节点文献

火电厂热工过程优化控制策略及应用研究

Optimization Control Strategies and Application Studies in Thermal System of Power Plant

【作者】 刘建民

【导师】 韩璞; 开平安;

【作者基本信息】 华北电力大学(北京) , 热能工程, 2009, 博士

【摘要】 针对火电机组单机容量和参数不断提高,热工过程和系统更加复杂,过程非线性、时变性、大迟延、大惯性、强耦合性和不确定性越来越严重,建模十分困难的情况,对适应于大型火电机组热工过程控制系统的优化策略进行了研究,以解决目前优化控制中的一些难点问题,提高火电机组自动化水平,确保大型火电机组运行的安全性、经济性和可控可调性。提出了控制系统“时间尺度”的慨念,并结合工程实践积累的经验,给出了一套PID控制器参数工程整定方法。该方法完全摆脱了被控对象数学模型的限制,现场运用简单、有效。对二阶系统的特性进行了深入的分析,应用梯度最速下降方法进行系统辨识,给出了二阶系统参数变化扰动的补偿公式和基于此公式设计的[Opt优化策略]。构建了一种基于经典力学匀加速运动方程的系统状态观测器OD。证明了该观测器在合适选择参数后是一致渐近稳定的,并且是对完全可观系统位置、速度和加速度的无偏估计。并在此基础上设计了运用OD组态的控制系统SCOD,其能够自动消除内、外干扰的影响,鲁棒性能良好,具有工程实用价值。针对确定性大时滞对象,给出了PID控制系统参数的另一种工程整定方法,即参数整定两步法。该方法只需要调整控制器的一个前置系数,即可克服系统纯滞后的影响,使得大时滞控制系统PID参数整定简单化。针对时变大时滞对象,给出了两种改进设计方法。一是在史密斯预估控制系统中加入了[Opt优化策略],这个优化策略能够在被控对象模型未知的情况下直接补偿系统,消除系统参数不确定因素和干扰的影响。二是通过把大时滞系统分解成多个小时间常数的一阶系统,设计了一种基于极点配置的状态观测器,并基于此构建了控制系统,此法可以有效消除系统的动态偏差,提高系统的控制品质。基于系统动力学特性,设计了一种通用控制器。指出鲁棒控制器和内模控制器虽然设计方法和理论不同,但它们的传递函数与本文设计的通用控制器等价,在经典动力学意义上具有统一性。并根据统一性分析,对二阶系统进行状态扩展,引入加速度控制项,改进了二次型优化算法,求解了PID控制参数。提出了一种鲁棒控制器加权函数的通用程序化设计方法。该方法使控制工程师像设计PID控制器一样去设计鲁棒控制器,不再去设计复杂的加权函数。同时设计了一种非线性PID和自校正PID控制器,改进了普通线性PID控制器的不足。对非线性控制系统设计理论上的难点问题进行了探讨,就非线性控制系统的反馈线性化、系统间隙度和非线性系统的优化控制进行了分析。

【Abstract】 Developing with the tendency to large capacity,high parameters and high efficiency of ssil-fired unit,the most thermodynamic objects have large inertia and delay,and are nonlinear and slow time-variable;some of them are multi-variable has large uncertainty and strong coupling.This makes it difficult to construct accurate mathematic models,and makes PID controllers with fixed parameters unadapted to higher and higher control request in production.So the optimization control strategies of thermodynamic objects are studied in the paper to improve the control performance.After the concept of "time scale" about control system is presented,the method of engineering setting PID controller parameter designed which can complete overcome the limitation of the controlled object.And the Optimization Control Stratedies(Opt) is designed on the basic of analysing the characters of time-domain and frequency-domain of second order system.A state observer(OD) based the classical dynamic equation is designed,it is proved that the OD is correspondence,asymptotic and steady,and the SCOD control system is designed on the basic of OD,which can not only cancel the inner and outside disturbance of system,but also obtain better robust quality.Aiming at the time-unvarying delay control system,the two-step method of engineering setting PID controller parameter is designed,which can overcome effectively the influence of time-delay by adjusting an advance coefficient of PID,this makes adjust PID controller parameter of time-delay control object simplity.Aiming at the time-varying delay control system,two kinds of improved methods are presented.One is the improved Smith-predictor,which may direct compensate the system and cancel the influence of uncertainty factors by adding the Opt optimization strategy.The other is a state observer based the pole assignment,which can avoid effectively the dynamic decination and improve control performance by decomposing the time-varying delay system with several small time constant first-order system.A General Controller is designed on the basic of classical dynamics equation; which can make the controlled plant output is controlled and the desired output can be obtained by using feedback control of position,velocity and acceleration of the General Controller.And the paper analyses that PID controller,state feedback controller,robust controller,inner model controller and GPC/MPC controller are identical in classical dynamics.The programmed designing method weighting function in robust controller designing is presented,which can make the robust controller designing simplity and utility.And the non-linearity PID and self-correction PID controllers are designed, which can improve the shortcoming of PID controllers and the robust performance of system.The feedback linearization,system lacunarity and optimization control of the nonlinear control system are discussed.

  • 【分类号】TP273;TM621.6
  • 【被引频次】6
  • 【下载频次】573
  • 攻读期成果
节点文献中: 

本文链接的文献网络图示:

本文的引文网络