Desenvolvimento de um sensor online por inferência para a medição indireta de DQO e sólidos suspensos em uma corrente de filtrado de amido

Abstract

Considering the growth of the industrial sector of the modified starch food industry and the environmental impact of the effluent generated in the washing process, it is very important to use a measurement system that ensures accuracy, precision, repeatability, linearity and stability of sampling. Therefore, this study proposes the development of turbidity sensors for the identification of suspended solids (starch) and conductivity for the identification of salts, in the control of the load generated for the WWTP, in Arduino, using a C++ programming language, from a highly automated and analogical measurement system, indirectly correlating the COD and suspended solids in the starch filtrate stream. Thus, studies were conducted about the problems encountered in the company as the large loss of starch to the WWTP, high load value to be treated by the anaerobic reactors and the generation of high concentration of salts in the washing stage. Thus, turbidity analysis were performed using the TB 210 IR equipment and the developed sensor, conductivity with the AK151 device, COD and total suspended solids for starch suspended in distilled water and starch filtrate. The results obtained proved the correlations between turbidity, conductivity, COD and total suspended solids, indicating a linear model, with excellent r2 values close to 100%. Finally, a financial analysis was simulated to indicate the benefits and costs of installing the online sensor by inference, reproducing revenue values for each sensor produced and a profit of R$53,192.84 per year for the company.