Data Pubblicazione:


This paper describes the methodology used for selecting the most influential parameters on the energy performance of a building, using limited computing power.

Proceedings of BS2015: 14th Conference of International Building Performance Simulation Association, Hyderabad, India, Dec. 7-9, 2015.

Elisa Olivero 1, Emmanuel Onillon 1, Patrick Beguery 2, Romain Brunet 2, Sophie Marat 2, Marc Azar3

1 Centre Suisse d'Electronique et Microtechnique SA, Switzerland

2 Schneider Electric Industries SAS, France

3 KTH, Building Services and Energy Systems Division BYV, Sweden


This paper describes the methodology used for selecting the most influential parameters on the energy performance of a building, using limited computing power.

Detailed building energy performance models development and their manual calibration are depicted. Novel interfaces for the connection of a detailed building simulation software with advanced analytics are then presented. After a first screening of the parameters by domain experts, two techniques are deployed on the models to reduce the number of parameters to be considered in an automatic calibration process: boundaries check and Morris method. The methodology is applied on two non-residential buildings.

Finally, the study results are presented, providing deep insight on the buildings energy performance and thus setting the basis for automatic model calibration and faults detection.


Detailed Building Energy Performance Simulation (BEPS) models are increasingly being used as methods for information management along the entire lifecycle of buildings.

Over the last 50 years, many BEPS programs have been developed and enhanced to integrate the latest researches in the field, including novel HVAC systems and controllers.

These models, if used throughout the buildings commissioning phase, might become powerful means in helping building operators and facility managers to assess building energy performance, detect anomalies and suggest management improvement. Today, the numerous available tools may differ in several ways: thermodynamical models, graphical user interfaces, purpose, life-cycle capabilities, transparency and ability to communicate with other programs (Crawley, 2008). Despite the increasing reliability of these programs, one of the major limits to their wide adoption is the number of parameters required for model development. In fact, each model requires hundreds of configuration parameters, which are generally difficult and costly to obtain and which can be responsible for large variance of the model output if not estimated with sufficient accuracy (New, 2012).

For this reason, it is important to select a subset of parameters, which are most likely to explain the model deviation from reality and should thus be determined more carefully for calibrating the model. In the frame of the European project Tribute ( the authors have collaborated to develop a new version of BEPS tools. This new-generation software aims at minimizing the gap between computed and measured energy consumption, thus improving the predictive capability of the tool. In order to achieve this goal, advanced algorithm connecting measured and simulated data for automatic calibration of the buildings parameters will be developed. In order to limit the cost of sensors deployment and the complexity of the optimization algorithm to be implemented, the number of parameters considered for calibration should be limited. For this reason, it is important to select the most influential parameters to be optimized and discard the others. The method chosen for the parameters screening, as well as the insight provided by the results on the buildings under analysis, will be described in the following sections.


Building  modelling Two existing public buildings were modelled. One library located in Torino, Italy and one office building located in La Rochelle, France. The software used was IDA Indoor Climate and Energy (IDA-ICE), developed by EQUA. IDA-ICE is a whole building simulation tool, based on dynamic multi-zone calculations, and providing results on thermal indoor climate and energy consumption.

Simulations are made on variable time steps. The IDA-ICE models were developed by energy simulation experts according to available data on building geometry, construction materials and HVAC systems (Diallo, 2015). It should be noted that the development of a building model consists of different phases, with continuous refinement of subsystems and increasing results accuracy.  ...

>>> READ the proceeding