摘要
镁砂熔炼过程具有多工况、群炉并行生产、高能耗等特点.在全厂供电容量约束下,为了最大化能源使用效率,需要根据全厂每台炉子的工况变化实时分配电能,实现全厂镁砂单位能耗与平均品位的多目标优化.本文基于最小二乘支持向量机技术建立了镁砂熔炼过程全厂电能分配优化模型.根据不同工况下降低镁炉供电量对镁砂熔炼过程的影响程度,提出了基于工况优先级的电能分配策略.根据主熔工况下镁砂产量与品位指标函数的特性分析,推导出主熔工况下电能分配模型决策变量维数缩减的条件.为了提高多目标优化算法的运行效率,设计了一种快速非支配解集构造方法,用来提高传统多目标粒子群优化算法的寻优效率.基于标准测试问题与现场实际例子对所提出的方法进行了检验.基于现场例子的实验结果证明所提出的方法能够避免工厂出现的用电超容量情况,并且提高了全厂用电效率.
The process of smelting magnesia grain is multiple-modes parallel-operating and high energy-consuming. To maximize the energy efficiency within the constraint of power capacity, it is desired to allocate the limited amount of electric energy to a number of furnaces for optimizing the plant-wide yields and grade of magnesia. First the allocation optimization model is developed by the least squares support vector machine. The priority of operating mode and dimension reduction strategies are proposed to simplify the optimization problem. A novel ranking approach is devised to find the non-dominating set in a population efficiently. The approach is used to improve the reM-time performance of traditional multi-objective particle swarms optimization. The proposed approach is validated by a factual case.
出处
《自动化学报》
EI
CSCD
北大核心
2014年第1期51-61,共11页
Acta Automatica Sinica
基金
国家重点基础研究发展计划(973计划)(2009CB320601)
国家自然科学基金(61020106003
60904079
61134006)资助~~
关键词
电熔镁炉
维数缩减
非支配解集构造
多目标优化
Electric smelting magnesia furnace, dimension reduction, non-dominated set, multiobjective optimization
