Figuration exceeding model / diffuse the ambiguous control method of join furnace temperature

Keywords: 　 of Flow Chart of 　 of SPF/DB 　 Fuzzy Control exceeds model figuration / diffuse join (SPF/DB) is to develop the aviation that rise to make a technology in recent years. Figuration exceeding model / diffuse the furnace temperature control of join heating furnace has important place in SPF/DB engineering technology. The product appearance that makes as a result of SPF/DB is complex, performance demand is high, process time is long, the temperature of the SPF/DB figuration mould inside heating furnace measures lag sex to be had greatly relatively nonlinear characteristic, build mathematical mathematical model to compare difficulty, accordingly, use ambiguous control more appropriate. It experiments to shortening periodic, improve a product it is quality, managing the sources of energy and raw material, important to reduce labor intensity to will rise action. The article exceeds model figuration only / diffuse the implementation method that faintness controls the furnace temperature of join heating furnace to go up in the computer undertakes elaborating. 1 ambiguous control implements a method 1.

Structure of 1 ambiguous controller and structure of controller of control target faintness see a picture 1. Graph Fig of sketch map of 1 ambiguous controller.

　 of 1 　 Fuzzy Controller dominates an end: To different pattern (basically be rectangular mould, the 80% above that this mould holds mould sum total) those who control furnace warm up, heat preservation, make energy is used up least, time is the shortest. 1.

2 parameter be detectinged undertake dominating an activity when the person when, must undertake judging to output deviation of the system above all, decide to adopt according to deviation next why to plant measure. Besides, still should be opposite below most circumstance deviation change ratio undertakes judging. That is to say, need fibrous root occupies deviation and deviation change ratio to have balance and judgement integratedly, such ability assure systematic pilot stability, can reduce quantity exceeding tone and oscillation phenomenon. So, when operation personnel undertakes controlling, of involves fuzzy concept basically have 3 by region: Deviation E; Deviation change is led; Control quantity of the person outputs U. Input parameter: .

In type of En=Yn-R of Δ of difference in temperature: Yn -- controlled quantity feedback value; R -- set-point. .

In type of ratio of change of difference in temperature: Tn -- sampling is periodic; Δ En-1 -- before one difference in temperature; Δ En -- this second difference in temperature. Output parameter: Voltage U. (Go to the lavatory to narrate, the following each medium deviation points to difference in temperature namely, deviation change ratio shows change of difference in temperature is led namely, control measures output to point to output voltage namely. ) 　 faintness controls a principle to see a picture 2. Graph 2 faintness control principle Fig.

2 　 Fuzzy Control Principle1.

3 faintness are changed 1.

3.

1 model gelatinization quantifies factor to decide those who set deviation E to be by region really [- X, x] , of the obscure part that deviation place takes it is by region (- N, - N+1, ... , 0, ... , n-1, n) , the faintness that can measure accurately in order to give out changes quantify factor K, 1.

3.

The practice of 2 models gelatinization is common course of action is, the span set of the deviation E that goes to observation above all is [- 6, + the successive quantity that changes between 6] , measure add of this in succession disperse to change next, be about to its cent is N archives, each files correspondence blurs collect, undertake blurring changing handling after that. If be in real work, the span that measures X accurately is not to be in [- 6, + between 6] , be in however [A, between B] , so, we can pass commutation type to will be in [A, b] the variable X translate into between [- 6, + 6] the variable Y between. (1) sets E to state the language is variable " deviation " . deviation (, negative) cent becomes a certain number of grade. For instance cent becomes 14 form, its code name is 　 (- 6, - 5, - 4, - 3, - 2, - 1, - 0, + 0, + 1, + 2, + 3, + 4, + 5, + then deviation talking region is 6) 　 X= ｛ - 6, - 5, - 4, - 3, - 2, - 1, - 0, + 0, + 1, + 2, + 3, + 4, + 5, + E of variable of 6 ｝ deviation has 8 languages to take a cost in X: Meaning: big, in the middle of, small, 0, lose 0, lose small, in losing, lose big mark: PB, PM, PS, PO, NO, NS, NM, NB gives out the each obscure part in X (I=1, 2, ... , the definition of 8) , see a table 1. The 　 　 class such as the define table variable of the obscure part in expressing 1 deviation to talk region X - 6-5-4-3-2-1-0+0+1+2+3+4+5+6 is subject the 　 that spend PB 00000000000.

10.

40.

81.

0PM0000000000.

20.

71.

00.

70.

2PS00000000.

30.

81.

00.

50.

100PO00000001.

00.

60.

10000NO00000.

10.

61.

00000000NS000.

10.

51.

00.

80.

30000000NM0.

20.

71.

00.

70.

2000000000NB1.

00.

80.

40.

10000000000 　 　 (2) manages together, set for " deviation change is led " the language is variable, its are Y= ｛ by region - 6, - 5, - 4, - 3, - 2, - 1, 0, + 1, + 2, + 3, + 4, + 5, + 6 ｝ assign deviation change rate into 13 grade namely. Set take 7 languages cost: Meaning: big, in the middle of, small, 0, lose small, in losing, lose big mark: PB, PM, PS, o, NS, NM, NB gives out to go up in the Y that consider region each obscure part (I=1, 2, ... , the definition of 7) , see the watch is shown 2 times. The 　 　 class such as the define table variable of the obscure part in expressing 2 deviation change to lead the Y that consider region - 6-5-4-3-2-10+1+2+3+4+5+6 is subject spend PB0000000000.

10.

40.

81.

0 　 PM000000000.

20.

71.

00.

70.

2 PS00000000.

91.

00.

70.

200 O000000.

51.

00.

500000 NO000.

20.

71.

00.

90000000 NM0.

20.

71.

00.

70.

200000000 NB1.

00.

80.

40.

1000000000 　 　 (3) similar ground, express with U " output controls an amount " the language is variable, its set ｛ of the Z= that it is 　 by region - 7, - 6, - 5, - 4, - 3, - 2, - 1, 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7 ｝ set U desirable 7 languages value: Meaning: big, in the middle of, small, 0, lose small, in losing, bear large certificate name: PB, PM, PS, o, NS, NM, NB gives each faintness market (I=1, 2, ... , the definition of 7) sees a table 3. The 　 　 class such as the define table variable of the obscure part in expressing 3 control to measure the Z that consider region - 7-6-5-4-3-2-10+1+2+3+4+5+6+7 is subject spend PB000000000000.

10.

40.

81.

0PM0000000000.

20.

71.

00.

70.

20PS00000000.

41.

00.

80.

40.

1000O0000000.

51.

00.

5000000NS0000.

10.

40.

81.

00.

40000000NM00.

20.

71.

00.

70.

2000000000NB1.

00.

80.

40.

1000000000001.

The makes ambiguous automation regulation with 4 ambiguous regular control is a handlers essentially the experience in controlling a process tries to sum up, sum up the relevant measure that must come by experience control regulation, make accordingly " ambiguous controller " , replace a person to undertake controlling to complex industrial process with the machine. To the input it is planar (deviation and) of deviation change ratio, output is one dimension (output controls an amount) ambiguous controller, if use language form to convey, have 　 " if and criterion " , or write into " If And Then " . For example, "Be like furnace temperature on the high side, and temperature ascendant rate is rapidder, criterion voltage drop is small " it is this sort model. As above, in controlling a language, , etc is fuzzy concept, accordingly they are to talk region X accordingly, y, z the ambiguous subclass that wait. Then, the control regulation of this kind of ambiguous controller can write a form: I=1 of 　 of If E= And = Then U= , 2, .

.

.

, m 　 J=1, 2, .

.

.

, n among them, , it is to be defined respectively in X, y, the obscure part on Z. Deviation E and deviation change ratio input variable for two, output control measures U to be output variable, their control regulation lists control regulation to express, see a table 4. Express 4 control regulation to express ENBNMNSOPSPMPBMBPBPMONMNSPMPMONSNOPSONSNMPOPSPSONMPMONMNBPB1.

5 faintness synthesize operation, beg obscure concern matrix afore-mentioned ambiguous condition statements can induce a relation of a faintness, namely according to ambiguous maths theory, here " × " the meaning of operation defines by next type: In type: μ -- subject spend a mark. If change of deviation, deviation is led,be taken respectively and, synthesize regulation according to ambiguous inference, output controls a quantity to be obscure part, namely such, be like foregone input, control an amount with output, can seek corresponding ambiguous relationship according to afore-mentioned regulation piece; Conversely, the obscure concern that is like a system is foregone, the basis is inputted and also can seek a control amount. 1.

6 generate control watch to beg after going out, but according to E=-6, - 5, - 4, - 3, - 2, - 1, - 0, + 0, + 1, + 2, + 3, + 4, + 5, + 6 with =-6, - 5, ... , 0, ... , 5, each element of subclass of 6 each archives is subject spend a watch, by control regulation seeks a corresponding control decision-making, control regulation to each faintness namely due and corresponding, i=1, 2, ... 21, control an amount (corresponding at, ) for. Next, press again increase advantageous position average standard gets corresponding control is measured, pass so much consideration to be able to reach control is expressed, see a table 5. Express E-6-5-4-3-2-10+1+2+3+4+5+6U-67676777442000-56666666442000-47676777442000-37666666320 of 5 control watch - 1 - 1 - 1 - 24445444100 - 1 - 1 - 1 - 14445441000 - 3 - 2 - 1 - 04445110 - 1 - 1 - 1 - 4 - 4 - 4 ＋ 04445110 - 1 - 1 - 1 - 4, when giving what cut truly to mix, its are corresponding subject spending the I order and degree that measures at place besides correspondence is 1 besides, the value of the others grade all is 0, can beg in order to simplify piece. 1.

The faintness of 7 output information adjudicates output of 　 ambiguous pilot is an ambiguous subclass, it mirrors control language to differ a kind of combination that takes a cost. But object be accusinged can accept amount of a control only, this needs to give this control to measure from the court decision in the ambiguous subclass of output. That is to say, want derivation to give to assemble by faintness common aggregate map, this map says for the court decision, by the court decision ability is gotten accurate control is measured. How to measure faintness the issue that translate into measures accurately, we are used add authority to adjudicate a law on average -- common increase advantageous position average standard. Its control the Umax that measure 　 to decide by next type: 1.

The 8 commutation considering region that measure data are in effective control, ratio of change of deviation, deviation and the change that control an amount are not X commonly, y, the element in Z, should undertake when using control to express region alternates talking. After taking ration to change factor K1, can measure E of any actual measurement deviation the some element in melt into X. Manage together, take calm another quantify factor K2, the deviation change that also can go to actual measurement leads the some element in melt into Y, can inquire control expresses the metabolic Uij that reachs control is measured right now, multiply it in order to quantify factor (or call scale factor) K3, attainable the controls quantity U change that is added actually to furnace. 1.

9 control regulation is amended well-known, be in usually, for example, the parameter of optimal control system finds actor, adjust from the parameter of adaptive control system, and the parameter that operates ambiguous control model artificially differentiates wait for a circumstance, all need to undertake adjustment to controlling regulation. The control that offers a kind of belt to amend factor for this people is regular: U= [α E+(1- α ) ] in type: α -- interpose at 0, the 1 real number between. Reach through computation control is expressed 6, express with control 5 can see control is expressed quite 6 smoother. The control after expressing 6 control regulation to amend expresses E-6-5-4-3-2-10+1+2+3+4+5+6U-67777777543200-57777777543200-46666666533100-35555555422100 - 255444442100 - 1 - 1 - 144433331-1-2 - 2 - 2 - 2 - 044433221 - 1-2 - 3 - 3 - 3 ＋ 0444320-1 - 1 - 2 - 2 - 3 - 3 - 4+1333320 - 2 - , can undertake correction to controlling regulation. Ought to wearing what point out again is, serve as with α adjust parameter to go to the lavatory to go easily not only, and the physical meaning that also has profundity, the extraction that is worth because of α is taken greatly small, show the rate increasing advantageous position of pair of deviation E and deviation change ratio directly, this mirrorred the thinking characteristic when handlers undertakes dominating an activity fitly. For example, in class of object be accusinged second taller when, the value increasing advantageous position that to deviation change leads should be more than the value increasing advantageous position of deviation E, because this α wants to take a few smaller value; Conversely, the class in object be accusinged second inferior when, the value increasing advantageous position that to deviation change leads should be less than the value increasing advantageous position of pair of deviation E, because this α wants,take a few bigger value. In addition, this kind of method can overcome sheet to choose to control regular blemish by experience, the empty shelves in avoiding to control regular definition or jump change phenomenon. So this kind of belt amends factor quantifying descriptive method is reasonable and feasible. Faintness of 　 of technological process of 2 ambiguous control controls flow to see a picture 3 graphs Fig of flow chart of 3 ambiguous control.

The article exceeds last word of 3 　 Fuzzy Control Flow Chart3 only the ambiguous control method of model figuration furnace temperature undertook inducing, come true on the computer. The definition that we know to deviation and deviation change lead ambiguous market and ambiguous control regulation reach his to amend all need to undertake many experiment ability obtains more economic data. This is the job that should have next. CNC Milling CNC Machining