Due to a variety of influence factors, outside dry-bulb temperature takes on a systematic and randomfluctuation. If a deterministic model is used to forecast the dry-bulb temperature, the predicted resultoften has a rough accuracy. Neural network can learn the internal regularity of the sample data bysample training; therefore it has very much adaptability and advantage in the aspects of forecast.The influence factors of outside dry-bulb temperature exist difference in the daytime and the nighttime,which makes the fluctuant regularity of outside dry-bulb temperature inconsistent. Therefore, in thispaper, daily dry-bulb temperature was divided into the daytime parts and the nighttime parts in terms ofa certain principle based on sunshine duration. Through updating the recurrent back propagationnetwork and combined with the characteristic of daytime parts of outside dry-bulb temperature, anhourly forecast model for the daytime parts of outside dry-bulb temperature was put forward inaccording to the dry-bulb temperature historic data of four records each day in meteorological station.
AN HOURLY FORECAST MODEL FOR THE DAYTIME PARTS OF OUTSIDE DRY-BULB TEMPERATURE
![](https://www.aivc.org/sites/default/files/default_images/default_image_5.png)
Year:
2007
Bibliographic info:
The 6th International Conference on Indoor Air Quality, Ventilation & Energy Conservation in Buildings IAQVEC 2007, Oct. 28 - 31 2007, Sendai, Japan