| Summary: | 碩士 === 國立雲林科技大學 === 營建與物業管理研究所 === 100 === Following the development of the economy, citizens’ demand for better life quality is also increasing, thereby exacerbating the problems related to the management of energy utilization. The need for power consumption increases as time passes, and the precise control of buildings’ power consumption has become increasingly important with the rise of environmental awareness and ever higher power rates. Therefore, power consumption analysis will be helpful in obtaining a better understanding of the changes in power utilization, as well as predicting future trends.
Having gathered the power consumption data from 2008 to 2010 for three cases: the library of National Yunlin University of Science & Technology (YunTech), a sample office building and a sample seafood restaurant, this study utilized the periodic traditional Fourier Series Model and the improved periodic and trend Fourier Series Model to make power consumption prediction. Then, by using a programming solution method to reduce the prediction errors, the applicability of the improved Fourier Series Model and the changes of power consumption data were known. Lastly, a comparison was made using the averaging method.
As the research results showed, in the case of YunTech library, due to the University’s highly effective implementation of an energy saving and carbon reduction policy, the decline in power consumption was very obvious, and the predicted result of the improved model, SSE(5.66×109), was better than the traditional model’s SSE(1.44×1010), with an annual precision as high as 95.15%. In the case of the sample office building, the power consumption exhibited a slight increasing trend due to the increasing power rate, and the predicted result of the improved model, SSE(2.5×107) was better than the traditional model’s SSE(2.7×107), with an annual precision of 99.10%. In the case of the sample restaurant, the study could only predict the added equipment as there were no obvious changes in trend, and the improved model’s SSE(1.6×108) was better than the traditional model’s SSE(2.3×108), with an annual precision of 87.70%. In addition, the prediction results of these three cases, by using the Fourier Series, all proved to be better than by using the averaging method.
This study used the improved Fourier Series model of both periodicity and trend for exploration purposes, which was unprecedented in domestic researches. This study aimed to use case studies to discuss its applicability. But as the study only stopped at the initial framework establishment stage, with a periodical comparison made based on subjective selection of arithmetic expression and the lack of historical data, therefore, future discussion is still needed.
|
|---|
