Knowledge-based holistic energy management of public buildings

Much recent work has been conducted towards reducing the energy consumption of mixed mode buildings through rule-based automation. This paper presents a novel rule generation methodology through the simultaneous use of historical sensor data and theoretical models. The theoretical rule generation methodology will be explained, encompassing the development of an energy model to the inclusion of rules within the building specific ontology. This will include the development of use-case scenarios, their refinement through a sensitivity analysis and the generation of rules through an artificial neural network embedded evolutionary optimisation algorithm. Following a review of relevant literature, the methodology will be contextualised within the architecture of the KnoholEM project and applied to the first demonstration object, a mixed use public building in the Netherlands. An example of a rule generated will then be presented and the system will be considered within the existing solution space before comments regarding future developments