<p>We present a robust and efficient algorithm for handling deformable solids coupling with smoothed particle hydrodynamics fluids. We introduce signed distance fields (SDF) and volume maps with variable boundary contributions during runtime and store them on a spatial grid. As the solid deforms, we need to efficiently update the representation of the solid boundary to reflect its new shape. To reduce computational cost, we propose a local update strategy. Instead of updating the entire simulation domain, we only update the regions near the solid boundary that are affected by the deformation. To update the SDF, we develop a grid node search algorithm based on dynamic triangular patches to obtain node change values. Furthermore, we propose a neighbor grid node update method based on triangular vertex movement to calculate increments of volume maps, quantifying the changes in solid shape. We further demonstrate that our approach is suitable for GPU parallel computing, and handles diversiform scenarios, even including solid fracture simulation mesh or aberrant deformation rectification.</p>