基于遗传算法的未知环境大气释放源项重建

Source Rebuild for Unknown Atmospheric Releasing Affair Based on Genetic Algorithm

为有效应对危险气体的未知释放事件,选择利用遗传优化技术良好的鲁棒性,结合蒙特拉罗粒子扩散模式、逆向粒子轨迹模式等技术手段,建立了基于环境监测数据与气象数据的源项估算模式。同时,以野外示踪试验数据为基础,为有效屏蔽模拟中涉及的其它不确定因素,开展了在设定条件下的数值试验,对源项估算方案可行性、模式结果有效性进行了分析。结果表明:模式选择遗传算法为主要优化手段,结合扩散/轨迹模拟技术,通过适宜的评判因子搜寻源项参量的技术方案是合理可行的;模式估算结果合理有效,已建立的数十千米尺度区域源项估算模式具备了基于环境监测数据、对未知源项基本参数的估算能力。最后,本文讨论了对全部有效监测数据应用的必要性,提出了模式未来需要改进的技术方向。

To deal with the unknown releasing event of dangerous gases effectively, the genetic optimization （GA） techniques with good robustness combined with Monte-Calo particle dispersion model, particle reverse trajectory models and other technical means were used to built the source rebuild model, to estimate the source parameters based on environmental monitoring data and meteorological information. Depending on the field experiment data, the results of numerical experiment show： Scheme of rebuilding source parameters in the model, which chooses genetic algorithms as its main optimization means, combined with diffusion/trajectory simulation tech- niques and appropriate evaluation factor, is reasonable and feasible. The results of model were effective, and the built 10 km-scale regional source parameters rebuild model had the estimating a bility based on the data of environmental monitoring and the unknown source basic parameters. At last, the necessity to use all significant monitor data was discussed.