| Mixed-flow pump is widely used in various pumping station projects as a kind of pump device with large flow and head variation range,wide high efficiency area and stable operation.However,in the actual work of the pump device,various off-design conditions often occur,such as rotating stall,rim leakage,internal cavitation,and off-design operation,which may lead to abnormal water pressure pulsation of the unit,aggravate the vibration of the unit and ultimately affect the safe and stable operation of the pumping station.Considering that the analysis of water pressure pulsation in the field of pumping station engineering mainly focuses on the time domain signal and the frequency domain signal after FFT transformation,in order to obtain the operation information of pump device reflected by more pulsation signals,based on computational fluid dynamics and chaos theory,this paper analyzes and studies the pressure pulsation and its chaotic characteristics under different blade tip clearance and different water level combination operation conditions,and realizes the judgment of unit operation state through support vector machine algorithm combined with chaos theory.The main research contents of this paper include:(1)Hydrodynamic and pressure pulsation characteristics of the unit under different tip clearances.The hydraulic characteristics,pressure pulsation characteristics and pressure pulsation chaos characteristics of the unit under different tip clearances are analyzed and studied.The results show that with the increase of the tip clearance,the head of the pump device in the full flow range decreases,and the larger the flow rate,the smaller the head difference caused by the tip clearance.The efficiency of the pump device also decreases with the increase of the tip clearance.The leakage vortex is generated in the nearwall region of the rim on the suction side of the blade.The strength and distribution range increase with the increase of the gap,but under the same gap condition,the range and strength decrease with the increase of the flow rate.In the time domain,the periodicity of the signal under different gaps is strong,and there are four peaks and troughs in a cycle.The overall time-domain change trend of the monitoring point signals in the same area is similar,only the amplitude changes;in the frequency domain,there are complex low-frequency components in the pulsation signals of the impeller inlet section,the leading edge and middle edge of the blade tip,and the impeller outlet section near the upper part of the guide vane.The energy distribution of the monitoring points P11 and P12 in the trailing edge of the blade tip and the monitoring points P13,P14 and P15 in the impeller outlet section near the middle and lower part of the guide vane is relatively clear and concentrated,and the main frequency is mainly 2 times and 4 times the blade frequency.Based on the quantitative and qualitative analysis of the relevant parameters of chaos theory,the pressure pulsation signals corresponding to the monitoring points 13 and 14 located at the outlet section of the impeller near the bottom of the guide vane show strong chaotic dynamic characteristics under different tip clearances,which can be used as the data basis for the subsequent operation state prediction of the unit.(2)Hydrodynamic and pressure fluctuation characteristics of the unit under different water level combination operation conditions.The hydraulic characteristics,pressure pulsation characteristics and pressure pulsation chaos characteristics of the unit under different water level combination operation conditions are analyzed and studied.The results show that the overall streamline of each working condition is relatively smooth,but there is an obvious recirculation zone near the left wall of the left flow channel inlet separated by the diversion pier.With the increase of head,the range and strength of the recirculation zone gradually decrease.The inlet velocity of the left flow channel is greatly affected by the operating conditions.There is a wide range of low-speed recirculation zones(concentrated in the surface and bottom layers),and the velocity uniformity is only about 0.5.Under irrigation conditions,the y-direction flow velocity inside the flow channel is the largest;under both drainage conditions,there is a low-speed zone from the inlet to the middle of the diversion pier,and the flow velocity in the channel of drainage condition 2 is the smallest.The data of pressure pulsation measuring points at different positions in the inlet duct are basically consistent.From the time domain point of view,the pulsation signals under the three operating conditions have four peaks and troughs in a cycle,and the amplitude of the monitoring point signal near the left wall of the inlet section of the flow channel is ’small in the middle layer and large in the bottom layer ’.From the perspective of frequency domain,the main frequency of pressure pulsation under irrigation conditions is 0.25 Hz,the secondary frequency is 5 Hz(4 times leaf frequency),and the amplitude difference between the primary and secondary frequencies is small.The fluctuating energy under drainage condition 1 is unevenly distributed in the low frequency range of 0-5Hz,and the primary and secondary frequencies correspond to 0.4 times leaf frequency and 0.2 times leaf frequency respectively.The distribution of pulsating energy in the flow channel under drainage condition 2 is relatively concentrated.The primary and secondary frequencies are the same as those under drainage condition 1,but the corresponding amplitudes are slightly different.Based on the quantitative analysis of the relevant parameters of the chaos theory,combined with the qualitative analysis of the three-dimensional phase trajectory map,it is only considered that the pulsation signal under the irrigation condition shows weak chaotic characteristics,and the other two drainage conditions do not have chaotic characteristics.(3)Operating condition prediction of pump device based on support vector machine algorithm.The phase space reconstruction technology based on chaos theory restores the pulsation signal to the coordinate point of the high-dimensional space,and uses the high-dimensional coordinate data to represent the influencing factors of the pump device state.Based on the training of SVC algorithm,the nonlinear mapping relationship between influencing factors and unit state is clarified,and RBF is selected as the kernel function of the model.Using MATLAB software to write and design a set of visual operation of pressure pulsation signal chaotic characteristics identification and tip clearance state classification interface based on support vector machine algorithm.The results show that for the signal data of 13 monitoring points under different tip clearances,the regular term C is 23.01.For the signal data of 14 monitoring points,the regular term C is 5.01,and the gamma value is 0.1,which can achieve the best prediction effect.By giving a total of 528 signal data at 13 different tip clearances at the monitoring points,the number of correct tip clearance values predicted by the model in this paper is 525,and the accuracy rate is 99.43%.By giving a total of 547 signal data at 14 different tip clearances at the monitoring points,the accuracy rate is 100%. |
PDF Full Text Request