CN104324861A - Multi-parameter time-varying robot spraying method - Google Patents

Abstract

The invention relates to a multi-parameter time-varying robot spraying method which comprises the following steps: firstly, establishing a free-form surface multivariable spraying model which uses flow of a spraying gun, a spraying distance and a spraying speed as model independent variables; then discretizing an original spraying path into a plurality of sections of minute sub paths, distributing initial values of time-varying spraying process parameters to each sub path and on the basis of the established multivariable spraying model, predicting distribution of an initial coating thickness; next, on the basis of a prediction result on distribution of the coating thickness, carrying out combined optimization on the time-varying process parameters to obtain the optimal process parameters on each section of sub path and finally, obtaining a multi-parameter time-varying spraying path of a surface to be sprayed. According to the method disclosed by the invention, various process parameters are used as variables; by the multivariable spraying model and a free-form surface coating thickness predicting method, the optimal process parameters on the sub paths of the discretized spraying path are obtained; dynamic optimization of the process parameters is implemented; the multi-parameter time-varying robot spraying method has the important effect of improving spraying operation efficiency, quality and safety of a robot.

Description

Control During Paint Spraying by Robot method is become during a kind of multi-parameter

Technical field

Become Control During Paint Spraying by Robot method when the present invention relates to a kind of multi-parameter, belong to Control During Paint Spraying by Robot technical field.

Background technology

In recent years, spray robot is widely used because of features such as its spray efficiency are high, path accuracy is high, reproducible, along with the expansion and deeply of Control During Paint Spraying by Robot application, workpiece shapes becomes increasingly complex, also more and more higher to the requirement of coating thickness and uniformity, Control During Paint Spraying by Robot of how making rational planning for track, Optimizing Process Parameters, the potential of abundant excavator, improves spraying operation quality, become spray robot will faced by important topic.

Traditional Control During Paint Spraying by Robot method will keep invariable as basic principle by technological parameter in whole spraying operation process, to ensure the stability of coating quality, method is: before spraying starts, air pressure-adjusting valve is regulated accurately to set atomization and spraying swath pressure, adjusting coating pressure regulator valve setting gun traffic, carries automatic spray gun to carry out continuous spray perpendicular to surface of the work, constant spray distance and spraying rate to surface of the work by spray robot.And in fact, modern advanced spray robot not only can realize the continuous control of spraying rate, further provided with the device such as gear pump or pneumatic proportional valve, the real-time adjustment of gun traffic can be realized.And, when spraying complex profile product or the region such as inner chamber, turning, in order to prevent collision, be objectively also difficult to ensure the constant of spray distance.Therefore, if the important technical parameters such as spray distance, gun traffic, spraying rate all can be carried out Combinatorial Optimization as variable, dynamic adjustments is realized in spraying process, the optimization space of spraying profile can be increased greatly, for guaranteeing that spraying operation safety, raising Control During Paint Spraying by Robot operating efficiency and coating quality have obvious value and significance.

But, by finding the investigation of published document, patent, industrial products, present great majority are only confined to the optimization to spraying path to the investigation and application of spraying profile optimization, though a few studies has related to the optimization of spraying rate, but be only one-parameter optimization, be not deep into the aspect of multi parameter combinatorial optimization, the optimization method adopted is also most for specific application scenario, lacks versatility.

Summary of the invention

The present invention is directed in conventional machines people spraying method and technological parameter is considered as constant, limit the limitation that spraying profile optimizes space, Control During Paint Spraying by Robot method is become when proposing a kind of multi-parameter, by spray distance, the technological parameter such as spraying rate and gun traffic regards as variable, by on the basis on surface to be sprayed and spraying path discretization, the technological parameter on every cross-talk path is obtained by multivariable spraying model and free form surface coating layer thickness Forecasting Methodology, achieve real-time change and the dynamic optimization of technological parameter in spraying process, effectively can improve complex-curved coating thickness and uniformity control accuracy.

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

1) foundation sprays model with the free form surface multivariable that multiple spraying parameter is model independent variable;

2) discretization of half-space surface to be sprayed is turned to a cloud, some sections of small subpaths are turned to by discrete for original spraying path, for every cross-talk path time become spraying parameter and carry out initializations assignment, spray model based on carried multivariable and primary coat thickness distribution predicted;

3) based on coating layer thickness forecast of distribution result, with coating layer thickness, uniformity for optimization aim on every cross-talk path time become spraying parameter and be optimized, become spraying path when finally obtaining the multi-parameter on surface to be sprayed.

The method of the invention with gun traffic, spray distance, spraying rate for time become spraying parameter.

The oval two β multivariable coating deposition rate model of free form surface that described free form surface multivariable spraying model is is model independent variable with gun traffic, spray distance, spraying rate:

$\mathrm{\τ}(x,y,z)={\mathrm{\τ}}_{\max }^0\·\frac{q}{q_0}\·\frac{1}{p^2}\·\frac{\cos (\mathrm{\θ}+\mathrm{\α})}{\cos \mathrm{\θ}}\·{[1-\frac{x^2}{{\left({\mathrm{pa}}_0\right)}^2}]}^{{\mathrm{\β}}_1-1}\·{[1-\frac{y^2}{{\left({\mathrm{pb}}_0\right)}^2(1-\frac{x^2}{{\left({\mathrm{pa}}_0\right)}^2})}]}^{{\mathrm{\β}}_2-1}$

Wherein, τ (x, y, z) be the coating deposition rate of any point S on free form surface within the scope of spray gun mist cone, (x, y, z) at the coordinate of a S under the local coordinate system of spraying swath center, local coordinate system with spray gun spraying swath center for the origin of coordinates, with spray gun axis direction for Z axis, with spray gun direction of advance for X-axis benchmark coatings deposition coefficient, q ₀benchmark spraying flow, d ₀benchmark spray distance, a ₀, b ₀be the spraying swath ellipse long and short shaft length under benchmark spray distance respectively, q is current gun traffic, and d is current spray distance, θ is the angle of S and spray tip line and spraying swath center local coordinate system Z axis, and α is the angle that a S surface method vows n (s) and spraying swath center local coordinate system Z axis, β ₁, β ₂for β breadth coefficient;

Q ₀, d ₀, a ₀and b ₀record by spraying experiment, β ₁, β ₂with by carrying out repeatedly dull and stereotyped spraying experiment of keeping straight under different technical parameters, obtained by least square fitting after recording coating profile thickness profile data:

$\begin{array}{cc}\min & \underset{v=v_{\min }}{\overset{v_{\max }}{\mathrm{\Σ}}}\underset{d=d_{\min }}{\overset{d_{\max }}{\mathrm{\Σ}}}\underset{q=q_{\min }}{\overset{q_{\max }}{\mathrm{\Σ}}}\underset{x=-a}{\overset{a}{\mathrm{\Σ}}}{(\mathrm{\Γ}(c,t_{\max }^0,{\mathrm{\β}}_1{\mathrm{\β}}_2)-\widetilde{\mathrm{\Γ}}\left(c\right))}^2\\ s.t.& t_1\≤t_{\max }^0\≤t_2\\ & {\mathrm{\β}}_{\min }\≤{\mathrm{\β}}_1\≤{\mathrm{\β}}_{\max },{\mathrm{\β}}_{\min }\≤{\mathrm{\β}}_2\≤{\mathrm{\β}}_{\max }\end{array}$

Wherein that the coating layer thickness that goes out according to carried model inference is about β ₁, β ₂with function expression, the coating profile thickness to be the centre-to-centre spacing of surveying out be c place, [v _min, v _max] be spraying rate adjustable range allowable, [d _min, d _max] be spray distance adjustable range allowable, [q _min, q _max] be gun traffic adjustable range allowable, [β _min, β _max] be β parameter optimization span, [t ₁, t ₂] be benchmark coatings sedimentation coefficient span;

The judgment formula whether some S is within the scope of spray gun mist cone is

$\left\{\begin{array}{c}\frac{x}{{\left({\mathrm{pa}}_0\right)}^2}+\frac{y}{{\left({\mathrm{pa}}_0\right)}^2}\≤1\\ p=1-\frac{z}{d_0}\end{array}\right..$

Described step 2) coating layer thickness Forecasting Methodology be:

First, obtain surface to be sprayed and its original spraying path, discretization of half-space surface to be sprayed is turned to a cloud, with Ω=[s ₁..., s _i..., s _n] ^trepresent, s _i(1≤i≤n) is i-th discrete point;

Then, by discrete for spraying path be the small subpath of m section, with L=[Δ δ ₁..., Δ δ _j..., Δ δ _m] ^trepresent, Δ δ _j(1≤j≤m) for jth cross-talk path, length be Δ l, spray gun is at subpath Δ δ _jinterior movement velocity is v _j, run duration Δ t _j=Δ l/v _j;

For discrete point s any on surface to be sprayed _i(1≤i≤n), first judges whether it is in mist cone coverage, if so, then calculates Δ t based on carried multivariate model _jpoint s in period _icoating deposition rate calculate s _ipoint is at Δ t _jin time, the coating layer thickness of deposition is otherwise, think s _iat Δ t _jcoating deposit thickness in period is zero;

Then spraying process terminates rear some s _icoating layer thickness σ _ibe predicted as

${\mathrm{\σ}}_i=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\frac{{\mathrm{\τ}}_i^j\mathrm{\Δl}}{v_j}$

The average thickness prediction of whole sprayed surface for

${\stackrel{\‾}{\mathrm{\Γ}}}_{\mathrm{\Ω}}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\frac{{\mathrm{\τ}}_i^j\mathrm{\Δl}}{v_j}.$

Described step 3) in variable spraying parameter optimization method on every cross-talk path be:

First, pair time become spraying parameter carry out initialization assignment, make every section of small subpath Δ δ _jon (1≤j≤m) time become spraying parameter be basic process parameter [q ₀, v ₀, d ₀];

Then, be optimized the spray distance on every section of small subpath, method is: for jth cross-talk path Δ δ _j(1≤j≤m), according to path starting point coordinate, the initial spray distance d of method resultant ₀calculate Burners Positions, ripe collision detection algorithm is utilized to carry out interference checking, judge that whether spray gun exists with surface to be sprayed, surrounding environment to interfere, in this way, then increase in the effective spray distance of spray gun or again carry out interference checking after minimizing spray distance, repeatedly perform above-mentioned steps until eliminate interference, obtain the optimum spray distance on every section of small subpath;

And then be optimized the spraying flow on whole spraying path, method is: on the basis optimizing spray distance, calculate the average thickness on surface to be sprayed based on carried coating layer thickness distribution forecasting method and according to

$q=\left\{\begin{array}{cc}\min \{q_{\max },\frac{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;q_0\}& \mathrm{if}{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\&GreaterEqual;{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\\ \max \{q_{\min },\frac{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;q_0\}& f{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}<{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\end{array}\right.$

Calculate the optimum spraying flow q of whole subpath on whole sprayed surface, wherein for expecting coating layer thickness, gun traffic adjustable range [q _min, q _max] determined by spraying experiment.

Finally, carry out two-step optimization to the spraying rate on every section of small subpath, method is: the first step, utilize carry coating layer thickness distribution forecasting method calculate gun traffic optimization after the average thickness on surface to be sprayed if still do not obtain expecting coating layer thickness, then pass through

$v=\left\{\begin{array}{cc}\min \{v_{\max },\frac{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;v_0\}& \mathrm{if}{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\&GreaterEqual;{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\\ \max \{v_{\min },\frac{{\stackrel{-}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;v_0\}& f{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}<{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\end{array}\right.$

Initial optimization is carried out, wherein spraying rate adjustable range [v to the spraying rate v of subpath whole on whole sprayed surface _min, v _max] determined by spray robot performance;

Second step, utilize carry the thickness σ that coating layer thickness distribution forecasting method calculates all discrete points on surface to be sprayed after preliminary spraying rate optimization _i(1≤i≤n), by solving the least square problem of band inequality constraints

$\left\{\begin{array}{c}\min \underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{({\mathrm{\&sigma;}}_i-{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}})}^2\\ s.t.v_{\min }\&le;v_j\&le;v_{\max }\\ -a_{\max }\&le;\frac{v_{j+1}-v_j}{\mathrm{\&Delta;}{t}_j}\&le;a_{\max }\end{array}\right.$

To optimization each cross-talk path Δ δ _jspraying rate on (1≤j≤m) to improve coating uniformity, wherein a _maxfor the maximum of spray gun acceleration allowable, determined by spray robot performance.

Compared with the prior art the present invention, has the following advantages and high-lighting effect:

1. breach conventional machines people spraying method and technological parameter is considered as normal quantitative limitation, the important technical parameters such as spray distance, gun traffic, spraying rate are proposed all to carry out Combinatorial Optimization as variable, change in real time and dynamic adjustments in spraying process, significantly increase the optimization space of spraying profile, for guaranteeing that spraying operation safety, raising Control During Paint Spraying by Robot operating efficiency and coating quality have obvious value and significance.

2. on proposed free form surface, multivariable coating deposition rate model has good generalization ability and precision, for coating layer thickness distribution Accurate Prediction, spraying parameter optimization provide theoretical foundation;

3. proposed discretization coating layer thickness distribution forecasting method can realize the coating layer thickness forecast of distribution on arbitrary surface, and the guarantee for coating layer thickness during Control During Paint Spraying by Robot and the uniformity provides and provides powerful support for.

Accompanying drawing explanation

Control During Paint Spraying by Robot method flow schematic diagram is become when Fig. 1 is multi-parameter.

Fig. 2 is spraying operation schematic diagram.

1-surface to be sprayed in figure, 2-sprays path, and the mist formed during 3-spray gun spraying is bored.

Fig. 3 is that free form surface affects schematic diagram to coating deposition rate.

Fig. 4 is coating profile thickness curve matching schematic diagram.

Fig. 5 is spray distance optimization method schematic diagram.

Detailed description of the invention

Become spraying method during so-called multi-parameter, namely in spraying process, have multiple technological parameter to be no longer constant, but time to become, varied curve when can be understood as.Therefore, core of the present invention is the time varied curve how obtaining technological parameter, is described further the principle of varied curve, method and the course of work when obtaining technological parameter below in conjunction with accompanying drawing.

1. set up multivariable coating deposition rate model

The principle of aerial spraying utilizes compressed air that coating is atomized into molecule, surface of the work is ejected in taper, the attachment of coating cloud particle forms an elliptical region on the surface of the workpiece, coating thick middle surrounding in elliptical region is thin, when spray gun is along spraying path continuous moving, just defining the thin coating in one continuous print thick middle both sides at sprayed surface, by controlling the distances between two spraying paths, uniform coating can be obtained.The forming process of coating is very complicated, relate to a lot of physical process such as fluid flowing, atomization, volatilization, deposition, the factor affecting coating layer thickness distribution is a lot, wherein environment and unit factor comprise spray gun structure, air form, feeder, feeding device, ambient temperature and humidity, dope viscosity, coating material temperature, workpiece temperature, ventilation situation, pin valve opening etc., and technological parameter comprises atomizing air, fan width air, gun traffic, spraying rate, spray distance, overlap distance etc.

Traditional spraying model comprises Cauchy's distributed model, Gaussian distribution model, oval distributed model, parabolic distribution model, β distributes, analyze sedimentation model and built-up pattern etc., its modeling method is under one group of specific spray technological parameter, first carry out spraying experiment, by matching actual measurement coating thickness data Confirming model, this causes model only effective to current process parameters combination, once spraying parameter changes, just need again to test under new argument, again matching modeling, a lot of inconvenience is brought to the technical arrangement plan of spray robot or optimization, therefore necessary foundation a kind ofly to be taken into account the controlled of Control During Paint Spraying by Robot and normal change factor, there is the Spray gun model of good universality and generalization ability.

Spraying operation schematic diagram as shown in Figure 1, point O is spray pattern center, point G is spray tip, for convenience of description, set up local coordinate system OXYZ at spraying swath elliptical center point O, modeling method of the present invention comprises sets up basic model, derivation multivariate model, spraying experiment Confirming model coefficient three steps:

(1) basis spraying model is set up

If sprayed surface is plane, when spray gun is perpendicular to surface of the work and when keeping static spraying, the attachment of coating cloud particle forms an elliptical region on the surface of the workpiece, the thickness distribution of the paint film in elliptical region is class ellipsoid shape, namely the paint film immediately below nozzle is the thickest, and all the other are thinning gradually along X, Y two direction of principal axis everywhere, and X to or Y-direction section on, the thickness curve shape of paint film is similar.Therefore, the present invention's hypothesis at X to on Y-direction section, varnish deposition rate curve all obeys β distribution, and β value is identical on the section be parallel to each other, oval two beta coating deposition rate model is used to describe the coating deposition of any point in elliptic region, shown in (1).

$\begin{array}{c}t(x,y)=t_{\max }\&CenterDot;{[1-\frac{x^2}{a^2}]}^{{\mathrm{\&beta;}}_1-1}\&CenterDot;{[1-\frac{y^2}{b^2(1-x^2/a^2)}]}^{{\mathrm{\&beta;}}_2-1}\\ s.t.\\ -a\&le;x\&le;a,\\ -b\sqrt{1-{(x/a)}^2}\&le;y\&le;b\sqrt{1-{(x/a)}^2}\end{array}---\left(1\right)$

Wherein t _maxfor the deposition maximum in elliptic region, the deposition namely located immediately below spray tip, a and b is oval major and minor axis length, β ₁, β ₂for β breadth coefficient.

(2) derivation multivariate model

When sprayed surface is free form surface, analyze from microcosmic angle, as shown in Figure 2, for piece minimum spraying area Ψ of around any point S in spraying area, if its surperficial method vows that n (s) and Z-direction are not parallel, spraying area is deformed into Ψ ', and the coating deposition rate relation can deriving the two by volume constancy is such as formula (2)

$\frac{t_{\max }\left(\mathrm{\&Omega;}\right)}{t_{\max }\left({\mathrm{\&Omega;}}^{\&prime;}\right)}=\frac{\cos (\mathrm{\&theta;}+\mathrm{\&alpha;})}{\cos \mathrm{\&theta;}}---\left(2\right)$

Wherein, θ is the angle of a S and spray tip line GS and Z axis, and α is the angle that the surperficial method of S point vows n (s) and Z axis under local coordinate system OXYZ.

In technological parameter, ambient parameter, coating parameter, Operation system setting etc. are for same spray robot or be generally constant with a collection of spraying operation, the change of atomizing air, fan width air then easily causes the change of coating quality, therefore selects gun traffic, spraying rate and spray distance as the variable element in Control During Paint Spraying by Robot operation in the present invention.First, it can be used as independent variable to introduce spraying model, derivation multivariable spraying model.Derivation is:

When other technological parameters are constant, only gun traffic q increases, the corresponding increase of coating cloud particle on workpiece is deposited in unit interval, therefore varnish deposition speed also increases thereupon, if the mobility scale of flow is limited, can think that paint film section still obeys identical β distribution, so deposition maximum t _maxbe directly proportional to gun traffic q, shown in (3).

$\frac{t_{\max }\left(q_1\right)}{t_{\max }\left(q_2\right)}=\frac{q_1}{q_2}---\left(3\right)$

When other technological parameters are constant, only spray distance d increases, because spray angle is constant, the major and minor axis length a of spraying swath, b is directly proportional to spray distance d, if the mobility scale of spray distance is limited, can think that spray gun painting rate is constant, from volume constancy principle, coating deposition maximum t _maxwith square being inversely proportional to, shown in (4) of spray distance.

$\frac{t_{\max }\left(d_1\right)}{t_{\max }\left(d_2\right)}={\left(\frac{d_2}{d_1}\right)}^2---\left(4\right)$

When other technological parameters are constant, only spraying rate v changes, due to the flying speed relative to coating cloud particle, the rate travel of spray gun is very little, and therefore varnish deposition speed has almost no change.But, spraying rate affect surface of the work time of covering by spray gun spray torch, that is spraying rate is larger, and the varnish deposition time is shorter, otherwise the varnish deposition time is longer.Change an angle, spraying rate is larger, and the region that spray gun scans within the unit interval is larger, and the film thickness of surface of the work is naturally also thinner.

Note then for any point S on Space Free-Form Surface, first whether be in mist cone according to formula (5) judging point S and shroud within scope,

$\left\{\begin{array}{c}\frac{x}{{\left({\mathrm{pa}}_0\right)}^2}+\frac{y}{{\left({\mathrm{pb}}_0\right)}^2}\&le;1\\ p=1-\frac{z}{d_0}\end{array}\right.---\left(5\right)$

Wherein (x, y, z) is the coordinate of a S under local coordinate system OXYZ, d ₀benchmark spray distance, a ₀, b ₀the spraying swath ellipse long and short shaft length under benchmark spray distance.

In this way, then the coating deposition rate of this point can be written as formula (6),

$\begin{array}{c}\mathrm{\&tau;}(x,y,z)=t_{\max }^0\&CenterDot;\frac{q}{q_0}\&CenterDot;\frac{1}{p^2}\frac{\cos (\mathrm{\&theta;}+\mathrm{\&alpha;})}{\cos \mathrm{\&theta;}}\&CenterDot;{[1-\frac{x^2}{{\left({\mathrm{pa}}_0\right)}^2}]}^{{\mathrm{\&beta;}}_1-1}\&CenterDot;{[1-\frac{y^2}{{\left({\mathrm{pb}}_0\right)}^2(1-\frac{x^2}{{\left({\mathrm{pa}}_0\right)}^2})}]}^{{\mathrm{\&beta;}}_2-1}\\ s.t.\\ d_{\min }\&le;d_0\&le;d_{\max },\\ q_{\min }\&le;q_0\&le;q_{\max }\end{array}---\left(6\right)$

Wherein, q ₀benchmark spraying flow, [d _min, d _max] be spray distance adjustable range, [q _min, q _max] be gun traffic adjustable range, the coating deposition coefficient under benchmark spray parameters, α=arccos (n (s) n (K)).

(3) Optimized model coefficient

After modeling completes, need, by the coefficient entry in the means Confirming model such as spraying experiment and matching optimization, to make model, to there is best precision.Concrete grammar is:

Benchmark spray distance d ₀intermediate value is selected according to the spray distance scope that spray gun product manual is recommended, spray distance adjustable range must not exceed the spray distance scope that spray gun product manual is recommended, for improving Generalization accuracy, this scope should not obtain very large usually, and need ensure that the change of spraying major and minor axis meets linear rule.

Spraying swath ellipse long and short shaft length a under benchmark spray distance ₀, b ₀determined by spraying swath experiment, experimental technique is: spray gun axes normal in plane test specimen and relative to workpiece keep motionless, then within very short time, complete switch rifle action (0.3 ~ 0.5 second), obtain an elliptoid coating, measure major and minor axis length respectively.

Spraying flow adjustment range is determined according to spraying effect, and must ensure that coating atomization is even, spray not sagging, be improve Generalization accuracy, spraying flow adjustment range is unsuitable excessive, and benchmark spray stream measures spraying flow adjustment range intermediate value.

Coating deposition coefficient under benchmark spray parameters determined by dull and stereotyped spraying experiment of keeping straight on β breadth coefficient, concrete principle and method are:

When on test specimen flat board, edge at the uniform velocity sprays with the straight line of long axis normal spray gun, obtain the linearity coating that thick middle both sides are thin, can think 1 O (x in spraying area, y) film thickness be varnish deposition speed to the oval overlay area that an O is sprayed by spray gun time of covering carry out the result of integration, shown in (7)

$\mathrm{\&Gamma;}(x,y)={\&Integral;}_0^{\frac{2b\sqrt{1-{(x/a)}^2}}{v}\mathrm{\&tau;}(x,y)\mathrm{dt}---\left(7\right)}$

Can in the hope of the expression formula of paint film sectional thickness curve according to formula (6), if the parameters such as spraying major and minor axis length, spraying rate, gun traffic, spray distance are known, then the centre-to-centre spacing function that to be the film thickness of c be only about coating deposition coefficient and beta coefficient on paint film section, can write therefore, by carrying out flat board craspedodrome spraying experiment, recording the coating layer thickness under each centre-to-centre spacing on coating cross sections, being designated as optimum coating deposition coefficient can be obtained by least square fitting with β breadth coefficient, as shown in the figure.

In order to improve the Generalization accuracy of model, getting some groups of gun traffics, spray distance, spraying rate carry out many group spraying experiments, getting whole measurement result and be optimized, shown in (8).

$\begin{array}{cc}\min & \underset{v=v_{\min }}{\overset{v_{\max }}{\mathrm{\&Sigma;}}}\underset{d=d_{\min }}{\overset{d_{\max }}{\mathrm{\&Sigma;}}}\underset{q=q_{\min }}{\overset{q_{\max }}{\mathrm{\&Sigma;}}}\underset{x=-a}{\overset{a}{\mathrm{\&Sigma;}}}{(\mathrm{\&Gamma;}(x,t_{\max }^0,{\mathrm{\&beta;}}_1{\mathrm{\&beta;}}_2)-\widetilde{\mathrm{\&Gamma;}}\left(c\right))}^2\\ s.t.& t_1\&le;t_{\max }^0\&le;t_2\\ & {\mathrm{\&beta;}}_{\min }\&le;{\mathrm{\&beta;}}_1\&le;{\mathrm{\&beta;}}_{\max },{\mathrm{\&beta;}}_{\min }\&le;{\mathrm{\&beta;}}_2\&le;{\mathrm{\&beta;}}_{\max }\end{array}---\left(8\right)$

Wherein [v _min, v _max] be spraying rate adjustable range, [β _min, β _max] be β parameter optimization span, [t ₁, t ₂] be that coating deposition coefficient optimizes span.

In order to reduce enchancement factor impact, several are got by the spraying flat board after each spraying experiment and measures cross section, each measurement cross section records the coating layer thickness under each centre-to-centre spacing respectively, then averages as measurement result.

2. free form surface coating layer thickness prediction

Obtain the time varied curve of technological parameter, need first to set up the relation that technological parameter and coating layer thickness distribute, provide theoretical foundation for spraying parameter is optimized.Due to free form surface expression formula often more complicated and be difficult to obtain, its expression formula is directly utilized to carry out calculating coating layer thickness distribution comparatively difficulty, the present invention proposes a kind of discretization coating layer thickness distribution forecasting method, be applicable to the sprayed surface of any shape, concrete grammar is:

1) cloud is turned to by discrete for sprayed surface Ω, with Ω=[s ₁..., s _i..., s _n] ^trepresent, wherein n is discrete point number.

2) obtain the spraying path of surperficial Ω to be sprayed, by its discrete be the small subpath of m section, with L=[Δ δ ₁..., Δ δ _j, K, Δ δ _m] ^trepresent, note jth cross-talk path is Δ δ _j(1≤j≤m), Δ δ _jlength be Δ l;

3) remember that spray gun is at Δ δ _jmoment during starting point is t _j, speed is v _j, then, when spray gun uniform motion, it is at Δ δ _jinterior run duration is Δ t _j=Δ l/v _j.For any one discrete point s on surface patch _i, judge s by formula (5) _iwhether point is in mist cone coverage, and if so, then through type (6) calculates at Δ t _js in period _ithe coating deposition rate at some place then s _ipoint is at Δ t _jin time, the coating layer thickness of accumulation is otherwise, s _iat Δ t _jcoating deposit thickness in period is zero.To s _ithe coating cumulative thickness of point on whole small subpath summation, is a s _itotal coating thickness σ after spraying process terminates _i, shown in (9).

${\mathrm{\&sigma;}}_i=\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&tau;}}_i^j\mathrm{\&Delta;l}}{v_j}---\left(9\right)$

4) average thickness of whole sprayed surface Ω can calculate according to formula (10).

${\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&tau;}}_i^j\mathrm{\&Delta;l}}{v_j}---\left(10\right)$

Try to achieve after coating layer thickness predicts the outcome, can be undertaken visual by MATLAB or 3D sculpting software etc., observe coating layer thickness distribution intuitively.

3. time, become spraying parameter optimization

After the discretization of spraying path, become during technological parameter and be further understood that each cross-talk path Δ δ _jtechnological parameter on (1≤j≤m) is different, is each cross-talk path allocation one group of technological parameter, finally obtains the technological parameter array identical with discretization path dimension, as robot controlling foundation.Concrete method for solving is as follows:

1) spray path to initialize

Import surface to be sprayed and initially spray path, turning to a cloud by discrete for sprayed surface, turning to some sections of small subpaths by discrete for spraying path, for every cross-talk path, for it distributes initial spraying parameter [q ₀, v ₀, d ₀], obtain initial coating layer thickness distribution based on carried coating layer thickness Forecasting Methodology.

2) spray distance is optimized

Spray distance change can affect spraying swath size, thus impact overlap joint rule, destroy coating uniformity.Therefore should ensure the constant of spray distance in spraying operation as far as possible, only at some in particular cases, just regulate spray distance, such as some narrow zone easily causes the collision of spray gun and surface to be sprayed or surrounding enviroment, now can realize keeping away barrier by increasing or reducing spray distance.The optimization method of spray distance is: subpath carries out interference checking piecemeal, for jth cross-talk path Δ δ _j(1≤j≤m), according to path starting point coordinate, the initial spray distance d of method resultant ₀calculate Burners Positions, ripe collision detection algorithm (as bounding box method, dangerous spot detection method etc.) is utilized to carry out interference checking, judge that whether spray gun exists with surface to be sprayed, surrounding environment to interfere, in this way, then increase or again carry out interference checking after minimizing spray distance, repeatedly perform above-mentioned steps until eliminate interference, obtain the optimum spray distance on every section of small subpath, as shown in Figure 5.

2) gun traffic optimization

Consider that the adjustment of gun traffic has certain hysteresis quality, to each cross-talk path is all carried out, although gun traffic optimization can realize in theory, but in actual job, usually spraying workpiece is divided into some pieces of surfaces to be sprayed, and overall gun traffic optimization is only carried out in the spraying path of one piece of surface to be sprayed or a section longer.Therefore, the concrete grammar spraying flow optimization in the present invention is: on the basis optimizing spray distance, based on carried coating layer thickness distribution forecasting method, calculate the average thickness of sprayed surface Ω according to formula (9), if average thickness is less than expectation coating layer thickness, then increase gun traffic, if average thickness is greater than expection thickness, then reduce gun traffic, obtain the optimum spraying flow of whole subpath on whole sprayed surface, specific formula for calculation is such as formula shown in (11).Getting the implication be worth most in formula (11) is ensure that the spraying flow after optimizing does not exceed gun traffic adjustable range,

$q=\left\{\begin{array}{cc}\min \{q_{\max },\frac{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;q_0\}& \mathrm{if}{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\&GreaterEqual;{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\\ \max \{q_{\min },\frac{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;q_0\}& f{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}<{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\end{array}\right.---\left(11\right)$

Wherein q _min, q _maxfor minimum of a value and the maximum of gun traffic adjustable range, for expecting coating layer thickness.

3) spraying rate optimization

Spraying rate adopts two step optimization methods:

The first step, if gun traffic optimization does not obtain expecting coating layer thickness by range of flow restriction, then regulated by bulk velocity and improve the uniformity: based on through the spraying path of spraying flow optimization, coating layer thickness distribution is predicted, the average thickness of sprayed surface Ω is calculated according to formula (9), if average thickness is less than expectation coating layer thickness, then reduce spraying rate, if average thickness is greater than expection thickness, then increase spraying rate, global optimization is carried out to the spraying rate of subpath whole on whole sprayed surface, specific formula for calculation is such as formula shown in (12).Spraying rate after optimization can not exceed spray gun rate adaptation scope.

$v=\left\{\begin{array}{cc}\min \{v_{\max },\frac{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;v_0\}& \mathrm{if}{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\&GreaterEqual;{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\\ \max \{v_{\min },\frac{{\stackrel{-}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;v_0\}& f{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}>{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\end{array}\right.---\left(12\right)$

Second step, calculates the thickness σ of all discrete points on sprayed surface Ω according to formula (8) _i, set up coating layer thickness formula of variance on curved surface, the uniformity problems of coating be converted into the problem of minimum thickness variance D, shown in (13).

$D=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{({\mathrm{\&sigma;}}_i-{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}})}^2---\left(13\right)$

Consider that robot end exists velocity and acceleration constraint, spraying rate can be optimized, to each cross-talk path Δ δ by the least square problem solving band inequality constraints _jspraying rate on (1≤j≤m) carries out single optimization, improves coating uniformity, shown in (14).

$\left\{\begin{array}{c}\min \underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{({\mathrm{\&sigma;}}_i-{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}})}^2\\ s.t.v_{\min }\&le;v_j\&le;v_{\max }\\ -a_{\max }\&le;\frac{v_{j+1}-v_j}{\mathrm{\&Delta;}{t}_j}\&le;a_{\max }\end{array}\right.---\left(14\right)$

Wherein, v _minand v _maxfor minimum of a value and the maximum of spray gun speed allowable, a _maxfor the maximum of spray gun acceleration allowable, v _min≤ v _j≤ v _maxfor constraint of velocity, for acceleration constraint.

Through above step, finally obtain the discretization spraying path of sprayed surface technological parameter array [q corresponding with it _j, v _j, d _j] (1≤j≤m) (i.e. discrete form of Alternative parameter time varying curve), control program according to different vendor and model spray robot requires to be translated into corresponding motion planning and robot control code, gets final product when control carries out multi-parameter and becomes spraying operation.

Claims (5)

1. become a Control During Paint Spraying by Robot method during multi-parameter, it is characterized in that the method comprises the following steps:

1) foundation sprays model with the free form surface multivariable that multiple spraying parameter is model independent variable;

2) discretization of half-space surface to be sprayed is turned to a cloud, some sections of small subpaths are turned to by discrete for original spraying path, for every cross-talk path time become spraying parameter and carry out initializations assignment, spray model based on carried multivariable and primary coat thickness distribution predicted;

3) based on coating layer thickness forecast of distribution result, with coating layer thickness, uniformity for optimization aim on every cross-talk path time become spraying parameter and be optimized, become spraying path when finally obtaining the multi-parameter on surface to be sprayed.

2. become Control During Paint Spraying by Robot method during multi-parameter according to claim 1, it is characterized in that with gun traffic, spray distance, spraying rate for time become spraying parameter.

3. become Control During Paint Spraying by Robot method during multi-parameter according to claim 1 and 2, it is characterized in that: the oval two β multivariable coating deposition rate model of free form surface that described free form surface multivariable spraying model is is model independent variable with gun traffic, spray distance, spraying rate:

$\mathrm{\&tau;}(x,y,z)=t_{\max }^0\&CenterDot;\frac{q}{q_0}\&CenterDot;\frac{1}{p^2}\&CenterDot;\frac{\cos (\mathrm{\&theta;}+\mathrm{\&alpha;})}{\cos \mathrm{\&theta;}}\&CenterDot;{[1-\frac{x^2}{{\left({\mathrm{pa}}_0\right)}^2}]}^{{\mathrm{\&beta;}}_1-1}\&CenterDot;{[1-\frac{y^2}{{\left({\mathrm{pb}}_0\right)}^2(1-\frac{x^2}{{\left({\mathrm{pa}}_0\right)}^2})}]}^{{\mathrm{\&beta;}}_2-1}$

Wherein, τ (x, y, z) be the coating deposition rate of any point S on free form surface within the scope of spray gun mist cone, (x, y, z) at the coordinate of a S under the local coordinate system OXYZ of spraying swath center, local coordinate system OXYZ with spray gun spraying swath center O for the origin of coordinates, with spray gun axis direction for Z axis, with spray gun direction of advance for X-axis benchmark coatings deposition coefficient, q ₀benchmark spraying flow, d ₀benchmark spray distance, a ₀, b ₀be the spraying swath ellipse long and short shaft length under benchmark spray distance respectively, q is current gun traffic, and d is current spray distance, θ is the angle of S and spray tip line and spraying swath center local coordinate system Z axis, and α is the angle that a S surface method vows n (s) and spraying swath center local coordinate system Z axis, β ₁, β ₂for β breadth coefficient;

Q ₀, d ₀, a ₀and b ₀record by spraying experiment, β ₁, β ₂with by carrying out repeatedly dull and stereotyped spraying experiment of keeping straight under different technical parameters, obtained by least square fitting after recording coating profile thickness profile data:

$\min \underset{v=v_{\min }}{\overset{v_{\text{max}}}{\mathrm{\&Sigma;}}}\underset{d=d_{\min }}{\overset{d_{\max }}{\mathrm{\&Sigma;}}}\underset{q=q_{\min }}{\overset{q_{\max }}{\mathrm{\&Sigma;}}}\underset{x=-a}{\overset{a}{\mathrm{\&Sigma;}}}{(\mathrm{\&Gamma;}(c,t_{\max }^0,{\mathrm{\&beta;}}_1,{\mathrm{\&beta;}}_2)-\widetilde{\mathrm{\&Gamma;}}\left(c\right))}^2$

$s.t.t_1\&le;t_{\max }^0\&le;t_2$

β _min≤β ₁≤β _max,β _min≤β ₂≤β _max

Wherein that the coating layer thickness that goes out according to carried model inference is about β ₁, β ₂with function expression, the coating profile thickness to be the centre-to-centre spacing of surveying out be c place, [v _min, v _max] be spraying rate adjustable range allowable, [d _min, d _max] be spray distance adjustable range allowable, [q _min, q _max] be gun traffic adjustable range allowable, [β _min, β _max] be β parameter optimization span, [t ₁, t ₂] be benchmark coatings sedimentation coefficient span;

The judgment formula whether some S is within the scope of spray gun mist cone is

$\left\{\begin{array}{c}\frac{x}{{\left({\mathrm{pa}}_0\right)}^2}+\frac{y}{{\left({\mathrm{pb}}_0\right)}^2}\&le;1\\ p=1-\frac{z}{d_0}\end{array}\right..$

4. become Control During Paint Spraying by Robot method during multi-parameter according to claim 1, it is characterized in that: described step 2) coating layer thickness Forecasting Methodology be:

First, obtain surface to be sprayed and its original spraying path, discretization of half-space surface to be sprayed is turned to a cloud, with Ω=[s ₁..., s _i..., s _n] ^trepresent, s _i(1≤i≤n) is i-th discrete point;

Then, by discrete for spraying path be the small subpath of m section, with L=[Δ δ ₁..., Δ δ _j..., Δ δ _m] ^trepresent, Δ δ _j(1≤j≤m) for jth cross-talk path, length be Δ l, spray gun is at subpath Δ δ _jinterior movement velocity is v _j, run duration Δ t _j=Δ l/v _j;

For discrete point s any on surface to be sprayed _i(1≤i≤n), first judges whether it is in mist cone coverage, if so, then calculates Δ t based on carried multivariate model _jpoint s in period _icoating deposition rate calculate s _ipoint is at Δ t _jin time, the coating layer thickness of deposition is otherwise, think s _iat Δ t _jcoating deposit thickness in period is zero;

Then spraying process terminates rear some s _icoating layer thickness σ _ibe predicted as

${\mathrm{\&sigma;}}_i=\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&tau;}}_i^j\mathrm{\&Delta;l}}{v_j}$

The average thickness prediction of whole sprayed surface for

${\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}=\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&tau;}}_i^j\mathrm{\&Delta;l}}{v_j}.$

5. become Control During Paint Spraying by Robot method during multi-parameter according to claim 1, it is characterized in that: described step 3) in variable spraying parameter optimization method on every cross-talk path be:

First, pair time become spraying parameter carry out initialization assignment, make every section of small subpath Δ δ _jon (1≤j≤m) time become spraying parameter be basic process parameter [q ₀, v ₀, d ₀];

Then, be optimized the spray distance on every section of small subpath, method is: for jth cross-talk path Δ δ x (1≤j≤m), according to path starting point coordinate, the initial spray distance d of method resultant ₀calculate Burners Positions, ripe collision detection algorithm is utilized to carry out interference checking, judge that whether spray gun exists with surface to be sprayed, surrounding environment to interfere, in this way, then increase in the effective spray distance of spray gun or again carry out interference checking after minimizing spray distance, repeatedly perform above-mentioned steps until eliminate interference, obtain the optimum spray distance on every section of small subpath;

And then be optimized the spraying flow on whole spraying path, method is: on the basis optimizing spray distance, calculate the average thickness on surface to be sprayed based on carried coating layer thickness distribution forecasting method and according to

$q=\left\{\begin{array}{cc}\min \{q_{\max },\frac{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;q_0\}& \mathrm{if}{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\&GreaterEqual;{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\\ \max \{q_{\min },\frac{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;q_0\}& f{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}<{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\end{array}\right.$

Calculate the optimum spraying flow q of whole subpath on whole sprayed surface, wherein for expecting coating layer thickness, gun traffic adjustable range [q _min, q _max] determined by spraying experiment;

Finally, carry out two-step optimization to the spraying rate on every section of small subpath, method is: the first step, utilize carry coating layer thickness distribution forecasting method calculate gun traffic optimization after the average thickness on surface to be sprayed if still do not obtain expecting coating layer thickness, then pass through

$v=\left\{\begin{array}{cc}\min \{v_{\max },\frac{{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;v_0\}& \mathrm{if}{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\&le;{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\\ \max \{v_{\min },\frac{{\stackrel{-}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}{{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}}\&CenterDot;v_0\}& f{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}>{\stackrel{\&OverBar;}{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}}\end{array}\right.$

Initial optimization is carried out, wherein spraying rate adjustable range [v to the spraying rate v of subpath whole on whole sprayed surface _min, v _max] determined by spray robot performance;

Second step, utilize carry the thickness σ that coating layer thickness distribution forecasting method calculates all discrete points on surface to be sprayed after preliminary spraying rate optimization _i(1≤i≤n), by solving the least square problem of band inequality constraints

$\left\{\begin{array}{c}\min \underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{({\mathrm{\&sigma;}}_i-{\widehat{\mathrm{\&Gamma;}}}_{\mathrm{\&Omega;}})}^2\\ s.t.v_{\min }\&le;v_j\&le;v_{\max }\\ -a_{\max }\&le;\frac{v_{j+1}-v_j}{\mathrm{\&Delta;}{t}_j}\&le;a_{\max }\end{array}\right.$

To optimization each cross-talk path Δ δ _jspraying rate on (1≤j≤m) to improve coating uniformity, wherein a _maxfor the maximum of spray gun acceleration allowable, determined by spray robot performance.