Modular Structure Parameter Generation Algorithm for Rapid Assembly and Disassembly in Temporary Environments

In view of the urgent demand for rapid assembly and disassembly of temporary modular structures, this paper proposes a parametric generation algorithm that places disassembly feasibility as a core constraint of scheme generation, prior to the verification stage at the end of design. A linkage representation system for geometric, interface, and logic parameters is established, and the feasible region of parameters is defined by the non interference disassembly criterion and the connection reuse threshold as hard constraints. On this basis, a hierarchical cooperative strategy for macro topology generation and micro parameter optimization is proposed, and two way information transfer between the two levels is realized via a differentiable surrogate model. A reversibility guided reinforcement learning reward function is designed to enable the generator to evaluate the blocking risk of an action on the subsequent disassembly path in real time during the module by module addition process. A dynamic connector adapter is developed to adaptively match connector parameters according to local stress distribution. Experiments show that, compared with the standard GNN benchmark in three typical temporary scenarios, the algorithm reduces disassembly time by 24.2%-27.2%, the constraint satisfaction rate reaches 93.2%, and the connection reuse rate increases to 86.4%, and the structural safety margin is maintained in the range of 0.38-0.52. The significant advantages of the proposed mechanism in the co optimization of disassembly efficiency and structural performance are verified.