Mandau, M. (2019). Development of a concept for implementing predictive maintenance on a pharmaceutical solida packaging line [Master Thesis, Technische Universität Wien]. reposiTUm. http://hdl.handle.net/20.500.12708/80038
predictive maintenance; digital; pharma; packaging; production
de
predictive maintenance; digital; pharma; packaging; production
en
Abstract:
Digitalization is on the rise and is being processed in all branches of industry. Even within production, the value of digital solutions is increasing drastically. One area that has been rarely the focus of digitization initiatives, however, is maintenance. Companies just start to realize that there is big potential in digital approaches in that area. Especially Maintenance offers great potential for cost reduction, as with around 40% of the plant costs it accounts for a large proportion of the total production costs. In this work, therefore, special attention will be paid to the field of maintenance. A concept will be developed to establish a predictive maintenance strategy in the solida packaging department within the pharmaceutical industry. For this purpose, a pilot area is defined at Bayer AG's production site in Berlin, and the most critical plant is worked out through systematic criticality analysis of the packaging plants. The subsequent in-depth analysis of the selected packaging system provides important information about the existing weak points and is supplemented by the consideration of an FMEA and the maintenance plans. The weak points found are then differentiated as to whether it can be eliminated by one-time measures or whether it makes sense to follow a predictive approach. For this purpose, systematic interviews with line operators and experts are conducted and methods such as the cause-effect analysis (Ishikawa) and the specific question of 5x Why the actual causes of error are worked out. Based on the elaborated potentials of the system, additional sensors including the corresponding positioning are recommended for the implementation of predictive maintenance. In some cases, existing sensors are used and data collection is expanded. Corresponding correlation hypotheses are set up for possible algorithms. For the implementation of the predictive maintenance a manual is created, in which is explained, how the typical procedure should look like, which responsibilities exist and with which expenditure is to be counted. Finally, based on a payback calculation, a recommendation for a successful implementation is given and the current cost factors pointed out.