Abstract: A model of association rule mining for the mid-term water cooling parameters is established to tackle the interference of redundant construction data on analyzing the correlation between daily cooling rate and water-cooling parameters. First, the main factors affecting the daily cooling rate are determined, and then the quantitative association rules are extracted from the history monitoring data of dam temperature. On such basis, the main control parameters are determined so as to optimize the mid-term water cooling measures. Results suggest that the water-cooling temperature should be controlled within (14.6,14.8]℃ so as to meet requirements, the initial temperature of concrete should be in (17.7,18.5]℃, and water flow rate between (15.0,22.5] L/min. Water flow rate should be regarded as control focus, so the initial value of cooling flow rate can be preset as within (15.0,22.5] L/min. The research finding is of guiding significance for deciding mid-term cooling measures.

Key words: temperature control of concrete, mid-term cooling phase, cooling rate, data mining, association rules, Xiluodu arch dam