CN105595391A - Advanced control method for comas tower dryer (CTD) - Google Patents

Abstract

The invention relates to an advanced control method for a comas tower dryer (CTD). The method comprises the steps that a predictive PI control algorithm and a double integral link are used for series connection, and the predictive PI control algorithm conducts correction on constant interference; when a deviation is reduced, the double integral link is used for slowly adjusting a secondary controller, a main actuator returns back to the ideal set value, and therefore a system has the good static property and finally reaches a stable state. According to the advanced control method for the comas tower dryer (CTD), the effect of removing interference can be achieved, and higher control precision of outlet moisture of the comas tower dryer (CTD) can be achieved.

Description

A kind of advanced control method about CTD air current type tobacco shred drying machine

Technical field

The present invention relates to the control method of CTD air current type tobacco shred drying machine, particularly relate to a kind of about CTD air current type tobacco shred drying machineAdvanced control method.

Background technology

Conventionally having a controlled variable in industry spot adopts two or more manipulated variable to controlSituation, generally all they can be controlled as single-loop system respectively, fix one of them variable, control another one becomeAmount, or both parallel control, but the selection of manipulated variable should be considered the reasonable and economic of technique, considers again to controlThe rapidity of performance, the two usually exists in a control procedure simultaneously, such as the moisture in cut tobacco dryer of tobacco Hong Si workshop sectionIn control system, by stationary magazine creel wall temperature, or regulate humidity discharging amount to control, so just can not be organically in conjunction with the twoThe advantage of control mode.

In control system, thereby people are by analyzing the input and output of system or obtaining object by mathematical modelingNominal model. But no matter nominal model is set up how accurately, and real system always can exist unknown dynamic characteristic, i.e. mouldThe uncertainty of type. The deviation of practical object and nominal model can regard as system in disturb, in control system, also oftenThere are various external disturbances, such as controlled quentity controlled variable disturbance or measurement noise, in these of control system, disturb and disturb general designation outwardFor total disturbance of system.

In the industrial process such as petrochemical industry, iron and steel, tobacco, inevitably there will be various external disturbances. In these disturbancesIn, existing normal value is disturbed, and also has PERIODIC INTERFERENCE. Disturb for normal value, some conventional control algolithms are as Smith Prediction ControlAlgorithm, Internal Model Control Algorithm just can play good inhibitory action to these interference, makes system output reach stable state. ButFor frequency conversion PERIODIC INTERFERENCE problem or other more complicated PERIODIC INTERFERENCE, said method just don't work. Therefore, this type of is askedTopic also just becomes a difficult point place of Industry Control.

CTD (ComasTowerDryer) is the novel tower-type air flow type cut tobacco drier of Italian COMAS company development, this machineAdopt the advanced design theory of permanent dehydration, maintaining is convenient. Have that tower body height is low, expansion and the advantage such as drying integrated.Use in Longyan and Xiamen cigarette enterprise at present.

But because CTD air flow type cut tobacco drier uses seldom at home, be also in the process of continuing to optimize, in the time of actual productionHave many problems, as feed system putty problem, the discharging gas lock steam outward amount of emitting is large etc. Particularly export moisture content of cut tobaccoFluctuation ratio is larger, and control accuracy is not high, and the quality that has seriously restricted product improves. Although producer has done a large amount of to this problemExplore and improvement, as: hot blast temperature is adopted to the Cascade Control Plan with respect to furnace temperature, reduced the ripple of hot blast temperatureMoving, still there is following problem: 1. the precision of moisture content of cut tobacco control is low, and standard deviation is large, and it is larger that mean value departs from set point.The result of 63 batches of creation data statistics is: standard deviation is 0.2633 to the maximum, and minimum is 0.1589; In setting value ± 0.75 scopeInterior qualification rate is 97.6 to the maximum, and minimum is 87.6; Mean value departs from set point and is 0.3102 to the maximum. At head due to moistureCan not detect, moisture control relies on operating personnel's experience completely, causes head moisture stabilization poor, and the fluctuation time is long, dry headPipe tobacco is many. 3. because PID controls fluctuation greatly, the amplitude of adjustment is too fierce, larger to the impact of equipment and executing agency, easilyCause putty and equipment fault.

Summary of the invention

Technical problem to be solved by this invention is to provide a kind of advanced control method about CTD air current type tobacco shred drying machine,Can remove interference effect, make the moisture content of outlet that dries silk reach higher control accuracy.

The technical solution adopted for the present invention to solve the technical problems is: provide a kind of about CTD air current type tobacco shred drying machineAdvanced control method, use that predictive PI control algolithm is connected with dual-integration link, makes predictive PI control algolithm to normal valueInterference is proofreaied and correct; When deviation reduces, described dual-integration link, for slowly regulating submaster controller, makes main actuatorValue is returned in ideal set value, thereby makes system have good static properties, and finally makes system reach a stable state.

The defeated entry/exit of described predictive PI control algolithm is closed and is: $u\left(t\right)=K\[1+\frac{1}{{\mathrm{\λT}}_{i}s}\]e\left(t\right)-\frac{1}{{\mathrm{\λT}}_{i}s}\[u\left(t\right)-u(t-L)\],$ ItsIn, λ is an adjustable parameter, the directly speed of Control loop system responses, and K is gain, T_iIntegration time constant, e (t) are defeatedEnter and the deviation of output, the output that u (t) is the t-L moment for output, u (t-L).

Described dual-integration link is taked taking humidity discharging control as main, hot blast temperature control as both are coordinated system by auxiliary modeOne.

Described dual-integration link is controlled moisture content of outlet in short-term with humidity discharging, when the valve position of humidity discharging exceedes range of set valueTime, will adjust hot blast temperature, by the change of hot blast temperature, make exhausting moisture valve bit recovery to range of set value.

Beneficial effect

Owing to having adopted above-mentioned technical scheme, the present invention compared with prior art, has following advantage and actively imitatesReally: the present invention is incorporated into dual-integration link in the system with predictive PI controller, the system of guarantee have good dynamicallyRespond, reach the effect of " symptomatic treatment in acute condition ". When deviation reduces, dual control system is brought into play again submaster controller and is slowly regulatedEffect, main actuator value is returned in ideal set value, thereby makes system there is good static properties, reach operation excellentThe object of changing, and finally make system reach a stable state, improve system accuracy, stablize humidity discharging valve position, do not need too much artificialIntervene.

Brief description of the drawings

Fig. 1 is unity feedback system structure chart.

Fig. 2 is predictive PI control algolithm structure chart.

Fig. 3 is that predictive PI adds dual-integration link block diagram.

Detailed description of the invention

Below in conjunction with specific embodiment, further set forth the present invention. Should be understood that these embodiment are only for illustrating the present inventionLimit the scope of the invention and be not used in. In addition should be understood that after having read the content of the present invention's instruction art technology peopleMember can make various changes or modifications the present invention, and these equivalent form of values fall within equally the application's appended claims and limitScope.

Embodiments of the present invention relate to a kind of advanced control method about CTD air current type tobacco shred drying machine, by predictive PI controlThe algorithm processed use of connecting with dual-integration link, makes predictive PI control algolithm that normal value is disturbed and be proofreaied and correct; Reduce in deviationTime, described dual-integration link, for slowly regulating submaster controller, is returned in ideal set value main actuator value, fromAnd make system there is good static properties, and finally make system reach a stable state.

Principle of the present invention is:

(1) dual-integration link is combined with predictive PI control algolithm, utilize dynamic response that dual controlling unit is good andStatic properties, use that predictive PI control algolithm is connected with dual-integration link, can either improve the control accuracy of moisture, again canIn enough situations about changing at feed state (as entrance moisture, inlet flow rate), in the tower that keeps dry, moisture field is relatively constant,And then the processing quality quality that ensures to dry silk, optimize the control model of head and afterbody simultaneously, reduce to greatest extent head andThe dry pipe tobacco amount of afterbody.

(2) definition: as shown in Figure 2, predictive PI controller, is made up of two parts independently, PI controls and prediction estimatorPoint, the defeated entry/exit of controller is closed and is:

$u\left(t\right)=K\[1+\frac{1}{{\mathrm{\λT}}_{i}s}\]e\left(t\right)-\frac{1}{{\mathrm{\λT}}_{i}s}\[u\left(t\right)-u(t-L)\]$

Wherein the former PI controller, the latter is predictive controller. Wherein, K is gain, T_iIntegration time constant, e (t) areInput is that output, u (t-L) are the output in t-L moment with deviation, the u (t) of output.

In order to make the moisture content of outlet that dries silk reach higher control accuracy, the advanced control algorithm of design replaces originalPid control mode, sets up a new pattern of the advanced control of brand-new CTD air current type tobacco shred drying machine thus. Present embodiment adoptsKey technology comprise: predictive PI technology, Kalman filtering technique, dual (multiple) control technology etc. These technology allSuccessful Application in industry, its advance obtains the Large Scale Process such as petrochemical industry, iron and steel, chemical fibre, glass and cigarette with reliabilityThe long-term inspection of industry.

Predictive PI control technology

For a long time, traditional pid control algorithm is due to simple in structure, and the parameter advantage such as easily adjust is widely used in workIn industry control system. But for the system of some multivariables, large time delay, non-linear, close coupling, pid algorithm often can notObtain well-content control effect. By Advanced process control APC (AdvancedProccessControl) raising systemPerformance has become a kind of trend.

In numerous APC, since predictive PI control algolithm was proposed by Hagglund from 1992, obtain sending out progressivelyExhibition and improvement, and successful Application is in the control of some complex objects. Predictive PI control algolithm can be divided into two kinds: one is bandHave the PID controller of forecast function, its essence or a kind of PID controller, just design according to some advanced control mechanismsController parameter. Another kind is also that this programme adopts, and is the controller that prediction algorithm and pid algorithm merge, and isA kind of predictive PI control algolithm truly.

Predictive PI control algolithm has been successfully applied to petrochemical industry, iron and steel, chemical fibre, glass, biofermentation, redrying and volumeOn more than 100 control loop in cigarette industry, the input and output transfer function of this controller can represent:

$U\left(s\right)=\frac{1}{\mathrm{\λ}K}(1+\frac{1}{Ts})E\left(s\right)-\frac{1}{\mathrm{\λ}Ts}(1-e^{-\mathrm{\τ}s})U\left(s\right)$

The leading time constant that in formula, T is process, K generally elects process gain as, and τ is system lag time, and λ is adjustable ginsengCount, its objective is the closed loop response speed of adjustment System. Sort controller is structure, and its Section 1 is PI controller,Section 2 is predictive controller. Thereby this control algolithm had both had, and pid control algorithm is simple in structure, the advantage easily of adjusting, againThe function with prediction, control accuracy is high, and robust performance is good. Also very convenient in the realization of algorithm.

Kalman filtering technique

Owing to adopting infrared mode to measure charging and discharging moisture, and this mode certainty of measurement is low, is subject to multipleThe impact of disturbing factor is larger, measures noise large, poor stability. Therefore be necessary the data of infrared moisture to process rootAccording to infrared moisture measuring principle and Kalman filtering technique, measured data are processed, statistical angle,To the most real moisture data, for the stable operation of CTD air current type tobacco shred drying machine is given security.

Kalman filtering is a kind of high efficiency recursion filter (autoregressive filter), and it can be from a series of not completeEntirely comprise in the measurement of noise, estimate the state of dynamical system. In simple terms, Kalman filter is " an optimalRecursivedataprocessingalgorithm (optimization autoregression data processing algorithm) ". For solving very large portionThe problem of dividing, it is optimum, most effective or even the most useful. Its extensive use has exceeded 30 years, comprises robotNavigation, control, Data Fusion of Sensor are even followed the trail of etc. at radar system and the guided missile of military aspect. More answered in recent yearsFor Computer Image Processing, for example head and face identification, image are cut apart, Image Edge-Detection etc.

5 formula of Kalman filter are its core contents. In conjunction with modern computer, Kalman's program phase in factWhen simple, be only appreciated that its those 5 formula, it is very simple that practical application also just becomes.

The research object of present embodiment is the moisture content of outlet of CTD air current type tobacco shred drying machine. Rule of thumb judgement, this moistureTo have certain dynamic relationship, the namely moisture in lower moment with jetting steam caused flow, hot blast temperature, wind speed, humidity discharging amount etc.It is variation that can these technique amounts and can estimating. But experience estimate be not 100% accurately, may have deviation. ThisA little deviations can be considered as white Gaussian noise, and namely deviation is that it doesn't matter with surrounding time, and meet Gauss's distribution. SeparatelyOutward, it is also inaccurate that infrared moisture is measured, and measured value can have deviation than actual value. Also this deviation is regarded as to Gauss's white noiseSound. Now, at a time there are two about moisture content of outlet value: experiential predicted value (predicted value of system) and infrared rayThe value (measured value) of moisture measurement. Use these two values to estimate the actual value of moisture content of outlet in conjunction with their noises separately, thisTime this actual value of obtaining be the most accurately and reliably from statistical angle.

Dual (multiple) control technology

In existing moisture control, the general humidity discharging amount of fixing that adopts, controls moisture by adjusting hot blast temperature. ThisScheme does not combine adjustment humidity discharging and hot blast temperature, can not really ensure that in drying tower, humidity is relatively stable.

It is main that this programme is taked humidity discharging control, and hot blast temperature control is auxiliary, and by both coordinating and unifyings. In short-term with humidity dischargingControl moisture content of outlet, because humidity discharging control moisture has response speed fast, can improve greatly like this moisture controlThe precision of system, and when the valve position of humidity discharging is too small or when excessive, by the speed adjustment hot blast temperature with suitable, like this due to hot blastThe change of temperature, will make exhausting moisture valve bit recovery to desired setting value. This control model can either improve the control of moisturePrecision, in the situation that can change at feed state (as entrance moisture, inlet flow rate) again, moisture field in the tower that keeps dryRelatively constant, and then the processing quality quality that ensures to dry silk. Concrete scheme as shown in Figure 3.

According to successful case in the past, after CTD air current type tobacco shred drying machine is implemented advanced control program, can reach following orderMark:

1. control procedure is completely automatic, really accomplishes operating in a key, avoids manual intervention in process of production, reducesThe randomness of manipulation strength and operation.

2. overcome interference performance strong, control accuracy is high, and the qualification rate that moisture content of outlet is controlled at setting value ± 0.5% is greater than97%, standard deviation is less than 0.15, and moisture content of outlet mean value departs from set point and is less than 0.1.

3. there is adaptive control function, by on-line study and accurate Mathematical Modeling, guarantee head moisture at headArrive fast setting value, neither leap high, not on the low side yet.

4. by accurate Mathematical Modeling, optimize the control model of head and afterbody, reduce to greatest extent head and tailThe dry pipe tobacco amount of portion, dry pipe tobacco amount reduces 30% on the original basis.

5. coordinate hot blast temperature and humidity discharging amount to the association between moisture adjustment, eliminate the coupling between them, avoid adjustingHave suffered the vibration in journey.

6. by the design of system, the system adjustment of making is level and smooth as much as possible, reduces the fluctuating range of control system, reduces stifledMaterial and equipment fault number of times, reduce pipe tobacco making in process of production broken.

Claims (4)

1. about an advanced control method for CTD air current type tobacco shred drying machine, it is characterized in that, by predictive PI control algolithm and dualIntegral element series connection is used, and makes predictive PI control algolithm that normal value is disturbed and be proofreaied and correct; When deviation reduces, described twoMultiple integral link, for slowly regulating submaster controller, is returned in ideal set value main actuator value, thereby system is hadGood static properties, and finally make system reach a stable state.

2. the advanced control method about CTD air current type tobacco shred drying machine according to claim 1, is characterized in that, described pre-The defeated entry/exit pass of surveying PI control algolithm is: $u\left(t\right)=K\[1+\frac{1}{{\mathrm{\λT}}_{i}s}\]e\left(t\right)-\frac{1}{{\mathrm{\λT}}_{i}s}\[u\left(t\right)-u(t-L)\],$ Wherein, λ is oneIndividual adjustable parameter, the directly speed of Control loop system responses, K is gain, T_iIntegration time constant, e (t) are input and outputDeviation, u (t) for output, u (t-L) be the output in t-L moment.

3. the advanced control method about CTD air current type tobacco shred drying machine according to claim 1, is characterized in that, described twoMultiple integral link is taked taking humidity discharging control as master, hot blast temperature control are as auxiliary mode is by both coordinating and unifyings.

4. the advanced control method about CTD air current type tobacco shred drying machine according to claim 3, is characterized in that, described twoMultiple integral link is controlled moisture content of outlet in short-term with humidity discharging, in the time that the valve position of humidity discharging exceedes range of set value, will adjust hot blastTemperature, by the change of hot blast temperature, makes exhausting moisture valve bit recovery to range of set value.