Machine learning vs. statistical model for prediction modeling and experimental validation: Application in groundwater permeable reactive barrier width design.

Permeable reactive barrier (PRB) is an effective in-situ technology for groundwater remediation. The important factors in PRB design are the width and reactive material. In this study, the beaded coal mine drainage sludge (BCMDS) was employed as the filling material to adsorb arsenic pollutants in groundwater, aiming to design the width of PRB. The design methods involving traditional continue column experiments and empirical formulas, as well as machine learning (ML) predictions and statistical methods, which are compared with each other. Traditional methods are determined based on breakthrough curves under several conditions. ML method has advantages in predicting the width of mass transfer zone (WMTZ), which simultaneously consider the characteristics of material, pollutant, and environmental conditions, with data collected from articles. After data preprocessing and model optimizing, selected the XGBoost algorithm based on the high accuracy, which shows good prediction for WMTZ (R2 = 0.97, RMSE = 0.15). The experimentally derived WMTZ values were also used to validate the predictions, demonstrating the ML low error rate of 7.04 % and the feasibility. Subsequent statistical analysis of multiple linear regression (MLR) showed the error rate of 39.43 %, interpret superiority of ML due to the complexity of influencing factors and the insufficient precision of math regression. Compared to traditional width design methods, ML can improve design efficiency and save experimental time and manpower. Further expansion of the dataset and optimization of algorithms could enhance the accuracy of ML, overcoming existing limitations and gaining broader applications.

Utilizing machine learning for reactive material selection and width design in permeable reactive barrier (PRB).

Permeable reactive barrier (PRB) is an important groundwater treatment technology. However, selecting the optimal reactive material and estimating the width remain critical and challenging problems in PRB design. Machine learning (ML) has advantages in predicting evolution and tracing contaminants in temporal and spatial distribution. In this study, ML was developed to design PRB, and its feasibility was validated through experiments and a case study. ML algorithm showed a good prediction about the Freundlich equilibrium parameter (R2 0.94 for KF, R2 0.96 for n). After SHapley Additive exPlanation (SHAP) analysis, redefining the range of the significant impact factors (initial concentration and pH) can further improve the prediction accuracy (R2 0.99 for KF, R2 0.99 for n). To mitigate model bias and ensure comprehensiveness, evaluation index with expert opinions was used to determine the optimal material from candidate materials. Meanwhile, the ML algorithm was also applied to predict the width of the mass transport zone in the adsorption column. This procedure showed excellent accuracy with R2 and root-mean-square-error (RMSE) of 0.98 and 1.2, respectively. Compared with the traditional width design methodology, ML can enhance design efficiency and save experiment time. The novel approach is based on traditional design principles, and the limitations and challenges are highlighted. After further expanding the data set and optimizing the algorithm, the accuracy of ML can make up for the existing limitations and obtain wider applications.

Comparison of effects of multiple oxidants with an ultrasonic system under unified system conditions for bisphenol A degradation.

Bisphenol A (BPA) as a trace contaminant has been reported, due to widespread use in the plastics industry. This study applied the 35 kHz ultrasound (US) to activate four different common oxidants (H2O2, HSO5-, S2O82-, and IO4-) for BPA degradation. With increasing initial concentration of oxidants, the degradation rate of BPA increased. The synergy index confirmed that a synergistic relationship between US and oxidants. This study also examined the impact of pH and temperature. The results showed that the kinetic constants of US, US-H2O2, US-HSO5- and US-IO4-decreased when the pH increased from 6 to 11. The optimal pH for US-S2O82- was 8. Notably, increasing temperature decreased the performance of US, US-H2O2, and US-IO4- systems, while it could increase the degradation of BPA in US-S2O82- and US-HSO5-. The activation energy for BPA decomposition using the US-IO4- system was the lowest, at 0.453nullkJnullmol-1, and the synergy index was the highest at 2.22. Additionally, the ΔG# value was found to be 2.11 + 0.29T when the temperature ranged from 25 °C to 45 °C. The main oxidation contribution is achieved by hydroxyl radicals in scavenger test. The mechanism of activation of US-oxidant is heat and electron transfer. In the case of the US-IO4- system, the economic analysis yielded 271 kwh m-3, which was approximately 2.4 times lower than that of the US process.

Zero-valent iron based materials selection for permeable reactive barrier using machine learning.

The zero-valent iron (ZVI) based reactive materials are potential remediation reagents in permeable reactive barriers (PRB). Considering that reactive materials is the essential to determining the long-term stability of PRB and the emergence of a large number of new iron-based materials. Here, we present a new approach using machine learning to screen PRB reactive materials, which proposes to improve the efficiency and practicality of selection of ZVI-based materials. To compensate for the insufficient amount of existing machine learning source data and the real-world implementation, machine learning combines evaluation index (EI) and reactive material experimental evaluations. XGboost model is applied to estimate the kinetic data and SHAP is used to improve the accuracy of model. Batch and column tests were conducted to investigate the geochemical characteristics of groundwater. The study find that specific surface area is a fundamental factor correlated with the kinetic constants of ZVI-based materials, according to SHAP analysis. Reclassifying the data with specific surface area significantly improved prediction accuracy (reducing RMSE from 1.84 to 0.6). Experimental evaluation results showed that ZVI had 3.2 times higher anaerobic corrosion reaction kinetic constants and 3.8 times lower selectivity than AC-ZVI. Mechanistic studies revealed the transformation pathways and endpoint products of iron compounds. Overall, this study is a successful initial attempt to use machine learning for selecting reactive materials.

Evaluation of anode materials in sonoelectrochemistry processes: Kinetic, mechanism, and cost estimation.

Eco-friendly pollutant treatment technology has a developing tendency in future. The combination of ultrasound (US) and electrochemical (EC) is a promising technology, because they are efficient, clean and environmentally friendly. In this study, the impacts of anode material have been investigated in US (300 kHz) and EC (10V) system. The results of all systems revealed that the kinetic constant decreased with increasing pH. The results are also shown that ΔG# > 0 and ΔH# > 0 during PCP degradation in EC or US-EC systems are non-spontaneous and endothermic reactions. Meanwhile, in the US-EC system, TiO2, Ti4O7, PbO2, SnSb, RuIr, and BDD, except for TiO2, all the anode materials showed a synergistic index (SI) of 106-197%, and the activation energies were 19.32, 33.4, 33.74, 32.84, 10.41, 36.44 kJ mol-1, respectively. In EC and US-EC systems, PCP can be completely mineralized by BDD anode within 30 min. TBA scavenger experiments verified that hydroxyl radicals were the main oxidant in each system using BDD and PbO2 anode. As a result of estimating the cost according to the anode material when removing PCP using the EC or US-EC system, BDD was the smallest in the two systems, 1.58 and 1.12 $ m-3, respectively. Finally, this study may serve as a reference for implementation of US-EC system in wastewater treatment.

Evaluation of self-oxidation and selectivity of iron-based reductant in anaerobic pentachlorophenol contaminated soil.

Soil contamination due to chlorinated organics prompts an important environmental problem; however, the iron-based reduction materials and complicated ground environment are the main barriers to implementation and promotion of in situ soil remediation. Therefore, this study aims to evaluate the reductants zero-valent iron (ZVI) and its activated carbon composite (AC-ZVI) in terms of their self-oxidation and selectivity in soil experiments. The results indicated that saturated moisture conditions were beneficial for degradation due to the dispersal of the pollutants from soil particles. Particularly, increasing the water/soil ratio to the over-saturated state would decrease the selectivity of ZVI and AC-ZVI. Meanwhile, increasing the reductant loading decreased the selectivity of ZVI and AC-ZVI, whereas the high initial concentration increased the selectivity of AC-ZVI. In addition, the self-oxidation of ZVI (3.0 ×10-3 h-1) is 4.2 times higher than that of AC-ZVI (0.7 ×10-3 h-1), and the selectivity of AC-ZVI (48%) is 6.9 times higher than that of ZVI (7%), which confirmed that AC-ZVI is a superior iron-based amendment in saturated moisture conditions. Therefore, this study provides a reliable and feasible evaluation method for in situ remediation process, and deepens the understanding of the effects of moisture contents.

Arsenic adsorption study in acid mine drainage using fixed bed column by novel beaded adsorbent.

The downflow fixed-bed column adsorption-desorption of arsenic by the beaded coal mine drainage sludge-Youngdong (BCMDS-YD) adsorbent was experimentally studied. The specific surface area of BCMDS-YD synthesized using inorganic binding was 178 m2 g-1, and the pHIEP was 5.32. The XRD analysis revealed that it was composed of calcite and schwertmannite. As a result, an increase in the inflow rate resulted in an earlier column exhaustion. The breakthrough curve indicated that a smaller adsorbent particle size and lower influent pH prolonged the column life span. Thomas logistic model was applied to fit the breakthrough curve by linear and nonlinear regression. Under the condition of an influent flow rate of 2.65 mL min-1 (EBCT 40 min), an influent arsenic average concentration of 0.5-1 mg L-1, an influent pH of 7.6, an adsorbent mass of 100 g, an adsorbent grain size of 1.40-1.70 mm, and an operating temperature of 25 °C, the equilibrium adsorption capacity reached 4.56 mg g-1. The mechanism of arsenic adsorption is adsorption and precipitation. As a result of the adsorbent reuse experiment, it was judged that it could be reused with good results in all three cycle experiments. The cost of treating arsenic with the BCMDS-YD adsorbent was 0.145 $ per m-3. The results of this study show examples of sustainable development concepts in mining drainage, and BCMDS-YD can effectively remove arsenic and other heavy metals from acid mine drainage.

Quantification of perfluorooctanoic acid decomposition mechanism applying negative voltage to anode during photoelectrochemical process.

Perfluorooctanoic acid (PFOA) is a carcinogen with a high binding energy between fluorine and carbon and is symmetrically linked, making it difficult to treat. In this study, a self-doped TiO2 nanotube array (TNTA) was used as the anode and platinum as the cathode to quantify the PFOA removal mechanism using a photoelectrochemical (PEC) system. The external voltage was negative compared to that of the anode. In addition, NO3- and t-BuOH were used as scavengers to quantify the PFOA oxidation/reduction mechanism in the PEC system. As a result of the study, TNTA crystals are TiO2 anatase, and the band gap energy was 3.42. The synergy index of PEC was 1.25, and the best electrolyte was SO42-. The PFOA decomposition activation energy corresponds to 70.84 kJ mol-1. Moreover, ΔH# and ΔS# correspond to 68.34 kJ mol-1 and 0.190 kJ mol-1 K-1, respectively. When the external negative voltage was 1 V, the contributions of the oxidation/reduction reaction during PFOA decomposition were 60% and 40%, and when the external negative voltage was 5 V, the contributions of the redox reaction were 45% and 55%. As the external negative voltage increased, the contribution of the reduction reaction increased as the number of electrons applied to the anode increased. When PFOA was decomposed, the by-products were C7F13O2H, C6F11O2H, C5F9O2H, and C4F7O2H, respectively. This study is expected to be used as basic data for research on the effects of other factors on the oxidation/reduction as well as the selection of anode and cathode materials on the decomposition of pollutants other than PFOA when using a PEC system.