基于GM(1,1)与灰区间估计的SPE通量水平长期预测

太阳质子事件（SPE，Solar Proton Events）是干扰日地空间最主要的源，大规模质子事件会影响在轨空间站实验设备的可靠性，有时甚至会威胁空间站的安全运行.提出一种基于灰色GM（1，1）和区间估计太阳质子事件预测方法；首先对1976—2010年SPE通量水平数据进行预处理，分别建立以发生时间为序列的一般SPE通量序列与极端SPE通量序列；之后将灰色GM（1，1）与区间估计相结合建立预测方法，融合反映一般SPE通量水平随周期性波动的活跃性调节系数，建立SPE通量水平长期预测模型；然后叠加不同SPE类型所得结果合成预测年份的SPE通量水平，给出未来一年或几年间SPE通量水平的变化范围；最后选取1976—2010年太阳质子事件年均值数据，分多批次预测1996—1998年和1999—2001年等SPE通量均值区间，结果表明各年实际发生SPE的通量均值均位于预测区间内，并且多年预测区间偏差最大值小于26%，实验结果还表明单次预测时长以2~3年为宜.

Novel method for predicting SPE flux levels based on GM(1,1) model and grey interval predictions

Solar proton event (SPE) is the main source interfering with solar-terrestrial space, large proton event could affect the reliability of on-orbit space station experimental facilities, and sometimes it even threatens the safe operation of space station. An SPE predicting method which is based on GM(1, 1) and interval estimation was proposed. Firstly, the SPE flux levels sequence which consists of the flux data from 1976 to 2010 were preprocessed, a normal SPE flux levels sequence and an extreme SPE flux levels sequence were established according to time occurrence. Then combined with GM(1, 1) and grey interval estimation, a new forecasting method was built up. Integrated with the active adjustment coefficient that reflecting SPE flux level of cyclical fluctuation, long-term forecasting model of SPE flux level was set up. By fusing the results obtained from different SPE type, the mean value of flux levels was obtained, then SPE flux level range was given out in next year or several years. The annual average SPE data from 1976 to 2010 were selected, SPE flux mean intervals were predicted in multiple batches between 1996 to 1998, and 1999 to 2001.The results show that all the SPE flux levels are located in the prediction interval and the maximum deviation of prediction interval for many years is less than 26%. The experiment results also show that the optimum length of a single prediction is two or three years.