An integrated method for evaluating and predicting long-term operation safety of concrete dams considering lag effect

Effective operation safety evaluation of concrete dams is critical for ensuring the longevity and quality service of a dam. This paper introduces a novel method for quantifying the safety status of concrete dams and predicting future long-term safety performance, considering lag effect of indices. First, lag effect of operation indices is quantified using the modified moving average-cosine similarity method, based on which a comprehensive safety evaluation index system is established. Second, analytic hierarchy process is used to determine the subjective weighting of each index. Considering data correlation, a new method named coefficient of discreteness and independence is proposed to calculate the objective weighting of each index using maximal information coefficient. The final actual weighting of each index is assumed to be a linear combination of the above subjective and objective weightings. Third, based on the long-term monitoring data of a concrete dam, the safety score of a concrete dam can be quantified using technique for order preference by similarity to an ideal solution. Finally, neural networks (NN) are used to predict future long-term safety performance as a faster and simpler way to obtain future safety score. The effectiveness of this proposed method is verified through a case study. The case study showed that structural safety, environmental safety, and total safety scores of a concrete dam can fluctuate periodically, but the overall performance trend is relatively stable, as expected in real-world cases. NN were found to be accurate in predicting future safety performance.