基于改进萤火虫算法优化的电力光纤线路状态预测模型研究

Research on state prediction model of power optical fiber line based on firefly algorithm optimization

摘要:

为了提升电力通信系统光纤线路状态预测的准确率，提出一种基于改进萤火虫(Firefly Algorithm, FA)算法优化的电力光纤线路状态预测模型(FA-ARIMA-GRU)。首先，针对FA寻优过早收敛和寻优精度低等问题，对整体距离进行指数加权平均，设计了迭代衰减步长因子，并且对偏差修正进行考虑，从而改善上述FA所存在的问题；其次，将改进的FA用于ARIMA-GRU光功率预测模型输入参数的优化，从而在一定程度提升ARIMA-GRU预测模型输入参数的准确性；最终，通过仿真试验对FA-ARIMA-GRU预测模型的效果进行验证，结果表明FA-ARIMA-GRU预测模型具有较优的预测效果，精准预测光功率值，提前掌握电力光纤的线路状态，预知光纤线路故障、有效规避故障和保障电力通信传输通畅不间断。

Abstract:

In order to improve the accuracy of optical fiber line state prediction in power communication system, a power optical fiber line state prediction model (FA-ARIMA-GRU) optimized based on improved firefly algorithm (FA) algorithm is proposed. Firstly, aiming at the problems of premature convergence and low optimization accuracy of FA optimization, the overall distance is exponentially weighted average, the iterative attenuation step factor is designed, and the deviation correction is considered to improve the problems of FA; Secondly, the improved FA is used to optimize the input parameters of ARIMA-GRU optical power prediction model, so as to improve the accuracy of the input parameters of ARIMA-GRU prediction model to a certain extent; Finally, the effect of fa-arima-gru prediction model is verified through simulation test. The results show that FA-ARIMA-GRU prediction model has better prediction effect, accurately predict the optical power value, master the line state of power optical fiber in advance, predict optical fiber line faults, effectively avoid faults and ensure smooth and uninterrupted power communication transmission.