【作者】 田亮；

【导师】 刘吉臻；

【作者基本信息】 华北电力大学（河北） ， 热能工程， 2005， 博士

【摘要】 许多火电机组都面临协调控制系统优化这一课题,相关研究工作主要集中在对象建模和控制算法两个方面,而建立机组负荷-压力对象模型、深入分析其特性是这一工作的基础。本文选择的研究方向为:建立当前应用最为广泛的300~600MW 级、采用直吹式制粉系统亚临界压力一次中间再热自然循环汽包锅炉、凝汽式汽轮机单元机组、50%~100%不投油助燃负荷范围内的负荷-压力对象简化非线性动态模型,为协调控制系统的设计提供依据。建模过程包括:机理建模、模型简化、参数确定、实验验证和特性分析几个方面。进行的主要工作包括: 研究单元机组各个组成部分的特点,做出合理简化,依据基本物质平衡和能量平衡关系建立各个部分机理模型,在机组运行参数附近对其进行简化拟合,获得具有解析形式的模型表达式。并且在同一类型不同容量的达旗330MW、神头500MW、盘山660MW机组上多个负荷-压力工作点进行燃料、汽轮机调门扰动实验和模型与机组运行数据长期对比实验,验证模型的准确性,同时总结出一套有效的求取模型参数的方法。在现有模型基础上,重点对制粉过程和过热器差压特性进行分析。将制粉动态和水冷壁动态合并简化为一阶惯性加纯迟延过程,在能够反映机组主要动态的前提下降低了模型阶次;考虑到蒸汽在过热器内是吸热膨胀的过程并且有减温水喷入,对其差压与流量的非线性关系进行了修正,并以锅炉吸热量代替主蒸汽流量建立模型;还对本模型与?str?m 模型的差别进行分析,指出?str?m 模型中压力的9/8 次幂随压力升高逐渐减小并趋近于1。以机理分析为基础,结合扰动实验及大量运行统计数据分析,得到了非线性模型中主要参数随工作点变化的规律,为:燃料指令增益随机组负荷升高而增加;锅炉蓄热系数随汽包压力升高而减小;制粉惯性在较大范围内随机变化。利用这些结论对模型进行修正,可以更为真实地反映机组实际情况。对非线性模型进行线性化,得到不同负荷-压力工作点下对象的线性模型,将对象系统非线性转化为线性模型参数时变特性,以便于用传统方法分析。线性模型传递函数中包含输入变量,表现为对象增益和惯性随工作点不同而变化。文中给出了此变化的数学描述,并通过330MW 机组的扰动实验验证了其正确性。文中还结合模型分析了协调控制系统设计中存在的困难,结合模型比较了一些理论方法的优点与不足。着重分析了解耦设计、PID 参数鲁棒整定、发热量校正、凝结水节流这些工程技术上的实用方法。更多还原

【Abstract】 Lots of power plant faced the problem of coordinated control system optimization. The correlative study mainly focused on the two aspects: object modeling and control arithmetic. Constructing the unit load-pressure object model and thoroughly analyzing the characteristic of the model are the basic of the study. The study direction of this his paper is that constructing current most far-ranging unit plant model. The unit plant is 300–600MW、direct firing pulverizing system、subcritical pressure、first medium reheating、natural circulation drum boiler and condensing steam turbine unit plant. And the model is load-pressure simplified nonlinearity dynamic model on the 50%-100% load scope without infusing oil to assist combustion. The model in this paper is the basis of CCS design. The modeling process include mechanism modeling、model simplifying、parameters confirming、experiment validating and characteristic analyzing. The main task of this paper include as following: Studying the characteristic of every members of unit plant the model is simplified reasonably, then according to basic substance balance and energy balance every members mechanism model are construct. Nearby the unit running parameter the model are simplified and fitted, then the analytic model expression ware got. There are three same type and different capacity units, DaQi 330MW、ShenTou 500MW and PaShan 660MW. On the three units the experiments of fuel disturbance and turbine valve disturbance on multi load-pressure running point are made. Comparing the model and units running data the model veracity is validated. And a set effectively acquiring model parameters method is summarized. On the basic of existing model the pulverizing process and the characteristic of superheater differential pressure were analyzed. The pulverizing dynamic and water wall dynamic are united and were simplified to be a one-order inertia plus pure lag process. The order of model was depressed in the precondition of reflecting the main dynamic of unit. The steam is decalescence and swelling process in the superheater pipeline, and there is attemperation water sprayed. Revising the nonlinearity of the differential pressure and the flow, and the main steam flow was replaced by boiler decalescence to construct model. The difference between this model and ?str?m model were analyzed, then the rule was presented that 9/8 power of the pressure in ?str?m model decreased with pressure increasing and will become 1. On the basic of mechanism analysis, combining disturbance experiment and large running statistical data, the rule of the main parameters of nonlinear model changing with running point was gained. That is fuel instruction gain increased with unit load increasing. Boiler lumped storage coefficient decreased with drum pressure increasing. Pulverizing dynamic inertia time change randomly on a big scope. Using the conclusions to revise the model can reflect the practical thing of the unit more truly. Through linearizing the nonlinear model the linear model on the different load-pressure running point are gained. The nonlinearity of the model can translate into the characteristic that model parameters changing with time. So we can analyze the system with traditional method. The linear transfer function includes input variable. The representation is that the object gain and inertia change with the different running point. The mathematical description is given in this paper, and through the disturbance experiment of 330MW unit the correctness of the model is validated. Using the model the difficulty of the current CCS design are analyzed, the merit and shortage of some theoretical method are discussed. Some engineering applied methods, such as: decoupling design、PID parameters robust tune、heat value emendation and water of condensation throttle are analyzed.更多还原