能量吸收网络中太阳能的优化回归分析预测方法

Solar energy prediction method based on optimized regression analysis for energy harvesting network

在能量吸收无线传感网络中,为了满足能量管理与网络协议设计的需要,设计了较低计算复杂度的精确太阳能吸收预测方法。利用通径分析对太阳能吸收能量回归方法预测进行研究,首先将影响吸收太阳能量的关键因素进行关于太阳能量的通径分析,建立多变量回归模型,然后判断不同变量间的影响关系,最后对多变量回归模型进行优化。每次回归建模有效性都需通过F检验等。利用权威实验室2013年至2014年每日的数据,用Matlab软件进行数据拟合和预测,验证了预测方法的有效性。该方法不需要存储大量的历史数据,只需递推或者即时接收到的天气信息,经过极为简单的运算即可实时预测,归一化均方误差仅为2.55%。该方法运算量极小,精度较高,具有一定的应用价值。

In energy harvesting network,precise prediction of harvesting solar energy is necessary for energy management and network protocol design. The regression modeling prediction method for harvesting solar energy utilizing path analysis was studied. Firstly,the path analysis of the key factors with respect to harvesting solar energy was performed,and the multi-variable regression model was built. Secondly,the influences of the respective factors were judged. Finally,the multi-variable regression model was optimized according the above analysis. The effectiveness of regression modeling was validated by several tests,such as F check-out / test. Data fitting and prediction was performed by MATLAB using the daily data of years 2013 and 2014 from Solar Radiation Monitoring Laboratory,validating the effectiveness of the proposed prediction method. The method was able to predict with high precision of general root mean square error 2. 55% and amazing low computational complexity,in need of only received or deduced instant weather data,free of storing large amount history data. The method was so low in computation complexity,and so high in prediction precision,thus indicating significant application value.