| The hydraulic turbines can recycle the pressure energy from high-pressure fluid into mechanical energy of the turbine rotor, which can be further used to drive the working machines such as a pump, a fan and so on, or to generate into electricity. Pump as turbine(for short:PAT)is an important form for hydraulic turbine, including single-stage and multi-stage. One of the hot topics for the research of hydraulic turbine is how to choose a right pump to work as a hydraulic turbine. The high-pressure fluid in the actual production process often contains a certain amount of gas, and when the fluid pressure is greater, it need to use a multi-stage pump to work as a hydraulic turbine to recover its energy. So in this paper artificial neural network (for short:ANN) is used to predict the performance of hydraulic turbine, and the theoretical analysis and numerical simulation are used to study the external characteristics and internal flow field of gas-liquid two-phase multi-stage hydraulic turbine.1. BP and GA-BP neural network are established to predict the head and efficiency of a centrifugal pump working as a hydraulic turbine. The parameters of63centrifugal pumps acting as turbine are used as sample for two neural network training. Select outlet diameter of impeller D2, wrap angle of blade φ, outlet width of blade b2, outlet angle of blade β2, the number of blade z and specific speed of pump ns as input layer of the network, head H and efficiency77of turbine as output layer. Six groups of data extra samples are used to test the two trained neural network. The analysis on the relevance and linear regression between predicted values and experimental values of the two networks is made. The results show that the average error of head and efficiency forecasted with BP network are5.33%and0.85%and the average error of head and efficiency predicted by GA-BP network are3.56%and0.46%. GA-BP network has great predictive accuracy, close correlation between predicted results and experimental results, and its time spent by prediction is1/3of that of BP network. GA-BP network is more suitable for performance prediction of a pump working as hydraulic turbine.2. In the hydraulic turbine, the fluid flows through the device for a short time. And generally the equipment is made thermal insulation, it can be considered the thermodynamic process in a thermodynamic system turbine is reversible and adiabatic and no heat exchange with the outside. Further, since the share of the gas medium is relatively small, the internal energy change of the media caused by by gas expansion and compression is small and negligible. Considering the work for the expansion of compressible ideal gas, the formula for the head of PAT with gas-liquid two-phase medium is further deduced based on energy equation of the PAT with only water medium.3. A DG85-80five segmental centrifugal pump is chosen to work as a turbine. In order to get each external characteristic curve of the PAT, the numerical simulation in the case of multi-stage turbine work with gas-liquid two-phase media with different GVF of0.05,0.10,0.15,0.20is made based on CFD software. The results show that the work of gas expansion has great impact on the energy characteristics of the hydraulic turbine. With the increasing of mass flow, the shaft power and head of multi-stage turbine with gas-liquid two-phase medium increase. The GVF at the inlet of the turbine is lower and the area of mass flow with high efficiency is relatively larger. As the GVF at the inlet of turbine increasing, the head and power of turbine at the optimal operating conditions increase and the hydraulic efficiency and mass flow decrease. The curve for conversion factor of pump with water and turbine with gas-liquid two-phase medium shows that as the GVF at the inlet of turbine increasing, the flow coefficient h and power coefficient p increase and efficiency coefficient λ value decrease. The q is between1.5-1.7and the X is between0.89and1.1.4. Analyse the internal flow of the multi-stage hydraulic turbine with gas-liquid two-phase. For Multi-stage turbine with gas-liquid two-phase medium, the pressure decrease progressively from the entrance of positive vanes to the exit of the impeller. The pressure on the convex of blade is greater than the pressure on the concave. The flow within the impeller and vanes is relatively stable, the speed on the concave of blade is greater than the speed on the convex, and there is a whirlpool area near the convex of the blade inlet. The GVF increases progressively from the entrance positive of vanes to the exit of the impeller. The GVF on the convex of vanes is smaller than the GVF on the concave of vanes, and the GVF on the concave of blade is larger than the GVF on the convex of blade. For the impeller and vanes of the same stage, as the GVF at the inlet of the turbine or the mass flow increasing the asymmetry and unevenness of the pressure in the impeller and vanes exacerbates and the pressure gradient becomes larger. The speed difference from the entrance of positive vanes to the exit of the impeller becomes larger, the speed gradient in the in channel of impeller and vane is large. The overall GVF within the impeller and vane increasing, the distribution asymmetry exacerbates in the impeller and vane of turbine. |
PDF Full Text Request