Investigation of long memory in concrete fracture through acoustic emission time series analysis under monotonic and fatigue loading

The significant dependence of current events on past events is known as a long memory phenomenon in stochastic processes. The fracture process in concrete is interesting due to its complexity arising from long-range interactions of multi-scale cracks which may manifest in the long memory. We quantify the long memory effect in the concrete fracture process using the Hurst exponent for acoustic emission measurements. The magnitude and inter-event time series are studied for three different beam sizes under monotonic and fatigue loading. The long memory is strongly observed under fatigue loading. Local variations in the Hurst exponent reveal the fundamental difference between monotonic and fatigue loading, which can be used for damage detection in concrete.