Aniza Ria, Chen Wei-Hsin, Pétrissans Anélie, Hoang Anh Tuan, Ashokkumar Veeramuthu, Pétrissans Mathieu
Department of Aeronautics and Astronautics, National Cheng Kung University, Tainan, 701, Taiwan; International Doctoral Degree Program on Energy Engineering, National Cheng Kung University, Tainan, 701, Taiwan.
Department of Aeronautics and Astronautics, National Cheng Kung University, Tainan, 701, Taiwan; Research Center for Smart Sustainable Circular Economy, Tunghai University, Taichung, 407, Taiwan; Department of Mechanical Engineering, National Chin-Yi University of Technology, Taichung, 411, Taiwan.
Environ Pollut. 2023 May 1;324:121363. doi: 10.1016/j.envpol.2023.121363. Epub 2023 Feb 28.
Biowaste remediation and valorization for environmental sustainability focuses on prevention rather than cleanup of waste generation by applying the fundamental recovery concept through biowaste-to-bioenergy conversion systems - an appropriate approach in a circular bioeconomy. Biomass waste (biowaste) is discarded organic materials made of biomass (e.g., agriculture waste and algal residue). Biowaste is widely studied as one of the potential feedstocks in the biowaste valorization process due to its being abundantly available. In terms of practical implementations, feedstock variability from biowaste, conversion costs and supply chain stability prevent the widespread usage of bioenergy products. Biowaste remediation and valorization have used artificial intelligence (AI), a newly developed idea, to overcome these difficulties. This report analyzed 118 works that applied various AI algorithms to biowaste remediation and valorization-related research published between 2007 and 2022. Four common AI types are utilized in biowaste remediation and valorization: neural networks, Bayesian networks, decision tree, and multivariate regression. The neural network is the most frequent AI for prediction models, the Bayesian network is utilized for probabilistic graphical models, and the decision tree is trusted for providing tools to assist decision-making. Meanwhile, multivariate regression is employed to identify the relationship between experimental variables. AI is a remarkably effective tool in predicting data, which is reportedly better than the conventional approach owing to its characteristics of time-saving and high accuracy. The challenge and future work in biowaste remediation and valorization are briefly discussed to maximize the model's performance.
生物废弃物修复与增值以实现环境可持续性为目标，重点在于预防而非清理废弃物产生，通过生物废弃物到生物能源的转化系统应用基本的回收概念，这是循环生物经济中的一种合适方法。生物质废弃物（生物废弃物）是由生物质制成的废弃有机材料（如农业废弃物和藻类残渣）。由于生物废弃物来源丰富，它被广泛研究作为生物废弃物增值过程中的潜在原料之一。在实际应用方面，生物废弃物原料的多变性、转化成本和供应链稳定性阻碍了生物能源产品的广泛使用。生物废弃物修复与增值利用了人工智能（AI）这一新兴理念来克服这些困难。本报告分析了118篇在2007年至2022年间发表的将各种AI算法应用于生物废弃物修复与增值相关研究的作品。在生物废弃物修复与增值中使用了四种常见的AI类型：神经网络、贝叶斯网络、决策树和多元回归。神经网络是预测模型中最常用的AI，贝叶斯网络用于概率图形模型，决策树则被信赖用于提供辅助决策的工具。同时，多元回归用于识别实验变量之间的关系。AI是预测数据的一种非常有效的工具，据报道，由于其省时和高精度的特点，它比传统方法更好。本文简要讨论了生物废弃物修复与增值中的挑战和未来工作，以最大限度地提高模型性能。
