Noise of collapsing minerals: Predictability of the compressional failure in goethite mines

Compression experiments of goethite samples from an iron ore mine in New Caledonia revealed the collapse of the porous samples to follow a power law behavior. The porosity varies between 54 and 84%. The collapse under compression occurs in a series of individual events (avalanches). Each avalanche leads to a jerk in sample compression and an equivalent acoustic emission signal. The probability to find an acoustic emission signal with an energy within E and E + dE is p(E)dE, which has a power law distribution p(E) similar to E-epsilon, and reveals avalanche criticality. The energy exponent E varies systematically with the porosity of the sample between 1.6 and 2. The results are compared with previous measurements of porous silica (Vycor), which showed a slightly smaller exponent of 1.4. We observe fore- and after-shocks of the largest events. Significant correlations between the largest avalanches and fore-shocks were found in samples with high porosity. These correlations open the possibility for the prediction of a major collapse by acoustic detection of noise.