Prediction of Cavitation Characteristics of Throttle Orifice Plate Based on Improved BP Neural Network
-
摘要: 为高效地获取核级管道中节流孔板附近的空化特性，构建了可靠的改进反向传播（BP）神经网络预测模型。首先提取了节流孔板的几何特征参数，并使用拉丁超立方抽样（LHS）方法生成了上述几何特征参数的样本库；然后通过计算流体力学（CFD）方法得到了各个样本对应的最小空化数，以该无量纲参数作为输出响应；最后针对原始BP神经网络预测模型的不足，结合遗传算法建立了节流孔板空化特性的改进预测模型。结果表明，孔板开孔直径和前开角度对最小空化数具有较强的全局敏感度；通过遗传算法优化后的BP神经网络预测模型的预测精度得到了大幅提升，误差均方根降低约36.4%。
-
关键词:
- 节流孔板 /
- 空化 /
- 计算流体力学（CFD） /
- 反向传播（BP）神经网络 /
- 遗传算法
Abstract: In order to obtain the cavitation characteristics near throttle orifice plate in nuclear pipeline efficiently, a reliable improved backpropagation (BP) neural network prediction model was constructed. Firstly, the geometric feature parameters of the throttle orifice plate were extracted, and the sample data base of these parameters was generated by using the Latin Hypercube Sampling approach. Then, the minimum cavitation number corresponding to each sample is obtained by computational fluid dynamics (CFD) method, and the dimensionless parameter is used as the output response. Finally, in view of the deficiency of the original BP neural network prediction model, an improved prediction model of throttle orifice plate cavitation characteristics is established by combining genetic algorithm. The results show that the orifice diameter and front opening angle have strong global sensitivity to the minimum cavitation number; the prediction accuracy of the BP neural network prediction model optimized by genetic algorithm has been greatly improved, and the root mean square error is reduced by about 36.4%. -
图 2 BP神经网络拓扑结构
m—输入层神经元数，即输入参数的个数；n—隐含层神经元数；q—输出层神经元数；${w_{2,n-3,1}}$—从第2层的第n个神经元到第3层的第1个神经元的权值；x1,…,xm—输入层输入量的分量；h1,… ,hn—隐含层输出量的分量；o1,… ,oq—输出层输出量的分量
Figure 2. Topology Structure of BP Neural Network
图 4 节流孔板结构
Df—孔直径；ts—孔板厚度；tf—孔板平直段厚度；$\alpha $和$\beta $—孔板前、后开角
Figure 4. Structure of Throttle Orifice Plate
表 1 参数信息
Table 1. Parameters Information
参数名 参数值范围 均值 初始值 Dp/mm 180.3 180.3 180.3 ts/mm 10 10 10 ${\xi _1}$ 0.1~0.7 0.4 0.407 ${\xi _2}$ 0.2~0.8 0.5 0.5 ${\xi _3}$ 0.1~0.4 0.25 0 ${\xi _4}$ 0.1~0.4 0.25 0 
下载: 导出CSV
表 2 误差均方根对比
Table 2. Comparison of Error Root Mean Square
对比量 未改进预测模型 改进预测模型 相对偏差/% erms 0.33 0.21 36.4
下载: 导出CSV
-
[1] TANG T F, GAO L L, LI B R, et al. Cavitation optimization of the throttle orifice plate based on three-dimensional genetic algorithm and topology optimization[J]. Structural and Multidisciplinary Optimization, 2019(60): 1227-1244. [2] 袁少波,李振,徐伟祖,等. 核电厂小支管裂纹泄漏原因分析和改造[J]. 核动力工程,2019, 40(S1): 97-99. [3] FILHO J A B, SANTOS A A C, NAVARRO M A, et al. Effect of chamfer geometry on the pressure drop of perforated plates with thin orifices[J]. Nuclear Engineering and Design, 2015(284): 74-79. [4] MENTER F R. Two-equation eddy-viscosity turbulence models for engineering applications[J]. AIAA Journal, 1994, 32(8): 1598-1605. doi: 10.2514/3.12149 [5] 林名润,王杰,闫大鹏,等. 改进BP神经网络的滚珠丝杆故障诊断研究[J]. 机械设计与制造,2020(6): 173-176. doi: 10.3969/j.issn.1001-3997.2020.06.042 [6] MOOSA M, HEKMAT M H. Numerical investigation of turbulence characteristics and upstream disturbance of flow through standard and multi-hole orifice flowmeters[J]. Flow Measurement and Instrumentation, 2019(65): 203-218. -
计量
- 文章访问数: 262
- HTML全文浏览量: 102
- PDF下载量: 35
- 被引次数: 0
