Rapid Analysis And Prediction On Temperature Of Dam In The Concreting Process Based On Data Mining

With the rapid pouring of mass concrete blocks of high dam without longitudinal joint, the temperature control and crack prevention of dam in the concreting process have become the key technical problem of dam safety. With the advent of the era of big data, the dam project is gradually realizing the transformation from "traditional dam" to "digital dam" to "intelligent dam". The establishment of construction period database based on the massive construction data and monitor data of dam in the concreting process, and the built of rapid and precise data analysis and prediction model based on data mining, which can feedback the temperature of dam during construction period and provide decision support for real-time temperature control,are important for dam construction. This research is based on the background, whose achievements are concluded as follows.1 A construction period database is established, effectly restoring and managing the data, and providing fast query and extraction data interface for data analysis.The construction information and monitor data are cleaned, eliminating the noise value to ensure the accuracy of the data. Then they are fused, making it convenient to store and analysis.2 The measured temperature values are made qualification check on the basis of the temperature control standard(the allowable maximum temperature, the basic temperature difference, the temperature difference between the upper and lower concrete, the temperature difference between the inner and surface concrete) of the specification for roller compacted concrete dam.Through data analysis, the allowable maximum temperature is the main criterion to determine whether the thermometer points is qualified, and the pouring time is the key influence factors of whether satisfying the requirements of maximum temperature.3. The three dimensional finite element is used to simulate of the temperature and stress field of dam. The measured temperature process of thermometer points are close to the simulated temperature process and the measured equivalent strainprocess of strain gauge measuring points are also close to the simulated strain process, demonstrating the model has high precision.The dam had “high temperature range”in the concreting process, and the temperature declined to reasonable range after a month of completed construction, and the compressive stresses meet the requirements of the compressive strength, the tensile stresses basicly meet the requirements of the tensile strength,excep the inner region of the upper dam, which can guarantee the safe operation of dam.4. The associated rules between the highest temperature of dam concrete during the construction period and the pouring information are established based on Rough Sets and Neural Network.When the ratio of the training samples and predict samples is 4:1, the 50% residuals of the forecast maximum temperature is within 1℃, 73% residuals is within 3℃.The established Neural Network model has high prediction accuracy, which can be used to predict the maximum temperature of concrete during the construction period.5. According to measured temperature process, the associated rules between temperature process of concrete during the construction period and the pouring information are established based on Rough Sets and Support Vector.When the ratio of the training samples and predict samples is 4:1, the residuals of the temperature of concrete during the whole construction period fluctuates in the vicinity of 3℃, the established Support Vectormodel demonstrating high prediction accuracy and good stability. Compared to finite simulation model, the Support Vetor predicition model has great advantage on computational efficiency, which can real-time feedback and predict temperature process of dam in the concreting process.