摘要
针对钢铁企业实际生产过程中,采用单一预测模型进行预测难以把握大规模启停设备用电规律,预测精度较低等问题,根据生产-检修阶段的实际工艺情况,将生产,检修问题采用随机近似贪婪搜索RAGS对复杂特征进行特征选择,建立了一个自适应数据质量的多模型择优预测框架进行建模;将其应用于宝钢电网。仿真结果表明,提出多模型择优预测框架可以准确预测钢铁企业电力日负荷,为实现电力系统能源调度提供决策依据。
To solve the issues that asingle prediction model was applied to predict that iron and steel enterprises in the actual production process. It is difficult to grasp the law of the mass electricity equipment from start and end, the accuracy of prediction is low and so on. A method of random approximation greedy search ( RAGS ) is adopted to handle the problem that the complex features were selected in the actual process conditions based on the produc- tion - the maintenance plans. What is more, establish a merit prediction modeling that depend on multi-model frame- work of adaptive data quality. Simulation results of Baosteel grid show that the merit framework of multi-model can accurately predict electricity daily load for iron and steel enterprises. Above all, It can provide the basis for deci- sion-making to realize the energy scheduling of power system.
出处
《科学技术与工程》
北大核心
2014年第4期74-79,共6页
Science Technology and Engineering
基金
国家863高技术基金项目(2013AA04070)资助