抽水蓄能机组甩负荷试验时尾水锥管压力

抽水蓄能机组水力调节过渡过程计算控制核心要求之一是尾水锥管压力不超过设计值。设计值是根据水力过渡过程理论的一维数值模拟"计算值"加上一定"压力脉动修正"量和 "计算误差"后计算获得。长期以来，尾水锥管压力计算值与尾水锥管压力测量值之间仍存在一定偏差，导致采用实测数据对计算结果进行评判时不能得到合理的解释与评价。为解决调节保证计算与试验验证之间的分歧，该研究对调节保证计算时尾水锥管压力最小值含义进行了阐释，梳理了调节保证计算与试验中涉及尾水锥管压力的相关国内标准。在此基础上，以洪屏抽水蓄能电站调试阶段四台机组甩额定负荷时的实测尾水锥管压力为研究对象，首先分析了压力测点的测量条件，采用短时傅里叶变换进行频率特性分析，验证了测试数据的有效性；其次采用Savitzky-Golay滤波器分离出了表征一维数值模拟断面平均压力计算值的压力趋势与表征流动复杂性的压力脉动，针对压力脉动项研究了峰峰值与时长的关系并获得了压力脉动项最大峰峰值，验证了趋势项与一维数值模拟之间的一致性；随后采用压力脉动项最大峰峰值对数值模拟和实测压力趋势项极值进行修正；最后总结形成了尾水锥管压力调节保证设计值的修正流程。案例研究表明：采用截止频率为0.1～0.2倍转频的低通滤波器可以有效分离出与一维数值模拟一致的尾水锥管压力趋势项；采用3～4个旋转周期数对应的压力脉动数据进行分析可以获得压力脉动项的最大峰峰值；利用压力脉动最大峰峰值对一维数值模拟极值进行修正能够实现调节保证的有效设计与验证。该研究为抽水蓄能机组调节保证设计与验证提供了有效支撑。

Draft cone tube pressure of pumped-storage power unit in load rejection test

Abstract: One of the key control parameters in hydraulic transient process of pumped-storage power unit is draft cone tube pressure, which should be within the design value. The design value is calculated by one dimension mathematical model of hydraulic transient theory with considering on certain pressure pulsation correction and calculation error. For a long time, there is non-negligible deviation between the calculated value by one dimension numerical simulation and the measured value of draft cone tube from on-site load rejection test. Henceforth, once engineers use measured values to check calculation results, they can not get reasonable explanation and evaluation between simulation results and measured values. In order to settle the difference between regulation guarantee calculation value and measured value, minimum pressure of draft cone tube was clarified from multiple perspectives in this paper, and Chinese standards or codes relating to hydraulic transient calculation were reviewed. The standard research indicates that for pumped storage unit, the minimum draft cone tube pressure should not below -0.08MPa in any circumstances. Measured draft cone tube pressure data of the rated load rejection tests performed in Hongping pumped storage power station during commissioning period was investigated. Firstly, field measuring conditions of draft cone tube pressure was thoroughly assessed and confirmed, and the validity of measured pressure was verified by using short-time-Fourier-transform to analyse the frequency characteristics. The frequencies of four draft cone tube pressure are similar, of which are low frequency component and rotor-stator frequency. The main frequencies are lower than first order eigen frequency of measuring pipes. The obtained data can be regarded as near wall measured pressure. Secondly, Savitzky-Golay filter was employed to separate trend and pulsation of measured draft cone tube pressure. The separated pressure trend can be deemed as the average section pressure of draft cone tube in one dimension numerical simulation, and the pressure pulsation represents the fluid complexity in the hydraulic transients. For pressure pulsations, relationship between maximum peak-to-peak values and time lengths were studied. For Hongping pumped-storage units, the maximum peak-to-peak values and time lengths of each unit are 0.489 MPa/0.46 s, 0.477 MPa/0.09 s, 0.532 MPa/0.66 s and 0.486 MPa/0.11 s, the average is 0.496 MPa/0.33 s. Consistency between pressure trends and one-dimension simulation results was verified. The verfication indicates that the guide vane closing principles are the same and pressure trends comply with the simulation results well. Thirdly, maximum peak-to-peak value of pressure pulsation was used to correct the one-dimension simulation results and trend items. The correction of draft cone tube has negligible errors with on-site measured values. Finally, the correction process was summarized for the hydraulic transient calculation of draft cone tube. The case study indicates that the pressure trend, which is consistent with one-dimension simulation result, can be effectively separated from measured draft cone tube pressure by using low-pass filter with cutoff frequency of 0.1-0.2 times rated rotational frequency; the maximum peak-to-peak value of pressure pulsation can be obtained by selecting data of 3-4 rotational periods; hydraulic transient calculation of draft cone tube can be effectively corrected and verified by superposing the maximum peak-to-peak value of measured pressure pulsation. The research provides effective support for regulation guarantee design and verification of pumped-storage power units.