摘要
随着气象现代化建设的深入开展,大量观测设备投入并应用到气象预报和防灾减灾的服务中。为确保这些观测设备数据的可靠性和准确性,实现计量业务工作的信息化管理,提高计量业务的实时性和计量数据的可用性,通过建立指数函数回归模型和深度学习的方法,对不同运行时长的风速计所记录的启动风速进行统计,以建立启动风速与运行时长之间的相关数学模型并进行对比。结果表明,指数函数回归模型和自行设计的BP神经网络模型均能胜任最大使用年限的回归任务,且相对于回归模型,BP神经网络不需要分别对各区域数据分别建立模型,对风向传感器的计量检定具有重要意义。
With the deepening of meteorological modernization,a large amount of observational equipment has been inves-ted in and applied to meteorological forecasting and disaster prevention and reduction services.To ensure the reliability and accuracy of data from these observation devices,achieve information-based management of measurement business work,and improve the real-time performance of measurement business and the availability of measurement data.This ar-ticle establishes an exponential regression model and deep learning methods to statistically analyze the start-up wind speed recorded by anemometers with different operating hours,to establish a mathematical model of the correlation be-tween start-up wind speed and operating hours and compare them.The results indicate that both the exponential function regression model and the self-designed BP neural network model are capable of performing the regression task for the maximum service life.Compared to the regression model,the BP neural network does not require separate models for each region's data,which is significant for the metrological calibration of wind direction sensors.
作者
胡苍龙
李晗琪
王家珉
杨涛
王海江
HU Canglong;LI Hanqi;WANG Jiamin;YANG Tao;WANG Haijiang(Sichuan Meteorological Observation Data Center,Chengdu 610072,China;College of Electronic Engineering,Chengdu University of Information Technology,Chengdu 610225,China)
出处
《成都信息工程大学学报》
2026年第2期154-159,共6页
Journal of Chengdu University of Information Technology
基金
四川省科技计划重点研发资助项目(2023YFG0170)
高原与盆地暴雨旱涝灾害四川省重点实验室科技发展资助项目(SCQXKJYJXMS202319)。
关键词
风向传感器
计量方法
回归模型
神经网络
wind direction sensor
measurement method
regression model
neural network
