Application of neural networks in predicting airtightness of residential units

Krstić, Hrvoje; Koški, Željko; Otković, Irena Ištoka; Španić, Martina

doi:10.1016/j.enbuild.2014.08.007

Cited by 37 publications

(27 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several studies have been developed about short-term load forecasting [5][6][7], often using statistical models [8,9], regression methods [10], state-space methods [11], evolutionary programming [12], fuzzy systems [13] and artificial neural networks (ANN) [14][15][16]. Among these algorithms, ANN has received more attention because of its clear model, ease of implementation and good performance.…”

Section: Introductionmentioning

confidence: 99%

Electrical consumption forecasting in hospital facilities: An application case

Bagnasco

Fresi²,

Saviozzi

et al. 2015

Energy and Buildings

102

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Electrical consumption forecasting in hospital facilities: An application case

Bagnasco

Fresi²,

Saviozzi

et al. 2015

Energy and Buildings

102

View full text Add to dashboard Cite

“…Previous work has shown that neural networks can be used to create a predictive model for measuring the air-tightness of residential units in local conditions, and that these networks show good response [9]. In further work, one could compare the results from a designed mathematical model and from predictive models based on a neural network with a database of measured air-tightness values of buildings in the field.…”

Section: Resultsmentioning

confidence: 99%

“…number of experiments = n k−m (9) where m is the number of defining contrasts. Full FD and fractional FD could be used to fit response surfaces.…”

Section: Types Of Factorial Designmentioning

confidence: 99%

Experimental Design Applied to Modeling of the Air-Tightness of a Building

Miličević

Otković

Koški

2015

e-GFOS

Self Cite

View full text Add to dashboard Cite

Experimental design applied to modeling of air-tightness of a building Abstract: The paper presents experimental designs that can be used in modeling of the air-tightness of buildings as second-order functions using response surface method and corresponding experiment designs. The factors supposed to be significant for a model of building air-tightness-and thus those used in experiment designs-are the heat transfer coefficient for external walls, the heat transfer coefficient of the windows, and the position of the housing units with respect to the building envelope. We compared three-level full factorial and fractional factorial designs, as well as central composite designs and the Box-Behnken design.Keywords: air-tightness, design of experiment, factorial designs, central composite designs, Box-Behnken design VRSTE PLANOVA EKSPERIMENTA ZA MODELIRANJE ZRAKOPROPUSNOSTI ZGRADA Sažetak: U radu su prikazane vrste planova eksperimenta koje se mogu koristiti za modeliranje zrakopropusnosti zgrada, a definirane su preko funkcije drugog reda pomoću metode odzivne površine i odgovarajućeg plana eksperimenta. Faktori koji bi trebali biti važni za model zrakopropusnosti zgrade te stoga upotrijebljeni u planu eksperimenta, su koeficijent prolaska topline za vanjske zidove, koeficijent prolaska topline za prozore i položaj stambenih jedinica u odnosu na ovojnicu zgrade. Uspoređeni su potpuni faktorski plan i djelomični faktorski plan na tri razine, te centralno kompozitni i Box-Behnkenov plan eksperimenta.Ključne riječi: zrakopropusnost, plan eksperimenta, faktorski plan, centralno kompozitni plan, Box-Behnkenov plan

show abstract

“…The hidden layer is constituted by a group of radial basis functions. What is relevant to every node in the hidden layer are the parameter vector and the width [13]. The combination mode of AFSA and RBFNN is shown as Figure 3.…”

Section: Application Of Artificial Fish Swarm Algorithm In Rbfnnmentioning

confidence: 99%