CAJ | 학술논문

针对钢铁生产流程中能耗预测模型建立困难、预测精度低等问题，提出了一种基于蚁群优化的小波神经网络的钢铁生产物流能耗预测模型，首先对钢铁生产过程以及影响生产能耗的因素进行分析，确定输入参数构成特征空间，然后利用小波变换重构特征空间，接着利用神经网络模型建立能耗预测模型，最后采用蚁群算法对预测模型参数进行优化。在炼铁、炼钢以及轧钢工序的能耗预测实验表明，提出的方法具有较好的普适性，提高了预测精度，为钢铁企业提前了解能耗需求提供了指导。
침대강철생산류정중능모예측모형건립곤난、예측정도저등문제，제출료일충기우의군우화적소파신경망락적강철생산물류능모예측모형，수선대강철생산과정이급영향생산능모적인소진행분석，학정수입삼수구성특정공간，연후이용소파변환중구특정공간，접착이용신경망락모형건립능모예측모형，최후채용의군산법대예측모형삼수진행우화。재련철、련강이급알강공서적능모예측실험표명，제출적방법구유교호적보괄성，제고료예측정도，위강철기업제전료해능모수구제공료지도。
To overcome the difficulties of establishing a prediction model of energy consumption and to improve the predicting accuracy,a prediction model of material and energy consumption in steel production based on ant colony optimization wavelet neural network is proposed in this paper.The steel production process and the factors which impact energy consumption are firstly analyzed.The input parameters are then determined to constitute fea-ture space,which is reconstructed by wavelet transform.A prediction model of energy consumption with neural network is constructed thereafter.Finally the solving process is optimized by using ant colony algorithm.The ener-gy consumption prediction experiments in the processes of iron -making,steel -making and steel -rolling show that the proposed method has better universality,and at the same time it improves the prediction accuracy and provides guidance for steel enterprises to understand energy needs in advance.