Application of Improved BP Neural Network Model in Uplift Pressure Monitoring
The network structure, initial weights and initial thresholds were optimized to solve some problems, such as over-fitting and slow convergence rate in standard BP Neural Network. Combining the base seepage character of concrete dam, a uplift pressure monitoring model is established in this paper with measured data of a actual concrete dam . The advantage of the presented model is tested and validated by actual examples. It has positive significance in the actual application.
Y. T. Mei et al., "Application of Improved BP Neural Network Model in Uplift Pressure Monitoring", Advanced Materials Research, Vol. 304, pp. 24-30, 2011