Understanding the macro-kinetics of coal–oxygen reactions is the theoretical foundation for combating coal spontaneous combustion, with focus on obtaining kinetic parameters. There are still open questions, including which thermal analysis kinetics processing methods (TAKPMs) output more confident kinetic parameters, and how these TAKPMs respond in terms of computational accuracy when experimental conditions change. This study evaluated these questions quantitatively, and yielded the following noteworthy findings. (1) Integral processing-based multi-scan methods were more data-friendly than differentials, which produced over 40% error due to thermal noise. (2) Combined 29 widely used most probable reaction mechanism functions (MPRMFs), like Jander, Ginstling–Brounshtein, and Zhuralev–Lesokin–Tempelman, with five popular single-scan methods, and ran 870+ calculations under six experimental conditions. Using traditional model fitting to identify MPRMFs can introduce errors of 4.08–275.49% (median 121.99%). (3) Applied spectroscopy–Málek–Popescu coupling for high-confidence MPRMFs. Its integration into single-scan methods improved computational accuracy, reducing error to the range of 1.84–76.20%. However, a median of 37.76% error indicates that some single-scan methods still failed to yield satisfactory results. (4) Notably, the accuracy of single-scan methods correlated significantly with heating rate, with various methods showing positive or negative correlations. From 5 to 15 K/min, the increase/decrease in accuracy ranged 10.63–244.25% (with median of 51.40%). These findings underline the need for a case-by-case selection of TAKPMs based on experimental conditions, ensuring more confident kinetic parameters. This study introduced a workflow, exemplified with bituminous coal, to assist researchers in selecting the optimal TAKPMs for diverse experimental scenarios.

理解煤-氧反应的宏观动力学是防治煤自燃的理论基础，重点是获取动力学参数。仍然存在一些悬而未决的问题，包括哪些热分析动力学处理方法 (TAKPM) 输出更可靠的动力学参数，以及当实验条件发生变化时，这些 TAKPM 在计算精度方面如何响应。这项研究定量评估了这些问题，并得出了以下值得注意的发现。 (1) 基于积分处理的多重扫描方法比差分方法更利于数据处理，差分方法由于热噪声而产生超过 40% 的误差。 (2) 将 Jander、Ginsling-Brounshtein 和 Zhuralev-Lesokin-Tempelman 等 29 个广泛使用的最可能反应机理函数 (MPRMF) 与 5 种流行的单扫描方法相结合，并在 6 个实验条件下进行了 870 多次计算。使用传统模型拟合来识别 MPRMF 可能会产生 4.08–275.49% 的误差（中位数 121.99%）。 (3) 应用光谱-Málek-Popescu 耦合实现高置信度 MPRMF。其与单扫描方法的集成提高了计算精度，将误差降低到 1.84-76.20% 的范围。然而，37.76%的中位误差表明一些单次扫描方法仍然未能产生令人满意的结果。 (4) 值得注意的是，单次扫描方法的准确度与升温速率显着相关，各种方法均呈现正相关或负相关。从 5 到 15 K/min，准确度的增加/减少范围为 10.63-244.25%（中位数为 51.40%）。这些发现强调需要根据实验条件逐案选择 TAKPM，以确保更可靠的动力学参数。本研究引入了一个以烟煤为例的工作流程，以帮助研究人员为不同的实验场景选择最佳的 TAKPM。