Prediction And Early-warning Technology Of Environmental Disaster Based On Intelligent Analysis

With the continuous development of China’s industrialization level and the gradual improvement of people’s living standards,the emission of pollutants such as automobile exhaust,domestic sewage,industrial waste gas and wastewater increases,resulting in increasingly serious environmental disasters,which greatly affect the development of China’s natural ecological environment,and even endanger people’s health.In recent years,the national and local government agencies at all levels are very concerned about this,in order to be able to monitor the state of air and water quality in various regions,planning and deployment of many corresponding monitoring systems,the collection and storage of a large number of monitoring data.How to effectively use these data to ensure air quality,water quality,reduce environmental disasters,and timely forecast and early warning of environmental disasters has always been an urgent practical problem to be solved.Based on the environmental disaster data based on time series,aiming at the monitoring data of air and water quality,this paper introduces the improved LSTM neural network into the environmental pollution data prediction,and establishes the meteorological environmental disaster prediction model based on LSTM.Based on the water quality data of multiple parameters,the ADAM method based on learning rate gradient optimization was proposed,and the water quality data based on time series were analyzed and predicted,the gradient optimization LSTM prediction model was established,and the data query and analysis system was designed and developed.The main work of this paper is as follows:(1)The research background of environmental disaster data prediction is studied and analyzed,and the characteristics and prediction methods of environmental disaster monitoring data are analyzed and studied.(2)This paper analyzes the research status of environmental disaster data prediction at home and abroad,analyzes the effects of different theoretical basis of environmental monitoring data prediction methods,focuses on the analysis of LSTM related calculation methods and theoretical models,and gives different improvement strategies for LSTM.(3)In view of air pollution data of environmental disasters,an improved method for analysis and prediction of IGRA disaster data based on correlation is proposed.A time series prediction model based on input gate,output gate and forgetting gate was established,and experiments were carried out using the air pollution monitoring data after desensitization.The experimental results show that the prediction method proposed in this section has a high accuracy.(4)According to the characteristics of water quality in the monitoring area of an ecological reserve in China,a gradient descent optimization water quality prediction model based on the time series pollution monitoring data was proposed by using the desensitized data.In order to avoid the local optimality of the data and the efficiency of the algorithm,the algorithm applicable to the data of the ecological protection area was studied in detail.The accuracy and practicability of Adam adaptive gradient descent method were proved by experimental prediction,and the data analysis system was established.