煤炭的自燃隐患一直是影响煤炭安全生产的重大威胁之一，是自然界存在的一种客观现象，并且这种现象十分隐蔽，不容易被觉察。煤炭自燃如果不及时预防，那么给人类的经济、财产、安全等带来的损失是无法估量的。有数据表明，由煤矿自燃所引发的火灾占火灾总数的90%以上。因此，如何提前预防煤炭的自燃就成了各国科研学者研究的首要的任务。 为了能够及早的探测自燃发火点，并采取有效的防治措施，本文在理论层面上将在弱信号检测领域具有广泛应用的混沌理论应用在煤炭自燃的检测上。由于混沌理论对周期信号非常敏感，而对噪声干扰免疫，使得这一理论研究成为可能。主要研究内容和成果如下： 1) 针对传统的煤炭自燃预测方法的不足和不及时性等问题，结合煤炭自燃过程中伴生气体的特性，通过比较煤炭自燃过程中产生的CO气体和烃类气体的性质，分析并选取CO气体作为预测煤炭自燃隐患点的指标气体。 2) 对CO的时间序列进行了相空间的重构，并利用最大Lyapunov指数法验证了煤炭自燃所释放出的CO气体的混沌特性。 3) 建立了Duffing混沌振子的仿真模型，描述了Duffing振子检测CO浓度的原理，以此来实现对煤层自燃隐患点的预测。 4) 设计了混沌Duffing振子实验仿真系统，通过该系统可以直观的观察出各参数的变化对系统状态的影响。

Coal’s spontaneous combustion troubles are one of the greatest threats to the coal’s safe production. Spontaneous combustion is an objective phenomenon in the nature, but is invisible. If this phenomenon is not prevented, losses of economy, property and safety will be immeasurable. By statistics, conflagrations resulted from spontaneous combustion take up of 90 percent of the total. So, how to precaution against spontaneous combustion in advance is the most important task for researchers and scholars in the different countries. In order to be capable of early detecting the spontaneous combustion source and take effective measures in time, this thesis applies chaos theory which is widely used in weak signal detection to detect spontaneous combustion in theory. The main research content and results are as follows: 1) In view of problem that traditional method’s defect in spontaneous combustion, considering gas nature produced in the progress of spontaneous combustion of coal, by comparing the properties of the CO gas with the properties of the Non-hydrocarbon gases produced in the process of the coal’s spontaneous combustion, this thesis analyzes and selects CO gas as the index gas predicting spontaneous combustion. 2) The time series of CO is reconstructed and the biggest Lyapunov index method is used to prove the chaos characteristics of CO gas. 3) The simulation model of Duffing chaos oscillator is established and the principle of Duffing oscillator testing CO is described. These can realize the prediction to Coal’s spontaneous combustion troubles. 4) Duffing simulation system is designed. Ones the system’s parameters change, the system’s state can be changed immediately. By this system, parameters playing what king of roles can be presented easily.