The estimation of the life of rolling element bearings (REBs) is crucial to determine when maintenance is required. This paper presents a methodology to calculate the fatigue life of REBs considering non-stationary conditions. Instead of taking a constant value, the paper considers cyclic loading and unloading processes, as well as increasing and decreasing values of the speed of rotation. It employs a model-based approach to calculate contact loads between the different elements of the bearing, with a finite element model (FEM) used to calculate the contact stresses. Using this information, it then performs a fatigue analysis to study overloading in faulty bearings.

Safety in transport is a key. Railway and aerospace sectors have a need for ways to predict the behaviour of trains and aircraft, respectively. With this information, maintenance tasks for the correct operation of the assets can be carried out, reducing the number of failures that can cause an accident. However, the lack of enough data of the faulty state of those systems makes this to be difficult. Because of that either hidden faults or unknown faults can occur. As regulations in transport are very restrictive, components are usually substituted in early states of their degradation, which implies a loss of useful life of those components.In this article a methodology to overcome this limitation is presented. This methodology consists in the fusion of data obtained from two sources: data acquired from the real system, and synthetic data generated using physical models of the system. These physical models should be constructed in such a way that they can reproduce the main failure modes that can occur in the modelled system. This data fusion, that creates a hybrid model, not only allows to classify the condition of the system according to the aforementioned failure modes, but also to define new data that do not belong to any of those failure modes as a new failure mode, improving diagnosis and prognosis processes.

La seguridad en el campo del transporte es un punto crítico. Así, el sector ferroviario y el de la aeronáutica precisan de formas para predecir el comportamiento de trenes y aeronaves, respectivamente. Con esta información se pueden llevar a cabo las gestiones de mantenimiento necesarias para el correcto funcionamiento de los activos y reducir el número de fallos que puedan causar un accidente.Sin embargo, la falta de datos suficientes sobre estados con fallo de dichos sistemas hace que esta tarea sea complicada.Esta carencia de información hace que se puedan producir fallos ocultos o fallos desconocidos. Al tratarse la normativa del sector del transporte muy restrictivaen este aspecto, se tiende a reemplazar los componentes en estados tempranos de su degradación, lo que supone un desaprovechamiento de la vida de dichos componentes.En el presente artículo se propone una metodología para abordar esa limitación. Dicha metodología consiste en la fusión de datos de dos fuentes: por un lado, los datos adquiridos del sistema real; y, por otro lado, datos sintéticos generados a través de modelos físicos. Dichos modelos físicos han de estar construidos de forma que sean capaces de reproducir los principales modos de fallo que pueden ocurrir en dichos sistemas.Esta fusión de datos, que formaun modelo híbrido, permite no sólo clasificar el estado del sistema según los modos de fallo previamente estipulados, sino también definirnuevos modos de fallo que no concuerden con ninguno de los modos de fallo anteriores, mejorando los procesos de diagnosis y prognosis.

Diagnosis and prognosis processes are necessary to optimise the dependability of systems and ensure their safe operation. If there is a lack of information, faulty conditions cannot be identified and undesired events cannot be predicted. It is essential to predict such events and mitigate risks, but this is difficult in complex systems.Abnormal or unknown faults cause problems for maintenance decision makers. We therefore propose a methodology that fuses data-driven and model-based approaches. Real data acquired from a real system and synthetic data generated from a physical model can be used together to perform diagnosis and prognosis.As systems have time-varying conditions related to both the operating condi- tions and the healthy or faulty state of systems, the idea behind the proposed methodology is to generate synthetic data in the whole range of conditions in which a system can work. Thus, data related to the context in which the system is operating can be generated.We also take a first step towards implementing this methodology in the field of rolling element bearings. Synthetic data are generated using a physical model that reproduces the dynamics of these machine elements. Condition indicators such as root mean square, kurtosis and shape factor, among others, are calculated from the vibrational response of a bearing and merged with the real features obtained from the data collected from the functioning systemFinally, the merged indicators are used to train SVM classifiers (support vector machines), so that a classification according to the condition of the bearing is made independently of the applied loading conditions even though some of the scenarios have not yet occurred.

Rolling element bearings have been studied for decades, but more research is required into their dynamics, especially failure due to different kinds of damage in the context of condition monitoring. The appearance of a failure in an element of a bearing, as well as its degradation, can entail not only a malfunction in the system in which it is located, but also a catastrophic failure. This work presents a multi-body model of a rolling element bearing with the objective of analysing the dynamics of the bearing and emphasising the effect of defects in any of its element. The study models the metal-metal contacts between the bearing’s elements using the Hertz contact and the elastohydrodynamic lubricationtheories, both of which are theories of nonlinearity. It also considers the non-stationary regime of bearings and local geometric damage. Its results are compared with results in the literature. Finally, it includes a set of additional results showing different aspects of the response of the bearing.

Using condition-based maintenance (CBM) to assess machinery health is a popular technique in many industries, especially those using rotating machines. CBM is relevant in environments where the prediction of a failure and the prevention and mitigation of its consequences increase both profit and safety. Prognosis is the most critical part of this process and the estimation of Remaining Useful Life (RUL) is essential once failure is identified. This paper presents a method of synthetic data generation for hybrid model-based prognosis. In this approach, physical and data-driven models are combined to relate process features to damage accumulation in time-varying service equipment. It uses parametric models and observer-based approaches to Fault Detection and Identification (FDI). A nominal set of parameters is chosen for the simulated system, and a sensitivity analysis is performed using a general-purpose simulation package. Synthetic data sets are then generated to compensate for information missing in the acquired data sets. Information fusion techniques areproposed to merge real and synthetic data to create training data sets which reproduce all identified failure modes, even those that do not occur in the asset, such as Reliability Centered Maintenance (RCM), Failure Mode and Effect Analysis(FMEA). This new technology can lead to better prediction of remaining useful life of rotating machinery and minimizing and mitigating the costly effects of unplanned maintenance actions.