8/14/2019 Mop Hong

1/31

Finite Element Simulation of Elevator Car Frame Buffer Crash Using

ANSYS, Altair HyperMesh and RADIOSS

by

Loi Cheng

An Engineering Project Submitted to the Graduate

Faculty of Rensselaer Polytechnic Institute

in Partial Fulfillment of the

Requirements for the degree of

MASTER OF MECHANICAL ENGINEERING

Approved:

_________________________________________Ernesto Gutierrez-Miravete, Project Adviser

Rensselaer Polytechnic Institute

Hartford, CT

May, 2011

8/14/2019 Mop Hong

2/31

ii

Copyright 2011

by

Loi Cheng

All Rights Reserved

8/14/2019 Mop Hong

3/31

iii

CONTENTS

LIST OF TABLES ............................................................................................................ iv

LIST OF FIGURES ........................................................................................................... v

KEYWORDS .................................................................................................................... vi

ACKNOWLEDGMENT ................................................................................................. vii

ABSTRACT ................................................................................................................... viii

1. Introduction .................................................................................................................. 1

1.1 Background ........................................................................................................ 1

1.2 Problem Description........................................................................................... 3

2. Methodology ................................................................................................................ 6

2.1 Geometry ............................................................................................................ 6

2.2 System Properties ............................................................................................... 8

2.3 Static Structural Analysis ................................................................................. 10

2.4 Explicit Dynamic Analysis (without Acceleration) ......................................... 11

2.5 Explicit Dynamic Analysis (with Acceleration) .............................................. 12

3. Results........................................................................................................................ 13

3.1 Static Structural Analysis ................................................................................. 13

3.2 Explicit Dynamic Analysis (without Acceleration) ......................................... 15

3.3 Explicit Dynamic Analysis (with Acceleration) .............................................. 17

4. Conclusion ................................................................................................................. 19

5. References.................................................................................................................. 20

6. Appendix .................................................................................................................... 21

6.1 Calculation of Buffer Material Properties ........................................................ 21

6.2 Expected Dynamic Behavior (without gravity) ............................................... 22

6.3 Expected Dynamic Behavior (with gravity) .................................................... 23

8/14/2019 Mop Hong

4/31

iv

LIST OF TABLES

Table 1 Structural C-Shapes Shape Used for Car Frame Analysis ................................... 8

Table 2 System Properties ................................................................................................. 8

Table 3 Car Frame Material Properties ............................................................................. 9Table 4 Buffer Material Properties .................................................................................... 9

8/14/2019 Mop Hong

5/31

v

LIST OF FIGURES

Figure 1. Elevator System (Courtesy Otis Elevator) (Strakosch & Caporale, 2010) ........ 1

Figure 2. Car Frame (Courtesy ThyssenKrupp Elevator) .................................................. 2

Figure 3. Car Frame in Contact with Buffer (Courtesy www.servingotherdeck.com) ...... 3Figure 4. Polyurethane Buffer (Left) and Oil Buffer (Right) (Courtesy Courtesy ACLA

Werke, China Ningbo Xinda Group Co., Ltd.,) ................................................................. 3

Figure 5. ASME A17.1 Required Static Analysis for Buffer Engagement ....................... 4

Figure 6. Dynamic Analysis without Gravity Acceleration .............................................. 4

Figure 7. Dynamic Analysis with Gravity Acceleration ................................................... 5

Figure 8. Pro/Engineer Model of Car frame and Buffer .................................................... 6

Figure 9. Pro/Engineer Model of Car frame and Buffer, Bottom View ............................ 7

Figure 10. Overall Size of System (in mm) ....................................................................... 7

Figure 11. Structural C-Shape (Courtesy http://www.engineersedge.com) ...................... 8

Figure 12. ANSYS Static Structural Boundary Conditions ............................................. 10

Figure 13. ANSYS Mesh for Static Structural Analysis ................................................. 11

Figure 14. HyperMesh Mesh Detail ................................................................................ 11

Figure 15. Static Analysis Overall Stresses of Car Frame ............................................... 13

Figure 16. Static Analysis Peak Stresses of Car Frame ................................................... 14

Figure 17. Static Analysis Total Displacement of Car Frame ......................................... 14

Figure 18. Dynamic Analysis (no Gravity) Overall Stresses of Car Frame .................... 15

Figure 19. Dynamic Analysis (no Gravity) Peak Stresses of Car Frame ........................ 16

Figure 20. Dynamic Analysis (no Gravity) Displacement and Stress over Impact

Duration ........................................................................................................................... 16

Figure 21. Dynamic Analysis with Gravity, Overall Stresses of Car Frame ................... 17

Figure 22. Dynamic Analysis with Gravity, Peak Stresses of Car Frame ....................... 18

Figure 23. Dynamic Analysis with Gravity, Displacement and Stress Over Duration of

Impact .............................................................................................................................. 18

8/14/2019 Mop Hong

6/31

vi

KEYWORDS

ANSYS

HyperWorks

HyperMeshRADIOSS

FEA

Finite Element Analysis

Static Structural Analysis

Explicit Dynamic Analysis

Buffer Crash

8/14/2019 Mop Hong

7/31

vii

ACKNOWLEDGMENT

I would like to thank Professor Ernesto Gutierrez-Miravete for his help on this project.

I would also like to thank United Technologies for their financial support of my

education.

8/14/2019 Mop Hong

8/31

viii

ABSTRACT

The purpose of the project is to study the behavior of impact loading on an

elevator car frame using static structural analysis and explicit dynamic analysis under

different assumptions of boundary conditions.Using a representative elevator car frame, a static analysis of buffer impact was

performed using ANSYS Workbench 12.1, where the maximum stresses and deflection

were calculated. With the same car frame, two dynamic analyses were performed using

Altair HyperMesh and RADIOSS, where the stresses and deflections over the duration

of the impact were calculated.

The stresses of the static and dynamic analysis were very similar, which indicates

that both the static and dynamic methods are consistent in calculating stresses of the car

frame under buffer impact. However, the dynamic analysis allows for a more in depth

study of the behavior of the system over the duration of impact, whereas the static

method onlyprovides a single snapshot of the systems behavior.

8/14/2019 Mop Hong

9/31

1

1. Introduction1.1 Background

In a traction elevator system, the car is suspended by steel ropes, which are

wrapped around the machines drive sheave at the top of the hoistway, as shown in

Figure 1 (1). Electrical power is supplied to the machine to generate sheave rotation

and vertical movement of the car. The controller regulates the power supplied to the

machine to precisely bring the car from one landing to another. The machine is

equipped with brakes to stop the car once its reached a landing.

Figure 1. Elevator System (Courtesy Otis Elevator) (Strakosch & Caporale, 2010)

8/14/2019 Mop Hong

10/31

2

Multiple mechanisms are in place to safely stop the car in emergency situations.

An overspeed governor located at the top of the hoistway is connected to a set of safeties

in the car by a wire rope. The pulley in the governor rotates a speed proportional to the

velocity of the car. If the car moves beyond a certain speed, the governor activates the

safeties, which stops the car.

If the safety cannot stop the car before it strikes the pit, a buffer sits in the

hoistway pit and reduces the impact of the car. Depending on the rated speed of the car,

the buffer can be made of polyurethane, spring, or an oil and piston rod assembly (see

Figure 4). It is designed to reduce the deceleration of the impact to not cause injury to

passengers or damage the car frame and cab.

The car frame is the structural part of the car and holds the platform and the cab

that carries the passengers, as shown inFigure 2 (2). The platform is directly connected

to the plank of the car frame. The plank is connected to the vertical stiles, which are

connected the roller guides and to the crosshead. The hoisting ropes are attached to the

crosshead. Traditionally, car frame members are made up of structural steel beams. A

strike plate is attached to the bottom of the plank. As shown inFigure 3,this strike plate

is located directly above the buffer in the pit. The buffer and strike plate limits the

damage to the car frame when it collides into the pit.

Figure 2. Car Frame (Courtesy ThyssenKrupp Elevator)

8/14/2019 Mop Hong

11/31

3

Figure 3. Car Frame in Contact with Buffer (Courtesywww.servingotherdeck.com)

Figure 4. Polyurethane Buffer (Left) and Oil Buffer (Right) (Courtesy Courtesy ACLA Werke,

China Ningbo Xinda Group Co., Ltd.,)

1.2 Problem DescriptionTo study the effects of a buffer crash load on a car frame, multiple finite element

simulations are performed with different methodologies. Each methodology uses a

different assumption on loading and boundary conditions, which leads to different

results.

1.2.1 Static Structural AnalysisThe ASME A17.1 Elevator Safety Code requires a static structural analysis of the

car frame under buffer loading. To account for impact, the normal static loads are

doubled, as shown inFigure 5 (3). The analysis is the equivalent to a having a car with

twice its normal weight resting on the buffer, which is fixed to the ground. This

simplification allows the use of hand calculations to solve the problem.

http://www.servingotherdeck.com/http://www.servingotherdeck.com/http://www.servingotherdeck.com/http://www.servingotherdeck.com/

8/14/2019 Mop Hong

12/31

4

Figure 5. ASME A17.1 Required Static Analysis for Buffer Engagement

1.2.2 Explicit Dynamic Analysis (without acceleration)The explicit dynamic approach is not required by the Elevator Safety Code. In this

approach, an initial velocity is applied to the mass of the fully loaded car. As shown in

Figure 6,the car moves downwards at a constant speed until it strikes the buffer that is

fixed to the ground. Effects of gravity are ignored for this analysis. This approach

assumes that the machine is capable of controlling the system to a certain degree during

the impact, such that the car does not accelerate due to gravity.

Figure 6. Dynamic Analysis without Gravity Acceleration

2x Weight of Fully

Loaded Car CAR

BUFFER

Mass of Fully

Loaded Car CAR

BUFFER

Initial Velocity

8/14/2019 Mop Hong

13/31

5

1.2.3 Explicit Dynamic Analysis (with acceleration)In this final method, the mass of a fully loaded car moves downwards at an initial

velocity, and is also accelerating due to gravity, as shown in Figure 7. This approach

assumes that as the car frame comes in contact with the buffer, the hoisting ropes

become completely slack. Without any rope tension to counteract gravity loads, the car

frame is subject to gravity acceleration.

Figure 7. Dynamic Analysis with Gravity Acceleration

Mass of Fully

Loaded Car CAR

BUFFER

Initial Velocit

Gravity

8/14/2019 Mop Hong

14/31

6

2. Methodology2.1 Geometry

Using Pro/Engineer, a simplified car frame model is created. The car frame used

has a size that is typical of a 3500lb duty system and is made of imperial structural steel

shapes. Figure 8 shows a general view of the car frame. The uprights are made of

structural C-Shape C6x8.2. The crosshead and plank are made of C8x11.5. The

platform stringers are C3x4.1. The shape of the C-Shapes used and their sizes are shown

in Figure 11 and Table 1. The first number of the designation is the depth d of the

section, and the second number is the weight of the shape, in pounds per ft. For

example, a C6x8.2 has a cross section 6 inches deep and weights 8.2 lbs/ft. The

platform is 6mm thick steel and the strike plate is 25mm thick steel. The buffer issimplified as a cylinder with 200mm diameter and 300mm height.

Figure 8. Pro/Engineer Model of Car frame and Buffer

UPRIGHT C6X8.2

CROSSHEAD C6X8.2

PLATFORM 8mm

8/14/2019 Mop Hong

15/31

7

Figure 9. Pro/Engineer Model of Car frame and Buffer, Bottom View

Figure 10. Overall Size of System (in mm)

PLANK C6X8.2STRINGERS C3X4.1

BRACE 6mm

8/14/2019 Mop Hong

16/31

8

Figure 11. Structural C-Shape (Courtesyhttp://www.engineersedge.com)

Table 1 Structural C-Shapes Shape Used for Car Frame Analysis

2.2 System PropertiesThe car frame is designed for a net moving mass of 3175kg, of which 1588kg

(3500lb) is the duty load, 579kg is the car frame mass, and 1008kg is the cab. The

operating speed of the car is 1m/s. Table 2 summarizes the system properties used for

all the analysis. The car frame is made entirely of steel, andTable 3 lists the material

properties that are used.

Table 2 System Properties

Speed 1 m/s 200 ft/min

Duty Load 1588 kg 3500 lb

Mass of Car Frame 579 kg 1276 lb

Mass of Cab 1008 kg 2222 lb

Mass of Empty Car 1587 kg 3499 lb

Mass of Fully Loaded Car 3175 kg 7000 lb

Area DepthA d

Thickness Width Thickness

tw (in) bf (in) tf (in)

Crosshead,

Plank

C8x11.5 3.38 8.00 0.22 2.26 0.39

Upright C6x8.2 2.40 6.00 0.20 1.92 0.34

Stringers C3x4.1 1.21 3.00 0.17 1.41 0.27

Where used

(in 2) (in)

Designation Web Flange

http://www.engineersedge.com/http://www.engineersedge.com/http://www.engineersedge.com/http://www.engineersedge.com/

8/14/2019 Mop Hong

17/31

9

Table 3 Car Frame Material Properties

Car Frame

Material Steel

Density 7850 kg/m^3

Elastic Modulus 200 GPaPoisson's Ratio 0.29

Material properties of the buffer were calculated such that it the colliding car

would decelerate at an average of 9.8m/s. This rate is the maximum that is allowed by

the A17.1 Safety Code (3). Table 4 summarizes the properties used. Section 6.1

describes how these values were calculated.

Table 4 Buffer Material Properties

Buffer

Material Spring

Spring Constant 1221 N/mm

Equivalent Elastic Modulus 11.43 MPa

8/14/2019 Mop Hong

18/31

10

2.3 Static Structural AnalysisThe static structural analysis was done with ANSYS Workbench (4). The model

that was created in Pro/Engineer was imported into ANSYS for preprocessing. For

simplicity, all components of the car frame are bonded together. Since this is a static

structural analysis, the strike plate is also bonded to the buffer. Figure 13 shows the

mesh used for the analysis. Solid mesh was used for the structural C channel parts and

the buffer. For flat elements including the platform, braces and strike plate, a solid shell

mesh is used. The maximum mesh size is set to 30mm.

Figure 12. ANSYS Static Structural Boundary Conditions

An acceleration of 2g is applied to the masses in the system. To account for the

items that are not modeled, two times the normal duty and cab weights are applied as a

load distributed over the top surface of the platform. A fixed constraint is applied at the

bottom surface of the buffer. Figure 12 shows the boundary conditions applied to the

system.

8/14/2019 Mop Hong

19/31

11

Figure 13. ANSYS Mesh for Static Structural Analysis

2.4 Explicit Dynamic Analysis (without Acceleration)To perform an explicit dynamic analysis, the Pro/Engineer model is imported into

HyperMesh for preprocessing (5). Figure 14 shows the meshed imported geometry in

HyperMesh. The car frame is imported as a single solid entity and the buffer is imported

as a separate entity, each of them are meshed separately. All the components are mesh

using tetrahedral elements, with a maximum size of 30mm.

Figure 14. HyperMesh Mesh Detail

8/14/2019 Mop Hong

20/31

12

To account for cab and duty load, which are not modeled, point masses are

created on the nodes on the top surface of the platform. The combined masses are

equally distributed among these nodes. An initial velocity is applied on the all nodes

that make up the car frames. A fixed constraint is applied on the nodes that make up the

bottom surface of the buffer, as shown in Figure 14. There is an initial 1mm gap

between the car frame and the buffer. The expected dynamic behavior of the system is

calculated in Section6.2

2.5 Explicit Dynamic Analysis (with Acceleration)To setup the analysis with acceleration, the same model and mesh from Section2.4

is used. Additional 1g acceleration is added to all the nodes of the car frame to account

for gravity. All other loads and boundary conditions are identical to Section2.4. The

expected dynamic behavior of the system is calculated in Section6.3.

RADIOSS was used to solve the HyperMesh models. The program calculates

stresses and deflections at predetermined nodal stability time steps that would minimize

errors (6). The time step is related to the nodal mass M and nodal stiffness K of the

system by the following equation:

K

Mt

*2

The models use for this project required 0.028 m/s time increments. The

simulation was set to run for 400ms. Stresses and deflections were recorded every

0.5ms to an output file.

8/14/2019 Mop Hong

21/31

13

3. Results3.1 Static Structural Analysis

As shown in Figure 15, the static structural analysis shows that the high stress

areas are near the strike plate. Peak stresses are concentrated at the platform stringers in

the region between the two planks. In Figure 16, stress shown in this area is

approximately 105-163MPa. Stresses decrease to about 37-75MPa as we move away

from this region. The uprights and crosshead has very little stress at less than 12Mpa.

The A17.1 allowable stress for ASTM A36 steel is 190MPa (3). The results indicate that

the requirement is met.

Figure 17 shows the displacement of the car frame. The maximum buffer

compression is 52mm, which consistent with the value predicted in section 2.2. Thedeformation of the car frame is very low relative to the compression of the buffer.

Figure 15. Static Analysis Overall Stresses of Car Frame

8/14/2019 Mop Hong

22/31

14

Figure 16. Static Analysis Peak Stresses of Car Frame

Figure 17. Static Analysis Total Displacement of Car Frame

8/14/2019 Mop Hong

23/31

15

3.2 Explicit Dynamic Analysis (without Acceleration)As shown inFigure 18 andFigure 19,the areas of maximum stress are located on

the stringers in the same location found in section3.1. The maximum stress of 80MPa

occurs at 75ms, as shown in Figure 20, and the maximum deflection is about 50mm.While calculated deflection is consistent with the static analysis, the maximum stress is

about half of the value calculated by static analysis in section3.1. The results suggest

that during a buffer impact, dynamic forces account for about half of the total stresses in

the system, while static gravity forces (absent in this analysis) account for the other half

of the stresses. The static method, which used 2x static loads, would therefore produce

stresses consistent with a dynamic method where gravity is included, as demonstrated in

Section3.3.

Figure 18. Dynamic Analysis (no Gravity) Overall Stresses of Car Frame

The plot in Figure 20 shows the stresses and displacement over the duration of

impact for the stringer elements between the planks. The plot indicates an oscillatory

response in the stress of the car frame, which resembles a system of two springs in series

where one spring is much stiffer than the other. The small oscillation is governed by the

natural frequency of the car frame, which is much stiffer than the buffer. The

oscillations are somewhat consistent, with a period of about 10ms and peak to peak

8/14/2019 Mop Hong

24/31

16

amplitude of about 10MPa. The buffer acts as the softer spring and drive the major

oscillation from 0ms to 150ms. After 150ms the car has rebounded off the buffer and

the stresses have dropped to 0-10MPa, which comes from the vibration of the car frame.

Because of the absence of gravity in the calculation, the car frame will continue to move

upward at a constant velocity. Therefore, the results after 150ms are not considered to

be realistic.

Figure 19. Dynamic Analysis (no Gravity) Peak Stresses of Car Frame

Figure 20. Dynamic Analysis (no Gravity) Displacement and Stress over Impact Duration

Displacement

(mm)

Stress

(MPa)

Time (ms) Time (ms)

8/14/2019 Mop Hong

25/31

17

3.3 Explicit Dynamic Analysis (with Acceleration)Similar to the static structural analysis, this explicit dynamic analysis shows high

stresses in the same region of the stringers between the planks, as shown in Figure 21

and Figure 22. Figure 23 tracks the stress and displacement of the two representativeelements in this region over 400ms. Element E993987 has the highest stresses in the

system while its nearby element E1049750 has slightly lower stresses. The results

indicate that stresses reach a peak of 155MPa at about 90ms, where displacement is a

maximum of approximately 72mm. The displacement is greater than the value

calculated by static analysis, but is close to the 80mm predicted by the differential

equations in Section6.3,below.

Figure 21. Dynamic Analysis with Gravity, Overall Stresses of Car Frame

Similar to section3.2,the plot inFigure 23 indicates an oscillatory response in the

stress of the car frame that closely resembles a system of two springs in series. Again,

the oscillations are consistent with a period of about 10ms and peak to peak amplitude of

about 10MPa. The buffer drives the major oscillation from 0ms to 175ms. From 175ms

to 380ms, the car has rebounded off the buffer and the stresses have dropped to 0-

10MPa, which comes from the vibration of the car frame. At 380ms, stresses begin to

8/14/2019 Mop Hong

26/31

18

rise as the car frame comes into contact with the buffer again. The stresses are expected

to behave in the same pattern shown from 0-400ms, since there is no permanent

deformation or any other loss of energy in this ideal system. In an actual buffer impact

scenario, damping forces in the system would eventually bring the oscillating system to

a stop.

Figure 22. Dynamic Analysis with Gravity, Peak Stresses of Car Frame

Figure 23. Dynamic Analysis with Gravity, Displacement and Stress Over Duration of Impact

Displacement

(mm)

Stres

(MPa

Time (ms) Time (ms)

8/14/2019 Mop Hong

27/31

19

4. ConclusionThe stresses predicted by the static method were closest to the dynamic method

in which gravity was included. Both methods yield peak stresses in the range of

160MPa. However, the dynamic method yielded larger overall displacement than thestatic method. The results suggest that the static method is valid in predicting stresses

but may not be accurate in predicting displacements for car frame buffer impact.

The dynamic method without gravity (Section 3.2)yielded similar displacement

as the static analysis, but the maximum stress was half of the static method. The lack of

gravity forces in the system most likely under predicted stresses in the system.

Compared to the dynamic method with gravity, this method without gravity does not

seem to accurately predict car frame stresses and deflection under buffer impact.

To further validate the analysis presented, an actual test would need to be

performed. The car frame and buffer should closely match the properties used for the

analysis, and the masses and impact velocity should be the same. Stresses recorded with

strain gages located in the region of interest can be compared to the results in this report

to show which method is the most valid. The actual deflection of the buffer can also be

used to assess validity of analysis.

8/14/2019 Mop Hong

28/31

20

5. References1. Strakosch, G. R. and Caporale, R.S.The Vertical Transportation Handbook. New

York : John Wiley and Sons, 2010.

2. ThyssenKrupp.Vertical Transportation: A Glossary. [Online] [Cited: April 8, 2011.]http://www.thyssenkruppelevator.com/glossary.html.

3. American Society of Mechanical Engineers.A17.1 Safety Code for Elevators and

Escalators.New York : ASME, 2007.

4. ANSYS, Inc. ANSYS Workbench User's Guide. [Online] April 5, 2011.

http://www1.ansys.com/customer/content/documentation/121/wb2_help.pdf.

5. Altair Engineering.Hypermesh User's Guide. Troy, MI : Altair Engineering, 2009.

6..RADIOSS 10.0 User's Guide. Troy, MI : Altair Engineering, 2009.

7. Frontal Crah and Airbag Deployment Simulation Using RADIOSS. Barman,

Amlanjyooti, Kodwani, Ravi and Siddegowda, Nagaraj. 2008, Altair HyperWorks

Technology Conference.

8. Crash Simulation of an F1 Racing Car Front Impact Structure. Heimbs, S., et al., et

al.2009, 7th European LS-DYNA Conference, Salzburg, Austria.

8/14/2019 Mop Hong

29/31

21

6. Appendix6.1 Calculation of Buffer Material Properties

With an initial velocity of 1m/s, the distance the car frame would travel under 1g

deceleration can be calculated, as according to the method outlined in the A17.1 Safety

Code (American Society of Mechanical Engineers, 2007):

smV /1 (1)

mmg

VS 51

2

2

(2)

Therefore, in order for the car to decelerate at 1g, it will compress the buffer 51mm.

By equating the initial kinetic energy of the car frame with the potential energy of the

buffer at maximum compression, the spring constant can be calculatedkgm 3175 (3)

JmvKE 15882

1 2 (4)

PEKE (5)

2*2

1SkPE (6)

mmNS

PE

k /1221

*22

(7)

This spring constant also yields the same buffer compression under 2x static load

mmk

gmx 51

**2 (8)

The spring constant k is used to calculate the appropriate elastic modulus for the cylinder

geometry used for the buffer.

mmL 294

(9)

2^314204

)200(* 2

mmmm

A

(10)

MPaA

kLE 43.11

* (11)

8/14/2019 Mop Hong

30/31

22

6.2 Expected Dynamic Behavior (without gravity)The following calculations show the amount of buffer compression y(t), the velocity

yp(t) and acceleration ypp(t) of the car frame over time. They are based on solving the

following differential equation.

yKdt

ydmass **

2

2

(12)

8/14/2019 Mop Hong

31/31

6.3 Expected Dynamic Behavior (with gravity)The following calculations show the amount of buffer compression y(t), the velocity

yp(t) and acceleration ypp(t) of the car frame over time. The same method from6.2 is

used