Design optimization of drive

screw
linear actuator
1. Introduction:
Linear actuators are the mechanical devices which creates motion in straight line.
Linear actuators are widely used in machine tools, industrial machinery and household
appliances. The system assembly of drive screw linear actuator consists of electric motor,
drive gear, pinion gear, drive screw, load-carrying nut and chassis providing bearing surfaces
and supports as shown in figure 1. The present project work is intended to optimize the
design of drive screw linear actuator to reduce its total cost. Since linear actuator has wide
variety of applications it is necessary to select suitable material and dimensions for drive
screw so that cost associated with it is minimized while ensuring it satisfies various
constraints imposed on it by the functional and space requirements.

Fig. 1. Linear actuator

Fig. 2. Drive screw design

1.1 Objective of the report:

The objective of this project work is to find drive screw dimensions so that cost
associated with material requirement of drive screw is minimized while it satisfies all
functional and geometric constraints.

5
1.2 Report outline:

The project report starts with defining the problem statement for optimization
problem and then converting it into the standard optimization problem. The design
optimization problem is then solved by using optimization method Monotonicity analysis
which is followed by computational optimization using SQP algorithm in MATLAB and GRG
(Generalized Reduced Gradient) in Excel solver. The results of these methods were
discussed and compared in the later section of report along with sensitivity analysis. At last
conclusion made based on discussion.

2. Mathematical model:
2.1 Objective function:

The objective function is to minimize product cost which consist of both material and
manufacturing cost. But the cost associated with manufacturing is considered as fix for
relatively small changes in design hence only material cost is included in the objective
function. From the geometry of drive screw as shown in figure 2 we can write objective
function as product of cost per unit volume of material and total volume of component.

( )

Where,

Cm = material cost ($/inch3) assume Cm= 1 for design purpose

d1, d2, d3 = diameters of gear/drive screw interface as shown in figure 2.

L1, L2, L3 = length of segments of drive screw as shown in figure 2.

2.2 Assumptions:

 Manufacturing cost assumed to be fixed as we are planning for mass production

small change in dimensions does not affect manufacturing cost.
 Standard unified threads are used.
 Assembly force for drive screw is concentrated at the midpoint.
 Frictional forces are only considered between threads and load nut and everywhere
else it is assumed to be negligible.
 Drive screw is considered as beam assuming L/di≥ 10.

6
2.3 Design variable:

 d1 = Diameter of shaft/gear interface.

 d2= Diameter of threaded portion.
 d3= Diameters of bearing surface segments.
 L2 = Lengths of screw having diameters as d2.
 L3 = Lengths of screw having diameters as d3.
 Nm= Number of teeth on drive gear.
 NS= Number of teeth of pinion gear.
 Nt = Number of threads per inch.

2.4 List of parameters:

 f = friction coefficient = 0.35 (for steel on plastic)

 Fa = Force required to assembly drive screw into the assembly = 6 lb
 τall = Maximum allowable shear stress = 22,000 psi
 σall =Maximum allowable bending stress = 20,000 psi
 K = Stress concentration factor = 3
 L1 =Length of gear/drive screw interface = 0.405 inch
 S = Linear cycle rate = 0.0583 inch/sec
 Sm = Motor speed = 300 rpm
 Tm = Motor torque = 2 inch-ounce
 W =Drive screw load = 3 lb
 αn =Thread angle in normal plane = 60˚

7
2.5 Constraints:

 Strength against bending stresses:

Where,

 Fatigue failure in shear due to torsion:

Where,

K= stress concentration factor= 3

T = applied torque =

, Tm = motor torque = 2 inch-ounce

 Specified linear cycle rate of oscillation:

Where,

C4 = 60-1 (no. of threads/rev) (min/S)

8
  No slip condition

(W* )[

Where,

W = drive screw load

 Space constraints: packaging consideration

 Upper bound on number of threads per inch imposed by mass production

 Limit on number of teeth on gears based on design consideration to avoid

interference

9
3. Optimization problem:
The above problem can be rewritten in the standard negative null form after eliminating
intermediate variables M,L,T,I,C1,C2 and J together with corresponding defining equalities.

Objective function:

Minimize ( )

Subject to:

10
DESIGN OPTIMIZATION
OF FLAT
HEAD CHAMBER OF AN
I.C. ENGINE
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 3

Abstract

This report discusses in detail, design optimization of a flathead chamber of an I.C. engine from
thermodynamic point of view. Detailed problem formulation, derivation of mathematical
expression for objective function and constraints has been discussed in detail. The goal is to obtain
preliminary values for a set of combustion chamber variables that maximize the power output per
unit displacement volume while meeting specific fuel economy and packaging constraints.
Optimization has been carried out using MATLAB’s interior-point optimization algorithm and
symbolic computation.
Constraints has been studied to identify feasible ranges of design variable and further feasibility
for their monotonicity has been studied along with objective function to understand behavior.
Interaction between design variables with respect to objective function has been studied.
Multiple initial conditions have been used to verify optimized result and the optimized result has
further been checked against design validity constraints. Activity and sensitivity of constraints has
been studied based on Lagrange multipliers and further parametric study has been done to
understand effects of engine’s thermodynamic and geometrical parameter on optimum solution.
Symbols used in model are listed in table-1.

Table-1 Symbols used in model.

Nomenclature for Engine model

Af air/fuel ratio
b cylinder bore, mm
BKW brake power/liter, kW/1
BMEP break mean effective pressure
cr compression ratio
Cs port discharge ratio
dE exhaust valve diameter, mm
dI intake valve diameter, mm
EGR Exhaust gas recirculation, %
FMEP Friction mean effective pressure, bars
h Compound valve chamber deck height, mm
H Distance dome penetrates head, mm
IMEP indicated mean effective pressure, kPa
isfc Indicated specific fuel consumption, g/kW-h
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 4

Nomenclature for Engine model

Nc Number of cylinders
Nv Number of valves
Q Lower heating value of fuel, kJ/kg
r Radius of compound valve chamber
curvature, mm
s Stroke of piston, mm
Sv Surface to volume ratio, mm-1
V, v Displacement volume, mm3
vc Clearance volume, mm3
vd Dome volume, mm3
Vp Mean piston speed, m/min
ω Revolution per minute at peak power, x 10-3
Zb RPM factor in volumetric efficiency
Zn Mach index of port and chamber design
g Ratio of specific heats
ηv Volumetric efficiency
ηvb Base volumetric efficiency
ηt Thermal efficiency
ηtad Adiabatic thermal efficiency
ηtw Thermal efficiency at representative part load
point
ρ Density inlet charge, kg/m3
φ equivalence ratio

1. Introduction
Brake power is actual power output at crank shaft of an I.C. engine, whereas indicated
power is power output obtained at piston due to expansion of combustion products.
However due to friction losses, the brake power is always less than indicated power.
Superior engine design is that one which gives maximum brake power while meeting
specific fuel economy and packaging constraints.

Brake power depends on several factors like thermal efficiency, volumetric efficiency,
Engine RPM, lower heating value of fuel, inlet charge density, air-fuel ratio, equivalence
ratio. However, these factors are not totally independent of each other, i.e. they interact
with each other. These factors are interrelated to each other by means of some basic engine
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 5

design variables like, engine bore diameter ‘b’, inlet-valve diameter ‘dI’, outer-valve
diameter ‘dE’, compression ratio ‘cr’ and RPM ‘ω’ along with some other parameters
(described later).
In this project maximization of brake power per unit engine displacement has been done
with respect to above mentioned design variables, while subjected to constraints which
have been derived in section of mathematical formulation.

In this report complex dependencies of brake power on above mentioned design variables
has been studied and interaction with the design variables has been depicted. Constraints
have been formulated from thermodynamics and packaging point of view and they have
been investigated for monotonicity and estimation of feasible domain for finding optimum
solution. Then MATLAB Symbolic computation is used for framing optimization
problem inside MATLAB environment. Advantage of symbolic computation is that it is
easy to calculate gradient and Hessians for objective and constraints function. Then
interior-point algorithm for optimization has been used to find optimum solution.

2. Mathematical formulation
Table-2 Values of constant used in model are listed in table below
Parameter Value Specification
Ko 1 Unit conversion, 4-stroke engine
120
K1 1.2 Cylinder separation as % bore
K2 2 Engine height as multitude of stroke
K3 0.82 valve spacing as % of bore (flat head)
K4 0.83 Lower bound on valve ratio
K5 0.89 Upper bound on valve ratio
K6 0.6 Upper bound on Mach-index
K7 6.5 Upper bound on RPM
K8 230.5 g/kW-h Upper bound on isfc
K9 2.4(106) mm3 Upper bound on displacement volume
K10 1.6(106) mm3 Lower bound on displacement volume
K11 1 Bore fraction spec. for deck height
64
K12 0.125 Bore fraction spec. for valve distance
L1 400 mm Upper bound on engine block length
L2 200 mm Upper bound on engine block height
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 6

Geometry of flat head design is shown in Figure-1

Figure-1 Schematic for flat-head design geometry

Objective is to Maximize BKW/V:

𝐵𝐾𝑊
= 𝐾𝑜(𝐵𝑀𝐸𝑃)𝜔 (1)
𝑉

1
where 𝐾𝑜 = is a unit conversion constant and
120

BMEP = IMEP – FMEP (2)

𝜌𝑄
Here 𝐼𝑀𝐸𝑃 = 𝜂𝑡 𝜂𝑣 ( ) (3)
𝐴 𝑓

𝜌𝑄
The term ( ) is amount of energy available in fuel-air mixture per unit volume. The volumetric
𝐴𝑓

𝜌𝑄
efficiency accounts for flow losses and the product. 𝜂𝑣 ( ) is the energy per unit volume
𝐴𝑓

available in the mass inducted into the combustion chamber. The thermal efficiency accounts for
the thermodynamics associated with the Otto cycle. The volumetric efficiency can be expressed as
1 + 𝑧𝑏2
𝜂𝑣 = 𝜂𝑣𝑏 ( )
1 + 𝑧𝑛2
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 7

where ηυb is the base volumetric efficiency, Zb is the rpm factor in volumetric efficiency, and Zn is
the Mach Index of the port and chamber design. The base volumetric efficiency for a “best-in-
class” engine may be expressed empirically with a curve-fitting formula in terms of w.

1.067 − 0.038ⅇ 𝜔−5.25 , 𝜔 ≥ 5.25

𝜂𝑣𝑏 ={
0 ⋅ 637 + 0.13𝜔−0.014 𝜔2 + 0.00066𝜔3 , 𝜔 ≤ 5.25

Also, empirically

Zb = (7.72 x 10-2) ω

𝑏 2 −5
(9.428) × 10 𝜔𝑠 ( )
ⅆ𝐼
𝑍𝑛 =
𝐶𝑠

Cs is port discharge coefficient, a parameter characterizing the flow losses of a particular

manifold and port design.
The thermal efficiency is expressed as follows

1.5
𝜂𝑡 = 𝜂𝑡𝑎𝑑 − 𝑆𝑣 √( )
𝜔

Where 𝜂𝑡𝑎𝑑 is adiabatic thermal efficiency given by

(1−𝛾)
0.9(1−𝑐𝑟 )(1⋅18−0.225𝜙) ,𝜙≤1
𝜂𝑡𝑎𝑑 = {
0.9(1−[𝛾1−𝑣 ])(1.655 −0.7𝜙) ,𝜙>1

In the optimal design model stoichiometry will be assumed, so that ϕ = 1 and the two expressions
in above equation will give the same result. The thermal efficiency for an ideal Otto cycle is (1 −
cr(1-γ)) and the 0.9 multiplier accounts empirically for the fact that the heat release occurs over finite
time, rather than instantaneously as in an ideal cycle. It is assumed valid for displacements in the
order of 400 to 600 cc/cylinder and bore to stroke ratios typically between 0.7 and 1.3.
Heat transfer is accounted for by the product of the surface-to-volume ratio of the cylinder, Sυ, and
an RPM correction factor. The surface-to-volume ratio is expressed as
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 8

𝜋𝑁
(0.83) [(8 + 4𝑐𝑟 ) + 1 ⋅ 5(𝑐𝑟 − 1) ( 𝑐 ) 𝑏 3 ]
𝑠𝑣 = 𝑣
[(𝑐𝑟 + 2)𝑏]

Finally, the FMEP is derived with the assumption that the operating point of interest will be near
wide open throttle (WOT), the point used for engine power tests, and that engine accessories are
ignored. Under these conditions pumping losses are small and the primary factors affecting engine
friction are compression ratio and engine speed. The resulting expression is

𝐹𝑀𝐸𝑃 = (4 ⋅ 826(𝑐𝑟 − 9.2)) + (7.97 + 0.253𝑉𝑝 + 9 ⋅ 7 × 10−𝜎 𝑉𝑝2 )

where Vp is the mean piston speed given by
𝑉𝑃 = 2𝜔𝑠
Constraints have been imposed based on packaging and efficiency requirements.
Packaging constraints are as follows. A maximum length of L1 = 400 mm for the engine block
constrains the bore based on a rule of thumb that the distance separating the cylinders should be
greater than a certain percentage of the bore dimension. Therefore, for an in-line engine
K1Ncb ≤ L1 (1c)

where the constant K1 = 1.2 for a cylinder separation of at least 20% of the bore, and Nc is the

number of cylinders in the block. Similarly, an engine height limit of L2 = 200 mm constrains
the stroke
K 2 s ≤ L2 (2c)

where K2 = 2. For a flat cylinder head, geometric and structural constraints require the intake and
exhaust valve diameters to satisfy the relationship
dI + dE ≤ K3b (3c)

where K3 = 0.82, and the ratio of exhaust valve to inlet valve diameter is restricted as
𝑑𝐸
𝐾4 ≤ ≤ 𝐾5 (4c)
𝑑𝐼

where K4 = 0.83 and K5 = 0.87. Finally, the displacement volume is a given parameter
related to design variables by
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 9

𝜋𝑁𝑐 𝑏2 𝑠
𝑉=
4
𝑉
With 400 𝑐𝑐 < < 600 𝑐𝑐 as model validity constraints
𝑁𝑐

Now efficiency related constraints are examined. To preclude significant flow losses due to
compressibility of the fuel/air charge during induction the Mach Index of the port and chamber
design must be less than K6 = 0.6.

𝑍𝑁 ≤ 𝐾6
The knock-limited compression ratio for 98 octane fuel can be represented by
𝑐𝑟 ≤ 13.2 − 0.045𝑏 (5c)
The rated rpm at which maximum power occurs should not exceed the limits of the torque
converter in conventional automatic transmissions, K7 = 6.5. Therefore

ω ≤ K7 (6c)
Fuel economy at part load (1,500 rpm, Af = 14.6) is a representative restriction on overall fuel
economy. Therefore, a constraint is imposed on the indicated specific fuel consumption, isfc, at
this part load:
3 ⋅ 6 × 106
ⅈ𝑠𝑓𝑐 = ≤ 𝐾8
𝜂𝑡𝑤 𝑄
where 𝜂𝑡𝑤 is the part-load thermal efficiency and K8 = 240 g/kW-h
In order to assign parameter values, we select specifications for a 1.9L four-cylinder engine. For
this typical engine configuration maximizing power is of definite importance. The following
values are then used for the parameters:

ρ = 1.225 kg/m3, V = 1.859(106) mm3, Q = 43,958 kJ/kg

Af = 14.6, Cs = 0.44, Nc = 4, γ = 1.33
The ratio of specific heat has been calculated using following formula with zero recirculation
assumed
𝐸𝐺𝑅
𝛾 = 1.33 + 0.01 ( )
30
All of the above constraints are valid only in limited range of bore to stroke ratios
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 10

𝑏
0⋅7≤ ≤ 1.3
𝑆
So, this is one more model validity constraint.
3. Model for optimization
Here objective function and constraints as derived in above section have been cast out into
negative null form. So, optimization problem is as follows
Minimize
𝜌𝑄
𝑓(𝑐𝑟 , 𝜔, ⅆ𝐼 , 𝑏) = 𝐾0 (𝐹𝑀𝐸𝑃(𝑐𝑟 , 𝜔, 𝑏) − ( ) 𝜂𝑡 (𝜔, 𝑏, 𝑐𝑟 )𝜂𝑣 (𝜔, ⅆ𝐼 )) 𝜔
𝐴 𝑓

where
𝐹𝑀𝐸𝑃 = (4 ⋅ 826(𝑐𝑟 − 9.2)) + (7.97 + 0.253𝑉𝑝 + 9 ⋅ 7 × 10−𝜎 𝑉𝑝2 )

1.5
𝜂𝑡 = 𝜂𝑡𝑎𝑑 − 𝑆𝑣 √( )
𝜔

(1−𝛾)
0.9(1−𝑐𝑟 )(1⋅18−0.825𝜙) ,𝜙≤1
𝜂𝑡𝑎𝑑 = {
0.9(1−[𝛾1−𝑣 ])(10655−0 7𝜙) ,𝜙>1

𝜋𝑁
(0.83) [(8 + 4𝑐𝑟 ) + 1 ⋅ 5(𝑐𝑟 − 1) ( 𝑐 ) 𝑏 3 ]
𝑠𝑣 = 𝑣
[(𝑐𝑟 + 2)𝑏]

1 + 𝑧𝑏2
𝜂𝑣 = 𝜂𝑣𝑏 ( )
1 + 𝑧𝑛2
Zb = (7.72 x 10-2) ω

𝑏 2
−5
(9.428) × 10 𝜔𝑠 ( )
ⅆ𝐼
𝑍𝑛 =
𝐶𝑠

1.067 − 0.038ⅇ 𝜔−5.25 , 𝜔 ≥ 5.25

𝜂𝑣𝑏 = {
0 ⋅ 637 + 0.13𝜔−0.014 𝜔2 + 0.00066𝜔3 , 𝜔 ≤ 525
subject to
DESIGN OPTIMIZATION OF FLAT HEAD CHAMBER OF AN IC ENGINE 11

g1 = P3ω - dI2 ≤ 0
g2 = P4 - (0.8595(1 – cr-0.33)) + Sv ≤ 0
g3 = b – P1 ≤ 0
g4 = P2 – b ≤ 0
g5 = dI + dE – K3b ≤ 0
g6 = K4dI - dE ≤ 0
g7 = dE – K5dI ≤ 0
g8 = cr – 13.2 + (0.045b) ≤ 0
g9C1 = ω – K7C1 ≤ 0
g9C2U = ω – K7C2 ≤ 0
g9C2L = K7C1 – ω ≤ 0
Note: Constraint g9C1 is for case-C1 and constraint g9C2U and g9C2L are for case C2 and put lower and upper
Bound on ω

where Pi’s are parametric function for parametric study of model, which are defined as follows
𝜌𝑄
𝑃0 = = 3688 𝑘𝑃𝑎
𝐴𝑓
𝐿1
𝑃1 = = 83.33 𝑚𝑚
𝐾1 𝑁𝑐

4𝑘2 𝑉
𝑃2 = √ = 76.90 𝑚𝑚
𝜋𝑁𝑐 𝐿2

4𝑉𝐾6 𝐶𝑠
𝑃3 = 9.428 × 10−5 ( ) = 211.322 𝑚𝑚2 𝑠
𝜋𝑁𝐶
3 ⋅ 6 × 106
𝑃4 = = 1.1339(106 )𝑘𝑔−1
𝐾8 𝑄
Design Exploration via
Topology Optimization

Ethan Thompson, Team Lead – Ansys Discovery

© Copyright 2023 ANSYS, Inc.

 Traditional Method of Design and Manufacturing
• Traditional way of manufacture has influenced the
way products are designed

• Experience drives design (a good thing)

• Traditional design made from Boolean operations

such as subtraction, considers the tooling needs and
processes [1] Vehicle upright – manufactured via CNC Milling

• Leads to limitations in design and optimization

benefits

[1] Muzzupappa, Maurizio & Barbieri, Loris & Bruno, Fabio. (2011). Integration of topology optimization tools and knowledge management into
the virtual Product Development Process of automotive components. Int. J. of Product Development. 14. 14 - 33. 10.1504/IJPD.2011.042291.

Best COUB OIC/YouTube

2 ©2023 ANSYS, Inc.

 What is Topology optimization?
Topology optimization is a physics
driven approach to identify what the
optimal material distribution is based
on certain constraints and loads.

3 ©2023 ANSYS, Inc.

 Why do it?

• Improved product performance and design

• Light-weighting
• Reduced material waste and part consolidation (+AM)
• New material properties (+AM)
• Guide product design

4 ©2023 ANSYS, Inc.

 What steps to take?

Back to CAD
Topology Optimization

Easy to use
& Validate
Seamless workflow
FE Analysis

Topology optimization helps you create

new, innovative or improved designs.

Ansys makes it easy to use for every

engineer

5 ©2023 ANSYS, Inc.

 Ansys Topology Optimization Solutions
• Ansys Discovery
- GPU powered topology optimization
- Structural simulation only
- Integrated geometry operations for further operations
• Ansys Mechanical
- Higher fidelity capability
- Advanced options
• Solver options, controls, meshing, responses, etc.
• LS-Dyna
- LS-Opt and LS-TaSC
• Ansys Fluent
- Adjoint Solver

6 ©2023 ANSYS, Inc.

 Ansys Discovery – Upfront Simulation and Geometry Tool

• GPU powered solver

- Robust voxel-based meshing
- “Live” physics
• Structural optimization only
• Integrated geometry modeling
- Facets/STL toolkit
- Reverse engineering capabilities
- Subdivisional modeling
• Connection with Ansys Mechanical

7 ©2023 ANSYS, Inc.

 Journal Bearing - Overview Discovery

Topolgy Optimization: No Meshing

Objective & Constraints Run on GPU
Reverse Engineering
One GUI

Structural setup: Overall < 5 min

BCs & Load Validate & Compare → WB

8 ©2023 ANSYS, Inc.

 Ansys Mechanical

• Thermal optimization
- Thermal compliance
• Structural optimization
- Compliance, Stress, Mass, etc.
• Lattice Optimization
• Optimization Approaches
- SIMP and Level Set
• Topography Optimization (new in 2023R1)
- Fabricated structures

9 ©2023 ANSYS, Inc.

 Overview: Knuckle – Topology Optimization
Level Set SIMP
1 1 3
2

Bushing
Spherical 1
2
3
1 3
1 2

Revolute
Spherical 2

Level Set Based Density based - SIMP

• New, 10-year technology
• Boundary is parameter and • Matured, 30-year technology
moved to optimal shape • Density between 0 and 1 applied on
mesh elements
• Complex load scenarios possible • Densities are calculated for each
• Available in Discovery – powered element
by GPU & Workbench driven by • “Power law” filter is applied, and
CPU most needed elements are kept.

10 ©2023 ANSYS, Inc.

 Overview: Alcoa Bracket – Lattice Optimization

Lattice Optimization
• Easy & fast validation • Solid material is substitute by beam structures
• Automatic generated & depending on FEA calculation
applied variables on initial • Outer shape doesn’t change
geometry • Knockdown factors are a fast approach for
validation, properties of lattices are mapped
onto solid elements

11 ©2023 ANSYS, Inc.

 Ansys Mechanical Results

Knuckle Original TO Alcoa Bracket Original Lattice

Weight [kg] 14.54 8 Weight [g] 859 448
Max. Stress [MPa] 225 205 Max. Stress [MPa] 855 821
Saving [%] 55 Max. Disp. [mm] 1.17 2.15
Saving [%] 52
*Alcoa Bracket Challenge from GrabCAD

12 ©2023 ANSYS, Inc.

 Design Optimization With LS-DYNA
LS-OPT
▪ Size optimization
o shell thickness is designed per part basis
▪ Shape optimization (using mesh morphing tools)
▪ Reliability, robustness, and reliability-based design optimization
▪ Material parameter identification (including digital image correlation)

LS-TaSC
▪ Topology optimization
o Obtain the best material layout for given loading and boundary conditions
o MDO capabilities
▪ Topometry optimization
o shell thickness is designed per element basis
▪ Shape optimization
o a free shape of the outer surface contour is chosen

13 ©2023 ANSYS, Inc.

 General Capabilities – LS-Dyna
▪ Topology and Topometric Optimization

• Optimal material layout (solids/shells)

• Nonlinear LS-DYNA models, large deformations, contact

• Design variables: Element density/thickness

• Local Objective: Maximize stiffness, fundamental freq. Bumper

parts
Crash
box
• Final topology is LS-DYNA input file (or STL file)

Ansys LS-TaSC Other tools

14 ©2023 ANSYS, Inc.

 Overview of the Fluent Adjoint Workflow
• The workflow can be viewed as a
four-step process
1. CFD Run (known process)
2. Calculate the derivatives (gradients)
3. Sensitivity data Mesh Update CFD Run
• Mapping sent back by derivatives
4. Update the shape (Mesh) Sensitivity Adjoint
• Based on the sensitivity data data calculation
• Based on the environment constraints

• This four-step process can be run multiple

times to reach an optimum evolution for
the design…

15 ©2023 ANSYS, Inc.

 Adjoint-Driven Optimization Process CFD Solver

Repeat...
Export geometry
Flow solution
(STL format)
New Geometry
Mesh Local
Morph optimum Adjoint
Solver
x4x5
x
x2 3
x1
x0

Shape Sensitivities

Design Optimal
Tool Constrained
Design Change

16 ©2023 ANSYS, Inc.

 A Simple Example – Pipe U-Bend

17 ©2023 ANSYS, Inc.

 Adjoint Optimization Example – HVAC Duct
• Rear Cabin Automobile HVAC Duct - Minimize Total Pressure Drop

33% reduced pressure drop

• Small geometry changes determined by Adjoint optimization / mesh morphing results

in significant performance improvement!
- This is a result of computing node displacements for specified portion of domain based on the adjoint solution.

18 ©2023 ANSYS, Inc.

 Adjoint Optimization Example – Full Aircraft Lift/Drag
• Airplane – full scale - Maximize Lift/Drag
- Optimize shape of the wing to increase lift to drag
ratio
- Again, just small changes result in a significant
improvement…

55% increased lift

Sensitivities of Normal Optimal Displacement

19 ©2023 ANSYS, Inc.
 Discovery – Correct Tool for Identifying Trends
Ansys Discovery Ansys Flagship Products
- is easy to use - need skilled / trained users
- provides immediate feedback - require calculation time
- helps to find out the rough direction to go - can provide exact numbers
- focuses on non-complex features - are versatile in features

... like a COMPASS which helps ... like a GPS, which shows exactly
- not to get lost - distance and time to destination
- find out which direction to go - path to follow to exact destination

20 ©2023 ANSYS, Inc.

20
 A Topology Optimization Solution for Every Engineer

Advanced
Capabilities

Ansys Flagship

Basic Ansys Discovery

Capabilities
FEA Topology Geometry Validation Advanced
Optimization Analysis

21 ©2023 ANSYS, Inc.

21