Open AccessComparison of grey-box model and artificial neural network – prediction of surface condensation in residential space
Author(s) -
Eun Ji Ju,
J H Lee,
S H Park,
C S Park,
Myoung Souk Yeo
Publication year - 2019
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/609/3/032016
Subject(s) - artificial neural network , range (aeronautics) , schedule , condensation , humidity , model predictive control , computer science , environmental science , control theory (sociology) , meteorology , artificial intelligence , engineering , control (management) , physics , aerospace engineering , operating system
To apply real-time predictive control using automated devices for minimizing the risk of surface condensation in a residential space, the authors first developed a nodal network model that simulates the flow of moist air and the thermal behavior of a target area with the given boundary conditions of a space. The lumped model was enhanced using a parameter estimation technique based on the measured temperature, humidity, and schedule data. However, the humidity model prediction performance was still outside the valid range. A data-driven model was then developed using an artificial neural network (ANN) with the measured data that was formerly used to enhance the lumped model. Taking into consideration the possible uncertain characteristics of moist air, it was found that the data-driven model was a more suitable option for predicting the condensation as compared to the physics-based and grey-box models. With a stable range of errors between the simulation outputs and measured data, the ANN model could be useful for model predictive control.