基于BP神经网络和SSA SVM的接地网腐蚀速率组合预测

Combination Forecasting for Corrosion Rate of Grounding Grid Based on BP Neural Network and SSA SVM

摘要:

为提高接地网腐蚀速率预测精度，提出了一种接地网腐蚀速率组合预测方法。首先，采用SSA算法对SVM进行优化，建立接地网SSA SVM腐蚀预测速率模型；然后，采用6-11-1的BP神经网络对SSA SVM模型的预测残差进行修正，建立了基于BP神经网络和SSA SVM的接地网腐蚀速率组合预测模型；最后，采用接地网腐蚀实验数据进行算例分析。结果表明，所提接地网腐蚀速率组合模型预测结果的均方根误差、平均相对误差和相关系数分别为0.192、4.98%和0.974 6，在模型稳定性、预测精度、预测结果与实际值的相关性均优于其他模型，验证了所提模型的正确性和优越性。

Abstract:

In order to improve the prediction accuracy of grounding grid corrosion rate, a combined prediction method of grounding grid corrosion rate was proposed. Firstly, the SSA algorithm is used to optimize the SVM, and the SSA SVM corrosion prediction rate model of the grounding grid is established. Then, the prediction residual of the SSA SVM model is modified using the 6-11-1 BP neural network, and a combined prediction model of the grounding grid corrosion rate based on BP neural network and SSA SVM is established. Finally, an example is analyzed by using the experimental data of grounding grid corrosion. The results show that the root mean square error, average relative error and correlation coefficient of the prediction results of the combined model of grounding grid corrosion rate proposed in this paper are 0.192, 4.98% and 0.974 6, respectively. The model stability, prediction accuracy, and correlation between the prediction results and the actual values are better than other models, which verifies the correctness and superiority of the model.