当前位置:主页 > 科技论文 > 机电工程论文 >

泵性能测试曲线分段最小二乘多项式拟合算法

发布时间:2018-01-11 18:14

  本文关键词:泵性能测试曲线分段最小二乘多项式拟合算法 出处:《排灌机械工程学报》2017年09期  论文类型:期刊论文


  更多相关文章: 泵性能测试 最小二乘拟合 马鞍形特性曲线 分段曲线拟合算法 插值拟合


【摘要】:针对现有水泵性能测试曲线拟合算法单一、对水泵非稳定性状态不规则、马鞍形等形状数据拟合精度差、近似效果不理想甚至丢失真实试验数据特性的现状,通过比较水泵性能曲线的各种拟合方法,分析了不同拟合方法在水泵特性曲线拟合中的优缺点,并针对特殊的水泵马鞍形性能曲线,设计了分段最小二乘多项式拟合方法.为了保证分段拟合曲线的连续可导性,在分段最小二乘拟合曲线的基础上,对各相邻分段曲线的分段点采用插值拟合的方法进行连接.最后通过试验验证,分别采用分段三阶最小二乘拟合算法和五阶最小二乘拟合算法对试验数据进行拟合,并对拟合结果进行对比分析.结果表明:采用分段最小二乘多项式的拟合方法对复杂特性曲线进行拟合,可以提高拟合曲线的拟合精度,并保证了拟合曲线的连续性.
[Abstract]:In view of the existing pump performance test curve fitting algorithm is single, the pump instability state is irregular, the saddle shape data fitting accuracy is poor, the approximate effect is not ideal and even lost the real test data characteristics. By comparing various fitting methods of pump performance curve, the advantages and disadvantages of different fitting methods in water pump characteristic curve fitting are analyzed, and the special pump saddle performance curve is analyzed. The piecewise least square polynomial fitting method is designed. In order to ensure the continuous derivation of the piecewise fitting curve, the piecewise least square fitting curve is based on it. The piecewise points of each adjacent piecewise curve are connected by the method of interpolation fitting. Finally, the test is carried out. The experimental data were fitted by piecewise third-order least square fitting algorithm and five-order least square fitting algorithm respectively. The results show that the fitting accuracy of the fitting curve can be improved by using the piecewise least square polynomial fitting method to fit the complex characteristic curve. The continuity of fitting curve is ensured.
【作者单位】: 江苏大学国家水泵及系统工程技术研究中心;
【基金】:江苏省“六大人才高峰”高层次人才项目(2015-JNHB-015)
【分类号】:TH38
【正文快照】: 汤跃 对水泵进行性能试验时,需要采用曲线拟合的方法对试验数据进行拟合,从而得到水泵的性能曲线,但不同的曲线拟合方法得到的性能曲线在曲线的精度及误差等方面各有差异.在泵性能曲线拟合方法中,插值拟合算法比较简单且拟合曲线与试验数据比较贴合,但是当曲线摆动较大,试验

本文编号:1410655

资料下载
论文发表

本文链接:https://www.wllwen.com/jixiegongchenglunwen/1410655.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户8ad7f***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com