Automatic quality compliance checking in concrete dam construction: Integrating rule syntax parsing and semantic distance

The compliance checking of concrete dam construction quality often relies on the manual reference of specifications, which is time-consuming, labor-intensive, and prone to errors. Automating compliance checking is an effective means of ensuring the quality of dams. However, using computers to match and compare quality record texts with specification provisions remains challenging. Due to the numerous specifications and diverse forms of constraints in dam construction, as well as unstructured and non-standard quality records, semantic differences exist between quality records and specifications that prevent timely quality control, making it necessary to develop a complete and efficient construction quality compliance checking framework. To address these issues, an automatic construction quality compliance checking method based on rule syntax parsing and phrase semantic distance was proposed in this work. First, key entity knowledge was determined and automatically extracted. Then, syntax parsing rules were defined, and the parsing algorithm was used to assemble the key entities into syntax trees. Then, specification and quality record syntax trees were aligned based on node semantic distance. Finally, a scoring process was designed to achieve automatic compliance checking of concrete dam construction quality. The experiments showed that the entity knowledge extraction F1 value of the proposed method was 12.89% higher than similar models, and the checking accuracy was 88.89%, with excellent applicability to both quantitative and qualitative specification constraints. The proposed method constructed a complete quality compliance checking framework for the construction field, automatically obtaining precise scores as checking results, saving considerable time compared to manual checking, and promoting compliance checking throughout the entire lifecycle of construction.