Experimental assessment of crack prediction methods in international design codes for edge restrained walls

Through cracking resulting from external restraint of early-age thermal and long-term shrinkage strain is a significant issue in the construction industry as it causes leakage in water retaining and resisting structures. Concerningly, a recent field study found restraint induced crack widths to frequently exceed crack widths calculated in accordance with UK design practice (BS EN 1992-3 and CIRIA C766). Due to a lack of pertinent data, the reasons for this are uncertain. This paper compares measured and predicted crack widths in a series of 12 full-scale edge restrained walls constructed in the laboratory. The tests examine the influence on cracking of key parameters including concrete mix design, wall reinforcement ratio, wall aspect ratio and relative wall to base cross-sectional area. The measured and calculated crack widths are compared at first cracking and at the end of monitoring. Two types of behaviour were noted in the tests, dependent on when the first cracks formed. Cracking either occurred at early age, within 24 h of stripping the formwork, or later due to restraint of combined early age thermal contraction and shrinkage. The final crack widths were greatest, by a considerable margin, in walls where cracks formed at early age, despite the initial cracks being very narrow. BS EN 1992-3 gives the best estimates of crack width in the two walls that cracked at early age. Crack widths in these walls were significantly underestimated by C766. In the other 10 walls, which cracked later, C766 tends to give the best estimate of crack width.