Study On Inverse Problems Of Hydraulics In Water Diversion Projects

The water diversion project is the most effective and direct countermeasure to the supply-demand contradiction of water resource. With the science and technology's development, people are not satisfied with system forecast and hope that they can identify and control the system by all means, that is to solve hydraulic inverse problems. A water diversion system is considered to work well if it runs safely, reliably and flexibly. To meet these requests, the system must be identified effectively and be designed and controlled reasonably. This is the strong drive to study the inverse problems of hydraulics.In this paper, a long-distance water diversion project is taken as the study object. Based on theory of the hydraulics, system identification inverse problem is studied. The roughness in the free-pressure sect and the water-hammer velocity in the pressure sect of the project are identified, and pipeline leakage detection is studied. Some achievements of the thesis are listed below:1. The roughness coefficient in the free-pressure sect of the project is identified by optimization method. First, the numerical models of water surface free-pressure sect, with measuring end water level and flow as input parameters, and then the head water level is gained by model computation. When the calculating value is most close to the measuring value, the roughness value can be regarded as the roughness of the project. The precision can meet demand in theory.2. The water-hammer velocity in the pressure sect is identified. Based on the fluid transient theory, a numerical model of a sect in the project is established by FORTRAN, and the calculating result is used as character parameters of water-hammer identification. A three-layer LMBP neural network with one concealment level is set up, the water-hammer velocity can be gained by this method with the max water level and flow measured in this sect. Comparing the identifying value and academic value, the max comparative error is only 0.66%, which meets precision demand.3. The pipeline leakage is detected by the method of configuration mode identification based on fluid transient theory. First a transient numerical model of leakage in a sect of the project is established. Then the result is analyzed and used in the leakage detection by configuration mode identification method.All the research above can offer technical grounds for the design and management of the project, which is worth to be widely used.