Vous êtes sur la page 1sur 3

2011 International Conference on Mechatronic Science, Electric Engineering and Computer August 19-22, 2011, Jilin, China

Application of Fuzzy Controller in Drum Water-level Control


ZHuo Wang
School of Electric and Information Engineering Beihua University Jilin, China wwzz_6666@163.com
AbstractConventional Proportional Integral Derivative (PID) control is often used in present industrial field. Because the control parameters are fixed, the tuning results are difficult to meet the demands and the system is unstable or even out of control. In order to solve these problems, fuzzy control technique is presented to meet the demands of intelligent controlling. In this paper, Matlab is used to confirm that the fuzzy control system has good adaptability and anti-jamming capability. Thus, it is proved that the control methods is successful, meanwhile, it is valid for the problems about analyzing and applying for boiler waterlevel control. Keywords- Drum water-level of boiler ; PID control ; Fuzzy control ; MATLAB

Qiang Wang
School of Information and Technology Beihua University Jilin, China

II.

DESIGN OF DRUM WATER-LEVEL FUZZY CONTROL SYSTEM

A. Design of fuzzy controller Confirm the input/output variables of fuzzy controller The conventional 2-D fuzzy controller is used in fuzzy control system of boiler drum water-level. Input variables denote the deviation and the value of deviation change. e(k ) = yr ( k ) y (k ) (1) (2) ec ( k ) = e( k ) e( k 1) Equation 1 denotes the water-level deviation, where yr(k) means given water-level. When y(k) is higher than yr(k), actual water-level is higher than given water-level and the deviation is negative. The higher the actual waterlevel, the greater the negative deviation; Instead, when y(k) is lower than yr(k), actual water-level is lower than given water-level and the deviation is positive. The lower the actual water-level, the higher the positive deviation. Equation 2 denotes the change of water-level deviation. If the water-level at point k is higher than the value at point k-1, it means that the water flow variation is increased. In other words, actual water-level is higher than before and the deviation is negative. Output of the fuzzy controller equals to the variation of valve opening u(k) which directly effects the change of water-level y(k). Variable of output control depends on the change of valve opening: the bigger the valve is opened, the bigger the injected water flow and vice versa. Fuzzification of input/output variables During fuzzy control, above input/output parameters should be expressed as fuzzy subset of fuzzy language parameters. Suppose fuzzy set E denotes the input deviation, i.e E = {HB, HM, HS, NO, LS, LM, LB} Where, H (high) means that actual water-level higher than given water-level, L (low) means that actual waterlevel lower than given water-level, HB means big high, HM means middle high and HS means small high. Likewise, LS, LM, LB and NO mean different water-level. Then, fuzzy region corresponding to input deviation expresses as E = {-8 -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8}

I.

INTRODUCTION

Boiler is important power equipment which widely used in industrial production and daily life. Because of big change in load, frequently starting and stopping, it could not long-term work safely and stability in economical operation mode through manual operation boiler. Drum water-level control system is one of the most important parts in boiler control system. Drum water-level is an important index to ensure the safety and stable operation for industrial boilers and the main controlled aim of automatic control system. Keeping water-level in a given range is importance for enhancing steam quality, reducing wear and tear of equipment and operation cost[1]. During the boiler running process, there are amounts of random interference factors, i.e. load impact, the distribution voltage wave and the variation of water supply, air supply and oil supply. Even if the water supply flow and evaporation flow maintain a balance, the drum water-level is possible to change. In addition, instability phenomenon of fuel is subjected to the quality of the fuel, which made the drum water-level has complex dynamic properties. Currently, conventional PID control method is often used in most of the drum water-level control. However, PID control parameters are fixed and disturbance which almost would not find out the various control parameters be adapt to. Hence, its control effect is difficult to meet the requirements, causing the system unstable or even out of control. Fuzzy control method was introduced in this paper, which would enhance the anti-jamming of the boiler system, improve the control effect and keep water-level stability[2].

978-1-61284-722-1/11/$26.00 2011 IEEE

174

Suppose fuzzy set of fuzzy language parameters EC denotes the change of input deviation, i.e EC = {NB, NM, NS, ZO, PS, PM, PB} Corresponding fuzzy region is EC ={-8, -7, -6, -5, -4, -3, -2, -1, 0, +1,+2, +3, +4, +5, +6, +7, +8} Suppose fuzzy set of fuzzy language parameters U denotes the change of controlled variable, i.e U = {CF, CM, CS, Freez, OS, OM, OF} Where, C means valve closed, CF means close valve fast, O means valve open, Freez means keep valve in regular opening. Corresponding fuzzy region is U = {-8 -7 -6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 +7 +8} Determine the membership function of linguistic Choose membership function of input/output variables as trigonometric function and input fuzzy variable build relationship through fuzzy rule and output fuzzy variable [3]. Language description of fuzzy control rules As a rule of thumb, control rules could be expressed with the following fuzzy conditional statements: if E = HB and EC = NB, then U = CF if E = HM and EC = NS, then U = CM if E = HS and EC = PM, then U = OS if E = NO and EC = ZB, then U = Freez The whole control rules are presented in the following state table of fuzzy control.
TABLE 1. STATE TABLE OF FUZZY CONTROL U E HB EC NB NM NS ZO PS PM PB CF CF CF CM CM CS Freez CF CF CM CM CS Freez OS CF CM CM CS Freez OS OM CF CM CS Freez OS OM OF CM CS Freez OS OM OM OF CS Freez OS OM OM OF OF Freez OS OM OM OF OF OF HM HS NO LS LM LB

When the fuzzy control rule is considered, rule of control variables selection principle is, for big or other error situations, choose control variable has mainly focused on eliminate error as soon as possible; for small error situations, based on system stability, choose control variable should be pay attention to avoid overshoot. Hence, when EC is negative, there is an upward tendency in water-level. If the water-level is high, valve should be closed to decrease the water injection; otherwise, valve should be opened. In view of the above, 49 rules were received in Table1. B. Design and simulation of fuzzy control system Operators need three procedure parameters to control drum water-level deviation between actual drum waterlevel, variation of confluent and steam flow. According to above parameters, operators observe water-level, evaluate the equilibrium relation between water injection and steam flow, and then operate the master control injection valve to increase/decrease water volume. Above empirical process of ideation of human being should be reflected by designing of fuzzy controller [4]. The experience form previous experiments showed that regulating drum water and changing steam flow should be at the same time. When designing fuzzy control system, it just considers the relationship between steam flow and variation of water level in system adjusting structure, i.e. just reflects double impulse control system, and the estimation and adjustment of feed water flow was not considered in drum level major loop system. According to this analysis, fuzzy control system as shown in Fig. 1 was built. In this system, because limited change of water-level under field situation could not change in the whole range of drum water-level, a limiting structure joined in step of water-level deviation. Suppose the change of amplitude in the range of [-0.3 0.3]. Inputs of Mux1 are water-level deviation and variation of water-level deviation, Ke is the quantification factor of water-level deviation, Kc is the quantification factor of variation of water-level deviation, Ku is the quantification factor of output of fuzzy controller. Inputs of Mux2 are the system given signal and system output signal, the track response ability is easy to observe through the results of numerical simulations by using oscilloscope.

Figure1.Structure of fuzzy control system.

175

Because the value of Ku, which is the output scaling factor of fuzzy controller, influences greatly the system performance, it is able to meet system fast response requirement with larger Ku. System response curve is received and shown in Fig2 at the time of keeping the value of K =1, Kc=10, Kh=40. At this moment, there is no overshoot, have small setting time and response time, but remains about 2.5% static error in system [5]. It can be observed form Fig.2 that the properties of fuzzy control system are improved obviously. Moreover, it just needs to adjust two parameters to achieve better performance after establishing fuzzy rule.

Figure 4.Conventional system response graph with interference.

III.

CONCLUSION

Figure2.Fuzzy control system response graph.

C. Anti-jamming capability of fuzzy control system Add small amplitude sinusoidal interference signal as shown in Fig3 response curve of conventional fuzzy control system is represented in Fig4.

Through simulation curves, it could be indicated the effectiveness of fuzzy control strategy, which is based on fuzzy control theory and controls the drum water-level system of boiler. It has short transient time, small overshoot and high anti-jamming capability. Meanwhile, it would overcome the effect of steam flow disturbance on drum water-level which full reflects the characteristics of fuzzy control. During running process, even though there have some problems such as steady-state error and small fluctuate, they could be settled completely through further increasing the number of input/output fuzzy subset, modifying the fuzzy control query table and working point. Obviously, fuzzy control method has a good application prospect in drum water-level system. REFERENCES
[1] Tao Wen-zhen, ZHou Xi-zheng, Analysis and application of boiler drum water level measurement and control system, Automation & Instrumentation, vol.24No.12, pp. 44-47. 2009. Wu Ming-yong, Wang Guo-wei, Intelligent control system of boiler drum water level based on fuzzy adaptive PID, Industrial Boiler, vol.6No.5, pp. 28-32. 2010. Luo ZHi-yong, Application of fuzzy-pid control in drum water level control of industrial boilers, Computer Applications and Software vol.4No. , pp. 200-202. 2009. Lou Wei, Liu Xiang-dong, The application for fuzzy control in the boilers drum water-Level control system, Boiler Technology, China, vol.1No.1, pp. 27-31.2010. Ge Wei, ZHu ZHang-qing, Fuzzy self-tuning PID control of the boiler drum level, Techniques of Automation and Applications, vol.36No.6, pp. 22-24. 2008. Xue Li-li, SHen Xiao-qun, The design and simulation of boiler drum water level based on fuzzy control, Science & Technology Information, vol.2No.11, pp. 124-125. 2010.

[2]

[3]

[4]

[5] Figure3.Interference signal with small amplitude.

Compare Fig2 with Fig4, it is shown that conventional fuzzy system has minor fluctuation around desired value and the overshoot of fluctuation is in the range of permitted errors. However, there is little change in rise time and accommodation time. All these facts suggested that fuzzy control system posses better dynamic performance and good robustness [6].

[6]

176