Kim S, Lee H, Kim J, Kim C, Ko J, Woo H, Kim S
Dept. of Environmental Engineering, Pusan National University, Korea.
Water Sci Technol. 2002;45(4-5):405-11.
The genetic algorithm (GA) has been integrated into the IWA ASM No. 1 to calibrate important stoichiometric and kinetic parameters. The evolutionary feature of GA was used to configure the multiple local optima as well as the global optimum. The objective function of optimization was designed to minimize the difference between estimated and measured effluent concentrations at the activated sludge system. Both steady state and dynamic data of the simulation benchmark were used for calibration using denitrification layout. Depending upon the confidence intervals and objective functions, the proposed method provided distributions of parameter space. Field data have been collected and applied to validate calibration capacity of GA. Dynamic calibration was suggested to capture periodic variations of inflow concentrations. Also, in order to verify this proposed method in real wastewater treatment plant, measured data sets for substrate concentrations were obtained from Haeundae wastewater treatment plant and used to estimate parameters in the dynamic system. The simulation results with calibrated parameters matched well with the observed concentrations of effluent COD.
遗传算法（GA）已被集成到国际水协活性污泥1号模型（IWA ASM No. 1）中，以校准重要的化学计量和动力学参数。利用遗传算法的进化特性来配置多个局部最优解以及全局最优解。优化的目标函数旨在使活性污泥系统中估计的和实测的出水浓度之间的差异最小化。模拟基准的稳态和动态数据均用于采用反硝化布局的校准。根据置信区间和目标函数，所提出的方法提供了参数空间的分布。已收集现场数据并将其用于验证遗传算法的校准能力。建议采用动态校准来捕捉进水浓度的周期性变化。此外，为了在实际污水处理厂中验证该方法，从海云台污水处理厂获取了底物浓度的实测数据集，并将其用于动态系统中的参数估计。校准参数后的模拟结果与实测的出水化学需氧量（COD）浓度匹配良好。
