A combustion control support expert system for a coal-fired boiler

更新时间:2023-05-27 04:14:01 阅读量: 实用文档 文档下载

说明:文章内容仅供预览,部分内容可能不全。下载后的文档,内容与下面显示的完全一致。下载之前请确认下面内容是否您想要的,是否完整无缺。

A COMBUSITON CONTROL SUPPORT EXPJBT'SYSTEM FOR A COAL-FIRED BOILER

Toshio Miyayama Shinji Tanaka Tatsuya Miyatake. Tetsuya UmeW: Energy Plant Engineering Division, Kawasaki Heavy Industries, Ltd. 4-25. 2-Chome Minamisuna, Koto-ku, Tokyo, 136 Japan Yuichi Miyamoto. Kimiyoshi Nishino Techno logy Development Group, Kawasaki Heavy Industries, Ltd. Eiichi Harada Akashi Technical Institute Kawasaki Heavy Industries, Ltd.

Abstract Recently coal-fired boilers a r e required with using many kinds of coals and t h e characteristics on coals are diversified. Coal-fired boilers have so complicated combustion processes t h a t t h e system has many f a c t o r s t o determine t h e i r characteristics. Therefore, t h e r e are a plenty of know-how and expertise f o r coal-fired boilers. This paper r e p o r t s a newly developed combustion c o n t r o l support expert system (K-MIND, KAWASAKI-Combustion Control M_ulti-ktelllgent Development Support Expert System) which aims t h e attainment of environmental regulation and high efficiency operation of t h e thermal plants by applying Artificial technology t o them. 1. Introduction A s coal-fired large-scale Intelligence (AI)

(3) Diagnosis by expert system f o r boiler and

combustion system(4) Automatic unburned carbon monitoring system

(5) Monitoring system using communication line2. Configuration of t h e combustion c o n t r o l

support expert system{I) Optimal combustion support using f u z z y logic

There are many kinds and wide ranges of coals t o be burned in boilers. O t h e o t h e r hand, low NO, emission. low n unburned carbon content in ash and low oxygen operation a r e required.@ NO, emission. Unburned carbon content in ash

boilers

a r e complicated and

and

systems,

suitable

effective

treatments have t o be performed from synthetic judgements based on a number of d a t a which obtained from analog and d i g i t a l sensors. Therefore, operation support expert systems generally cover equipment maintenance and diagnosis, d a t a analysis, improvement of c o n t r o l performance and so on. The operation support expert system which has been developed in combustion c o n t r o l has t h e following five main functions. (aOptima1 combustion support using fuzzy logic (2) Dynamic simulation f o r NO, emissionClassi I icr

Pig.1 Low-NOx combustion system ( t w o stasc firing 1

One of t h e environmental problems t o bee considered i s a restriction of NOx emission. W apply low-NO, type coal-burners, a two-stage combustion system, and a low oxygen combustion system. Fig.1 shows t h e low N x combustion system O

I ECON '91

1513

CH2976-9/91/0000-1513$1.00 01991 IEEE

f o r t h e coal-fired boiler. Fly ash discharged from boilers is sold as a material f o r cement, so t h a t unburned carbon content in ash has t o be reduced below a c e r t a i n level. The reduction of unburned carbon content in ash leads high efficiency operation of t h e power plants. Unburned carbon content in as

h changes widely with t h e fineness of pulverized coal t o be burned in furnace. The main f a c t o r t h a t determines t h e fineness of coal is t h e position o r rotation of t h e m i l l classifier.

W e

can

see t h a t

t h e relationship

changes

depending on t h e coal species. Skilled operators have qualitative feelings about t h i s relationship. However, as plant conditions change from time t o time, it is difficult to determine optimal operation conditions on t h e basis of synthetic judgement.

x

1°/1

@ The relation between process values and t h e factors Fig.:! shows t h e relation between plant s t a t e s such as N O x emission, unburned carbon content in ash. m i l l motor power, and c o n t r o l positions such as t h e r a t i o of two s t a g e combustion and t h e position o r rotation of t h e m i l l classifier. There a r e mutually interfered relations. For example, if t h e operator increases t h e two-stage combustion r a t i o during low oxygen combustion condition in o r d e r t o decrease t h e NOx emlssion, unburned carbon content in ash increases. And if, in o r d e r t o reduce unburned carbon content in ash, the operator a d j u s t s t h e classifier position and increases t h e fineness of coal, t h e m i l l motor power increases and t h e mill r o l l e r abrasion is adversely affected. These relationships change with coal properties, mill load and mill r o l l e r abrasion.

loo

20

30

40

TS ration

1461

Fig.3

NOx and unburned carbon content in ash

@ Optimization using fuzzy logic

In this system, t h e procedures of operation by experts and with past operation d a t a a r e supported by fuzzy logic, and t h e system can display quantitative guidance on t h e basis of t h e present operating point. The N O x emission, unburned carbon content in ash and mill motor power have certain respective control limits, so t h a t necessary manipulating

omission

Pig.2

The r e l a t i o n b e t w e e n process v a l u e s and t h e factors

@ Combustion database W have carried out many combustion t e s t s with e many kinds of coal, accumulated much experience and built up a database f o r t h e f e a t u r e of combustion processes. Fig.3 shows t h e relation between NOx emission and unburned carbon content in ash.

U n b u r n c d cnrbon cnrbon c o n l o n t in nEli

M i l l motor pcwor

1,

r

111

- Furzylosic

TS

rkxxl111

S'4 hlD CG

.

1 K i n d o f coal I K i n d o f coal

1 Coilcr dcmand I Coilcr dcmand

1

I

Fig.4 Fuzi..y c o n t r o l s y s t e m

1514

I ECON '91

variables such as two s t a g e combustion r a t i o and t h e position or rotation of t h e mill classifier which navigates them i n t o each c o n t r o l l i m i t a r e shown by t h e use of f u z z y logic. As f o r making rules. t h e r e are many operating variables in t h e database f o r every p l a n t condition and coal property, so t h a t they have been registered as member-ship functions. and t h e causality betwee

n operation and response has been t r a n s l a t e d into IF-THEN rules. Fig.4 shows t h e fuzzy c o n t r o l system. Applying s t r u c t u r i n g design methods f o r plant conditions, t h e efficiency of t h e system design has highly increased. For example, as shown in Fig.5, one case Is t h a t both t h e NO, emission and unburned carbon in content are in abnormal zones, and another case

Fie.6 Example of man-machine intellace of fuzzy logic

is t h a t t h e NO, emission is in an abnormal zone but t h e unburned carbon content is In a normal zone. According to t h i s s t r u c t u r i n g method, t h e finer t h e mesh s t r u c t u r e, t h e more optimal t h e system f specification is. Fig.6 shows t h e example o man-machine interface of fuzzy logic.

combustion ratio, air/coal ratio. There a r e learning functions of basic t a b l e f o r coal properties and parameters. If operators f e e l t h e difference between t h e model and process value. t h e dynamic parameter of t h e system is learned by operational d a t a and it t o be used in changing p l a n t characteristic. J3) Diagnosis by emert system for boiler and combustion system An example of t h e man-machine interface of t h e boiler and auxiliary equipment f a u l t diagnosis, i s showed i n Fig.7. If some abnormal conditions w i l l happen in boilers and auxiliary equipment, estimation of t h e cause is performed by t h e expert system using t h e d a t a obtained from sensors, and then recovery guidance f o r t h e estimation of t h e cause i s disp layed. Particularly, In t h e case of coal-fired boilers, t h e expert system is mainly applied t o

L.

dU U

Unburnod osrbon:Norms1 o i b o r n l i n t r o n o

CL

n a 3

Fi[:-5 1 l c l:~ l i o nb c l w c c n N O x cniission a n t 1 i i n b i i r n c i l c a r b o n c o n l c n t i n ash

J2)Dynamic simulation for N x emission O There is t h e problem t h a t NOx emission over a regulated value in t h e case of transient s t a t e . Especially, when a mill s t a r t u p and shutdown or switch over,NOx emission tends t o increase by t h e unbalance of air/coal and t h e change of combustion condition. When a mill change in present condition,the system display NOx emission forecast in several mlnutes and navigates t h e two-stageFig.7 Exnmplo of mn.n-machine interface (foul1 dingnosis for tho mill)

I ECON '91

1515

the

mill

system

becape

there

are

many

15) Monitoring system using c o u u n i c a t i o n h e

complicated operation items in t h e combustion system including mills.

1001

lo

A-coal 0-coal C-coal D-coa I

Using communication line, on-line monitoring operation situations from remote sites become possible. Data processing, analysis. and software development a l s o become possible. Fig. 9 shows t h e system configuration of combustion c o n t r o l support system.3 In-closing .

W explain t h e configuration of t h e combustion e c o n t r o l support expert system (K-MIND). This expert syst

em has been developed by using Kawasaki Inference System (KIS). which is Kawasaki's general-purpose t o o l f o r building expert systems.and this t o o l is easy t o add new functions and revise existing functions. This combustion control support expert system has been t e s t i n g a t t h e coal-fired power plant now. The demand of coal tends t o increase by diverse f u e l from now on.the requisition of a combustion management support expert system which enclose a plenty expertise w i l l be growing. of know-how and

50'

0.1

'

I

* 1 1 1 1 '

1

'

'

'

"""

10

'

'

00

Fig.8

u n b u r n e d carbon content in a s h (%) C h a r a c t e r i s t i c s of p h o t o s e n s o r o u t p u t

{rl)Automatic unburned carbon monitoring system The measurement of unburned carbon content in ash is important from t h e viewpoint of boiler efficient c o n t r o l and effective utilization in combustion ash. Analytical method is usually t h a t w e sample ash and d r y and measure ignition loss. So this method needs some hours and t h e length i s obstructive in combustion control. Unburned carbon monitoring system utilizes a correlation unburned carbon with ash's hue,and measures t h e s t r e n g t h of reflex light using photosensor. Fig.8 shows t h e relation between unburned carbon content in ash and sensor output.

Reference l)Kurosaki, Miyamoto, Nishino, Hayashi, Harada, Ihara"Development of Operation Support Expert System", Kawasaki Heavy Industry Technical Report".[104].'89 S)Tanaka, Miyatake. Umeki, Yamamoto, Miyamoto. Nishino"An operation support expert system f o r a coal-fired boiler u s i n g fuzzy inference" Proceeding of t h e 30th SICE Annual Conference '91 J)Tanaka, Miyatake, Yamamoto, Teramoto"Middle-Sized Operator Training Simulator f o r Coal-Fired Power Plant" Electric Power Research Institute International Conference Report,'91 4)Shiraha. Mori. Harada, Mihara, Hamada, Takano, Kitaguchi, Hasegawa"Development of Unburned Carbon Monitoring System", Kawasaki Heavy Industry Technical Report",[109],'91

Unburnd mrtmn mooibrinl m m bTelepbooommputcr

network

Network

Conlml

Fig.9

Coniigration o i combustion control support expert s y s t e m

1516

I ECON '91

本文来源:https://www.bwwdw.com/article/ibw4.html

Top