Modelling and Analysis of Reliability and Availability by stochastic modelling techniques considering continuously the statistical quality of the initial data

如今，技术系统的可靠性和可用性建模和仿真已采用有效的方法进行。这些方法不局限于特定的系统，并且能够对不同方面进行建模。然而，可靠性和可用性的计算是基于均值分布的。计算数据基础的统计质量，即分布函数的置信水平既不能集成到模型的输入数据中，也不能在分析中考虑。因此，分析得到的结果也是平均值。通常，给出有限的信息来描述系统的行为，例如，系统故障时间的有限样本大小。因此，需要不断考虑输入数据的统计质量，即在可靠性和可用性分析中需要考虑输入数据的置信水平。因此，研究项目的目标是在整体方法中不断整合统计质量，即在建模和技术系统分析中不断考虑。开发这样一种通用的计算与仿真方法是一个具有现实意义的研究目标，具有严肃的科学目的。该方法将建模方法的性能与置信水平的表达性相结合。它具有很高的潜力，可用于各种应用，例如可靠性的规划和分配、测试计划和一般可靠性和可用性值的演示，以及由预后健康管理（PHM）给出的基于建模的预测领域。大大降低了实际可靠性和可用性预测论证过程的不确定性。实现了可靠性和可用性分析的新质量。在研究项目的最后，给出了一个新的方法，作为一个新的Petri网类。基于MATLAB的软件工程原型实现了新方法的首次应用。此外，它们是新研究的基础。通过几个应用阐明了新方法对系统制造商的附加价值。

Today, modeling and simulation of reliability and availability of technical systems are carried out with efficient methods. The methods are not restricted to specific systems and enable the modelling of different aspects. Nevertheless, the calculation of reliability and availability is based on mean value distributions. The statistical quality of the data basis of the calculations, i.e. the confidence levels of the distribution functions can neither be integrated into the input data of the models nor be considered within the analysis yet. Thus, the results earned by the analysis are mean values, too.Usually, limited information to describe the systems’ behaviour are given, e.g. limited sample sizes of failure times of the system. Hence, the continuously consideration of the statistical quality of the input data is necessary, i.e. the confidence level of the input data need to be considered within the analysis of reliability and availability.Therefore, objective of the research project is a continuously integration of the statistical quality within the holistic methodology, i.e. the continuously consideration within the modelling as well as within the analysis of technical systems. The development of such a universal calculation and simulation method represents a practically relevant research goal and has a serious scientific aim. The method combines the performance of the modelling methods with the expressiveness of confidence levels. It has a high potential to be used in varied applications, e.g. the planning and allocation of reliability, the test planning and the demonstration of general reliability and availability values as well as the area of modelling-based prognostics as given by the Prognostic Health Management (PHM). The uncertainty of actual reliability and availability prediction and demonstration procedures can be reduced significantly. A new quality of reliability and availability analysis can be achieved.At the end of the research project, a new methodology is given as a new Petri net class. Prototypical software engineering works with MATLAB enable a first application of the new method. Furthermore, they are the basis of new research. The added value of the new methodology for a system manufacturer are clarified by a few applications.