This action might not be possible to undo. Are you sure you want to continue?
MISSILE FIN PLANFORM OPTIMIZATION FOR IMPROVED PERFORMANCE Daniel J. Lesieutre, Marnix F. E. Dillenius, Teresa 0. Lesieutre Nielsen Engineering & Research, Inc. 526 Clyde Avenue Mountain View, California 94043-2212 USA
1. SUMMARY The aim of the research described herein was to develop and verify a fast running optimization-based aerodynamic/structural design tool for missile fin and configuration shape optimization. The developed software was used to design several missile fin planforms which were tested in the wind tunnel. Specifically, this paper addresses fin planform optimization for minimizing fin hinge moments, as well as aeroelastic design (flexible fin structures) for hinge moment control. The method is also capable of shape optimization of fin-body combinations with geometric constraints. The inclusion of aerodynamic performance, geometric constraints, and structural constraints within the optimization software facilitates multidisciplinary analysis and design. The results of design studies and wind tunnel tests are described. 2. LIST OF SYMBOLS AR
3. BACKGROUND This paperdescribes recent research performed by Nielsen Engineering& Research”2’3 aimedat developing practical methods for missile control fin design and for missile configuration shape optimization. Some background information is presentedwhich describesthe importance and difficulties of predicting and designing efficient control fins. This is followed by a description of the technical approach and design code developed. Results from the designcode and wind tunnel tests are presented. Missile control fins have been, and are arguably still, the most efficient meansof controlling a tactical missile and guiding it to a target. They can efficiently generate the required maneuvering force either by a direct action near the center of gravity, as in a mid-wing control missile, or through rotation of the missileto higher a, as in canardor tail control missiles. Affecting all of theseaerodynamically controlled configurations are the sizing and power requirementsof the control surface actuators.Other means of control, such as thrust vector control and control jets are also important to high performance missiles.Thrust vector control can improve both the initial engagementof a threat, including engagementof a rear target, and the end gamemaneuvering(if thrust is still available). Control jets, depending on placement, can be utilized to translate or rotate a missile.Both thrust vectoring and control jets provide fast responseand also provide control at high altitudes where aerodynamiccontrol becomesineffective. Lacau4 details the advantages and disadvantages of different missile control configurations. The primary effects of control fins on missile system design are the available maneuvering force and the time response associated with maneuvering. In terms of subsystemdesign, the control fins determine the actuator sizing. The actuatorsinfluence the missile weight directly through their size and power requirements. Briggs’ describes the performanceparameters which affect control fin actuator design and size. These include frequencyresponsebandwidth, stall torque, rated torque, and fin deflection rate at rated torque. The stall torque is the maximumexpected “worst case” applied torque felt by the actuator and is composed of the sum (multiplied by a factor of safety) of the aerodynamichinge momentand the frictional bearing torque associated with the fin root
fin aspect ratio of two tins joined at root chord, tin normal-force coefficient, normal force/q,SXf
f ii IP MOO %a %in Sref
t ‘CdcR ‘HL ycds
fin normal-force coefficient based on fin area, normal force/q,!& root chord tip chord design objective equality constraint inequality constraint index of performance (cost function) Mach number freestreamdynamic pressure exposed planform area of one fin reference area, body cross-sectional area exposed fin span fin thickness fin axial center of pressure,measured from root chord leading edge, normalized by root chord fin hinge line location aft of fin leading edge fin spanwisecenter of pressure,measuredfrom root chord, normalized by exposedfin span body angle of attack, degrees fin deflection angle, degrees fin polar angle location, 0” = horizontal, 90” = windward meridian, -90” = leeward meridian fin taper ratio, %I%
1998 by Nielsen Engineering
& Research (NEAR). Published with permission of the authors.
Paper presented at the RTO AVT Symposium on “Missile Aerodynamics”, Italy, 11-14 May 1998, and published in RTO MP-5.
held in Sorrento,
Reductions in hinge moments can significantly reduce this mass fraction.0.. Figure 2. Much of the deflected x&n variation is associated with nonlinear effects due to the fin-body gap which are extremely difficult to predict. 4. control forces and hinge moments are obtained from wind tunnel tests” (1988). 10 windward side roll angles. A = 1/2) for M.. Some examples of fins developed under considerable effort by manufacturers to minimize center-of-pressure travel are reproduced from Lacau4 in Figure 1. Internal carriage sets limits on fin span due to stowage requirements. “Theoretical estimate of these moments is not yet possible because the control forces center of pressure cannot be calculated with the needed accuracy. Aerodynamic nonlinearities such as those depicted present a strong challenge to designers of highly maneuverable missiles which operate from subsonic to hypersonic speeds. Therefore. and 9 deflection angles from -40” to +40”.‘2*13 are employed as the aerodynamic prediction modules within the design code. The VTXCHN14 methodology is used to model circular and noncircular body shapes within the SUBDL and .. it is not strictly necessary that the aerodynamic prediction accurately model all the nonlinearities present. Therefore. nonlinearities are not so important. Fin deflection rate capability must permit three axis missile control up to the structural load limit or maximum value of total normal force acting on the missile. and are difficult to predict with computational methods which lack experimental empiricism. Powell’s Conjugate Directions Method. Promising designs were analyzed with CFD for verification prior to wind tunnel testing. However. this variation is even greater since the center of pressure is further forward. This has been accomplished through the choice of the most beneficial location of the hinge line over the expected flight envelope. This is especially true for small values of hinge moment (desired).lq2 Program OPTIvES’** for missiles with arbitrary cross section bodies and up to two fin sections was developed under a U.. Lacau4 mentions. Compared to FIN52. A U.S.‘19p’0 The Nielsen Engineering & Research (NEAR) subsonic and supersonic panel method-based aerodynamic prediction modules.. from 0” to 90”./cR with deflection angle: up to 14% of CR. Hinge-moment coefficients typically increase for lower aspect ratio fins due to larger variations in the axial center-of-pressure travel with both load and Mach number. it must estimate the relative performance of fins adequately. The reduced span results in lower bending moments thus making the frictional bearing torques small compared to the aerodynamic hinge moments. when hinge moments are small. for both zero and nonzero deflections. Results for Triservice FIN42 (AR=l. Historically. Rated torque is the maximum expected applied torque (friction + aerodynamic) over a nominal flight envelope. for Triservice FIN52 (AR = 2.S. Rated torque multiplied by deflection rate determines the power requirements of the actuator. Nielsen6 states that. for undeflected fins in the absence of strong vertical effects. hinge moments have always been considered in missile designs.fs are nonlinear with the flow conditions and deflection angles. This results in fins with reduced aspect ratios. TECHNICAL APPROACH A numerical optimization shell has been coupled with subsonic and supersonic fast running panel method-based missile aerodynamic prediction programs which include nonlinear high angle of attack vertical effects. CL. xcrJcR and yc. 4. Navy SBIR effort investigated the extension to and design of flexible composite tin structures which aeroelastically minimize hinge moments3 A description of the methodology employed follows. It was shown7 that +. Lesieutre and Dillenius7 documented and correlated the axial and spanwise tin center of pressure for fins in the Triservice experimental data base. Data for 6 = 0” are shown as solid circles and correlate fairly well with C.. The primary goal was to design fins with improved performance over that of the initial or baseline fin. SUBDL” and SUPDL. $rr. However. “Zt is often contended that calculations of hinge moments are not reliable because of frequent nonlinear variation of hinge-moment coejicient with control deflection and angle of attack” (1960).1 Summary of Methodologies Employed The optimization algorithm implemented in the OPTMS* design software is a direct search algorithm. Not much has fundamentally changed since 1960 or 1988 in regards to the prediction or estimation of hinge moments. With deflection. The approach described herein1*2’3 to design control fins with improved performance is a practical one which utilizes numerical optimization and nonlinear aerodynamic prediction methods. Nielsen notes that. Current and future air-to-air missiles are being designed for internal carriage. They are highly nonlinear with respect to M.4-2 bending moment. A=%) are shown in Figure 3.. Air Force Small Business Innovative Research (SBIR) contract. the fin-body gap is physically larger for FIN42 than for FIN52 due to different root chord lengths. 4. Actuator mass is determined primarily by the power requirements and can account for 10% of the missile mass.8 C. When lower Mach numbers are considered. There are 990 data points plotted corresponding to 11 angles of attack from 0” to 45”. = 3.. this lower aspect ratio fin shows more variation of xC./cR with C. Figure 2 depicts the experimental xc& versus C. There is considerable variation of xC. and 6./cn and y&s correlate with C. and a structural finite element code.
This formulation is an extension of the Sequential Unconstrained Minimization Technique (SUMT) of Fiacco and McCormick.1 VTXCHN Body Modeling Methodology (2) The aerodynamic analysis of a body by VTXCHN. and inequality constraints. (xl 1 (1) where the indices m. are monotonically decreased during the optimrzation procedure.f. body. and overall configuration. includes objectives and constraints for up to two fin sections. Ipbody. In OPTMIS.2 Optimization Problem Formulation where IPoved.W) =IPoverall k. The following SUMT Index of Performance is employed: rP(r. fin weight. and natural mode frequencies. w) =C. each have the form of Eqn. objectives. Typically. and IP.2 the index of performanceformulation given by Eqn.were chosenas appropriate aerodynamic codes for inclusion in the aerodynamic optimization tool. and 4) if a secondfin set is present. equality constraints. and fin objectives and constraints.n "k w* . 4.l +c. IPi.then the optimization procedurewill not allow a design change in a direction where an inequality constraint is violated.w*) + rp. respectively.l +C. General descriptions of programs SUPDL and SUBDL follow. and constraints with respect to geometric variables. Program OPTMZS2 has two methods for handling the inequality constraintsspecified. stresses. steps 2 and 3 are repeated.4-3 SUPDL modules. "J +IP*ody(x. The completeform of the IP is given by: IP(X. The constraint weights. and . If there are no inequality constraints.. If an initial feasible design is specified. fin displacements are calculated with the flat-fin (rigid) load distribution. (1) and correspond to overall. 2) fin section loads are calculated including the effects of forebody vorticity. km) /w.14 including effects of vortex shedding.elementsof linear and slenderbody theory. and inequality constraints. The secondis as a side constraint..The first is in the manner specifiedin Eqn. and k represent sumson the numberof flow conditions. The VTXCHN code14hasreplacedthe body model within SUBDL and SUPDL and can model circular and noncircular cross section bodies including those with chines.This procedure is depicted below. and the aerodynamic loads are recalculated.the minimization problem being solved is an unconstrained minimization of f(x) when Wj is large. 1 4. subiterations between the aerodynamic and structural analysis module CNEVALFEMODSlW3 are performed to ensure a consistent load distribution and deformed fin shape. This representation of the Index of Performance is very versatile and allows single and multiple point designsto be investigated.comprisesconformal mapping. “[j and wk. i. The structural constraints are included through the CNEVAL-FEMODS’J module which employs automatic gridding and structural finite elements to compute displacements.” The SUMT formulation was enhanced so that multiple objective functions and multiple design point studies could be included. through a penalty within the IP. The fin displacements are used to define a new fin shape for the aerodynamic load calculation. body. Initially. (1) is further divided into three terms governing design objectives and constraints applicable to the fin. as well as nonlinear shock expansionand Newtonian analyses. The inequality constraints gk(x) add a large positive value to the IP if gk(x) approaches zero. As Wj decreases toward zero. (I) /w. The original SUBDL and SUPDL codesmodeledaxisymmetric bodies.[C. [Jr. This is the manner in which all structural constraints computed by the CNEVALFEMODSlT3 module are handled. and this iterative process is continued until the changes in displacements arc less than a user-specified tolerance. the equality constraints become important. The NEAR nonlinear panel method-basedmissile aerodynamic prediction programs SUBDL” and SUPDL12*13 which include models of body and fin shed vorticity at high angles of attack. j.The aerodynamic calculation proceedsstepwiseas follows: 1) VTXCHN computes the forebody loads including vortex shedding and tracking. Fin displacements are determined with the updated loads.3.equality constraints. The OPTMIS2 design software minimizes an Index of Performance (cost function) which includes objectives. 4. 3) vorticity shed from the forebody and the fin set is tracked aft including additional vortices shed from the afterbody. [WJS. For aeroclastic design studies. objectives are formulated with respect to aerodynamic performance variables. respectively. (l).3 Aerodynamic Modeling This section gives a brief summary of the body and fin aerodynamicmodelingmethodologies usedin the OPTMZS code.
5. Another nonlinear effect is related to nonlinear compressibility. The fins are modeled by supersonic panels laid out in the chordal planes of the fins.3. The analysis proceeds from the nose to the base. in excess of approximately 2. additional normal force along the leading and the side edge. The combined vortex gains strength and rises above the fin as shown in the sketch which follows. The pressure distribution calculated at the second cross section in the physical plane includes nonlinear effects of the vortices shed from the first cross section. contributions from free stream due to angle of attack. The panel method is based on the Woodward constant pressure panel solution16 for modeling lift. this step is omitted. The constant u-velocity panels on the interference shell only experience the mutual interaction with the constant u-velocity panels on the fins and fin thickness effects.and side-edge separation vorticity as the angle of attack is increased. The locations of the shed vortices are transformed to the mapped plane. On the fin. The resulting pressure distribution is integrated to obtain the aerodynamic forces and moments. Fins may have arbitrary planform. If this is the case. For subsonic flow threedimensional point sources/sinks are used. body-induced effects (upwash). Along the leading edge. The strengths of the shed vortices are related to the imposition of a stagnation condition at the contour comer or chine points in the mapped plane. The result is a distribution of nonlinear. Effects of tin thickness can be included by thickness panels in the chordal plane of the fin. The vortices are then tracked aft to the next cross section in the mapped plane. vortices are positioned slightly off the body close to the comer or chine points in the crossflow plane. and for supersonic flow three-dimensional line sources/sinks are used.” the suction is converted to normal force in proportion to vortex lift factors.and side-edges are accompanied by an augmentation to normal force which is nonlinear with angle of attack seen by the tin. The strengths of all of the constant u-velocity panels in a tin section are obtained from a solution of a set of simultaneous equations. This sketch shows how SUPDL models the path of the combined leadingand side-edge vortex by locating it above the fin plane at an angle equal to one-half of the local angle of attack (as seen by the fin). the vertical wake is represented by a cloud of point vortices with known strengths and positions. If the actual body is axisymmetric. vorticity can be generated at angles of attack as low as 5”. Fins can develop nonlinear leading. be located off the major planes. This nonlinearity is modeled by calculating the suction distribution along the leading and side edges. If the cross section has sharp comers or chine edges. In accordance with an extension17 of the Polhamus suction analogy. the leading-edge vortex joins the side-edge vortex. SUPDL12713 is a panel method-based program which together with the VTXCHN’4 body module can analyze an arbitrary cross section body with a maximum of two fin sections in supersonic flow. For smooth cross sectional contours.2 Supersonic Aerodynamic Prediction Method in the fin section. the flow tangency boundary condition includes mutual interaction with all other constant u-velocity panels L Tmilingulgc vata The vertical phenomena along the leading. velocity components are computed at points on the transformed body and transformed back to the physical plane. The axisymmetric body is modeled by three-dimensional sources/sinks for linear volume effects and by two-dimensional doublets for linear upwashlsidewash effects. the code makes use of the Stratford separation criterion applied to the pressure distribution to determine the separation points. vorticity can be generated at supersonic speeds provided the leading edge lies aft of the Mach cone emanating from the root leading edge (a subsonic leading edge). In addition. for example). As a result. The procedure for the first cross section is repeated. This situation can . If the side edge is long (similar in length to the root chord. a set of panels is laid out in a shell around the body over the length of the tin root chord to account for lift carry-over. 4. the fin leading edge shock may lie close to the surfaces (usually the lower surface) of the fin. Noncircular cross sections are transformed to corresponding circles in the mapped plane. Along the body. At a cross section near the nose. The circumferential pressure distribution is determined in the physical plane using the compressible Bernoulli equation. and vertical wakes from upstream fins and body flow separation. Each panel has a control point at which the flow tangency condition is applied.4-4 nonlinear vertical modeling. The strengths of the thickness panels are directly related to the local thickness slopes. For M. In SUPDL this panel is designated the constant u-velocity panel because the pressure on the panel is computed using the compressible Bernoulli velocity/pressure relationship. and be attached at arbitrary angles to the body surface. an axisymmetric body is created in the mapped space.
The weight of each subsequent design is ratioed to this initial weight.4-5 also occur at low supersonic Mach numbers if the angle of attack is high. up to 10 upper bounds and their associated node numbers can be specified. 4. The strengths of the lifting surface singularities are obtained from a set of linear simultaneous equations based on satisfying the flow tangency condition at a set of discrete aerodynamic control points. with exposed span of 0. In the simplest model. For all-movable fins. There are three degrees of freedom per structural node: two rotations in the plane of the fin. In either case. In the second option. the maximum value of the von Mises bending stress is found. and aeroelastic fin design. The number of violated constraints and the corresponding mode numbers and frequency ratios are recorded. with exposed . However. A unique feature is the option to include strip-on-strip interference based on the linear constant u-velocity panel solution to correct the flow angle used in either the shock expansion or Newtonian pressure calculation methods.72 diameters. Structural Constraint Evaluation. For the weight constraint. The parameters for any intermediate section are defined by linear interpolation. and the node number and displacement ratio are recorded. or supported on a shaft to represent an all-movable control surface. Additional design studies are described in References 1. The horseshoe vortices on the interference shell around the body are used only to model the carryover forces between the body and fins (the body volume and angle-of-attack effects are obtained from the three-dimensional sources and doublets and conformal mapping procedure in the VTXCHN module). 4. The nonlinear vorticity effects associated with fin edges described above for SUPDL are also modeled in SUBDL. FIN1 . 2.3 Subsonic Aerodynamic Prediction Method quadrilateral patches. The fin can be cantilevered at the root. five parameters for the root and five parameters for the tip define the thickness distributions. For the fin designs tested in the wind tunnel. plus a transverse displacement. and both elements are uniform. TX. If this value exceeds the allowable. consistent inertia elements from Reference 22 are used. are given below. For the stress constraint. There are two options for displacement constraints.3. 5. since nonconforming elements do not reproduce the proper symmetry properties for a rectangular or a square planform. only a single upper bound for the maximum absolute value of any displacement is specified. the fin loading prediction based on the constant u-velocity panel method and the Bernoulli velocity/pressure relationship is no longer adequate. then the number of violated displacement constraints is set to unity. each patch is represented by two bending elements. the number of violated displacement constraints is incremented. and the node number and displacement ratio are recorded. A frequency constraint is considered violated when the frequency for any specified mode becomes less than its bound. Displacement ratios (actual/allowable) are calculated at the specified nodes. 25 The meshed fin is divided into Descriptions of two tin planform optimization designs which were tested in the Lockheed-Martin High Speed Wind Tunnel in Dallas. and two (2) large span fins. Up to five lower-bound frequency constraints can be imposed by specifying the lower bounds and their mode numbers. The lifting surfaces and the portions of the body spanned by the lifting surfaces are modeled with planar horseshoe vortex panels. These elements are also described in detail in Reference 22. The fin is modeled with constant-thickness. if any ratio is greater than unity. wind tunnel tests. FIN5 and FIN6. verification of aerodynamic performance prediction. As an option. For dynamic problems.4 Fin Structural Modeling For fin structural modeling. the pressures acting along chordwise strips can be calculated with nonlinear shock expansion or Newtonian theories. the control shaft is modeled with a beam in bending and a rod in torsion.FIN4.1 Fin Phmform Optimization Design Studies Program SUBDL” is a panel method-based program which together with the VTXCHN14 body module can analyze an arbitrary cross section body with a maximum of two tin sections in subsonic flow. the constraint-violation flag is set to unity and the associated node number and stress ratio are recorded. 5. In the first option. and 3. The addressable geometries are the same as those described for SUPDL previously. the weight of the initial design is saved. there is an option to model each patch with two pairs of elements which eliminates any asymmetries. RESULTS This section describes results including fin planform design studies. four (4) small span tins. Details can be found in References 17 and 19. If this bound is exceeded. The generic section is a symmetric truncated double wedge with finite thicknesses at the leading and trailing edges and is illustrated in the sketch below.20 with modifications to allow for anisotropy. No transverse shear effects are included. triangular nonconforming bendin elements.
FIN1 was the small span trapezoidal reference fin used to start the design optimization for FIN2.. The model consisted of a twocaliber tangent ogive nose and a cylindrical body 5. FIN6 was the large span trapezoidal reference fin used to start the design optimization for FINS. The deflected results for the supersonic Mach number. and C.0 and 6 = 0 and 20” as a function of a. 1. corrections based on CFD calculations could be included. Further details can be found in Reference 1. . 5. Existing test hardware consisting of a body with tin strain-gage balances was utilized. However.0. number. = 0. high angle-of-attack condition gave an aft center-ofpressure location. for 6 = 0” are in good agreement for both Mach numbers. and 10 depict the fins described herein. the low M.C.I was minimized. 5. The subscript “2” refers to the supersonic design flow condition./cR with a are shown for M m = 0. The normal force is underpredicted in this case. the OPTMIS results for C. The tests included Mach numbers of 0. The predicted axial center of pressure is forward of the experimental result for angles of attack above 10”. and FIN4. (1) normal force. The comparisons of the measured and predicted C. The axial’ center-of-pressure location is also predicted well for the b = 0” conditions. whereas the supersonic Mach number. The subsonic prediction module.. C.. 5. and moment data. The results for FIN3 are similar to FINl. This is most likely due to inadequate modeling in OPTMIS of the gap between the deflected fin and the body which changes the fin loads near the root chord leading or trailing edge. FIN3 and FIN5 were designed using OPTME* to minimize the fin axial center-of-pressure travel from subsonic to supersonic flow. OVERFLOW slightly underpredicts C. the ratio 1~~ . To achieve this objective.4 diameters aft of the nose tip. C. respectively.s at M. The predicted aerodynamic results for 8 = 20” are not in as good an agreement with the experiment. 2. For large deflections this places the leading and trailing edges in different local flow fields. This accounts for both the overprediction of normal force and the forward location of the center of pressure.2 Wind Tunnel Test Description The fin planforms described above were tested in the Lockheed-Martin High Speed Wind Tunnel in Dallas. and 3. The comparison of the measured and predicted Cms for 6 = 0” are in good agreement for both Mach numbers.5. does model deflection effects through geometric deflection of the fin. = 2. = 2.8. TX. and (3) hinge moment. 15’).0 and for 6 = 0” and 20”. (2. A pair of fin balances were positioned 3. and the subscript “1” refers to the subsonic design flow condition. CL) = (0.5. For M. were reduced to provide fin axial and spanwise center-ofpressure locations.4 diameters were tested. Predicted results from OPTMIS* are shown as solid symbols with solid lines. Experimental data are shown as open symbols. The design objective was to minimize this center-of-pressure travel. The fin normal force based on fin area was to be maintained. and fin deflection angles of 0” and 20” were tested. SUBDL.. low a design condition gave a center of pressure forward on the tin. The variation of fin normal force C. FIN3. and results from the NASA OVERFLOW Navier-Stokes solve?* (zero deflection only) are shown as solid symbols with dashed lines. = 0. The design studies for FIN3 and FIN5 are described in this paper.0. were the only model data collected. agree fairly well at low angles of attack but do not have the correct stall behavior as angle of attack increases. 6. All tests were conducted with identical fins on the left and right balances to insure symmetry. This design objective also tends to give a flat xcP response with increasing fin normal force..4-6 span of 1. the flow field can vary significantly circumferentially around the body.5 and 2.4 Prediction Verification for Optimized FIN3 The predicted and measured aerodynamic performance of the small span reference fin FIN1 is shown in Figures 4 and 5. The local flow fields behind the bow shocks close to the body surface can only be predicted well by Euler or Navier-Stokes flow solvers. The supersonic prediction module.0 and a = 20”. In spite of the above. OPTMIS slightly overpredicts C.xcPll/lC~ . respectively. show the opposite trend. at M. The three-component outputs for the fins.3 Prediction Verification for Reference FIN1 axial center of pressure xC. The design flow conditions were: (M. during the period March 3 . SUPDL.0. currently models the effects of deflection through the boundary conditions and not through geometric deflection of the tin. The panel method-based programs are not capable of predicting these local flow conditions. 2”). M..5. and fin The predicted and measured performance of FlN3 is shown in Figures 6 and 7. (2) root-bending moment. For the reference fins. the axial center of pressure is predicted well by OPTMZS.5 and 2. 1997. The angle of attack range was -12” to 22”. the nonlinear flow field (local Mach number and local dynamic pressure variations) present behind the nose bow shock can be important when the fin is close to the nose. and xcp/cR are shown for M. All design studies have been performed at 6 = 0”. The fin force.. x&n and y&s. The internal structure of the body permitted mounting the fins at deflection angles from -20” to +20” at 5” intervals.0.. C. For this forward fin position.5. The design variables were third-order Chebyshev polynomials describing the leadingand trailing-edge shapes. = 0. However. The resulting geometries of FIN3 and FIN5 are shown in Figures 4 and 10..5 and a = 20”. Figures 4.2 calibers long.
good agreement between the predictions and the experiment. due to the smaller fin area. The experimental data shown in Figure 11 confirm this result. The predicted and measured performance of the large span fins FIN5 and FIN6 is shown in Figure 10.5 and 2. C. A description of the design and testing of an aeroelastic fin structure13 used to demonstrate the potential of chordwise flexibility to control center-of-pressure location is described. The design calculations indicated that x&cR could be shifted forward.0 for 6 = 0” as a function of angle of attack. Predicted results are shown from the OPTME code and the OVERFLOW22 code. Details of the structural modeling of the composite layup and structural properties can be found in Reference 3. and the trailing edge region. For supersonic Mach numbers (1. FIN3 produces less normal force than FIN1 for the same angle of attack. with grooves in a near spanwise direction. 5. A single orientation can be chosen. along with predicted results.7 Aeroelastic Fin Design A detailed comparison of experimental x&CR data for reference FIN1 and optimized FIN3. The comparisons of the measured and predicted CNF. 15’). The predictions for axial center of pressure do not agree with experiment for 6 = 20”.. of the composite fin lay-ups.5 for a = 5”.0. Predictions are designated TAILOR in Figure 11.0.. There is.5.4-7 However.. without appreciable change in C. for the flexible and rigid fins as a function of a for M. Figure 9 compares the FINI and FIN3 axial center-ofpressure location for all four test Mach numbers and for 6 = 0” and 20”. The axial center-of-pressure location is predicted slightly aft of the experimental value for moderate angles of attack (unstalled). FIN3 shows only slight variations of xCr..5 and 2. and the principal stiffness axis orientation. Measured and predicted results for FIN3 (optimized) and FIN1 (reference) are shown for 6 = 0”.5.3 FIN3 has 50% less center-of-pressure travel ‘NF than FINI. The experimental data. 5. the normal force can be increased by a higher angle of attack or fin deflection without adversely affecting center-of-pressure travel. The configuration modeled and the design variables governing the aeroelastic design are shown in Figure 12.0 and match the normal force well..5. the middle portion of the fin. The axial center of pressure is plotted as a function of C.13 an aeroelastic tailoring procedure was developed based on the SUPDL12*‘3 code and a structural finite element code FEMOD. OPTMIS slightly overpredicts the normal force at a = 20”. or the fin can be modeled as composed of up to three different layup orientation regions: the leading edge area of the fin. OVERFLOW results are shown for M. and the CFD results indicate that the optimized FIN3 has less center-of-pressure travel from subsonic to supersonic speeds and that the optimized fin has a flatter axial center-of-pressure variation with increasing C.32. a) = (0. 5.5 and 2.5 flow condition./cR with either a or 6 compared to the reference FINI. within 5% of cR. ( based on base diameter)..0 indicate that FIN5 has reduced center-of-pressure travel from subsonic to supersonic speed up to the onset of stall of the reference fin FIN6. In the earlier study.6 Results for Optimized FIN5 and Reference FIN6 Aeroelastic design studies have been performed to improve missile fin performance through beneficial passive deformations of the fin structure under aerodynamic load. The objective of the recent study3 was to minimize the tin axial center-of-pressure travel over a Mach number range of 1.. the results from the OPTMZS code. OPTMIS does not predict the stall characteristics for the M. respectively. The grooved aluminum trapezoidal fin is shown in Figure 11(a). The optimized tin FIN5 delays stall and reaches a higher peak normal force than the reference fin at subsonic speeds. C.13 The design procedure was successfully applied to a grooved aluminum lifting surface resulting in grooves in essentially the spanwise direction. The predicted results for 6 = 20” are similar to those of FIN1 in terms of C. This is followed by a recent study3 aimed at using aeroelastically tailored composite fins. The planform shape was fixed and the fin was undeflected. = 1.. However. The axial center-of-pressure results for M. p.5 Comparison of FIN1 and FIN3 FIN5 was designed to have a reduced center-of-pressure travel from subsonic to supersonic speeds.5.2 to 2. Structural displacement and stress constraints ensure that realistic fin structures are considered during the optimization process.0. = 0. The reasons for the lack of agreement given above for FIN1 apply here also. Both fins have similar normal force characteristics. The axial center-ofpressure location is also predicted well for the 6 = 0” conditions. For = 0.. the design objective for FIN3 was to minimize axial center-of-pressure travel from subsonic to supersonic speeds. Again. and x&CR are shown in Figure 11(b) and 11(c). and 3. = 0.. for 6 = 0” are in good agreement for both Mach numbers. = 2.0). in general. 2”) and (2. The design flow conditions were: (M.5. The vertical aXiS (x&CR) for both graphs in Figure 6 spans 0. 2. The design variables governing the tin structure are the fin thickness parameters at the fin root and the fin tip. and 3. 2.. and x&R are shown for M.. The design objective was to shift x. = 0. . as compared to the reference fin. are shown in Figure 8 for the design Mach numbers 0.. within 2% of CR./cR forward to the maximum possible extent by varying the direction of the grooves.
McIntosh Structural Dynamics. indicated in Figure 15 for the aeroelastic fin. Lesieutre.. “Minimization without Constraints”.O. = 1. Flight Vehicle Branch WLMNAV.O. Fred Davis of the Air Force Research Labs. Dillenius.. and hence hinge moments. Lesieutre. Powell.S. NEAR TR 530. M. in “Optimization and Design”. 5. D. REFERENCES Lesieutre. Prentice-Hall. “Remote Control Missile Model Test”. Nielsen. The space marching NEARZEUS results shown in Figure 15 extends the normal force prediction to high Mach The reduced center-of-pressure travel is numbers. 141. in Tactical Missile Aerodynamics: General Topics. M.5 are shown in Figure 14. New Jersey. Dr. Significant improvements to center-of-pressure travel. 8.J. 1988. Mountain View. 1.. in Proceedings. M. T. September 1997. CA..W. 1964. R. M. 4. 6.J.4-8 To start the optimization. “Missile Aerodynamics”. D. R. also. NEARZEUS predicts a similar forward shift of the center of pressure for the flexible fin.J. “PlanformKonfiguration Optimization Program OPTMIS For Arbitrary Cross Section Configurations With Up To Two Fin Sets . Craig Porter from NAWCWPNS. 10. Nielsen Engineering & Research...C. Progress in Astronautics and Aeronautics.. Jr. T. Dillenius. M. and Wilde.J. 7.- Rigid and Optimized Performance Flexible The optimized fin has nearly the same normal force characteristics of the rigid fin but the center-of-pressure travel over the Mach number range is reduced 56%. Both conventional circular body and unconventional noncircular body configurations can be designed and analyzed.. 1973. Computation Journal. “Optimal Aerodynamic Design of Advanced Missile Configurations With Geometric and Structural Constraints”. D.F. The deformation of the fin midplanes at M. 8. The thickness distribution of the optimized fin is depicted in Figure 13. and Sawyer. “A Survey of Missile Aerodynamics”. a constant thickness fin was specified. AGARD CP 493. April 1990. and Lesieutre. Samuel McIntosh.E. The speed and multidisciplinary capabilities of the method make it an excellent tool for preliminary design..J. “An Efficient Method for Finding the Minimum of a Function of Several Variables Without Calculating Derivatives”.E.D... Lesieutre. Rijckaert. Love. Initial studies of aeroelastic fin structures indicate that significant improvements to fin performance can be obtained through the use of flexible structures. S.M. Andy Sullivan and Mr. October 1997. and Lesieutre.H. Englewood Cliffs. J. pp 155-162. AIAA.. M. and Perkins. (Eds. December 1997. 3. W. J. “Systematic Tactical Missile Design”. Figure 15 indicates that the optimized flexible fin maintains the normal force of the rigid fin.C. The principal structural axes for thisfinare pLE= 2. NEAR TR 520. ACKNOWLEDGEMENTS The authors would like to thank Dr..7” for x/cR I 3/s and pTE = -48. NEAR Conference on Missile Aerodynamics. “Control of Hinge Moment by Tailoring Fin Structure And Planform”. NEAR TR 519. 1991. AGARD CP 451..Software Programmer’s Manual”. Palo Alto. at Eglin AFB for their support of this work under Air Force Contract FO8630-94C-0054.. Paper 30.J.. Shaw.. 3.2 and 2.J. Briggs.. 9.N. October 1988.F. J. and Dillenius. The normal force and axial center-of-pressure performance of the fin are shown in Table 1 and Figure 15.Software User’s Manual . Table l. M..M. Reprint. CONCLUSIONS An optimization-based design tool for missile tin and configurations design and analysis has been developed. can be obtained through planform optimization.F.3” for x/cR 2 a/8. 2. Dr. was responsible for the structural modeling described. China Lake for sponsoring the aeroelastic Iin design effort under Navy Contract N68936-97-C-0152. D. McGraw-Hill. M..E. Lacau..). Dillenius. . A large deformation of the fin at the root chord leading edge is indicated.G. May 1988. 6.E. CA. D. Sargent.F. Avriel. Paper 17. 7.F. New York. 1960. “Chordwise and Spanwise Centers of Pressure of Missile Control Fins”... The design capabilities of the method for fin planform optimization have been verified with CFD calculations and with a wind tunnel test. D. Allen. 7.
AIAA Paper 92-0080... M. 21... Jr.M. “Computer Prediction Program AMICDM. Dillenius. JAircraft..5 0”1$. D. New York.J. S. pp 193-199. Hegedus. NWC TP 6648... 8.. October 1966.. TERRIER.6be” unpublished NASA document. “Program SUBSAL and Modified Subsonic Store Separation Program for Calculating NASTRAN Forces Acting on Missiles Attached to Subsonic Aircraft”.4 n CROTALE TARTAR. F. T. Vol. NASA CR 3883. April 1985. FIN42 AR=1 k.. Polhamus. Jr. Chan.F.P. Vol..4-9 11. Lesieutre.. Army Missile Command Technical Report RD-CR-84-15. 22. M.. April 1987. 20. NASA TN D-3685. “Prediction of Vortex-Lift Characteristics Based on a Leading-Edge Suction Analogy”. NEAR TR 459.C.= 3.~90° 0”Sa. 0% @r5 90” and-40% 6 1+40° (solidsymbols are d = 0’). “Aeroelastic Tailoring Procedure to Optimize Missile Fin Center of Pressure Location”. and Baltakis. U.H.. W. “An Integrated Approach to the Analysis and Design of Wings and Wing-Body Combinations in Supersonic Flow”.F. S. and Flutter Performance Requirements”.30 I 4 ’ 1 ’ ’ a b ’ ’ lCNFS12 ’ ’ ’ ’ 3 ’ ’ Figure3. May 1992.N. McGraw-Hill. Przemieniecki.5 -40°. OY @r5 90” and-40% 8 %40” (solidsymbols are 5 = O”). September 1987.s65+40" -AA 0. at M. John Wiley 8z Sons.C.SaS45" . G. 0. I: Operational Instructions.5 Figure3. Perkins.. Improved Aerodynamic Prediction Program for Supersonic Canard-Tail Missiles With Axisymmetric Bodies”. Perkins. E.F. STANDARD Figure l.V.“. II: Sample Cases...7 -M * mm_ = 3. 1968.7 o A” a 0.. Jr. and Woodward.. A.C.TriserviceFlN42. “Nonlinear Programming”.F. 14.C. = 3. AIAA Paper 97-0041. NAWCWPNS TM 7319. Carmichael.Version 1. and Whittaker. Dillenius.F. Vol.6 30.G. McIntosh.. S._ * -A A% a o. Wardlaw. 19. Aerodynamic Program for Supersonic Army Type Missile Configurations with Axisymmetric Bodies”. S..E..E.. “Theory of Matrix Structural Analysis”. “Optimization and Tailoring of Lifting Surfaces with Displacement. 13. April 1971. May 1994.J. M. February 1996. x&n asfunction of C.0 0°S$. and McCormick.A.. M.0 for 0”s a < 45”.TriserviceFlN42. P. Fiacco.Controlsurfaces with limitedcenter-of-pressure shifts. S. A. Frequency.0 for 0% a 5 45”. Fam[ly (canard) MAGIC dl (canard) SUPER 530 (tail) V. T. x&n asfunctionof C. Canning. “OVERFLOW User’s Manual .W.O. Dillenius.E. III: Boundary Layer Code ZEUSBL”. 17. January 1997. M. M.C... Buning. and Lesieutre.. Jr. D.s at M.. C.” E FIN52 M_ = 3.~900 0°.L. Lesieutre. and McIntosh. Naval Air Warfare Center Report NWC TP6834. J.F.S. 12.=O. June 1984. and Perkins.E. et al..E.. Dillenius.8 . F. 18.C. 15.S. Dillenius. “VTXCHN: Prediction Method For Subsonic Aerodynamics and Vortex Formation on Smooth and Chined Forebodies at High Alpha”.0 A AR=2 h=0. January 1992. “NEARZEUS User’s Manual.. and Dillenius.. F. Inc.E.. .S4S0 -40”565+40” 0. “Program LRCDM2. 1968. R. “Modified NWCDM--NSTRN and Supersonic Store Separation Programs for Calculating NASTR4N Forces Acting on Missiles Attached to Supersonic Aircraft”.C. M. 16. New York. 23. Priolo.
o .Fin &0 00 00 t / Cl 0 ---m--- + --m-- CXP. OPTWIS OPTWIS 0vnrLw rl r1 r1 r1 I1 Me . = 0. and .5 . 4 --m-.Comparison of measured andpredicted C.I .0 ~PTWIS rl 2.o . 0pm1s %‘...0 .. CXP. = 0...5 I) 6 Fin r3 r3 r3 OPTYIS r3 LW r3 cx~.o 20. FIN3 1 1 1 L 0=0.o 10..s .3 r. Figure6. FIN3 at M.2 lo a 20 0 "a 20 Figure5.. xC..s $ .o .I.o ..kR for FIN1 at M. cxtp.o .0 .5 . .0 0..0 z 0=0. .o ..Comparison of measured andpredicted CNFS and xC.o .a .5 a 0 0. .o ..5 $0.o .s .kR for FIN3 at M.o .0 -0 .2W 0 10 D-5$0.5 rl a...o .0 .0 . "a 20 FIN3 MB = 2.o .o 0vcrLw ri a.. Figure7. = 2..yigy a 20 0 .o 20.o .o 10 20 FIN1 I 0..5.0 OPTYIS ri a.5. and xCplcR for FIN1 at M. and xC.0 0 --W-0vnrLw r3 a.0 .o 30.0.s .Comparison of measured andpredicted C..s 0 0.o 6 20.o . = 2.0 Figure4.6 I.o .kR for.0.I .7 q 10 20 FIN3 Me= 2.o 0.s .4 0..Comparison of measured andpredicted C..
.8 Rigid Exp.5 and 2. Figure 1l. 6 = 0” and20”.0 FIN1 Y-I 0.Oo”-M_=3. I. 0. .0.Continued.5. M..Comparison of measured andpredicted C. with angleof attack.0 FIN3 OPNIS ____________ O”. FIN1 20 a Figurelo. FIN1 0 II a 0.5 . 1.5 and2.2 : 0 I- : l q 0.5 FIN3 @ Y-I 2.. Figure9. xc& for FIN5 andFIN6 at M. .-A+ for reference FIN1 andoptimized FIN3. Rigid TAILOR Flexible ..C-J ou” 00 0 b n 0.. = 0. .. 1..0. . = 0.Comparison of measured andpredicted xC. . 2. Flexible Exp. .44 0 10 and (a) Groovedflexible fin. . Figure11.2 q N qOO." Figure 8. . = 0.Measured x..RAW.JcRfor reference tin FIN1 andoptimizedtin FIN3. 0. M. .Performance of rigid andgroovedflexible tins.5 (rEJ0 *@o”” . .0.TAILOR nL=1.! Ax&.5 FIN1 111s I 2.5.. 10 a 15 20 ( b ) Variationof C.0 and3.ZRFLOW 0 0 0 a $I.0. .6 t 0 @ CNF I..
Configuration and design variables used for the aeroelastic fin design study.5 Aeroelasticaliy Optimbed Fin Figure 13. t. Fbxibb END. 0: 0 Rigid Exp.Predicted aeroelastic deformations.5 8 n Figure 14..4-12 0. Figure 11.Optimized aeroelastic fin thickness. Body r with Control Fin I 0.: 0.0 n o ’ n 0 1% 0 Cl 1’5 Cl no io O. .. Rigid TAILOR Fbxbb TAILOR M_= 1. Figure 12.3 ppoa’ve v Ply Orientation p-00 f3 egativs 1 2 M"_ 4 5 Figure 15 Comparison of rigid and flexible fin performance.Concluded.6 t a-5" 0.! ( c ) Variation of xc& with angle of attack.. CONTROL FIN ON BODY %O 2 00 0 0e0 m 0 5 0 0 0 o 0.5 OPTMIS Predicted Deformation 0 8 0 : $0 >y n qo 0 0 O 0 0 0 O 0 0 0. Fbxibb Exp.5 Rigid Exp.Z15r 0 0 0 n Rigid Exp.6 0 : m no.. FbxBb Exp. Rigid TAILOR Fbxibb TAILOR Mm = 2. Rigid TAILOR Fbxibb TAILOR M_ = 3.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.