The overall aim of this innovation project is to perform a Market and Economic Feasibility Study for commercialising our market disruptive Signals2Solutions (S2S) software platform for cost-effective Predictive Maintenance (PM) analysis targeting the global offshore and civil aviation industries.
The Signals2Solutions (S2S) platform is focused on analyzing non-stationary data and interpreting its results, based on a further deployment of the Hilbert-Huang Transform (HHT) methodology. Comparing with stationary signals (i.e. stable data along time), non-stationary signals (i.e. based on time-varying coefficients) generally better reflect real production cases, where the machines' use intensity varies according to different time schedules, productivity needs and working conditions. Therefore, HHT is the most indicated Predictive Maintenance method for those machinery and equipment characterized by high variability conditions, such as, for instance, drilling equipment and aircrafts & helicopters, where intensity use strongly vary based on e.g. the ground conditions or the aircraft condition. Until now the PM techniques used for those tasks have been basically two: Fourier Transform (FT) and Machine Learning (ML). Nevertheless, they both present evident limits when it comes to analyze equipment and machinery in real-use conditions, as they start from the assumption that signals by default are linear and stationary (FT) or that data generation mechanism of an equipment does not change over time (ML). On the other hand, Simulation & Lab tests of other signal analysis techniques based on HHT - i.e. Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEDMAN) and Ensemble Empirical Mode Decomposition (EEMD) - have shown that those techniques cannot be applied to Predictive Maintenance, as they currently require too much computing power and are therefore too low to effectively detect machinery problems.