A Flexible Software Framework for Self-Adapting Algorithm-Based Fault Detection and Diagnosis in Solar Heating Systems
Abstract
A software framework for algorithm-based fault detection in solar heating systems is described. It automatically links modular algorithms in order to perform the best applicable fault detection methods considering the data available for each monitored system. It follows an object-oriented design, where algorithm objects communicate their needed data and parameter inputs, offered outputs as well as constraints via corresponding objects to a path finder. Special effort is made to enable the creation of flexible measurement data requests that do not need to know sensor names in advance. In this paper the basic concepts and implementation approaches to achieve this flexibility in the software framework are described. Finally, exemplary results are shown for a set of algorithms that were implemented in the software framework to determine gradual leakage in a monitored system and bring the capabilities of the framework to life.