The unsteady flow inside a large centrifugal pump with stay vanes was analyzed in this study. The static performance and pressure fluctuations in the pump were numerically predicted and were compared with experimental data. Considering the relative positions of the impeller to the volute tongue and stay vanes, the static performance which was obtained using a full unsteady calculation was compared with traditional steady calculation results. A comparison of the results with the experimental data showed that the operation condition farther from the design condition resulted in larger differences between the steady simulation and experimental results, with errors beyond reasonable limits, while the performance curves obtained by the unsteady calculations were closer to the experimental data. A comparison of the pressure fluctuations at four monitoring points with the experimental data showed that the amplitudes at HVS1 and HVS2 are much larger than at HD1 and HD2. The main frequency for these four monitoring points, which agreed well with the experimental data, was the blade passing frequency. The relative obvious errors in pressure fluctuations for HD1and HD2 were due to the inlet flow rate variation of the simulation. Thus, unsteady numerical simulations can be used to predict the pressure fluctuations when designing a pump.

Rolling bearings are the most fault-prone parts in rotating machinery. In order to find faults in time and reduce losses, this paper presents an intelligent diagnosis method for rolling bearings. At present, the deep residual network (RESNET) is the most widely used convolutional neural network (CNN) and has become one of the hotspots in fault diagnosis. However, the fully connected layer of the deep residual network has the disadvantage of too many training parameters, which makes the model training and testing time longer. So, we proposed a new network structure which the global average pooling (GAP) technology replaces the fully connected layer part of the traditional RESNET. It effectively solves the problem of too many parameters of the traditional RESNET model, and uses data enhancement, dropout, and other deep learning training techniques to prevent the model from overfitting. Experiments show that the accuracy of fault diagnosis of the improved algorithm reaches 99.83%, training time has been shortened. Also, the whole process of rolling bearing fault detection does not need any manually extract features, and this “end-to-end” algorithm has good versatility and operability.

The rotating stall is an unstable flow phenomenon of pump turbines in pump mode, which is of increasing concern to scientists and engineers working on pump turbines. However, at present, various studies are carried out based on CFD (computational fluid dynamics) simulation, while directly measured data and experimental research on flow fields are seldom reported. By utilizing PIV (particle image velocimetry) measuring equipment, the flow field within the guide vane zone of a low specific speed pump turbine in pump mode was measured. By measuring and analyzing the transient flow field, the evolutionary process of the rotating stall within the guide vane passages was determined. We found that for all three tested guide vane openings, regardless of whether the positive slope appeared or not, a pre-stall operating point was found for each opening in the process of decreasing the flow rate. The analysis of the loss within the flow field indicated that the dissipation-induced loss increased greatly after the rotating stall appeared. The pump performance curves at the three guide vane openings showed an inflection at the pre-stall point. When the flow rate is larger than that of the pre-stall point, the head of the pump turbine dramatically increases as the flow rate decreases. However, when the flow rate is smaller than the pre-stall point, such increases noticeably slows down. The research results showed that whether the positive slope on the pump performance curve occurred or not, instability caused by the rotating stall should be of great concern.

The pressure fluctuations in both the rotating runner and the other fixed components in a model Francis turbine under various loads were experimentally measured by means of onboard measuring equipment in the runner and data storage device on the shaft in this study. Large pressure fluctuations were observed under both small guide vane opening and large guide vane opening conditions. Flow separation at the blade suction surface led to large pressure fluctuations for small guide vane openings, the unsteady flow around the inlet on the blade pressure side led to large pressure fluctuations for large openings. The pressure fluctuations correlation between the runner and other components of the turbine, mainly the draft tube, was analyzed in detail for both small guide vane opening (12 deg) and large guide vane opening (30 deg). The results show that the pressure fluctuations in the runner space increased by the superposition of draft tube vortex rope pressure fluctuations and runner inter blade vortices pressure fluctuations, resulting in much larger pressure fluctuations in the runner space than in other components.

Sediment abrasion may cause the removal of material and efficiency loss to hydraulic machinery (hydro-turbine, pump, etc.), which may worsen the internal flow and endanger the safe operation of the machines. As abrasion is caused by the impact of sediment on the solid surface, the movement of particles determines the wear intensity. To understand the wear mechanism of hydraulic machinery, the motion characteristics of the asymmetric solid–liquid two-phase flow in a sediment-laden flow around NACA0012 cascades were studied. The laser particle imaging velocimetry (PIV) method was used to measure the flow of quartz particles with median diameters d₅₀ of 82.7 μm, 65.9 μm, and 31.8 μm near the wall of cascades at an impact angle of 10° in a Venturi circuit. The flow characteristics and velocity slip between solid and liquid phases, as well as the effects of particle size and the Reynolds number on velocity slip, were analyzed. The results showed that: (1) the flow is asymmetrically distributed in front of the cascade’s leading edge at a 10° impact angle, and there is strong velocity slip between solid and liquid phases; (2) under the influence of particle inertia, the velocity of the solid phase is higher than that of the liquid phase in the deceleration stage, while the velocity of the solid phase is lower than that of liquid phase in the acceleration stage; (3) in the process of approaching the leading edge, the velocity difference between the solid and liquid phase increases by about 10% and the angle difference increases by about 8.8°.

An advanced comprehensive test system, designed for testing synergetic erosion due to cavitation and sand abrasion in hydraulic machinery, is presented in this paper. This system includes an integrated test rig, control platform, and state-of-the-art measurement etc. For the integrated test system, there are three test modes, Venturi-section water tunnel, rotating disc and rotating disc with jet nozzle. The maximum velocity is 45 m/s for Venturi-section water tunnel test mode, and 85 m/s for rotating disc test mode. The pressure range for those two test modes can be regulated within -0.09 MPa~0.6 MPa. The highest flow relative velocity is 120 m/s for rotating disc with jet nozzle test mode. All key parameters measured from the test rig, such as flow discharge, pressure, sand concentration, temperature etc, can be displayed online and processed in the control platform. This new test system provides researchers with the possibility to measure cavitation erosion, sand abrasion and the synergetic damage in hydraulic machinery. Further, flow visualization analysis, weight loss measurements and erosion outline measurements are available using the system.