Intervention of Artificial Intelligence to Predict the Hydrogen Production and the Degradation of pollutants in Wastewater through Nano photo-catalysis

Abstract

As the world grapples with the dual challenges of water scarcity and efficient energy sources, innovative solutions like wastewater treatment with simultaneous hydrogen generation offer a beacon of hope for a sustainable future. Transforming wastewater from a liability to a valuable resource, simultaneous hydrogen generation revolutionizes the way, the world approach water management and energy production. No doubt, pharmaceutical industry has contributed a lot in the human development by improving the quality and span of life, but the release of micro-pollutants by the industry, either intentional or unintentional, has been disrupting not only the ecological balance, but also adversely affecting the human health. Hydrogen production from pharmaceutical wastewater presents a groundbreaking opportunity to address environmental challenges and promote sustainability. This approach not only reduces pollution and emissions but also provides a sustainable solution for the pharmaceutical industry's wastewater management. Through hydrogen production, pharmaceutical wastewater's hidden potential is unlocked, offering a healthier planet and a low-carbon future. As the industry embraces this eco-friendly practice, it sets a new standard for environmental stewardship, proving that even the most unlikely sources can be transformed into valuable resources. As the Amoxicillin (AMX) belongs to the penicillin class and is a widely prescribed medicine. For several decades, AMX has been extensively utilized in disease control and livestock feed due to its exceptional therapeutic properties. Moreover, agriculture is recognized as a significant contributor to AMX pollution and another prominent source is human excretion. Previous studies have revealed that human body poorly metabolizes AMX, leading to the excretion of 40-80% of un-metabolized AMX following oral administration. The AMX concentration typically ranges from ng/L to µg/L in sewage except in some rare cases that showed mg/L levels. AMX concentration in treated wastewater has been reported to be within the range of 10 to 200 mg/L. A good number of investigations documented in the existing literature have elucidated that conventional wastewater treatment is ineffective in degrading the pharmaceutical compounds and their metabolites. In this regard, several methods have been used to treat antibiotics; the Advanced Oxidation Process (AOP) is the most efficient method, generating hydroxyl radicals and mineralizing the organic compounds into harmless products. This research work assesses the efficacy of the lab synthesized catalyst Ni2P-TiO2 (NPT) using Artificial Neural Network (ANN) for the degradation of AMX in aqueous suspension under Ultraviolet (UV) irradiation. Thorough experimentation was undertaken at 50 ppm antibiotic concentration, using three different compositions of the synthesized catalyst NPT (1:9, 3:7, and 5:5) for five hours. Of the various catalysts tested, the optimum pH conditions, dose, and time of treatment were attained as natural pH, 0.25 g/L, 2 hours respectively. The degradation and mineralization emerged the highest with the respective percentages of 83.00% and 70.00% through NPT (1:9). ANN was applied with the Swish activation function to predict AMX degradation. Chemical Oxygen Demand (COD) removal was considered the key parameter for determining AMX degradation using a three-layer backpropagation neural network. The results obtained through the ANN were similar to the experimental results, and their correlation coefficient was 0.96. The findings show that all the input variables such as pH, catalyst dose, and irradiation time have an immense effect on the degradation efficiency. The study demonstrates that Neural Network (NN) modeling can successfully predict and simulate the degradation process. The Hydrogen production was also assessed at the optimized catalyst dose and optimized catalyst with the same pollutant concentration. The gradient boosting regression modeling was deployed to predict the hydrogen production. The model accuracy was defined through different statistical measures and 93.00% correlation coefficient represented that the experimental results and predicted results were in good agreement. Further, the investigation on the application of an ANN to forecast the efficiency of photocatalytic degradation of the antibiotic AMX was done in the presence of Zirconium Dioxide (ZrO2) as a catalyst under UV radiation. The evaluation of photocatalytic degradation efficiency is based on the reduction in COD. The experiments were conducted by varying the pH and using catalyst doses ranging from 0.05 to 0.25g/L with the duration of 180 minutes. Remarkably, under natural pH conditions and a catalyst dose of 0.20g/L, a degradation efficiency of 66.66% was achieved within 30 min. Furthermore, under optimized experimental parameters, the photocatalytic process produced 62.7 µmol/L of hydrogen gas after 90 minutes. The ANN model effectively predicted degradation efficiency by taking pH, catalyst dose, and time as input variables, and COD removal as the output variable. The number of the hidden neuron count that provided the best fit results were also optimized. The study achieved a strong correlation coefficient of 95.00% between the predicted and experimental results, confirming the model’s suitability for forecasting the degradation process. The study evaluated the effectiveness of the laboratory-made catalyst Ni2P-ZrO2 (NPZ) in the degradation of an antibiotic in an aqueous suspension when exposed to UV light. The degradation experiments were conducted utilizing two distinct photo-catalyst compositions of NPZ in the proportions of 1:9 and 2:8. The most effective experimental results were obtained using a natural pH, a catalyst concentration of 0.20 g/L, and reaction duration of 30 min. after testing the different catalysts. The degradation of AMX was predicted using time series forecasting through the ensemble gradient boosting model. Experimental data were used for training, validating, and confirming time series predictions. The use of ensemble technique highly affected the experimental findings. The model's performance was quite satisfactory in terms of correlation coefficient (94.00%), Normalized Mean Square Error (0.01), and Mean Square Root Error (0.0911) which significantly contributed to the model's accuracy. The study has demonstrated that time series forecasting can be used for predicting the degradation process precisely. Different Machine Learning (ML) algorithms have been used for the prediction of antibiotic removal and hydrogen production. These algorithms have provided excellent results by confirming their applicability in the photocatalytic process.