Challenge: All Systems Go! – Predicting and optimizing maintenance for military platforms

Challenge Statement

The Department of National Defence and the Canadian Armed Forces (DND/CAF) are seeking innovative solutions for fleet-wide, automated, proactive Health and Usage Monitoring Systems (HUMS) for military platforms. The goal is to support a movement to Condition-Based Maintenance (CBM), and ultimately, predictive maintenance, to optimize limited maintenance resources, and to increase the availability of operational platforms.

Background and Context

The usual approach to maintaining in-service, military equipment (i.e., aircraft, vehicles, ships) is routine or time-based (preventative maintenance) or failure-based (corrective maintenance). Maintenance operations are scheduled on a routine basis or arranged based on the mileage or usage of equipment regardless of its condition. Preventative maintenance is a costly but necessary strategy for avoiding critical failures. Performing corrective maintenance for a single failure can be time-consuming to repair and can impact or even shut down an operation.

The DND/CAF is moving towards CBM in order to optimize maintenance operations, reduce costs, increase safety, and meet operational demands. It is a strategy for monitoring the actual condition of equipment and for scheduling maintenance tasks only when needed. It relies on HUMS to monitor operating parameters and record the status of critical systems in order to permit the early detection and remediation of faults. Predictive maintenance is the next evolution of CBM by using available data, algorithms and statistical models to predict when equipment is likely to fail.

The basis of the challenge is to develop a generic yet sophisticated HUMS that can be installed on virtually any military platform and be able to “learn” how to identify faults in order to report on the condition of that system. The second aspect of the challenge is to devise an on-board system that can fuse, interpret and analyze operational data from multiple sources, including on-board sensors as well as feedback from operators and maintenance personnel. Since it is not feasible for each platform to transmit massive volumes of data to be subsequently processed and analyzed, this would likely need to be done on-board the platform, which means size, weight, and power are considerations. The final aspect of the challenge involves means to interpret and visualize the data in order to support maintenance decisions regarding “what to do when” and how important it is to undertake maintenance actions “now or soon.” Understanding the relationship between platform usage and “health” in order to permit the shift to a much more efficient and predictive CBM strategy for DND/CAF fleet maintenance is a key aspect of the challenge.

Desired Outcomes

The DND/CAF is looking for fleet-level, on-board, real-time CBM and HUMS solutions that can be integrated into existing and new Land, Navy and/or Air platforms. Innovative research, tools, technologies and/or processes are sought that address, but are not limited to the following:

• A generic, standardized and sophisticated HUMS that can be integrated into different platforms to permit usage monitoring that combines sensor data from multiple sources with feedback from operators, maintenance staff on critical failures and major repair factors, etc.;
• Development of algorithms and statistical models for detecting and tracking incipient faults. Solution proposals might include the development of better sensors, modelling of fault patterns, and/or incorporating multiple condition measurements to derive a better understanding of the time between when a potential failure is identified and when it occurs;
• A HUMS capable of taking available data to “learn” in real-time how to detect and track faults in order to report on the “health” of the system, with a view to minimizing the upfront need to input technical data. Solution proposals are sought that make use of machine learning to create an evolving model of a system, which can be used to predict system behaviours. The solution would need to be integrated into the on-board HUMS, so size, weight, and power are a concern.
• A straightforward and effective means to visually represent CBM and HUMS data in order to support the identification of recommended actions and maintenance decisions;