CN101830059B - Method for controlling injection speed of screw of injection molding machine - Google Patents

Abstract

The invention discloses a method for controlling the injection speed of a screw of an injection molding machine, which comprises following steps of: (1) carrying out phase-step open-loop test on an injection process, (2) carrying out displacement segmentation on a displacement area of the screw, (3) confirming the set value of an identification signal, recording measured values of the identification signal, injection speed data and displacement data of the screw in the whole identification test at different moments, (4) confirming time segmentation data corresponding to the identification test, then modeling and segmenting to identify linear model of each partial model segmentation span, (5) establishing global nonlinear model which takes screw injection displacement as scheduling variable, (6) measuring injection displacement of the screw on line to linearize the global nonlinear model to obtain a linear model, and (7) taking the linear model as a prediction model of a dynamic matrix model prediction controller to carry out closed-loop control on the screw injection speed. By adopting the invention, the nonlinear model can be linearized conveniently, and the set point of the injection speed can be tracked effectively. The modeling process is simple and the identification cost is low.

Description

A kind of method for controlling injection speed of screw of injection molding machine

Technical field

The invention belongs to the injection mo(u)lding field of engineering technology, be specifically related to the screw speed control method in the injection process.

Background technology

Injection mo(u)lding is an important techniques in polymer process industry, and the many plastic products of one-tenth output that this process can economical and efficient, these goods are in industry, agricultural, and electronics also has occasions such as household to have a wide range of applications.An injection machine generally includes injecting unit, the switch form unit, and hydraulic power unit also has control module.

Plastic polymer is melted in injecting unit, heating and stirring.The polymer of fusion is squeezed in the mould cavity that belongs to the switch form unit under the effect of injection pressure then.Through cooling, the polymer cure of fusion becomes the goods with die cavity shape.Hydraulic power unit is used for the controlling of injecting unit, and control module is responsible for the monitoring to whole injection moulding process through put forward The whole control command signal and control sequence to machine.

The injection moulding process stage:

Injection moulding process is a typical cycle repetitive process.Each injection cycle all comprises three stages independently:

Injection stage, packing stage, preformed and cooling stage, its concrete process sketch is as shown in Figure 2.

(1) injection stage

At first answer close die, then nozzle move forward and with tight-lipped close connection of the running channel of mould, last under given injection speed, screw rod is with the certain speed start injection that travels forward.In this process, will guarantee to produce correct melt filling speed and loading to the control of screw position.After mould cavity is full by the polymer filling, injection screw will keep certain pressure on polymer, and injection process finishes, and switches to packing stage.

(2) packing stage

Controlled at this stage cavity pressure at a higher stress level, as pressure compensation, prevent the contraction of goods.Do like this is in order in the middle of cooling procedure, to guarantee correct product weight.In order to prevent pressurize, guarantee the quality of product, generally pressurize is divided into several stages, in the end (or several) stage pressure remains on a lower level, until gate freeze.Process just gets into the plasticizing cooling stage.

(3) plasticizing cooling stage

The polymer of fusion cooling forming in die cavity in this stage; After the screw rod of injecting unit is retreated through the rotation of a period of time; With the mixed with polymers of existing molten state in solid-state polymer and the barrel, the polymer that is plastified will used in the injection next time.After goods reached certain rigid, this process finished, moulds open, and goods are ejected.

In the actual industrial production process, above several stages constantly repeats.

The procedure parameter that needs control:

In the several stages of injection cycle process, in order to obtain the accurate geometry of required product, microstructure, density etc., we must effectively control the procedure parameter of several keys.These procedure parameters mainly comprise: polymer melt temperature, polymer filling rate and cavity pressure.

The procedure parameter of more than mentioning will directly have influence on quality of item.Different temperature can change polymer viscosity significantly; Watering before the road junction freezes, injection rate has determined to inject what of amount of the polymer in the mould; Cavity pressure determined action of injection the size of the viscous friction that will overcome, also determined to inject the amount of mould polymer simultaneously.The melt viscosity that factor is a polymer that product quality is exerted an influence; If viscosity is too low, the polymer of molten state will experience longer distance in die cavity under same pressure, just can cause pressurize; Increase the weight of goods, also can produce more overbottom pressure simultaneously.On the other hand, too high injection rate also can cause pressurize and have influence on the molecularly oriented of injecting products.

That summarizes says that the injection machine Variable Control mainly comprises following aspect:

Ground floor: machine variable.Mainly comprise

Temperature: barrel temperature, nozzle temperature and condensation temperature;

Pressure: dwell pressure, back pressure, maximum injection pressure;

Logic sequence and motion: die sinking, locked mode, filling, pressurize, preformed, the position switching point that ejects etc., screw of injection speed, injection capacity etc.

The second layer: process (independence) variable.Mainly comprise melt temperature, melt pressure (nozzle and cavity pressure), maximum shear stress, the heat condensation rate etc. of becoming estranged.

The 3rd layer: quality variable.Comprise mainly that product weight and thickness, contraction and bent stuck up, the products appearance at indenture, the place of condensing and intensity etc.

In the middle of above-described variable, the injection speed that we select ground floor is a controlled variable, because it is playing the part of a vital role aspect the product quality influencing.It has characterized at injection stage, and the polymer of fusion gets into the speed of die cavity.Injection speed can influence in the die cavity very greatly to a great extent presses and the resultant articles quality, mainly shows the following aspects: overbottom pressure, shrink, and impact strength, product form also has surface characteristics.Output is also directly related with the injection speed setting value.So a kind of speed value of can accurate tracking setting responds closed loop control algorithm fast, play crucial effects for the quality of final products.In this experiment, we mainly control the screw of injection speed in the ground floor, also are injection speed.

At present, through to the updating of process model and control algolithm, a lot of injection speed closed loop control methods have been arranged by people's Application Research.Like non-minimum phase adaptive control algorithm, based on the iterative learning PI controller and the Sliding-Mode Control Based of nonlinear model based on model.But above algorithm does not pass through experimental verification, and the nonlinear characteristic of process also has been left in the basket.Someone has proved through experiment and has had non-linear, time-varying characteristics in the injection process; And designed PID controller based on fuzzy logic ordination, through experiment show this controller have very strong speed Tracking performance and robustness (to different dies, different temperature; Different polymer can both obtain satisfied control effect); But the priori that the design demand of this controller is a lot of needs through a large amount of experiment accumulation, and the design of controller is difficulty relatively.Someone has proposed, and single model is simplified PREDICTIVE CONTROL and multi-model is simplified two kinds of methods of PREDICTIVE CONTROL.Wherein multi-model process mainly is to the non-linear proposition of process; Come identification model through the multistep step test, when carrying out the multi-model segmentation, segmentation criteria is not confirmed; Be not sure of the location point that step changes; The step response model difference that the model that picks out like this obtains at the diverse location place is also bigger, and the adaptive bigger problem of model can appear in the model that obtains probably when prediction algorithm is used, to the quick tracking performance of injection speed also variation thereupon.This method be through inject time long large-scale injection product test; Its response tracking time to setting value is long about 300ms; Concerning the small sized product of lacking inject time, obviously inapplicable, and concerning the processing of small sized product; The method of setting up multi-model through the multistep step is very difficult, and inconvenience realizes fast and the effectively tracking to injection speed.

Summary of the invention

Non-linear, time-varying characteristics to injection speed existence in the present injection machine injection process; And existing control method can not realize the effective control to the injection machine injection speed well; Technical problem to be solved by this invention provides a kind of method for controlling injection speed of screw of injection molding machine, and this method has realized the effective tracking to the screw of injection speed setting value.

The inventive method has adopted a kind of non-linear modeling method based on the scheduling variable---with the position scheduling variable; Set up the overall interpolation nonlinear model of position-based for the scheduling variable; Take the method for local linearization again; Local linear model to after the process linearisation adopts predictive control algorithm that injection speed is carried out closed-loop control again, thereby has realized the effective tracking to the injection speed setting value.

The technical solution adopted for the present invention to solve the technical problems is following:

This method for controlling injection speed of screw of injection molding machine mainly comprises the steps:

(1) utilize high-speed data acquiring device that the injection process of injection machine is carried out the step open-loop test; Obtain gathering the screw of injection speed constantly and the system oil pressure data of screw displacement and injection machine, and further obtain the rate of change of said system oil pressure based on the system oil pressure data of injection machine at each;

(2) according to the rate of change of said system oil pressure and the relation between the screw displacement, to screw rod in the injection process the displacement section of process carry out displacement subsection, obtain the required displacement subsection interval censored data of modeling;

(3) the step open-loop response characteristic corresponding that obtains based on step (1) with screw of injection speed; Confirm the identification signal setting value in order to carrying out identification test, and utilize the identification of high-speed data acquiring device record to test each identification signal measured value, screw of injection speed data and screw displacement data constantly in the whole process;

(4) confirm the time slice data that corresponding identification is tested according to the displacement subsection interval censored data in the step (2) earlier; According to said time slice data resulting identification signal measured value of step (3) and screw of injection speed data are carried out the modeling segmentation again; Obtain between three partial model piecewise intervals and two transition regions, and pick out three linear models corresponding

with each partial model piecewise interval

(5) according to the identification signal measured value between said three partial model piecewise intervals and two transition regions and screw of injection speed data, the screw displacement data of synchronization; What the linear model of three partial model piecewise intervals described in the integrating step (4), foundation were shown below serves as the overall nonlinear model of scheduling variable with the screw of injection displacement:

$y\left(t\right)=({\mathrm{\α}}_1\left(w\right){\hat{G}}^1\left(q\right)+{\mathrm{\α}}_2\left(w\right){\hat{G}}^2\left(q\right)+{\mathrm{\α}}_3\left(w\right){\hat{G}}^3\left(q\right))u\left(t\right)+v\left(t\right)$

Wherein, weight function α ₁(w), α ₂(w) and α ₃(w) be the cubic spline function of screw of injection displacement w;

(6) on-line measurement screw of injection displacement, the screw of injection displacement data that utilizes current time to record carries out linearisation to said overall nonlinear model, obtains linear model;

(7), and utilize this predictive controller that screw of injection speed is carried out closed-loop control with the forecast model of said linear model as the dynamic matrix model predictive controller.

Further, identification signal according to the invention is the M sequence.

Further, the sampling period of M sequence according to the invention is 10ms.

Compared with prior art; Beneficial effect of the present invention mainly shows: ⑴ is through introducing the rate of change of the central system oil pressure of screw of injection process; Analytic process produces nonlinear reason, the position waypoint has been carried out reasonably confirming, and with this process has been set up local linear model; ⑵ serve as the scheduling variable with the screw of injection displacement; Utilize the local linear model of segmentation; Add the test data and the transit data of whole process; Having set up the nonlinear parameter model of the overall situation on a large scale, is parameter with the real time position at scene in use again, can be very easily with the nonlinear model linearisation; ⑶ utilizing the later model of local linearization, sets up the DMC controller and carry out emulation experiment, obtained satisfied control effect; ⑷ that model is set up is simple, the identification cost is low, and online convenience of calculation, control resultant effect are good, have realized the effective tracking to the injection speed setting value.

Description of drawings

Fig. 1 is the screw speed control device structural representation of injection machine;

Fig. 2 is the course of work sketch of injection machine;

The fundamental diagram of Fig. 3 screw of injection speed controlling of the present invention;

Fig. 4 is screw displacement and system pressure rate of change open loop step response graph of a relation;

Fig. 5 is a screw of injection velocity step response diagram;

Fig. 6 is modeling when test screw displacement variation relation figure in time;

Fig. 7 is a modeling test data piecewise graph;

Fig. 8 is the global data fitted figure of the corresponding linear model of partial model piecewise interval, wherein,

A: the matched curve of first section linear model;

B: the matched curve of second section linear model;

C: the matched curve of the 3rd section linear model.

D: real process curve;

Dynamic matrix Model Predictive Control speed closed loop analogous diagram when Fig. 9 is the single linear model, wherein, solid line: setting value curve; Dotted line: speed Tracking curve.

Dynamic matrix Model Predictive Control speed closed loop analogous diagram when Figure 10 is nonlinear model, wherein, solid line: setting value curve; Dotted line: speed Tracking curve.

Among the figure, 1. die cavity, 2. cast gate, 3. running channel, 4. sprue gate, 5. nozzle; 6. heating collar, 7. screw rod, 8. barrel, 9. hopper, 10. hydraulic motor; 11. reversal valve, 12, safety valve, 13. hydraulic oil containers, 14. motor speed encoders; 15. permanent magnetic synchronous electrical motor, 16. servo driver of motor, 17. displacements, velocity sensor, 18. high-speed data acquiring devices, 19. gear pumps.

The specific embodiment

The oil electricity that the present invention can adopt Ningbo HaiTai Machine Manufacturing Co., Ltd to produce mixes energy-saving injection machine, and its model is HTL68/JD, and it is as shown in Figure 1 among the present invention screw of injection speed to be carried out close loop control circuit.Wherein, AI, AO are data collecting plate card (PCI1710HG/PCI1720) analog quantity input/output port; S600 is the permanent-magnet synchronous electric motor driver; PMSM is a permanent magnetic synchronous electrical motor; P and V represent MTS sensing station measurement output signals and tachometric survey output signal respectively.High-speed data acquiring device of the present invention can adopt the high-speed data acquiring device of being made up of the PXI-1042Q cabinet of NI company and NI PXI-5122 high-speed digitization appearance module and LABWIEW software; Also can adopt with the PCI1710HG that grinds magnificent company and PCI1720 data collecting plate card and be equipped with the high-speed data acquiring device that the PC of real time operating system RTAI-LINUX is formed, realize injection speed, screw of injection displacement etc. in the injection process are realized the data acquisition task of Millisecond.Polymer can be selected high density polyethylene (HDPE) (HDPE) for use, and barrel is heated to be four sections heating collar heating, and temperature is set and is 200 ℃.

Among the present invention with screw of injection speed in the injection process as control variables, the input of motor driver is as performance variable, screw displacement is as the scheduling variable.Fig. 1 is the screw speed control device structural representation of injection machine; Concrete operation principle is: behind given servo driver of motor 16 a certain speed setting values; Permanent magnetic synchronous electrical motor 15 driven gear pumps 19 rotate with certain rotating speed; Hydraulic oil in the fuel tank is pressed in the injection cylinder through reversal valve 11, and lead-screw 7 is to left movement, when screw rod is moved; Be installed in the displacement on the screw rod, displacement, the rate signal that velocity sensor 17 just can produce screw rod, through high-speed data acquistion system 18 the output analog signals of sensor carried out acquisition process then.Because what select for use in this embodiment is position, the velocity sensor of the output of MTS voltage; There are the vibrations of machine in industry spot; Motor, frequency converter, the electromagnetic interference of transformer etc.; Make the data that collect have High-frequency Interference, the radio-frequency component that therefore can use lowpass digital filter will collect in the signal filters out.Can select Butterworth LPF, through the emulation experiment of off-line, the data sampling frequency is 1KHZ in test, when cut-off frequency is decided to be 40HZ, just can obtain comparatively ideal data and curves.

Before carrying out Model Distinguish, through utilizing high-speed data acquiring device the injection machine motor driver is applied step voltage input, injection process is carried out the step open-loop test, confirm screw rod in the injection process the displacement section of process carry out displacement subsection.Wherein be input as open loop step test motor driver setting value with the 5V step; And the system oil pressure data and the screw displacement data of process are carried out acquisition and recording; System oil pressure numerical value to collecting carries out difference, obtains corresponding rate of pressure change, and the gained test result is as shown in Figure 4.Then, according to its variation characteristics, as shown in Figure 4 it is divided into three sections: first section slowly reduces the interval for rate of pressure change, and second section is the rate of pressure change interval that increases sharply, and the 3rd section slowly reduces the interval for rate of pressure change.Between first section and second section, and there is certain transit data section between second section and the 3rd section.The piecewise interval of determining screw displacement is 157mm-153.5mm (first section), 153.5mm-152.5mm (between transition region one), 152.5mm-150mm (second section), 150mm-149mm (between transition region two), 149mm-143mm (the 3rd section).

After screw displacement piecewise interval data are confirmed, just can carry out the realization of overall nonlinear model foundation and closed-loop dynamic matrix control method.Concrete realize being divided into modeling and control two large divisions, below this two large divisions is introduced respectively.

(1) the modeling method flow process is following:

(1) utilizes above-mentioned screw of injection velocity step response curve (as shown in Figure 5); The amplitude of design input signal (M sequence) is 5V; The M sequential sampling cycle is 10ms; Carry out open loop models identification test,, carry out filtering through Butterworth LPF with screw of injection speed that obtains and screw displacement data.

(2) utilize each waypoint (157,153.5,152.5,150,149,143 of above-mentioned screw displacement piecewise interval; Unit: mm); Calculate each position waypoint time corresponding point data (137,247,279,371,400,585 in the Model Distinguish test process; Unit: ms), as shown in Figure 6.

(3) according to obtaining the time slice point data in (2), the datagram that divided ring identification test obtains carries out segmentation, and segmentation result is as shown in Figure 7.In MATLAB; Utilize the performance variable (motor driver voltage given) and the speed output data of each piecewise interval; Carry out Model Distinguish, pick out the ARX model of 3 second orders:

the corresponding linear model overall situation fitted figure of every partial model piecewise interval as shown in Figure 8.And pick out a single overall linear model, be used for doing than than usefulness.

(4) identification overall situation nonlinear model utilizes the transit data between three partial model piecewise intervals and each interval to set up overall nonlinear model.

A. the nonlinear model of the overall situation is shown below:

$y\left(t\right)=({\mathrm{\α}}_1\left(w\right){\hat{G}}^1\left(q\right)+{\mathrm{\α}}_2\left(w\right){\hat{G}}^2\left(q\right)+{\mathrm{\α}}_3\left(w\right){\hat{G}}^3\left(q\right))u\left(t\right)+v\left(t\right)$

Wherein, weight function α ₁(w), α ₂(w), α ₃(w) be the nonlinear function of scheduling variable position w.

is the operating point linear model that picks out in (3).

B. weight function α ₁(w), α ₂(w), α ₃(w), adopt cubic spline function, the order s=51 of cubic spline function, accumulation collection: K={k ₁, k ₂..., k _s, k ₁=157<k ₂<...<k _j<...<k ₅₁=143.

Concrete function representation is following:

${\mathrm{\&alpha;}}_1\left(w\right)={\mathrm{\&beta;}}_1^1+{\mathrm{\&beta;}}_2^{1}w+\underset{j=2}{\overset{s-1}{\mathrm{\&Sigma;}}}{\mathrm{\&beta;}}_{j+1}^1{|w-k_j|}^3$

${\mathrm{\&alpha;}}_2\left(w\right)={\mathrm{\&beta;}}_1^2+{\mathrm{\&beta;}}_2^{2}w+\underset{j=2}{\overset{s-1}{\mathrm{\&Sigma;}}}{\mathrm{\&beta;}}_{j+1}^2{|w-k_j|}^3$

${\mathrm{\&alpha;}}_3\left(w\right)={\mathrm{\&beta;}}_1^3+{\mathrm{\&beta;}}_2^{3}w+\underset{j=2}{\overset{s-1}{\mathrm{\&Sigma;}}}{\mathrm{\&beta;}}_{j+1}^3{|w-k_j|}^3$

Wherein, s=51, k _j=k ₁+ (j* (k ₅₁-k ₁)/50), order $\mathrm{\&theta;}={[{\mathrm{\&beta;}}_1^1,{\mathrm{\&beta;}}_2^1,...,{\mathrm{\&beta;}}_{\mathrm{<figure-callout\; id="1"\; label="s"\; filenames="HSA00000087516000052.png"\; state="\{\{state\}\}">s</figure-callout>}}^1;{\mathrm{\&beta;}}_1^2,{\mathrm{\&beta;}}_2^2,...,{\mathrm{\&beta;}}_{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="HSA00000087516000011.png,HSA00000087516000031.png"\; state="\{\{state\}\}">s</figure-callout>}}^2;{\mathrm{\&beta;}}_1^3,{\mathrm{\&beta;}}_2^3,...,{\mathrm{\&beta;}}_s^3]}^T,$ Parameter for the needs estimation.

C. the transit data between each operating point test data and each operating point is designated as

Z ^N={u(t),y(t),w(t),t=1,2,...,N}

Use data set Z ^NEach operating point model of difference emulation, predicted output accordingly:

${\hat{y}}^1\left(t\right)={\hat{G}}_1^1\left(q\right)u\left(t\right)$

${\hat{y}}^2\left(t\right)={\hat{G}}_1^2\left(q\right)u\left(t\right)$

${\hat{y}}^3\left(t\right)={\hat{G}}_1^3\left(q\right)u\left(t\right)$

D. through minimizing prediction of output error criterion function, confirm the parameter vector θ among the B:

$V\left(\mathrm{\&theta;}\right)=\frac{1}{N}\underset{t=1}{\overset{N}{\mathrm{\&Sigma;}}}e{\left(t\right)}^2=\frac{1}{N}\underset{t=1}{\overset{N}{\mathrm{\&Sigma;}}}{(y\left(t\right)-[{\mathrm{\&alpha;}}_1\left(w\right){\hat{y}}^1\left(t\right)+{\mathrm{\&alpha;}}_2\left(w\right){\hat{y}}^2\left(t\right)+{\mathrm{\&alpha;}}_3\left(w\right){\hat{y}}^3\left(t\right)\left]\right)}^2$

The data vector of note scheduling variable is:

Then prediction of output error can be expressed as:

Because prediction of output error e (t) is the linear function of parameter vector θ, optimize proposition so and can be summed up as a linear least-squares problem, its separate into:

$\widehat{\mathrm{\&theta;}}={\left[{\mathrm{\&Phi;}}^T\mathrm{\&Phi;}\right]}^{-1}{\mathrm{\&Phi;}}^{T}Y$ Wherein, Y=[y (1), y (2) ..., y (N)] ^T

Utilize said method, pick out parameter vector θ, then at screw displacement w=w ^x(143<w ^x<157) the corresponding linear model in place is expressed as:

$G(q,w^x)={\mathrm{\&alpha;}}_1\left(w^x\right){\hat{G}}^1\left(q\right)+{\mathrm{\&alpha;}}_2\left(w^x\right){\hat{G}}^2\left(q\right)+{\mathrm{\&alpha;}}_3\left(w^x\right){\hat{G}}^3\left(q\right)$

(2) the control method flow process is following:

(1) the optimization time domain P of setting dynamic matrix model predictive controller and control time domain M when emulation, make P=30; M=1. also weighted matrix R is controlled in initialization, error weighting matrices Q, dynamic matrix A (be null matrix this moment); Prediction of output vector ym0; Process output vector yp0, output error vector ey, transposed matrix S etc.

(2) confirm scheduling variable w ^x, w when working control ^xDirectly measure by displacement transducer, combine simulation step length, per 5 step-lengths 0.084mm (during actual the realization is the screw displacement measured value) that successively decreases during emulation. simultaneously according to w ^xValue confirm real process model plant (153.5<w ^x<157, plant=s1; 149<w ^x<153.5, plant=s2; 143<w ^x<149, plant=s3), s1 wherein, s2, s3 represent divide three partial models between the lane place.

(3) with the w that confirms in (1) ^xThe substitution formula $G(q,w^x)={\mathrm{\&alpha;}}_1\left(w^x\right){\hat{G}}^1\left(q\right)+{\mathrm{\&alpha;}}_2\left(w^x\right){\hat{G}}^2\left(q\right)+{\mathrm{\&alpha;}}_3\left(w^x\right){\hat{G}}^3\left(q\right)$ Calculate w ^xLocal linear model G (q, the w of position ^x), and calculate G (q, w ^x) pairing dynamic matrix A ^Wx, with A ^WxDynamic matrix as the current time controller.

(4) through type d=C ^T(inv (A ' * Q*A+R)) * A ' * Q calculation control vector d.

(5) calculate predicted vector ym0, process output vector yp0, and give y. as the actual output assignment of emulation with yp0 (1,1)

(6) calculate output error vector ey.

(7) calculation control effect increment Delta u=d*ey, u=u (t)+Δ u is applied on model and the object as next time performance variable.Note also should carrying out amplitude limit with window function to u, the later u value of amplitude limit is adopted by Practical Calculation, also should be with the u value substitution computing behind the amplitude limit when calculating next time.

(8) repeat above (2)--the step of (7) in this part.

Fig. 9 is a dynamic matrix Model Predictive Control speed closed loop analogous diagram when utilizing the single linear model pick out, control cycle T=1ms wherein, P=30, N=1.Can find out that obviously aircraft pursuit course all certain overshoot can occur at each set point change place.And under pre-set parameters such as same T, P, N, adopt position-based better than adopting the single linear model for the dynamic matrix Model Predictive Control speed closed loop control effect of the overall nonlinear model of scheduling variable.Figure 10 is its control design sketch, and there is certain overshoot at first step place in beginning, but wants much little with respect to what Fig. 7 showed, and two step places of back do not have overshoot basically.Through the contrast of this two width of cloth figure, proved that screw of injection method for control speed of the present invention has good tracking performance to the setting value of injection speed.

The present invention can be based on the energy-saving injection machine servo-drive system; PC and data collecting plate card and RTAI-LINUX data acquisition platform; Adopt the nonlinear model modeling of position-based for the scheduling variable; And, utilize predictive control algorithm through local linearization, the screw speed in the injection process is carried out closed-loop control.Position with screw rod is the scheduling variable, and combines the rate of pressure change in the injection process that injection process is carried out segmentation modeling, has set up overall large-scale nonlinear parameter model; When process is controlled, with overall nonlinear model, carry out local linearization through the method for substitution location parameter, re-use the DMC control algolithm object is controlled, thinking is simple, implements convenient, safety, reliable.Therefore application prospect is boundless.

Claims (3)

1. a method for controlling injection speed of screw of injection molding machine is characterized in that comprising the steps:

(1) utilize high-speed data acquiring device that the injection process of injection machine is carried out the step open-loop test; Obtain gathering the screw of injection speed constantly and the system oil pressure data of screw displacement and injection machine, and further obtain the rate of change of said system oil pressure based on the system oil pressure data of injection machine at each;

(2) according to the rate of change of said system oil pressure and the relation between the screw displacement, to screw rod in the injection process the displacement section of process carry out displacement subsection, obtain the required displacement subsection interval censored data of modeling;

(3) the step open-loop response characteristic corresponding that obtains based on step (1) with screw of injection speed; Confirm the identification signal setting value in order to carrying out identification test, and utilize the identification of high-speed data acquiring device record to test each identification signal measured value, screw of injection speed data and screw displacement data constantly in the whole process;

(4) confirm the time slice data that corresponding identification is tested according to the displacement subsection interval censored data in the step (2) earlier; According to said time slice data resulting identification signal measured value of step (3) and screw of injection speed data are carried out the modeling segmentation again; Obtain between three partial model piecewise intervals and two transition regions, and pick out three linear models corresponding

with each partial model piecewise interval

(5) according to the identification signal measured value between said three partial model piecewise intervals and two transition regions and screw of injection speed data, the screw displacement data of synchronization; What the linear model of three partial model piecewise intervals described in the integrating step (4), foundation were shown below serves as the overall nonlinear model of scheduling variable with the screw of injection displacement:

$y\left(t\right)=({\mathrm{\&alpha;}}_1\left(w\right){\hat{G}}^1\left(q\right)+{\mathrm{\&alpha;}}_2\left(w\right){\hat{G}}^2\left(q\right)+{\mathrm{\&alpha;}}_3\left(w\right){\hat{G}}^3\left(q\right))u\left(t\right)+v\left(t\right)$

Wherein, weight function α ₁(w), α ₂(w) and α ₃(w) be the cubic spline function of screw of injection displacement w;

(6) on-line measurement screw of injection displacement, the screw of injection displacement data that utilizes current time to record carries out linearisation to said overall nonlinear model, obtains linear model;

(7), and utilize this predictive controller that screw of injection speed is carried out closed-loop control with the forecast model of said linear model as the dynamic matrix model predictive controller.

2. a kind of method for controlling injection speed of screw of injection molding machine according to claim 1 is characterized in that: said identification signal is the M sequence.

3. a kind of method for controlling injection speed of screw of injection molding machine according to claim 2 is characterized in that: the sampling period of said M sequence is 10ms.

Images