Study On Optimization Layout Two-stage Method Of Low Impact Development

Many cities have faced serious waterlogging hazards on account of the acceleration of urbanization and the lag of urban stormwater drainage system.In order to remit it effectively,government has propelled the forceful construction of sponge city——establishment of low impact development of rainwater systems.The key of building sponge city is planning layout scientifically and choosing Low Impact Development(LID)facilities and its combined system.However,LID facilities are diverse,choosing and layout is restrained by local natural geographical conditions,hydrogeological characteristic,water resources status,rainfall patterns and so on.Hence,choosing single facility or its combined system and determining optimized layout scheme is a great challenge for structuring low impact development of rainwater systems.To meet the challenge,it’s impending to research layout approach of LID facilities.This research will provide theoretical support and scientific guidance for layout of LID facilities and construction of sponge city.Based on Storm Water Management Model(SWMM)and Matrix Laboratory(MATLAB),taking a drainage in Chizhou city of national-level sponge city pilot city as study subject,starting with prevention and treatment waterlogging,pollution reduction and cost-effectiveness,this thesis build evaluation model and indexes,two-stage optimization framework(analytical module and iteration module)for single LID facilities(storage tank)and LID facilities combinatorial system(Bio-retention cell,permeable pavement,green roof)is proposed and discussed by simulating,main achievements and conclusions are as follows:(1)The quantification method of preliminary scheme is proposed,which combine waterlogging,pollution reduction,cost and other indicators.Comparative independence of indicators are considered to put forward quantification indicator of node which include flood depth and flood duration based on comprehensive analysis of multiple waterlogging indicators.Analytic hierarchy process is elected to analyze nodes quantificationally,considering sensibility of nodes and obtaining final score of nodes through quantitative calculation.Flood reduction efficiency,pollution reduction efficiency and their formulas is defined to establish cost-effectiveness indicator system.Considering final score of nodes and cost-effectiveness indicator system to simulate quantification of preliminary scheme of LID facilities,which determine preliminary scheme afterwards.(2)Two-stage optimization framework is developed based on analytical module.In iteration module,generalized pattern search and genetic algorithm are study intensively and present two two-stage optimization frameworks include analytical module + generalized pattern search and analytical module + genetic algorithm.To test two frameworks,optimized layout of single LID facilities and LID facilities combinatorial system are applied respectively in study region.(3)For single LID facilities,the preliminary scheme based on quantificational computation can screen more reasonable scheme in all flooding nodes,it can achieve final scheme in three iterations.Compared with the preliminary scheme,the optimal scheme can receive more 38.85% runoff and 42.29% Total Suspended Solid(TSS)with the same cost and recover hydrologic environment of study region to pre-development level.Analytical module + generalized pattern search can achieve optimized layout perfectly.(4)Return period and score criteria in the analytic hierarchy process are changed to test the reliability of proposed framework.Optimal schemes can still obtain the desired effects and receive more runoff and TSS than preliminary schemes under 50-and 100-yr return periods.Changing indicator scoring methods has had no impact on the locations identified for the previous scheme,which verify the reliability of framework.(5)For LID facilities combinatorial system,two two-stage optimization frameworks are verified respectively.The results indicate that storage cost(one of optimized objectives)is declined at different levels and annual volume-based control rate(another optimized objective)is declined(1.30%~1.67%)which is optimized by generalized pattern search while genetic algorithm is risen(0.40%~1.26%)with the increase of dimension(8 subcatchments increased to 32 subcatchments).The performance of analytical module + genetic algorithm is more gratifying when optimizes multi-dimension problem.