Department of Marine Environmental Engineering, National Kaohsiung University of Science and Technology, Kaohsiung City, Taiwan; Department of Seafood Science, National Kaohsiung University of Science and Technology, Kaohsiung City, Taiwan.
Department of Seafood Science, National Kaohsiung University of Science and Technology, Kaohsiung City, Taiwan.
Sci Total Environ. 2023 Aug 15;886:163972. doi: 10.1016/j.scitotenv.2023.163972. Epub 2023 May 8.
In view of the global climate change concerns, the society is approaching towards the development of 'green' and renewable energies for sustainable future. The non-renewable fossil fuels may be largely replaced by renewable energy sources, which could facilitate sustainable growth, energy development and lessen the reliance on conventional energy sources. The traditional methods employed in biorefineries to estimate the data values for the biofuel production systems are often complicated, time-consuming and labour-intensive. Modern machine learning (ML) technologies hold enormous potential in managing high-dimensional complex scientific tasks and improving decision-making in energy distribution networks and systems. The data-driven probabilistic ML algorithms could be applied to smart biofuel systems and networks that may reduce the cost of experimental research while providing accurate estimates of product yields. The current review demonstrates a thorough understanding of the application of different ML models to regulate and monitor the production of biofuels from waste biomass through prediction, optimization and real-time monitoring. The in-depth analysis of the most recent advancements in ML-assisted biofuel production methods, including thermochemical and biochemical processes is provided. Moreover, the ML models in addressing the issues of biofuel supply chains, case studies, scientific challenges and future direction in ML applications are also summarized.
鉴于全球气候变化的担忧，社会正在朝着开发“绿色”和可再生能源以实现可持续未来的方向发展。不可再生的化石燃料可能会被可再生能源大量取代，这将有助于可持续增长、能源开发，并减少对传统能源的依赖。生物炼制厂中用于估计生物燃料生产系统数据值的传统方法通常很复杂、耗时且劳动密集型。现代机器学习 (ML) 技术在管理高维复杂科学任务和改进能源分配网络和系统的决策方面具有巨大潜力。数据驱动的概率 ML 算法可应用于智能生物燃料系统和网络，这可能会降低实验研究的成本，同时提供产品产量的准确估计。目前的综述展示了对不同 ML 模型在通过预测、优化和实时监测来调节和监控从废生物质生产生物燃料方面的应用的透彻理解。提供了对 ML 辅助生物燃料生产方法的最新进展的深入分析，包括热化学和生物化学过程。此外，还总结了 ML 模型在解决生物燃料供应链问题、案例研究、科学挑战以及 ML 应用的未来方向方面的应用。
