Prediction of indoor environmental parameters for naturally ventilated building using artificial neural network: a reflection of outdoor parameters
This research investigates the environment condition within two types of buildings, naturally ventilated and air-conditioned to gain better understanding of the indoor environment condition of the selected buildings. The main purpose of this study is to develop predictive model to forecast indoor en...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | http://eprints.uthm.edu.my/1728/1/24p%20SURAYA%20GHAZALI.pdf http://eprints.uthm.edu.my/1728/2/SURAYA%20GHAZALI%20WATERMARK.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This research investigates the environment condition within two types of buildings, naturally ventilated and air-conditioned to gain better understanding of the indoor environment condition of the selected buildings. The main purpose of this study is to develop predictive model to forecast indoor environmental parameters using Artificial Neural Network (ANN) technique. Field measurements were conducted at four traditional Malay houses in Peninsular Malaysia to acquire the actual indoor and outdoor data; and to provide data for network training during model development. Hourly time-series data of three indicators including: air temperature, relative humidity and air velocity were used to forecast the indoor environmental parameters. The performance of the developed model was evaluated using R squared (R2) and Mean Square Error (MSE). Network testing was performed to validate the models developed. The accuracy of the model was measured using the Mean Absolute Percentage Error (MAPE). Results from the research show that twelve ANN models with the best structure were successfully developed to forecast indoor temperature, humidity and velocity. The MAPE values for the comparison between the actual and predicted for naturally ventilated building is less than 20 percent for indoor temperature and humidity which can be considered acceptable as suggested by many researchers. However, the MAPE value is more than 20 percent for indoor velocity. As for air-conditioned building, the MAPE values exceed 20 percent for all parameters. It was found that the developed models only applicable for naturally ventilated building. The models in general could predict indoor temperature and humidity pattern with modest accuracy. However, it is not applicable for air-conditioned building due to the different building characteristics.
Keywords: indoor environment, natural ventilation, prediction, Artificial Neural Network. |
|---|