Energy Performance Characterization of A Test Case “Twin House” in Holzikirchen Based on Trnsys Simulation And Grey Box Model
I. Rehab,Philippe André
DOI: https://doi.org/10.26868/25222708.2015.2688
2015-12-07
Abstract:In the frame of the IEA Annex 58 project, this paper presents an exercise of building energy performance characterization based on full scale dynamic measurements. First focus of the exercise is the verification and validation of the numerical TRNSYS BES-model of the case study test house in Holzkirchen. Second focus is on the modelling of the house through a second order inverse “grey box” model in order to determine reliable performance indicators which include UA-value, total heat capacity, and solar aperture. Final issue is the comparison of predicted indoor temperatures of free floating period, results of TRNSYS and “grey box” models simulation. INTRODUCTION Different models and known methodologies for energy performance characterization can be summarised in three categories of models: white-box, grey-box and black-box models (Bohlin, 1995) (Madsen et al., 1995) (Kristensen et al., 2004). TRNSYS model is typically a white box model based on a complete description of the physical properties of the building. Grey-box model is used when the knowledge of these properties is not comprehensive enough. It is based on a partial dataset and partially on empiricism (Kramer and al., 2012). Black box model is used when parameters have no direct physical meaning. No physical properties knowledge is required for this model (De Coninck et al., 2014). This paper presents an exercise of verification and validation of a test case house using white box and grey-box models. First section describes the test case house, experimental set up and data sets. Second section concerns TRNSYS modelling, according to the "modelling specification report" provided in the exercise. Inputs of the model are the measured outdoor climate data. Part of outputs, are indoor temperatures which will be compared to real measured temperatures of each zone of the house. Third section deals with the modelling of the house as a second order inverse grey box model. Data from a 32-day-long experiment is analyzed and used to fit lumped parameter models formulated as coupled stochastic differential equations. Outputs of the model are the indoor air temperatures. The model is fitted using PEM (prediction error method) techniques with MATLAB. The estimated physical parameters which include UA-value, total heat capacity, and solar aperture for the building are discussed. Last part of the paper presents a simulation of the white and grey box models to predict indoor temperatures of a free floating period. Results of both simulations are compared and discussed. EXPERIMENT SET UP Description of the test case house The experiment was undertaken on a test case house named “House O5” situated at Holzkirchen, Germany (near Munich). The latitude and longitude are respectively 47.874 N, 11.728 E. The elevation above mean sea level (MSL) is 680m. Figure 1 shows an East view the house. Figure 2 shows a vertical section and the internal layout. For the experiment, the layout was divided into north and south areas. South side includes: the living room, the children’s bedroom, the corridor and the bathroom. North side includes: the parent’s bedroom, the lobby and the kitchen. A full specification of the house, including: constructions, windows and roller blinds description, systems of ventilation, heating and cooling, air leakage, ground reflectivity and weather data, was provided in the "modelling specification report" of the exercise.(Strachan et al., 2014). Figure 1 East view of the test case “house 05” Proceedings of BS2015: 14th Conference of International Building Performance Simulation Association, Hyderabad, India, Dec. 7-9, 2015.
Computer Science,Environmental Science,Engineering