 2018)). This does not reflect reality and the RUL predictions that guide the O&M decisions are deemed to be incorrect.…”