Objective Aiming at the challenge that implicit knowledge in ship design drawings is difficult to express and extract effectively, this study intends to explore potential design rules from existing design data and provide data support for the optimization of ship fire protection design.

Methods Taking 400 fire protection design drawings of civil passenger ships in inland river basins as the research object, the drawings were first standardized and preprocessed to extract key indicators such as fire compartment area, safety exit width, and fire wall fire resistance rating, and a transaction dataset was constructed. By comparing the time consumption of three association rule algorithms (Apriori, FP-growth, and Eclat) on small datasets, the Apriori algorithm was selected. The dataset was divided into a training set of 300 samples and a test set of 100 samples. Reasonable minimum support and confidence thresholds were set to mine frequent itemsets and strong association rules from the data.

Results The proposed method can effectively identify implicit knowledge in design drawings. Core rules such as "the layout of service desks is closely associated with escape exits", "the proportion of fire compartment area ranges from 10% to 30%", and "the fire resistance rating of fire walls increases with the expansion of building area" were discovered. Moreover, the minimum accuracy of 8 key rules in the test set reached 95%.

Conclusion The effectiveness of the Apriori algorithm in extracting implicit knowledge from ship design was verified. The obtained rules can supplement the current explicit design specifications, provide decision support for fire protection design optimization, and promote the transformation of ship design from experience-driven to data-driven. In addition, the method can be extended to other fields of ship design.