Prognosis of Mission-Aware Remaining Useful Life for ROV Thrusters Using a Physics-Consistent Simulation Framework

Abstract

Remotely operated vehicles (ROVs) are now widely used to perform underwater missions that vary in duration, maneuvering intensity, and propulsive load. However, these differences, although they typically have a significant impact on how propulsion components degrade over time, are not considered by predictive models. As a result, the remaining useful life (RUL) of the thrusters is typically estimated assuming that future operating conditions will be like past ones, an assumption that is often unrealistic in mission-driven underwater operations.

In our work, mission-aware prediction of the remaining useful life of ROV thrusters is investigated, focusing on how planned mission characteristics shape the degradation progression and RUL estimation. We use a physics-consistent simulation environment representative of the BlueROV2, with applicability to related platforms such as the BlueBoat surface vehicle, in order to study this interaction in a controlled and repeatable manner where mission profiles are defined through a set of descriptors that capture thrust demand, load variability, duty cycle, and maneuvering aggressiveness, allowing for systematic comparison of different operational scenarios. Degradation is introduced by gradually modifying the electromechanical performance parameters of the thrusters, producing distinct degradation trajectories under the same initial conditions.

Health indicators derived from simulated measurements, such as motor current, rotational speed, temperature, and thrust, are used to monitor the progression of degradation, and predictive estimates are obtained by propagating the estimated health state forward across the candidate mission profiles rather than assuming a single RUL independent of the future mission. The results show that missions with similar time durations, but different propulsion command characteristics can lead to substantially different RUL predictions, even when the initial health state is the same.

 The simulation framework, in addition to observing these differences, allows us to have a systematic investigation of the sensitivity of the prognosis results to the parameters of each mission. By varying the mission descriptors separately, we can determine which aspects of a mission, such as sustained high thrust versus intermittent peak loads, dominate the degradation behavior and leading to uncertainty in the RUL predictions. This type of analysis is difficult to perform only with field data, and it is quite important for understanding the limits of prognosis in operational environments. Also, these results suggest that the RUL should be considered as dependent on how a system is expected to be used, rather than as a single fixed value. Since different mission characteristics have been shown to affect degradation in different ways, considering the planned mission leads to predictive estimates that are easier to interpret and more directly linked to operational choices. In practice, the framework allows for the comparison of different mission options based on their expected impact on the thruster life and highlights missions that are likely to accelerate degradation while offering the ability to prioritize maintenance actions based on expected operational requirements.

Finally, our study through simulations offers a simple and repeatable way to investigate this behavior in underwater robotic systems, without requiring extensive real-world failure data, and can serve as a basis for future work on mission-aware PHM methods.