Experimental and Modeling of Residual Deformation of Soil–Rock Mixture under Freeze–Thaw Cycles

Projects in seasonal frozen soil areas are often faced with frost heaving and thawing subsidence failure, the foundation fill of most projects is a mixture rock. Therefore, taking soil–rock different rock contents as research objects, residual deformation under multiple freezing–thawing cycles studied. In addition, deep learning method based on artificial neural network was pioneered combined test mixture, Long short-term memory (LSTM) model established to predict results test. The LSTM has been verified be feasible exploration freeze–thaw cycle law which can not only greatly reduce period test, but also maintain high prediction accuracy certain extent. study found that will repeatedly produce heave thaw action cycles, initial changes hugely. With increase number decreases then becomes steady. Under condition content block more than 80%, content, caused by gradually decrease due skeleton function rock, while content’s further deformation. Furthermore, an effectively freezing short term, shorten time required for improve efficiency