基于神经网络与遗传算法的炼焦单元能耗优化方法.docVIP

硕士学位论文 摘要 炼焦单元是钢铁生产流程中的重要生产和耗能单元，利用智能方法对其能耗进行分析和优化，改变目前仅靠人工经验实现能源调配和过程状态参数设定状况，可为炼焦单元的科学用能提供指导，降低能耗。 通过对炼焦单元能耗问题的分析，本文从能耗预测和COG、BFG配比对炼焦单元能耗影响两个方面展开研究。针对炼焦单元能耗预测问题：首先对炼焦单元的能耗影响因素进行分析，定性给出其原材料参数和过程状态参数；然后采用灰色关联度方法计算这些参数与炼焦单元总能耗的关联程度，选择与炼焦单元能耗关联度较大的因素进行能耗预测；最后利用人工神经网络挖掘出炼焦单元原材料参数、炼焦单元总能耗与炼焦单元能源产消量、关键过程状态参数的关联关系，利用该模型对炼焦单元的历史能耗进行检测，效果分析表明所建模型满足预测要求。针对COG和BFG配比对炼焦单元能耗影响分析问题：首先利用人工神经网络挖掘出炼焦单元中COG和BFG配比与能源产消量的关联关系；然后利用该关联关系建立炼焦单元能耗优化模型；最后通过遗传算法求解得到优化的工序能耗和相应的COG和BFG配比。 关键词：炼焦单元，能耗影响因素，能耗预测，能源替换关系，能耗优化  ABSTRACT The coking unit is an important sector in the steel production process and is also an energy-consuming unit, energy efficiency is analyzed and optimized using intelligent methods, the current situation in energy deployment and process parameters setting can be changed for guidance in energy deployment and lowing energy consumption. Through the analysis of energy efficiency in coking unit, this essay focused on two aspects: one is the prediction of energy; the other one is the designation of COG and BFG; To the prediction of coking unit: the coking process was firstly introduced, and the mechanism analyses which are influencing the coking unit are discussed. Then the gray incidence degree (GID) between these parameters and energy consumption of coking unit was calculated using gray relation method. The numbers of GID above 0.5 are selected as the input and output. Finally the parameters of washed coal, the energy consumption of coking unit, the consumption and production of different kinds of energy and the key process parameters are used to set a BP Artificial Neural Network (ANN) model and the model was examined by historical data, and the result was met up with the demands. To solve the analysis of energy consumption of coking unit using the replacement of COG and BFG: the substitutability of COG and BFG to the influence of coking unit was analyzed qualitatively; then the BP ANN was used to set up the relation between the proportion of BFG and the consumption