• No results found

Aerodynamic design optimization of helicopter rotor blades in hover performance using advanced configuration generation method

N/A
N/A
Protected

Academic year: 2021

Share "Aerodynamic design optimization of helicopter rotor blades in hover performance using advanced configuration generation method"

Copied!
10
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

AERODYNAMIC DESIGN OPTIMIZATION OF HELICOPTER

ROTOR BLADES IN HOVER PERFORMANCE USING

ADVANCED CONFIGURATION GENERATION METHOD

Ngoc Anh Vu, Ho-Jung Kang, Abdulaziz Irgashevich Azamatov, Jae-Woo Lee, Yung-Hwan Byun Dept. of Aerospace Information Eng., Konkuk Univ., Seoul, Korea

Abstract

This study proposed a process to get optimal helicopter rotor blade shape regarding aerodynamic performance discipline by using a new geometry representation algorithm CST which uses both the Class Function/Shape Function Transformation. By this approach, airfoil shape was considered as design variables. This optimization process was constructed by integrating several programs which were developed by Department of Aerospace Information Engineering of Konkuk University. The design variables include twist, taper ratio, point of taper initiation, blade root chord, coefficients of airfoil distribution function. Aerodynamic constraints consist of limits on power available in hover and forward flight. While, the trim condition must be attainable.1 This paper considers rotor blade configuration for hover flight condition only, so that power required in hover was chosen as objective function of optimization problem. Sensitive analysis of each design variable showed that airfoil shape has an important role in performance of rotor. The optimum rotor blade reduced hover power required 7.4% and increased figure of merit 6.5% are a good improvement for rotor blade design.

1. INTRODUCTION

Regarding to blade aerodynamic performance design, there are two common approaches. First, most of researchers now focus on blade shape design to optimize the aerodynamic performance of rotor blades by selecting the point of taper initiation, root chord, taper ratio, and maximum twist which minimize hover power while not degrading forward flight performance.1 This approach usually deals with integration of several programs to build an optimization process. Second, some works tried to solve this problem by CFD methods. These CFD methods are reasonable for hover case but long time consuming. Moreover, in forward flight, the flow filed passes the blade is very complex to apply CFD method. Therefore, the CFD method is not suitable for preliminary design phase because of quick estimation requirement. With the target of quick estimation for preliminary design phase, this study follows the first approach with advanced improvements. In this study, a new geometry representation algorithm which uses both the Class Function/Shape Function Transformation (CST) method was applied to take a consideration of airfoil shape. The advanced points of this CST method are high accuracy and few variables of geometry representation.2 Therefore, this work dealed with the same problem of blade aerodynamic performance design was mentioned above and some additional design variables came from airfoil shape consideration.

The satisfactory of aerodynamic performance design

was defined by the following requirements which must be right for any flight conditions: the required power must be less than the power available, and the rotor blade must be trimmed.1

The process of the design was represented in figure 7. This process includes sizing module also. After getting a size of helicopter, helicopter rotor blade shape optimization process would be performed as next step of design process. Following this process, a set of initial values for design variables are chosen from sizing module. The airfoil base line which is airfoil NACA0012 was chosen for the first step of design process. Then, blade shape such as chord distribution, twist distribution, and airfoil points coordinate are generated. The power required for hover and forward flight were computed by KHDP program, and trim condition is checked. Airfoil analysis is performed by 2KFoil program to generate airfoil aerodynamic characteristics in C81 format. Some others additional codes to generate airfoil coordinates, chord distribution and twist distribution were implemented in order to build a full framework for optimization process in Model Center software. Model Center is a powerful tool for automating and integrating design codes. Once a model is constructed, trade studies such as parametric studies, optimization studies, and DOE (Design Of Experiment) studies may be performed.3

The required power in hover analysis is performed by using blade element method which considers the airfoil characteristics.

(2)

2. ANALYSIS PROGRAM

2.1. KHDP (Konkuk Helicopter Design Program)

KHDP is a helicopter sizing, performance analyis and trim analysis program was developed in Konkuk University.

Fig.1. KHDP Program process. 

These codes were developed to be used in the conceptual design phase and hence used empirical formulas to reduce computing times.4

2.1.1. Sizing module

The sizing process based on graphical design techniques developed during the 1950’s and 1960’s and were initially utilized with nomographs.

A graphical design method, called the fuel ratio or RF method is typical of the developed techniques.5, 6. This method was used to construct the whole process of sizing as shown in figure 1.

This process starts from mission analysis and performance requirements analysis. The performance requirements are specified in terms of hover capability and cruise speed requirement at a specified altitude and temperature, which reflect the environment where the vehicle is expected to operate. The mission requirements address the critical relationship between payload, range and hover time, which will determine the type of rotary wing aircraft.

Based on empirical formulas, this module is able to generate following sizing results.

+ Geometry data: main rotor, tail rotor, fuselage, tail fin.

+ Weight data: gross weight, empty weight, fuel weight.

+ Mission performance data: Each mission segment fuel requirement.

The sizing module is validated by comparing with the data of existing helicopter UH-60A as shown in table 1 below.

Table 1. Sizing module validation

Parameter UH-60A Data

from Sizing Module Gross Weight lb 22000 21409 Empty Weight lb 10901 11241 Disk loading lb/ft2 9.7 10.0 Main rotor diameter ft 53.66 54.2

Solidity 0.084 0.089

Main rotor tip speed ft/sec

725 722.3 Number of main rotor 4 4 Tail rotor diameter ft 11 11.7 Tail rotor tip speed

ft/sec

685 703.6 Fuselage length ft 50.6 48

(3)

2.1.2. Performance analysis module

To quickly understand and image the helicopter behavior, the performance analysis module was developed. An analytical method was used to provide the designer with a reliable tool of sufficient fidelity to assist in the design process. The module based on an energy approach, and it has been written to yield results quickly and inexpensively.7, 8 This module is able to predict following performance: + Vertical Climb (Maximum)

+ Rate of Climb (Maximum Rate of climb) + Ceiling (Hover Ceiling, service Ceiling) + Dash speed, Maximum cruise speed

+ Cruise Flight (Maximum range, cruise speed for maximum range, Endurance, cruise speed for maximum endurance…)

+ Descent + Autorotation

+ Height-Velocity diagram + Acceleration and deceleration + Turn maneuver

Fig. 2. An algorithm for performance analysis in cruising flight.

The module is able to yield not only a specific performance but also series of helicopter performance in different operating environment such as weight of helicopter, altitude, temperature, and forward velocity.

Momentum and blade element theory were applied to calculate the power required in different operations of helicopter which are hover, climb, cruise, descent, autorotation. 8, 9

Figure 2 showed an example of an algorithm was

developed to predict performance behavior of helicopter in cruising flight by momentum theory and blade element theory (BET). BET needs to call trim module analysis to get power required.

A Validated results of this module were shown in table 2. The validated results were got from AS332 L1 helicopter data .

Table 2. Performance analysis module validation Performance AS332 L1 Performance

Module Max Speed kts 150 129 Max Rate of Climb

ISA, SL, ft/m

1618 at 70kts

1682 at 80kts Hover ceiling ISA,

SL, IGE ft

10663 11000

Hover ceiling ISA, SL, OGE ft 7546 7000 Service Ceiling ft > 9.500 16800 Service Ceiling, OEI ft 5906 5750 Max Range nm 454 at 136kts 477 at 136kts Max Endurance hrs 4.24 4.31

2.1.3. Trim analysis module

Controlling the helicopter is a nonlinear control system design problem. an important part of practical helicopter control is the ability to determine the trim states for the helicopter over all flight conditions. the aircraft is trimmed when the desired balance is achieved or the aircraft enters a desired steady state. The controls to be trimmed are the actuators and the dependent states to be trimmed are the pitch, roll, and yaw, i.e., These states are trimmed for the desired specified steady-state translational velocities and angular rates, variables representing desired steady states.10

Harmonic balance method was used in this trim code.

The following flight condition was implemented in this code:

+ Straight flight (with or without: climb, sideslip) + Vertical flight

+ Turning flight (with or without: climb, coordinated or uncoordinated)

Figure 3 showed some validated results of this trim code.11

Atmosphere Conditions Design Gross Weight

Power Required V > Vmax

Vstart Vstart +VStep

Specific Range Max Range Max Endurance Velocity for SRM & FFM True

False

Fuel Flow (FF) (lb/h) Specific Range (SR) (n.m/lb) Specific Range Max (SRM) = 0 Fuel Flow Min (FFM) = Large Number

SRM = SR SR > SRM True False FFM = FF True False FF < FFM Set up flight condition Trim analysis code Call

(4)

2 2 o T s c g c T e a T 2 w c 2 a. Trim b. A Fig. 3 Tri 105 Helic 2.2. 2KFoil 2KFoil is an isolated airfo be suitable w of this code methods w interaction me The inviscid f stream funct compressibili good compre conditions.12 The Viscous layers and w equation lagg and an envelo The airfoil is 2KFoil as air will generate correspond number, and 2.3. Others Some additio 0 50 100 -6 -5 -4 -3 -2 -1 0 1 2 3 4 Airsp P itc h A tt itude ( d eg) 0 50 100 8 8.5 9 9.5 10 10.5 11 11.5 12 12.5 13 Airsp C o lle c tiv e (d e g) code program AGARD GAT m and contro copter. l Program n airfoil analy oils was modi

with present s is a combin ith a fully ethod.12 formulation o tion panel m ty correction essible predic formulations wake which ged dissipatio ope entransit going to be a rfoil coordinat lift, drag, mo to specific Reynolds nu s Program onal codes to 0 150 200 250 peed (km/h) 0 50 1 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 Ai B an k A ngl e ( deg) / HY HO )O L J K W 7U L P DQ G 0 150 200 250 peed (km/h) 0 50 1 -2 -1.5 -1 -0.5 0 0.5 1 Ai La ter al C y c lic ( deg) m results TUER experim ol parameters ysis program ified from Xfo study. The m nation of hig -coupled vi f 2KFoil is a method. A K is incorpora ctions all the

s come from are describe on integral bo ion criterion.1 analyzed will te points, an ment coefficie angle of a mber. generate airf 100 150 200 250 rspeed (km/h) G &R Q W U R O 3DU DP HW HU V 0 50 0 1 2 3 4 5 6 7 8 9 10 T a il R o to r C olle c tiv e ( de g ) 100 150 200 250 rspeed (km/h) 0 50 -6 -5 -4 -3 -2 -1 0 1 Lo ngi tud in al C y clic ( d eg) ment data Results of Bo m of subsoni oil program to main algorithm h-order pane scous/inviscid linear vorticity Karman-Tsien ated, allowing way to soni the boundary ed with a two oundary laye 2 be inputted to d then 2KFo ent CL, CD, CM attack, Mach foil coordinate 100 150 200 250 Airspeed (km/h) 100 150 200 250 Airspeed (km/h) o c o m el d y n g c y o er o il M h e point, ch impleme optimizat These co chord d distributio 3. GEO MET CST rep Figure 4 expressio shapes.2 “shape fu By using distribute ) / (x c y Where N1, N2 c x S( / ) x: non-d c: curve c is chord For for polynom i x S( ) Where n : orde i : numb CST m Figure 4 non-dime exponen defined. coefficien shape fu function a Fig. 4. method. hord distributi nted in order tion process i odes provide distribution on… OMETRY REP THOD 13 resentation m 4. This met ons to repre The compo unction” and “ g CST metho ed by followin ) / ( 1 2 x c S CN N ˜ 2 1( ) ( N N x c x C : Exponents

>

¦

N ˜ i i i x c A 0 ) / ( dimensional v e length (if sh d length) rmulation of

ials are used

i n i ix x K (1 )  ( ! i i n K ¸¸¹ · ¨¨© § { er of Bernstei bers 0 to n method follows 4. First, the ensional valu ts and deg Then, in fi nts are calc unction and c are obtained. Representa

ion and twist r to build a in Model Cen various way such as: PRESENTAT method’s proc thod is bas esent and m onents of th “class function

od, the curve g equation: ) / (x c S 1 ) 1 ( )N x c N c 

@

: shape func values from 0 hape is like a f CST me as a shape fu i )! ( ! i n n  : binom n polynomial s the proces given data p es. Second, gree of sha tting process ulated. Final class functio ation proced t distribution full framewor ter software. s to present linear, non ION CST cedure is show sed on anal modify the va is function a n”. e coordinates 2: Class func ction, 0 to 1. airfoil upper c ethod, Bern unction. mial coefficien s which show points conver the class-fun ape function s shape fun ly, by multip n, the distrib dure using were rk for twist, linear wn in lytical arious are – s are tion curve, nstein nts wn in rts to nction are nction plying bution CST

(5)

Figure 5 shows the airfoil geometry is represented by using CST method and NURBS. In this case, control variables are the coordinates of control points (5 variables for upper curve and 5 for lower curve). CST method with 4 control variables has a better fit to existing airfoil than NURBS which uses 10 control variables.13

Fig. 5. RAE 2822 Airfoil representation. Figure 6 shows the absolute errors of an airfoil generation with CST and NURBS (5 control points for each curve, 4th order blending functions). Generation by NURBS at the tail part of airfoil has bigger errors.

Fig. 6. Absolute errors in airfoil generation.

The advantage of CST method in comparison with others method such as Spline, B-Splines or NURBS is can represent curves and shapes using few scalars control parameters and very accurately. In this study, airfoil baseline was chosen as NACA 0012. With the given data coordinate points in Cartesian coordinate space, a curve fitting was generated using 4th order Bernstein polynomials. The Class function for the airfoils was:

) 1 ( ) (x x0.5 x C 

Airfoil distribution function defined as upper curve and lower are presented sequently as below.

» » ¼ º « « ¬ ª         4 4 3 3 2 2 2 3 1 4 ) 1 ( 4 ) 1 ( 6 ) 1 ( 4 ) 1 ( ) ( x A x x A x x A x x A x A x C x y l l l l lo l

» » ¼ º « « ¬ ª         4 4 3 3 2 2 2 3 1 4 ) 1 ( 4 ) 1 ( 6 ) 1 ( 4 ) 1 ( ) ( x A x x A x x A x x A x A x C x y u u u u uo u Where: Au0 = 0.1718; Au1 = 0.15; Au2= 0.1624; Au3 = 0.1211; Au4 = 0.1671; Al0 = -0.1718; Al1 = -0.15; Al2 = -0.1624; Al3= -0.1211; Al4 = -0.1671 4. DESIGN PROCESS 4.1. Design Considerations

Helicopter hover performance is expressed in terms of power loading or figure of merit (FM). In this study we assume rotor thrust and weight of helicopter are equal. Therefore, the hover power required should be made as small as possible.The hover power requried to drive the main rotor is formed by two components: induced power, profile power (to overcome viscous losses at the rotor). The induced power and the profile power primarily influence the blade aerodynamics performance design.7

The conventional approach of blade aerodynamics performance design is started with the selection of the airfoils which could be applied over various regions of the blade radius. The choice of airfoils is controlled by the need to avoid exceeding the section drag divergence Mach number on the advancing side of the rotor disc, exceeding the maximum section lift coefficients on the retreating side of the rotor disc.1

The present work considers the effect of blade airfoil shape on power required. Therefore, a baseline airfoil NACA0012 was chosen as a unique airfoil through blade to simplify the process of optimum design. Moreover, this approach can deal with various helicopters which operate in various velocity ranges. The considerations of selection of base line airfoil are skipped.

A changes of A0 and A4 coefficients of CST method are sufficient for airfoil shape modification.2 These coefficients were also the design variables of examined optimal problem.

With above mentions, this approach leads the induced and profile power are functions of twist, taper ratio, point of taper initiation, blade root chord, A0 and A4 coefficients of airfoil distribution function. Satisfactory aerodynamics performance is defined by the following requirements:1

+ The power required must be less than the power available.

+ The helicopter must be able to trim at hover flight condition. 0.08 0.06 0.04 0.02 0 0.02 0.04 0.06 0.08 0 0.2 0.4 0.6 0.8 1 NURBS C urve C S T C ontrolPoints 0.005 0 0.005 0.01 0.015 0.02 0.025 0 0.2 0.4 0.6 0.8 1 NURBS upper curve NURBS lower curve C S Tuppercurve C S Tlowercurve

(6)

Fig. 7. Design synthesis process

4.2. Design synthesis process

The design synthesis process was showed in figure 7. The rectangular with dash line represented a module which is integrated in model center software. Each module is connected with others module by data input/output flow which are the mutual parts. The arrows pointing into modules present the design variables, while the one pointing out present the objective function.

There are 4 modules were implemented in this optimization framework which are chord, twist, radius distribution generation module; airfoil points coordinate generation module; airfoil characteristic library with C81 format module; sizing, trim, performance analysis module. The chord twist, radius distribution were generated by a code in which the geometry representation can be change, such as linear, non linear function. In this study, chord distribution generated base on root chord, the point of taper initiation, and taper ratio. Twist distribution is assumed varies as linear function along the blade. Radius distribution was divided by equal annulus area of rotor disk. These distributions would be the input data for trim code in trimming process.

Four coefficients of airfoil distribution function were

defined as initial input data of design process after getting the fitting curve of airfoil baseline

NACA0012. Then, airfoil coordinate points were generated by using CST function. With airfoil coordinate points, 2KFoil will generate airfoil characteristics library with C81 format which are the airfoil lift, drag, moment coefficients with respect to angle of attack for different mach numbers (from 0.1 to 0.9).

KHDP program with performance analysis module

provides many options for objective function. The objective function of this study was chosen as power required in hover. Helicopter data are going to be analyzed by performance code come from either sizing module or user inputs.

After getting the trim condition, namely, trim condition was attainable and performance as well, the power required will be evaluated to proceed the next loop of optimization process. So, a new set of initial data (root chord, the point of taper initiation, taper ratio, pretwist, A0 and A4 coefficients of airfoil distribution function) will be generated depending on optimization algorithm. This loop will be stop until convergence condition is satisfied.

5. OPTIMIZATION FORMULATION AND

METHOD

5.1. Design variables

The design variables are maximum pretwist, taper ratio, point of taper initiation, blade root chord, A0 and A4 coefficients of airfoil distribution function. The blade is rectangular to station of the point of taper initiation and then tapered linearly to the tip.14 The twist varies linearly from the root to the tip. NACA0012 was chosen as baseline airfoil, and two coefficients A0, A4 are design variables of airfoil shape.

5.2. Constraints

The power required in hover must be less than power available.

The trim constraint in hover is implemented by expressing the constraint in terms of the number of trim iterations iter, the maximum number of trim iterations allowed itermax.

(7)

A c W c T T d 5 T f o A w f s Another cons chord does n Where ct is th chord allowed This constrai range shown The magnitu distribution fu 5.3. Object The perform function of op In this study, objective func Fig. 8. Des Center. All modules w which is a integrating d framework an shown in figu straint used to ot become to he tip chord a d. c dt nt can be dec in table 4. de of A0 an unction are les

gi A0, tive function mance modul ptimization pro power requir ction. ign framewo was wrapped powerful t design code nd link data ure 8. Design o ensure that oo small.

and ctmin is the

min t c cribed in term d A4 coeffici ss than 1. 0 1 4  d A n and optimiz e provides oblem can be red in hover w rk and link d d in Model Ce tool for aut es. A snap among each Explorer too

t the blade tip e minimum tip m of taper ratio ents of airfo zation tool the objective e very various was chosen as data in Mode enter program tomating and shot of this module were ol was used to p p o il e s. s el m d s e o perform t Design systemat space us and the problem surrogate expensiv make glo systems The sur Design E To crea executes multiple t a table. runs are space (u Fig. 9. T Design E The proc Explorer models a optimizat mechanis minimum best des Fig. 10 the optimizati Explorer's tic and effic sing design o intelligent u analysis an e models serv ve and "nois obal analysis practical. rrogate The Explorer are ate a surrog s analysis c times, and st The input va e chosen to sing an Ortho The process Explorer. cess of usin tool was sho are selective tion process sms are im m. A final patt ign found is a . Design Expl ion search us key techno cient samplin of experiment use of "surro nd optimizat ve as substitu sy" computer s and optimiz surrogate interpolating gate model, code (Mode tores the resu ariable values efficiently ca ogonal Array) of using su ng surrogate own in figure ly updated a s progresses mplemented ern search g at least a loca

lorer for optim

sing Model Ce ologies are ng of the d ts (DOE) met ogate" model tion. the sm utes for poten r simulations zation of com models used Kriging mod Design Exp l Center m ults of each r s for this seri anvas the d . urrogate mod model in D e 9. The surro and refined a s. Global se to avoid uarantees tha al minimum. mization probl enter. the esign thods ls for mooth ntially s and mplex d by dels.15 plorer model) run in ies of esign del in esign ogate s the earch local at the em.

(8)

A gradient based optimization algorithm (sequential quadratic programming) is in conjunction with the surrogate models to predict the optimum design for design problem.3

The Design Explorer tool was used for this problem was shown in figure 10. And boundary of each design variable was summarized in table 4.

6. RESULTS AND CONCLUSION

Figure 11 shows sensitive analysis of each design variable on objective function. These analysis tell us that those coefficients coming from airfoil distribution function had an important role on performance of rotor. Therefore, this study has told us that airfoil shape should be considered as design variables. This optimizaion problem was applied on rotor blade of Bo 105 LS helicopter.

Table 3 shows optimum results in which the objective function reduced 7.4% and FM increased 6.5%.

Table 3. Optimum results.

Baseline Optimization Improvement

AU0 0.1718 0.28027 AU4 0.1671 0.2293 AL0 -0.1718 -0.0755 AL4 -0.1671 -0.1256 TAPR 1 0.2 POTAP 1 0.5 CHOR 0.27 0.24 TWIST -8 -15.3 FM 0.72 0.77 6.5% ITM 6 12 Power HP 687.44 636.22 7.4% Where:

TAPR: Taper ratio; POTAP: Position of taper initiation; CHOR: Chord length; FM: Figure of merit; ITM: Number of trim iteration; AU0, AU4, AL0, AL4: Coefficients of airfoil shape distribution function. This study was performed for hover case only. Therefore, we can see that the optimum taper ratio and position of taper are on boudary of these design variables.

The optimum blade shape may have smaller solidity in comparison with baseline. In this case, twist was decrease from -8deg to -15.3deg.

In any case of airfoil baseline, airfoil shape represented by 2 coefficients for upper curve AU0, AU4, and 2 coefficients for lower curve AL0, AL4, are always show an important role in effective performance of rotor. By using CST method, we can represent airfoil curve with few coefficients that is resonable to perform an optimization problem. A further study on rotor blade design in forward flight

and maneuver flight also need to take a consideration of airfoil shape. The requirements that the airfoil section not stall in forward flight and that the drag divergence mach number be avoided.

REFERENCE

[1] Joanne L. Walsh, ‘Performance Optimization of Helicopter Rotor Blades’, NASA Technical Memorandum 104054, 1991.

[2] Brenda M. Kulfan, `A Universal Parametric Geometry representation Method – “CST”`, AIAA-2007-62, 45th AIAA aerospace Sciences Meeting and Exhibition, Reno, Nevada, USA, 8-11 January, 2007.

[3] Model Center software, ‘manual’.

[4] Ho-Jung Kang, Hyeong-Uk Park, Vu Ngoc Anh, Jae-Woo Lee, Chang-Joo Kim, Yung H. Yu, Chul-Ho Kim, Yoo-Sang Hwang, and Yong-Jin Chang, ‘Development of Robust Design and Optimization Process for Unmanned Rotorcraft Design’, AHS international 65th Annual Forum & Technology Display, 2008.

[5] AMCP 706-201 Engineering Design Handbook: Helicopter Engineering, Part One Preliminary Design, HQS U.S. AMC, August 1974.

[6] Simonds, R.M., “A Generalized Graphical Method of Minimum Gross Weight Estimation”, The National Conference of the Society of Aeronautical Weight Engineers, Inc., San Diego, CA, May 1956.

[7] J. Gordon Leishman, “Principles of Helicopter Aerodynamics”, Maryland University, 2006. [8] W. Z. Stepniewski, C. N. Keys, “Rotary-Wing

Aerodynamics”, 1909.

[9] Performance Data Report, Hiller Aircraft corporation, 1955.

[10] Russell Enns and Jennie Si, “Helicopter Trimming and Tracking Control Using Direct Neural Dynamic Programming”, IEEE Transaction on Neural Networks, Vol 14, No 4, July 2003.

[11] KHP project final report by Konkuk university. [12] Xfoil program, ‘Tutorial’

[13] Abdulaziz Irgashevich Azamatov, Jae-Woo Lee, Yung-Hwan Byun, Sang-Ho Kim, ‘Advanced Configuration Generation Technique for the Complex Aircraft Geometry’, Advanced Intelligent Mechatronics(AIM) 2008. IEEE/ASME International Conference, China, 2-5 July 2008 [14] Joanne L. Walsh, William J. LaMarsh II, Howard

M. Adelman, ‘Fully Integrated Aerodynamic /Dynamic Optimization of Helicopter Rotor Blades’, NASA Technical Memorandum 104226, 1992.

[15] Booker, a. j., "design and analysis of computer experiments," 7th aiaa/usaf/nasa/issmo symposium on multidisciplinary analysis &

(9)

optimization, st. louis, mo, (sept. 2-4, 1998) pp. 118-128. aiaa-98-4757.

APPENDIX

Fig. 11. Sensitive analysis of design variables

AU0 0.3 0.25 0.2 0.15 0.1 PO W ER _ H O VER 1000 980 960 940 920 900 880 860 840 820 800 780 760 740 720 700 680 AU4 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 PO W ER _ H O VER 703 702 701 700 699 698 697 696 695 694 693 692 691 690 689 688 687 686 685 AL0 -0.1 -0.15 -0.2 -0.25 -0.3 -0.35 -0.4 PO W E R_ HO VER 720 715 710 705 700 695 690 685 AL4 -0.05 -0.1 -0.15 -0.2 -0.25 -0.3 -0.35 -0.4 PO W ER _ H O VER 701 700 699 698 697 696 695 694 693 692 691 690 689 688 687 686 685 684 683 TAPR0.75 0.80.85 0.90.95 1 0.7 0.65 0.6 0.55 0.5 P O W E R_ HO V E R 684 681 678 675 672 669 POTAP 1 0.95 0.9 0.85 0.8 0.75 0.7 0.65 0.6 0.55 0.5 CHOR 0.3 0.29 0.28 0.27 0.26 0.25 0.24 0.23 0.22 0.21 0.2 PO W E R _ HO VER 690 688 686 684 682 680 678 676 674 672 TWIST 16 14 12 10 8 6 PO W ER _ HO VER 686 685.9 685.8 685.7 685.6 685.5 685.4 685.3 685.2 685.1 685 684.9 684.8 684.7 684.6 684.5 684.4 684.3 684.836 682.858 680.879 678.901 676.922 674.944 672.965 670.987 669.008 667.03

(10)

Table 4. Design variables range.

minimum maximum Baseline

AU0 0.05 0.5 0.1718 AU4 0.05 0.5 0.1671 AL0 -0.5 0.05 -0.1718 AL4 -0.5 0.05 -0.1671 TAPR 0.2 1 1 POTAP 0.5 1 1 CHOR 0.2 0.35 0.27 TWIST -16 -5 -8 FM 0.69 1 0.69 ITM 1 15 6

Fig. 13. Optimum results

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -0.1 -0.05 0 0.05 0.1 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 -0.2 -0.1 0 0.1 0.2 Baseline Optimization Shape

Referenties

GERELATEERDE DOCUMENTEN

The multicenter randomized-controlled trial, Remote Ischemic Conditioning in Renal Transplantation-Effect on Immediate and Extended Kidney Graft Function (CONTEXT) trial, examines

Een moeder vertelt wat het met haar gevoel van eigenwaarde deed toen haar kinderen niet meer bij haar thuis mochten slapen: ‘Nou dan begin je jezelf af te vragen van: wat

It was really nice to work with you all and I shall remember my time in this group. I would like to thank Yvonne and Hilda for their help in administrative

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright

Here, we report on the design and engineering of a complete in vitro phospholipid biosyn- thesis pathway using eight puri fied (membrane) proteins, to realize the enzymatic conversion

In Tanzania, the socio economic setting from where an entrepreneur originates influences how the person can access resources such as cultural capital (formal education, information

After the digestion by BamHI and XbaI, mtg-6his was fused after AmyQ signal peptide encoding gene resulting in pHT43 mtg where the recombinant SP amyQ -mtg-6his.. will be controlled

Het zorg- en vastgoedbedrijf zijn vaak twee aparte entiteiten, daarom beschouwen wij drie mogelijkheden om de (financiële) verrekening met het zorgbedrijf vorm te geven: markthuur