摘要
为了在延迟焦化塔中制取高密度沥青焦,首先需要对煤沥青进行热处理以除去不溶物质和进行组分调制.采用管式炉、真空闪蒸塔组织沥青改性处理过程,数字化控制系统监控整个工艺流程,通过合适的神经网络结构和训练算法对重闪蒸油产出率及相关工艺参数之间的关系进行辨识,并应用稳定性理论对网络误差、隐层个数以及网络收敛趋势进行讨论.研究表明:适当提高管式炉出口温度可以增加重闪蒸油的产量;过高的温度控制会使物料黏度急剧上升,导致热处理条件恶化;有限量参数的采集和非线性曲线拟合可以弱化原始数据的随机性,并且通过模型的不断修正而收敛到满意的精度,从而在较短的时间内搜索到最佳热处理条件.
In order to produce the high density pitch coke in delayed coking tower, the coal tar pitch should be pretreated so that the insoluble matter can be eliminated and components can be modulated. The process of coal tar pitch modification was organized by means of the tube furnace and vacuum flash distillation tower, meanwhile digitalization control system was adopted to monitor all round process flow. Relationship among yield rate of heavy flash oil and other related parameters were identified by appropriate network structure and the net train algorithm, the network error and the number of ulterior layer and the convergence tendency of network were discussed applying for the stability theory at the same time. The results show that raising the outlet temperature of tube furnace properly can increase the yield of heavy flash oil, but too higher temperature can increase the viscosity of the material and lead to heat treatment condition degenerates rapidly. The limited parameters＇ sampling and the nonlinear curve fitting can weaken the randomization of the original data and it will converge to the satisfactory precision through model modifying uninterruptedly, in this way the optimal thermal treatment condition can be searched in a short period.
出处
《煤炭学报》
EI
CAS
CSCD
北大核心
2009年第12期1687-1692,共6页
Journal of China Coal Society
关键词
煤沥青
热处理工艺
沥青焦
收敛
coal tar pitch
thermal treatment technology
pitch coke
convergence
作者简介
张中明（1959-），男，上海人，高级工程师．E-mail：zhangzhongming-8@163．com
