Advanced search
Start date
Betweenand

Development of regression models as an assessment tool for social housing's thermal performance in early design stages: considerations on natural ventilation

Grant number: 14/09225-8
Support type:Scholarships abroad - Research Internship - Master's degree
Effective date (Start): November 26, 2014
Effective date (End): May 25, 2015
Field of knowledge:Applied Social Sciences - Architecture and Town Planning - Architecture and Urbanism Technology
Principal Investigator:Karin Maria Soares Chvatal
Grantee:Michele Marta Rossi
Supervisor abroad: S. Ranji Ranjithan
Home Institution: Instituto de Arquitetura e Urbanismo de São Carlos (IAU). Universidade de São Paulo (USP). São Carlos , SP, Brazil
Local de pesquisa : North Carolina State University (NC State), United States  
Associated to the scholarship:13/16768-5 - An assessment tool for social housing's thermal performance: considerations on natural ventilation, BP.MS

Abstract

Building thermal performance simulation tools are important in a design process in all its stages, mainly in the early ones. However, some barriers prevent the effective implementation of these tools in architecture offices such as time, resources and the expertise knowledge required. Thus, regression models (meta-models), which are derived from a large number of robust simulations, are presented by some studies as an accurate and fast option to predict the building's performance. The presented project refers to the development of some activities related to a master national research project [1] by the student in North Carolina State University (NCSU), Raleigh, NC, USA. The national project focuses on natural ventilation strategies. It is part of a greater research [2]that proposes the development of a set of meta-models for the assessment and verification of the thermal performance of social housing still in the early stages of the design process in three Brazilian climatic zones. The methodology stages to be developed by the student in NCSU consist in finalizing the definition of the input data for the parametric simulations that will be used to elaborate and to validate the meta-models. Followed by simulations using EnergyPlus with the Monte Carlo method. Finally, regression analysis in simulations' results will be used to elaborate, validate and apply these meta-models with search algorithms to identify good combinations of design parameter values that optimize thermal performance. Such activities will be aided by a multidisciplinary team with professors who are experts in numerical methods, architectural design and EnergyPlus modeling. The emphasis is on the importance of this internship at NCSU in order to acquire knowledge and learn a procedure that is not fully known at the Institute of Architecture and Urbanism of São Carlos (IAU-USP). The partnership between NCSU and the IAU-USP has already been established. [1] FAPESP 2013/16768-5 - "An assessment tool for social housing's thermal performance: considerations on natural ventilation". [2] This master's degree is linked to a major research. Also bound to the same main research there are two other masters and four professors from North Carolina State University (NCSU). The method presented is the same for the three masters, each of which is responsible for addressing a specific aspect: natural ventilation (specific emphasis of the Master in question), building systems and a building's solar exposure and shading. (AU)

Scientific publications
(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)
ROSSI, MICHELE MARTA; OLIVEIRA FAVRETTO, ANA PAULA; GRASSI, CAMILA; DECAROLIS, JOSEPH; CHO, SOOLYEON; HILL, DAVID; SOARES CHVATAL, KARIN MARIA; RANJITHAN, RANJI. Metamodels to assess the thermal performance of naturally ventilated, low-cost houses in Brazil. ENERGY AND BUILDINGS, v. 204, DEC 1 2019. Web of Science Citations: 0.

Please report errors in scientific publications list by writing to: cdi@fapesp.br.