基于ARMA模型的序贯概率比改进算法

针对常规的序贯概率比检验方法（sequenceprobabilityratiotest，SPRT）在故障检测时存在的较大时间延迟，本文提出一种通过对序贯概率比时间序列运用自回归滑动平均（auto—regressivemovingaverage，ARMA）模型，建立预测反馈控制机制的SPRT改进算法。该算法的核心思想是根...

Improved Algorithm of Sequential Probability Ratio Based on ARMA Model

Aiming at the large time delay existing in the period of fault detection of normal Sequential Probability Ratio Test （SPRT）, an improved SPRT algorithm of building a forecast feedback control mechanism is proposed in this paper by using ARMA model for the time series of sequence probability ratio. The core idea is that the recursion formula of the different sequence probability ratio is selected as per the variation tendency of SPR, so as to weaken the cumulative effect of SPRT in the normal state to achieve the purpose of decreasing the time delay. Finally, the improved method has been applied in the fault detection of white noise series. The simulation result proves that the method can shorten the time delay effectively.