摘要
烧结终点预报对于提高烧结矿强度和产量、降低能耗具有重要意义,但是烧结终点状态受多种因素影响,无法直接检测,只能由操作工依据经验进行判断,严重影响了烧结生产的稳定运行。本系统运用K均值聚类分析的样本优选方法对海量数据进行处理,选择具有代表性的样本,从而有效缩小样本空间、改善样本质量。使用风箱温度曲线计算废气温度上升点和烧结终点软测量值,以台车速度和点火温度作为输入,采用BP神经网络模型,对烧结终点位置进行预报。在实际应用中,该模型预报结果较准确地反映了烧结终点位置的变化,起到了稳定生产、节约能源的作用。
The prediction of burning through point（BTP） is important to improve the strength and yield of sinter and reduce the energy consumption, but the BTP state under the influence of various factors can not be detec- ted directly, only be judged by the operator on the basis of experience, which has a strong impact on the stable operation of sintering production. This system uses optimized sample selection method based on K-means clus- ter analysis to process the massive data, select the representative samples, thus the sample space is effectively shrunk, the quality of samples is improved. Using the bellows temperature curve to calculate the burning rising point（BRP） and BTP soft measurement value, taking pallet velocity and ignition temperature as input, BTP po- sition is predicted by using BP neural network model. In practical el more accurate reflect the change of BTP, which plays a role in application, the predicted results of the mod- stabilizing production and saving energy.
出处
《测控技术》
CSCD
2017年第3期86-89,共4页
Measurement & Control Technology
关键词
烧结终点
废气温度上升点
BP神经网络
K均值聚类
样本优选
burning through point
burning rising point
BP neural network
K-means cluster
optimized sampleselection
