ROLE OF DIFFERENT PARAMETERS IN THE QUANTIFICATION OF GENERATED SLUDGE IN THE OXYLATOR UNIT OF WATER TREATMENT PLANTS, USING ARTIFICIAL NEURAL NETWORK MODEL (CASE STUDY OF JALALIEH WATER TREATMENT PLANT, TEHRAN, IRAN).

NAEAMIKHAH, N.; NASRABADI, T.; SIRDARI, Z. Z.

In recent years, smart imaging and error detection to achieve optimum performance in water and sewage treatment plants using Artificial Neural Network (ANN) detection technique has had a very active research field and has been regarded as one of the modern methods of modeling. This study has focused on predicting the sludge produced by the oxylator unit of the water treatment plant using an artificial neural network with multilayer Perceptron method. The measured input data: water flow, pH, turbidity, temperature, free residual chlorine, conductivity, consumed lime, coagulant (ferric chloride), and magnum and the sludge produced by the oxylator were examined as output data of the model in two consecutive years 2014 and 2015. At the end, the results of neural network analysis were used to show the effect of each of the parameters in the production of sludge. According to the results, turbidity and pH are the most effective parameters in predicting the sludge produced by the oxylator unit of the water treatment plant. This study indicated that artificial neural network could achieve a 0.99 validity with a regression of 0.9881 and mean square error of 0.006 among the observed and predicted output variables of the model. In addition, artificial neural networks provide an effective tool to analyze the data to understand and simulate the nonlinear behavior of treatment plants.

SEWAGE sludge digestion; WATER treatment plants; ARTIFICIAL neural networks; IMAGING systems; ERROR detection (Information theory); EQUIPMENT & supplies

Applied Ecology & Environmental Research, 2017, Vol 15, Issue 4, p129

1589-1623

Academic Journal

10.15666/aeer/1504_129142