Liu Wei,M Xin,Ji Zhenyun,Lu Weno,Wng Zhengqu,Chen Ling,Liu Weixio,Lu Jiwen

aKey Laboratory for Precision and Non-traditional Machining Technology of the Ministry of Education,Dalian University of Technology,Dalian 116024,China

bAVIC Aerodynamics Research Institute,Shenyang 110034,China

An experimental system for release simulation of internal stores in a supersonic wind tunnel

Liu Weia,*,Ma Xina,Jia Zhenyuana,Lu Wenbob,Wang Zhengqub,Chen Linga,Liu Weixiaoa,Lu Jiwena

aKey Laboratory for Precision and Non-traditional Machining Technology of the Ministry of Education,Dalian University of Technology,Dalian 116024,China

bAVIC Aerodynamics Research Institute,Shenyang 110034,China

Binocular vision system;Pose measurement;Release mechanism;Wind tunnel

Aerodynamic parameters obtained from separation experiments of internal stores in a wind tunnel are significant in aircraft designs.Accurate wind tunnel tests can help to improve the release stability of the stores and in-flight safety of the aircrafts in supersonic environments.A simulative system for free drop experiments of internal stores based on a practical project is provided in this paper.The system contains a store release mechanism,a control system and an attitude measurement system.The release mechanism adopts a six-bar linkage driven by a cylinder,which ensures the release stability.The structure and initial aerodynamic parameters of the stores are also designed and adjusted.A high speed vision measurement system for high speed rolling targets is utilized to measure the pose parameters of the internal store models and an optimizing method for the coordinates of markers is presented based on a priori model.The experimental results show excellent repeatability of the system,and indicate that the position measurement precision is less than 0.13 mm,and the attitude measurement precision for pitch and yaw angles is less than 0.126°,satisfying the requirements of practical wind tunnel tests.A separation experiment for the internal stores is also conducted in the FL-3 wind tunnel of China Aerodynamics Research Institute.

1.Introduction

Nowadays,stealth performance,supersonic cruise performance and large overload maneuvering capability are highly required for the new generation of aircrafts.1Because of the increasingly hostile environment in which these aircrafts are required to operate,it is significant to increase the aircraft speed and decrease its radar cross section to encourage survivability.2One of the primary measures of the effectivenessof military combat aircraft is the ability to release stores(weapons,fuel tanks,pods,etc.)safely and accurately.3Therefore,it is significant to measure the aerodynamic parameters of released internal stores to guide the proper design of aircrafts and release mechanisms.

Aerodynamic parameters of internal stores can be measured with some existing methods such as free drop testing,aerodynamic grid testing,captive trajectory systems(CTS)and computational fluid dynamics(CFD).However,disadvantages exist in the above methods as well.For example,the grid testing method would cause mass of data to be calculated for an experiment,which requires excessive test items and long experimental periods.Both the aerodynamic grid testing method and CTS may require expensive and time-consuming tests in ground facilities and both approximate the dynamic event using quasi-steady assumptions.4Likewise,CFD predictions of the store separation event have been limited to feasibility demonstrations and to occasional problems that are extremely difficult to simulate in the wind tunnel,such as the mutual interaction during the release of multiple stores.5On the other hand,free drop testing has the advantage of including the dynamic effects without limitations on the store attitude or the number of stores released.6Hence,it has been extensively adopted in multi-bodies’separation experiments to measure their aerodynamic parameters.

Static pressure gradient forms in the payload bay as it opens,which brings significant upward force moment towards the store.With the influence of force moment,the store release is not able to be realized.The store even has a risk of flying back into the payload bay when ignited,threatening the safety of the aircraft.Particularly,when the cavity opens during supersonic cruise,the interaction between the strong vortex sheet inside the cavity and the shear layer outside may create aerodynamic noise.Resonance caused by aerodynamic noise may damage or interfere with the electronic components inside the released store,reducing the operational capability.Simulative experiments of the free separation of internal stores in wind tunnels can provide critical parameters for aircraft internal store separation,which contribute significantly to the safety of aircrafts.

For simulative experiments,extensive achievements have been realized in the study of internal model release mechanism.Keen et al.proposed a separation system using an explosive bolt to ensure suspension and a single spring to provide ejection load.7The mechanism has a simple structure and could be driven conveniently.The release force can be loaded when the restraint disappears,and then the store is ejected at a certain initial velocity actuated by the spring pre-tightening force.However,the release force cannot be controlled precisely,and is not able to be adjusted after being assembled.Besides,the store has a tendency to roll backward,influencing the safety of the separation.Johnson et al.proposed a separation system for internal stores.8The store is fixed on the system by an explosive bolt and driven by double springs.The advantage of this system is that the kinetic parameter of the store can be adjusted by changing the preload of the springs,and the tendency to roll backward is avoided effectively.However,it is difficult to ensure the synchronization of two springs,affecting the steady performance of the system.Murray et al.presented an ejection system in which a pneumatic actuator with two air cylinders provided power and a ball brake driven by a cylinder to provide suspension.9Firstly,the two cylinders in the pneumatic actuator are inflated to a predetermined value,and then the ball brake is driven backward by the cylinder,breaking the suspension between the system and the store.Then,the store is released at a high speed by the release force provided by the pneumatic actuator.The system provides adjustable aerodynamic parameter with subtle wide-range control by changing the predetermined value of the actuator cylinders,and is highly repeatable.However,it is hard to control the two air cylinders instantaneously,and perform an instantaneous release force load when the ball brake activates a release.Besides,the system maintenance will be a tough job with its complex configuration.

For the pose measurement of the internal store,due to the small volume,high speed and rolling signature of the store,higher requirement is needed from the measurement system.Compared with other measurement methods,the photogrammetry is advantageous to measure pose parameters of high speed rolling small objects in narrow adverse measuring environment because of fast frequency response,noncontact and high precision characteristics.10,11The OptotrakTMmeasurement system,developed by the Northern Digital Inc.of Canada,has been successfully applied to the aerodynamic elasticity and angle measurement of wind tunnel models.12Three precisely calibrated linear CCDs are used as displacement sensor.The 3D coordinates of each marker can be obtained in real time through capturing the near-infrared light given out by the active luminous markers.The system has some advantages in measurement accuracy and stability.However,the big volume of embedded markers applied in the system cannot satisfy the high speed pose measurement demand of small volume objects.Murray employed optical tracking system based on the particle image velocimetry(PIV)to obtain the pitch angle,angular velocity and related pose information of the high speed target.13However,rolling angles cannot be obtained due to some technical constraints.NASA LaRC developed a pose measurement system for wind tunnel models based on binocular vision.14Images of the model are captured continuously and synchronously by two CCD cameras through the observation window.The pose parameters of the wind tunnel model can be calculated after measuring the 3D centroid coordinates of circular markers.However,this method cannot solve the problem of hidden phenomenon when the model is rolling.Martinez et al.employed a system based on a single camera to get the 3D displacements and rotations of targets.15The system is useful in the case of limited optical access.However,measurement accuracy is majorly associated with the position precision of markers,which affects the measurement accuracy of the system,especially when measuring the displacement of the model’s center of mass along the camera optical axis.

In this paper,we design an experimental system for the simulation of internal stores separation in supersonic wind tunnels,realizing internal store release and high speed pose measurement.This paper is organized as follows:Section 3 introduces the simulative release mechanism for internal stores,which can ensure stable controllable release of the scaled model.Section 4 introduces the scaled model according to the geometric and dynamic characteristics of an actual munition whose type is highly classified.The dynamic characteristics of the model is then measured and adjusted utilizing a binocular-vision system.Section 5 outlines the pose parameters acquisition method of small-volume high speed rolling targets in dark wind tunnels based on a binocular-vision system developed in this paper.In Section 6,experiments are conducted to verify the precision of the measurement system as well as the stability of the simulative separation mechanism in both the laboratory and the wind tunnel.Finally,the conclusion is presented in Section 7.

2.Experimental method

In this paper,we employ the free drop testing method to measure the aerodynamic parameters of the internal stores.The internal store is released from the aircraft model in a wind tunnel,and then its moving trajectory is captured by two high speed cameras.The schematic diagram of the simulative experiment system is shown in Fig.1.Firstly,the internal store model is installed on the release mechanism with markers attached properly on its surface.Secondly,the model is separated stably by the release mechanism through the control system.Meanwhile,the pose measurement system is triggered to capture the real-time images of the store.Finally,the position and attitude measurement is achieved by processing the sequential images using an image processing center(IPC).

3.Design of release mechanism

Design requirements of the store release mechanism are listed as follows:(A)the ejector must provide reliable suspension before the release process;(B)the store release mechanism must be installed in a limited space because of the narrow space inside the aircraft model;(C)the mechanism should be able to realize a small stroke release within a limited space,and a fine adjustability of the end-of-stroke velocity and angular velocity is required.Given those requirements,a release mechanism for internal stores is designed.With this mechanism,high speed separation of the model in a narrow space can be realized.The end-of-stroke velocity can be adjusted from 1.5 m/s to 8 m/s,while the angular velocity is adjustable within the range of 50(°)/s to 250(°)/s.A release hook is designed to suspend the model and a single pressurized piston is added to provide loading force.A six-bar linkage with two synchronizing bars is designed to ensure the initial and endof stroke position of the model.The concrete structure of the release mechanism is shown in Fig.2.

The lengths of the linkages in the mechanism are calculated with the method of coordinate resolution,using the mechanism’s initial and terminal states as the constraints.Fig.3 shows the initial and terminal states of the mechanism(synchronizing bars are not shown in the figure).

As shown in Fig.3,the connecting rodsl2andl′2synchronize with each other and rotate at the same angular speed,so dol3andl4.The constraint equations for the mechanism in its initial state and terminal state can be set up as follows:

wherel1is the length of the supporting structure;l2,l′2,l3,l4are the lengths of the connecting rods labeled in Fig.3 as well;l5is the length of the motion link;α1is the angle between the initial position of the connecting rod,l2,and the horizontal plane,and the corresponding value is α2whenl2is in its terminal position;β1is the angle between the initial position of connecting rod,l3,and the horizontal plane,while β2is the angle between its terminal position and the horizontal plane;γ1is the angle between the initial position of the connecting rod,l4,and the horizontal plane,while γ2is the angle between its terminal position and the horizontal plane;his the retracted length of the mechanism;sis the extended length of the mechanism;δ is the angle between the terminal position of the motion link and the horizontal plane.α1，β1，γ1can be obtained with Eq.(1),and α2，β2，γ2can be obtained with Eq.(2).The Eq.(3)provides the final expressions for α2，β2and γ2:

Combining the above equations with the constraints when mechanism is in its initial condition,we can easily obtain the unknown parameters.Then,the synchronizing bars can be designed based on the synchronism demandsofthe mechanism.

4.Design of store model

The release characteristics of the scaled model should be highly identical to those of the real store.In order to meet this requirement,the design of the model’s geometric profile and parameters such as mass,center-of-gravity and moment of inertia,should obey the dynamic similarity principle.Generally,scaled models are designed according to different mass properties of the stores.Namely,a model’s center-of-gravity and moments of inertia must be adjustable to obtain models with different mass properties.Therefore,we use counterweights made of metal installed inside the model shell which is made of small density material.Then the adjustment of the mass properties can be realized by changing the location of the counter weights.The specific structure of the scaled model is shown in Fig.4.

In order to measure the model’s moment of inertia,the model is installed on a special fixture to create a pendulum mechanism shown in Fig.5.The ‘pendulum” is pulled back to a small angular displacement and then released,which causes it to swing back and forth.Simultaneously,the trajectory of the model is captured by the high speed binocular cameras.14Then,an eigenvalues based fit plane method is adopted to analyze the trajectory.Utilizing a curve fit method considering the effects of friction moment from the bearing and air resistance,we acquire the trajectory in a single plane.Finally,the model’s moment of inertia is calculated based on the compound pendulum principle.

Since the entire ‘pendulum” is made up of the special fixture and the store model,the model’s moment of inertia,Job,can be obtained from the pendulum equation below:

whereJdis the moment of inertia of the special fixture relative to the models’center-of-gravity;gis the gravitational constant;mobis the mass of the store model;mdwhich is measured via an electronic balance is the mass of the special fixture;lobis the length between the model’s center-of-gravity and the horizontal axis;ldis the length between center-of-gravity of the special fixture and the horizontal axis;Tobis the period of oscillation in entire ‘pendulum”.Therefore,the model’s moment of inertia,Job,can be obtained afterTobbeing measured with the binocular vision method proposed in Ref.16.

Models can be adjusted to satisfy different parameter requirements according to practical models using the proposed method.We use two kinds of models with different moments of inertia and masses,which are named 1 and 2,respectively.

5.Pose measurement system

Photogrammetry has the advantage of conducting a high speed noncontact measurement in complex experimental environment against many other measurement methods,providing an effective solution to the pose measurement of high speed and rolling objects in wind tunnels.Thus,in this paper,a binocular vision measurement method is adopted to measure the instantaneous pose parameters of internal store separation.We employ Zhang’s calibration method to complete camera calibration.17A compound calibration target which is the combination of a cross-shaped target and a planar target is utilized in this paper.Its real structure is shown in Fig.6.The compound calibration target has the advantage of covering the entire field of view,reducing the impact of lens distortion which may cause low measuring precision.Besides,markers on the target are made of low-cost retroreflective material,which is advantageous in dark wind tunnels.

To conduct the experiment,retrore flective tape markers are firstly arranged on the model surface as designed.18The image sequence of the moving target is captured by the binocular cameras with low illumination light sources surrounding the camera lenses providing the light environment.19,20The results of the image acquisition are shown in Fig.7(a).Then,image processing techniques are used to trace markers on the model,and the extraction of the coordinate of center-of-gravity at sub-pixel accuracy is realized utilizing the gray centroid method.19Fig.7(b)shows the trajectories of markers obtained after image processing.A limit-restraint method based on epipolar constraint and the layout constraint are introduced to recognize and match the markers in the left and right images.18In this way,the 3D coordinates of the markers on the internal store model are calculated.

In this paper,utilizing the 3D coordinates of the markers on the model surface,the pose information of the high speed object can be obtained indirectly.An optimizing method for the coordinates of markers based on a priori model which will be presented below is proposed in this paper.Since the target is cylindrical,the distances from the target axis to points on its surface are equal.Besides,a priori model can be established when the target is in a static state.In this paper,the target axis and the coordinates of feature points are optimized with the constraints of the equal distances and priori model to improve spatial precision of markers.The optimization objective function is given by

Then,as the position relationship between the markers and the store model is given as designed,the position of the model’s center-of-gravity can be calculated precisely.Finally,a local coordinate system on the model is established.18By computing the coordinate transformation between the initial reference coordinate system(IRCS)and the local reference coordinate system(LRCS),we acquire the pose information of the target.The coordinate transformation process is shown in Fig.8.

The relationship between the point P=[XYZ]Tin IRCS and the corresponding P′=[X′Y′Z′]Tin LRCS is given by

where R and T represent the rotation and translation matrix between IRCS and LRCS.

And

where-θZ,-θXand-θYrepresent the three attitude parameters of the target in the world coordinate system—the yaw,pitch and roll,respectively;tX,tYandtZrepresent the displacements of the target in three directions:theX,YandZ,respectively.According to the coordinates of the chosen points mentioned above,the position and attitude parameters of the target can be computed.

Real-time processing of the images,accurate tracking of the markers and reestablishment of the coordinate system can be realized through the software system.Then,with accurate estimation of pose information of the target,the trajectory of the released model can be reproduced precisely.

6.Verification experiments in laboratory and wind tunnel

6.1.Experimental system for store separations

The simulative experimental system consists of a pose measurement system,an ejection mechanism and a control system(Fig.9).The ejection mechanism is utilized to eject the target out of the tank at a particular velocity,angular velocity and angle.At the same time,the vision measurement system is triggered to measure the pose parameters of the target.

The wind tunnel is a dark narrow enclosure and internal test space which can only be monitored through an observation window of a particular shape.Considering all the constraints caused by the actual wind tunnel structures,we set up ahigh speed binocularvision measurementsystem(Fig.10).The binocular vision measurement system based on the proposed measurement method contains two high speed cameras (FASTCAM SA-X), two wide-angle lenses(Nikon17-35),two low angle lights,a shockproof platform,two electronic control platforms,and a processing terminal.To restrain the influence of vibration caused by the wind tunnel at runtime,two cameras are placed symmetrically on the customized air flotation shockproof table.The camera system is installed on the electric control platform so that cameras can be translated and rotated to a proper position.Lights surrounding the lenses are employed to meet the light requirements of images being captured.

6.2.Repeatability testing of release mechanism

The repeatability of a mechanism has significant influence on the experimental results.Therefore,not only should the mech-anism guarantee a stable separation,it should also present repeatable results in many test runs.An optical tracking method was used to test the repeatability of the store drop mechanism.Several experiments in the same condition were conducted in this paper.

A model of a particular type was dropped many times to verify the repeatability of the mechanism.The position and angle of the released model in each test were measured using the proposed vision measurement system.And repeatability testing results of the model are shown and compared in Table 1.It can be noticed that the standard deviations of the model’s velocity and angular velocity at the time when it ejects from the aircraft are 0.015 m/s and 2.6(°)/s,indicating excellent repeatability of the release mechanism.

Table 1 Repeatability testing of mechanism system.

6.3.Determining relations between ejection load and initial motion parameters(both velocity and angular velocity)for different types of models

In order to guarantee that the model ejects out of the cabin with preset velocities and angular velocities,relations between the ejection load and initial velocity and angular velocity for different types of models have to be studied beforehand.However,for circumstances with the same mechanism and pressure,changes of the model’s mass and other parameters will lead to the change of the velocity and angular velocity.Therefore,experiments for models with different parameters need to be conducted.In this paper,experiments for two types of models named 1 and 2 are given,and curves indicating the relations between the air pressure and model’s velocity and angular velocity are shown in Figs.11 and 12.

As shown in Figs.11 and 12,the velocity and angular velocity of both Model 1 and Model 2 increase as the ejection pressure increases.Since Model 2 is heavier than Model 1,its initial velocity and angular velocity are lower than those of Model 1 under the same ejection pressure.On the other hand,the models’motion parameters are not in perfect linear relations with the ejection pressure,because the ejection pressure varies during the accelerating stroke.

6.4.Accuracy testing of measurement system

We have made improvements for both the calibration target and computation method comparing with the method proposed in Ref.13.As a result,the precision stability for the measurement has been improved although the measurement accuracy was not increased largely.In this paper,targets are fixed on the high-precision motion control stages.The position measurement precision in theX,Y,Zdirections is veri fied through capturing the images of the moving target when the motion control stages move at fixed distances in theX,Y,Zdirections with high precision.Then the attitude measurement precision of the measurement system is verified via rotating the target around theX,Y,Zaxes at fixed angles.

Experimental facilities are shown in Fig.13.Each high precision motion control stage is built up with a displacement platform and a rotary platform.The motion control stage with a position precision of 0.001 mm is used to control the target so that it moves at a specified distance.The displacement accuracy veri fication tests are conducted in 24 groups,and the target is moved 20 mm in each group.The angle accuracy veri fication tests include 24 groups of models,and each rotates at five degrees around the three axes.Precision of the measurement results after being optimized with the priori model is shown in the figures below,among which Fig.14(a)is the precision of displacement measurement and Fig.14(b)is the precision of angular measurement.

As shown in Fig.14,the mean square error of the displacement measurement inXdirection is 0.091 mm,inYdirection 0.102 mm,and inZdirection 0.098 mm.The errors mainly come from the errors of the extraction of markers and camera calibration.The mean square error of the pitch angle precision is 0.101°,while that of the yaw angle precision is 0.106°.The high precision solution of the measurement system for the pitch angle and yaw angle results from the high precision of axis fitting.The roll angle precision is 0.642°for the model’s diameter is so small that it is more difficult to measure its roll angle with high accuracy.In conclusion,the experimental results indicate that the proposed method exhibits high accuracy for the position and attitude measurement of high speed rolling targets.

6.5.Verification experiment in supersonic wind tunnel

To calculate the pose parameters of models,experiments are conducted in the closed supersonic wind tunnel(FL-3).Specifically,the aircraft model is installed in the test section of wind tunnel with the ejection mechanism installed inside the aircraft model.Besides,the measurement system is arranged outside the test section to measure the pose parameters of the model in realtime through the observation window.Pose measurement experiments of high speed rolling models with different moments of inertia at different Mach wind speeds(Ma=0.2,0.4,1.5)are completed in a transonic wind tunnel.Afterwards,the pose parameters of wind tunnel models expressed in the world coordinate system can be obtained.Additionally,a particular point on the door of the airplane cabin is selected as the origin of the world coordinate system and theX-axis is established according to the right-handed system.Fig.15 shows the position measurement results of Model 1 and Model 2 with the air pressure of 9.3×105Pa and the wind speed of 1.5Ma.Fig.16 shows the angular measurement results of Model 1 and Model 2 with the air pressure of 9.5×105Pa and the wind speed of 1.5Ma.

Experiments for scaled models were conducted in the wind tunnel environment and the position and attitude information of the model was successfully measured,which indicates the ability of high speed and intellectualized pose measurement of the proposed method and measurement system.It can be analyzed from the pitch angle measurement results that Model 1 obtained a pitch up attitude within a small value coming through the shear layer while Model 2 obtained relatively steady attitude.Whether they will obtain pitch up attitudes or not,the pose information of different types of models in different release conditions can be measured using the proposed system,ensuring the safety of experiments for real aircrafts.There are three main factors affecting the accuracy of experimental results:(A)the air pressure will not remain constant when the cylinder moves forward and a higher ejection speed needs higher air pressure,which affects the stability of the ejection mechanism and results in unstable release characteristics of the models;(B)problems such as the large measuring range and limited resolution of the ultra-high speed cameras directly affect the pose measurement precision of the target;(C)the synchronism of the system would be difficult to control due to the complex wind tunnel environment and the long distance between the control system and the ejection mechanism.

7.Conclusions

An experimental system for the simulation of internal stores’separation in supersonic wind tunnels is designed,solving problems caused by the influence of the high speed flow and the restricted installation space.With this system,the experiment of internal stores’separation is conducted successfully.The specific innovation is shown as follows:

An internal store release mechanism is proposed,realizing the high speed controllable separation with the model’s velocity repeatability of 0.015 mm/s and angular velocity repeatability of 2.6(°)/s.Besides,a scaled model of the internal store is designed reasonably,and an efficient measurement method based on a binocular vision system for the model’s moment of inertia is proposed to realize a short-time adjustment of its aerodynamic parameters of the models.

A high speed binocular vision measurement system is established to measure the pose parameters of internal store models in wind tunnel measurement tests,and an optimizing method for the coordinates of markers is presented based on a priori model.Its position measurement accuracy reaches 0.102 mm,and the attitude measurement accuracy for pitch and yaw angles is less than 0.106°.

The proposed experimental system can meet the requirement of wind tunnel tests.A stable result is obtained in the simulation experiment for internal store release in the FL-3 supersonic wind tunnel.The proposed system is capable of conducting release tests for different types of models with different release characteristics in both transonic and supersonic environment with reliable measurement results.

Future studies may involve the following researches:(A)to study the effect of high speed air flow on the measurement accuracy of the system;(B)to develop a more effective method for the accuracy verification of the system.

Acknowledgements

The authors thank the anonymous reviewers for their critical and constructive review of the manuscript.This work was supported by the National Natural Science Foundation of China(Nos.51375075 and 51227004),the Scientific Research Fund of Liaoning Provincial Education Department of China(No.L2013035),and the Science Fund for Creative Research Groups of China(No.51321004).

1.Liu W,Shang ZL,Ma X,Zhang Y,Li X,Jia ZY,et al.Position and attitude measuring method of auxiliary tank based on colorcoding method in wind tunnel environment.Acta Aeron Astron Sin2015;36(5):1556–63[Chinese].

2.Wilcox FJ,Jr.Tangential,semi-submerged,and internal store carriage and separation at supersonic speeds.Reston:AIAA;1991.Report No.:AIAA-1991-0198.

3.Keen KS,Morgret CH,Langham Jr TF,Baker WB.Trajectory simulations should match flight tests and other lessons learned in 30 years of store-separation analysis.47th AIAA aerospace sciences meeting including the new horizons forum and aerospace exposition.Reston:AIAA;2009.

4.Cline DM,Rinert WC,Jolly BA.Calculation of generic store from an F-16 aircraft.14th Applied aerodynamics conference;1996.p.631–41.

5.Sickles WL,Denny AG,Nichols RH.Time-accurate CFD predictions for the JDAM separation from an F-18C aircraft.Reston:AIAA;2000.Report No.:AIAA-2000-0796.

6.Jordan JK,Denny AG.Approximation methods for computational trajectory predictions of a store released from a bay.Applied aerodynamics conference;1997.

7.Scott Keen K,Morgret Charles H,Frank Langham T,Baker William B,Jr.Trajectory simulations should match flight tests and other lessons learned in 30 years of store-separation analysis.Reston:AIAA;2009.Report No.:AIAA-2009-0099.

8.Johnson RA,Stanek MJ,Grove JE.Store separation trajectory deviations due to unsteady weapons bay aerodynamics.Reston:AIAA;2008.Report No.:AIAA-2008-0188.

9.Murray NE,Jansen BJ,Gui L.Measurements of store separation dynamics.47th AIAA aerospace sciences meeting including the new horizons forum and aerospace exposition.Reston:AIAA;2009.p.5–8.

10.Liu W,Ma X,Li X,Chen L,Zhang Y,Li XD,et al.Highprecision pose measurement method in wind tunnel based on laseraided vision technology.Chin J Aero2015;28(4):1121–30.

11.Ma X,Liu W,Chen L,Li X,Jia ZY,Shang ZL.Simulative technology for auxiliary fuel tank separation in a wind tunnel.Acta Aeron Astron Sin2016;29(3):608–16[Chinese].

12.Watzlavick RL,Crowder JP,Wright FL.Comparison of model attitude systems:active target photogrammetry,precision accelerometer,and laser interferometer.Reston:AIAA;1996.Report No.:AIAA-1996-2252.

13.Murray EN,Jansen JB,Gui L.Measurements of store separation dynamics.Proceedings of 47th AIAA aerospace sciences meeting includingthenewhorizonsforumandaerospaceexposition.Reston:AIAA;2009.

14.Jones TW,Lunsford CB,Graves SS.Design and development of a real-time model attitude measurement system for hypersonic facilities.43rd AIAA aerospace sciences meeting.Reston:AIAA;2005.

15.Martinez B,Bastide M,Wey P.Free flight measurement technique in shock tunnel.30th AIAA aerodynamic measurement technology and ground testing conference.Reston:AIAA;2014.

16.Liu W,Zhang Y,Ma X,Shang ZL,Jia ZY.Measurement method for moment of inertia based on binocular vision.Chin J Sci Instrum2014;35(9):1972–8.

17.Zhang ZY.A flexible new technique for camera calibration with distortion coefficients considered.IEEE Trans Pattern Anal Mach Intell2000;22(11):1330–4.

18.Jia ZY,Ma X,Liu W,Lu WB,Li X,Chen L,et al.Pose measurement method and experiments for high speed rolling targets in a wind tunnel.J Sensors2014;14(12):23933–53.

19.Liu W,Ma X,Jia Z,Zhang Y,Shang ZL,Li X.Position and attitude measurement of high speed isolates for hypersonic facilities.J Measur2015;62:63–73.

20.Liu W,Liu SJ,Zhang Y,Shang ZL,Ma X.An image acquiring method for position and pose measurement of high speed target in wind tunnel.J Sensors Transducers2013;161(12):635–44.

14 January 2016;revised 22 February 2016;accepted 3 April 2016

Available online 21 December 2016

?2016 Chinese Society of Aeronautics and Astronautics.Production and hosting by Elsevier Ltd.This is anopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

*Corresponding author.

E-mail address:lw2007@dlut.edu.cn(W.Liu).

Peer review under responsibility of Editorial Committee of CJA.