关键词: 堆石料, 压实质量, 评估模型, 碾压参数, 现场碾压试验, 径向基神经网络

Abstract: Compaction quality of earth-rock dams is crucial to dam safety, and for rockfill material it can be monitored in real time via a compaction quality assessment model. Rolling parameters are kept constant during traditional compacting construction using a fixed scheme, while in intelligent compaction they are adjusted and optimized based on the compaction state in-situ. Thus a good model for compaction quality assessment should take rolling parameters into account. This paper presents an analysis of the correlation of compaction meter value (CMV) with the relative density of rockfill materials and rolling parameters based on field compaction tests. Results show that CMV is strongly correlated with the relative density and it can be used as a good index for compaction quality assessment of rockfill materials. And it is significantly influenced by roller vibrating frequency and roller speed, while driving direction is an insignificant factor. Using the field test data and a radial basis function (RBF) neural network, we develop a compaction quality assessment model involving rolling parameters which achieves a high accuracy verified by the field test results.

Key words: rockfill materials, compaction quality, assessment model, rolling parameter, field compaction test, radial basis function neural network