The invention discloses a heterogeneous discrete element particle
motion simulation method,
system and device based on a graph neural network, a medium and a product, and relates to the field of discrete element particle
motion simulation.The method comprises the steps that state information and interaction information of heterogeneous discrete element particles at the
current time step are obtained, and a heterogeneous discrete element particle
motion simulation model is used for simulating the state information and interaction information of the heterogeneous discrete element particles at the
current time step; determining the acceleration of the heterogeneous discrete element particles at the next
time step; the heterogeneous discrete element particle motion
simulation model is obtained by training an initial model by using a training
data set; the initial model comprises a vertex
encoder, an edge
encoder, a graph neural network and a decoder; and determining the speed and position of each
time step in the future set time period by using an explicit Euler integral formula according to the acceleration of each
time step in the future set time period, and obtaining the motion trail of the heterogeneous discrete element particles in the future set time period. According to the method, the
simulation efficiency and the
simulation precision are improved, and accurate prediction of the motion trail of the heterogeneous discrete element particles is realized.