Prediction of electromechanical equipments state nonlinear and non-stationary condition effectively is significant to forecast the lifetime of electromechanical equipments. In order to forecast electromechanical equipments state exactly, support vector regression optimized by genetic algorithm is proposed to forecast electromechanical equipments state. In the model, genetic algorithm is employed to choose the training parameters of support vector machine, and the SVR forecasting model of electromechanical equipments state with good forecasting ability is obtained. The proposed forecasting model is applied to the state forecasting for industrial smokes and gas turbine. The experimental results demonstrate that the proposed GA-SVR model provides better prediction capability. Therefore, the method is considered as a promising alternative method for forecasting electromechanical equipments state.
With the rapid development of science and technology, the composition and structure of electromechanical equipment are more and more complicated. Once a certain equipment breaks down, chain reaction will be caused (Acevedo-Rodríguez et al., 2009, Katagiri and Abe, 2006 and Zhu et al., 2009). The whole production system cannot operate smoothly, then huge economic loss will be caused. Prediction of electromechanical equipments state nonlinear and non-stationary condition effectively is significant to forecast the lifetime of electromechanical equipments (Elish and Elish, 2008, Hämäläinen et al., 1996, Valentini, 2002 and Whitley et al., 1990). Running conditions of large-scale electromechanical equipment are complicated. When the equipment is in the fault state, its thermal dynamics characteristic exerts complexity and nonlinear. As artificial neural networks have general nonlinear mapping capabilities, it becomes a popular prediction technique in the prediction of electromechanical equipments (Jain et al., 2007 and Nandi et al., 2004). But the prediction results of artificial neural networks are affected by their drawbacks, such as lack generalization and local optimization solution.
Support vector regression (SVR) is a novel learning machine based on statistical learning theory, which has been successfully used for nonlinear systems modeling. Compared with artificial neural networks, SVM provides more reliable and better performance under the same training conditions (King et al., 2000 and Yuan et al., 2009). How to choose the best training parameters is an important problem for SVR because this problem will directly affect its regression accuracy. In our work, genetic algorithm (GA) is used to optimize the model parameters, and so the generalization ability and forecasting accuracy are improved. Based on the Darwinian principle of ‘survival of the fittest’, GA can obtain the optimal solution after a series of iterative computations (Chang et al., 2007, Hardas et al., 2008 and Weile and Michielssen, 2000). Therefore, support vector regression optimized by genetic algorithm (GA-SVR) is proposed to forecast electromechanical equipments state. In the model, GA (Jagielska et al., 1999, Kwon et al., 2003 and Pereira and Lapa, 2003) is employed to determine training parameters of support vector machine, and the SVR forecasting model of electromechanical equipments state with good forecasting ability is obtained. The data of electromechanical equipments state are used to test the accuracy of the proposed model. The experimental results show that the GA-SVR model is considered as a promising alternative method for forecasting electromechanical equipments state.
In this paper, Section 2 introduces the theory of support vector regression. In Section 3, GA-based optimization of SVR model is introduced. The structure of electromechanical equipment state forecasting model is introduced in Section 4. In Section 5, experimental analysis for electromechanical equipment state prediction is gained. Finally, the conclusion is gained in Section 6.
The method employs a hybrid GA-SVR approach for the forecasting of electromechanical equipments state, where genetic algorithm is used to select suitable parameters of SVR. Genetic algorithm consists in maintaining a population of chromosomes, which represent potential solutions to the problem to be solved. And the leave-one-out cross-validation (LOOCV) of SVR is adopted to evaluate fitness. The proposed GV-SVR is more robust, rapid and accurate compared with ANN. The data of electromechanical equipments state are used to test the accuracy of the proposed model. The experimental results show that GA-SVR can achieve greater forecasting accuracy than artificial neural network in forecasting the electromechanical equipments state.
