High-crustal-stress soft rock mass crustal stress calculation method based on deep neural network

The invention discloses a high-crustal-stress soft rock mass crustal stress calculation method based on a deep neural network. The method comprises the following steps: establishing a three-dimensional crustal stress calculation model of drill holes with different diameters; substituting the rock mass material parameters into the three-dimensional ground stress calculation model to obtain positive strain data at different positions of the hole wall; different parameters of the rock mass material and positive strain data at different positions of the hole wall are substituted into a deep neural network algorithm, an elastic-plastic mechanical stress relieving calculation method is established, and finally the relation between the hole wall stress and the far-field ground stress is obtained; a local wall surface stress relieving method is adopted to measure and obtain enough hole wall positive strain data; taking out a rock core of a deep test rock stratum by using side wall coring equipment, and carrying out a