TL;DR
This paper presents a novel photo-Fenton process for efficiently degrading styrene gas in nitrogen-sealed storage tanks, achieving over 96% removal under optimized conditions. They also developed a BAS-BP neural network model to predict styrene residuals, offering a valuable tool for real-time monitoring and control in industrial settings.
What Problem Does This Paper Address?
Styrene is a common volatile organic compound (VOC) used in various industries, but its storage in nitrogen-sealed tanks poses safety hazards due to its reactivity and potential for degradation. Traditional methods like combustion and activated carbon adsorption have drawbacks, including safety risks, increased costs, and clogging issues with activated carbon due to styrene's self-polymerization. Existing advanced oxidation methods for VOCs often lack specific applications for styrene in nitrogen-sealed environments, especially considering the need to minimize oxygen presence for safety. This paper aims to address this gap by developing an effective and safe method for styrene degradation in such conditions.
Methods
The core of the study is the photo-Fenton degradation process. This method utilizes the combination of UV light, ferrous ions (Fe²⁺) as a catalyst, and hydrogen peroxide (H₂O₂) to generate highly reactive hydroxyl radicals (•OH). These radicals then break down styrene into less harmful substances. The researchers systematically investigated the impact of key parameters on styrene removal efficiency:
- Ferrous ion (Fe²⁺) concentration: Explored from 0 to 0.6 mmol/L.
- Hydrogen peroxide (H₂O₂) concentration: Varied from 0 to 30 mmol/L.
- pH: Tested at 3.0, 5.0, and 7.0.
The experimental setup involved a quartz glass reactor with a diving UV lamp, where nitrogen gas carrying styrene vapor was bubbled through the solution. By-products were monitored using a TOC analyzer, and styrene concentration was measured with a gas chromatograph.
A significant methodological contribution is the development of a BAS-BP neural network model. This model combines the Beetle Antennae Search (BAS) algorithm with a Backpropagation (BP) neural network. The BAS algorithm is used to optimize the weights and thresholds of the BP neural network, aiming to improve its goodness-of-fit and predictive capabilities. This approach is intended to provide a more robust and accurate prediction of styrene residuals under various operating conditions, even those difficult to replicate experimentally.
Datasets
The study utilized experimental data generated from their laboratory setup. This data served as the training set for the BAS-BP neural network model. They collected 17 sets of experimental data for training and performed four sets of calibration tests for evaluation. No pre-existing public datasets were explicitly mentioned for training or benchmarking the neural network.
Key Insights and Ideas
- Synergistic effect of UV and Fenton: The photo-Fenton process (UV + Fenton) demonstrated significantly higher styrene removal efficiency compared to individual components (UV, Fenton, H₂O₂, Fe²⁺ alone or in pairs). This highlights the synergistic benefits of combining UV irradiation with the Fenton reaction.
- pH is critical: The study found that an acidic pH of 3.0 is optimal for styrene degradation. Higher pH values lead to the precipitation of iron hydroxides, reducing the availability of Fe²⁺ for catalysis and thus decreasing •OH radical generation.
- Optimal reagent concentrations: The research identified specific optimal concentrations for H₂O₂ (20 mmol/L) and Fe²⁺ (0.3 mmol/L) that maximize styrene removal while considering economic factors and by-product formation.
- BAS-BP model's predictive power: The BAS-BP neural network model showed improved performance over a standard BP network, offering better accuracy in predicting styrene degradation. This suggests its potential for real-time monitoring and control in industrial applications.
- Nitrogen atmosphere is key: The entire process was conducted under a nitrogen atmosphere, crucial for safety in storage tank environments and to avoid potential side reactions with oxygen.
Key Results
- Optimal Conditions: The best styrene removal efficiency of 96.23% was achieved under the following conditions: H₂O₂ at 20 mmol/L, Fe²⁺ at 0.3 mmol/L, and pH 3.0.
- Degradation Time: This optimal condition achieved high removal efficiency within approximately 100 minutes.
- By-product Control: The final by-product (TOC) concentration under optimal conditions was 22.45 mg/L, indicating effective degradation.
- BAS-BP Model Performance: The BAS-BP model demonstrated a significant reduction in evaluation indicators (MSE, RMSE, MAPE) compared to the standard BP network, indicating improved prediction accuracy and convergence rate. For instance, MSE was reduced by 64.22%.
Limitations and Future Work
- Laboratory Scale: The study was conducted at a laboratory scale. Scaling up the process to industrial applications may present new challenges and require further optimization.
- Limited By-product Analysis: While TOC was measured, a detailed identification and quantification of all degradation by-products were not fully explored. Further analysis could provide a more comprehensive understanding of the degradation pathway.
- Model Generalizability: The BAS-BP model was trained on a specific set of experimental data. Its generalizability to a wider range of styrene concentrations or different VOC mixtures would need further validation.
- Long-term Stability: The long-term stability and efficiency of the photo-Fenton process in a continuous operation mode were not extensively investigated.
The authors suggest that the developed BAS-BP model can provide valuable real-time data and serve as an early warning system for instrument damage or process deviations, contributing to safer and more efficient operation of styrene storage systems. Future work could focus on pilot-scale studies and a more in-depth analysis of degradation by-products.
Summary updated