考虑内外检测数据对齐的管道完整性评价技术研究

Research on pipeline-integrity evaluation technology considering alignment of internal and external inspection data

针对目前内外检测数据的对齐主要依靠人工核查和开挖验证,无法充分利用挖掘检测信息的问题,分别从内检测和外检测数据中提取特征要素信息和里程信息作为输入、输出变量,采用极限梯度提升(XGBoost)算法建立两者的映射关系,预测内检测点在外检测点的里程,以内检测信息为基准,采用逐步平移实现内外检测数据的对齐,并根据对齐结果进行综合分析。结果表明,数据对齐后所有站场和阀室的里程误差均在5 m以内,经纬度分布变化高度一致,对齐结果点满足精度要求;实例分析中管道受交、直流干扰程度较小,管道全线阴极保护正常,虽然土壤腐蚀性较强,但存在金属损失且防腐层破损处未见明显的腐蚀坑和腐蚀产物,说明管道运行状态良好。研究结果可为提升管道完整性管理水平提供理论依据。

As the alignment of internal and external inspection data mainly relies on artificial judgment and excavation verification, the inspection data cannot be fully utilized and excavated. Considering this problem, this paper extracts the characteristic elements and mileage information from the internal and external inspection data as input and output variables, respectively, and adopts the extreme gradient boosting(XGBoost) algorithm to establish the mapping relationship of the two to predict the mileage of the internal inspection point at the external inspection point. With the internal inspection data as a benchmark, the alignment result of internal and external inspection data is obtained by gradual translation, and the comprehensive analysis is made according to the alignment result. The results reveal that the mileage errors of all stations and valve chambers are within 5 m after data alignment, and the latitude and longitude distribution changes are highly consistent, which means the alignment result points meet the accuracy requirements. In the case analysis, the pipeline is slightly affected by alternating current(AC) and direct current(DC) interference, and the cathodic protection of the whole pipeline is normal. Although the soil is highly corrosive, no obvious corrosion pits and corrosion products can be found in places with metal losses and anticorrosion layer damage, which indicates the good operating condition of the pipeline. The research results can provide a theoretical basis for improving the management level of pipeline integrity.