Research on the Stability Problem of Hydroelectric Station Penstock under External Pressure Based on Neural Network

摘要

The stability of steel penstock under external pressure is a main factor of penstock design in hydroelectric station. Since the assumption conditions and the boundary conditions are simplified in the currently existing methods, the obtained computing results have low accuracy level and every method has its limitation. In this present paper, we apply artificial neural network model and improved BP algorithm to this problem with more complicated conditions and factors. By analyzing the influences on losing stability and breach for reinforcing ring steel penstock caused by diameter-to-thickness ratio of the steel tube, size and space of reinforce rings, a new research approach is obtained and test results show this method is feasible and its validity is verified. This paper provides an efficient solving path for stability problem of reinforcing ring steel penstock under external pressure. Furthermore, we use this method to Chinese Yachihe hydroelectric power station design, the computing results show this method is reliable, economic and fully satisfies the project requirements.