Evaluation and modelling of UPM pulp mill wastewater treatment plant in Fray Bentos, Uruguay

The main objective of this research was to assess the performance of the UPM wastewater treatment plant (WWTP) by applying standard wastewater treatment modelling tools. Different operational and configuration alternatives to enhance UPM's WWTP performance were investigated. Modelling of activated sludge systems in municipal wastewater treatment plants has become a useful tool and well established practice despite that over the years the complexity of the models has increased. However, modelling of industrial wastewater treatment plants (specifically pulp and paper mill industries) still exhibits some challenges. Over the years different models have been developed as follows: (i) the activated sludge models no.1 (ASM1) describing carbon oxidation and nitrogen removal; (ii) the activated sludge models no.2 (ASM2) incorporating nitrogen and biological phosphorus removal, (extended to the model ASM2d incorporating the denitrifying PAOs); and (iii) the activated sludge models no.3 (ASM3) correcting the deficiencies which were recognized in the ASM1 model. This research has been conducted at the UPM pulp mill facilities in Fray Bentos, Uruguay. The existent UPM pulp mill in Uruguay is one of the most modern pulp mills in the world. It has an annual production capacity of 1.1 million tons of fully bleached eucalyptus market pulp. The WWTP consists of a conventional aerobic activate sludge system designed for an influent flow of approximately 73,000 m3/day. The UPM pulp mill effluent exhibits a complex mixture of high concentration of organic matter and low concentration of nutrients such as nitrogen (e.g. urea is added as a nitrogen source to cope with the nutrient requirements for biological cell growth). The simulations of the biological wastewater treatment plant were carried out using the software BioWin v. 3.0 (EnviroSim Associates Ltd., Canada). BioWin developed its own integrated activated sludge/anaerobic digestion (AS/AD) model which is applied in the software. This model is a combination of the international models ASM1, ASM2d, and ASM3 proposed by the International Water Association (IWA) incorporating the anaerobic digestion model (ADM) as well. A sampling program and measuring campaign was carried out at the UPM WWTP to characterize the wastewater according to the STOWA protocol. The model calibration was obtained by adjusting some wastewater fractions, and kinetic and stoichiometric parameters as follows: (i) kinetic parameters: aerobic decay rate and aerobic denitrification DO half saturation; and (ii) stoichiometric parameters: nitrogen and phosphorous in biomass, and endogenous residue. In addition to the wastewater characterization, an aerobic batch evaluation was carried out at the laboratory facilities of the UPM WWTP aiming at obtaining the necessary information to calibrate the model. Once of the model was calibrated, the performance of the WWTP was properly predicted. Wastewater data from different periods of time was used aiming at validating the model. A good model that describes the performance of thewastewater treatment plant was obtained. The model describes well both the current operation of the plant, as well as the historical performance. The following key effluent parameters were used to validate the model: COD, TSS, TN and TP. Modelling the plant allows the prediction of different future scenarios. Different potential scenarios were evaluated to reduce the TP concentration in the treated effluent. Including an anaerobic phase prior to the aerobic process may reduce the TP concentration in the treated effluent. In addition, chemical P precipitation at the sweep coagulation range showed that the P concentrations can be further reduces to concentrations below 0.1 mg/L. The WWTP at UPM was properly modelled and the model allows the prediction of actual and future plant scenarios. The model shows that alternatives to the current WWTP configurations can be proposed to improve the plant performance. In addition, chemical P precipitation exhibit great opportunities to reduce P concentrations in the treated effluent even further