Adaptive and Optimal Control of the Water Gate of a Dam Using a Finite Element Method

From MaRDI portal
Publication:4360892





The authors develop a methodology for flood control based on adaptive and optimal control of the water gate of a dam, using the shallow water equation which is discretized by the finite element method. The determination of the water volume flowing through the dam gate is usually heavily dependent on experience, based on some information, for example about the water elevation and discharge. This paper presents sequel control methods which are independent of such experience. Though there are many numerical control methods, dynamic programming, the conjugate gradient method, and the Sakawa-Shindo method are used herein because of computational efficiency. Stochastic optimal control is also introduced. The method of stochastic optimal control is dynamic programming which is one of the most straightforward methods in conventional optimal control theory. The dynamic programming for a stochastic system is applied to discrete and continuous systems. Random inflow discharge, the only case, is treated herein. For taking into consideration the constraints of the capacity of the water gate and the necessity to avoid overflow in the downstream area of the dam, the clipping-off conjugate gradient method is used. Two adaptive control methods are proposed for real time control. One is to choose the steady-state control. This method uses steady results obtained by means of the Riccati equation. The other is the predictive control method. As a numerical example, the Agigawa dam is analysed. Because the conjugate gradient method needs the calculation of the gradient of the performance function, it takes more calculation time to obtain the optimal control volume than the Sakawa-Shindo method. The results of optimal control show the reduction of water elevation in the reservoir. However, it is found difficult to apply the theory, because the inflow discharge must be known in advance of the whole control duration. The adaptive control methods proposed in this paper solve the difficulties of optimal control theory, though the quality of the control is a little inferior. It can reduce the water elevation by more than 6 times than that without control, and it only takes about 20\% of the computational time of the optimal control theory.











This page was built for publication: Adaptive and Optimal Control of the Water Gate of a Dam Using a Finite Element Method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4360892)