PE-HMM Based Operation Status Evaluation Method for Aqueduct
With the continuous development of long-distance,high-flow,and large-span aqueduct projects,the moni-toring and evaluation of the aqueduct's operating status has become increasingly important.The Changgangpo aqueduct project in Luoding City,Guangdong Province is used as the project for the study.Based on the discharge vibration dis-placement data of the aqueduct,a method based on Permutation Entropy(PE)and Hidden Markov Model(HMM)for e-valuation of aqueduct operation damage state was proposed.Firstly,the Permutation Entropy algorithm and the K-means method were adopted to extract the basic features of the vibration data for generating the observed state sequences of the HMM model.Secondly,the model parameters were trained using the HMM algorithm,and the mean absolute percent-age error was used as a metric to filter the best model parameters.These parameters as the initial values were used to re-train to obtain the probability distribution of the hidden state during the operation of the aqueduct.Finally,combining the score level and probability value corresponding to the hidden state of the aqueduct during the operation period,the desired value of the operational status of the aqueduct was obtained to quantitatively assess the operational status of the aqueduct.The evaluation results of aqueduct operation state based on PE-HMM method are consistent with the results of field in-vestigation.It can be seen that the PE-HMM method can truly reflect the running state of aqueduct structure from the per-spective of aqueduct vibration displacement data,and has high evaluation accuracy and engineering guiding significance.
aqueductevaluation of operating statuspermutation entropy algorithmHidden Markov Model
万方数据
维普
