This paper focuses on the simulation of expected vibration signal of a faulted bearing. The importance of a proper simulation data in the early steps of condition monitoring is well-known. Nowadays the availability of a real test-bench is not so common, and this is proven by the number of scientific papers validated on few online available data centres (i.e. the CaseWestern University). As a consequence, the early validation of a new diagnostic technique on a proper simulated signal is crucial, in order to provide a feedback to the researcher and increasing the chances of getting a positive result in the real case. While dozens of comprehensive models of ball bearing have proposed in literature so far, the complexity of these models accordingly increased. As supposed, the scientific papers just outline the theoretical foundations of assumptions and features of the model, leaving the reader the task of converting all in lines of code, i.e. leaving several degrees of freedom to the programmer. The aim of this paper is to detail step-by-step an analytical model of faulted bearing that the reader could freely and immediately use to simulate different faults and different operating conditions. This project has been developed under a Creative Commons licence (Attribution-ShareAlike 4.0 International) and the vision of the project is a set of tools accepted by the community of researchers on condition monitoring, for the preliminary validation of new diagnostics techniques. The tool proposed in this paper is focused on ball bearing, and it is based on well-known model published by Antoni in 2007. The features available are the following: selection of the location of the fault (e.g. outer ring, inner ring, etc...), stage of the fault (e.g. punctual fault, distributed fault, etc...), cyclostationarity of the signal, random contributions, deterministic contributions, effects of resonances in the machine and working conditions (stationary and non-stationary). The script is provided for the open-source Octave environment. The output signal is finally analysed to prove the expected features.
Step-by-step algorithm for the simulation of faulted bearings in non-stationary conditions
D'ELIA, Gianluca;MUCCHI, Emiliano;DALPIAZ, Giorgio
2016
Abstract
This paper focuses on the simulation of expected vibration signal of a faulted bearing. The importance of a proper simulation data in the early steps of condition monitoring is well-known. Nowadays the availability of a real test-bench is not so common, and this is proven by the number of scientific papers validated on few online available data centres (i.e. the CaseWestern University). As a consequence, the early validation of a new diagnostic technique on a proper simulated signal is crucial, in order to provide a feedback to the researcher and increasing the chances of getting a positive result in the real case. While dozens of comprehensive models of ball bearing have proposed in literature so far, the complexity of these models accordingly increased. As supposed, the scientific papers just outline the theoretical foundations of assumptions and features of the model, leaving the reader the task of converting all in lines of code, i.e. leaving several degrees of freedom to the programmer. The aim of this paper is to detail step-by-step an analytical model of faulted bearing that the reader could freely and immediately use to simulate different faults and different operating conditions. This project has been developed under a Creative Commons licence (Attribution-ShareAlike 4.0 International) and the vision of the project is a set of tools accepted by the community of researchers on condition monitoring, for the preliminary validation of new diagnostics techniques. The tool proposed in this paper is focused on ball bearing, and it is based on well-known model published by Antoni in 2007. The features available are the following: selection of the location of the fault (e.g. outer ring, inner ring, etc...), stage of the fault (e.g. punctual fault, distributed fault, etc...), cyclostationarity of the signal, random contributions, deterministic contributions, effects of resonances in the machine and working conditions (stationary and non-stationary). The script is provided for the open-source Octave environment. The output signal is finally analysed to prove the expected features.I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
