Chapter 1 Bearing Faults Simulations Through a Parametric Model of a Gearbox S. Cinquemani, F. Rosa, and E. Osto Abstract The paper deals with a project aimed to improve the reliability of a condition monitoring system applied on gearboxes installed on rolling mills. In this context, to properly set up the algorithm, it is necessary to have measurements associated both to standard operating conditions and to malfunctioning. Since the experimental determination of the latter is obviously not cost and time effective, they can be simulated by means of numerical models of the mechanical system in several operating conditions. The outputs generated, corresponding to different fault conditions of the more critical elements of the system, will provide a useful data base to tune the algorithm of condition monitoring. Keywords Condition monitoring • Bearing • Mechanical transmission • Rolling mill • Parametric model 1.1 Introduction Gearboxes for rolling mills are complex machines that cannot be treated as commodities, since they are parts of a complex drive system that in case of failure could seriously affect the plant productivity [1–3]. The root causes of a geared system failure can sometimes be quite different from the appearing ones. From this point of view, an early detection of improper operating conditions, as condition monitoring can provide, gives a good chance to plan extraordinary maintenance to prevent sudden stop of production and to identify primary failure causes instead of secondary ones. It worth to be mentioned that preventive maintenance to solve primary failure causes implies negligible cost compared to ones related to secondary failure causes (e.g. bearing replacement vs. gear replacement). Bearings represent a typical source of gearboxes failures or improper operation [4, 5]. Also for bearing, due to the vast number of different failure modes, specific standards have been introduced as ISO 15243. Gradual deterioration of the operating behaviour is normally the first signal of bearing damage. Failures due to poor ordinary maintenance (lack of lubrication, for example) and improper mounting are relatively infrequent, and very often lead quickly to machine downtime. On the other hand, depending on the operating conditions, a few weeks, or under some circumstances, even a few months, may pass from the time damage begins to the moment the bearing actually fails because of the contact fatigue damaging mechanisms. This typical progressive evolution of the damage makes bearings especially suitable for continuous condition monitoring applications. From the point of view of failure detection, the main effects of bearing damage impacts on operating temperatures, lubricant contamination and vibrations. In principle, all these information can be used for condition monitoring application, but the techniques based on the analysis of vibration signals are the most efficient as they can provide an early identification of the specific bearing involved, of the part of the bearing affected by damage and on the degree of the damage itself, thanks to the information coming from frequency and amplitude data. In particular damages related to contact fatigue affecting the races or the rolling elements can be detected and identified [6]. Moreover they are suitable to be modelled by means of models which can provide preliminary information in terms of expected frequencies and amplitudes in the signal analysis. Condition monitoring algorithms are based on signals in different working conditions. To properly set up the algorithm, it is necessary to have measures associated both to standard operating conditions and to malfunctioning. The latter, not being experimentally determinable, can be simulated by developing numerical models of the machine under varying S. Cinquemani ( ) • F. Rosa Department of Mechanical Engineering, Politecnico di Milano, Via La Masa 1, 20156 Milano, MI, Italy e-mail: simone.cinquemani@polimi.it E. Osto Siemens VAI Metals Technologies Srl, Via Luigi Pomini, 92, Marnate, Italy © The Society for Experimental Mechanics, Inc. 2015 C. Niezrecki (ed.), Structural Health Monitoring and Damage Detection, Volume 7, Conference Proceedings of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-15230-1_1 1
RkJQdWJsaXNoZXIy MTMzNzEzMQ==