A novel encapsulation of 16 polycyclic aromatic hydrocarbons in petroleum sludge with palm oil fuel ash binder; an optimization study and sensitivity analysis using machine learning application.

Palm oil fuel ash (POFA) has limited use as a fertilizer, while contribute effectively to the environmental contamination and health risks. Petroleum sludge poses a serious effect on the ecological environment and human health. The present work aimed to present a novel encapsulation process with POFA binder for treating petroleum sludge. Among 16 polycyclic aromatic hydrocarbons, four compounds were selected for the optimization of encapsulation process due to their high risk as carcinogenic substrates. Percentage PS (10-50%) and curing days (7-28 days) factors were used in the optimization process. The leaching test of PAHs was assessed using a GC-MS. The best operating parameters to minimize PAHs leaching from solidified cubes with OPC and10% POFA were recorded with 10% PS and after 28 days, at which PAH leaching was 4.255 and 0.388 ppm with R2 is 0.90%. Sensitivity analysis of the actual and predicted results for both the control and the test (OPC and 10% POFA) revealed that the actual results of the 10% POFA experiments have a high consistency with the predicted data (R2 0.9881) while R2 in the cement experiments was 0.8009. These differences were explained based on the responses of PAH leaching toward percentage of PS and days of cure. In the OPC encapsulation process, the main role was belonged to PS% (94.22%), while with 10% POFA, PS% contributed by 32.36 and cure day contributed by 66.91%.

Metronidazole photocatalytic degradation by zinc oxide nanoparticles synthesized in watermelon peel extract; Advanced optimization, simulation and numerical models using machine learning applications.

Antibiotics in water systems and wastewater are among the greatest major public health problem and it is global environmental issues. Herein a novel approach for the photocatalytic degradation of metronidazole (MTZ) by using eco-green zinc oxide nanoparticles (EG-ZnO NPs) which biosynthesised using watermelon peels extracts has been investigated. Mathematical prediction models using an adaptive neuro-fuzzy inference system (ANFIS), artificial neural networks (ANN) and response surface methodology (RSM) were used to determine the optimal conditions for the degradation process. The FESEM analysis revealed that EG-ZnO NPs was white with a spherical shape and size between 40 and 88 nm. The simulation process for the mathematical prediction model revealed that the best validation performance was 55.35 recorded at epoch 2, the coefficient (R2) was 0.9967 for training data, as detected using ANN analysis. The best operating parameters for MTZ degradation was predicted using RSM to be: 170 mg L-1 of EG-ZnO NPs, 20.61 mg 100 mL-1 of MTZ, 10 min exposure time, and a pH of 5, with 77.48 vs 78.14% corresponding to the predicted and empirically measured respectively. The photocatalytic degradation of MTZ was fitted with pseudo-first-order kinetic (R2 > 0.90). MTZ lost the antimicrobial activity against Bacillus cereus (B. cereus) and Escherichia coli (E. coli) after degradation with EG-ZnO NPs at the optimal conditions as determined in the optimization process. These findings reflect the important role ANFIS and ANN in predicting and optimising the efficacy of engineered nanomaterials, including EG-ZnO NPs, for antibiotic degradation.