Biomass Characteristics and Biomethane Production Database
BCBPD is an integrated platform developed by Beijing University of Chemical Technology (BUCT) for searching, comparing, and predicting the physicochemical characteristics and biomethane production performance of lignocellulosic wastes (LW).
Platform Features
Data Query
Search and compare the physicochemical characteristics (TS, VS, C, O, H, N, cellulose, hemicellulose, lignin, organic loading, S/I ratio) and cumulative biomethane production curves of 178 untreated and 254 pretreated lignocellulosic waste samples collected from peer-reviewed literature.
Biomethane Prediction
Predict the cumulative biomethane production curve for uncollected or hypothetical lignocellulosic wastes using machine learning models (XGBoost) optimized via TPOT. Enter raw material properties (TS, VS, elemental composition, lignocellulosic fractions, etc.) to obtain a day-by-day biomethane yield prediction curve.
About BCBPD
The Biomass Characteristics and Biomethane Production Database (BCBPD) was developed to support researchers and engineers in the field of anaerobic digestion and biogas production. The platform provides two main functions:
- Data Query: A searchable database containing the physicochemical properties and cumulative biomethane production of 178 species of untreated lignocellulosic wastes and 254 samples of chemically pretreated materials. Users can filter by category, name, or specific properties and view detailed biomethane production curves over the full digestion period.
- Prediction: For lignocellulosic wastes not included in the database, users can input raw material characteristics to obtain a predicted biomethane production curve using an XGBoost model optimized by TPOT (Tree-based Pipeline Optimization Tool). The pretreated prediction model additionally accepts pretreatment agent type, concentration, and duration as inputs.
The machine learning models were trained on standardized datasets with rigorous cross-validation, achieving high prediction accuracy. Feature importance analysis via SHAP (SHapley Additive exPlanations) and partial dependence plots are available to help users understand the key factors influencing biomethane production.
Related Publications
- Chao Song, Chang Chen, et al. "Machine learning-assisted prediction and optimization of biomethane production from lignocellulosic wastes." Bioresource Technology, 2023, 390: 129953. https://doi.org/10.1016/j.biortech.2023.129953
- Chao Song, Chang Chen, et al. "Chemical pretreatment of lignocellulosic wastes for enhanced biomethane production: A data-driven approach." Waste Management, 2024, 186: 114–125. https://doi.org/10.1016/j.wasman.2024.07.004
If you use BCBPD in your research, please cite the above publications.