A graph theory-based automated design method for nuclear power plant corridors

Corridor design is a critical yet challenging task in nuclear power plant layout planning. To improve efficiency, this study proposes a graph theory-based automated corridor design method. The approach models path connectivity via an adjacency matrix, encoding all possible inter-partition routes. Specifically, a Path Clustering Algorithm (PCA) employing Dijkstra’s algorithm prioritizes paths by path overlap ratio and merges key routes into an optimized layout, which is subsequently expanded to meet width requirements. The method iteratively constructs path graphs, clusters routes, and expands corridors until all equipment rooms connect to the entrance. Experimental results demonstrate that the algorithm reduces total corridor length by up to 57.47% compared to Dual-Source Shortest Path approaches across various layout scales. When integrated with equipment layout optimization, the method maintains high space utilization, with effective area occupancy exceeding 87% for corridor arrangements. These findings validate the efficiency and general applicability of the proposed approach.