Stefan Flad

LineMET - Automatic model based efficiency analysis of bottling plants

Description of Topic

Bottling plants are complex lines of several linked machines and conveyors. Ways to increase their efficiency, e. g. measured as OEE (Overall Equipment Effectiveness), are enhancing the output quality, the performance or the availability of the filling machine. To find negative effects on the performance and the availability a model-based algorithm was developed, that automatically localizes down time causing components. Thus, it supports the introduction and execution of a TPM concept (Total productive maintenance concept).

Materials and methods

The diagnostic algorithm is based on a compositional model that describes the behavior of the components and their interaction over time. The component models are parameterizable and generic, enabling easy adaptation to nearly all existing plants. The algorithm compares the behavior of the physical plant with the normal behavior predicted by the model. If there are inconsistencies, the algorithm identifies the components that caused the downtime. Observations of the actual plant behavior are supplied by a common automatic data acquisition system, that records data in compliance with WS (Weihenstephaner Standard for data acquisition). No additional sensors are required.


The model-based algorithm has already been realized as a demonstrator application which is running on a bottling plant. It is able to identify technical causes of downtime in the plant as well as in simulated scenarios with a correctness of 89 per cent.


Further goals are finding reasons for reduced speed of the filler, considering also deficiencies in the logistic context of the plant, identifying data faults automatically, evaluating the existing algorithm in different plants, and enhancing the accuracy of the algorithm.

Actual materials of the resaerch project

Das IGF-Vorhaben 16116 BG der Forschungsvereinigung Industrievereinigung für Lebensmitteltechnologie und Verpackung e. V., Schragenhofstr. 35, 80992 München, wurde über die AiF im Rahmen des Programms zu Förderung der Industriellen Gemeinschaftsforschung und -entwicklung vom Bundesministerium für Wirtschaft und Technologie aufgrund eines Bundestagbeschlusses gefördert.