中文摘要：

为了对镁碳质耐火材料损伤声发射信号进行有效的特征提取,研究了利用非线性领域的分形方法进行特征提取的可行性。首先利用仿真声发射信号研究了多重分形谱参数（Δα、Δf、K）,广义分形维数均值Mean Dq等参数表征信号频谱结构特征的能力,结果表明多重分形谱宽Δα能够有效的表征信号的频谱特征;然后对镁碳质耐火材料进行单轴压缩试验,采集损伤声发射信号,并计算出两种典型损伤声发射信号的多重分形谱宽Δα,通过多重分形谱宽的分布证实了利用其进行特征提取是可行的。

英文摘要：

In order to extract the feature of damage acoustic emission signal of magnesia carbon refractory effective, it studied the feasibility of feature extraction method by usingfractal theory of nonlinear field, A t first, it utilized the simulation signal to study the signal spectrum structure features characterizing capability of parameters of multi-fructal spectrum （△α, △f, K）, general dimension mean MeanDq et al, and the results show that the width values of multi-fractal spectrum can characterized the signal spectrum characteristics effectively; then the magnesium carbon refractory is subjected to uniaxial compression test, and collects damage acoustic emission signals, and calculates the multi-fractal spectral width △α of two typical damage acoustic emission signals, according to the distribution of multi-fractal spectrum width it confirmes that the feature extraction by using it is feasible.