Multidisciplinary Optimisation of a Business Jet Main Exit Door Hinge for Production by Additive Manufacturing

Multidisciplinary Optimisation of a Business Jet Main Exit Door Hinge for Production by Additive Manufacturing Martin Muir Research Engineer, EADS Innovation Works Airbus Operations Operations Ltd, Building Building 20A1 New Filton House, Filton, Bristol, BS34 7QT [email protected]

Abstract Additive Manufacturing (AM) using Powder Bed Fusion (PBF) processes is a novel and rapidly expanding manufacturing field which eliminates many conventional conventional constraints from f rom the manufacturing process. With the reduction of design inhibitors such as tool paths, AM becomes an ideal process for the fabrication of topology optimised (TO) structural designs. However, the relatively modern processes, particularly for metallic technologies, introduce additional design complexities which must be accounted for during optimisation in order to realise the maximum potential of the structural design. This work details the process from inception to delivery of the design optimisation of the MED (Main Exit Door) hinges for the Gulfstream G280 aircraft. Through several examples, a comparison of stress/mass and compliance topology optimisation for static and fatigue cycled structural parts with novel material properties and constraints will be evaluated. The report shows that significant mass reductions can be achieved even with heavily dimensionally constrained structures, subjected to high static and dynamic loadings. Finally the work assesses the suitability of AM PBF processes for the manufacture of TO components suitable for civil aerospace structural classes. The final summary highlights benefits not only from in service mass reduction, but also aesthetic considerations of visible structural parts and the combinatory benefits from the linked use of TO and AM. Keywords: Additive Manufacturing, Power Bed Fusion, topology optimisation,

1.0

Introduction

1.1

Introduction and Context

Optimisation has, in one form or another, been a constant factor in engineering design since the earliest foundations of the creed. Indeed the search for the best from a given set of conditions is something to which almost everyone, in almost every disciple can relate. Facing some of the most strenuous constraints, in demanding often competing environments, aviation engineering has forever been at the forefront of engineering design and technical innovation. Speed, range, survivability, and versatility have often been a the driving criteria, the objective if you will, for which many designs have been catered; despite this, and in most of not all cases, cost has been a considerable function to be included and equated into each design case. Historically, cost has been a function in the required design plan…recently however, cost, economy, efficiency efficie ncy have become the function, the objective for the next generation of commercial aviation platforms (Krog, Tucker, Kemp, & Boyd, 2004). 2004). © Altair Engineering 2013

1

In response to this economic and environmental trend, several joint European projects and commissions have arisen whose aim is to reduce (through various means) the overall contribution that aviation makes to global CO2 production. Cleanskies is one such initiative, and the one under which this research is undertaken. 1.2

Aim/Objective

The principal aim of the project was to assess the feasibility for a manufacturing change from a complex twin part casting and welding operation, to a single piece design produced using additive manufacturing. manufacturing. Further objectives are aimed toward component redesign in order to affect further efficiency savings in either the manufacturing or in-service phases of t he product life cycle.

2.0

Methodology

2.1

Existing Part Analysis Analysis and Investigation

The Israeli Aircraft Industries (IAI) Gooseneck Hinge(s) are a relatively small component (200*240*50mm) installed on the main exit door (MED) of Gulfstream Aerosp ace’s G200/G250/G280 Aircraft. The door is of an Airstairs Plug Type design which rotates upon two Gooseneck hinges (GNH) aligned along the main aircraft longitudinal longitudinal axis, and attached att ached in proximity to the t he fuselage/cabin floor. The installed hinges must allow unobtrusive function of both the door and its incorporated stairs, allowing safe operation and simple entry/egress from the cabin. Major considerations for the design of the hinges are kinematic, though with a single airworthiness consideration related to the accidental opening of the door during take-off phase, and a secondary consideration related to the fatigue loads caused by repeated deployment/stowage of the hatch/stairs and subsequent passenger usage. The installed setup can be seen in Figure 1. 1.

F i g u r e 1 - MED MED of Gulfstream A erospace G250 G250 show ing GN hin ges

© Altair Engineering 2013

Multidisciplinary Multidisciplinary Optimisation of a Business Jet Main Exit Door Hinge for Production by Additive Manufacturing

2

2.2

Current Material and Design Drivers:

Presently, the GNH is fabricated from 15-5PH H1025 (15%Cr, 5%Ni with precipitate hardening) Stainless Steel which has a Yield Strength (YS) of 1000MPa and a Ultimate Tensile Strength (UTS) of 1070MPa when loaded either longitudinally or transversely. The current design is a gestalt entity composed of cast and welded sections allowing for a hollow core in the neck region and solid attachment points for the door hinge point and the main rotational hinge point. Two major loading conditions exist in conjunction with a single boundary condition included for each of GNs. The loads to be included included are; an aero load based on a failure failure case, and a fatigue load based on cyclical loading of the MED during use. Both of the applied load cases were provided by IAI and verified prior to application. The complexities of the current design mean that care must be taken, assuring that high stresses are not formed in areas around the main weld seams, such as that running down the length of the neck. The current iteration of the design has been modified in order to address a susceptibility to fracture on the lower hinge (deployed position) of the gooseneck due to cyclic stress.

aircraft no se) F i g u r e 2- MED of Gulfstream A erospace G250 (looking aft fro m aircraft show ing GN Hinges in both stow ed (left) (left) and deploy ed (right) (right) condition s.



3

MED of Gulfstream Aero space G250 G250 (from (from above) showing GN hinge F i g u r e 3 - MED placement and distance between hinges 555mm

F i g u r e 4 – Goo seneck Hin ge (revised) fro m G200/250/ G200/250/280 280 MED

2.3

The Aero Failure Load:

F i g u r e 5 - MED MED of Gulfstream A erospace G250 G250 show ing c entre of area



4

1.1.1. Aerodynamic Aerodynamic Drag Force Assuming sea level conditions and with the force acting directly along the aircraft longitudinal axis (x) the component forces are as follows:

Where V = aircraft aircraft takeoff velocity and A is the area of the door exposed to the air.

The additional component force acting in the lateral/transverse lateral/transverse aircraft axis (y) is based on a moment force spread over a distance of application and has the following magnitude:

The force (Fy) is in a positive direction (tension) when acting on the forward hinge and a negative direction (compression) when acting on the rearward hinge due to the rotational forces applied to the MED by the drag force. 1.2.

The Fatigue Load

This load case is one defined by anticipated usage, MED mass, and the distance of action for that mass from the lower hinge line.

F i g u r e 6 - MED of Gulfstream Gulfstream A erospace G250 G250 show ing centre of m ass and distance (982 (982mm mm ) from pivot © Altair Engineering 2013


5

Due to the nature of the hinge, there is both a static force (A) and coupled, equal, parallel reaction force (B) imposed on the hinge. Both of these forces (A, B) have been calculated (IAI Report) using standard numerical calculation to be ~12KN, and are were applied at 100% load over 80000 cycles as per Figure 7

A

F i g u r e 7 – GN showin g location of applied fatigue load and and reaction forces

1.3.

Boundary Conditions

There is a single static connection point based around the upper hinge pivot ( Figure 4) 4) connected using either a rigid spider with a single 6DOF constraint (somewhat unrealistic) or carefully selected and applied 6DOF BCs applied to the loaded surface on the inner edge of the upper hinge pivot. This spider was eventually selected as it most closely resembled the conditions in the IAI report detailing the fatigue case. 1.4.

Requested Solution

IAI have chosen this part as the focus of an investigation designed to utilise new materials and green metallic technologies and approaches as part of the B2 Low weight green metallic fuselage section. The involvement of EADS Innovation Works was intended to reduce the mass of the part through the inclusion of structural topology optimisation with manufacture through powder bed AM processes, and more specifically direct metal laser sintering (DMLS). 1.5.

Summary



6

3.0 Suitability for for AM and Selection Selection of Process Process

F i g u r e 8 – GN Hinge show n in optim al build orientation along with assoc iated iated X and Y dimensions

Several manufacturing factors must be considered before an optimisation process is undertaken, as their inclusion or neglect can have significant impact upon the converged design. First amongst these factors is that given to the geometry of the component and its optimal orientation for build; immediately the GNH presents a problem as the optimal orientation (Figure 8) 8) gives a maximum height of almost 280mm, exceeding the available platform height in some machines; thereby limiting available manufacturing options. Further considerations must be given to powder removal in the post build environment; this is especially important for this design as the current GNH is structurally shelled within the neck region. As all AM processes would deposit irremovable powder within this region, design modification is required to support the use of the AM processes. Finally, additional consideration must be given to the requirement for process-specific scaffolding structures which are a pre-requisite of powder bed AM processes. 3.1 Conclusions Whilst both the DMLS and electron beam melting (EBM) processes are capable of fabricating the part in the optimal orientation, it is on t he very fringe of each of the machines operating capabilities and would likely increase the chances of build problems. Furthermore, whilst both processes are capable of building variants steel, the requirement to reduce structural mass leans favourably toward the application of a material change to Ti6/4. Finally, due to the know deficiencies of as built AM components when subjected to fatigue loads, particularly in DMLS processes, a design optimisation using the processing factors and established material properties for titanium fabrication using EBM was selected (Tomlin & Meyer, 2011). 2011) . © Altair Engineering 2013


7

4.0 Preliminary Analysis 4.1 Initial Finite Element Model Validation An initial investigation was undertaken in order to establish a suitable baseline of comparison for the EADS IW FE model against established results from IAI. Using material data for 15-5Ph Steel applied to an unstructured TET10 element mesh, grid independence was established for the operable domain and a comparable result attained. Figure 9 and Figure 10 show the model comparisons with the newer design showing comparable trends when subjected to the failure case loading.

– IAI GNH F i g u r e 9 GNH FE stress m odel (ksi) Whilst comparable, the EADS model does show a higher area of stress in the lower neck region Figure 10 than is visibly present in the IAI model data. However, this result is consistent with the expected loads, material and geometry.



8

– Results of analysis show ing areas of high str ess attained attained using IAI BCs F i g u r e 10 and results demon strating grid independence independence

Though no constraints upon component displacement were provided by IAI, the validated model was analysed for maximum displacements under both established loading conditions in order to determine additional optimisation constraints to be used within the solver.

F i g u r e 11 – Maximum displacement of the hinge when sub jected jected to the Aero Load

The maximum deflection (Figure 11) 11) of the hinge under the failure load case is ~6mm at the tip, which led to the application of a 10mm (total) maximum displacement constraint within the optimisation. optimisation.

5.0 Problem Definition 5.1

Requirement for Design Modification

Whilst the adoption of new manufacturing methods was the principal aim of the project, it was clear from the outset that some form of design modification would be required. When © Altair Engineering 2013


9

coupled with the increased likelihood of a required material change in order to meet target specifications, the case for a full redesign became unarguable (Wong & Hernandez, 2012). 2012) . 5.2

Design Brief

IAI were quite specific in their requirements for a new/derivative design, stating unequivocally that that the design must satisfy both loading conditions, whilst maintaining the current attachment and kinematic constraints. Furthermore, IAI had indicated that a sizeable mass reduction would be advantageous to the demonstrator program. The heavy constraint on the design domain coupled with the requirement for mass reduction within a structure already known to be operating at, and beyond its limit proved to be a challenging set of objectives. As an already lightweight design, it was known that the immovable design domain would present a particular challenge (Muir, 2012), 2012), if any mass reduction (beyond that achievable though material change) was to be attained (Turner, Clough, Martin, & Topp, 1956). 1956). 5.3

Optimisation Objective

Two methods of structural optimisation were selected for trials: a conventional compliance based optimisation, coupled with global stress, fatigue, volume and displacement constraints, and a minimum mass optimisation problem with similar constraints. It was known from the outset that the differing objective would yield differing topologies (REF) but a direct comparison between the two design outputs can often prove fruitful in establishing further design directions. 5.4

Optimisation Constraints

As previously previously alluded to, the fatigue performance of as built Titanium AM components is poor when compared to established manufacturing techniques. However, with post processing activities and careful design, the fatigue performance can be greatly improved. As such a maximum fatigue fatigue constraint constraint of 600MPa at 80000 cycles cycles was used for constraint constraint 1 (C1). A global constraint on maximum stress was applied at 800MPa giving constraint 2 (C2), with a deflection constraint of 10mm (max total) applied as constraint 3 (C3). Finally, a varied constraint for volume fraction was applied to the design domain under the compliance objective. 5.5

Design Domain

Though largely externally constrained, with the attachment points being excluded from the optimisation (Figure 12) 12), the operable design domain has at least been expanded to fill the entirety of the neck region with useable design space.



10

F i g u r e 12 - Showin g the separation of the hinge into designable (yellow) (yellow) and no n- desig nable (red/pu (red/pu rple) areas areas

1.6.

Grid Definition

Using an unstructured domain, it was found that grid independence was achieved using a 4mm element size when using TET10 Elements or with a 1.5mm element size when using TET4s. TET10s were ultimately used for validation and approximation stages with a highly dense TET4 (Figure 13) 13) mesh being used for the final optimisation in order to achieve a smoother output.

TET4 Grid w ith 1.5 element si ze F i g u r e 13 - TET4

6.0 Optimisation Results: The primary design result (Figure 15) 15) was obtained using a compliance based optimisation approach with an iteratively attained final volumetric fraction of 37% (63% mass removed from the neck region). Conversely, under near-identical optimisation constraints the mass objective topology optimisation failed to converge without significant relaxation of the operable constraints. Routinely, the structural optimiser would (at least partially) infringe © Altair Engineering 2013


11

upon one of the active constraints in order to satisfy another. These infringements were often minimal, but could not be completely removed without constraint relaxation. Figure 14 shows results from the mass objective topology optimisation which demonstrates a vastly different shape to the compliance based optimisation.

F i g u r e 14 - Mass objective topolog y optim ization ization result

F i g u r e 15 - Optimal design from a compliance based optim isation solver und er the given loads and constraints

The presented design had a predicted weight of ~650g compared to the original design’s 1530g representing a mass reduction of almost 58% when compared to the original design whilst exhibiting only a 10% increase in maximum stress. A second analysis was undertaken in which a symmetry plane was included along the largely symmetrical centre-line of the hinge. This was included in order to alleviate IAI concerns related to the asymmetric nature of the optimisation output and the perception that passengers would have upon seeing such a design upon entry and egress from fr om the aircraft. Figure 16



12

F i g u r e 16 - Optimal design design w ith the inclusion o f a longitudinal sym metry plane

The resulting symmetric design pays a substantial penalty for the inclusion of the additional constraint with the resulting mass saving being reduced from ~58% to only ~54% whilst exhibiting higher stresses than the optimal design when subjected to the same loading conditions.

7.0 Design Extraction and Validation Though heavier and exhibiting higher stress than the optimal design, design 2 with the inclusion of a symmetry plane was selected for extraction Figure 17 and production due to aesthetic reasons.

F i g u r e 17 - Extracted Extracted design based on the second ary design outp ut



13

F i g u r e 18 - Design validation show ing areas of stress wh en subjected to the aero load case

F i g u r e 19 – Design validation validation sh owin g areas areas of stress wh en subjected to the fatigue fatigue load

Design validation was completed complet ed using Altair’s OptiStruct OptiS truct program, with the secondary design offering significant reductions, in mass with only moderate increases in stress and deflection under both the aero (Figure 18) 18) and the fatigue (Figure 19) 19) load cases. Whilst it can be seen that significant significant mass reductions have been attained, the have been substantial reductions in the calculated reserve factors (Eui W. Lee, Charles S. C. Lei, & William E. F, 2001) especially within the chosen design.



14

Stress (MPa) Design

Mass (g) Peak

RF

Averaged

RF

Fatigue

RF

Disp (mm)

% Mass Reduction

Original

1547

670

1.42

610

1.56

646

1.16

6

NA

Orig Ti

866

725

1.17

689

1.24

711

1.1

11.2

44

Design 1

650

591

1.44

574

1.48

620

1.26

8.2

58

Design 2

710

740

1.15

650

1.31

607

1.29

9.1

54

F i g u r e 20 - Table Table showing th e maximum stresses in each each design and und er each each load case along along w ith the total mass reductions

8.0 Conclusions The optimisation of an already lightweight design is often a difficult process, no matter the application. application. If constrained constrained by conventional thinking and manufacturing manufacturing technology, it is likely that further mass reduction would either dramatically increase fabrication cost, or decrease component life. The freedom of design offered through the use of AM production helps to alleviate both detractors. The optimisation demonstrated within this report clearly highlights the benefits of a combined approach to design and manufacture demonstrating combined savings beyond the reach of either process. Though a significant mass saving is demonstrated in the results of this paper, it is beyond doubt that the material change has significantly significantly helped the overall mass reduction. However, though Titanium offers a 44% direct density saving over 15-5Ph Stainless Steel, this mass saving reduces to approximately 33% for the same deflection characteristics. However, the results clearly show that for a moderate decrease in stiffness, a substantial mass saving can be affected on even a lightweight failure prone structure. Whilst it is clear to see that there are potential advantages to be gained from the combined use of TO and AM, there are several factors which must be consider and included within the optimisation parameterisation if one wishes to ensure success. As built mechanical performance can be directly affected by myriad factors such as, processing parameters, feedstock material, build orientation, etc. can all influence as built material properties. However, it is an understanding of AM processes and its direct influence on component design which has the largest impact upon part performance. It is quantification and inclusion inclusion of this design aspect within the optimisation that will have the greatest influence on the further integration of structural optimisation optimisation and additive manufacturing. manufacturing.

9.0 References [1]

Eui W. Lee, Charles S. C. Lei, & W illiam E. F. (2001). Applications, (2001). Applications, Benefits, Benefits, and Implementation Implementation of Ti-6Al-4V Castings. Paper presented at the Cost Effective Application Application of Titanium Alloys Alloys in Military Platforms”,, Platforms”,, Loen, Norway,. Norway,.

[2]

Krog, L., Tucker, A., Kemp, M., & Boyd, R. (2004). Topology Optimisation of Aircraft Wing Box Ribs 10th AIAA/ISSMO Multidisciplinary Multidisciplinary Analysis and Optimization Conference: Conference: American Institute of Aeronautics and Astronautics. Astronautics.



15

[3]

Muir, M. (2012). Civil Aerospace Mass Reduction Through Automated Topology Optimization and Advanced Manufacturing. . Paper presented at the 9th ASMO UK/ISSMO Conference on Engineering Engineering Design Optimization, Optimization, Cork, Ireland.

[4]

Tomlin, M., & Meyer, J. (2011). Topology Optimization of an Additive Layer Manufactured (ALM) Aerospace Part. Paper presented at the The 7th Altair CAE Technology Conference, Gaydon, UK, 10th May.

[5]

Turner, M. J., Clough, R. W., Martin, H. C., & Topp, L. J. (1956). Stiffness and Deflection Analysis of Complex Structures. Journal of the Aeronautical Sciences (Institute of the Aeronautical Sciences), 23(9), 23(9), 805-823. doi: 10.2514/8.3664 10.2514/8.3664

[6]

Wong, K. V., & Hernandez, A. (2012). A Review of Additive Manufacturing. ISRN Mechanical Engineering, 2012 , 10. doi: 10.5402/2012/208760



16

Multidisciplinary Optimisation of a Business Jet Main Exit Door Hinge for Production by Additive Manufacturing

Recommend Documents