Nitrogen prediction model through new hybrid model using ant colony optimization and Elman neural network / Pavitra Kumar

• Main Author: Pavitra , Kumar
• Format: Thesis
• Published: 2021
• Subjects: TA Engineering (General). Civil engineering (General)
• Online Access: http://studentsrepo.um.edu.my/14057/1/Pavitra_.pdf; http://studentsrepo.um.edu.my/14057/2/Pavitra_Kumar.pdf; http://studentsrepo.um.edu.my/14057/
• Institution: Universiti Malaya

The prediction of nitrogen not only assists in monitoring the nitrogen concentration in streams but also helps in optimizing the usage of fertilizers in agricultural fields. A precise prediction model guarantees the delivery of better-quality water for human use, as the operations of various water treatment plants depend on the concentration of nitrogen in streams. Considering the stochastic nature and the various hydrological variables upon which nitrogen concentration depends, a predictive model should be efficient enough to account for all the complexities of nature in the prediction of nitrogen concentration. For two decades, artificial neural networks (ANNs) and other models (such as autoregressive integrated moving average (ARIMA) model, hybrid model, etc.), used for predicting different complex hydrological parameters, have proved efficient and accurate up to a certain extent. The current research work presents the methodology of development of artificial neural network models for the prediction of nitrate-nitrogen and ammonia nitrogen in streams. A new training method is proposed which paves the way of selection of models among the number of models having different combinations of internal parameters providing different levels of prediction accuracy. Using the new training approach, four optimum models have been selected for prediction of both the nitrogen compounds at two different measuring stations i.e. LUI and KAJANG, in Langat River basin in Malaysia. These models provided considerable accuracy of prediction. However, for real time prediction the accuracy needed to be enhanced. To enhance the prediction of the developed artificial neural network models, a new hybrid model was developed. Ant colony optimization and Elman neural network was integrated to form a new ACO-ENN hybrid model. ACO performs a decision-making task and ENN performs learning task. ACO selects a best combination of internal parameters and provides it to ENN for an enhanced starting point, leading to a better result. The same four models were developed based on the same historical data using hybrid modelling technique. The prediction results obtained from the hybrid models had higher accuracy than that obtained from standalone artificial neural network models.