基于输入延迟支持向量机的氮气管网压力预测

发布时间：2018-04-24 19:26

  本文选题：因果关系 + 影响因素延迟　； 参考：《信息与控制》2016年06期

【摘要】：在钢铁企业能源系统的低压氮气使用过程中,由于氮气使用单元分散且在管网中的位置不同,对管网压力的影响会出现短时间的延迟.鉴于此种情况,本文提出了一种基于影响因素输入延迟的多核最小二乘支持向量机对管网压力进行建模预测.该方法首先对低压氮气压力影响因素的延迟时间进行确定,提出一种基于因果关系的影响因素延迟时间计算方法,同时根据不同的影响因素和对应的延迟时间分别构造训练样本,进而建立基于最小二乘支持向量机的预测模型.通过对某钢铁企业现场低压氮气管网压力的两种不同情况,即正常工况和超限工况分别进行建模仿真验证,说明了本文提出的方法在压力预测上具有较高的精度.
[Abstract]:In the process of low pressure nitrogen use in the energy system of iron and steel enterprises, the influence of nitrogen gas use units on the pressure of pipeline network will be delayed for a short time because of the dispersion of nitrogen gas using units and the different positions in the pipe network. In view of this situation, a multi-kernel least squares support vector machine (LS-SVM) based on the influence factor input delay is proposed to model and predict the pipe network pressure. In this method, the delay time of the influencing factors of low pressure nitrogen pressure is first determined, and a method of calculating the delay time of influencing factors based on causality is proposed. At the same time, the training samples are constructed according to different influencing factors and the corresponding delay time, and then the prediction model based on least squares support vector machine is established. By modeling and simulating two different situations of pressure of low pressure nitrogen pipe network in an iron and steel enterprise, that is, normal working condition and over-limit working condition, it is proved that the method proposed in this paper has high precision in pressure prediction.
【作者单位】： 大连理工大学控制科学与工程学院;
【基金】：国家自然科学基金资助项目(61304213,61473056)
【分类号】：TF083;TP181
，

本文编号：1797910