多孔介质材料破坏失效的雪崩特征与其破坏机制密切相关。为探究基于声发射统计的混凝土拉压破坏过程中的雪崩特征，开展了混凝土单轴压缩和巴西劈裂试验，并监测加载过程中的声发射信号，对比分析了声发射信号的幅值、绝对能量、持续时间、等候时间等参数的统计特征。结果表明：拉压破坏下声发射能量最大值均分布在峰值应力附近，但劈裂过程中的信号数量远低于单轴压缩。b值均呈现先增大后减小趋势，而劈裂破坏下的b值下降更早，变化幅度更大，说明大裂纹出现的更早，更容易破坏。拉压破坏下的声发射绝对能量、幅值、持续时间的概率密度分布均符合幂律，但对应的雪崩指数值不同，与破坏过程中的失效机制相关。其中劈裂破坏的绝对能量分布接近纯幂律分布，而压缩破坏符合带阻尼的幂律分布，所对应的能量指数值更大。不同时间段内单轴压缩和巴西劈裂破坏的能量指数值呈现下降趋势，说明随荷载增加试件内部大能量信号占比增大。拉压破坏下的等候时间概率密度分布较为接近，表现出稳定性，也说明了统计规律的适用性。

The avalanche characteristics of destruction of porous dielectric materials are closely related to its failure mechanism. In order to investigate the avalanche characteristics of concrete during tensile and compressive failure based on acoustic emission statistics, the concrete uniaxial compression and Brazil splitting tests were carried out. And the acoustic emission signals in the loading process were monitored, and the statistical characteristics of the acoustic emission signals such as amplitude, absolute energy, duration and waiting time were compared and analyzed. The results show that the maximum acoustic emission energy is distributed near the peak stress, but the number of signals in the splitting process is much lower than that in uniaxial compression. All the b-value increases first and then decreases, and the b-value decreases earlier and the change range is larger under splitting failure, indicating that the large cracks appeared earlier and is more easily destroyed. The probability density distributions of absolute energy, ?amplitude and duration of acoustic emission under tensile and compressive failure conform to the power law, but the corresponding avalanche index values are different, which is related to the failure mechanism during the failure process. The absolute energy distribution of the splitting failure is close to the pure power law distribution, while the compressive failure accords with the damped power law distribution, and the corresponding energy index value is larger. The energy index values of uniaxial compression and Brazilian splitting failure in different time periods show a downward trend, indicating that the proportion of large energy signals in the specimen increase with the increase of load. The probability density distribution of waiting time under tensile and compressive failure is close, which shows the stability, and also shows the applicability of statistical law.

国家自然科学基金（52204201）