The geometric shape of compression ridges is easily recognizable in profiles, and

thrust faults often appear simultaneously with compression ridges, forming a series of raised peaks and depressed valleys that alternate between peaks and valleys. The seismic profile shows strong amplitude reflection characteristics ( Figure 4(A) ). Squeezing ridges are widely developed in the toe area due to the downward transport of gravity flow sediments. Most sediments accumulate at the toe to form thrust faults, and under certain slope conditions, the accumulation of sediments at the toe can also form a certain scale of squeezing ridges, which are commonly developed in block transport in deep water and on land ( He et al., 2018 ). The development of thrust faults causes uneven terrain and is a favorable area for turbidity current deposition unloading.

The bottom interface is a weak surface formed by MTDs under the action of gravity, generally with good continuity. During the process of gravity flow sedimentation moving downwards along the bottom interface, due to the varying degrees of resistance, consolidation ability at different locations, and erosion resistance of the bottom interface, the bottom interface will be damaged by sediments, forming erosion scratches. The concave part of the damaged bottom interface is called an erosion groove ( Figure 4(B) ). The width of the erosion groove discovered in the research area is up to 3 km. The upper part of the erosion scratch often develops structures such as thrust faults, and the distribution of erosion grooves at the bottom interface can be clearly seen through the RMS property map. There are two reasons for their formation: 1) erosion of underlying strata by high-speed fluids; 2) The floating blocks carried by high-speed fluids jump forward, causing erosion of the bottom strata ( Qin, 2012 ).

1) Volume fraction equation:

Based on the Euler two-phase flow simulation used in the article, the two phases are considered as interpenetrating continuum, therefore, a volume fraction is required to represent the space occupied by each phase in the control volume. The volume occupied by the q-th phase is defined as:

$V_q={\displaystyle \underset{V}{\int }{\alpha }_q\text{d}V}$ (1)

The total volume fraction of two different phases is 1:

$\underset{q=1}{\overset{n}{{\displaystyle \sum }}}{\alpha }_q=1$ (2)

Therefore, the effective density of the q-th phase: ${\hat{\rho }}_q={\alpha }_q{\rho }_q$ In the equation, α_q is the volume fraction of the q-phase, and at the same time, the Euler model shares a common pressure value ( Zeng et al., 1997 ).

2) Continuity equation:

The Euler method is to take any fluid contained in a control body (or control element) as the research object in a space filled with fluid particles. For a certain control body, the law of conservation of mass can be expressed as: the increase in fluid mass in the control body per unit time = ∑[the mass of fluid flowing into the control body per unit time − ∑the mass of fluid flowing out of the control body per unit time]. Therefore, the continuity equation for the q-th phase is ( Hu et al., 2022 ):

$\frac{\partial }{\partial t}\left({\alpha }_q{\rho }_q\right)+\nabla \cdot \left({\alpha }_q{\rho }_q{\stackrel{\to }{v}}_q\right)=\underset{p=1}{\overset{n}{{\displaystyle \sum }}}\left({\stackrel{\dot{}}{m}}_{pq}-{\stackrel{\dot{}}{m}}_{qp}\right)＋S_p$ (3)

( ${\stackrel{\dot{}}{m}}_{pq}$: transfer from the p-th phase to the q-th phase; ${\stackrel{\dot{}}{m}}_{qp}$: transfer from the q-th phase to the p-th phase; $S_p$: Source item, default is 0; ${\stackrel{\rightharpoonup }{V}}_q$: velocity of the q-th phase.)

3) Momentum conservation equation:

Any flow problem must satisfy the law of conservation. According to Newton’s second law, write the momentum change of the q-th phase in the control body: the rate of change of momentum over time + the momentum change caused by the fluid passing through the surface of the control body = the force acting on that phase in the control body [26]. The momentum equation is:

$\begin{array}{l}\frac{\partial }{\partial t}\left({\alpha }_q{\rho }_q{\stackrel{\to }{v}}_q\right)+\nabla \cdot \left({\alpha }_q{\rho }_q{\stackrel{\to }{v}}_q{\stackrel{\to }{v}}_q\right)\\ =-{\alpha }_q\nabla p+\nabla \cdot {\overline{\stackrel{¯}{\tau }}}_q+{\alpha }_q{\rho }_q\stackrel{\to }{g}+\underset{p=1}{\overset{n}{{\displaystyle \sum }}}\left({\stackrel{\to }{R}}_{pq}+{\stackrel{\dot{}}{m}}_{pq}{\stackrel{\to }{v}}_{pq}-{\stackrel{\dot{}}{m}}_{pq}{\stackrel{\to }{v}}_{pq}\right)+\left({\stackrel{\to }{F}}_q+{\stackrel{\to }{F}}_{lift,q}+{\stackrel{\to }{F}}_{vm,q}\right)\end{array}$ (4)

( ${\overline{\stackrel{¯}{\tau }}}_q$: stage stress of phase q; $\stackrel{\to }{g}$: gravitational acceleration; ${\stackrel{\to }{F}}_q$: volume force; ${\stackrel{\to }{F}}_{lift,q}$: lift force; ${\stackrel{\to }{F}}_{vm,q}$: virtual mass force; ${\stackrel{\to }{R}}_{pq}$: the interaction force between phases; ${}_{\stackrel{\to }{\text{v}}}$: Interphase velocity.)

The professional modeling software for geometric structures on the market includes CAD, Solid Works, etc. However, considering software compatibility issues and subsequent secondary modifications, Fluent’s built-in Space Claim component is used. Open this component and based on the principle of geometric similarity, while ensuring the same Froude number (Fr = V/√ gl), set the scale of the 2D model to 1:1000 and create a rectangle with a length of 30 m and a width of 18 m. Divide the model into two regions, Flow up and Flow down, and establish a mesh model of the canyon wall ( Figure 5 ). After establishing the geometric model proportionally in Space Claim, import it into the mesh for mesh partitioning, set the mesh density to 2 mm × 2 mm, and have a total of 72,585 nodes (65,344 elements).

The collapse of canyon walls is often closely related to various factors such as

sediment accumulation, heavy precipitation, and geological activity. In order to better explore the dynamic response of canyon wall collapse, this paper sets up a transverse waterway model of canyon wall collapse ( Figure 5 ). In order to observe the process of canyon wall collapse more intuitively, the canyon wall collapse model is simplified. The initial left and right slopes of the canyon wall are set differently, with the left wall steep and the right wall gentle.

In the general setting, we use transient simulation to observe the sedimentary distribution of turbidity currents. After the control equation is discretized, it can be solved. Here, the pressure velocity coupling scheme uses the most widely used flow field calculation method, implicit SIMPLE algorithm, which is mainly used for prediction and adjustment to achieve convergence faster. In order to make the turbidity current sedimentation process more complete, the iteration number is uniformly set to 200 steps, and the step size is set to 0.1. Add water and sand to the material (where sand refers to turbidity particles, specific parameters are set in Table 2 , refer to Wang Yue’s [9] settings for turbidity particle size and related parameters). Set up solid-liquid mixing inflow at the entrance.