CN103914581A - Optimization method for plastic injection molding technological parameter - Google Patents

Abstract

The invention discloses an optimization method for a plastic injection molding technological parameter and belongs to technological parameter optimization methods, solving the problems that an existing optimization method for the plastic injection molding technological parameter is large in actual deviation, difficult to adjust accurately and tedious in a technological parameter setting process. The optimization method comprises an initial optimization step and a secondary optimization step, wherein primary and secondary factors of the technological parameter and optimal levels of the factors can be rapidly established in the initial optimization step, the obtained technological parameter serves as the initial technological parameter of the secondary optimization step, and thus testing times on an injection molding machine are greatly reduced; the secondary optimization step adopts a learning algorithm to perform online optimization, so that an optimizing iteration process can be sped up, the convergence of the iteration process can be ensured, and the optimal injection molding technological parameter can be obtained through several times of testing.

Description

A kind of plastic injection moulding process parameter optimization method

Technical field

The invention belongs to process parameter optimizing method, particularly a kind of plastic injection moulding process parameter optimization method.

Background technology

Injection-moulding plastic is to make first evenly plasticizing in heating cylinder of plastics, is then pushed through cooling curing moulding in the die cavity of close die by plunger or screw rod, obtains plastic products.The equipment of injection-moulding plastic is injection machine.

In injection-moulding plastic production run, the factor that affects plastic products quality is many and complicated, comprise starting material, injection machine, mould and moulding process, but after starting material, injection machine, mould etc. are determined, determine that the main factor of plastic products quality is moulding process.Molding technique parameter also becomes the factor that injection-molded item quality is the most quick, cost is minimum of controlling.Therefore regulate and optimize injection-mold process parameter, the injection-molded item for acquisition with stable high-quality amount is significant.

Injection machine technological parameter arranges with optimization method mainly by die trial, traditional die trial adopts hit-and-miss method, rule of thumb constantly carry out parameter trial, it depends critically upon people's experience and level, complicated technology control ability is poor, causes that production efficiency is low, yield rate is low, repeatable accuracy is poor, difficult quality guarantee.Existing injection machine process parameter optimizing method has:

(1) numerical simulation software such as finite element, boundary element proposing according to Polymer Rheology scheduling theory, as the injection-moulding plastic simulation softward HSCAE of Central China University of Science and Technology's research and development, and the HSCAE software of autodesk, inc. etc., although can obtain the technological parameter of recommending and optimizing, but have that counting yield is low, a problem such as material, goods, die information incompleteness in production practices, and optimize the technological parameter obtaining and must manually be input in injection machine.

(2) method of test design, as Orthogonal Experiment and Design, the method can well reduce experiment number, but the method can only be limited in fixed level, rather than optimal case in certain trial stretch, in actual injection moulding production run, operating process is loaded down with trivial details.

(3) expert system, as artificial neural network, case-based reasoning, rule-based reasoning etc., but there is the bottleneck of sample collection in these methods, the performance of correct and sufficient learning sample guarantee neural network.

In addition, said method is off-line and obtains technological parameter, major defect is that the setting of technological parameter is not to synchronize and carry out at die trial scene with optimizing, and can not effectively consider that the time variation presenting between technological parameter and the product quality of the generations such as machine difference such as the external environment condition such as room temperature, coolant water temperature, hydraulic oil pump oil temperature, machine loss, gear train fatigue is related to problem.Technological parameter and actual deviation that off-line obtains are larger, are difficult to accurate adjustment, and technological parameter setting up procedure is also loaded down with trivial details.

In order to understand the present invention, below Conceptions is explained:

Flow of articles is long: the extreme length that refers to the least significant end that flows from cast gate to goods.

Flow channel length: flow channel length is the extreme length of nozzle to cast gate.

Runner weight: the general assembly (TW) of whole flow passage system.

Goods general assembly (TW): the general assembly (TW) of goods in a mould, for single product weight is multiplied by the quantity of goods in a mould.

Projected area: along the direction of folding mould, the area that goods and runner are shared.

Water route quantity: the number of mold cooling water passage.

Body force: body force is to pass through the contactless force of space behavior on all element of fluids, as gravity, inertial force, electromagnetic force etc.

Source item: be a generalized quantity, it has represented all other sums in those transient terms, convective term and diffusion terms that can not be included in governing equation.

Short penetrating: claim again to short, that glue is discontented with, is walked to insufficient fill, product is uneven etc., refers to that die cavity is not full of completely, makes product not full, profile incompleteness is imperfect.

Bubble: the inner phenomenon that forms small volume or bunchiness hole of goods.

Shrink: claim again shrink mark, contraction etc., refer to product surface out-of-flatness, inwardly produce shallow hole or lacuna.

Welding line: claim again that weld mark, welding are bad, Knit-lines, suture line etc., refer to the linear trace forming at product surface when each plastic melt front end meets.Weld line not only affects the image appearance of goods, and affects the mechanical property of goods.

Flow marks: plastic surface produces year colyliform, spiral fashion or the rough phenomenon of nebulous waveform centered by cast gate.

Overlap: claim again flash, overflow edge, burr, batch cutting edge of a knife or a sword etc., refer to that the excessive filling in discontinuous place (normally die joint, vent port, exhaust thimble, slide mechanism etc.) at mould causes the excessive flaw of plastics.

Warpage: refer to that goods produce rotation or twisted phenomena, there is fluctuating in smooth place, and straight flange inwardly or bend outwards or distortion.

Burn: claim again to stick with paste spot, blackspot, black line etc., refer to the dead color point or the dark-coloured striped that occur at product surface.Blackspot and black line are the flaws of same type, the order of severity difference of just burning.

Top is white: there is obvious albinism at the position that ejects of plastic, occurs white shape microcrack.

WLF-Cross model seven parameters: WLF-Cross model is that temperature, shear rate and the pressure model to viscosity influence is described in a kind of conventional being used for, and its expression formula is:

$\mathrm{\η}=\frac{{\mathrm{\η}}_0(T,P)}{1+{\left(\frac{{\mathrm{\η}}_0\stackrel{\·}{\mathrm{\γ}}}{{\mathrm{\τ}}^{*}}\right)}^{1-n}},$

In formula, η is viscosity, η ₀for zero-shear viscosity, τ is material constant, and unit is pa, and n is non-newtonian index, and T is temperature, and P is pressure, for shear rate;

${\mathrm{\η}}_0=D_1\exp \left[\frac{-A_1(T-D_2-D_{3}P)}{A_2+T-D_2}\right],$

In formula, material constant D ₁unit is Pa s, material constant D ₂unit is K, material constant D ₃unit is K/Pa, material constant A ₁dimensionless, material constant A ₂unit is K;

N, τ ^*, D ₁, D ₂, D ₃, A ₁and A ₂be WLF-Cross model seven parameters, its occurrence is determined by plastics classification, the trade mark, manufacturer.

Act on behalf of geometric model: in actual applications, most of working of plastics are all the uniform thin-gage goods of wall thickness, determine the long l of flow of articles and the average wall thickness b of part according to the geometric properties of part and gate location, and complex-shaped surface mould corresponding goods is simplified to the rectangular flat of end band cast gate according to volume equal principle, what the rectangular flat of end band cast gate was plastic products acts on behalf of geometric model.

Mesh generation: two dimension continuous on space or 3D region are carried out to subdivision, it is divided into many triangle subareas or tetrahedron subregion, and determine the process of the node in every sub regions.

Summary of the invention

The invention provides a kind of plastic injection moulding process parameter optimization method, solve that existing plastic injection moulding process parameter optimization method exists large with actual deviation, be difficult to accurate adjustment, the loaded down with trivial details problem of technological parameter setting up procedure.

A kind of plastic injection moulding process parameter optimization method provided by the present invention, comprises initial optimization step and Optimization Steps again, it is characterized in that:

(1) initial optimization step, comprises following sub-step:

(1.1) determine technological parameter and span thereof:

Described technological parameter comprises melt temperature T _p, mold temperature T _m, injection rate V, dwell pressure P, dwell time t _p, cool time t _c;

Melt temperature T _pwith mold temperature T _mspan is corresponding with product material type, from plastics database, determines; Described product material type represents with plastics classification, the trade mark, manufacturer;

Injection rate V, dwell pressure P, dwell time t _p, cool time t _cspan, according to product material type and mold feature, from instance database, obtain by similarity search;

Described mold feature comprises flow of articles length, goods thickest, product weight, cast gate quantity, flow channel length, runner weight, projected area, water route quantity, specific discharge;

Described plastics database comprises some records, every record comprises plastics classification, the trade mark, manufacturer, minimum injection temperature, maximum injection temperature, recommend injection temperature, minimal die temperature, maximum mold temperature, recommend mold temperature, the maximum shear stress, maximum shear speed, melting density, solid-state density, specific heat capacity, eject temperature, heat-conduction coefficient, recommend back pressure, recommend screw lines speed and WLF-Cross model seven parameters, each record derives from plastics physical property table that plastics-production producer provides or the plastics database of injection-moulding plastic CAE simulation software,

Described instance database comprises some records, and every record comprises mold feature, product material type and corresponding V, P, t _p, t _cspan, each record derives from the successful case in analysis case or the injection mo(u)lding manufacturer actual production process of process window module of injection-moulding plastic CAE simulation software;

(1.2) Orthogonal Experiment and Design:

With 6 technological parameter T to be optimized _p, T _m, V, P, t _p, t _cfor factor, the span of each technological parameter is divided into 5 levels, adopt the orthogonal test table L of 6 factors, 5 levels ₂₅(5 ⁶), be identified for 25 groups of combination of process parameters of analog computation, wherein: the symbol that L is orthogonal arrage, 25 line numbers that are orthogonal arrage, represent test number (TN), 5 be the number of levels of the factor in orthogonal arrage, 6 columns that are orthogonal arrage, the number of expression empirical factor;

(1.3) analog computation obtains Forming Quality value, comprises following process:

(1.3.1) whether the geometric figure file that judges goods exists, and is the geometric model in geometric figure file to be carried out to mesh generation; Otherwise according to the average wall thickness of the volume of goods, goods, flow of articles length, cast gate quantity, flow channel length, construct the geometric model of acting on behalf of of goods, carry out mesh generation to acting on behalf of geometric model; The geometric figure file of described goods is provided by user oneself design or die making shop man;

(1.3.2) in the injection machine database of inquiry injection-moulding plastic CAE simulation software, whether there is the injection machine using in the actual production of injection mo(u)lding manufacturer, using this injection machine as analog computation injection machine used, otherwise, in the injection machine database of injection-moulding plastic CAE simulation software, add the injection machine using in the actual production of injection mo(u)lding manufacturer, then set it as analog computation injection machine used;

Described injection machine database comprises some records, every record comprises injection machine model, manufacturer, maximum injection amount, maximum injection speed, maximum injection pressure, maximum dwell pressure, maximum injection stroke, maximum clamp force, screw diameter, maximum screw speed, and each record derives from the machine parameter that injection machine manufacturer provides;

(1.3.3) determine boundary condition according to determined each group of combination of process parameters, employing Finite Element Method, Finite Volume Method or finite difference method solve the geometric model of goods or act on behalf of the NS equation of geometric model, the analog computation result that obtains every group of combination of process parameters, analog computation result comprises that maximum pressure falls temperature before lowest stream maximum shear speed the maximum shear stress τ _maxthe longest cool time

The geometric model of goods or act on behalf of the NS equation of geometric model:

$\frac{\∂\left(\mathrm{\ρ}{u}_i\right)}{\∂t}+\frac{\∂\left({\mathrm{\ρu}}_j{u}_i\right)}{\∂j}=\frac{{\∂\mathrm{\τ}}_{\mathrm{ij}}}{\∂j}+f_i$

$\frac{\∂\mathrm{\ρ}}{\∂t}+\frac{\∂\left(u_i\mathrm{\ρ}\right)}{\∂i}=0$

$\mathrm{\ρC}(\frac{\∂T}{\∂t}+u_i\frac{\∂T}{\∂x_i})=\mathrm{\Φ}+\frac{\∂}{\∂x_i}\left(k\frac{\∂T}{\∂x_i}\right);$

In formula, ρ is plastic melt density, u _ifor plastic melt is at the speed component of i direction, u _jfor plastic melt is at the speed component of j direction, τ _ijfor plastic melt is at i-j plane viscous stress, f _ifor plastic melt is in the suffered body force of i direction, i=1,2,3, j=1,2,3, wherein 1 is rectangular coordinate system in space directions X, and 2 is rectangular coordinate system in space Y-direction, and 3 is rectangular coordinate system in space Z direction; T is temperature, and t is the time, and C is plastic melt specific heat capacity, and k is plastic melt coefficient of heat conductivity; Φ is source item;

(1.3.4) calculate the Forming Quality value of each group of combination of process parameters;

Obtain according to each group of combination of process parameters analog computation result τ _maxwith obtain the Forming Quality value Q of each group of combination of process parameters _r:

$Q_r={\mathrm{\ω}}_1{P}_{\max }^d+{\mathrm{\ω}}_2{T}_{\min }^f+{\mathrm{\ω}}_3{\stackrel{\·}{r}}_{\max }+{\mathrm{\ω}}_4{\mathrm{\τ}}_{\max }+{\mathrm{\ω}}_5{t}_{\max }^c,$

The group sequence number r=1,2 of r group combination of process parameters ..., 25, weights ω ₁, ω ₂, ω ₃, ω ₄, ω ₅all any values in scope (0,1);

(1.4) determine factor order and the combination of theoretical optimal procedure parameters, comprise following process:

(1.4.1) calculating 6 factor 5 levels are respectively organized the signal to noise ratio (S/N ratio) η of combination of process parameters _r:

η _r=-10log ₁₀(Q _r ²), in formula, r=1,2 ..., 25;

(1.4.2) calculate respectively the signal to noise ratio (S/N ratio) extreme difference R of each factor _m:

$R_m=K_{m_{\max }}-K_{m_{\min }},$

$K_{m_{\max }}=\max (K_{1m},K_{2m},K_{3m},K_{4m},K_{5m}),$

$K_{m_{\min }}=\min (K_{1m},K_{2m},K_{3m},K_{4m},K_{5m}),$

In formula, the sequence number m=1 of factor, 2 ..., 6,

K ₁₁＝η ₁+η ₂+η ₃+η ₄+η ₅，

K ₁₂＝η ₁+η ₆+η ₁₁+η ₁₆+η ₂₁，

K ₁₃＝η ₁+η ₁₀+η ₁₄+η ₁₈+η ₂₂，

K ₁₄＝η ₁+η ₉+η ₁₂+η ₂₀+η ₂₃，

K ₁₅＝η ₁+η ₈+η ₁₅+η ₁₇+η ₂₄，

K ₁₆＝η ₁+η ₇+η ₁₃+η ₁₉+η ₂₅，

K ₂₁＝η ₆+η ₇+η ₈+η ₉+η ₁₀，

K ₂₂＝η ₂+η ₇+η ₁₂+η ₁₇+η ₂₂，

K ₂₃＝η ₂+η ₆+η ₁₅+η ₁₉+η ₂₃，

K ₂₄＝η ₂+η ₁₀+η ₁₃+η ₁₆+η ₂₄，

K ₂₅＝η ₂+η ₉+η ₁₁+η ₁₈+η ₂₅，

K ₂₃＝η ₂+η ₈+η ₁₄+η ₂₀+η ₂₁，

K ₃₁＝η ₁₁+η ₁₂+η ₁₃+η ₁₄+η ₁₅，

K ₃₂＝η ₃+η ₈+η ₁₃+η ₁₈+η ₂₃，

K ₃₃＝η ₃+η ₇+η ₁₁+η ₂₀+η ₂₄，

K ₃₄＝η ₃+η ₆+η ₁₄+η ₁₇+η ₂₅，

K ₃₅＝η ₃+η ₁₀+η ₁₂+η ₁₉+η ₂₁，

K ₃₆＝η ₃+η ₉+η ₁₅+η ₁₆+η ₂₂，

K ₄₁＝η ₁₆+η ₁₇+η ₁₈+η ₁₉+η ₂₀，

K ₄₂＝η ₄+η ₉+η ₁₄+η ₁₉+η ₂₄，

K ₄₃＝η ₄+η ₈+η ₁₂+η ₁₆+η ₂₅，

K ₄₄＝η ₄+η ₇+η ₁₅+η ₁₈+η ₂₁，

K ₄₅＝η ₄+η ₆+η ₁₃+η ₂₀+η ₂₂，

K ₄₆＝η ₄+η ₁₀+η ₁₁+η ₁₇+η ₂₃，

K ₅₁＝η ₂₁+η ₂₂+η ₂₃+η ₂₄+η ₂₅，

K ₅₂＝η ₅+η ₁₀+η ₁₅+η ₂₀+η ₂₅，

K ₅₃＝η ₅+η ₉+η ₁₃+η ₁₇+η ₂₁，

K ₅₄＝η ₅+η ₈+η ₁₁+η ₁₉+η ₂₂，

K ₅₅＝η ₅+η ₇+η ₁₄+η ₁₆+η ₂₃，

K ₅₆＝η ₅+η ₆+η ₁₂+η ₁₈+η ₂₄，

(1.4.3) by the signal to noise ratio (S/N ratio) extreme difference R of each factor _mby sequence from big to small, signal to noise ratio (S/N ratio) extreme difference is larger, shows that Forming Quality is more responsive to this factor;

(1.4.4) select level corresponding to maximum signal to noise ratio in each factor, form theoretical optimal procedure parameters combination x _oe, be designated as vector form:

$x_{\mathrm{oe}}={({\hat{T}}_p,\hat{V},\hat{P},{\hat{t}}_p,{\hat{t}}_c,{\hat{T}}_m)}^T,$

In formula, be respectively the corresponding T of level of maximum signal to noise ratio _p, V, P, t _p, t _c, T _m;

(2) Optimization Steps again, comprises following sub-step:

(2.1) the first die trial of injection machine, comprises following process:

(2.1.1) the combination of process parameters x of the 1st die trial ₁=x _oe, by x _oein each technological parameter be set on injection machine guidance panel, operation injection machine carries out die trial, judges whether to exist product defect, is the process of carrying out (2.1.2); Otherwise complete die trial;

Described product defect is the open defect that technologist can directly observe, be divided into+1 class defect of product defect and-1 class defect ,+1 class defect comprise shortly penetrate, bubble, shrink, welding line, flow marks;-1 class defect defect comprises overlap, warpage, it is white to burn, push up, the demoulding is difficult; + 1 class defect and-1 class defect mutual exclusion;

(2.1.2) record and preserve combination of process parameters x ₁and corresponding product defect, carry out sub-step (2.2):

(2.2) adjustment of technological parameter and die trial continuously, comprises following process:

(2.2.1) according to die trial and corresponding product defect, the responsive order of factor of establishing in cohesive process (1.4.3), with the regulation rule of drafting, in each technological parameter span of determining in sub-step (1.1), adjust respectively each technological parameter, form the combination of process parameters x of the w time die trial _w, w=2,

(2.2.2) by combination of process parameters x _win each technological parameter T _p, V, P, t _p, t _c, T _m, being set on injection machine guidance panel, operation injection machine carries out die trial, judges whether to exist product defect, is to carry out (2.2.3), otherwise completes die trial;

(2.2.3) judging whether die trial number of times >=3 and comprise the die trial of existence+1 class defect and the die trial of existence-1 class defect, is to carry out sub-step (2.3), otherwise turns over journey (2.2.1);

(2.3) machine learning obtains actual optimum technological parameter, comprises following process:

(2.3.1) according to the combination of process parameters x of die trial _wand corresponding product defect classification y _w, solve each Lagrange multiplier α by following formula _woptimal value:

$\left\{\begin{array}{c}\min \frac{1}{2}\underset{w=1}{\overset{f}{\mathrm{\Σ}}}\underset{g=1}{\overset{f}{\mathrm{\Σ}}}{y}_w{y}_g{\mathrm{\α}}_w{\mathrm{\α}}_g\left(x_w{x}_g\right)-\underset{w=1}{\overset{f}{\mathrm{\Σ}}}{\mathrm{\α}}_w\\ \underset{w=1}{\overset{f}{\mathrm{\Σ}}}{y}_w{\mathrm{\α}}_w=0\\ {\mathrm{\α}}_w\≥0;w=1,\·\·\·,f\end{array}\right.,$

Obtain optimal value set: α ^*=(α ^* ₁, α ^* ₂..., α ^* _f) ^t,

In formula, product defect classification product defect classification die trial number of times sequence number w=1,2 ..., f, g=1,2 ..., f, f is total die trial number of times; α ^* ₁, α ^* ₂..., α ^* _ffor each Lagrange multiplier α _woptimal value;

(2.3.2) construct optimum lineoid: v ^*x+b ^*=0,

Wherein, weight v ^*=(v ₁, v ₂..., v ₆) be 6 dimension row vectors, x is 6 dimension combination of process parameters variablees, the corresponding corresponding technological parameter of every one dimension, eigenwert α ^* _hfor optimal value set α ^*in any one positive component, x _hfor α ^* _hcorresponding combination of process parameters, y _hfor α ^* _hcorresponding defect classification;

(2.3.3) calculate each combination of process parameters x _wdistance s with optimum lineoid _w:

s _w＝v ^*□x _w+b ^*，w＝1，…，f，

Relatively | s _w| size, obtains minor increment s _opt, according to s _optobtain corresponding combination of process parameters x _opt=(T _popt, T _mopt, V _opt, P _opt, t _popt, t _copt);

(2.3.4) calculate combination of process parameters x _optsubpoint x on optimum lineoid _e:

x _e＝(T _pe，T _me，V _e，P _e，t _pe，t _ce)，

$T_{\mathrm{pe}}=T_{\mathrm{popt}}-\frac{\frac{\left|v_1\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_1},$

$T_{\mathrm{me}}=T_{\mathrm{mopt}}-\frac{\frac{\left|v_2\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_2},$

$V_e=V_{\mathrm{opt}}-\frac{\frac{\left|v_3\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_3},$

$P_e=P_{\mathrm{opt}}-\frac{\frac{\left|v_4\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_4},$

$t_{\mathrm{pe}}=t_{\mathrm{popt}}-\frac{\frac{\left|v_5\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_5},$

$t_{\mathrm{ce}}=t_{\mathrm{copt}}-\frac{\frac{\left|v_6\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_6},$

By x _eas actual optimum combination of process parameters, be set on injection machine guidance panel, operation injection machine carries out die trial, judges whether to exist product defect, is by x _eas x _w, rotor step (2.3), otherwise complete die trial.

Described plastic injection moulding process parameter optimization method, is characterized in that:

In the process (2.2.1) of described sub-step (2.2), described in the regulation rule drafted refer to the rule of each technological parameter being adjusted according to product defect:

A. for+1 class defect:

The short defect of penetrating, increases melt temperature T _p, mold temperature T _m, injection rate V, dwell pressure P, dwell time t _p, reduce t cool time _c;

Air blister defect, reduces melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, increase dwell pressure P, dwell time t _p, reduce t cool time _c;

Shrink defect, reduces melt temperature T _p, mold temperature T _mconstant, injection rate V is constant, increases dwell pressure P, dwell time t _p, reduce t cool time _c;

Welding line defect, increases melt temperature T _p, mold temperature T _m, injection rate V, other parameter constant;

Flow marks defect, increases melt temperature T _p, mold temperature T _m, injection rate V is constant, increases dwell pressure P, dwell time t _p, reduce t cool time _c;

B. for-1 class defect:

Overlap defect, reduces melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, increase t cool time _c;

Warpage defect, melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, increase t cool time _c;

Burn defect, reduce melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, cool time t _cconstant;

Push up white defect, reduce melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, increase t cool time _c;

Demoulding difficulty defect, reduces melt temperature T _p, mold temperature T _mconstant, reduce injection rate, dwell pressure P, dwell time t _p, increase t cool time _c.

The optimizing process of injecting molding machine is divided into initial optimization step and Optimization Steps again by the present invention, initial optimization step is considered material, goods, the impact of die factor, adopt offline optimization, searching process is without die trial, adopt orthogonal experiment design method, accurately consider material, goods, the impact of die factor, can establish fast the secondary factors of technological parameter and the excellent level of factor, obtain the formability of technological parameter, secondary factors relation can also be used for again the reasoning process of the technological parameter of Optimization Steps, the technological parameter obtaining is as the initial technological parameter of Optimization Steps searching process again, greatly reduce the die trial number of times of carrying out on injection machine.

Optimization Steps is considered the impact of the uncertain factors such as injection machine, environment again, adopt learning algorithm on-line optimization, take the technological parameter of die trial and quality of item as data sample, can accelerate the iterative process of optimizing, and guarantee the convergence of iterative process, by die trial several times, just can obtain optimum injecting molding machine.

The present invention is realized on injection molding machine of plastic, can simplify the operation, realize the robotization of die trial process.

Accompanying drawing explanation

Fig. 1 is FB(flow block) of the present invention.

Embodiment

Below in conjunction with embodiment, the present invention is further described.

Experiment is the FANUCS2000i150A of Japan's FA NUC company with injection machine, and experiment material is selected POM (polyoxymethylene) plastics.

As shown in Figure 1, embodiments of the invention, comprise initial optimization step and Optimization Steps again;

(1) initial optimization step, comprises following sub-step:

(1.1) determine technological parameter and span thereof:

Described technological parameter comprises melt temperature T _p, mold temperature T _m, injection rate V, dwell pressure P, dwell time t _p, cool time t _c;

Selected plastics classification: POM, the trade mark: POM Generic Estimates, manufacturer: CMOLD Generic Estimates, can obtain 180 ℃≤T according to the plastics database in the injection-moulding plastic CAE HSCAE of simulation software _p≤ 235 ℃, 50 ℃≤T _m≤ 105 ℃;

Flow of articles is long is 383mm, and goods thickest is 3mm, and product weight 50g, cast gate quantity are 1, flow channel length 110mm, runner weight 15g, projected area 3000, water route quantity 3, specific discharge 36L/min.According to mold feature and plastics classification, mould and the instance database of laboratory, retrieve reasoning and obtain 10cm from China of the Central China University of Science and Technology ³/ s≤V≤150cm ³/ s, 30MPa≤P≤90MPa, 1s≤t _p≤ 13s, 5s≤t _c≤ 33s.

(1.2) Orthogonal Experiment and Design:

With 6 technological parameter T to be optimized _p, T _m, V, P, t _p, t _cfor factor, the span of each technological parameter is divided into 5 levels, in table 1.

Table 1

Adopt the orthogonal test table of 6 factors, 5 levels, in table 2.

Table 2

(1.3) analog computation obtains Forming Quality value, comprises following process:

(1.3.1) whether the geometric figure file that judges goods exists, because the geometric figure file of described goods is by applicant oneself design, therefore the geometric model in geometric figure file is carried out to mesh generation;

(1.3.2) in the inquiry injection-moulding plastic CAE HSCAE of simulation software injection machine database, there is not FANUCS2000i150A type, the machine parameter of FANUCS2000i150A is joined in the injection machine database in the injection-moulding plastic CAE HSCAE of simulation software, set it as analog computation injection machine used; Injection machine model FANUCS2000i150A, the FANUC of manufacturer, maximum injection amount 108g, maximum injection speed 161cm ³/ s, maximum injection pressure 270MPa, maximum dwell pressure 270MPa, maximum injection stroke 145mm, maximum clamp force 150T, screw diameter 32mm, maximum screw speed 400rpm.

(1.3.3) determine boundary condition according to determined each group of combination of process parameters, adopt the injection-moulding plastic CAE HSCAE of simulation software to simulate, this injection-moulding plastic CAE simulation software employing finite element and finite difference method solve the NS equation of the geometric model of goods, obtain the analog computation result of every group of combination of process parameters: maximum pressure falls temperature before lowest stream maximum shear speed the maximum shear stress τ _maxthe longest cool time in table 3;

Table 3

(1.3.4) calculate the Forming Quality value of each group of combination of process parameters;

Obtain according to each group of combination of process parameters analog computation result with obtain the Forming Quality value Q of each group of combination of process parameters _r, in table 3;

$Q_r=0.6P_{\max }^d+0.2T_{\min }^f+0.01{\stackrel{\·}{r}}_{\max }+0.1{\mathrm{\τ}}_{\max }+0.5t_{\max }^c;$

(1.4) determine factor order and the combination of theoretical optimal procedure parameters, comprise following process:

(1.4.1) calculate the signal to noise ratio (S/N ratio) η of each group of combination of process parameters _r, in table 4;

(1.4.2) calculate the signal to noise ratio (S/N ratio) extreme difference R of each factor _m, in table 4;

Table 4

(1.4.3) by the signal to noise ratio (S/N ratio) extreme difference R of each factor _mby sequence from big to small, obtained R by table 4 ₃> R ₁> R ₅> R ₆> R ₂> R ₄so, the sensitivity V > T of Forming Quality to corresponding process parameters _p> t _p> t _c> T _m> P;

(1.4.4) select level corresponding to maximum signal to noise ratio in each factor, for T _p, K ₂₁=-234.3 maximums, corresponding level 2 optimums, for T _m, K ₄₂=-236.7, corresponding level 4 optimums, for V, K ₁₃=-229, corresponding level 1 optimum, for P, K ₅₄=--236.6, corresponding level 5 optimums, for t _p, K ₅₅=--236.9, corresponding level 5 optimums, for t _c, K ₁₆=--236.2, corresponding level 1 optimum, so the combination of theoretical optimal procedure parameters

x _oe＝(194℃，90℃，30cm ³/s，90MPa，10s，5s)；

(2) Optimization Steps again, comprises following sub-step:

(2.1) the first die trial of injection machine, comprises following process:

(2.1.1) by x ₁=x _oe=(194 ℃, 90 ℃, 30cm ³/ s, 90MPa, 10s, 5s) be set on injection machine, operation injection machine carries out die trial, and there is the short defect of penetrating in resulting product;

(2.1.2) record and preserve combination of process parameters x ₁and the corresponding short defect of penetrating of goods, carry out sub-step (2.2);

(2.2) adjustment of technological parameter and die trial continuously, comprises following process:

(2.2.1), according to die trial and the corresponding short defect of penetrating of goods, the sensitivity of the Forming Quality in cohesive process (1.4.3) to each technological parameter, with the corresponding short regulation rule of penetrating defect in the regulation rule of drafting, increases melt temperature T _p, mold temperature T _m, injection rate V, dwell pressure P, dwell time t _p, reduce t cool time _c, the combination of process parameters x of the 2nd die trial of formation ₂=(220 ℃, 95 ℃, 120cm ³/ s, 90MPa, 12s, 4s);

(2.2.2) by the combination of process parameters x after adjusting ₂be set on injection machine guidance panel, operation injection machine carries out die trial again, judges that overlap defect appears in resulting product, carries out (2.2.3);

(2.2.3) now, total die trial number of times=2 and comprise the die trial of existence+1 class defect and the die trial of existence-1 class defect, but the condition of die trial number of times >=3 in do not satisfy condition " die trial number of times >=3 and comprise the die trial of existence+1 class defect and the die trial of existence-1 class defect ", therefore

According to die trial and the corresponding short defect of penetrating of goods, the sensitivity of the Forming Quality in cohesive process (1.4.3) to each technological parameter, with the regulation rule of corresponding overlap defect in the regulation rule of drafting, reduces melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, increase t cool time _c, the combination of process parameters x of the 3rd die trial of formation ₂=(200 ℃, 50 ℃, 70cm ³/ s, 60MPa, 6s, 15s), by combination of process parameters x ₃be set on injection machine guidance panel, operation injection machine carries out die trial again, judge that the short defect of penetrating appears in resulting product, now total die trial number of times=3 and comprise the die trial of existence+1 class defect and the die trial of existence-1 class defect, carry out sub-step (2.3);

(2.3) machine learning obtains actual optimum technological parameter, comprises following process:

(2.3.1) according to the combination of process parameters x of die trial _w, and corresponding product defect classification y _w, in table 5, solve each Lagrange multiplier α by following formula _woptimal value:

$\left\{\begin{array}{c}\min \frac{1}{2}\underset{w=1}{\overset{3}{\mathrm{\Σ}}}\underset{g=1}{\overset{3}{\mathrm{\Σ}}}{y}_w{y}_g{\mathrm{\α}}_w{\mathrm{\α}}_g\left(x_w{x}_g\right)-\underset{w=1}{\overset{3}{\mathrm{\Σ}}}{\mathrm{\α}}_w\\ \underset{w=1}{\overset{3}{\mathrm{\Σ}}}{y}_w{\mathrm{\α}}_w=0\\ {\mathrm{\α}}_w\≥0;w=1,\·\·\·,3\end{array}\right.,$

Obtain optimal value set: α ^*=(α ^* ₁, α ^* ₂, α ^* ₃) ^t=(0.1,0.12,0.2) ^t,

In formula, product defect classification product defect classification die trial number of times sequence number w=1 ..., 3, g=1 ..., 3, f is total die trial number of times; α ^* ₁, α ^* ₂, α ^* ₃for each Lagrange multiplier α _woptimal value;

Table 5

(2.3.2) construct optimum lineoid: v ^*x+b ^*=0,

Weight $v^{*}=(v_1,v_2,\·\·\·{,v}_6)=\underset{w=1}{\overset{3}{\mathrm{\Σ}}}{y}_w{{\mathrm{\α}}^{*}}_w{x}_w=(-5.8,-0.4,-3.4,-7.8,-\mathrm{1.24,2.02}),$ X is 6 dimension combination of process parameters variablees, the corresponding corresponding technological parameter of every one dimension,

Choose optimal value set α ^*in any one positive component α ^* ₃, x ₃for α ^* ₃corresponding combination of process parameters, y ₃for α ^* ₃so corresponding defect classification eigenwert $b^{*}=y_3-\underset{w=1}{\overset{3}{\mathrm{\Σ}}}{y}_w{{\mathrm{\α}}^{*}}_w(x_w.x_3)=0.4;$

(2.3.3) calculate each combination of process parameters x _wdistance s with optimum lineoid _w:

s _w＝v ^*□x _w+b ^*，w＝1，…，f，

Relatively | s _w| size, obtains minor increment s _opt, according to s _optobtain corresponding combination of process parameters x _opt=x ₃=(200 ℃, 50 ℃, 70cm ³/ s, 60MPa, 6s, 15s);

(2.3.4) calculate combination of process parameters x _optsubpoint x on optimum lineoid _e:

x _e＝(T _pe，T _me，V _e，P _e，t _pe，t _ce)，

$T_{\mathrm{pe}}=T_{\mathrm{popt}}-\frac{\frac{\left|v_1\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_1}=207,$

$T_{\mathrm{me}}=T_{\mathrm{mopt}}-\frac{\frac{\left|v_2\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_2}=77,$

$V_e=V_{\mathrm{opt}}-\frac{\frac{\left|v_3\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_3}=90,$

$P_e=P_{\mathrm{opt}}-\frac{\frac{\left|v_4\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_4}=75,$

$t_{\mathrm{pe}}=t_{\mathrm{popt}}-\frac{\frac{\left|v_5\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_5}=7,$

$t_{\mathrm{ce}}=t_{\mathrm{copt}}-\frac{\frac{\left|v_6\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_6}=10,$

So,

x _e＝(T _pe，T _me，V _e，P _e，t _pe，t _ce)＝(207℃，77℃，90cm ³/s，75MPa，7s，10s)

By x _eas actual optimum combination of process parameters, be set on injection machine guidance panel, operation injection machine carries out die trial, and success, completes die trial.

Claims (2)

1. a plastic injection moulding process parameter optimization method, comprises initial optimization step and Optimization Steps again, it is characterized in that:

(1) initial optimization step, comprises following sub-step:

(1.1) determine technological parameter and span thereof:

Described technological parameter comprises melt temperature T _p, mold temperature T _m, injection rate V, dwell pressure P, dwell time t _p, cool time t _c;

Melt temperature T _pwith mold temperature T _mspan is corresponding with product material type, from plastics database, determines; Described product material type represents with plastics classification, the trade mark, manufacturer;

Injection rate V, dwell pressure P, dwell time t _p, cool time t _cspan, according to product material type and mold feature, from instance database, obtain by similarity search;

Described mold feature comprises flow of articles length, goods thickest, product weight, cast gate quantity, flow channel length, runner weight, projected area, water route quantity, specific discharge;

Described plastics database comprises some records, every record comprises plastics classification, the trade mark, manufacturer, minimum injection temperature, maximum injection temperature, recommend injection temperature, minimal die temperature, maximum mold temperature, recommend mold temperature, the maximum shear stress, maximum shear speed, melting density, solid-state density, specific heat capacity, eject temperature, heat-conduction coefficient, recommend back pressure, recommend screw lines speed and WLF-Cross model seven parameters, each record derives from plastics physical property table that plastics-production producer provides or the plastics database of injection-moulding plastic CAE simulation software,

Described instance database comprises some records, and every record comprises mold feature, product material type and corresponding V, P, t _p, t _cspan, each record derives from the successful case in analysis case or the injection mo(u)lding manufacturer actual production process of process window module of injection-moulding plastic CAE simulation software;

(1.2) Orthogonal Experiment and Design:

With 6 technological parameter T to be optimized _p, T _m, V, P, t _p, t _cfor factor, the span of each technological parameter is divided into 5 levels, adopt the orthogonal test table L of 6 factors, 5 levels ₂₅(5 ⁶), be identified for 25 groups of combination of process parameters of analog computation, wherein: the symbol that L is orthogonal arrage, 25 line numbers that are orthogonal arrage, represent test number (TN), 5 be the number of levels of the factor in orthogonal arrage, 6 columns that are orthogonal arrage, the number of expression empirical factor;

(1.3) analog computation obtains Forming Quality value, comprises following process:

(1.3.1) whether the geometric figure file that judges goods exists, and is the geometric model in geometric figure file to be carried out to mesh generation; Otherwise according to the average wall thickness of the volume of goods, goods, flow of articles length, cast gate quantity, flow channel length, construct the geometric model of acting on behalf of of goods, carry out mesh generation to acting on behalf of geometric model; The geometric figure file of described goods is provided by user oneself design or die making shop man;

(1.3.2) in the injection machine database of inquiry injection-moulding plastic CAE simulation software, whether there is the injection machine using in the actual production of injection mo(u)lding manufacturer, using this injection machine as analog computation injection machine used, otherwise, in the injection machine database of injection-moulding plastic CAE simulation software, add the injection machine using in the actual production of injection mo(u)lding manufacturer, then set it as analog computation injection machine used;

Described injection machine database comprises some records, every record comprises injection machine model, manufacturer, maximum injection amount, maximum injection speed, maximum injection pressure, maximum dwell pressure, maximum injection stroke, maximum clamp force, screw diameter, maximum screw speed, and each record derives from the machine parameter that injection machine manufacturer provides;

(1.3.3) determine boundary condition according to determined each group of combination of process parameters, employing Finite Element Method, Finite Volume Method or finite difference method solve the geometric model of goods or act on behalf of the NS equation of geometric model, the analog computation result that obtains every group of combination of process parameters, analog computation result comprises that maximum pressure falls temperature before lowest stream maximum shear speed the maximum shear stress τ _maxthe longest cool time

The geometric model of goods or act on behalf of the NS equation of geometric model:

$\frac{\∂\left(\mathrm{\ρ}{u}_i\right)}{\∂t}+\frac{\∂\left({\mathrm{\ρu}}_j{u}_i\right)}{\∂j}=\frac{{\∂\mathrm{\τ}}_{\mathrm{ij}}}{\∂j}+f_i$

$\frac{\∂\mathrm{\ρ}}{\∂t}+\frac{\∂\left(u_i\mathrm{\ρ}\right)}{\∂i}=0$

$\mathrm{\ρC}(\frac{\∂T}{\∂t}+u_i\frac{\∂T}{\∂x_i})=\mathrm{\Φ}+\frac{\∂}{\∂x_i}\left(k\frac{\∂T}{\∂x_i}\right);$

In formula, ρ is plastic melt density, u _ifor plastic melt is at the speed component of i direction, u _jfor plastic melt is at the speed component of j direction, τ _ijfor plastic melt is at i-j plane viscous stress, f _ifor plastic melt is in the suffered body force of i direction, i=1,2,3, j=1,2,3, wherein 1 is rectangular coordinate system in space directions X, and 2 is rectangular coordinate system in space Y-direction, and 3 is rectangular coordinate system in space Z direction; T is temperature, and t is the time, and C is plastic melt specific heat capacity, and k is plastic melt coefficient of heat conductivity; Φ is source item;

(1.3.4) calculate the Forming Quality value of each group of combination of process parameters;

Obtain according to each group of combination of process parameters analog computation result with obtain the Forming Quality value Q of each group of combination of process parameters _r:

$\mathrm{Qr}={\mathrm{\ω}}_1{P}_{\max }^d+{\mathrm{\ω}}_2{T}_{\min }^f+{\mathrm{\ω}}_3{\stackrel{\·}{r}}_{\max }+{\mathrm{\ω}}_4{\mathrm{\τ}}_{\max }+{\mathrm{\ω}}_5{t}_{\max }^c,$

The group sequence number r=1,2 of r group combination of process parameters ..., 25, weights ω ₁, ω ₂, ω ₃, ω ₄, ω ₅all any values in scope (0,1);

(1.4) determine factor order and the combination of theoretical optimal procedure parameters, comprise following process:

(1.4.1) calculating 6 factor 5 levels are respectively organized the signal to noise ratio (S/N ratio) η of combination of process parameters _r:

η _r=-10log ₁₀(Q _r ²), in formula, r=1,2 ..., 25;

(1.4.2) calculate respectively the signal to noise ratio (S/N ratio) extreme difference R of each factor _m:

$R_m=K_{m_{\max }}-K_{m_{\min }},$

$K_{m_{\max }}=\max (K_{1m},K_{2m},K_{3m},K_{4m},K_{5m}),$

$K_{m_{\min }}=\min (K_{1m},K_{2m},K_{3m},K_{4m},K_{5m}),$

In formula, the sequence number m=1 of factor, 2 ..., 6,

K ₁₁＝η ₁+η ₂+η ₃+η ₄+η ₅，

K ₁₂＝η ₁+η ₆+η ₁₁+η ₁₆+η ₂₁，

K ₁₃＝η ₁+η ₁₀+η ₁₄+η ₁₈+η ₂₂，

K ₁₄＝η ₁+η ₉+η ₁₂+η ₂₀+η ₂₃，

K ₁₅＝η ₁+η ₈+η ₁₅+η ₁₇+η ₂₄，

K ₁₆＝η ₁+η ₇+η ₁₃+η ₁₉+η ₂₅，

K ₂₁＝η ₆+η ₇+η ₈+η ₉+η ₁₀，

K ₂₂＝η ₂+η ₇+η ₁₂+η ₁₇+η ₂₂，

K ₂₃＝η ₂+η ₆+η ₁₅+η ₁₉+η ₂₃，

K ₂₄＝η ₂+η ₁₀+η ₁₃+η ₁₆+η ₂₄，

K ₂₅＝η ₂+η ₉+η ₁₁+η ₁₈+η ₂₅，

K ₂₃＝η ₂+η ₈+η ₁₄+η ₂₀+η ₂₁，

K ₃₁＝η ₁₁+η ₁₂+η ₁₃+η ₁₄+η ₁₅，

K ₃₂＝η ₃+η ₈+η ₁₃+η ₁₈+η ₂₃，

K ₃₃＝η ₃+η ₇+η ₁₁+η ₂₀+η ₂₄，

K ₃₄＝η ₃+η ₆+η ₁₄+η ₁₇+η ₂₅，

K ₃₅＝η ₃+η ₁₀+η ₁₂+η ₁₉+η ₂₁，

K ₃₆＝η ₃+η ₉+η ₁₅+η ₁₆+η ₂₂，

K ₄₁＝η ₁₆+η ₁₇+η ₁₈+η ₁₉+η ₂₀，

K ₄₂＝η ₄+η ₉+η ₁₄+η ₁₉+η ₂₄，

K ₄₃＝η ₄+η ₈+η ₁₂+η ₁₆+η ₂₅，

K ₄₄＝η ₄+η ₇+η ₁₅+η ₁₈+η ₂₁，

K ₄₅＝η ₄+η ₆+η ₁₃+η ₂₀+η ₂₂，

K ₄₆＝η ₄+η ₁₀+η ₁₁+η ₁₇+η ₂₃，

K ₅₁＝η ₂₁+η ₂₂+η ₂₃+η ₂₄+η ₂₅，

K ₅₂＝η ₅+η ₁₀+η ₁₅+η ₂₀+η ₂₅，

K ₅₃＝η ₅+η ₉+η ₁₃+η ₁₇+η ₂₁，

K ₅₄＝η ₅+η ₈+η ₁₁+η ₁₉+η ₂₂，

K ₅₅＝η ₅+η ₇+η ₁₄+η ₁₆+η ₂₃，

K ₅₆＝η ₅+η ₆+η ₁₂+η ₁₈+η ₂₄，

(1.4.3) by the signal to noise ratio (S/N ratio) extreme difference R of each factor _mby sequence from big to small, signal to noise ratio (S/N ratio) extreme difference is larger, shows that Forming Quality is more responsive to this factor;

(1.4.4) select level corresponding to maximum signal to noise ratio in each factor, form theoretical optimal procedure parameters combination x _oe, be designated as vector form:

$x_{\mathrm{oe}}={({\hat{T}}_p,\hat{V},\hat{P},{\hat{t}}_p,{\hat{t}}_c,{\hat{T}}_m)}^T,$

In formula, be respectively the corresponding T of level of maximum signal to noise ratio _p, V, P, t _p, t _c, T _m;

(2) Optimization Steps again, comprises following sub-step:

(2.1) the first die trial of injection machine, comprises following process:

(2.1.1) the combination of process parameters x of the 1st die trial ₁=x _oe, by x _oein each technological parameter be set on injection machine guidance panel, operation injection machine carries out die trial, judges whether to exist product defect, is the process of carrying out (2.1.2); Otherwise complete die trial;

Described product defect is the open defect that technologist can directly observe, be divided into+1 class defect of product defect and-1 class defect ,+1 class defect comprise shortly penetrate, bubble, shrink, welding line, flow marks;-1 class defect defect comprises overlap, warpage, it is white to burn, push up, the demoulding is difficult; + 1 class defect and-1 class defect mutual exclusion;

(2.1.2) record and preserve combination of process parameters x ₁and corresponding product defect, carry out sub-step (2.2);

(2.2) adjustment of technological parameter and die trial continuously, comprises following process:

(2.2.1) according to die trial and corresponding product defect, the responsive order of factor of establishing in cohesive process (1.4.3), with the regulation rule of drafting, in each technological parameter span of determining in sub-step (1.1), adjust respectively each technological parameter, form the combination of process parameters x of the w time die trial _w, w=2,

(2.2.2) by combination of process parameters x _win each technological parameter T _p, V, P, t _p, t _c, T _m, being set on injection machine guidance panel, operation injection machine carries out die trial, judges whether to exist product defect, is to carry out (2.2.3), otherwise completes die trial;

(2.2.3) judging whether die trial number of times >=3 and comprise the die trial of existence+1 class defect and the die trial of existence-1 class defect, is to carry out sub-step (2.3), otherwise turns over journey (2.2.1);

(2.3) machine learning obtains actual optimum technological parameter, comprises following process:

(2.3.1) according to the combination of process parameters x of die trial _wand corresponding product defect classification y _w, solve each Lagrange multiplier α by following formula _woptimal value:

$\left\{\begin{array}{c}\min \frac{1}{2}\underset{w=1}{\overset{f}{\mathrm{\Σ}}}\underset{g=1}{\overset{f}{\mathrm{\Σ}}}{y}_w{y}_g{\mathrm{\α}}_w{\mathrm{\α}}_g\left(x_w{x}_g\right)-\underset{w=1}{\overset{f}{\mathrm{\Σ}}}{\mathrm{\α}}_w\\ \underset{w=1}{\overset{f}{\mathrm{\Σ}}}{y}_w{\mathrm{\α}}_w=0\\ {\mathrm{\α}}_w\≥0;w=1,\·\·\·,f\end{array}\right.,$

Obtain optimal value set: α ^*=(α ^* ₁, α ^* ₂..., α ^* _f) ^t,

In formula, product defect classification product defect classification die trial number of times sequence number w=1,2 ..., f, g=1,2 ..., f, f is total die trial number of times; α ^* ₁, α ^* ₂..., α ^* _ffor each Lagrange multiplier α _woptimal value;

(2.3.2) construct optimum lineoid: v ^*x+b ^*=0,

Wherein, weight v ^*=(v ₁, v ₂..., v ₆) be 6 dimension row vectors, x is 6 dimension combination of process parameters variablees, the corresponding corresponding technological parameter of every one dimension, eigenwert α ^* _hfor optimal value set α ^*in any one positive component, x _hfor α ^* _hcorresponding combination of process parameters, y _hfor α ^* _hcorresponding defect classification;

(2.3.3) calculate each combination of process parameters x _wdistance s with optimum lineoid _w:

s _w=v ^*□x _w+b ^*,w＝1，…，f，

Relatively | s _w| size, obtains minor increment s _opt, according to s _optobtain corresponding combination of process parameters x _opt=(T _popt, T _mopt, V _opt, P _opt, t _popt, t _copt);

(2.3.4) calculate combination of process parameters x _optsubpoint x on optimum lineoid _e:

x _e＝(T _pe，T _me，V _e，P _e，t _pe，t _ce)，

$T_{\mathrm{pe}}=T_{\mathrm{popt}}-\frac{\frac{\left|v_1\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_1},$

$T_{\mathrm{me}}=T_{\mathrm{mopt}}-\frac{\frac{\left|v_2\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_2},$

$V_e=V_{\mathrm{opt}}-\frac{\frac{\left|v_3\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_3},$

$P_e=P_{\mathrm{opt}}-\frac{\frac{\left|v_4\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_4},$

$t_{\mathrm{pe}}=t_{\mathrm{popt}}-\frac{\frac{\left|v_5\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_5},$

$t_{\mathrm{ce}}=t_{\mathrm{copt}}-\frac{\frac{\left|v_6\right|}{\underset{u=1}{\overset{6}{\mathrm{\Σ}}}\left|v_u\right|}\×s_{\mathrm{opt}}}{v_6},$

By x _eas actual optimum combination of process parameters, be set on injection machine guidance panel, operation injection machine carries out die trial, judges whether to exist product defect, is by x _eas x _w, rotor step (2.3), otherwise complete die trial.

2. plastic injection moulding process parameter optimization method as claimed in claim 1, is characterized in that:

In the process (2.2.1) of described sub-step (2.2), described in the regulation rule drafted refer to the rule of each technological parameter being adjusted according to product defect:

A. for+1 class defect:

The short defect of penetrating, increases melt temperature T _p, mold temperature T _m, injection rate V, dwell pressure P, dwell time t _p, reduce t cool time _c;

Air blister defect, reduces melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, increase dwell pressure P, dwell time t _p, reduce t cool time _c;

Shrink defect, reduces melt temperature T _p, mold temperature T _mconstant, injection rate V is constant, increases dwell pressure P, dwell time t _p, reduce t cool time _c;

Welding line defect, increases melt temperature T _p, mold temperature T _m, injection rate V, other parameter constant;

Flow marks defect, increases melt temperature T _p, mold temperature T _m, injection rate V is constant, increases dwell pressure P, dwell time t _p, reduce t cool time _c;

B. for-1 class defect:

Overlap defect, reduces melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, increase t cool time _c;

Warpage defect, melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, increase t cool time _c;

Burn defect, reduce melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, cool time t _cconstant;

Push up white defect, reduce melt temperature T _p, mold temperature T _mconstant, reduce injection rate V, dwell pressure P, dwell time t _p, increase t cool time _c;

Demoulding difficulty defect, reduces melt temperature T _p, mold temperature T _mconstant, reduce injection rate, dwell pressure P, dwell time t _p, increase t cool time _c.