Zhang Qinzhao, Liang Kaihong and Cao Shuliang of the State Key Laboratory of Water and Sediment Science and Hydraulic and Hydropower Engineering, Department of Thermal Power Engineering, Tsinghua University think that it is very important to study the hydraulic design of high-speed mixed-flow impeller impellers and to develop the hydraulic model of high specific speed mixed-flow pumps , And introduced a series of their research results. Figure 1. Axial velocity distribution Francis pump is a pump with a higher specific speed and is widely used in the fields of farmland irrigation, waterlogging and flood control, sewage treatment and power station cooling system. In the large flow, low lift applications, axial flow pump head, the flow range of small, high-efficiency narrow area, poor resistance to cavitation, the use of mixed flow pump instead of axial flow pump, axial flow pump to play on the basis of the advantages of compensation These shortcomings, so as to achieve better results. Design Method of High Specific Speed ​​Mixed Flow Pump Axial flow calculation uses the streamline iterative method to solve the axial flow. The streamline iteration method is based on the axial velocity gradient equation along any quasi-orthogonal line and the continuous equation of fluid motion. According to the principle of equal flow of each sub-flow channel, iteratively solving the axial velocity gradient equation along the quasi-orthogonal line can be obtained Axial flow network. By streamline iteration, after determining the profile of the axial flowpath, an accurate axial flow network can be obtained by a very short iterative calculation process. Figure 1 is the distribution of axial flow velocity along the axial flow corresponding to flow network. Figure 2. Confirmation of overcurrent cross-section formation line Over-current cross-sectional inspection. By changing the traditional method of inscribed circle to approximate the area of ​​overcurrent cross section, using the characteristics of streamline iteration method, another convenient method is used to calculate Overcurrent cross-sectional area. In the axial flow calculation, the axial flow network formed by the quasi-orthogonal lines and the axial flow lines has been determined, and the slope of the axial flow lines at each grid point has been calculated. Based on these initial data, the axial flow The slope at any point on the line can be calculated by the interpolation method. Taking into account the perpendicularity of the flow line formed by the overcurrent section and the axial surface, the slope of the overcurrent section line passing through this point can be determined. As the initial cross-section of the formation of cross-section line, and adjust the formation of line-axis and axial flow intersection point until it meets the vertical relationship, which can determine the flow cross-sectional shape of the line shape and location, as shown in Figure 2. Then use the numerical integration can accurately calculate the cross-sectional area of ​​the flow. Overcurrent cross-sectional area of ​​the change rule shown in Figure 4. This improved method has the advantages of high area calculation accuracy and greatly reducing the workload of modification. Leaf blade profile Figure 3 Blade axial projection The design of the impeller uses the leaf blade differential equation: where Cn is the circumferential component of the absolute velocity of the particle ω is the axial velocity of the impeller ω, the axial velocity Cm corresponds to The radius r is known and the angular velocity ω is given. Therefore, as long as the distribution law of the streamline Cnr along the axial plane is given, the wrap angle of the blade can be obtained by using the integral formula (1), and the blade axial cross-section can be obtained. Due to the binary theory of Ωu = 0, the velocity moment along the same axis is equal to the constant, therefore only the law of velocity moment distribution on the streamline of any axis can be given. Generally, can. The guide vane is designed using the differential equations of the blade bones: it is necessary to obtain the angle of the blade by the integral formula (2) after the distribution angle of the blade placement angle α along all the axial lines, and then the blade axial intercept is obtained. Impeller blades and vanes of the axial cross-section of the blade shown in Figure 4. Figure 4. Overcurrent cross-sectional area change rule Paddle thickening and rounding Thickening and rounding are carried out in the conformal transformation plane. The basic method is as follows: transform the space streamline conformal flow in the flow surface to the cylinder unfolding surface, and thicken the blade bone line in the unfolding surface of the cylinder to obtain the unfolding line of flow lines on the back of the blade, Angle take point and converted to the axial projection map, the smooth connection of the same name points for the back of the blade axial line, as shown in dashed line in Figure 3; at the same time in the cylindrical surface of the blade head and tail rounded, and then Cylindrical expansion surface of the round head of each coordinate point converted to the flow point on the surface space, so as to complete the blade head and tail rounding work. After rounding the three-dimensional shape of the blade head shown in Figure 5. The optimization of the impeller takes the maximum efficiency of the impeller as the objective function, and uses the genetic algorithm to search the law of the optimal speed moment distribution to obtain the impeller with good performance in theory. Genetic algorithm optimization process is simple and fast, and the search solution is generally near the optimal point or the optimal point, it changes the traditional manual modification and design process, improve the design efficiency. Woodworking drawing and 3D modeling