Abstract: In view of the shortcomings of single model used to simulate and forecast the data sequence of dam monitoring effect quantities, a new combinatorial model is constructed and an engineering example is given in this paper. In this combinatorial model, trend component, periodic component and random component of the monitoring data sequence are respectively simulated by Verhulst model,periodic extensional model and AR(p) model. The forecast methods for monitoring effect quantities can be enriched and the overall forecast accuracy can be improved with this combinatorial model, and our understanding of the variation regularity of monitoring effect quantities can also be deepened.

Key words: dam monitoring, combinatorial forecast, Verhulst model, periodic extensional model, AR(p) model