Three-step treatment of real complex, variable high-COD rolling wastewater by rational adjustment of acidification, adsorption, and photocatalysis using big data analysis

Abstract

• An effective simple three-step process for treating rolling emulsified wastewater. • An intelligent big data analysis model for predicting operation parameters. • Employing g-C 3 N 4 as the adsorbent to remove paraffin in oily water efficiently. • Combining ANN and GA algorithm for deep learning and achieving prediction. • Easy regeneration of adsorbent and photocatalyst for reusing. A three-step process of acidification, adsorption, and photocatalysis is developed to treat real rolling wastewater (RRW) from the metal-processing industry, and big data analysis is employed to rationally solve the problem of varying wastewater composition. The results show that the chemical oxidation demand (COD) of RRW with initial values between 50,000 and 74,000 mg/L decreases to below the standard value of 70 mg/L after three cascaded steps: acidification with H 2 SO 4 to recover paraffin for reuse, adsorption by g-C 3 N 4 to remove most organics with long hydrocarbon chains, and photodegradation by TiO 2 under sunlight to completely photodecompose residual organics. An artificial neural network and a genetic algorithm are combined to model RRW treatment and to precisely predict the usable operating parameters of the three-step treatment. An effective technique for completely treating high-COD oily water (this three-step cascade process) and a potential strategy for intelligently solving wastewater variability and complexity (big data analysis) are provided as a reference for wastewater treatment.