| Summary: | 博士 === 國立成功大學 === 建築學系碩博士班 === 93 === In order to predict urban energy consumption, this dissertation focuses on the prediction model of the electricity consumption in urban residential area. The electricity consumption characteristics of 6 categories, including residential town houses, roadside stores, apartment buildings, parks, road lamps and traffic lights, are discussed in this study.
27 residential blocks with 1386 residential town houses in Tainan, Taiwan are selected as samples. Among them, 599 houses are valid samples. The conclusions are as follows: 1.The average EUI is 33.29 kWh/㎡•yr; 2.Concerning the eletricity consumption density, the two- story building is 1.56 times larger than that of 5-story building.Considering orientations, the west-facing houses are 1.33 times the average electricity consumption of those facing north; 3.The air-conditinion electricity consumption accounts for 15.56% of the total. In addition, a simplified prediction model for electricity consumption of residential blocks is developed in this study. For utterly residential buildings, the deviation of the predicted data is about 10.14% to 15.62%. The coefficient of correlation(R value) between the predicted value and the sample is 0.78. The electricity consumption of residential blocks with west-east major axis is truly lesser than that of north-south axis by 12.89%.
Regarding the roadside stores, 59 residential street blocks 555 commercial town houses in Tainan, Taiwan are selected as samples. Among them, 434 houses are valid samples. The conclusions are as follow: 1.The average story still in business usage is 1.31; 2.The average EUI is 153.73 kWh/㎡•yr with a high standard deviation of.306.59 which implies a large difference among samples; 3.All the samples are divided into 38 categories. The EUI of the highest category is 54 times larger than the lowest one; 4. Outdoor temperature was used as the main factor to predict the EUI of the 38 store categries; the R value is 0.861; 5. To enhance the predictability of the model, we put in more factore, including air condition status, width od the road in front, whether or not situated at the corner of the block, the orientation of the bilding. By doing this, the R value increases to 0.932
21 apartment buildings are chosen as samples are used to derive the prediction model. The EUI of private and public utilized area is ised to predict the total electricity consumption.
74 parks, the road lamps are the traffic lights of the entire city are used to derive the prediction model. The average EUI of parks is 2.92 kWh/㎡•yr, and that of road lamps and traffic lights are 3.87 kWh/㎡•yr and 0.34 kWh/㎡•yr, respectively.
Concerning the residential area, 4 prediction models, with 6 categories and 12 formulas, are used to apply to different precision and conditions. 3 models, with 4 categories and 8 formulas, can be apply to larger scale residential areas. The last model is for single block and site prediction.
|
|---|
