基于CEEMD-TCN模型的变压器油中溶解气体浓度预测方法

Prediction Method of Dissolved Gas Concentration in Transformer Oil Based on CEEMD-TCN Model

变压器油中溶解气体浓度是了解变压器运行状态、判断变压器是否发生故障的重要指标。针对油中溶解气体浓度序列非线性、非平稳性的特点,数据直接训练模型会明显降低预测精度,因而提出了一种基于互补集合经验模态分解和时间卷积网络相结合的预测方法。首先,将原始序列分解成不同尺度的子序列分量,经过预处理后训练时间卷积网络,并优化网络超参数,各分量的预测结果叠加重构从而获得最终预测结果。通过实验验证表明,该模型预测误差小,预测精度高。

The concentration of dissolved gas in transformer oil is an important indicator for understanding the operating status of the transformer and judging whether the transformer is faulty.In view of the non-linear and non-stationary characteristics of the dissolved gas concentration sequence in the oil,the data training model directly will significantly reduce the prediction accuracy,so a prediction method based on the combination of complete ensemble empirical mode decomposition and temporal convolutional network is proposed.First,the original sequence is decomposed into sub-sequence components of different scales,the temporal convolutional network is trained after preprocessing,and the network hyperparameters are optimized,and the prediction results of each component are superimposed and reconstructed to obtain the final prediction result.Experimental verification shows that the model has small prediction errors and high prediction accuracy.