史密斯预估器matlab仿真,（毕业论文）史密斯预估器设计.doc

(毕业论文)史密斯预估器设计

摘 要

大迟延对象的控制一直是控制领域研究的焦点问题。加热炉温度控制便属于这类复杂的控制对象。传统的加热炉温度控制系统采用的是原料油出口温度同燃料油流量或同炉膛温度的串级控制，但由于燃料油流量存在波动，使得温度控制效果较差。而且由于近年来炉膛改造，炉膛容积增大，使得控制系统主副被控对象均存在较大的时间滞后。对于无滞后或滞后比较小的系统，通常采用PID控制。对于大滞后系统，PID控制效果并不好，需要另加补偿，因此提出了Smith预估补偿控制系统。而 Smith 预估算法则在模型匹配时具有好的性能指标 ,但是由于这种算法严重依赖模型的精确匹配 ,而在实际中这是很难做到的 ,当模型失配时,Smith 预估算法就难以取得良好的控制效果,因此提出了改进型Smith控制系统。

本文研究的重点是设计与实现适用于燃烧控制过程的控制规律和控制算法。具体讨论了纯滞后系统的Smith预估器及工程实现方法，着重对这种控制算法进行了较深入的讨论，并提出了一种改进型Smith预估控制器，该控制器把自适应控制与史密斯预估器有机地结合起来，实现对控制系统的参数自整定，而且还通过仿真对设计和改进的结果进行了分析。仿真实验中，若采用PID控制算法，系统会出现较大的超调量，采用史密斯预估补偿控制没有超调量。若保持控制器和模型的参数不变，改变对象参数，使估计模型与之失配，此时史密斯预估算法出现振荡，系统稳定性被破坏。改进型Smith算法不仅能够保持系统的稳定，而且振荡次数少，收敛速度快。

关键词:加热炉；增益自适应；史密斯预估器

Abstract

Large delay time has been the focus of research in the control field. Furnace temperature control is belong to this complex control object.Traditional furnace temperature control system uses raw oil outlet temperature with the fuel oil flow cascade control or furnace temperature, but due to volatility in fuel flow, making the temperature control less effective. In recent years, due to the furnace chamber volume increases, the control system exists in a large time lag. For no or less delay, the system usually adopts PID control. For large delay system, PID control effect is not good, need additional compensation. Thus puts forward Smith-predictor compensator control system. The Smith-predictor in the matching model has good performance when, but this algorithm heavily depends on the exact match model, but it is difficult to do in practice. Smith-predictor algorithm is difficult to obtain good control effect when the model mismatch. Thus put forward an improved Smith-predictor control system. This stduy focuses on design and realize the control laws and control algorithm that is applicable to furnace combustion control process. Specific discuss the method of Smith- predictor engineering realization for pure time-delay system, focusing on such a control algorithm for a more in-depth discussion, and puts forwar