Comparison of artificial neural network and multi-kinetic models to predict optimum retention time for dairy wastewater treatment in the integrated fixed-film activated sludge
Fatemeh Shokri Dariyan, Akbar Eslami, Ehsan Aghayani, Mojtaba Pourakbar, Ali Oghazyan
Abstract
Dairy industry wastewater with high levels of nutrients and organic matters can adversely affect water bodies, and ecosystems. Therefore, there is an immediate need to treat such wastewaters before discharge. The dairy wastewater is acceptable to aerobic biological treatment. Given that these processes are generally failed by organic loading shocks, this work attains to solve the issue through innovative operation modes of integrated fixed-film activated sludge process for aiming to break high organic loads and better biodegradation of high-strength wastewaters. In order to achieve the purpose, two novel bioreactors were designed and operated with employing cycle times of 2.5–4.5 h at an OLR of 2.93–16.8 KgCOD/m3.d. The kinetic models and artificial neural networks (ANNs) were applied to reach a better realisation of the data and to predict the optimum retention time. The results show that the bioreactors obtained more than 97% of organic matter removal. In addition, based on kinetic coefficients, the high performance of the bioreactors is proved. Therefore, the investigated processes could be of interest as a suitable treatment process.