WSEAS Transactions on Signal Processing
Print ISSN: 1790-5052, E-ISSN: 2224-3488
Volume 12, 2016
Research on Data-Driven Energy Efficiency Optimization for Copper Flash Smelting Process
Authors: , , , ,
Abstract: The characteristics of the copper flash smelting process include: multiple variable, nonlinearity, strong coupling, long delay and large fluctuations. With the development of computer technology and industrial automation, the complex industrial process has produced a large number of production data, which contains rich information for the mining of their patterns. In order to improve energy efficiency of the copper flash smelting process, this paper presents a method for minimizing energy consumption with meeting three technical indexes (matte grade, matte temperature and ratio of Fe to SiO2) as a constraint. Our method is composed of two main parts: firstly, the least square support vector machine model (LS-SVM) is used to predict three technical indexes and we compare it with back propagation (BP) neural network; secondly, the comprehensive energy consumption model based on particle swarm optimization (PSO) is used to find the operational-pattern of lowest energy consumption. Experimental results on practical production data show that our energy efficiency optimization method can accurately predict three technical indexes and find the operational-pattern leading to the lowest energy consumption.
Search Articles
Keywords: Copper Flash Smelting, Energy Consumption, Energy Efficiency Optimization, Three Technical Indexes, Least Square Support Vector Machine (LS-SVM), Particle Swarm Optimization (PSO)
Pages: 58-67
WSEAS Transactions on Signal Processing, ISSN / E-ISSN: 1790-5052 / 2224-3488, Volume 12, 2016, Art. #8