基于Fluent和LSTM神经网络的超声波测风仪阴影效应补偿研究
|
摘要
超声波测风仪因其结构坚固,维修成本低等优点,在气象、生活及农业等领域有着广泛应用。但由于其结构特点造成的阴影效应,会导致其风速测量精度下降,是当前测风领域中不可忽视的问题。针对该问题,提出一种基于Fluent软件以及LSTM长短期记忆神经网络的超声波阴影效应的补偿算法,对不同风速风向以及不同温度下的阴影效应进行补偿。利用Fluent仿真得到样本数据完成LSTM预测模型训练;基于Fluent仿真数据对SVR和MLR等模型与LSTM模型对超声波测风仪阴影效应进行对比实验,验证LSTM算法模型的有效性及优越性;通过风洞数据对LSTM神经网络修正算法的可行性进一步验证。实验结果表明:该算法可对阴影效应所造成的误差进行有效补偿,其精确度得到显著提高,为减小超声波测风仪的阴影效应提供了一定的参考价值。
Because of its strong structure and low maintenance cost, ultrasonic anemometer has been widely used in meteorology, life and agriculture.However, due to the shadow effect caused by its structural characteristics, the accuracy of wind speed measurement will be reduced, which is a problem that cannot be ignored in the field of wind measurement.In order to solve this problem, this paper proposed a compensation algorithm based on Fluent software and LSTM long-term and short-term memory neural network to compensate the shadow effect under different wind speed, wind direction and temperature. The LSTM prediction model was trained by using the sample data obtained by fluent simulation. The shadow effects of the SVR and MLR models were compared with the LSTM model proposed in this paper. The validity and superiority of the LSTM algorithm model were verified. The feasibility of the LSTM neural network correction algorithm was further validated by wind tunnel data.The experimental results show that the proposed prediction model based on Fluent and LSTM neural network can effectively compensate the error caused by the shadow effect, and its accuracy is significantly improved, which provides a certain reference value for reducing the shadow effect of the ultrasonic anemometer.
Because of its strong structure and low maintenance cost, ultrasonic anemometer has been widely used in meteorology, life and agriculture.However, due to the shadow effect caused by its structural characteristics, the accuracy of wind speed measurement will be reduced, which is a problem that cannot be ignored in the field of wind measurement.In order to solve this problem, this paper proposed a compensation algorithm based on Fluent software and LSTM long-term and short-term memory neural network to compensate the shadow effect under different wind speed, wind direction and temperature. The LSTM prediction model was trained by using the sample data obtained by fluent simulation. The shadow effects of the SVR and MLR models were compared with the LSTM model proposed in this paper. The validity and superiority of the LSTM algorithm model were verified. The feasibility of the LSTM neural network correction algorithm was further validated by wind tunnel data.The experimental results show that the proposed prediction model based on Fluent and LSTM neural network can effectively compensate the error caused by the shadow effect, and its accuracy is significantly improved, which provides a certain reference value for reducing the shadow effect of the ultrasonic anemometer.
引文
[1] Ishida H,Yoshikawa K,Moriizumi T.Three-dimensional gas-plume tracking using gas sensors and ultrasonic anemometer[C]//Sensors,2004.Proceedings of IEEE.IEEE,2004.
[2] Olmos P.Ultrasonic velocity meter to evaluate the behaviour of a solar chimney[J].Measurement Science & Technology,2004,15(7):N49.
[3] Brody W R,Meindl J D.Theoretical analysis of the CW doppler ultrasonic flowmeter[J].IEEE transactions on bio-medical engineering,1974,21(3):183.
[4] 邝建新,梁心雄.基于地铁复杂环境超声波测风仪器优势的应用分析[J].电脑知识与技术,2018(6):195-197.
[5] 王晓蕾,郭俊,陈晓颖,等.两种测风仪的动态比对试验及分析[J].解放军理工大学学报:自然科学版,2014(3):283-289.
[6] 陈伟,张有为.2D型超声波测风仪在风力发电机上应用的分析[J].仪器仪表用户,2016,23(8):90-92.
[7] Wyngaard J C,Zhang S F.Transducer-Shadow Effects on Turbulence Spectra Measured by Sonic Anemometers[J].Journal of Atmospheric and Oceanic Technology,1985,2(4):548-558.
[8] Rajita G,Mandal N.Review on transit time ultrasonic flowmeter[C]// International Conference on Control,Instrumentation,Energy & Communication.IEEE,2016:88-92.
[9] Horst T W,Semmer S R,Maclean G.Correction of a Non-orthogonal,Three-Component Sonic Anemometer for Flow Distortion by Transducer Shadowing[J].Boundary-Layer Meteorology,2015,155(3):371-395.
[10] 行鸿彦,吴红军,徐伟,等.三维超声波换能器测风阵列研究[J].仪器仪表学报,2017(12):2943-2951.
[11] 行鸿彦,魏佳佳,徐伟,等.超声波换能器测风阵列的改进设计[J].仪器仪表学报,2017,38(8):1988-1995.
[12] Javadzadegan A,Moshfegh A,Fulker D,et al.Development of a Computational Fluid Dynamics Model for Myocardial Bridging[J].J Biomech Eng,2018,140(9):091010.
[13] 张志,李振山,蔡宁生.焦炭燃烧模型的改进及其Fluent实现与实验验证[J].中国电机工程学报,2015,35(7):1681-1688.
[14] 胡晓蕾,高岩.基于FLUENT模拟的通风屋顶流场分析[J].山西建筑,2018,44(7):123-124.
[15] Zhao K.The Study of Ultrasonic Finder with Time Difference Method[J].International Electronic Elements,2005(1):65-67.
[16] Williams G,Baxter R,He H,et al.A Comparative Study of RNN for Outlier Detection in Data Mining[C]// Proceedings of the 2002 IEEE International Conference on Data Mining.IEEE,2002:709-712.
[17] 林慧君,徐荣聪.组合ARMA与SVR模型的时间序列预测[J].计算机与现代化,2009(8):19-22.
[18] 何正义,曾宪华,曲省卫,等.基于集成深度学习的时间序列预测模型[J].山东大学学报(工学版),2016(6):40-47.
[19] 黎亚雄,张坚强,潘登,等.基于RNN-RBM语言模型的语音识别研究[J].计算机研究与发展,2014,51(9):1936-1944.
[20] 张东晓,陈云天,孟晋.基于循环神经网络的测井曲线生成方法[J].石油勘探与开发,2018(4):65-74.
[21] Greff K,Srivastava R K,Koutnik J,et al.LSTM:A Search Space Odyssey[J].IEEE Transactions on Neural Networks & Learning Systems,2015,28(10):2222-2232.
[22] Gers F A,Schmidhuber J,Cummins F.Learning to Forget:Continual Prediction with LSTM[J].Neural Computation,2000,12(10):2451-2471.
[23] 黄敏,卢会国,王保强,等.超声测风传感器在回路风洞中的测试[J].气象科技,2016,44(1):14-18.
[2] Olmos P.Ultrasonic velocity meter to evaluate the behaviour of a solar chimney[J].Measurement Science & Technology,2004,15(7):N49.
[3] Brody W R,Meindl J D.Theoretical analysis of the CW doppler ultrasonic flowmeter[J].IEEE transactions on bio-medical engineering,1974,21(3):183.
[4] 邝建新,梁心雄.基于地铁复杂环境超声波测风仪器优势的应用分析[J].电脑知识与技术,2018(6):195-197.
[5] 王晓蕾,郭俊,陈晓颖,等.两种测风仪的动态比对试验及分析[J].解放军理工大学学报:自然科学版,2014(3):283-289.
[6] 陈伟,张有为.2D型超声波测风仪在风力发电机上应用的分析[J].仪器仪表用户,2016,23(8):90-92.
[7] Wyngaard J C,Zhang S F.Transducer-Shadow Effects on Turbulence Spectra Measured by Sonic Anemometers[J].Journal of Atmospheric and Oceanic Technology,1985,2(4):548-558.
[8] Rajita G,Mandal N.Review on transit time ultrasonic flowmeter[C]// International Conference on Control,Instrumentation,Energy & Communication.IEEE,2016:88-92.
[9] Horst T W,Semmer S R,Maclean G.Correction of a Non-orthogonal,Three-Component Sonic Anemometer for Flow Distortion by Transducer Shadowing[J].Boundary-Layer Meteorology,2015,155(3):371-395.
[10] 行鸿彦,吴红军,徐伟,等.三维超声波换能器测风阵列研究[J].仪器仪表学报,2017(12):2943-2951.
[11] 行鸿彦,魏佳佳,徐伟,等.超声波换能器测风阵列的改进设计[J].仪器仪表学报,2017,38(8):1988-1995.
[12] Javadzadegan A,Moshfegh A,Fulker D,et al.Development of a Computational Fluid Dynamics Model for Myocardial Bridging[J].J Biomech Eng,2018,140(9):091010.
[13] 张志,李振山,蔡宁生.焦炭燃烧模型的改进及其Fluent实现与实验验证[J].中国电机工程学报,2015,35(7):1681-1688.
[14] 胡晓蕾,高岩.基于FLUENT模拟的通风屋顶流场分析[J].山西建筑,2018,44(7):123-124.
[15] Zhao K.The Study of Ultrasonic Finder with Time Difference Method[J].International Electronic Elements,2005(1):65-67.
[16] Williams G,Baxter R,He H,et al.A Comparative Study of RNN for Outlier Detection in Data Mining[C]// Proceedings of the 2002 IEEE International Conference on Data Mining.IEEE,2002:709-712.
[17] 林慧君,徐荣聪.组合ARMA与SVR模型的时间序列预测[J].计算机与现代化,2009(8):19-22.
[18] 何正义,曾宪华,曲省卫,等.基于集成深度学习的时间序列预测模型[J].山东大学学报(工学版),2016(6):40-47.
[19] 黎亚雄,张坚强,潘登,等.基于RNN-RBM语言模型的语音识别研究[J].计算机研究与发展,2014,51(9):1936-1944.
[20] 张东晓,陈云天,孟晋.基于循环神经网络的测井曲线生成方法[J].石油勘探与开发,2018(4):65-74.
[21] Greff K,Srivastava R K,Koutnik J,et al.LSTM:A Search Space Odyssey[J].IEEE Transactions on Neural Networks & Learning Systems,2015,28(10):2222-2232.
[22] Gers F A,Schmidhuber J,Cummins F.Learning to Forget:Continual Prediction with LSTM[J].Neural Computation,2000,12(10):2451-2471.
[23] 黄敏,卢会国,王保强,等.超声测风传感器在回路风洞中的测试[J].气象科技,2016,44(1):14-18.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |