Modeling of sakarya landfill leachate using gene expression programming

Over the years, the consumption of people increased in parallel with the increase in population. However, it has become an important problem in increasing the amount of waste. Many environmental problems arise by storing these kind of wastes. One of the problems arising out of damages to soil, people, animals and plants is leachate and wastewater. Leachate is a type of wastewater generated by the accumulation of liquids from snow, rain and waste in the solid waste field at the bottom of the solid waste storage area. Since leachate contains high organic and inorganic pollutant loads, careful and sensitive work is required in their disposal. One of the most common methods used in the disposal of leachate is to discharge it into sewer system. For this disposal method, it is very important to know the characteristic values of leachate and to monitor it in certain periods. The analyzes required to be carried out in accordance with the relevant legislation and the continuity of these analyzes require significant costs and time. In order to reduce the costs of these analyzes and save time, the models which can predict the related parameters can be developed by using the analysis data obtained previously. The frequency of the analyzes to be made with the developed models will be reduced and both the cost of analysis and time will be saved. In this study, it is aimed to develop a model that can predict Biochemical Oxygen Demand (BOD) parameter (output parameter) using Chemical Oxygen Demand (COD), Total Kjendal Nitrogen (TKN), Total Chromium (TK), Lead (Pb), Iron (Fe+2), Fluoride (F-), Copper (Cu), Zinc (Zn) and pH parameters (input parameters) of solid waste leachate by Gene Expression Programming (GEP) technique. Accordingly, scenarios with different combinations of input parameters have been developed and model equations which predict the GEP and BOD parameters have been developed. For this purpose, scenarios were developed for different combinations of input parameters in BOD estimation and model equations were developed with GEP. Among the equations obtained according to the scenarios, the most accurate estimation of BOD5 was reached by using COD, F-, Cu, Zn and pH parameters with values for training data set (R2 = 0.860) and test data set (R2 = 0.929). As a result, it is seen that the developed equation makes a good estimation.