Study on cascade hydropower alternative schemes based on multi-objective particle swarm optimization algorithm

Alternative scheme is the most important section in the environmental impact assessment (EIA). And the optimization of alternative schemes are the focus research fields in EIA. This paper define the content of alternatives of basin hydropower planning EIA (HPEIA), including the alternative generation about development patterns (single high stage dam or multistage dams) and development scales (dam height, installed capacity). The multi-objective model for alternatives generating of HPEIA has been built, which aims to reduce the impacts on socio-economic development and to increase the generated energy. Objectives of the model include ecological landscape change rate, installed capacity, cultivated field inundation area, and resettlement population, and water level is the decision variable. The basic data is obtained by remote sensing images and processed in ARCGIS: (1) obtain the water level-water surface area curve based on the DEM; (2) divide the landscape patterns in submerged area; (3) decide the centralism population distribution scope and their elevation boundary in the map. Moreover, multi-objective particle swarm optimization (PSO), by which the non-inferior solution sets of the multi-objective model are made, and combining with PSO and grey relational analysis. This model is tested in the hydropower planning alternativesgenerating in Hei Shanxia reach, Yellow River. The comparative analysis of the three optimal sequences was investigated. One is to get the theoretical solution which presents the highest relativity, the other is obtain the solution on the condition of maximum installed capacity, and the third is to achieve the practical solution that is consistent with the decision objectives. The results of the above analysis show that the multi-objective model of alternatives generating of HPEIA can meet the competitive principle. Application of the PSO is suggested to be a helpful tool that solve the multi-objective model. (C) 2019 Published by Elsevier Ltd.