Soft computing techniques for diagnostics and optimisation of building energetic behavior

Abstract

Understanding and controlling building dynamics is a major task and still an open issue for many researchers. In this work, the goal is to achieve energy e ciency improvement through application of soft computing techniques on real case applications. The pattern followed is based on Smart Cities paradigm, aiming to actively involve citizens increasing cities energy e ciency, their well-being and self-knowledge. Developing methodologies applied to real world problems is challenging, since in many cases the knowledge of the dynamics is incomplete, uncertainty is high and signals are noisy. Soft Computing techniques are well suited for coping those problems: as they are based on human mind pattern, they provide good generalisation, iterative learning and self adaptation. This dissertation focuses on diagnostics and optimisation issues applied mainly to buildings, a framework for high level building behavior is proposed based on fuzzy rules data fusion and multiobjective optimisation algorithm for fenestration design and thermal heating control are proposed. Some implementations of these techniques have been applied to ”Smart Village” R.C. Casaccia test case, a small prototype of a Smart City.