| This study is financial supported by the National Outstanding Youth Fund of China (No.50825902), National Science&Technology Pillar Program of China (No.2011BAF14B04) and Jiangsu Provincial Project for Innovative Postgraduate (No. CXZZ12-0679) of China. Because of increasing customer demands and stricter environmental noise level restrictions, the hydraulic design methods of centrifugal pumps with low noise and vibration by optimizing the geometric parameters have attached increasingly attentions in recent years. In this paper, the theoretical analysis, experimental and numerical investigation were used to study the the mechanism of rotor-stator interaction and flow-induced noise characteristics of centrifugal pumps. The purpose is to establish a set of hydraulic design methods for centrifugal pump with high efficiency and low noise. The main research contents of this dissertation are below:1. The classification and the origin of centrifugal pump noise are systematically summarized in this paper through studying in depth the current research literatures. It shows that conventional criterias of centrifugal pump noise evaluation is unable to meet the pump’s development. Flow-induced noise is the key to process the evaluation and measurement of the centrifugal pump’s noise. The source of flow-induced noise inside the centrifugal pumps is from the unsteady hydro-forces, which are caused by rotor-stator interaction between the rotary impeller and fixed part, radial force on the impeller and unsteady flow phenomenon at off-design conditions.2. A test rig of the flow-induced noise based on passive four terminal network method of the centrifugal pump was built to collect sound signals in various operating conditions. The paper investigated the sound pressure level of the model pump with variable speed, flow rate and the diameter of impeller. Moreover, noise characteristics under backflow and cavitation condition were also studied. The sound pressure level of the flow-induced noise is higher in the case of small flow rates. The sound pressure level initially decreases when the flow rate is greater than0.6Qd, reaches the minimum between Qd and1.2Qd and then subsequently increases with increasing flow rate. Noise levels at the inlet and outlet of model pump rise with the rotational speed increasing, in addition, the slope of noise level is larger downstream of the pump. As the cavitation coefficient is reduced, the overall sound pressure level of flow-induced noise is gradually increased, and decreased after reaching a maximum. An optimum value as 15%of the impeller radius has been found between impeller and volute tongue in order to reach a minium sound level. Cutting the impeller diameter could significantly reduce the noise levels and improve the cavitation performance of the model pump when the gap is less than the optimum. For the frequency domain, the blade passing frequency (290Hz) and the frequency doubling are the main frequencies of the flow induced noise when pump run within hign efficiency area. Peaks at the shaft frequency (48.3Hz) and frequency doubling are also observed. Below0.6Qd, the amplitude of the low-frequency noise increases with decreasing flow. In the working conditions of small flow, the inlet exhibits severe instability and generates backflow, which further results in greater fluctuations around shaft frequency. Noise spectrum of high-frequency domain exhibite more energy when cavitation occurs, but the low frequency domain, especially blade passing frequency and its harmonic peaks, gradually decreases.3. A hybrid algorithm combination computational acoustic with computational fluid dynamics based on the Lighthill equation theory was adopted to calculate the sound field of an IS65-50-165type pump. Moreover, vibro-acoustic effect was also analyzed during the simulation. The result shows that scale adaptive simulation (SAS) method could not only provide more accurate information of noise source, but also avoid the high demand for computing resources. Blade passing frequency and multiple are the main frequencies of the pressure fluctuation and it is strongest near the tongue, which means that the interaction between impeller blades and volute tongue is the main source of flow pulsation. Simulation values follow the experimental ones in the trend, with maximum difference3.1%, which could meet the needs of engineering optimization. In detail, acoustic boundary element method has superiority in solving noise levels on blade passing frequency and its harmonic. Although it is more complex in model building for the acoustic finite element method, it could directly show the distribution of turbulent noise source. The result is more consistent with the actual noise after considering turbulent noise solving. When pumps run at off-design condition, calculation results just based on dipole source is unable to accurately reflect the sound field characteristics. The influence from vibro-acoustic interaction on peak value at blade passing frequency could be ignored.4. Taken four indicators such as noise level at BPF, efficiency, head and shaft power as criteria, the study used matrix method to optimize multi-target orthographic experiment and the optimum plan is selected according to the weight of each factor from simulation works. After comparing test results between optimized impeller and the original one, it found that the optimum model satisfied all the standards, which verified that the matrix method combined with numerical simulation in pump optimization is feasible. Moreover, PIV method was used to compare the inner flow field and focus on the unsteady flow near the tongue. The results show that tongue always plays an important role on flow variables. Velocity inside the flow field changes periodically due to the interaction between the impeller blades and the volute tongue. This interaction effect could also influence the flow structure at the impeller inlet. The key to design an impeller with high efficiency and low noise is to keep a reasonable gap between impeller and tongue and form better control of the flow within the impeller in order to weaken the wake pulsation at impeller trailing edge.5. In order to reduce the radiated noise coming from pump operation, a water cooling system called jetting device was designed to replace the traditional fan cooling systems in pump motors. The concentration of the sound pressure level of radiated noise and performance of pumps were measured on a closed-type testing rig for3different models:the original pump, pump with a jetting pipe at normally designed diameter (dy=6mm) and another one that is deliberately larger (dy=6mm). The flow fields with the entire passage of the three pump models were also calculated to investigate the effect of the jetting device flow. Results show that radiated noise caused by the fan is an important component in the noise source of the pump system. The radiated sound pressure level of the model pump was significantly reduced by8.3dB after integrating the jetting pipe. In the condition with dy=6mm, no significant changes were observed in the efficiency, head, and shaft power curves, cavitation performance also improved at small flow rate. High-pressure water injection can effectively control inlet recirculation with decreased recirculation vortex strength and recirculation onset critical flow rate in case of with a normally designed jetting pipe. The performance in terms of the efficiency, head, shaft power curves, and cavitation of the pump with the larger jetting pipe worsened. However, the hump phenomenon of the model pump under a small flow rate was enhanced because of increased impeller flow.6. More precise internal flow analysis has been made by means of measurements on a centrifugal pump model in Laboratoire de Mecanique de Lille of France, including three-hole probe, PIV, unsteady static pressure on diffuser blades and flow-induced noise at the inlet of impeller. The internal flow characteristics and flow-induced noise were investigated under five flow rates to understand the influence of rotor-stator interaction. The results show that noticeable fluctuations in radial velocity component were observed, which means pump system act as a hydroacoustic source to the spread of noise. More obvious fluctuations of the tangential velocity component indicates that pressure fluctuations on the blades will result in dipole sources of hydroacoustic noise, which would be the main type of flow-induced noise in this pump system. The vaned diffuser could be divided into three parts based on pressure losses, such as vaneless part, region between diffuser inlet and throat, and region after the throat. The best pressure recovery factor is obtained at smallest flow rate. Around the area between impeller exit and diffuser throat, minimum pressure loss appeared near the design point of the diffuser. A large fluctuating separated region exists at the leading edge of the diffuser on the suction side at small flow rates, turbulence cores probably because of the rotor-stator interaction, originate from the impeller outlet, forced into the diffuser passage, and cause backflow at diffuser inlet section, which are the main reasons of loss. Backflow would spread from the impeller to the inlet pipe and form stall cores with the continued descent of flowrate, which would cause low frequency turbulence noise. Loss always originates from the flow separation at the pressure side after the diffuser throat at large flow rates, however, turbulence noise caused by these regions could not propagate upstream easily because of the impeller structure. |
PDF Full Text Request