报告题目：Data-driven analysis of nutritional content in food waste compost based on process parameters
报告人：LEE Chew Tin
邀请人：潘雄锋教授
报告时间及地点：2025年8月8日 8：30-11：30 学院B309
报告人简况：
Dr Chew Tin Lee received her PhD from the University of Cambridge, U.K. She currently serves as the Professor in Universiti Teknologi Malaysia (UTM), Malaysia. She is passionate on promoting sustainable organic waste management through 3R campaign. Lee is the Associate Editor for Energy(Elsevier). She is now the key driver and Co-chair for the International Conference of Low Carbon in Asia since 2015. She actively contributes to low carbon policy in Malaysia (i.e. Iskandar Malaysia region, and Kuala Lumpur City Council). She was the Erasmus Scholar in University Politechnica Madridand the Institution of Engineers Malaysia.
报告内容摘要：
Compost quality assessment plays a vital role in ensuring the agronomic and environmental safety of organic waste recycling processes. Among the various indicators, total nitrogen (TN) is a key determinant of compost nutrient value; however, its accurate prediction remains challenging due to variations in feedstock composition and process conditions. This study aimed to develop and evaluate multiple linear regression models to predict TN content in food waste compost using datasets compiled from published studies across diverse composting systems. Independent variables included both operational parameters and nutrient-based indicators, ensuring that the models were grounded in biologically relevant and widely available data. The final models, TN Model achieve a regression R² of 0.64 and adjusted R² of 0.59. Despite the prediction R² of 0.27, it produced a lower RMSE (0.58) and a reasonable MAPE (26.08 %), reflecting a strong balance between interpretability and predictive accuracy. The findings reinforce the benefits of integrating biologically meaningful and universally measurable parameters, such as aeration rate, TOC, pH, and compost volume. These parameters not only reflect core microbial and physicochemical processes that govern nitrogen transformation, but also routinely monitored in most composting operations, making the resulting models both scientifically robust and practically applicable. This study offers a scalable approach to compost quality monitoring and highlights the potential of data-driven modelling using harmonized literature datasets. The resulting models are well-suited for practical implementation in research, agricultural, and waste management settings.
【关闭】
