In today’s industrial environment, effectively monitoring assets throughout their entire lifetime is essential. Prognostic and Health Management (PHM) is a powerful tool that enables users to achieve this goal. Recently, sensor-equipped actuator electrification has been introduced to capture intrinsic key system variables as time series. This data flow has opened up new possibilities for extracting essential maintenance information. To leverage the full potential of these data, we have developed a novel algorithm for time series registration, which serves as the core of a new similarity-based prognostic method in a PHM context: Partial Time Scaling Invariant Temporal Alignment for Remaining Useful Life Estimation (PARTITA-RULE). Our algorithm transforms acceleration signals into a subset of descriptors for a new actuator, creating a time series. We can extract valuable maintenance information by aligning this time series with the one already labeled from past behaviors of the same actuator’s family of heterogeneous sizes and robust scaling factors. The unique aspect of our method is that we do not need to inject prior knowledge for registration intervals at this stage. Once the unknown series is aligned with all possible candidates, we create a weighting scheme to assign a relevance score with an uncertainty measurement for each aligned pair. Finally, we compute interpolants on the Wasserstein space to obtain the asset’s Remaining Useful Life (RUL). It is important to note that a relevant result in a PHM context requires a database filled with different labeled system behaviors. To test the effectiveness of our method, we use an industrial data set of vibration signals captured on an aeronautical electric actuator. Our method shows promising Remaining Useful Life (RUL) estimation results even with incomplete time segments.

The aerospace industry develops prognosis and health management algorithms to ensure better safety on board, particularly for in-flight controls where jamming is dreaded. For that, vibration signals are monitored to predict future defect occurrences. However, time series are not labeled according to severity level, and the user can only assess the system health from the data mining procedure. To that extent, a clustering algorithm using a deep neural network core is developed. Time series are encoded into pictures to be fed into an artificially trained neural network: U-NET. From the segmented output, one-dimensional information on cluster frontiers is extracted and filtered without any parameter selection. Then, a kernel density estimation finally transforms the signal into an empirical density. Ultimately, a Gaussian mixture model extracts the latter independent components. The method empowered us to reveal different degrees of severity faults in the studied data, with their respective likelihoods, without prior knowledge. It was then compared to state-of-the-art machine learning algorithms. However, internal clustering results evaluation for time series is an open question. As the state-of-the-art indexes were not producing relevant results, a new indicator was built to fulfill this task. We applied the whole method to an actuator consisting of an induction machine linked to a ball screw. This study lays the groundwork for future training of diagnosis and prognosis structures in the health management framework.

Electromechecanical actuators in the aerospace industry are gradually replacing hydraulic ones. In these circumstances, prognostics and health management are innovative frameworks to ensure better safety on board, especially in flight controls where jamming is dreaded. It allows the user to assess and predict system health in real-time. The first step is to collect temporal data from the monitored actuator and perform a data mining procedure to gain insight into its current health. Clustering encompasses several data-driven methods used to reveal patterns. However, getting a set of classes usually requires providing the algorithm with prior knowledge, such as the number of groups to seek. To avoid this drawback, we have developed a clustering algorithm using a deep neural network, as its core, to get the number of groups in data associated with their likelihood. Temporal sequences are reshaped into pictures to be fed into an artificially trained neural network: U-NET. The latter outputs segmented images from which one-dimensional information is extracted and filtered, without any need for parameter selection. A kernel density estimation finally transforms the signal into a candidate density. This new method provides a robust clustering result coupled with an empirical probability to label the times series. It lays the groundwork for future training of diagnosis and prognosis structures in the PHM framework.