基于正弦曲线拟合的快速限流预测算法

A quick current-limiting predicted algorithm based on sinusoidal curve fitting

传统的单纯由硬件电路实现的限流器针对电流突然上升的情况不能及时做出反应,而基于嵌入式软件实现的限流器能根据采集的电流数据通过算法处理及时做出预测,使电流峰值还未出现就可以计算出其幅值并及时做出反应。本文介绍了一种基于微控制单元(MCU)实现的快速限流预测算法,该算法可以把采集到的受噪声等影响的电流数据拟合成接近理想的正弦曲线,预测求解出电流幅度,从而快速判断电流的峰值是否达到限流器的限流标准,当电路中电流突然超过30 m A或5倍以上时,在规定时间内实现脱扣。经过多次实验可以验证本算法能够准确预测出电流幅值是否达到脱扣条件。

The traditional current limiters implemented by hardware can not respond timely when facing the sudden rising of the current, while those based on the software implementation can calculate in advance the amplitude of the current according to acquisition data and respond timely before appearing of the crest and the valley. A fast forecast algorithm for current limit is introduced based on Micro Control Unit(MCU) implementation, which can fit the current acquisition data influenced by noise to the sine curve close to the ideal solution as far as possible. It can predict the current amplitude, and quickly determine whether the current crest reaches the current limit. When the circuit current increases by more than 5 times or 30 m A, it would cut off timely. Through many experiments it is verified that the algorithm can accurately predict whether the current reaches the condition of the releasing or not.