基于环境识别记忆的离心泵多工况水力设计研究

项目名称： 基于环境识别记忆的离心泵多工况水力设计研究

项目编号： No.51209105

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 水利科学与海洋工程学科

项目作者： 王凯

作者单位： 江苏大学

项目金额： 25万元

中文摘要： 多工况水力设计已成为离心泵研究的热点之一。本项目拟对离心泵多工况水力设计中的关键科学问题展开研究，以提高其多个工况的运行效率。 本项目拟采用三维PIV技术研究离心泵全工况下的内部流动特征，采用小波神经网络提出三维PIV粒子图像去噪方法，并结合互相关算法和快速傅里叶变换算法研究粒子图像的精确处理算法。在分析叶轮出口处滑移速度和液流角偏移量的基础上，探索离心泵全工况滑移系数计算模型。基于泵内能量损失的全工况测量建立离心泵多工况性能计算模型。结合EIMS和GASA算法发展求解离心泵多工况性能计算模型的GASA-EIMS算法，研究环境识别因子对求解精度和收敛速度的影响。基于上述研究，提出一种离心泵多工况水力设计方法，并通过试验验证。 本项目旨在建立具有环境识别记忆的离心泵多工况水力设计方法。研究成果对促进我国节能减排目标的实现具有重要意义，同时也为其它叶片式流体机械的多工况水力设计提供借鉴。

中文关键词： 离心泵；多工况设计；图像处理；滑移系数；GASA-ELMS算法

英文摘要： Multi-conditions hydraulic design for centrifugal pumps has become one of the research hot spots.In order to improve performance under different run conditions, the key scientific problems on multi-conditions hydraulic design for centrifugal pumps will be researched. This project will study the internal flow characteristics of centrifugal pumps under full conditions with three dimensional PIV technology. Based on the wavelet neural network, the denoise method for three dimensional PIV images will be established. Accurate processing algorithm of particle images will be researched with the application of cross-correlation algorithm and fast Fourier transform algorithm. Calculation models of slip factor for centrifugal pumps under full conditions will be explored on the basis of analyzing the slip velocities and deviations of flow angle at the outlet of impeller. Multi-conditions energy performance calculating models of centrifugal pumps will be established based on measurement of energy losses under whole conditions. GASA-EIMS algorithm, which is suitable for solving multi-conditions energy performance calculating models of centrifugal pumps, will be developed with EIMS and GASA algorithm. The influence of environment identify factor on solution precision and convergence rate will be studied.On the basis of the ab

英文关键词： centrifugal pump；multi-conditions design；image preprocessing；slip factor；GASA-EIMS algorithm