Proceedings of the
The Nineteenth International Conference on Computational Intelligence and Security (CIS 2023)
December 1 – 4, 2023, Haikou, China
Soft Measurement Modeling of Electric Arc Furnace Current and Voltage Based on Iwoa Optimization Support Vector Regression
School of Electrical and Power Engineering, Taiyuan University of Technology, China.
ABSTRACT
A soft measurement method for arc current and arc voltage of an electric arc furnace based on improved whale algorithm optimization support vector regression (IWOA-SVR) is proposed to improve the soft measurement modeling method based on back propagation neural networks (BPNN) into a SVR model optimized by IWOA algorithm for the prediction of arc electric parameters. propagation neural networks (BPNN)-based soft measurement modeling method is improved to an IWOA algorithm-optimized SVR model for the prediction of arc electric parameters. The method takes advantage of the strong learning ability and low generalization error rate of the support vector regression model, and can effectively solve the characteristics of the arc furnace production process generated by the sudden change of data affecting the performance of the model. To overcome the optimization problem of the combination of penalty coefficients and kernel function parameters in the SVR model, a nonlinear dynamic weight improvement whale optimization algorithm is introduced for parameter optimization, which allows the SVR model to better learn the nonlinear relationship between the auxiliary variables and the dominant variables of the soft measurement model. Finally, the trained model is substituted into the test set for prediction performance testing, which verifies that the model proposed in this paper has high accuracy.
Keywords: Arc current, Arc voltage, Soft measurement, Whale optimization, Support vector regression, Prediction.
Download PDF
School of Electrical and Power Engineering, Taiyuan University of Technology, China.
ABSTRACT
A soft measurement method for arc current and arc voltage of an electric arc furnace based on improved whale algorithm optimization support vector regression (IWOA-SVR) is proposed to improve the soft measurement modeling method based on back propagation neural networks (BPNN) into a SVR model optimized by IWOA algorithm for the prediction of arc electric parameters. propagation neural networks (BPNN)-based soft measurement modeling method is improved to an IWOA algorithm-optimized SVR model for the prediction of arc electric parameters. The method takes advantage of the strong learning ability and low generalization error rate of the support vector regression model, and can effectively solve the characteristics of the arc furnace production process generated by the sudden change of data affecting the performance of the model. To overcome the optimization problem of the combination of penalty coefficients and kernel function parameters in the SVR model, a nonlinear dynamic weight improvement whale optimization algorithm is introduced for parameter optimization, which allows the SVR model to better learn the nonlinear relationship between the auxiliary variables and the dominant variables of the soft measurement model. Finally, the trained model is substituted into the test set for prediction performance testing, which verifies that the model proposed in this paper has high accuracy.
Keywords: Arc current, Arc voltage, Soft measurement, Whale optimization, Support vector regression, Prediction.
Download PDF