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

Hydrothermal liquefaction (HTL) of high-moisture biomass or biowaste to produce bio-oil is a promising technology. However, nitrogen-heterocycles (N-H) presence in bio-oil is a bottleneck to the upgrading and utilization of bio-oil. The present study applied the machine learning (ML) method (random forest) to predict and help control the bio-oil N-H, bio-oil yield, and N content of bio-oil (N_oil). The results indicated that the predictive performance of the yield and N_oil were better than previous studies, achieving test R-2 of 0.92 and 0.95, respectively. Acceptable predictive performance (test R-2 of 0.82 and RMSE of 7.60) for the prediction of N-H was also achieved. The feature importance analysis, partial dependence, and Shapely value were used to interpret the prediction models and study the N-H formation mechanisms and behavior. Then, forward optimization of N-H was implemented based on optimal predictive models, indicating the high potential of ML-aided bio-oil production and engineering.

Automated summary

In this study, machine learning method was successfully utilized to predict and control nitrogen-heterocycles, bio-oil yield, and nitrogen content in bio-oil, achieving good predictive performance. The interpretation and study of the prediction models provided insights into the formation mechanisms and behavior of nitrogen-heterocycles.