基于粒子群优化–长短期记忆网络模型的变压器油中溶解气体浓度预测方法

Prediction of Dissolved Gas Concentration in Transformer Oil Based on PSO-LSTM Model

电力变压器作为电网中传输和变换电能的主要设备,对油中溶解气体的浓度进行有效预测,可为变压器的故障诊断及状态评估技术提供一定的理论依据。鉴于此,提出一种基于粒子群优化算法(particle swarm optimization,PSO)与长短期记忆网络(long-shorttermmemory,LSTM)的变压器油中溶解气体浓度预测方法。首先该模型以油中溶解的7种特征气体浓度序列作为可视输入;然后通过使用粒子群优化算法对长短期记忆网络中相关超参数进行迭代优化;最后构建PSO-LSTM组合模型对油中溶解气体的浓度进行预测。该模型克服了依据经验选取参数而导致预测精度低的问题。算例分析结果表明,相较于传统预测算法,所提方法可以更好地追踪油中溶解气体浓度的变化规律,提高了预测精度,为电力变压器安全稳定运行提供了有力保障。

As the main equipment of power transmission and transformation in grid,the effective prediction of the dissolved gas concentration in the transformer oil can provide a certain theoretical basis for the transformer fault diagnosis and state assessment.Therefore,this paper proposes a method for predicting the concentration of the dissolved gas in the transformer oil based on the particle swarm optimization(PSO)and the long-short term memory(LSTM).Firstly,the model takes the concentration sequence of the seven characteristic gases dissolved in the oil as a visual input.Then,the(PSO)is used to optimize the relevant super parameters in the LSTM.Finally,a PSO-LSTM combined model is constructed to predict the concentration of the dissolved gas in the oil.The model overcomes the problem of low prediction accuracy caused by selecting parameters based on experience.The results of examples show that,compared with the traditional prediction algorithms,the proposed method in this paper can better track the variation rule of the dissolved gas concentration in the oil,effectively improve the prediction accuracy,and provide a strong guarantee for the safe and stable operation of the power transformer.