Research On The Generalization Model Of Typical Stream Segment On The Yellow River And Its Application

The research of the Yellow River generalized model is based on the high-developed system of the Yellow River and the practical demand of the unified allocation of water resources. As increasing artificial involvement in the water allotment system of the Yellow River, there are large amounts of issues to be lucubrated. This paper's concrete work can be summarized into following three parts.1)Research on the generalized model of the typical stream section on the Yellow River. The basic theory and development of the generalized theory are introduced. Following the predecessor's footprint, the author improves and eventually proposes the applicable generalized means to the drainage area of the Yellow River. This paper determines the feasible trunk stream generalization diagram of the Yellow River, depending on the prolific information which contains the arrangement plan of hydropower station hydrographical network, water deployment data and water allocation data in the latest decades. Then generalization model capable of abstracting the main cross-section in the natural watercourse is presented, which covers two styles of modes that are single channel mode and channel-beach banding mode. Thus, two particular generalized models are presented to pave the road to the following issue.2) Research on the flow propagation on the main stream of the Yellow River. By analyzing lots of relevant data, this paper fixes on the special period when happens"artificial plastic flood", uses the methods of the Computational Hydrodynamics, which is typical Saint-Venant format and expansion Saint-Venant format, and adopts the numerical modeling with the implicit difference scheme, eventually it is successfully used to simulate the flow propagation process at the right section. The simulation results show that the constructed models are feasibility and practicability.3) Research on the propagation time forecasting of the given section. The paper introduces the trend of development in the propagation time forecasting, and represents its principal forecast methods. In this paper, the feed-forward neural network technology with improved LM algorithm and with the fault tolerance algorithm is adopted, and two forecasting models of the propagation time are presented that exploits the existing real-time information available for the adjacent section to predict the propagation time of the different flood wave situation. By the application to the specific examples, the feasibility and validity of the constructed models is validated. The verification results showed that the neural network model is suitable for propagation time forecasting purposes. Furthermore, the results exhibits that the fault tolerance neural network model is better than the other, as it gives the improvement to remedy the shortage of low precision in break points.