Simulation of the biochemical and chemical oxygen demand and total suspended solids in wastewater treatment plants: Data-mining approach

Controlling environmental factors are one way to prevent environmental degradation. Wastewater treatment plants are the systems that can help the health of the industry and the environment if they function properly. If monitoring and evaluation are not done on these systems, they can have environmental impacts. In these systems, bringing the product to the standard level can be interpreted as cleaner production. To reduce the costs of monitoring, wastewater treatment processes, mathematical, statistical, and other simulators should be used to manage wastewater treatment systems. Due to the complexity of biological processes and the advancement of database methods, the artificial neural network (ANN) and M5 model tree were used to model the biochemical oxygen demand in a water sample during the period of 5 days at a temperature of 20 °C to degrade the water contents aerobically (BOD5), chemical oxygen demand (COD), total suspended solids (TSS) parameters of wastewater effluents from the treatment plant. The data mining models (ANN, M5 model tree) were applied to the wastewater treatment plant of Ramin thermal power covering 3 years (2013–2015) daily dataset. The appropriate architecture of the ANN and M5 model tree was determined through several steps of training and testing of the models. Statistical indicators were used to evaluate the models and the uncertainty of the models was checked. Results showed that ANN (with the coefficient of determination equal to 0.95, 0.95, and 0.97 for BOD5, COD, and TSS, respectively) had better performance than M5 model tree (with the coefficient of determination of 0.88, 0.90, and 0.83 for BOD5, COD, and TSS, respectively). The M5 model tree is tool for describing and analyzing the range of data and expressing how they communicate with each other. Both models showed robustness, reliability, and high generalization capability. Hence, the data mining techniques (ANN and M5 model tree) can be successfully used for environmental decisions and estimation of missing data in wastewater treatment plants. •Global concern for environmental management of wastewater treatment plant output.•Application of the data-mining models in the wastewater treatment plant.•Comparison of different architectures with the approach of determining the best architectures.•Successfully using artificial neural network and M5 model tree for wastewater data.•Better uncertainty of artificial neural network than M5 model tree.