The content of the invention
The purpose of the present invention is to be directed to influence of the composite parameters to composite suffered drill thrust in drilling
The needs of analysis, there is dispersed and randomness answer to overcome existing Local sensitivity analysis method not to be suitable for mechanical characteristic
The deficiency of condensation material, existing Global sensitivity analysis method is used for reference, invent a kind of global spirit of composite drilling process finite element
Basis of sensitivity analysis method, it considers composite each parameter in drilling and, for the influence degree of drill thrust, given based on side
The composite drilling process finite element Global sensitivity analysis method of difference.
The technical scheme is that:
A kind of composite drilling process finite element Global sensitivity analysis method, its feature are as follows:
First, multigroup random number is produced according to the distribution of composite relevant parameter to input as sample;
Secondly, the three-dimensional Drilling model of composite is established using finite element software;
3rd, above-mentioned three-dimensional Drilling model is called, emulating drilling process by finite element software produces drill thrust-time
Figure, obtain corresponding cutting force by the use of least squares method and exported as sample;
4th, each ginseng is asked according to " state-contingent parametric method " (States Dependent Parameter, SDP) respectively
Conditional expectation of the number for output;
5th, according to the Global sensitivity analysis method based on variance --- " sobol methods " analyzes different parameters for multiple
The size of the influence degree of material drill thrust.
The relevant parameter of described composite is all to obey the stochastic variable being just distributed very much, and the coefficient of variation is 0.06,
And it is uncorrelated, the random sample X that sample size is N is produced according to the distributed constant of each variable using Matlab softwarest(x1,x2,
x3…xi) (t=1,2,3 ... N) inputted as sample.
Described finite element software is ABAQUS/Explicit.
Three-dimensional Drilling model n times are called, corresponding cutting force Y is produced by Abaqus software emulations drilling processt(t=
1,2,3 ... N) exported as sample.
Parameters X is asked according to SDP methods respectively1, X2, X3,…XiFor sample output Y conditional expectation E (Y | X1), E (Y |
X2) ... E (Y | Xi) step is as follows:
According to control theory, establishing state space SDP models is:
observation equation:Yt=Pi,t+ei,t;
SDP model:Pi,t=Pi,t-1+di,t-1;
di,t=di,t-1+ηi,t, ηi,t~N (0, σ2);
In formula:Pi,tFor with state variable Xi,tThe SDP state relevant parameters of change, it is basic variable Xi,tFunction, et
(observation interference), ηi,t(system disturbance) be average be zero, variance σ2And ση 2White Gaussian noise input, its value can use
Maximum likelihood optimization is estimated to obtain.
Then recursive Kalman filtering (kalam Filtreing, KF) is used in circulation backward fit procedure and is accordingly passed
Smooth (Fixed Interval Smoothing, the FIS) rule of fixed interval is returned to estimate to obtain each state relevant parameter Pi,t。
Finally according to formula:
E(Yt|Xi,t)=Yt-Pi,t;
Calculate each parameter for output conditional expectation E (Y | Xi,t);
According to Global sensitivity analysis method --- " sobol methods " calculates each parameter XiGlobal sensitivity to exporting Y
Process is as follows:
According to formula:
Calculate foregoing gained sample output YtVariance V (Y) and each parameter XiFor Y conditional expectation E (Y | Xi) variance
V(E(Y|Xi));
The formula carried according to sobol methods:
Composite parameters are calculated in drilling process for the global sensitivity index S of cutting forcei。
Beneficial effects of the present invention:
The Global sensitivity analysis method of the present invention can meet composite parameters to composite in drilling when institute
By the demand of the impact analysis of drill thrust, existing Sensitivity Analysis Method is overcome not to be suitable for mechanical characteristic dispersiveness and random
The deficiency of composite, when have studied the uncertainty of different input variables (Compound Material Engineering parameter) to composite drilling
Export the influence size of response quautity (drill thrust).
The present invention is advantageous to, to Simplified analysis process, save the time, improves efficiency.
Embodiment:
The present invention is further illustrated with reference to the accompanying drawings and detailed description.
As Figure 1-3.
A kind of composite drilling process finite element Global sensitivity analysis method, it comprises the following steps:
First, multigroup random number is produced according to the distribution of composite relevant parameter to input as sample;
Secondly, the three-dimensional Drilling model of composite is established using finite element software ABAQUS/Explicit;
3rd, the corresponding cutting force of drilling process generation is emulated by finite element software and is used as sample output;
4th, each ginseng is asked according to " state-contingent parametric method " (States Dependent Parameter, SDP) respectively
Conditional expectation of the number for output;
5th, according to the Global sensitivity analysis method based on variance --- " sobol methods " analyzes different parameters for multiple
The size of the influence degree of material drill thrust.
Comprise the following steps that:
First, it is assumed that each parameter of composite is all to obey equally distributed stochastic variable, and it is uncorrelated, utilize
Matlab softwares produce the random sample X that sample size is N according to the distributed constant of each variablet(x1,x2,x3... xi) (t=1,2,
3 ... N) inputted as sample;
Secondly, the three-dimensional Drilling model of composite is established using finite element analysis software ABAQUS/Explicit;
3rd, call above-mentioned composite material model (i.e. the three-dimensional Drilling model of composite) n times, fortran language controls
Abaqus software emulations drilling process processed obtains the curve Y of corresponding cutting force time changet- t schemes, using in least square
Collimation method obtains the numerical value Y represented by the center line of curve in collimation method or arithmetic meant(t=1,2,3 ... N) exported as sample;
4th, parameters X is asked according to SDP methods respectively1, X2, X3,…XiFor sample output Y conditional expectation E (Y |
X1), E (Y | X2) ... E (Y | Xi) step is as follows:
According to control theory related content, state space SDP models are:
observation equation:Yt=Pi,t+ei,t;
SDP model:Pi,t=Pi,t-1+di,t-1
di,t=di,t-1+ηi,t,ηi,t~N (0, σ2);
In formula:Pi,tFor with state variable Xi,tThe SDP state relevant parameters of change, it is basic variable Xi,tFunction, et
(observation interference), ηi,t(system disturbance) be average be zero, variance σ2And ση 2White Gaussian noise input, its value can use
Maximum likelihood optimization is estimated to obtain.
Then recursive Kalman filtering (kalam Filtreing, KF) is used in circulation backward fit procedure and is accordingly passed
Smooth (Fixed Interval Smoothing, the FIS) rule of fixed interval is returned to estimate to obtain each state relevant parameter Pi,t。
Finally substitute into formula:
E(Yt|Xi,t)=Yt-Pi,t;
Try to achieve each parameter for output conditional expectation E (Y | Xi,t);
5th:According to Global sensitivity analysis method --- " sobol methods " calculates each parameter XiTo the global sensitive of output Y
Degree, its process are as follows:
Substitute into formula:
The gained sample of above-mentioned steps 2 output Y is tried to achieve respectivelytVariance V (Y), and each parameter X in step 3iFor Y condition
It is expected E (Y | Xi) variance V (E (and Y | Xi))。
The formula carried according to sobol methods:
Composite parameters are calculated in drilling process for the global sensitivity index S of cutting forcei。
With reference to T700 carbon fibre reinforced composites, details are as follows:
A kind of composite drilling process finite element Global sensitivity analysis method, its flow chart is as shown in figure 1, specific bag
Include following steps:
First, multigroup random number is produced according to the distribution of the relevant parameter of composite to input as sample:With T700 carbon
Exemplified by unidirectional (0 °) plate of fibre reinforced composites individual layer, such as Fig. 2.Its longitudinal stretching ultimate strength Y is obtained by testing1=
1900MPa, longitudinal ultimate compression strength Y2=1000MPa, cross directional stretch limiting chi1=84MPa, transverse compression limiting chi2=
250MPa, longitudinal Young's modulus E2=8.2GPa, horizontal Young's modulus E1=112GPa, Poisson's ratio ν=0.3, G=4500MPa.
Using these parameters as input variable, remaining parameter keeps constant, it is assumed that and each parameter is all to obey equally distributed stochastic variable,
And it is uncorrelated, produce the random sample X that sample size is 100 using uniform designt(t=1,2,3 ... 100) it is used as sample
Input, partial data such as table 1;
Secondly, the three-dimensional Drilling model of composite is established using finite element analysis software ABAQUS/Explicit, is such as schemed
Shown in 2, the cutter is that the cutter is 6 millimeters of drill bits, 118 ° of drift angle, 30 ° of helical angle.Kinematic constraint is in X and Y-direction, feeding speed
Degree is 1.5mm/s, drilling speed 314rad/s.Carbon fibre composite is multidirectional laminate, is 12mm*12mm*5mm's by size
40 are formed according to 0 °/- 45 °/90 °/45 ° directions stackings, and the laminate border is fixed.
3rd, control Abaqus finite element analysis softwares to call above-mentioned composite material model (i.e. multiple using fortran language
The three-dimensional Drilling model of condensation material) 100 times, pass through the song that software emulation drilling process produces corresponding cutting force time change
Line Yt- t schemes, and the numerical value Y represented by the center line of curve is obtained using collimation method in least squares method or arithmetic meant(t=1,
2,3 ... 100) export as sample, partial results such as table 1;
The material parameter of table 1. exports with drill thrust
4th, according to SDP methods ask respectively parameters for sample export Y conditional expectation E (Y | Xt) step is as follows:
According to control theory related content, state space SDP models are:
observation equation:Yt=Pi,t+ei,t;
SDP model:Pi,t=Pi,t-1+di,t-1
di,t=di,t-1+ηi,t;ηi,t~N (0, σ2);
In formula:Pi,tFor with state variable Xi,tThe SDP state relevant parameters of change, it is basic variable Xi,tFunction, et
(observation interference), ηi,t(system disturbance) be average be zero, variance σ2And ση 2White Gaussian noise input, its value can use
Maximum likelihood optimization is estimated to obtain.
Then recursive Kalman filtering (kalam Filtreing, KF) is used in circulation backward fit procedure and is accordingly passed
Smooth (Fixed Interval Smoothing, the FIS) rule of fixed interval is returned to estimate to obtain each state relevant parameter Pi,t。
Finally substitute into formula:
E(Yt|Xi,t)=Yt-Pi,t
Try to achieve each parameter for output conditional expectation E (Y | Xi,t);
5th:According to Global sensitivity analysis method --- " sobol methods " calculates each parameter XiTo the global sensitive of output Y
Degree, its process are as follows:
Substitute into formula:
The gained sample of above-mentioned steps 2 output Y is tried to achieve respectivelytVariance V (Y), and each parameter X in step 3iFor Y condition
It is expected E (Y | Xi) variance V (E (and Y | Xi))。
The formula carried according to sobol methods:
Composite parameters are calculated in drilling process for the global sensitivity index S of cutting forcei.Result of calculation
Such as table 2.
The Calculation of Sensitivity result of table 2.
For ease of analyzing to obtain block diagram, such as Fig. 3.Can with it is concluded that:The Young's modulus of material influences the most on drill thrust
Significantly, Poisson's ratio has negatively influencing to drill thrust size, and the coefficient of rigidity on drill thrust almost without influence.
Part that the present invention does not relate to is same as the prior art or can be realized using prior art.