中文摘要：

耐火材料损伤模式识别的研究在工业生产中具有重要意义,然而在利用声发射仪器对耐火材料进行采集得到的声发射信号饱受环境噪声、声发射通道以及传感器特性的影响。为了恢复损伤源的实际声发射信号,提出了利用粒子滤波技术对采集的声发射信号进行滤波。首先在地震波模型的基础上,建立了声发射源信号的状态空间模型,然后再对确定信号的进行仿真实验,证明了粒子滤波的有效性。最后将粒子滤波用于实际声发射源信号恢复,取得了一定的效果,为后续声发射信号特征提取打下了良好基础。

英文摘要：

The research of refractory damage pattern recognition has important significance in industrial production, but the acoustic emission signals which was collected from refractory using acoustic emission instrument suffered the impact of environmental noise, acoustic emission channels and sensor characteristics. In order to restore the actual acoustic emission signals of damage source, it proposed to filter the acquired acoustic emission signal with particie filtering technology. First, on the basis of seismic model, the AE source signal state space model was established and then the simulation experiments with deterministic signal was completed, the effectiveness of particle filter was demonstrated. Finally, particle filter was used for the recovery of actual acoustic emission source signal and achieved certain results which laid a good foundation for the acoustic emission signal feature extraction.