采空区煤自燃是影响矿井安全生产的主要灾害之一，不仅产生毒害气体，还会诱发瓦斯爆炸、煤尘爆炸等一系列次生灾害。随着我国煤炭开采深度和强度增加，自燃灾害防治日趋复杂，尤其在综放采空区，受遗煤量和漏风影响，自然发火危险性高。因此，掌握采空区遗煤氧化升温规律，构建高效的预警方法对于煤自燃灾害防治具有重要的意义。采空区煤自燃是遗煤与氧气发生复合反应的最终结果，而煤氧复合反应速率与温度及氧浓度均存在非线性关系。本文采用理论分析、实验测试、数值模拟和现场应用等方法，针对煤氧复合反应速率与氧浓度的非线性关系、煤阶段性降氧氧化特性、推采条件下综放采空区遗煤自燃演化规律以及煤自燃动态预警方法等方面开展了研究。成果如下：

（1）结合理论分析和程序升温实验测试，分析了煤低温氧化反应级数的变化规律，得到实验煤样反应级数随煤温具有先减小后增加的阶段性特征，随着温度升高，反应级数逐渐趋近于1；通过分段拟合得到了反应级数与煤温的函数关系，由此明确了煤氧化反应速率与氧浓度的非线性关系；推导了耗氧速率及气体产物生成速率的非线性数学模型，对比了非线性与线性模型得到的煤氧化特性参数随温度变化规律，结果表明在高温阶段非线性模型计算结果要低于线性模型计算结果，其差异性主要是由于非线性模型考虑了煤样罐中氧浓度在风流方向上实际衰减特征所导致的。

（2）通过标志性气体的非线性模型计算结果，结合煤氧复合反应级数和气体产率变化率分析了煤自燃进程的阶段性特征，将煤自燃过程划分为自热、临界、热解、裂解、裂变五个阶段；基于特征温度及工作面推采过程中采空区遗煤所处环境氧浓度的衰减特征，开展了阶段性降氧程序升温实验，由非线性氧化反应速率模型分析了不同降氧环境下的煤氧反应速率及动力学参数，得到在高氧浓度条件下煤氧化至一定温度，降氧并持续升温过程中，煤氧化反应速率相对同低氧水平测试结果有一定的提升，且提升程度与降氧温度成正相关；分析煤样耗氧速率及气体产率与氧浓度、温度的关系，得到了对煤氧化速率有明显影响的三个氧浓度值，分别为5%、9%和17%，而氧浓度对表观活化能及煤自燃进程的突变特征影响较小。

（3）基于煤氧复合反应速率与氧浓度的非线性关系，考虑综放工作面推采速度、采空区遗煤及孔隙空间分布特征、粒度影响函数等因素，构建了采空区遗煤四维非线性氧化数值模拟模型，结合实验测试确定的关键参数进行了模拟计算，并借助自然发火实验确定的自然发火期及现场气体监测结果验证了模型的准确性；模拟结果表明采空区遗煤升温具有明显的“分区”特征，即“两道”位置遗煤温度高，采空区中部温度低，采空区最高温度与推进速度之间存在非线性关系；氧浓度是影响采空区遗煤升温的关键因素之一，当氧浓度下降至小于5%后，遗煤升温速率及CO产率显著下降。

（4）基于不同氧环境下遗煤氧化阶段性特征和气态产物变化规律，确立了三个单气体指标和五个常用的煤自燃预警复合气体指标，优选并构建了以R₂、C₂H₄、C₂H₆为主指标，CO、R₁、R₃、R₄、R₆为辅助指标的多指标分阶段预警体系，确定了不同阶段多指标与煤温的量化关系；考虑采空区遗煤预测温度和氧浓度的变化，结合煤氧复合反应的非线性特征，构建了采空区遗煤自然发火期预测模型，可预测在不同温度和氧环境下，煤温达到不同预警级别所需的时间。

（5）以采空区煤自燃分级预警方法、遗煤温度及自然发火期预测模型为基础，综合考虑矿井自燃灾害监测与预警需求，研发了煤自燃动态预警系统，并在陕西某矿综放工作面进行了现场工业试验；该方法及系统能够初步实现采空区煤自燃风险的动态感知和预警，提高煤自燃信息的处理效率和预警结果稳定性，避免单一指标造成的结果失真，现场应用效果良好。研究成果对加强采空区自燃灾害的防控水平具有一定的实际意义。

The coal spontaneous combustion is one of the major disasters affecting the mine safety production in the goaf, which not only produces toxic gases, but also induces a series of secondary disasters such as gas or coal dust explosion. With the increase in the depth and intensity of coal mining in China, the prevention and control of spontaneous combustion disasters are becoming more and more complex. Especially in the fully mechanized caving goaf, affected by the amount of residual coal and air leakage, the risk of spontaneous combustion is strong. Therefore, it is of great significance to master the law of oxidation temperature rise of residual coal in goaf and construct an efficient early warning method for coal spontaneous combustion disaster prevention. Coal spontaneous combustion in goaf is the final result of the composite reaction of residual coal and oxygen, and the composite reaction rate of coal and oxygen has a nonlinear relationship with temperature and oxygen concentration. In this paper, theoretical analysis, experimental test, numerical simulation and field application are used to study the nonlinear relationship between coal-oxygen composite reaction rate and oxygen concentration, the characteristics of staged oxygen reduction and oxidation of coal, the evolution law of residual coal spontaneous combustion in fully mechanized caving goaf under push mining conditions and the dynamic early warning method of coal spontaneous combustion. The main results are as follows:

（1）Combined with theoretical analysis and temperature programmed test, the change rule of coal oxidation reaction order at low temperature was analyzed. It was found that the reaction order of experimental coal samples decreased first and then increased with coal temperature. With the increase of temperature, the reaction series gradually approached 1. The functional relationship between the reaction order and the coal temperature was obtained by piecewise fitting, thus clarifying the nonlinear relationship between the coal oxidation reaction rate and the oxygen concentration. The nonlinear mathematical models of oxygen consumption rate and gas product formation rate are derived, and the variation of coal oxidation characteristic parameters obtained by nonlinear and linear models with temperature is compared. The results show that the calculation results of nonlinear model are lower than those of linear model in high temperature stage, and the difference is mainly due to the nonlinear model considering the actual attenuation characteristics of oxygen concentration in coal sample tank in the direction of air flow.

（2）Through the calculation results of the nonlinear model in the iconic gas, combined with the coal-oxygen composite reaction order and the gas yield change rate, the stage characteristics of the coal spontaneous combustion process were analyzed. The coal spontaneous combustion process was divided into five stages: autothermal, critical, pyrolysis, cracking and fission. Taking into account the characteristic temperature and the attenuation characteristics of the environmental oxygen concentration of the residual coal in the goaf during the mining process of the working face, the staged oxygen reduction temperature programmed experiment was carried out, and the coal oxygen reaction rate and kinetic parameters under different oxygen reduction environments were analyzed by the nonlinear oxidation reaction rate model. It was obtained that when the coal was oxidized to a certain temperature under the condition of high oxygen concentration, the coal oxidation reaction rate had a certain improvement compared with the test results of low oxygen level during the oxygen reduction and continuous heating process, and the improvement degree was positively correlated with the oxygen reduction temperature. The relationship of oxygen consumption rate and gas production rate with oxygen concentration and temperature was analyzed, and three oxygen concentrations which had significant influence on coal oxidation rate were obtained, which were 5%, 9% and 17%, respectively. The oxygen concentration had little effect on the apparent activation energy and the mutation characteristics of coal spontaneous combustion process.

（3）Based on the nonlinear relationship between the coal-oxygen composite reaction rate and the oxygen concentration, the four-dimensional nonlinear oxidation numerical simulation model of the residual coal in the goaf was constructed considering the factors such as the mining speed of the fully mechanized caving face, the spatial distribution characteristics of the residual coal, pores in the goaf, and the particle size influence function. The simulation calculation was carried out combined with the key parameters determined by the experimental test, and the accuracy of the model was verified by the natural ignition period and the field gas monitoring results determined by the natural ignition experiment. The simulation results show that the temperature rise of residual coal in goaf has obvious “partition” characteristics, that is, the temperature of residual coal in “two lanes” position is high, the temperature in the middle of goaf is low, and there is a nonlinear relationship between the maximum temperature of goaf and the advancing speed. Oxygen concentration is one of the key factors affecting the temperature rise of residual coal in goaf. When the oxygen concentration decreases to less than 5 %, the temperature rise rate and CO yield of residual coal decrease significantly.

（4）Based on the stage characteristics of residual coal oxidation and the change law of gaseous products under different oxygen environments, three single gas indicators and five commonly used coal spontaneous combustion early warning composite indicators were established. A multi-indicator and stage-by-stage early warning system with R₂, C₂H₄, and C₂H₆ as main indicators and CO, R₁, R₃, R₄, and R₆ as auxiliary indicators was optimized and constructed, and the quantitative relationship between multi-indicator and coal temperature at different stages was determined. Considering the change of predicted temperature and oxygen concentration of residual coal in goaf, and combining with the nonlinear characteristics of coal-oxygen composite reaction, the prediction model of spontaneous combustion period of residual coal in goaf is constructed, which can predict the time required for coal temperature to reach different warning levels under different temperatures and oxygen environments.

（5）Based on the classification and early warning method of coal spontaneous combustion in goaf, the prediction model of residual coal temperature and spontaneous combustion period, the dynamic early warning system of coal spontaneous combustion was developed considering the monitoring and early warning requirements of mine spontaneous combustion disaster, and the field industrial test was carried out in the fully mechanized caving face of a mine in Shaanxi. This method and system can preliminarily realize the dynamic perception and early warning of coal spontaneous combustion risk in goaf, improve the processing efficiency of coal spontaneous combustion information and the stability of early warning results, and avoid the distortion of results caused by single index. The field application effect is good. The research results have certain practical significance for strengthening the prevention and control level of spontaneous combustion disasters in coal mine goaf.