原位转化技术是松辽盆地嫩江组中低熟页岩油藏最具前景的开采技术，转化过程涉及油页岩中有机质多步热解反应的耦合，因此建立可靠的有机质热解动力学模型需对该过程进行解耦。该文对0.5、1、2、4 ℃/min升温速率下嫩江组油页岩热解的热重曲线开展了反卷积分峰，基于分峰精度以及活化能均一性判据，将有机质热解过程解耦为3步反应并获得了各步反应的热解特性。随后对3步反应分别开展了动力学分析，获得的动力学方程对3步反应过程的预测精度均高于0.96。结合各步反应对应的产物分布及在有机质热解过程中的权重，建立了嫩江组油页岩多反应耦合有机质热解动力学模型。该模型对于0.2~4 ℃/min升温速率范围内有机质热解过程的预测精度高于0.994，远超有机质热解总包反应动力学模型(预测精度0.932)。采用该模型可较好地预测不同升温历程下油页岩有机质热解过程及产物组成，对于该油藏原位转化技术的开发具有指导意义。

Objective: The medium-low maturity shale oil reservoirs of the Nenjiang Formation in the Songliao Basin, China, exhibit considerable development potential. In-situ conversion technology is considered the most promising method for the extraction of oil from such reservoirs. This conversion process involves multiple coupled pyrolysis reactions of organic matter in oil shale, necessitating a valid kinetic model capable of accurately predicting the multiple reactions in the pyrolysis process and the resulting product distribution. The establishment of such a model is critical for optimizing in-situ conversion efficiency and guiding the technological development of this extraction method. Methods: In this study, pyrolysis experiments were conducted at heating rates of 0.5, 1, 2, and 4 ℃/min to analyze the pyrolysis characteristics of oil shale from the Nenjiang Formation in the Songliao Basin. The derivative thermogravimetric (DTG) curves were deconvoluted to decouple the pyrolysis process into three distinct reactions: kerogen conversion, primary pyrolysis of bitumen, and secondary pyrolysis of bitumen. The weight loss characteristics of these reactions were obtained at various heating rates. The results showed that with the increase in the heating rate, the weight loss process of each reaction shifted toward higher temperatures. However, the relative shape and corresponding weight loss rate of each peak remained largely unchanged. To meet the in-situ conversion requirements, we classified the pyrolysis products into six categories and determined their distributions for each reaction, leading to the establishment of equations for product distribution. Kinetic analyses were subsequently performed based on the Starink method, Coats-Redfern method, and the kinetic compensation effect to determine the initial values and distribution ranges of kinetic parameters for each reaction. The Bayesian optimization method was then performed to iteratively refine these parameters, which minimized prediction errors and yielded final kinetic parameters. Results: The resulting kinetic equations exhibited a high predictive accuracy, with R² values exceeding 0.96 for each reaction. The calculated relative contributions of the three reactions to the overall pyrolysis process of organic matter reached 0.229, 0.509, and 0.262. Through the application of the corresponding weights to the kinetic equations and incorporation of the product distribution equations, a kinetic model was established for the pyrolysis of organic matter in oil shale from the Nenjiang Formation. The model demonstrated a strong predictive accuracy at heating rates ranging from 0.5 to 4 ℃/min, achieving an overall correlation coefficient of R²>0.994 8. To further evaluate the model's applicability under slow heating conditions relevant to in-situ conversion, we conducted temperature-programmed pyrolysis experiments on the oil shale from the Nenjiang Formation at 0.2 ℃/min through thermogravimetric analysis. This model demonstrated a high prediction accuracy (R²>0.994) for the pyrolysis process of organic matter at a heating rate of 0.2 ℃/min, which substantially outperformed the global reaction kinetic model (prediction accuracy of 0.932). Model predictions for product distribution at this heating rate indicated that the product yields increased with temperature, with a gradual rise observed between 260-330 ℃, followed by a sharp increase between 350-430 ℃, which peaked at approximately 450 ℃. However, beyond 430 ℃, the increased presence of heteroatomic compounds and CO₂ suggested a potential decline in the economic efficiency of in-situ conversion. Conclusions: The kinetic modeling method proposed in this work displays a higher prediction accuracy compared with the traditional method, and the developed kinetic model effectively predicts the pyrolysis process and product distribution of Nenjiang oil shale under varying heating conditions. This model offers crucial theoretical insights into the optimization and implementation of in-situ conversion technology, which supports the efficient exploitation of shale oil reservoirs in the Nenjiang Formation.