Soft-computing Model For Effluent Ammonia Nitrogen Based On Density Clustering Self-organizing RBF Neural Network

In recent years, with the rapid development of society, the problem of water pollution has become more and more serious, and the eutrophication is an important cause of water pollution. The eutrophication of water body will destroy the ecological environment and affect human health. So it is one of the main objectives of the construction of urban sewage treatment plants in our country to avoid the eutrophication of water body. Because the mechanism of water eutrophication is complexity and the influence factors are numerous, it is difficult to establish its accurate mathematical prediction model, and as a result, it is hard to prevent the eutrophication of water body. The content of ammonia nitrogen in water is the key parameter of eutrophication, and its value can be used to evaluate water quality and prevent pollution. Therefore, in order to realize the real time measurement of ammonia nitrogen in wastewater treatment plant, this paper proposes a soft-computing model of the effluent ammonia nitrogen based on the density clustering self-organizing RBF neural network, and realizes the timely and accurate prediction of effluent ammonia nitrogen.The main contents of the study are as follows:1. According to the mechanism analysis and data processing of effluent ammonia nitrogen in wastewater treatment, a set of secondary variables is chosen for the effluent ammonia nitrogen soft-computing model. The secondary variables selection is the key step in the effluent ammonia nitrogen soft-computing model. In this paper, through the analysis of the mechanism of biological denitrification in the activated sludge process and the measurable variables of the actual sewage treatment plant, the 7 influent factors are employed to describe the effluent ammonia nitrogen. After the normalization of the data, using PCA method to analyze the 7 influent factors. Finally, the auxiliary variable dimension of the effluent ammonia nitrogen soft-computing model is reduced to 5 dimensions by 7 dimension.2. Aiming at the problem that the structure parameter of RBF neural network is difficult to be determined, a density clustering self-organizing RBF neural network is designed. This density clustering algorithm takes the self-organization adjustment of the RBF neural network structure with the density value and Euclidean distance between samples, so as to realize the determination of network structure. Then the gradient descent algorithm is used to train the network parameters to determine the final structure and parameters of the RBF neural network. The nonlinear system modeling simulation results show that the proposed self-organizing mechanism can optimize the structure of RBF neural network, and the prediction accuracy of network is improved.3. A kind of effluent ammonia nitrogen soft-computing model that based on the density clustering self-organizing RBF neural network is established to solve the effluent ammonia nitrogen online measure problem. The proposed density clustering self-organizing RBF neural network is applied to the soft-computing model of effluent ammonia nitrogen. Because the neural network based on density clustering self-organizing RBF can be adjusted according to the characteristics of the sample data, the soft-computing model is much closer to the actual sewage treatment process. The experimental results verify the effectiveness of the established model of the effluent ammonia nitrogen.4. An effluent ammonia nitrogen soft-computing platform is developed. The platform mainly includes the user registration and login module, data processing module, model training and prediction module. Firstly, using Lab VIEW 2012 software designed the interface, and it provides the users for network model selection, parameters initialization. Then, the Matlab and Access 2013 softwares are used to write the background program, and realized the soft-computing model calculation and the users’ information management. Finally, through the information transmission among the modules of user information, data processing and model call, the visualization of the soft-computing process of effluent ammonia nitrogen is realized.