- Extension and optimization of an expert-based system for fault detection in air handling units: An ontology-based approach
Liebhold, Max; Müller, Matthias S. (Thesis advisor); Müller, Dirk (Thesis advisor); Terboven, Christian (Consultant); Wassermann, Christian (Consultant); Blechmann, Sebastian (Consultant)
Aachen : RWTH Aachen University (2026)
Bachelor Thesis
Bachelorarbeit, RWTH Aachen University, 2025
Abstract
In a previous work, a rule-based fault detection and diagnosis (FDD) system for air handling units (AHU) was developed using ontologies. The aim was to develop a generalized and transferable FDD system that can be applied to systems from a wide range of manufacturers with reduced workload. For the FDD system based on expert knowledge, the APAR rule set was used and an implementation was created in Python. However, this program had very high runtimes of up to 26 minutes, which could make live operation in plants difficult. In this thesis, the program is analyzed in terms of structure and runtime and subsequently extended and optimized in terms of runtime, maintainability and expandability. Simulation data is used to test various approaches developed to improve runtime, such as the management of database connections and the use of local data. An extension with new rules and the development of functions to improve the robustness of the program are presented. Overall, the runtime is accelerated by an average speed-up of 28.52 or up to 59.61 and the program is made easier to maintain and expand thanks to new functions and modules. The optimized and expanded program makes it possible to check sensor data from air handling units for faults with a significantly shorter runtime. This simplifies the testing and further development of the program as a whole. The extensions also offer extended functionality and improve the generalization of the approach.
Institutions
- Faculty of Computer Science [120000]
- Chair of High Performance Computing (Computer Science 12) [123010]