Search

Operators from various industries have been pushing the adoption of wireless sensing nodes for industrial monitoring, and such efforts have produced sizeable condition monitoring datasets that can be used to build diagnosis algorithms capable of warning maintenance engineers of impending failure or identifying current system health conditions. However, single operators may not have sufficiently large fleets of systems or component units to collect sufficient data to develop data-driven algorithms. One potential solution to overcome the challenge of having limited representative datasets is to merge datasets from multiple operators with the same type of assets. However, directly sharing data across the company's borders yields privacy concerns. Federated learning (FL) has emerged as a promising solution to leverage datasets from multiple operators to train a decentralized asset fault diagnosis model while maintaining data confidentiality. However, the performance of traditional FL algorithms degrades when local clients' datasets are heterogeneous. The dataset heterogeneity is particularly prevalent in fault diagnosis applications due to the high diversity of operating conditions and system configurations. To address this challenge, this paper proposes a novel clustering-based FL algorithm where clients are clustered based on their dataset similarity. Estimating dataset similarity between clients without explicitly sharing data is achieved by training probabilistic deep learning models and having each client examine the predictive uncertainty of the other clients' models on its local dataset. Clients are then clustered for FL based on relative prediction accuracy and uncertainty. Experiments on three bearing fault datasets, two publicly available and one newly collected for this work, show that our algorithm significantly outperforms FedAvg and a cosine similarity-based algorithm by 5.1% and 30.7% on average over the three datasets. Further, using a probabilistic classification model has the additional advantage of accurately quantifying its predictive uncertainty, which we show it does exceptionally well. [ABSTRACT FROM AUTHOR]

Noise legislations and the increasing customer demands determine the NVH (Noise Vibration and Harshness) development of modern commercial vehicles. In order to meet the stringent legislative requirements for the vehicle noise emission, exact knowledge of all vehicle noise sources and their acoustic behavior is required. The development of new analysis tools to investigate the vibroacoustic behaviour within vehicle development process is of essential importance to achieve better products in combination with time and cost reduction. This paper discusses the application of an experimental multi-point and multi-DOF energy flow analysis (EFA) methodology for the NVH assessment and ranking of vibro-acoustic sources. The methodology consists of two steps in which the system mobility matrix are obtained by direct measurements whereas the dynamic loads under operational conditions are obtained through an inverse procedure. A finite difference technique is used to distinguish between degrees of freedom. A practical example is given for a gearbox, for which the noise paths from the gears through the gear-supporting shafts and bearings into the gearbox housing has been experimentally assessed in terms of energy flows. This paper also discusses the accuracy and limits of the methodology. [ABSTRACT FROM AUTHOR]

This paper proposes a method for detecting insufficient lubrication in rolling bearings, by utilising a low cost MEMS ultrasonic microphone. We run a set of grease lubricated bearings under load, alternating between good and insufficient lubrication conditions by adjusting speed and temperature, over an 18 day experiment, followed by a short validation experiment on a different set of bearings on a secondary test rig. Vibration signals are recorded from an ultrasonic microphone, using a purpose built prototype sensor, then band-pass and high-pass filtered and subsequently analysed using statistical time-domain features. The viscosity ratio Kappa of the lubricant is used as a metric for quality of oil-film formation, and is calculated throughout the experiment from rotating speed and temperature. The results are compared with the predictions from a linear regression model and a simple Neural Network trained on ensembles of three and ten selected features. The three Hjorth's Parameters activity , mobility and complexity calculated on high-pass filtered microphone signals from the experiments show a good correlation with the viscosity ratio Kappa. When combining these, using a Linear Regression model or a simple two layer Neural Network model, we are able to predict the viscosity ratio with good precision. The proposed method may be used to assess the quality of oil film formation in bearings during operation under varying speed and temperature conditions. [ABSTRACT FROM AUTHOR]

• Novel technique is proposed to diagnose vibration data from rolling bearing. • Method indicates perfect fault signature in the presence of multiple impulsive components. • Bi-frequency representation, matrix factorization, spatial denoising are used. • Presented results are based on real data from reciprocating compressor operating in industrial conditions. In this paper, authors present the original procedure for local damage detection in rolling bearings based on vibration data. The aim is to obtain envelope spectrum (ES) of the signal component related to damage, that is clear and easy to interpret. The method is especially aimed at cases, when multiple cyclic impulsive components are present and interfere with each other, which makes ES of such signal very difficult to evaluate. In order to deal with such situation properly, authors propose to choose Cyclic Spectral Coherence (CSC) map as a two-dimensional data representation that will be the basis for the analysis. Nonnegative Matrix Factorization (NMF) is used to analyze such map in two ways: first, it helps to initially separate cyclic components by producing a set of filters for input vibration data, and second, to identify proper damage-related frequency components in envelope spectrum. In addition, an intermediate step of spatial denoising allows enhancing the properties of CSC map. Finally, Monte Carlo simulation improves statistical significance of the result and increases robustness by reducing the impact of random initialization effects. The method has been evaluated using real-life vibration data measured on rolling bearing operating in the industrial gas compressor. [ABSTRACT FROM AUTHOR]

Highlights • Validated, comprehensive dynamic models. • Modeling contribution in diagnostics of bearings, gears and cardan joints. • Physics-based prognostics. • Insights on behavior of faulty bearings, gears and cardan joints. Abstract Prognostics of rotating machines parts is an important subject in the maintenance field. The paper focuses on the contribution of the dynamic models for the development of reliable prognostics and diagnostics algorithms. The dynamic models enable prediction of changes in dynamic behavior reflecting the fault type and severity. New condition indicators and signal processing techniques can be developed based on these insights. The models can be used to build databases of simulated signals that can be used to perform sensitivity analysis with respect to geometrical tolerances, environmental and operating conditions, severity, type, and location of faults. A research methodology towards physics-based prognostics of rotating machinery is proposed. The methodology combines experiments with comprehensive dynamic models, and it is demonstrated with examples of rolling-element bearings, gear transmissions, and cardan joints. It is important to well validate the models and to assure their completeness. Therefore, experiments and models are complementary: the experimental results should be used to validate the model assumptions, and the models enable comprehensive interpretation of the test results. The combination of the two allow the generalization of the conclusions and algorithms with a high level of confidence. [ABSTRACT FROM AUTHOR]

• The paper focuses on the Hjorth's parameters application for condition monitoring. • The Hjorth's parameters are three: the Activity, the Mobility and the Complexity. • A new parameter called Detectivity is introduced to merge Hjorth's parameters. • Effectiveness of Detectivity is proved on both experimental and synthetic datasets. Hjorth's parameters are statistical time-domain parameters used in signal processing and introduced by Bo Hjorth in 1970. These parameters are Activity, Mobility and Complexity. They are related to the variance of the signal and of its subsequent derivatives. They are commonly used in the analysis of electroencephalography (EEG) signals for feature extraction, but also in the tactile signal analysis in robotic area. In this paper, Hjorth's parameters are applied to vibration signals for fault detection in ball bearings. In particular, two open-access datasets are used: the NASA bearing dataset of the University of Cincinnati and the Polytechnic of Turin rolling bearing dataset. In the first part of the paper the three parameters are used for monitoring the health of the bearings along their lifetime, proving their effectiveness in condition monitoring. In the second part, a new parameter is introduced, named Detectivity, that merges the information carried by Hjorth's parameters in a single value and is suitable for continuous monitoring of machines. [ABSTRACT FROM AUTHOR]

This paper presents a data-driven prognostic framework for rolling-element bearings (REBs). This framework infers a bearing's health index by defining a degree-of-defectiveness (DD) metric in the frequency domain of bearing raw signal, named DD-based health index (DDHI). Then, we systematically apply least-square support vector machines (LSSVMs) in the forms of Bayesian inference-aided one-class LSSVM (Bayesian-OCLSSVM) for anomaly detection in order to define the time-to-start (TTS) point of RUL prediction and the recurrent least-square support vector regression (Recurrent-LSSVR) model for predicting future values of DDHI for calculating the RUL. In addition, this paper addresses several pertinent challenges, such as failure threshold determination during anomaly detection and RUL estimation, by developing adaptive thresholds. We conduct extensive experiments on publicly available two benchmark datasets using a run-to-failure experiment. The results demonstrate the efficacy of the proposed framework compared to state-of-the-art methods in terms of the accuracy and convergence of the RUL estimation of bearings. [ABSTRACT FROM AUTHOR]

Gear transmissions remain as one of the most complex mechanical systems from the point of view of noise and vibration behavior. Research on gear modeling leading to the obtaining of models capable of accurately reproduce the dynamic behavior of real gear transmissions has spread out the last decades. Most of these models, although useful for design stages, often include simplifications that impede their application for condition monitoring purposes. Trying to filling this gap, the model presented in this paper allows us to simulate gear transmission dynamics including most of these features usually neglected by the state of the art models. This work presents a model capable of considering simultaneously the internal excitations due to the variable meshing stiffness (including the coupling among successive tooth pairs in contact, the non-linearity linked with the contacts between surfaces and the dissipative effects), and those excitations consequence of the bearing variable compliance (including clearances or pre-loads). The model can also simulate gear dynamics in a realistic torque dependent scenario. The proposed model combines a hybrid formulation for calculation of meshing forces with a non-linear variable compliance approach for bearings. Meshing forces are obtained by means of a double approach which combines numerical and analytical aspects. The methodology used provides a detailed description of the meshing forces, allowing their calculation even when gear center distance is modified due to shaft and bearing flexibilities, which are unavoidable in real transmissions. On the other hand, forces at bearing level were obtained considering a variable number of supporting rolling elements, depending on the applied load and clearances. Both formulations have been developed and applied to the simulation of the vibration of a sample transmission, focusing the attention on the transmitted load, friction meshing forces and bearing preloads. [ABSTRACT FROM AUTHOR]