流域突发水污染事故应急处理技术识别研究

In dealing with gusty water pollution accidents in a drainage area, the prompt and suitable prior lash-up measures taken are decisive, which makes the whole pollution treatment process successful or failing. Currently, the lash-up measures based on pre-arranged planning cannot well satisfy the needs of dealing with all kinds of gusty pollution accidents in a drainage area. This study intends to establish the internal mechanism and causality of pollution accidents and the lash-up treatment technology using system theory and Bayes network theory; to combine the interval uncertainty method based on interval mathematics and Bayes network theory, break the limitation of sample information and the problem of multiple uncertain factors, and establish the case information model for gusty water pollution accidents in a drainage area and technical analysis model for lash-up treatment; to develop a method based on deep learning, establish recognition principles based on the correlation of each factors, then realize the selection of the lash-up treatment technology and achieve the recognition and generation of lash-up treatment method. This study faces the actual needs in our country’s environmental management of drainage areas, and based on the analysis of our country’s lash-up methods of dealing environmental incidents, the assimilation and coupling methods of statistics of different sources in a gusty water pollution accidents of a drainage area is studied and developed, and further a recognition method for the lash-up treatment technology for the gusty water pollution accidents in a drainage area is established. This study helps scientifically dealing with the gusty pollution accidents in a drainage area and minimizes the environmental damage, which is of great practical significance.

流域突发水环境污染事故时应急处理技术及时、得当与否，决定了污染事故全过程处置的成败，目前基于预案的应急处理方法无法很好地满足流域突发水污染事故的需求。本研究拟运用系统论和贝叶斯网络理论建立污染事故和应急处理技术的内在机理和因果关系；将基于区间数学的区间不确定性方法与贝叶斯网络理论相结合，突破样本信息有限，不确定因素多的难题，建立流域突发水污染事故案例信息模型和应急处理技术分析模型；发展基于深度学习的方法，根据要素的关联性建立识别原则，进行应急处理技术的搜索识别，实现应急处理技术的识别和生成。研究面向我国流域环境管理实际需求，在分析水环境事故应急处理方法实际情况的基础上，研发流域突发水污染事故中不同源数据的同化与耦合处理方法，建立一种流域突发性水环境污染事故应急技术识别的方法，对科学应对流域水环境突发性污染事故，最大限度降低水环境损害, 具有重大的现实意义。

为在最短时间内削减因突发事故进入水环境中的污染物，降低其对水环境的危害，研究针对最佳应急处理技术快速识别这一应急响应面临的关键问题，构建了包含97种化学品基础信息和污染应急处理技术的流域应急处理技术识别信息系统；结合不同的流域污染事故情景，突破案例信息有限的难题，分别构建了应急技术库中化学物污染事故的应急处理技术识别方法和应急技术库未涵盖化学物污染事故的应急处理技术识别方法。综合运用D-S证据理论、改进层次分析、加权秩和比、优劣解距离等研究方法，进行信息分析处理、 模型构建、应急技术提取。通过对历史流域污染事故的案例演算以及岷江流域设定污染事故的分析，验证了各优选方法的适用性与可行性。当突发水污染事故时，运用该识别方法进行“情景-应对”，可以快速对突发污染事故的污染态势进行甄别，通过数据信息的充分分析，快速提取有效的应急处理技术，实现应急响应工作的数字化、程序化，保障应急处理技术高尖精准，从而快速控制污染，最大限度降低水环境损害。