| Beijing-Tianjin-Hebei region, the Yangtze River Delta and the Pearl River Delta region are the three most developed areas in China with the highest urbanization level, so it is of great scientific and practical significance to do research on the relations between the climate effect and electricity demand in these three regions. Analyses were made based on three kinds of data:1) daily climate data from national base climatological stations during the period of 1959-2013; 2) social-economic statistical data such as population, GDP and electricity consumption during 1985-2013; 3) 22 model data from CMIP5, including historical modeling data and 3 projection simulations (RCP2.6/RCP4.5/RCP8.5). The main methods used in this paper are degree-day calculating, multiple linear regression and artificial neural network. Besides, the objectives of this paper are to analyze the climate effect of three main urban agglomerations and the correlation between climate indices and electricity demand by developing models, and in the same time, to predict the possible changes of electricity demand under the 3 emission scenarios.To begin with, temperature observations were used to analyze the feature of historical temperature changes in these three regions. The results reveal that:1) the temperature of Beijing-Tianjin-Hebei region increased the most during the past 55 years; and in all regions, temperature between 1980s-1990s and 1990s-2000s rose significantly; the temperature of winter seasons grew constantly while that of summer seasons used to decrease between 1960s-1970s; 2) In Yangtze River Delta, the temperature in coastal areas in Zhejiang Province increased the most, and in the north part of Beijing-Tianjin-Hebei region, temperature rose notably, while in the south part of the Pearl River Delta region that grew significantly.Then, the modeling capabilities of 22 CMIP5 models were tested. The results show that:l) For time changes, the BNU-ESM and CSIRO-MK3.6.0 are well off in simulating the temperature changes in China. MME (Multi-Model Ensemble) can properly reveal the trend of temperature changes in the threeurban agglomerations, but the fluctuations are weaker than that of temperature observations. In terms of spatial changes, MME can reflect the spatial gradient of temperature changes, and the simulations of MME are mostly close to the real temperature of the Pearl River Delta region, but cannot simulate temperaturein Beijing-Tianjin-Hebei region exactly.Afterwards, the power load variations of Beijing during 2006-2100 were examined and the future electricity consumption was predicted with the method of artificial neural network based on the projections of CMIP5 models. The results show that:1) there are obvious weekly/seasonal/festival characteristics in power load changing in Beijing; 2) artificial neural network can simulate the features of historical power load well, and that is also capable of predicting the future electricity demand. It turns out that the further away from now and the higher emission scenarios are, the severer the fluctuations are. That is, more electricity is need in summer seasons and less is need in winter seasons and holidays in the future.Finally, multiple linear regression and artificial neural network were used to simulate and predict the elec-consumption in the three regions. The results show that both of these methods can simulate the historical elec-consumption accurately, but in the case of lacking samples, it is better to use traditional methods like multiple linear regression. |
PDF Full Text Request