Using small scale tests to predict full scale ice loads

All large old sea-ice features show the imprint of the original first year-ice elements from which they were formed. The paper suggests how this can be used to prepare a mull-element engineering model. Chain assemblies are used to develop the mathematics for multi-element systems. Predicting loads from multi-element systems requires a strength distribution that represents the whole system. It is noted that, no small scale testing programs were designed with this in mind. The available data sets are biased sub-distributions of the potential strength population. A hypothetical strength distribution is used in worked examples ranging from large structure/ice interactions to local ice loads. These examples are compared with full scale experience. This suggests that the model provides a method of predicting full-scale loads from small scale tests. The mathematics also implies the possibility of zero strength. This is discussed and compared with experience. The implications for the development of 100 year design loads are examined. Finally, a cost effective method of gathering data to prepare representative strength distributions is proposed.