Yunlong Reservoir Dam Safety Monitoring Data Analysis Methods

Combining the prototype observation information of YunLong reservoir, on the basis of comprehensive summing up study of predecessor on statistical model, aiming at the shortage of least squares regression, complying with partial least squares regression method and recursion partial least squares regression, partial least squares regression statistical model and recursion partial least squares regression statistical model of dam safety monitoring are take out and the performance of both models are compared and analyzed in this article. This study work not only has important utilization value of engineering practice but also has significant meaning to improve management level of dam safety in China.Main study content and results of this article are shown as below:(1) Analyze the open question of least squares regression method during the process of present statistical modeling, point out that the serious multiple correlation among factors is the root reason resulting on structural instability and weak explanation of least squares regression.(2) On the theory of partial least squares regression method partial least squares regression statistical model is taken out to utilize the statistical model for modeling. The study and analysis indicate that this model can overcome serious multiple correlation among factors and statistical model with stable structure and clear explanation. After checking calculation of engineering field data and compare with least squares regression this model is proved as a powerful tool of modeling with multiple correlations among factors.(3) The traditional partial least squares regression has a shortage as below: The model will strike root as soon as it is constructed. When changes of process characteristic or operating condition takes place it can not be updated in time. To overcome this shortage recursion least squares regression method of block type is put forward. The actual case analysis shows that this model can adapt the change of dam operating state well and can explain safety state of dam more reasonably.