| Water is the source of life on earth,the blood of the biosphere,and the material basis for human survival and development.Affected by the substantial growth of the world’s population,rapid social and economic development,unreasonable development and utilization of water resources,and excessive discharge of industrial and agricultural pollutants,the deterioration of water quality has become increasingly serious,and sudden water environmental pollution accidents have occurred frequently.In this context,water quality safety early warning system has gradually become a focus of attention by environmental management departments and scientific research scholars in various countries.The conventional water quality safety early warning system can raise the alarm only for the gradient accidents and sudden accidents with the appearance of deterioration,but for sudden accidents with no apparent sign,it is powerless.This kind of accidents are usually caused by slight changes of multiple factors and erupt through time and space accumulation and their joint function.In response to this problem,this article firstly based on the traditional fuzzy comprehensive evaluation model,based on the transfer ladder theory of Petri nets,made structural adjustments to the logical topology network of the traditional index system,and improved the feedback mechanism of the model to the factor layer.The construction of an evaluation model that can reflect the sudden water environment pollution accidents caused by the subtle changes of hidden danger factors within the normal range through time and space accumulation and joint action.Then,the BP neural network model and the improved comprehensive evaluation model are organically combined,and the current monitoring data of the water quality indicators in the study area are used as input samples to check the model,and finally build a water quality safety early warning model.Finally,on the basis of the established early warning model,use computer language for programming,establish a safety early warning system including a visual interface.A water quality safety early warning system based on hazard identification and response is established.The model was applied to the evaluation and early warning research of Taihu Lake water quality safety from 2007 to 2020,and the results showed that:(1)Based on the inclusion of total nitrogen,total phosphorus,transparency,dissolved oxygen,chemical oxygen demand and other indicators into the step factor,the water quality safety early warning system constructed can respond in time to the water bloom incidents that have occurred in Taihu Lake.Therefore,based on the transfer ladder theory of Petri network,the improved comprehensive evaluation model solves the calculation defect of the relative weight in the traditional evaluation model.When responding to sudden water pollution accidents,the evaluation results can accurately reflect the actual development and changes of the safety state in a timely and accurate manner.(2)From 2007 to 2015,the water quality and safety of Taihu Lake improved year by year,and the water quality improved year by year,but each year the data showed a trend of low in the middle and high in the middle.This is mainly due to the relatively high water temperature of Taihu Lake from May to October each year,and sufficient nutrients such as nitrogen and phosphorus,which induce the massive growth and reproduction of cyanobacteria in the lake,resulting in a lower water quality safety level.(3)From 2016 to 2020,the water quality of Taihu Lake is generally at a light pollution level,and the water quality safety status is safe,which is consistent with the actual water quality status of Taihu Lake in recent years.However,considering the geographical location and water quality characteristics of Lake Taihu,the water quality management of Lake Taihu still cannot be taken lightly.It is necessary to continue to implement relevant preventive and control measures to ensure that the water quality of Taihu Lake can match the rapid economic development. |