为进一步提高接地网腐蚀速率的预测精度，解决传统模型易陷入局部最优且随机初始化模型参数影响预测准确性和稳定性的问题。文中首先将混沌映射、动态搜索半径策略和优化气味浓度判定公式引入果蝇优化算法(Fruit fly Optimization Algorithm, FOA)得到改进后的果蝇优化算法(Update Fruit fly Optimization Algorithm, UFOA)；然后利用UFOA优化BP神经网络的初始权值和阈值，建立基于UFOA优化的BP神经网络接地网腐蚀速率预测模型(UFOA-BP)；最后以重庆24座变电站的接地网数据进行实验验证。结果表明相对FOA优化的BP神经网络模型、BP神经网络模型、人工蜂群算法优化的支持向量机模型和广义回归神经网络模型，文中提出的UFOA-BP模型在预测精度和模型稳定性方面均优于其他四种模型，验证了该模型的有效性和可行性，为运维人员提前发现接地网安全隐患并安排检修进而保障电网安全稳定运行提供帮助。
英文摘要:
In order to solve the problem that the traditional prediction model is prone to fall into the local optimum and the parameters of the model are randomly initialized which affect the accuracy and stability of the prediction so that the prediction accuracy of corrosion rate of grounding grid is further improved. This paper, chaos mapping, dynamic search radius strategy and optimized odor concentration determination formula are introduced into fruit fly optimization algorithm (FOA) at first. Then, the weights and thresholds of BP neural network are adjusted adaptively by updated fruit fly optimization algorithm (UFOA) which has a strong global search for optimal solutions. Finally, a corrosion rate prediction model of grounding grid is established based on UFOA. In the measurement data of Chongqing 24 substations experimental verification, the results show that UFOA-BP model relative BP neural network model which is optimized by FOA, BP neural network model, support vector machine (SVM) model which is optimized by artificial bee colony (ABC) and generalized regression neural network (GRNN) model is better in prediction accuracy and stability. It illustrates that the UFOA-BP model which is proposed in this paper is effective and feasible to solve the problem of grounding grid corrosion rate prediction. Provide help for operation and maintenance personnel to find potential safety hazards of grounding grid in advance and arrange maintenance to ensure safe and stable operation of power grid.