PSUADE使用情况阶段报告.pptVIP

PSUADE使用情况阶段报告 张香香 2012年5月4日 操作步骤： 明确模型的物理过程 参数筛选 响应曲面分析 以Sacramento模型分析为例， 学习不确定性分析方法及PSUADE的使用 Sacramento模型示意图 上表中参数值为“PS_”的参数为待定参数，其余参数我们视其为常数赋值，下面我们通过3 种方法对这6 个参数进行敏感性分析来筛选参数。 Setup ? [hydrosystem@localhost source_codes]$ gfortran -o mopexcal.x *.for ? mv mopexcal.x ../ ? Modify the data path in file “ps_test01.sac.Tmplt” to your own path. ? Modify the parameters you want to analyze in file “ps_test01.sac.Tmplt” by changing the value of each parameter to an character string. Such as, change the value of UZTWM to PS_UZTWM. ? Declare all the parameters in file “driver.py”. Such as, inputNames = [“PS_UZTWM”] ? Change the path of executable file in file “driver.py”. Such as, sysComm = "/home/hydrosystem/2012/ganyj/test_problems/mopexcal.x“ ? Generate an input file (psuade.in). ? Run: psuade psuade.in /home/hydrosystem/2012/zhangxx/test_problems/ Generate an input file psuade> geninputfile Enter the name of the PSUADE input file : ps.in Enter the number of inputs (> 0) : 6 name for input 1 ? UZTWM lower bound for input 1 ? 5 upper bound for input 1 (> lower bound)? 300 name for input 2 ? UZFWM lower bound for input 2 ? 5 upper bound for input 2 (> lower bound)? 150 name for input 3 ? UZK lower bound for input 3 ? 0.1 upper bound for input 3 (> lower bound)? 0.75 name for input 4? LZTWM lower bound for input 4 ? 10 upper bound for input 4 (> lower bound)? 700 name for input5? LZFSM lower bound for input 5 ? 5 upper bound for input 5 (> lower bound)? 500 name for input 6 ?LZFPM lower bound for input 6 ?100 upper bound for input 6 (> lower bound)? 1200 Enter the number of outputs (> 0) : 1 name for output 1 ? output Sampling Method ? MC Sample size (> 0) ? 100 Driver program name : (enter NONE if not needed) ./SACdriver.py maximum number of jobs at one time = 1 How long is each job (in minutes) ? 0 Are the simulations deterministic? (y or n) y Create the file ? (y or n) y psuade ps.in 从上表可以看出，参数LZFPM，LZFSM，UZFWM 对输出结果的影响比较大，而参数LZFPM，LZFSM，UZFWM，UZK 同输出结果具有较强的非线