A Study Application of Gray System in the Water Resource Demand Forecasting Model Region of Southern TaiwanRegarding tap water as the instance

• Main Authors: Chin-Che Fu, 傅金車
• Other Authors: Chih-Ju George Jou,Ph.D
• Format: Others
• Language: zh-TW
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/08133427399647981236

碩士 === 國立高雄第一科技大學 === 環境與安全衛生工程所 === 90 === A Study Application of Gray System in the Water Resource Demand Forecasting Model Region of Southern Taiwan Regarding tap water as the instance Student：Chin-Che Fu Advisor：Chih-Ju George Jou,Ph.D Department of Safety, Health and Environmental Engineering National Kaohsiung First University of Science and Technology ABSTRACT Due the changes of the industry and economy structures, the population in Taiwan has increased quickly and the urbanization and industrialization has also developed rapidly. As a result, the water usage pattern has also changed dramatically in recent years. With the popularization of tap water, the demand for water resources has also increased. In particular, the total amount of water distribution has increased from 0.41 billion m3 in 1973 to 2.845 m3 in 2000, which is an increase of 590%. With the 2.494 billion m3 water distribution of 1999, the demand of water has increased 0.351 billion m3 (an increase of 14.07%) in 2000. The average amount of water distribution per person is about 0.345 m3 per day in 2000, which is 0.034 m3 more than the 0.311 m3 of 1999. The increase rate is 10.93%. As the demand grow stronger, the water shortage problem becomes more serious. This is particularly true for this year (2002). Explore new water resources to build new reservoir is the most time-consuming method, but it is the long-term solution for resolving the problem of future high water demand. However, under the resistance of residents, this does not seem to be a feasible strategy. An alternative solution is to coordinate the overall water resources so that water resources can be transferred among different areas. In particular, in large scale, water can be transferred from south Taiwan to north Taiwan and vice versa. In small scale, the water management department in each area can be integrated for mutual support. Different parts of Taiwan have different water usage patterns and demands. To fulfill these demands, an accurate, real-time water demand prediction model is required for each area of Taiwan. Therefore, with limited resources, this study tried to use time series, regression analysis and gray system prediction theory to resolve this problem. With the south Taiwan and tap water management departments as units, variables required for prediction are first collected. Built based on the data from 1970 to 1997, the prediction model is validated by data from 1998 to 2000. Based on the criterion of accuracy and reliability, the optimal prediction model is found to predict the water demand for 2001 to 2016. The prediction results can be summarized as follows. For the optimal time series model, the averages of the absolute error for samples years are 1.81%, 0.77%, 1.55& and 0.92% for the 5th, 6th, 7th and the south Taiwan, respectively. For the validated years, these averages are 16.55%, 1.66%, 5.36% and 7.22%. For regression model, these averages are 1.90%, 0.76%, 1.39% and 0.92% for the sample years and 14.82%, 1.58%, 6.69% and 6.09% for the validation years. For the gray system prediction model, these averages are 1.07%, 0.36%, 0.45% and 0.43%, for the sample years and 2.61%, 1.79%, 6.64% and 4.65% for the validation years. These results show that the regression analysis and the gray system prediction method are better suited for our tap water prediction problem. Despite that fact that there are many commercially available statistical packages, the regression analysis requires relatively large amount of data and is more complicated to implement and is thus more time-consuming. In contrast, the gray system prediction method requires only four data to perform computations. Concluded from the experimental result, in predicting the year water demand, the gray system prediction model provides the most ideal prediction result. It is an accurate, reliable, efficient and effective method.