CAJ | 학술논문

为了提高IS50-32-160低比转数离心泵在设计工况下的扬程和效率，采用数值模拟、试验设计、近似模型和遗传算法相结合的优化方法，选取了泵叶轮的叶片出口宽度、叶片出口安放角和叶片包角3个参数作为设计变量，采用最优拉丁超立方试验设计方法进行20组方案设计，应用ANSYS CFX 14.5软件对各方案进行定常数值计算，得到设计工况下的效率和扬程，并将效率和扬程作为设计目标，根据Kriging近似模型建立了设计目标与设计变量之间的近似函数，采用遗传算法对近似函数进行求解，得到最优的叶轮参数组合。研究结果表明：原始方案的外特性数值模拟结果与试验结果吻合程度较好，设计工况下预测扬程偏差为3.3%；优化后的泵水力效率提高了4.18%，而且近似模型在预测性能的准确性高；通过对比原始方案和优化方案的内流场特性，优化方案内部流动得到改善，优化的叶轮的漩涡区域比原始方案的较小；优化使得效率在主频和次频下的脉动幅值分别下降了1.52和0.84，叶轮内的较大压力脉动强度区域减小，隔舌附近监测点在主频下的压力脉动系数幅值下降了0.02。非定常压力脉动强度降低，从而泵的运行稳定性提高。提出的优化设计方法对低比转数离心泵高效以及无过载特性的优化具有一定的参考意义。
위료제고IS50-32-160저비전수리심빙재설계공황하적양정화효솔，채용수치모의、시험설계、근사모형화유전산법상결합적우화방법，선취료빙협륜적협편출구관도、협편출구안방각화협편포각3개삼수작위설계변량，채용최우랍정초립방시험설계방법진행20조방안설계，응용ANSYS CFX 14.5연건대각방안진행정상수치계산，득도설계공황하적효솔화양정，병장효솔화양정작위설계목표，근거Kriging근사모형건립료설계목표여설계변량지간적근사함수，채용유전산법대근사함수진행구해，득도최우적협륜삼수조합。연구결과표명：원시방안적외특성수치모의결과여시험결과문합정도교호，설계공황하예측양정편차위3.3%；우화후적빙수력효솔제고료4.18%，이차근사모형재예측성능적준학성고；통과대비원시방안화우화방안적내류장특성，우화방안내부류동득도개선，우화적협륜적선와구역비원시방안적교소；우화사득효솔재주빈화차빈하적맥동폭치분별하강료1.52화0.84，협륜내적교대압력맥동강도구역감소，격설부근감측점재주빈하적압력맥동계수폭치하강료0.02。비정상압력맥동강도강저，종이빙적운행은정성제고。제출적우화설계방법대저비전수리심빙고효이급무과재특성적우화구유일정적삼고의의。
In order to improve the efficiency and head of a centrifugal pump with low-specifc-speed under design flow rate, an optimization approach of performance of the low-specific-speed centrifugal pump of IS50-32-160 was proposed by combining numerical simulation, design of experiment, approximation model and genetic algorithm. Three geometrical parameters containing blade outlet width, blade outlet angle, and blade warp angle were chosen as the design variables, and then 20 impellers were designed by Optimal Latin Hypercube Sampling method (OLHS). Commercial software ANSYS CFX 14.5 was used to conduct the steady numerical simulation to calculate the head and efficiency under design condition, which were chosen as the optimal objectives. Approximation model was built by using the Kriging model between the objectives and design variables, and had good prediction accuracy with R-square values of 0.9513 for efficiency and 0.9294 for head. Finally, the best combination of impeller parameters was obtained by solving the approximation model with genetic algorithm. To demonstrate the improvement of performance, the velocity distribution obtained by steady simulation and the pressure fluctuation intensity distributions calculated by unsteady simulation were compared. The results showed that the performance curves obtained by experiment and numerical simulation had a good agreement and the head deviation under design flow rate was 3.3%. The optimization improved the hydraulic efficiency by 3.2%, while, the head did not improve, but it still met the requirement of design. Compared with the parameters of original impeller, the blade outlet width of optimized impeller was smaller, the blade wrap angle was larger and the blade outlet angle did not change. The optimized impeller was redesigned with the optimal parameters, and the performance was calculated. It can be found that the predicted head and efficiency deviations were 3.3% and 0.056%, respectively. The internal velocity flow characteristics in the optimal impeller were improved and the area of the vortexes in the optimized impeller was smaller. The main frequency of fluctuation of efficiency is 145 Hz determined by blade passage frequency. The fluctuation values of efficiency of original and optimized pump under main frequency were 5.13 and 3.61, respectively. The pressure fluctuation intensity was the lowest at the leading edge of impeller and increase gradually along with the passage. The biggest pressure fluctuation intensity was distributed at the pressure side near to the trailing edge. The area of largest pressure fluctuation intensity was smaller in optimized impeller than in original one. The pressure fluctuation intensity was large from the tongue to the second section of volute, while the pressure fluctuation intensity was small at the seventh and eighth sections of volute. The pressure fluctuation intensity in the tongue reduced in the optimized pump. The pressure fluctuation of the monitor set in the tongue is periodic and there are 6 peaks in one period. The pressure fluctuation coefficient of the monitor in the optimized pump is 0.01, lower than that in the original one. The optimization decreased the fluctuation caused by impeller-volute interaction, thus enhancing the operation reliability of centrifugal pump. The optimization method presented can provide references to the optimization of high efficiency and non-over-load design of low-specific-speed centrifugal pump.