Plant-scale biogas production prediction based on multiple hybrid machine learning technique

Yi Zhang; Linhui Li; Zhonghao Ren; Yating Yu; Yeqing Li; Junting Pan; Yanjuan Lu; Lu Feng; Weijin Zhang; Yongming Han

doi:10.1016/j.biortech.2022.127899

Plant-scale biogas production prediction based on multiple hybrid machine learning technique

Bioresour Technol. 2022 Nov:363:127899. doi: 10.1016/j.biortech.2022.127899. Epub 2022 Sep 6.

Authors

Yi Zhang¹, Linhui Li², Zhonghao Ren¹, Yating Yu¹, Yeqing Li³, Junting Pan⁴, Yanjuan Lu⁵, Lu Feng⁶, Weijin Zhang⁷, Yongming Han⁸

Affiliations

¹ State Key Laboratory of Heavy Oil Processing, Beijing Key Laboratory of Biogas Upgrading Utilization, College of New Energy and Materials, China University of Petroleum Beijing (CUPB), Beijing 102249, PR China.
² College of Artificial Intelligence, China University of Petroleum Beijing (CUPB), Beijing 102249, PR China.
³ State Key Laboratory of Heavy Oil Processing, Beijing Key Laboratory of Biogas Upgrading Utilization, College of New Energy and Materials, China University of Petroleum Beijing (CUPB), Beijing 102249, PR China. Electronic address: liyeqingcup@126.com.
⁴ Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, PR China.
⁵ Beijing Fairyland Environmental Technology Co., Ltd, Beijing 100094, PR China.
⁶ NIBIO, Norwegian Institute of Bioeconomy Research, P.O. Box 115, N-1431 Ås, Norway.
⁷ School of Energy Science and Engineering, Central South University, Changsha 410083, PR China.
⁸ College of Information Science & Technology, Beijing University of Chemical Technology, Beijing 100029, PR China.

PMID: 36075348
DOI: 10.1016/j.biortech.2022.127899

Abstract

The parameters from full-scale biogas plants are highly nonlinear and imbalanced, resulting in low prediction accuracy when using traditional machine learning algorithms. In this study, a hybrid extreme learning machine (ELM) model was proposed to improve prediction accuracy by solving imbalanced data. The results showed that the best ELM model had a good prediction for validation data (R² = 0.972), and the model was developed into the software (prediction error of 2.15 %). Furthermore, two parameters within a certain range (feed volume (FV) = 23-45 m³ and total volatile fatty acids of anaerobic digestion (TVFA_AD) = 1750-3000 mg/L) were identified as the most important characteristics that positively affected biogas production. This study combines machine learning with data-balancing techniques and optimization algorithms to achieve accurate predictions of plant biogas production at various loads.

Keywords: Anaerobic digestion; Extreme learning machine; Genetic algorithm; Graphical User Interface Software; Synthetic Minority Over-sampling Technique for Regression.

MeSH terms

Algorithms
Biofuels*
Machine Learning*
Software

Substances

Biofuels