基于大数据的嵌入式设备的超负荷状态估计

Overload State Estimation of Embedded Devices Based on Big Data Analysis

对嵌入式设备的超负荷状态进行准确估计,能够提高嵌入式设备的稳定性。由于嵌入式设备构造较为复杂,各个零件承担的电压电流负荷无法形成单一的稳定特征,会发现较大变化。传统的估计方法,仅仅以单个电路的负荷数据特征进行实时更新叠加,没有充分考虑负荷参数之间的相互影响,以及特征变化中的可识别周期性,计算的结果不准。提出采用大数据分析的嵌入式设备超负荷状态估计方法。对嵌入式设备的等效原理进行分析,建立嵌入式设备的等效模型;将各支路上的大数据功率数据作为超负荷状态估计的变量,得到稳定的嵌入式设备各支路上的电流,根据卡尔曼滤波原理建立嵌入式设备的超负荷状态估计的目标函数,将各支路电流作为卡尔曼滤波的输入量,进行泰勒级展开,最终获得准确的估计结果。仿真结果表明,改进算法能够提高超负荷估计的精度。

An overload state estimation method of embedded device based on big data analysis is proposed. To an- alyze the principle of equal effects of embedded device, an equal effect model of embedded device is established; the big data power data on each branch are used as the variables of the overload state estimation, the stable current in each branch of embedded device is got. The objective function of the overload state estimation of the embedded device is established according to the principle of Kalman filtering. The branch current is taken as the input of the Kalman filtering, to carry out the Taylor series expansion and finally obtain the accurate estimated results. The simulation re- sults show that the improved algorithm can increase the accuracy of overload estimation.