发明名称 Machine Learning Approach for Analysis and Prediction of Cloud Particle Size and Shape Distribution
摘要 Techniques for analysis and prediction of cloud particle distribution and solar radiation are provided. In one aspect, a method for analyzing cloud particle characteristics includes the steps of: (a) collecting meteorological data; (b) calculating solar radiation values using a radiative transfer model based on the meteorological data and blended guess functions of a cloud particle distribution (c) optimizing the cloud particle distribution by optimizing the weight coefficients used for the blended guess functions of the cloud particle distribution based on the solar radiation values calculated in step (b) and measured solar radiation values; (d) training a machine-learning process using the meteorological data collected in step (a) and the cloud particle distribution optimized in step (c) as training samples; and (e) predicting future solar radiation values using forecasted meteorological data and the machine-learning process trained in step (d).
申请公布号 US2014324350(A1) 申请公布日期 2014.10.30
申请号 US201313873601 申请日期 2013.04.30
申请人 INTERNATIONAL BUSINESS MACHINES CORPORATION 发明人 Hamann Hendrik F.;Lu Siyuan
分类号 G01W1/10 主分类号 G01W1/10
代理机构 代理人
主权项 1. A method for analyzing cloud particle characteristics, the method comprising the steps of: (a) collecting meteorological data; (b) calculating solar radiation values using a radiative transfer model based on the meteorological data and blended guess functions of a cloud particle distribution such that the solar radiation values are generated for each of the blended guess functions; (c) optimizing the cloud particle distribution by optimizing the weight coefficients used for the blended guess functions of the cloud particle distribution based on the solar radiation values calculated in step (b) and measured solar radiation values; (d) training a machine-learning process using the meteorological data collected in step (a) and the cloud particle distribution optimized in step (c) as training samples; and (e) predicting future solar radiation values using forecasted meteorological data and the machine-learning process trained in step (d).
地址 Armonk NY US