Abstract: An intelligent evaluation method for the interlayer bonding quality of roller compacted concrete (RCC) dam is proposed in the light of reliable evaluation and dynamic control of interlayer bonding quality of RCC dam. (1) An evaluation indicator system with moisture content and compaction degree of RCC thermal layers as evaluation parameters and splitting tensile strength at 90 d-age of RCC core samples as evaluation target is established to reasonably characterize the interlayer bonding quality in the field. (2) The inverse distance weighted (IDW) interpolation method was employed to simulate the discrete parameters obtained by sampling detection with self-developed intelligent devices in the entire work area of RCC dam, and the spatial uncertainty was analyzed quantitatively by comparing the parameter simulation accuracy of sample sequences with different quantities and different mesh sizes. (3) The intelligent evaluation model for the interlayer bonding quality was established based on Bagging algorithm integrated with back-propagation artificial neural network (BP-ANN). The model was applied to the dynamic evaluation of the interlayer bonding quality of a typical construction warehouse of Wunonglong RCC Dam. Results suggest that the present method accurately and dynamically evaluates the interlayer bonding quality in consideration of spatial uncertainty, and also integrates the intelligent perception, transmission and evaluation of construction information.

Key words: RCC dam, interlayer bonding quality, intelligent evaluation, Bagging, BP neural network