Content area
Full Text
ABSTRACT
Artificial intelligence and data-driven modeling are becoming more prominent in the building, and construction sectors. Physics-based models usually require significant computational power and a considerable amount of time to simulate output. Therefore, data-driven models for predicting the physical properties of buildings are becoming increasingly popular. The objective of this research is to introduce Artificial Neural Networks (ANNs) methods as a means of representing the physical properties of buildings. Achieving this goal will illustrate the future capacity of integrated neural networks in building performance simulations. The Annual Radiation Intensity Neural Network (ARINet) demonstrates the feasibility of using a 3D convolutional neural network to predict the surface radiation received by building façades. The structure of ARINet is composed of 3D convolution, fully connected, and 3D deconvolution layers. In this research, it was trained on 1,692 datasets and validated by 424 datasets generated by a physical simulator. ARINet showed errors in 0.2% of the validation sets.
INTRODUCTION
In the present of the Big Data era, it is becoming more and more common to employ data-driven models, especially when physical models may not fully explain the operational environment (Simon, 2019). In building physics, models are useful for clarifying a building's physical properties and when making inferences about the future, as well as for providing feedback on design changes and facilitating optimization. With recent increases in computing power and the substantial availability of data sources, the combined use of both modeling techniques is likely to be essential to the future of building performance simulation (BPS). Physicsbased models designed to examine the surface of the earth with conservation laws. Unlike conservation laws, models used empirical methods are mostly inductive and based on observable phenomena (Goldstein & Coco, 2015). For example, building a physical sky model requires a certain empirical model to calculate the local impact of diffuse solar radiation on a horizontal surface. However, it may also require a number of assumptions to predict the surrounding natural phenomena (Han, Malkawi, & Gajos, 2019).
As more data are made available, it is becoming increasingly difficult to incorporate all available sources and fewer assumptions into a single predictor. It can be argued that the empirical parameterization of numerical models should be conducted using ANNs methods because this type of tool is...