中文摘要：

针对矿用救生舱舱内环境动态评价的问题，采用支持向量机多重分类机制进行评价研究。确定了舱内环境评价的等级指标；将环境评价问题处理为分类和分级两部分，分别给出了相应的函数。使用粒子群优化算法进行了参数优化；制订了具体的评价方案，得出了动态评价结果。经现场试验证明，该环境评价方法可以对救生舱舱内环境进行有效、实时的动态评价，算法收敛速度较快，为救生舱舱内环境调控提供了可靠的理论基础和充分的现场数据。

英文摘要：

Aiming at the problem of dynamic environment assessment in coal mine refuge chamber, support vector machine multi-classification was adopted to assess research. The classification targets of inner environment were de- fined. Environment assessment of refuge chamber was divided into two parts, which were classification and grading The corresponding functions were given. The parameters were optimized by using particle swarm organization, for- mulates specific assessment scheme, and final dynamic assessment results were obtained. It was proved that the as- sessment scheme can assess inner environment of coal mine refuge chamber in real time and effectively by experiments, which are accomplished in a practical chamber. Furthermore, the convergence speed of assessment algorithm was improved. It provided reliable theoretical basis and adequate data.