摘要
以非接触式指尖密封结构为对象,采用CFX建立了非接触式指尖密封共轭传热数值模型,计算了动压靴的温度分布,分析了气体黏性对动压靴温度分布的影响,并进一步采用ABAQUS软件进行了非接触式指尖密封热变形的数值模拟,分析了气体黏性对热变形的影响.结果表明:气体黏性引起高速旋转气体与动压靴底部摩擦生热,造成动压靴温度分布不均匀且局部温度升高;随着温度升高,非接触式指尖密封的热变形程度增大;考虑气体黏性作用时,非接触式指尖密封产生更大的热变形.
Taking the 1 000 MW ultra-supercritical unit as an object of study,artificial neural network prediction models with high accuracy and good dynamic characteristics were established for the unit load and main steam pressure in consideration of its regenerative cycle system.Subsequently,an optimized intelligent predictive controller was proposed for the coordinate system,which was used to optimize the openings of deaerator water level control valve and steam turbine control valve,and to control the total fuel demand based on above prediction models under variable load conditions,so as to improve the coordinated control effect.Detail simulation tests were conducted on the optimized coordinated control with a full-scope simulator for the given 1 000 MW USC power unit.Results show that via the method,both the response speed of dynamic load and the load control accuracy can be effectively improved,with significant reduction in control deviation of the main steam pressure under varying load conditions,proving the method to have good practicability.
出处
《动力工程学报》
CAS
CSCD
北大核心
2017年第8期622-628,共7页
Journal of Chinese Society of Power Engineering
关键词
非接触式
指尖密封
气体黏性
温度分布
热变形
ultra-supercritical power unit
neural network
prediction model
condensate throttling
coordinated system
intelligent control optimization