Understanding Causes of Concrete Culvert Pipe Joint Separation

Summary

Joint separations affect approximately 20% of the Minnesota concrete pipe inventory (Taylor and Marr 2012). The goals of this research are to determine the factors that contribute to joint separation and recommend practices to mitigate this problem. Research efforts follow a three-pronged approach: (1) Examination of TAMS HydInfra database; (2) field survey of concrete culverts; and (3) computational modeling of soil-culvert systems. Database investigations were conducted using a Random Forest model along with individual feature analysis to determine factors that correlate with joint separations. Geographic features such as the county and route number of the culvert are more important for predicting joint separation than geometric features such as culvert size or cover depth. Field surveys reveal that separations typically occurred at the ends of pipes, or in the first untied joint from the end of the pipe if only some joints were tied. Joint separation was often observed alongside other distress such as infiltration or inslope voids. Computational modeling results show that embankment self-weight and traffic loading concentrate their highest demands under the center of the road, which did not match with field inspections. Soil freezing or changes of water-table level impose greater demands than traffic, and these maximum demands are located at the pipe ends, and thus these mechanisms are more likely to result in joint separation. Recommendations include tying all joints on installation, properly compacting the backfill, and limiting freezing expansion of the embankment materials, particularly if a cohesive soil cap is used.