Modelling Indoor Air Carbon Dioxide (Co2)Concentration Using Neural Network

J-P. Skön, Markus Johansson, Mika Raatikainen, Kauko Leiviskä, Mikko Kolehmainen

2012 · 25 citations · 10 references

DOIFull text

Open access

Concepts

Abstract

The use of neural networks is popular in various building applications such as prediction of heating load, ventilation rate and indoor temperature. Significant is, that only few papers deal with indoor carbon dioxide (CO2) prediction which is a very good indicator of indoor air quality (IAQ). In this study, a data-driven modelling method based on multilayer perceptron network for indoor air carbon dioxide in an apartment building is developed. Temperature and humidity measurements are used as input variables to the network. Motivation for this study derives from the following issues. First, measuring carbon dioxide is expensive and sensors power consumptions is high and secondly, this leads to short operating times of battery-powered sensors. The results show that predicting CO2 concentration based on relative humidity and temperature measurements, is difficult. Therefore, more additional information is needed.

References

10