Author(s): K. R. Croasdale; B. W. Graham; G. Comfort; R. W. Marcellus
Linked Author(s):
Keywords: No Keywords
Abstract: A project was conducted to evaluate how well various types of ice pressure sensors and strainmeters measured in-situ ice stresses, and thereby, how accurately they predicted ice loads on a structure. The project was conducted in a large open-air test basin with saline ice up to about 0.7 m (28 inches) thick. The ice sheet was approximately 30 m (100 ft) wide and 50 m (164 ft) long and was loaded against a reaction structure which was 3 m (10 ft) wide. During the 22 tests which were performed, loads up to about 155 tons (1386 KN) were applied through the ice sheet and measured at the structure. Both short-term (elastic) and long-term (creep) tests over periods up to three hours were performed. In total, 10 different types of ice sensors were used in the experiment. These included wide thin panels of various internal designs, small hard cylindrical sensors, small thin hydraulic disk sensors, and two types of surface strain meters. The loads on the reaction face were measured with three load cells which were supported by three frame structures fixed to the bottom of the basin. In the data analysis, the effect of the uneven load support system on the stress distribution in the ice sheet was predicted by finite-element analysis (for key tests) using the measured loads. A comparison between actual nominal ice pressure on the structure (based on the reaction loads) and that predicted from the sensor readings (combined with a finite-element analysis of stress distribution) was made for each sensor for all tests. The results of the experiment indicate that for most sensors, their overall ability to predict loads on a structure from in-situ ice stress (or strain) readings is quite good. The average percentage difference between the actual ice loads on the structure and loads predicted based on the sensor readings was better than about 30% for most of the cases.
Year: 1986