| 摘要: |
| 以空气-水为介质对槽式孔板进行了湿气计量特性试验研究，提出了一种新的基于神经网络的槽式孔板湿气 计量修正模型。模型以Lockhart-MartineUi参数X、气体弗劳德准数Fr8、密度比札、孔径比;8四个无量纲参数作为模 型的输人，“虚高”Ofl作为输出。结果表明,在表压为0.25 ~0. 35 MPa,；?为0.02?0.6，化为0_ 5 ~2.7，谷为0.5 ~ 0. 75的测试范围内,模型能够很好地预测实际“虚高",用新修正模型对由于液相存在而引人的气相流量误差进行修 正后，气相流量相对误差在95%的置信度下小于±4%,明显优于其他槽式孔板湿气计量修正模型，可以满足生产计 量的精度要求。 |
| 关键词: 神经网络 湿气 槽式孔板 计量修正模型 虚髙 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金项目(60672003);山东省自然科学基金项目(Y2006F49) |
|
| Wet gas flow metering correction model of slotted orifice based on neural network |
|
HUA Chen-quan1,2,WANG Chang-ming1,GENG Yan-feng2,WANG Xiao-zhi2
|
|
(1. College of Mechanical Engineering in Nanjing University of Science and Technology，Nanjing 210094, China ;2. College of Information and Control Engineering in China University of Petroleum, Dongying 2570611 China)
|
| Abstract: |
| The wet gas metering characteristics of slotted orifice meter were discussed by using air-water as media. A novel wet gas metering correction model based on BP neural network was proposed. In the model, Lockhart-Martinelli parameterX, gas Froude number, the gas to liquid density ratio Rd，and bore diameter ratio are the inputs, and over-reading is the output. The results show that the new correction model can predict the over-reading accurately, and correct the liquid-induced gas flow rate prediction error of the wet gas flow to 土 4% at 95% confidence level under the conditions of pressure from 0. 25 MPa to 0. 35 MPa, X from 0. 02 to 0. 6, gas Froude number from 0. 5 to 2. 7, and /3 from 0. 5 to 0. 75. The model is superior to other correction model and can satisfy the accuracy requirement of production metering. |
| Key words: neural network wet gas slotted orifice metering correction model over-reading |
