[A-HW18] 水循環・水環境

[AHW18-05] Development and Application of Distributed Multi-objective Simulation-Optimization Model for Water Resource Supply and Allocation across Multiple Watersheds

*Chien-Lin Huang¹、Nien-Sheng Hsu¹、Gene J.-Y. You¹ (1.National Taiwan University, Taiwan)

キーワード：Drought, water resource system, network flow model, water supply allocation, simulation optimization model, multi-objective operation

This research aims to analyze the transferability of river canals, the diffusion of water shortages in supply areas, the capacity of water purification scheduling, measures for anti-drought and water supply reduction, the replenishment rate of aided water sources, the release-runoff characteristics, withstand ability for water shortages, and the supporting allocation ability of water sources to develop governing equations for the network flow model of the water supply system across multiple watersheds in Taiwan. Accordingly, this enables the model construction of a multi-objective water supply allocation simulation optimization for various industries such as agriculture, industry, and people's livelihood. Furthermore, the impact of supply allocation models and unequal distribution of hydrological conditions on the water scarcity trade-off are assessed. To achieve this, the traditional optimization model was enhanced, and the objective function of water supply and allocation operations was modified to minimize the Modified Shortage Index (MSI) for various industries and agricultural water demand nodes. The decision variables are the amount of water released from the reservoir and the amount of water diverted by the water intake. The constraints include the water flow mass conservation continuity equation and discount supply limit for each node, such as reservoir, water demand/diversion/intake, lateral water source confluence, etc. We created a simulation optimization model for water supply deployment across northern Taiwan using the General Algebraic Modeling System (GAMS). The systematic model water source incorporates the Dahan Shihmen Reservoir, Sindian Feitsui Reservoir, and Touchian Baoshan II Reservoir. Also, it takes into account the impact of the Bansin Phase II and Taoyuan-Hsinchu backup supporting system. The simulation optimization model constructed by the GAMS is solved by the constrained nonlinear programming quasi-Newton method, which can optimize the spatiotemporal pattern of water supply allocation with a smaller MSI (=0.06) than that of linear programming solutions in the calibration period of 2001-2011-year water demand scenario. The water shortage rate of optimized water supply allocation for the verification period from 2012 to 2020 years shows that during the three drought years, the model tends to give priority to the upstream Shihmen irrigation area, causing the average water shortage rate in the downstream Taoyuan irrigation area (=15.4%) to be higher than that of Shihmen irrigation area (=6.76%) because the release operation becomes smaller and more conservative. To achieve a water shortage rate of less than 3% in tap water supply areas in drought years during the 50-year return period, Shihmen, Taoyuan, and Hsinchu agricultural irrigation areas need to be fallow to reduce water supply by 18.34%, 41.82%, and 54.38%, respectively. If the agricultural irrigation area is not left fallow in the 100-year return period drought scenario, the average water shortage rate in Hsinchu, Taoyuan, and Bansin water supply areas will reach 29.43%, 18.13%, and 12.58% respectively. Among them, the water shortage rate in Bansin is lower because its water supply comes from the conjunctive operation of Shihmen and Feitsui Reservoir. Testing scenarios from 2020 to 2022 show that opening the Taoyuan-Hsinchu tap water supporting connection pipe can reduce the water shortage rate in the Hsinchu water supply area by 4.16%-5.58%.