Prediction of Gas Intake of Iron and Steel Enterprises Owned Power Plants Based on HS-RVM Model

Author(s): 
H. J. Xiao †§, M. Y. Zhang ‡,S. B. Wang §

Affiliation(s): 
† Quality Development Institute, Kunming University of Science and Technology, Kunming Yunnan 650093, China
‡ Energy Industry Development Institute of Yunnan Province Energy Investment Group CO., LTD, Yunnan Kunming, 650021, China
§ State Key Laboratory of Complex Nonferrous Metal Resources Clean Utilization, Kunming University of Science and Technology, Kunming Yunnan 650093, China

Cite this paper
 H. J. Xiao, M. Y. Zhang, S. B. Wang, “Prediction of Gas Intake of Iron and Steel Enterprises Owned Power Plants Based on HS-RVM Model”, Journal of Mechanical Engineering Research and Developments, vol. 39, no. 1, pp. 30-38, 2016. DOI: 10.7508/jmerd.2016.01.005

ABSTRACT: To deal with the difficulties in accurate prediction and effective scheduling of the empirical model of gas amount supplied by self-owned power plants in iron and steel enterprises, by analyzing the historical data of gas amount supplied by self-owned power plants, this paper adopts Relevance Vector Machine (RVM) to explore the tendency and rules of gas supply amount change. Meanwhile, Harmony Search (HS) is used to optimize the hyper parameters affecting the performance of RVM prediction model. A HS-RVM prediction model of gas amount supplied by self-owned power plants is constructed. This model is verified based on the actual operating data of power plants. Results have shown that it enjoys better performance than HP-Elman and PSO-SVR models, with the predicted root-mean-square error of 30 (period), 45 (period) and 60 (period) as 1.867%, 1.442% and 1.376% respectively, which meets the gas scheduling and management demand in self-owned power plants of iron and steel enterprises. Wilcoxon sign rank test verifies the effectiveness of this HS-RVM prediction model. 

Keywords : Gas; Self-owned power plants; Relevance vector machine; Harmony search; Prediction

References
[1] T. Liu, X. Zhang, and S. W. Huang, “Recovery of Residual Heat Resources From Owned Power Plant in the Iron abd Steel Industry”, Journal of Iron and Steel Research, vol. 26, no. 11, pp. 29-33, 2014.
[2] H. Li, J. J. Wang, H. Wang, and H. Meng, “Establishment of PNN-HP-ON-LS SVM model to predict the amount of gas emergence”, The Chinese Journal of Process Engineering, vol. 3, pp. 451-457, 2013.
[3] H. Meng, J. J. Wang, and H. Wang, “Reflection on the imbalance of gas system in steel complex”, Iron and Steel, vol. 3, pp. 84-90, 95, 2015.
[4] H. Meng, J. J. Wang, and H. Wang, “Research on Forecasting Varying Tendency of Gas Supply in Self-Provided Power Plant Based on ARMA-ARCH”, Journal of Kunming University of Science and Technology (Natural Science Edition), vol. 3, pp. 66-72, 2014.
[5] H. J. Li, J. J. Wang, H. Wang, “An HP(2)-Elman Model for Prediction and Scheduling on Affluent Gas in Steel Enterprises”, Journal of Iron and Steel Research, vol. 25, no. 7, pp. 11-18, 2013.
[6] H. J. Li, J. J. Wang, and H. Wang, “An HP-Elman-LSSVM Model for Prediction and Adjustment on self-Provided Power Plant By-Product Gas Supply in steel Enterprises”, Iron and Steel, vol. 48, no. 8, pp. 75-81, 2013.
[7] H. Drucker, C. Burges, L. Kaufman, and V. Vapnik, “Support Vector Regression Machines”, Advances in Neural Information Processing Systems, vol. 28, no. 7, pp. 391 – 394, 1996.
[8] L. Zhang, W. D. Zhou, and P. C. Chang, “Iterated time series prediction with multiple support vector regression models”, Neurocomputing, vol. 99, no. 1, pp. 411-422, 2013.
[9] M. E. Tipping, “Sparse Bayesian Learning and the Relevance Vector Machine”, Journal of Machine Learning Research, vol. 1, no. 3, pp. 211-244, 2001.
[10] Q. F. Meng, Y. H. Chen, and Z. Q. Feng, “Nonlinear prediction of small scale network traffic based on local relevance vector machine regression model”, Acta Physica Sinica, vol. 62, no. 15, pp. 150-189, 2013.
[11] L. H. Wu, “Network Intrusion Detection Based on Relevance Vector Machine Optimized by Particle Swarm Optimization Algorithm”, Microelectronics & Computer, vol. 27, no. 5, pp. 181-184, 2010.
[12] Z. W. Geem and J. H. Kim, “A New Heuristic Optimization Algorithm: Harmony Search”, Simulation, vol. 76, no. 2, pp. 60-68, 2001.
[13] D. Zou, L. Ga, and S. Li, “An effective global harmony search algorithm for reliability problems”, Expert Systems with Applications, vol. 38, no. 4, pp. 4642-4648, 2011.
[14] J. Yan, Y. Liu, and S. Han, “Wind power grouping forecasts and its uncertainty analysis using optimized relevance vector machine”, Renewable & Sustainable Energy Reviews, vol. 27, no. 6, pp. 613-621, 2013.
[15] M. Mahdavi, M. Fesanghary, and E. Damangir, “An improved harmony search algorithm for solving optimization problems”, Applied Mathematics & Computation, vol. 188, no. 2, pp. 1567-1579, 2007.
[16] J. Lu and J. H. Gu, “Continuous function optimization based on improved harmony search algorithm”, Journal of Computer Applications, vol. 34, no. 1, pp. 194-198, 2014.
[17] W, Zhao and D. Wei, “Relevance Vector Machine Combined with Glowworm Swarm Optimization for Cam Relationship of Kaplan Turbine”, Advanced Science Letters, vol. 11, no. 4, pp. 244-247, 2012.
[18] H. Kong, E. Qi, and H. Li, “An MILP model for optimization of byproduct gases in the integrated iron and steel plant”, Applied Energy, vol. 87, no. 7, pp. 2156-2163, 2010.