On the correlation between Turbidity, Conductivity and COD

Abstract

Introduction COD is one of the most important parameters for wastewater quality. Measuring COD is an expensive and time consuming process. There is a need for an inexpensive, accurate, reliable and continuous technique to monitor COD. Turbidity can be used as a proxy for COD, but is not accurate enough for some applications. Some research suggests that conductivity measurements can be used to improve the turbidity prediction of the COD. This has led to the research question: “Can combined turbidity and conductivity measurements be used to accurately estimate the COD of wastewater in a Dutch sewer?” Research To research this fifty samples were taken from the sewer system of Ulvenhout, The Netherlands during three storm events in 2012. The samples are analyzed in a laboratory for COD concentration, turbidity and specific EC. The results are analyzed with MATLAB to find a correlation between turbidity, conductivity and COD. Twelve different fitting formulas were tested using a Levenberg-Marquart algorithm to find the best relation. Conclusions and Recommendations It is concluded that measurements of conductivity and turbidity can be used to estimate the COD of wastewater to some extent. The r2 of the correlation found during this research is about 0.65-0.69. The turbidity is dominant in the estimation of the COD and the added value of the conductivity measurements is limited. Although the conductivity could be used to make the fit a bit better, the improvement is normally not worth the effort of the additional measurements and may even be insignificant. The highest correlation is found with quadratic fits, but the r2 of linear fits is only slightly lower then the r2 of quadratic fits. Since the linear fit of turbidity to COD (without conductivity) has a similar r2, but is simpler, the use of the linear fit is preferred in many cases. The results of the study are only partly in line with literature.