Modelling and Analysis of Reliability and Availability by stochastic modelling techniques considering continuously the statistical quality of the initial data
Final Report Abstract
Modelling and Analysis of Reliability and Availability by stochastic modelling techniques considering continuously the statistical quality of the initial data The modelling and simulation of operational reliability and availability parameters of technical systems is carried out with powerful methods. These are not limited to special systems and enable the modelling of manifold aspects. However, the analysis is based on mean-valued functions, i. e. the statistical quality of the input data could not be considered in the analysis yet. The calculation results were also meanvalues. Since usually only limited data are available to describe the system behaviour, e. g. a limited sample of system downtimes, a continuous consideration of the statistical quality of the input data is necessary. The confidentiality of the data is thus captured and indicated with a confidence interval or the specification of a confidence level. Within the research project, a methodology to integrate the statistical quality into the modelling, simulation and analysis of technical systems was developed. The method includes the newly developed Petri grids as a high-level Petri net class as well as the newly developed bootstrap Monte Carlo simulation as an analysis method of operational parameters with confidence intervals. Petri grids are based on the extended coloured stochastic Petri nets and extend them with the aspect of confidence intervals. With Petri grids, a close-to-reality modelling of technical systems is possible, whereby the statistical uncertainty of the input data is captured. The flexible models allow the consideration of manifold aspects related to the operation of a technical system, e. g. different component states, ageing aspects or costs. The bootstrap Monte Carlo simulation combines the bootstrap method with the Monte Carlo simulation and thus allows the analysis of the Petri grids with a confidence interval. With the help of the bootstrap Monte Carlo simulation, a confidence level can be assigned to the analysis results, based on the statistical quality of the input data. The overall methodology was implemented in software for practical and research-oriented application. The bootstrap Monte Carlo simulation tool (BMCS-Tool) allows the application of the bootstrap Monte Carlo simulation. In addition, the REALIST (Reliability, Availability, Logistics and Inventory Simulation Tool) program package developed at the Institute of Machine Components (IMA) of the University of Stuttgart was further developed to enable the modelling with Petri grids. The method was also verified and validated on several levels. On the basis of two sample calculations, the modelling with Petri grids and the analysis by means of bootstrap Monte Carlo simulation was demonstrated and the added value of the confidence intervals was clarified. Furthermore, the transferability to general applications was demonstrated. In the use cases, the reliability of a dual-clutch transmission and the availability of an offshore wind turbine was analysed with confidence intervals.
Publications
- Availability Analyses of Maintained Systems with Confidence Intervals. In: Proceedings of 2019 Annual Reliability and Maintainability Symposium (RAMS 2019), Orlando, USA, 2019
Müller, Frank; Bertsche, Bernd
(See online at https://doi.org/10.1109/RAMS.2019.8769269) - Simulation and Analysis of Simple, Repairable Systems with a Confidence Interval. In: Beer, M.; Zio, E. (Hrsg.): Proceedings of the 29th European Safety and Reliability Conference 2019 (ESREL 2019), Hannover, Deutschland, 2019 [Best Student Poster Award der European Safety and Reliability Association]
Müller, Frank; Bertsche, Bernd
- Verfügbarkeitsanalyse reparierbarer Systeme mit Vertrauensbereich. In. VDI Berichte 2345, 29. VDI-Fachtagung Technische Zuverlässigkeit 2019 in Nürtingen bei Stuttgart, Düsseldorf: VDI-Verlag, 2019, S. 281-292
Müller, Frank; Bertsche, Bernd
(See online at https://doi.org/10.51202/9783181023457-281) - Coverage Probability of Methods for Steady-State Availability Inference with a Confidence Intervals. In: Proceedings of European Safety and Reliability Conference (ESREL 2020), Venedig, Italien, 2020
Müller, Frank; Zeiler, Peter; Bertsche, Bernd
