水电工程施工安全隐患关联规则挖掘

Mining association rules of near-misses of hydropower projects construction

为高效挖掘水电工程隐患排查中积累的大量数据,采用基于互信息和左右信息熵的短语提取技术,从非结构化的隐患数据中提取结构化关键短语;运用桑基图,描述隐患特征、隐患部位、隐患类型之间的信息流动;结合短语提取技术,优化Apriori关联算法,挖掘隐患部位与隐患类型的关联规则,绘制关联规则网络图,结合桑基图分析规则;验证某水电站13820条隐患数据案例。结果表明:开关箱、防护栏杆等是隐患排查的重点,易因排架电焊机、开关箱、电缆线凌乱等对象在闸门处发生触电;边坡和高处坠落等关联度较高,可为水电工程隐患排查和治理提供参考。

To efficiently mine the large amount of text data accumulated in the near-misses investigation of hydropower projects,phrase extraction technology based on mutual information and left-right information entropy was adopted to extract structured key phrases from unstructured near-misses data.Sankey diagram was used to describe information flow among characteristics,locations and types of potential hazards.Combined with phrase extraction technology,Apriori association algorithm was optimized,association rules of hidden danger locations and types were mined,association rules network diagram was drawn,and rules were analyzed with Sankey diagram.13820 hidden danger data cases of a hydropower station was verified.The results show that near-misses data switch box guardrails is key of hidden perils,such as easy for shelving electric welding machine switch box cable clutter object at gate,such as electric shock stick slope,falling of correlation degree is higher,study provides hidden perils in hydropower engineering and management.