| Summary: | 碩士 === 國立臺灣海洋大學 === 河海工程學系 === 107 === In recent year, rainfall has caused frequent urban floods. Therefore, central and local governments have enacted building construction regulations to install rainwater detention/retention facilities at newly constructed and renovated buildings to reduce urban flooding. However, rainwater detention/retention facilities have different effectiveness in flood mitigation due to different operation alternatives. Most existing rainwater detention/retention facilities are operated based on the water levels set by the owners which are generally called passive operations. To improve the operation effectiveness, this study considers comprehensive influence factors including storage capacity, types and operation alternatives and proposes a smart operation alternative which combines with rainfall forecasting model to compare between passive operation and smart operation using cases study.
Firstly, this study collects regulations related to the determination of storage capacity and discusses the method of design and arrangement of rainwater detention/retention facilities. Based on those data collected, rainwater detention/retention facilities can be classified into three categories: retention, gravity detention, and pumped detention. The flood peak reduction effectiveness and appropriate building coverage ratio and return period of rainfall are evaluated by the passive operation alternative of rainwater detention/retention facilities using selected build to open ratios and design storms. For those existing passive operations, the limitations can also be found. Furthermore, the conceptual model of smart operation is proposed to improve the effectiveness of passive operation. This study refers to the concept of Internet of Things to establish the framework of smart operation and formulates the strategies to establish smart operation mode for both gravity and pumped detention. For rainfall forecast, the grey theory is adopted and 18 different forecasting models are proposed and compared for three different types of rainfall. Based on the simulation results from both assumed and existing cases, the operation effectiveness of selected alternatives is examined for both rainfall forecast and operation results. The superior alternatives will then be recommended which will provide as a reference for future design.
The results show that the gravity or pumped detention with a capacity of 0.04~0.05m3/m2 and retention with a capacity greater than 0.07 m3/m2 achieve the best flood peak reduction for 5 years return period of rainfall with passive operation. To examine the accuracy of rainfall forecast, three indicators with selected score, which include cumulative rainfall error, efficiency coefficient and relative root mean square error, are used to examine forecast models. Rainfall forecast model is better with higher total score. Forecast models which forecast one unknown value have higher total score. Forecast models which forecast shorter unknown value perform better than those with longer unknown value. Among rainfalls due to frontal, typhoon and convective mechanism, rainfalls by typhoon have better forecast performance than others. Totally, time interval of 10 minutes with 3 known data (model 1) and time interval of 10 minutes with 4 known data with moving average (model 5) have the best forecast performance with total scores of 10.45 and 10.52, respectively. For operational evaluation aspect, five indicators, which include peak reduction ratio, total runoff volume reduction ratio, time lag of peak , storage ratio, and orifice (pump) operation time, are selected to evaluate the operation alternatives. For the case of gravity detention, smart operation is no difference from that of passive operation for peak reduction ratio. For pump detention, peak reduction ratio for smart operation is 20%~30% higher than that of passive operation. Total runoff volume reduction ratio and storage ratio for both gravity and pump detentions are more than 70 percent. Pump duration for smart operation is far less than that of passive operation. Those results reveal that smart operation is far better than that of passive operation. Overall, those alternatives in good performance of forecast will also have good performance in operation. The results obtained in this research will provide as the reference in smart operation for detention pond in the future.
|
|---|
