Pyrolysis of biomass is a crucial process for the production of renewable energy and a sustainable alternative to fossil fuels. The present study analyzed the pyrolysis process and the composition of the products of six representative lignocellulosic biomasses using simultaneous thermal analyzer, pyrolysis - gas chromatography/mass spectrometry (Py-GC/MS), liquid-phase nuclear magnetic resonance (NMR) spectroscopy and X-ray photoelectron spectrometer (XPS). Furthermore, based on the TG results, the kinetics of the pyrolysis of three components of the six biomasses were explored using kinetic modelling, triple-parallel reaction model and Asym2sig deconvolution function. Subsequently, kinetic models of the six biomasses were developed using a combinatorial kinetic method. Finally, an artificial neural network (ANN) model was developed to predict the pyrolysis behaviors of the studied biomasses. For example, corn straw (CS) revealed three primary pyrolysis stages: below 400 K (volatilization of small molecules), 400–670 K (decomposition of major components), and above 670 K (charring of the residual components and secondary decomposition of intermediates). The optimum kinetics models for CS components were: f₁(α₁) = α₁^−0.80697 (1 − α₁)^1.99689 [−ln(1 − α₁)]^1.15425, f₂(α₂) = α₂^0.43522 (1 − α₂)^1.29066 [−ln(1 − α₂)]^0.32644, and f₃(α₃) = α₃^−2.82644 (1 − α₃)³ [–ln(1 − α₃)]^−1.87480. Moreover, ANN23 showed the highest R² value (0.99908). Therefore, ANN23 is the most suitable model for predicting the pyrolysis of CS. The present research provides valuable references for the pyrolysis of biomass.

中文翻译：

---

基于三重平行反应模型、组合动力学、Py-GC/MS 和人工神经网络的木质纤维素生物质热解动力学、反应机理和产物分布


生物质热解是生产可再生能源的关键过程，也是化石燃料的可持续替代品。本研究使用同步热分析仪、热解-气相色谱/质谱 （Py-GC/MS）、液相核磁共振 （NMR） 光谱和 X 射线光电子能谱仪 （XPS） 分析了 6 种代表性木质纤维素生物质的热解过程和产物组成。此外，基于 TG 结果，使用动力学建模、三重平行反应模型和 Asym2sig 反卷积函数探索了 6 个生物质中 3 个组分的热解动力学。随后，使用组合动力学方法开发了六种生物量的动力学模型。最后，开发了人工神经网络 （ANN） 模型来预测所研究生物质的热解行为。例如，玉米秸秆 （CS） 揭示了三个初级热解阶段：低于 400 K（小分子挥发）、400-670 K（主要成分分解）和高于 670 K（残留成分炭化和中间体二次分解）。CS 组分的最佳动力学模型为：f₁（α₁） = α₁^−0.80697（1 − α₁）^1.99689[−ln（1 − α₁）]^1.15425，f ₂（α₂） = α₂^0.43522（1 − α₂）^1.29066[−ln（1 − α₂）]^0.32644 和 f₃（α₃） = α₃^−2.82644（1 − α₃）³[–ln（1 − α₃）]^−1.87480。此外，ANN23 显示出最高的 R² 值 （0.99908）。因此，ANN23 是预测 CS 热解的最合适模型。本研究为生物质的热解提供了有价值的参考。