| CPC G01N 29/0654 (2013.01) [G01N 29/12 (2013.01); G01N 29/14 (2013.01); G01N 33/383 (2013.01)] | 18 Claims |

|
1. An acoustic emission method for monitoring structural health of concrete comprising:
affixing at least one acoustic emission sensor to a concrete structure;
obtaining at least one acoustic emission waveform from the concrete structure;
employing at least one deep learning algorithm to analyze the at least one acoustic emission waveform to determine a condition of the concrete structure with respect to at least one Alkali-silica reaction progress wherein the at least one deep learning algorithm attributes the Alkali-silica reaction process to an Alkali-silica expansion range and an Alkali-silica reaction damage phase is determined for the concrete structure based on the at least one acoustic emission waveform from the concrete structure;
wherein the at least one deep learning algorithm includes at least one convolutional neural network and at least one stacked autoencoder:
wherein at least one heterogeneous ensemble learning network, including the at least one convolutional neural network, and at least one random forest classifies the Alkali-silica reaction damage phase; and
wherein the method is nondestructive to the concrete structure being monitored.
|