【作者】 张峰；

【导师】 刘教瑜；

【作者基本信息】 武汉理工大学 ， 检测技术与自动化装置， 2009， 硕士

【摘要】 目前,国内外大多数的火电厂的主汽温度控制仍采用传统的PID串级控制。众所周知,火电厂是调峰电厂,且随着随着单元机组容量的不断扩大,从而对自动化程度的要求很高,而主蒸汽温度是电厂最重要的性能参数,所以保持主蒸汽温度的控制精度对电厂的安全、经济、稳定的运行具有很重要的意义。常规的串级PID控制系统很难在非线性、大惯性、大滞后及时变性的工况下均保持优良的控制品质。超临界锅炉由于没有汽包,只有汽水分离器,汽水分离器所储存的水容量很小,所以启动速度快,锅炉各个部分的耦合型强:由于其没有汽包,下降管的数量也大大下降,所以整台锅炉的钢材耗量低,锅炉成本也低。其蒸汽压力、温度性能参数高,使发电效率有很大的提高,锅炉的供电煤耗较大的降低。正因为超临界锅炉具有以上特点,所以超临界锅炉是中国未来大型锅炉的发展趋势,600MW机组锅炉又是现在的超临界锅炉的主要型号,所以深入研究并掌握超临界600MW机组锅炉的控制技术是十分重要的。本文针对阿尔斯通武汉锅炉有限公司600MW超临界直流锅炉,设计出一种新型的主蒸汽温度控制策略。模糊神经网络是将人工神经网络与模糊逻辑系统相结合的网络,是智能控制理论研究领域中一个十分活跃的分支。这种策略利用模糊控制和神经网络控制等控制理论,结合传统的串级PID控制系统的基本结构,把模糊神经网络控制系统良好的动态特性体现在控制系统的主控制器的设计上。主控制器采用自适应模糊神经网络控制器,其中模糊规则是由典型模糊控制规则表确定的。副控制器采用PI控制器,副控制器和减温器系统及反馈回路一起构成副调节回路。主控制器和过热器系统以及反馈回路一起构成主控制器回路。上述主控制回路和副控制回路一起构成本文的600MW机组锅炉的智能控制策略。最后结合MATLAB的仿真工具,对本文所设计的智能控制系统和传统的串级PID调节系统进行比较,分别在几种典型的工况下进行仿真图形的比较,分析其他们的控制特性,发现本系统的控制品质明显高于传统的串级调节系统,具有很强的理论和工程价值。更多还原

【Abstract】 Currently,the main steam temperature control system in most domestic and overseas power plants still adopts the conventional PID cascade control scheme.As we know,the thermal powers are the regulate-peak power.The larger capacity and the higher control demand of the parameters the thermal power Turbo-generator unit has, so it has high demand to the automatic.The main steam temperature is the most important performance parameter,the good control precision has large importance to the power plant’s run of the safety,enconmy and stabilization.The worse control performance the conventional PID control scheme gets under different loads, nonlinear,large inertia,large delay and time-varying work condition.The supercritical once-through boiler only has separators but of drum.The Boiler has littler water capacity than the drum-boiler so it has high start-up speed,and the different section have the large coupling.It didn’t have the drum,and has littler downcomer,so it has littler metal consumption and lower cost.Because of the high pressure and temperature performance parameter,it has higher power-generation efficiency and lower coal consumption.The supercritical once-through boiler,which is the developmental trend of domestic large capacity boilers,has advantages of front applicability.The 600MW supercritical once-through boiler is the main type boiler. So it is very important to research and master the control technology deeply.I designed the new main steam control scheme to the 600MW once-through boiler of ALSTOM-WuHan Boiler Co.,Ltd in the paper.The Fuzzy-Neural Network is the unit of Fuzzy Control and Neural Network control.It is very active ramus of the intelligence theory reseach.Applying the Fuzzy Control and theory,the scheme in the paper combine the conventional cascade control’s basic structure,and add the good dynamic performance of the Fuzzy-Neural Network control system to the design of the main controller.The main controller is composed of a Adapt Fuzzy-Neural Network controller.Moreover,the fuzzy controller’s suitable fuzzy rules are confirmed by the fuzzy control rule table.The second controller is the PI controller.The second control loop makes up of the second controller,the de-superheater system and feedback part.The main control loop makes up of the superheater system,the main controller and feedback part.The main control loop combine the second control loop to the main Steam Temperature control strategy of 600MW Generating Unit Boiler of the paper.Finally,Combined the MATLAB toolbox and compared with the conventional cascade PID control scheme to the intelligence control system and compared the drawing in the several work conditions,and analyzed the performance.It is true that the system in the paper has a better control quality than conventional PID does.It has very large theory value and project value.更多还原