Application of RBF-ANN model in groundwater quality evaluation of Changchun region

Traditional groundwater quality evaluation methods often need to set the weight of evaluation factors, and the evaluation results were to a large extent influenced by the subjective factors and there would be a poor grade of evaluation result because of one factor that had a larger content. To solve above problems, the artificial neural network theory and ideology was introduced, and use the RBF neural network to construct the multi-layer feed-forward groundwater quality evaluation model. This model was used to evaluate the 25 deep confined groundwater samples´ water quality in Changchun region. According to the water quality characteristic of Changchun region and the demand of drinking water supply, choose 10 evaluation factors. The evaluation results showed that groundwater quality in Changchun region were more than Grade III, With GIS data fusion to analyze and express the spatial distribution information on water quality, from the distribution information, we can get that the ratio of the distribution area of Grade I, Grade II, Grade III water quality in Changchun City to the whole area are 6%, 72%, 22% respectively.