Զѧ
ACTA AUTOMATICA SINICA
1999 25 6 Vol.25 No.6 1999



¯ģ
̬Żо1)
ά塡
ժҪ˻Ԫ缼ӹҵ(¯)ģ.յҪ󣬷ֱָģͺͲƷģͣ˸ӹҵ̲Ʒģ.ڽģϣ¯̬Żо÷нϺõĽЧ.˽ŻƵĽ.
ؼʡƷģ̬Żƣ¯.
STUDY OF NEURAL NETWORK QUALITY MODELS AND
STEADY-STATE OPTIMIZING CONTROL FOR
NICKEL FLASH SMELTING FURNACE
WAN WeihanWAN Baiwu
(Systems Engineering Institute,Xian Jiaotong University,Xian710049)
YANG Jinyi
(Jinchuan Non-ferrous Metals Complex,Jinchang737000)
AbstractThe paper proposes an approach that uses soft-sensing method to set up the neural network models of the complex industrial processnickel flash smelting furnace.They are technological index quality models and yield model for the furnace.This opens up a new application field of neural network modeling.The paper also gives a study of steady-state optimizing control for the furnace.The results show that the modeling and optimization provide better effect in saving energy consumption.Finally,the paper suggests how to implement on-line steady-state optimizing control to the furnace in the future.
Key wordsSoft-sensing technique,neural network quality modeling,steady-state optimizing control,nickel flash smelting furnace.
1
չʮָӵĹҵ̣Ҫʵ߿ƾͱ˽ոùҵ̵ѧģ.¯ģ͵ķ÷ܹΪ׼ȷطӳ¯ʵй.ڽģĻϣ̬Żо.
2ҵ̱
ұһǿұ.ָѾȸķ״򻯾Ԥȿۼ̳ʵĻ¯Һ̬Ļұ.¯Ĺͼ1ʾ.


ͼ1¯Ĺ
3̬ģ͵Ľ
3.1ѡ
ͨԸ罨ģͺƽۣ¯Դص㣬ѡBPΪģ.
3.2̬ģ͵Ľ
¯̬ѧģָ¯ھ£̬ʱϻָ֮ѧϵ.ñʽyi=fi(x1,x2,,xn)ʾyiָx1,x2,,xn.fi(.)ƷָϵķԹϵ.
2ݵʵͼ2ʾ.Ȳһxiʵypjͨѵģ.¯йУʵypjͨXӫǺͽƽõģͺʵʣʱymjΪʵȥʵ̬Ż.ⶨypjĽypjȥģ.


ͼ2Ľṹͼ
ͨṹĶԱо˫BPṹڵΪ5-5-5-1.250000εĵѧϰ.ѧϰߺͼ36ʾʵΪ¯ʵߣΪѧϰ.


ͼ3Ʒλѧϰ


ͼ4¶ȵѧϰ


ͼ5ȵѧϰ


ͼ6ѧϰ
4̬Żʵ
4.1ԼµĿŻʵ
ܺEΪĿ꺯
E=1Fair+2Foxy+3Foil=1c1+2c2+3c3,(1)
аʵʱֵ(Ԫ/λ)1=0.982=0.5,3=0.015.1ΪָŻı仯Χ.
1ָŻķΧ

Ʒλ¶ȱۼ̳
maxi,cmaxi501 2501.322 00012 0001 20013.010.0
mini,cmini401 1501.02018 00010 0008509.07.50

ŻԱʾΪ
(2)
(3)
(4)
ԼתΪgi(ci)0ʽ.Ż⣬ԼתΪԼŻ.SUMT(ԼС)㷨
(5)
ݶȷ

ý2ʾ.
2Ż

ŻŻӳֵʼŻ
(Nm3/H)1=0.004m1=0.006m2=0.0061=18 292.051=19 522.90
(Nm3/H)2=0.004m3=0.006,m4=0.0062=12 325.372=10 018.69
(L/H)3=0.004m5=0.006,m4=0.0063=974.102=1 105.44
ܼ(T/H)4=0.004m7=0.006,m8=0.0064=11.104=13.13
̳(T/H)5=0.004m9=0.0065=7.465=10.13
(T/H)=25.89

ܺĿֵʼֵE=7393.2(Ԫ)Żֵ=6 385.5(Ԫ).ÿСʱԼ1 007.7(Ԫ).ģͲ30ݣ30Լ725544(Ԫ)ȫ갴10(300)ƣԼܺķ7255440(Ԫ).
4.2Լ¶ĿŻʵ
ܺΪŻĿ꣬ŻĻϣͼԼбһ㣬۲ܺĵı仯.ӣܺҲ.ΪΪṩһۺϿǵĽ.

5ָŻƣ4
ָָָͲ͸ŻΪߵڿ趨ṩָ.ŻԶڵģͽģ͸ºͲŻ㣬Żȫֱ͵Ϊ趨ֵ.Żͼ7ʾ.һҪ̬Żڵȷ.ͨݵ̬Żѡ2Ϊһ̬ŻڽΪ.


ͼ7Żƽṹͼ
6
оóĽΪ¯ʵ̬Ż˿ϲһ.һصǾ콫оɹʵʣоڵ.⣬¯Ѿ꣬˲ר֪ʶܷר֪ʶ룬רҿϵͳʹƸƣһ̽.
1)ҡƻɫ˾л.
߼飺ά1964ͨѧеѧԺеӹϵʿо.оΪӹҵ̵ĽģŻܿӦоеܹϵ.
塡19281951оҵϺͨѧо.ͨѧϵͳоڣʿʦ.ĿǰҪоΪϵͳݽŻƺܿơҵ̵̬ŻƵ.񲿼ƼһȽ2ΡȽ2κȽ1Σ270ƪ.
塡19351961ͨѧϵҵνɫ˾.
ߵλά塡ͨѧϵͳо710049
塡ɫ˾737000
ο
1壬ֽ.ģҵ̵̬ݽ׿.Զѧ199016(2)186192
2Իۣʫ嵣.̿Ƶķչչ.Ӧã199714(2)145151
3Pican N,Alexandre F.Artificial neural networks for the presetting of a steel temper mill.IEEE Expert Intelligent Systems and Their Applications.1996,11(1):2227
4壬.ҵ̼̬Ż..ѧ磬1998
ոڡ1997-09-10޸ĸڡ1998-03-16
