Machine learning prediction of nitrogen heterocycles in bio-oil produced from hydrothermal liquefaction of biomass

Enhancing biomass conversion to bioenergy with machine learning: Gains and problems

The growing concerns about environmental sustainability and energy security, such as exhaustion of traditional fossil fuels and global carbon footprint growth have led to an increasing interest in alternative energy sources, especially bioenergy. Recently, numerous scenarios have been proposed regarding the use of bioenergy from different sources in the future energy systems. In this regard, one of the biggest challenges for scientists is managing, modeling, decision-making, and future forecasting of bioenergy systems. The development of machine learning (ML) techniques can provide new opportunities for modeling, optimizing and managing the production, consumption and environmental effects of bioenergy. However, researchers in bioenergy fields have not widely utilized the ML concepts and practices. Therefore, a comparative review of the current ML techniques used for bioenergy productions is presented in this paper. This review summarizes the common issues and difficulties existing in integrating ML with bioenergy studies, and discusses and proposes the possible solutions. Additionally, a detailed discussion of the appropriate ML application scenarios is also conducted in every sector of the entire bioenergy chain. This indicates the modernized conversion processes supported by ML techniques are imperative to accurately capture process-level subtleties, and thus improving techno-economic resilience and socio-ecological integrity of bioenergy production. All the efforts are believed to help in sustainable bioenergy production with ML technologies for the future.

Machine learning application for predicting key properties of activated carbon produced from lignocellulosic biomass waste with chemical activation

The successful application of gradient boosting regression (GBR) in machine learning to forecast surface area, pore volume, and yield in biomass-derived activated carbon (AC) production underscores its potential for enhancing manufacturing processes. The GBR model, collecting 17 independent variables for two-step activation (2-SA) and 14 for one-step activation (1-SA), demonstrates effectiveness across three datasets-1-SA, 2-SA, and a combined dataset. Notably, in 1-SA, the GBR model yields R2 values of 0.76, 0.90, and 0.83 for TPV, yield, and SSA respectively, and records R2 of 0.90 and 0.91 for yield in 2-SA and combined datasets. The model highlights the significance of the soaking procedure alongside activation temperature in shaping AC properties with 1-SA or 2-SA, illustrating machine learning's potential in optimizing AC production processes.

Combination of ensemble machine learning models in photocatalytic studies using nano TiO2 - Lignin based biochar

Synergizing photocatalytic reactions with machine learning methods can effectively optimize and automate the remediation of pollutants. In this work, commercial Degussa TiO2 nanoparticles and lignin based biochar (LB) where used to prepare TiO2: lignin based biochar (TLB) composites using ultrasound-assisted co-precipitation method. The photocatalytic property of the TLB composites where studied by conducting the photocatalytic degradation of a Basic blue 41 (BB41) dye. The influence of calcination temperature, T:LB compositions, catalyst dosage, initial dye pH, initial dye concentration, and illumination time on photocatalytic dye degradation were experimentally studied. The degradation efficiency of 96.72 % was obtained under optimized conditions for the photocatalyst calcined at 500 °C containing a 1:1 wt percentage of TiO2 and LB. The experimental data was further used to predict the photocatalytic degradation efficiency using Gradient Tree Boosting (GTB) and Extra Trees (ET) models. The GTB model gave the highest prediction accuracy of 94 %. The permutation variable importance revealed catalyst dosage and dye concentration as the most influential parameters in the prediction of the photocatalytic dye degradation efficiency.

Machine learning applications for biochar studies: A mini-review

Biochar is a promising carbon sink whose application can assist in reducing carbon emissions. Development of this technology currently relies on experimental trials, which are time-consuming and labor-intensive. Machine learning (ML) technology presents a potential solution for streamlining this process. This review summarizes the current research on ML's applications in biochar production, characterization, and applications. It briefly explains commonly used machine learning algorithms and discusses prospects and challenges. A hybrid model that combines ML with mechanism-based analysis could be a future trend, addressing the ML's black-box nature. While biochar studies have adopted ML technology, current works mostly use lab-scale data for model training. Further work is needed to develop ML models based on pilot or industrial-scale data to realize the use of ML techniques for the field application of biochar.

Machine Learning Model Insights into Base-Catalyzed Hydrothermal Lignin Depolymerization

Lignin, an abundant component of plant matter, can be depolymerized into renewable aromatic chemicals and biofuels but remains underutilized. Homogeneously catalyzed depolymerization in water has gained attention due to its economic feasibility and performance but suffers from inconsistently reported yields of bio-oil and solid residues. In this study, machine learning methods were used to develop predictive models for bio-oil and solid residue yields by using a few reaction variables and were subsequently validated by doing experimental work and comparing the predictions to the results. The models achieved a coefficient of determination (R2) score of 0.83 and 0.76, respectively, for bio-oil yield and solid residue. Variable importance for each model was calculated by two different methodologies and was tied to existing studies to explain the model predictive behavior. Based on the outcome of the study, the creation of concrete guidelines for reporting in lignin depolymerization studies was recommended. Shapley additive explanation value analysis reveals that temperature and reaction time are generally the strongest predictors of experimental outcomes compared to the rest.

Machine learning assisted predicting and engineering specific surface area and total pore volume of biochar

Biochar produced from pyrolysis of biomass is a platform porous carbon material that have been widely used in many areas. Specific surface area (SSA) and total pore volume (TPV) are decisive to biochar application in hydrogen uptake, CO₂ adsorption, and organic pollutant removal, etc. Engineering biochar by traditional experimental methods is time-consuming and laborious. Machine learning (ML) was used to effectively aid the prediction and engineering of biochar properties. The prediction of biochar yield, SSA, and TPV was achieved via random forest (RF) and gradient boosting regression (GBR) with test R² of 0.89-0.94. ML model interpretation indicates pyrolysis temperature, biomass ash, and volatile matter were the most important features to the three targets. Pyrolysis parameters and biomass mixing ratios for biochar production were optimized via three-target GBR model, and the optimum schemes to obtain high SSA and TPV were experimentally verified, indicating the great potential of ML for biochar engineering.

Physics-informed machine learning methods for biomass gasification modeling by considering monotonic relationships

Machine learning methods have recently shown a broad application prospect in biomass gasification modeling. However, a significant drawback of the machine learning approaches is their poor physical interpretability when relying on limited experimental data. In the present work, a physics-informed neural network method (PINN) is developed to predict biomass gasification products (N₂, H₂, CO, CO₂, and CH₄). PINN simultaneously considers regression, structure, and physical monotonicity constraints in the loss function, providing physically feasible predictions. Specifically, the PINN models have outperformed prediction capability (average test R² 0.91-0.97) compared to five other machine learning methods through 50 times random sample classifications. Furthermore, it is demonstrated that the developed models can maintain correct monotonicity even if the feedstock characteristics or gasification conditions are outside the training data. By using a reliable physical mechanism to guide machine learning, the model can ensure better generalizability and scientific interpretability.

Advances in machine learning for high value-added applications of lignocellulosic biomass

Lignocellulose can be converted into biofuel or functional materials to achieve high value-added utilization. Biomass utilization process is complex and multi-dimensional. This paper focuses on the biomass conversion reaction conditions, the preparation of biomass-based functional materials, the combination of biomass conversion and traditional wet chemistry, molecular simulation and process simulation. This paper analyzes the mechanism, advantages and disadvantages of important machine learning (ML) methods. The application examples of ML in different aspects of high value utilization of lignocellulose are summarized in detail. The challenges and future prospects of ML in this field are analyzed.

Machine learning for hydrothermal treatment of biomass: A review

Hydrothermal treatment (HTT) (i.e., hydrothermal carbonization, liquefaction, and gasification) is a promising technology for biomass valorization. However, diverse variables, including biomass compositions and hydrothermal processes parameters, have impeded in-depth mechanistic understanding on the reaction and engineering in HTT. Recently, machine learning (ML) has been widely employed to predict and optimize the production of biofuels, chemicals, and materials from HTT by feeding experimental data. This review comprehensively analyzed the application of ML for HTT of biomass and systematically illustrated basic ML procedure and descriptors for inputs and outputs of ML models (e.g., biomass compositions, operation conditions, yield and physicochemical properties of derived products) that could be applied in HTT. Moreover, this review summarized ML-aided HTT prediction of yield, compositions, and physicochemical properties of HTT hydrochar or biochar, bio-oil, syngas, and aqueous phase. Ultimately, future prospects were proposed to enhance predictive performance, mechanistic interpretation, process optimization, data sharing, and model application during ML-aided HTT.

Hydrothermal Liquefaction of Lignocellulosic and Protein-Containing Biomass: A Comprehensive Review

Hydrothermal liquefaction (HTL) is a thermochemical depolymerization technology, also known as hydrous pyrolysis, that transforms wet biomass into biocrude and valuable chemicals at a moderate temperature (usually 200–400 °C) and high pressure (typically 10–25 MPa). In HTL, water acts as a key reactant in HTL activities. Several properties of water are substantially altered as the reaction state gets closer to the critical point of water, which can result in quick, uniform, and effective reactions. The current review covers the HTL of various feedstocks, especially lignocellulosic and high protein-containing feeds with their in-depth information of the chemical reaction mechanisms involved in the HTL. Further, this review gives insight and knowledge about the influencing factors such as biomass pretreatment, process mode, process conditions, etc., which could affect the efficiency of the hydrothermal process and biocrude productivity. In addition, the latest trends, and emerging challenges to HTL are discussed with suitable recommendations.