基于BEADS-ESMC组合算法的三相光伏并网逆变柜触点红外温度预测方法

Prediction Method for Electrical Components In frared Temperature of Three-phase Photovoltaic Grid-connected Inverter Cabinet Based on BEADS-ESMC Combination Algorithm

针对三相光伏并网逆变柜内触点红外温度噪声较大导致的温度预测难问题,提出了一种基于稀疏去噪和基线估计(baseline estimation and denoising with sparsity,BEADS),三次指数平滑法(exponential smoothing,ES)与马尔科夫链(markov chain,MC)相结合的红外温度预测组合算法(BEADS-ESMC)。通过样本序列分析确定基线估计参数和平滑系数,并根据样本统计规律建立状态转移概率矩阵,对重要触点的温度进行多步预测。实验表明,应用本方法的多步红外温度预测精度优于常用的ESMC方法,且具有较短的算法处理时间。

In order to improve the infrared temperature prediction accuracy for important electrical components of three-phase photovoltaic grid-connected inverter cabinet.An infrared temperature prediction combined algorithm (BEADS-ESMC)combining baseline estimation and denoising with sparsity (BEADS),triple exponential smoothing (ES)prediction and Markov chain (MC)error correction is proposed.The baseline estimation parameters and smoothing coefficients are determined by sample sequence analysis,and according to the statistics of the sample,a state transition probability matrix is established to predict the temperature of important electrical components.Experimental results show that the accuracy of multi-step infrared temperature prediction using this method is better than that of the commonly used ESMC method,and the algorithm processing time is shorter.