Research On Environmental Problems Of Ningdong Energy Chemical Industry Base Based On SVM

With the rapid development of industrialization, the social economic also has a fast growth. But the resulting environmental problems become more and more serious. In this paper, we use the embedded genetic algorithm optimization of support vector machine (short for GA-SVM)evaluating the ecological environment vulnerability of Ningdong base from macroscopic aspects. The main results is:(1) The overall evaluation of ecological environment vulnerability rating for the Ningdong is â…¡;(2) To analysis the main reasons for the cause of the evaluation result, there has two aspects:one is the objective natural climate factors; another is subjective human factors.Then, we put forward the prediction methods of PM10 concentration for the local residents learn about the future of air condition, reasonable arrangement of work travel reference is given. In this paper, we use the method of least squares and support vector machine (SVM), and put forward the least squares support vector regression machine (LS-SVR) model to predict the concentration of PM10, using the base value concentration, temperature, relative humidity and wind speed, pressure, rainfall, sunshine six factors affecting the concentration of PM10 environmental training test, and compared with the BP neural network (BP-ANN) forecasting model and multiple linear regression (MLR) provision compares the forecasting model. The main results are as follows:(1) The LS-SVR model compare with the BP-ANN model and the MLR model can more accurately to predict PM10 concentration, so it can become the preferred method to forecast the PM10 concentration.(2) PM10 concentration of the initial value had a great influence on PM10 concentration. So it should be controlled in the concrete operation of the enterprise, such as thermal power, coal, coal chemical industry of solid waste.(3) We can increase precipitation or class precipitation method to reduce the concentration of PM10(4) Ningdong base PM10 concentration significant seasonal variation, high performance for spring and winter, summer, autumn is low.