Artificial Neural Network (ANN) Approach to Modelling of Selected Nitrogen Forms Removal from Oily Wastewater in Anaerobic and Aerobic GSBR Process Phases

Comparison of benzotriazole ultraviolet stabilizers (BUVs) removal from wastewater after subsequent stages of sequencing batch reactor (SBR) treatment process

The research focused on benzotriazole ultraviolet stabilizers (BUVs) which are commonly used compounds despite being found dangerous, e.g. promoting breast cancer cell proliferation, damaging vital organs such as hearts, brains livers and kidneys. The aim of the study was to analyse the efficiency and removal rate of BUVs from wastewater depending on the quantity of tested compounds and SBR anaerobic-aerobic conditions. The study was conducted in sequencing batch reactors (SBRs - 17 L) with real flocculent activated sludge (8 L) and model wastewater (5 L) containing UV-326, UV-327, UV-328, UV-329 and UV-P from 50 to 600 μg∙L-1. The SBR were operated in 390 cycles of 7 h and 10 min over 130 days. The similarity of the technological parameters of the treatment process to those used in a real wastewater treatment plant was maintained. Efficiency removal of individual BUVs was strictly dependent on the dose of compounds introduced into wastewater and ranged from 68.2 to 97 %. Removal of UV-329 occurred with lowest efficiency (from 68.2 to 85.2 %) while UV-326 was most efficiently removed from the wastewater (from 94.1 to 97 %). UV-329 was removed from wastewater with the lowest (0.0968-0.9524 μg∙L-1∙min-1) average removal rate while UV-327 with the highest (0.16-1.3357 μg∙L-1∙min-1), irrespective of BUVs dose in the influent. Secondary release of BUVs into the wastewater occurred in SBR during the settling phase and was dependent on the type and concentration of the BUVs in the raw wastewater. This occurrence was noted for UV-326 ≥ 100; UV-327 = 600; UV-328 ≥ 200; UV-329 ≥ 50 and UV-P ≥ 100 μg∙L-1. The settling phase needs to be shortened to the required minimum. This is an important conclusion for WWTPs in regards to SBR cycle duration and technological parameters of the treatment process.

Predictive modelling and optimization of an airlift bioreactor for selenite removal from wastewater using artificial neural networks and particle swarm optimization

Selenite (Se⁴⁺) is the most toxic of all the oxyanion forms of selenium. In this study, a feed forward back propagation (BP) based artificial neural network (ANN) model was developed for a fungal pelleted airlift bioreactor (ALR) system treating selenite-laden wastewater. The performance of the bioreactor, i.e., selenite removal efficiency (RE_selenite) (%) was predicted through two input parameters, namely, the influent selenite concentration (IC_selenite) (10 mg/L - 60 mg/L) and hydraulic retention time (HRT) (24 h - 72 h). After training and testing with 96 sets of data points using the Levenberg-Marquardt algorithm, a multi-layer perceptron model (2-10-1) was established. High values of the correlation coefficient (0.96 ≤ R ≤ 0.98), along with low root mean square error (1.72 ≤ RMSE ≤ 2.81) and mean absolute percentage error (1.67 ≤ MAPE ≤ 2.67), clearly demonstrate the accuracy of the ANN model (> 96%) when compared to the experimental data. To ensure an efficient and economically feasible operation of the ALR, the process parameters were optimized using the particle swarm optimization (PSO) algorithm coupled with the neural model. The RE_selenite was maximized while minimizing the HRT for a preferably higher range of IC_selenite. Thus, the most favourable optimum conditions were suggested as: IC_selenite - 50.45 mg/L and HRT - 24 h, resulting in RE_selenite of 69.4%. Overall, it can be inferred that ANN models can successfully substitute knowledge-based models to predict the RE_selenite in an ALR, and the process parameters can be effectively optimized using PSO.

Modelling the biological treatment process aeration efficiency: application of the artificial neural network algorithm

The biological treatment process (BTP) is responsible for removing chemical oxygen demand (COD) and ammonia using microorganisms present in wastewater. The BTP consumes large quantities of energy due to the transfer of oxygen using air pumps/blowers. Energy consumption in the BTP is due to low solubility of oxygen, which results in low aeration efficiency (AE). The aim of the study was to develop an AE model that can be used to monitor the performance of the BTP. Multilayer perceptron artificial neural network (MLP ANN) algorithm was used to model AE in the BTP. The performance of the AE model was evaluated using R², mean square error (MSE), and root mean square error (RMSE). Sensitivity analysis was performed on the AE model to determine variables that drive AE. The results of the study showed that MLP ANN algorithm was able to model AE. R², MSE, and RMSE results were 0.939, 0.0025, and 0.05, respectively, during testing phase. Sensitivity analysis results showed that temperature (34.6%), COD (21%), airflow rate (19.1%), and OTR/K_La (15.7%) drive AE. At high temperatures, the viscosity of wastewater is low which enables oxygen to penetrate the wastewater, resulting in high AE. The AE model can be used to predict the performance of the BTP, which will assist in minimizing energy consumption.

Biosorption of LMW PAHs on activated sludge aerobic granules under varying BOD loading rate conditions

Polycyclic aromatic hydrocarbons belong to the main priority substances for the aquatic environment. One of the emission sources of these compounds to environment is wastewater discharged from conventional wastewater treatment systems, which are not designed to cope with this type of pollution. Thus, due to the widely discussed properties of aerobic granular activated sludge in the literature - a conducted study has proven its ability to remove LMW PAHs (naphthalene (Nap), acenaphthylene (Acy), acenaphthene (Ace), fluorene (Flu), phenanthrene (Phe) and anthracene (Ant)) from wastewater by biosorption process at varying loadings of organic compounds expressed as BOD (kg/kg·d) on the activated sludge mass. The maximum biosorption of Nap was 605 µg/kg_d.m., Acy equals to 134 µg/kg_d.m., Ace equals to 355 µg/kg_d.m. Flu equals to 104 µg/kg_d.m. Phe equal to 204 µg/kg_d.m. and Ant equal to 173 µg/kg_d.m. The study showed that the BOD loading rate is one of the factors affecting the biosorption process of LMW PAHs. However, as the amount of adsorbed LMW PAHs increased, the condition of aerobic granular activated sludge deteriorated, which was evidenced by gradual increase in the values of technological parameters of activated sludge (SVI, HRT, SRT) and a smaller increase in activated sludge dry mass.

A Numerical Study on Influent Flow Rate Variations in a Secondary Settling Tank

The secondary settling tank is an essential unit for the biochemical treatment of domestic sewage, and its operational effect influences the quality of effluent. Under the influence of the confluence of rainwater and sewage, wastewater use habits, etc., the inflow of the secondary sedimentation tank changes over time. In this paper, OpenFOAM, an open-source computational fluid dynamics package, was used to study the dynamic behaviors of solid–liquid two-phase flow in the tank under influent flow rate variations. A coupled method including a mixture model, drift equation and a dynamic boundary method is proposed. Numerical investigations were carried out for a 2D axisymmetric sedimentation tank using 12 cases. With increasing influent flow rate, sludge accumulates continuously in the bottom left side of the tank, sludge hopper, and inlet; the sludge blanket thickness near the right end of the tank increases continuously; and the sludge concentration in the tank approximately linearly increases with time, with a low slope. The developed framework is generic and is, therefore, expected to be applicable for modelling sludge sedimentation processes.