MATTIAS ELFGREN THOMAS STEINER
Development and
Validation of Adaptive Ride Down Safety Systems
MASTER OF SCIENCE PROGRAMME Mechanical Engineering
Luleå University of Technology
Department of Applied Physics and Mechanical Engineering Division of Computer Aided Design
2006:167 CIV • ISSN: 1402 - 1617 • ISRN: LTU - EX - - 06/167 - - SE
Development and Validation of
Adaptive Ride Down Safety Systems
Mattias Elfgren Thomas Steiner
Luleå University of Technology Division of Computer Aided Design
2nd May 2006
A BSTRACT
The work done in this master thesis can de divided into several categories, but the main idea was to optimize the principle of energy absorption that was developed in the Sirius course [1]. If possible, the principle would then be implemented in three applications in the interior of a personal vehicle. These applications were the steering column - which would get the main focus, the ride down seats and the shoulder belt load limiter.
A drop test rig, which tried to resemble a crash situation, was built in order to test different materials and cutting geometries. By input from VMCC, the mass of the weight and the height from which it was dropped could be calculated.
From previous experience it was known that a material that gave short chips gave large fluctuations in force, so a search was carried out to find materials that would give long chips. Also, the geometry of the cutting insert was thought to play a role in the stability of the resulting forces. Therefore, an investigation of cutting geometries was carried out.
This resulted in five different materials (aluminium, copper, brass, bronze and PTFE) and six cutting geometries that were to be tested.
A testing matrix was developed so that each cutting geometry would be tested three times on each material. This first test series lead to the conclusion that aluminium was the most suitable material, but it was also realized that the test rig was not stiff enough to decide the best geometry of the cutting insert. This understanding lead to a redesign of the test rig.
In the second test setup two cutting inserts were put with cutting edges pointing to- wards each other, and the aluminium test plates fell down between. This way the perpen- dicular forces were cancelled out and the cutting process got much more stable. Again, all cutting geometries were tested, but only with the aluminium test plates. After some statistical analysis, the choice of most optimal cutting geometry could be decided. With both the material and cutting geometry chosen, other parameters could be examined.
Different cutting depths were tested to see if a dependency between depth and force could be found. An almost linear relationship was found when the cutting depth was plotted with the cutting force, but a somewhat different behaviour was shown when the specific cutting force was plotted next to the force. This means that if both cutting inserts has nearly the same cutting depth the force is linear to the cutting depth, but if
iii
obviously was lower.
To summarize the energy absorbing principle, the tests performed in this work show that the stability of the force level is strongly depending of the material. The cutting geometry has only a small influence, while the temperature and the velocity has little to none.
Design proposals of the steering column, ride down seats and shoulder load limiter are presented at the end of this rapport, specially designed to take as little space as possible and using the same principle as in test setup two.
iv
P REFACE
Safety has always been a keystone in Volvo Car Corporation, and now, when fuel efficiency and environmental issues are more important then ever, the need to make smaller cars is essential. In order to offer the same protection in a small car as in a larger, new ways to keep the occupant safe must be developed. One way to do this is to let structures in the interior travel forward during a frontal crash, while consuming kinetic energy. If this energy can be absorbed in a controlled and reliable fashion, as well as be allowed to vary from crash to crash, from people to people, many lives can be saved.
This Master Thesis has been carried out at Luleå University of Technology, in collab- oration with Volvo Monitoring and Concept Center.
We would like to give a special thank to the following persons:
Our supervisor and examinator at LTU, PhD, Peter Åström, who always have supported us with good feedback and ideas.
Our supervisor at VMCC, PhD, Kolita Mendis, for believing in our ideas and showing a great deal of sympathy.
Tore Silver at the Division of Manufacturing Systems Engineering at LTU for his kind- ness and help throughout manufacturing of the test rig.
And a big thanks to the following people, who all have contributed and helped us carry through this project:
The crew at Centralverkstaden
Peter Norman and Esa Vuorinen, Division of Manufacturing Systems Engineering Magnus Löfstrand, PhD Student, Division of Computer Aided Design
Luleå 2006-05-02
... ...
Mattias Elfgren Thomas Steiner
v
C ONTENTS
Chapter 1: Introduction 1
1.1 Background . . . . 1
1.2 Volvo Monitoring and Concept Center . . . . 2
1.3 Purpose . . . . 2
1.4 Objectives . . . . 3
1.5 Principle of Energy Absorption . . . . 3
1.6 Delimitations . . . . 4
Chapter 2: Experimental Equipment 7 2.1 The Kistler Plate . . . . 7
2.2 The Test Rig . . . . 7
2.3 Test Rig Specifications and Calculations . . . . 9
2.4 Cutting Insert Geometry . . . . 10
2.5 Materials to be Tested . . . . 12
Chapter 3: Test Setup 1 15 3.1 Test Series 1 . . . . 15
3.2 Results . . . . 16
3.3 Conclusions . . . . 18
3.4 Measurement Stability . . . . 19
Chapter 4: Test Setup 2 21 4.1 Test Series 2 . . . . 21
4.2 Statistical Analysis of the Cutting Geometries . . . . 24
4.3 Other Parameters . . . . 24
Chapter 5: Results 31 5.1 The Correlation between Cutting Depth and Cutting Force . . . . 31
Chapter 6: Design Propositions 33 6.1 Steering Column . . . . 33
6.2 Ride Down Seats . . . . 39
6.3 Shoulder Belt Load Limiter . . . . 42
Appendix C:Push Rod Concept 59
Appendix D:Test Matrix for First Test Setup 67
Appendix E:Calculations of Mass in the Second Test Setup 69 Appendix F: Tests Performed to Decide Optimal Geometry 71 Appendix G:Tests Performed to Decide Optimal Material (1) 75
G.1 No Chip Breaking Groove and Neutral Cutting
Angle . . . . 75 Appendix H:Tests Performed to Decide Optimal Material (2) 79
H.1 No Chip Breaking Groove and Positive Cutting
Angle . . . . 79 Appendix I: Tests Performed to Decide Optimal Material (3) 83
I.1 No Chip Breaking Groove and Negative Cutting
Angle . . . . 83 Appendix J: Tests Performed to Decide Optimal Material (4) 87 J.1 Chip Breaking Groove and Neutral Cutting Angle . . . . 87 Appendix K:Tests Performed to Decide Optimal Material (5) 91 K.1 Chip Breaking Groove and Positive Cutting Angle . . . . 91 Appendix L: Tests Performed to Decide Optimal Material (6) 93 L.1 Chip Breaking Groove and Negative Cutting Angle . . . . 93 Appendix M:Graphs of Calculated Cutting Depth and Distance for
Aluminum 97
viii
C HAPTER 1 Introduction
1.1 Background
There exist many different possibilities to passively hinder or lower severe injuries in frontal car crashes. A combination of soft, ductile materials and even surfaces with no or a minimum of outward pointing details, is a good way of protecting the occupants in low to mid velocity crashes. In the new Volvo S40 for example, the ignition key is moved from the steering column to the instrument panel. This was to avoid knee injuries, as there was a risk of the knees hitting the sharp key and the hard lock cylinder behind it in a frontal crash. SAAB [2] has used a similar tactic for decades. By placing the ignition key between the two front seats they efficiently remove the problem.
Volvo standardized the 3-point shoulder belt for the front seats back in 1959 [3], and since that much has happened. Today Volvo uses a device that automatically adjusts the shoulder belt to suit the height of the person using it. In the event of a crash, the shoulder belt is pretensioned in the blink of an eye and resists forces up to a given load, and then let go. This way the shoulder belt will absorb some of the energy from the crash, but on the same time not decelerate the person to fast, which might result in chest and/or neck injuries. In the Volvo cars of today, this load can be varied between two predetermined values and only works for a very short period of time during the crash.
This feature is called the shoulder belt load limiter.
For years it has been a legal demand that the steering column must not be pushed back inwards the driver apartment in a crash. But if the steering wheel would go in the opposite direction when the body of the driver is thrown into it, it would also work as a deformation zone in the interior of the vehicle. In the new Volvo S40/V50 the steering column is allowed to translate to a maximum of 115 mm. This design has two levels of energy absorption, depending if the driver is using a seat belt or not, and the level must
1
needs to have a more developed safety system. In the small concept car Volvo introduced what they call "Ride Down Seats", which basically works as another deformation zone in the interior. In the event of a crash, the front seats translate forward while absorbing energy and thereby lowering the deceleration of the occupants.
The task to design this system was given to Luleå University of Technology 2003/2004 as a Sirius project [4]. The Sirius team developed a method of using hydraulic dampers and CES-valves [5]. In 2004/2005, another Sirius project started as a continuation of the first, and this time a safety system for the steering column was to be designed. A system with cutting inserts using mechanical cutting as an energy absorbing principle was developed, but the levels of force generated form the tests were not stable enough.
The design was somewhat big as well and is therefore in need of a redesign.
1.2 Volvo Monitoring and Concept Center
Volvo Monitoring and Concept Center (VMCC) is located in Camarillo outside Los An- geles since 1986. The purpose of VMCC is to be an idea reservoir for the parent company Volvo Car Corporation (VCC) and as well as for the entire Ford family. The goal for VMCC is to keep up with new trends and technologies and integrate them in new con- cepts and designs. The work done at VMCC helps VCC to grow in the competitive auto market of today. The center has been involved in many key Volvo projects, including the XC90, S80, S60, the Environmental Concept Car (ECC), and Safety Concept Car (SCC). And lately, when new environmental issues have been in focus, they designed the 3-Concept Car (3CC) that has won gold medals in environmental performance and the prestigious award in best prototype design.
1.3 Purpose
Volvo is interested to get into the market for smaller cars, but to keep the driver safe in case of a crash, Volvo needs new safety systems. In order to make the The Ride Down Seats useful, the steering column needs to work together with it, i.e. translate forward 200 mm.
By increasing the translation length of the steering column from today’s 115 mm to 200 mm, injuries to the driver can be reduced. The properties of such a steering column should be that the system must be designed to fit in the space behind the instrument panel, and, most importantly is that the absorbed levels of energy are stable during the
1.4. Objectives 3 crash. It is also desirably that the energy absorption is adaptive.
1.4 Objectives
The aims for the project are to refine the principle of absorbing energy by cutting of material, as well as investigate if the principle will be suitable in different parts of the interior, e.g. the shoulder belt load limiter, the Ride Down Seat and the steering column.
The primary goal is to adjust and optimize the principle of energy absorbing and improve the design for the steering column. Design proposals will be presented for each application.
To find out if this type of principle works in a car, it is needed to show that the levels of force depend on the cutting depth, and that the variation of absorbe energy will not fluctuate to much. In order to find the optimal conditions, the different variables which have a strong influence on the result must be found. The identified variables are listed below.
1. Energy absorbing material 2. Geometry of the cutting insert 3. Cutting depth
4. Cutting width 5. Mass (of body) 6. Velocity (of body) 7. Temperature
The work will be concentrated on finding the right material and the most suitable cutting insert geometry that together will give small fluctuations in forces. The other variables are set to a constant value in the beginning.
1.5 Principle of Energy Absorption
The principle used is much like the parting principle that can be done in turning oper- ations. The interesting part of this principle is the cutting force. The cutting force can be controlled if all parameters are known and set. The force is dependent on param- eters such as specific cutting force, cutting area and cutting velocity. The formula for calculating the force is shown in equation 1.1.
F = Kc(v) × AD (1.1)
4. v = Cutting velocity
As shown in equation 1.1, the specific cutting force is dependent of cutting velocity, but for high velocities the specific cutting force can be assumed constant. Figure 1.1 illustrates the behavior of specific cutting force as function of cutting velocity for a steel material. In the range 125 m/min up to 300 m/min in figure 1.1, the specific cutting
Figure 1.1: Specific cutting force as function of cutting velocity for a steel material.
force can be assumed constant. This behaviour explains why the principle is interesting for a safety system.
1.6 Delimitations
This thesis has its focus on the energy absorbing principle, and therefore no concerns have been taken into the following aspects:
• The steering column can move freely behind the instrument panel and is not affected by any other forces then from the cutting process.
• The part of the steering column that is attached to the chassis can be attached at the given points, see chapter 6.1.
1.6. Delimitations 5
• All cutting inserts can be adjusted by an external controller that decide the levels of force required the moment before the column starts to move and possibly even once after.
C HAPTER 2 Experimental Equipment
2.1 The Kistler Plate
To measure how much the force varies during the cutting process, a Kistler measur- ing plate [6], see Appendix A, have been used. The Kistler plate is a 3-Component Dynamometer, see figure 2.1. It is made for three main applications:
1. General 3-component force measurement (dynamic and quasistatic)
2. Cutting force measurements for optimization of the manufacturing process (tem- porary measurements)
3. Cutting force measurements (turning, milling, grinding) for training purposes It has the capacity of measuring forces in x, y, and z directions. The range in x and y is from -5 kN to 5kN, and -5 kN to 10 kN in z, the z-direction will be used in the tests.
The natural frequency for measurements in x and y is 2 kHz, and 3.5 kHz in z.
2.2 The Test Rig
To verify the idea of having an insert cutting through a material, a test setup had to be designed. While looking into a frontal car crash, roughly this is what happens; Due to a rapid and violent deceleration the driver is thrown forward, the belt stretches up and the airbag blows. When the driver hits the airbag, the body brings a load to the steering column, the column collapses and it absorbs energy. The energy absorbed decelerates the body. The test rig was designed to represent a real car crash situation, see figure 2.2.
A weight of about the same mass as an average torso is dropped from a predetermined 7
Figure 2.1: Kistler, 3-Component Dynamometer
height and hits the energy absorbing part, representing the steering wheel and steering column. By the energy developed from the falling mass, the energy absorbing material starts being cut while the force is measured at each 0.5 ms by the Kistler plate.
Figure 2.2: The new design of the test rig with separate guiders for the falling weight and the energy absorbing part.
2.3. Test Rig Specifications and Calculations 9
Equation f(x) [N] x1 [s] x2 [s] F(x) [au] Unit
1 f(x) = 254,68x - 8638,2 0,032 0,053 45,900 Au
2 f(x) = -200,27x + 16199 0,053 0,069 63,720 Au
3 f(x) = 186,01x - 10222 0,069 0,081 44,748 Au
4 f(x) = -169,13x + 18681 0,081 0,113 72,815 Au
Summary: Area below graph 227,185 Kg*M*s−1
Total mass: 52 Kg
Mass of falling weight 40 Kg
Delta V = X if the mass M = 52 kg 4,368 M*s−1
Table 1: Simplified integration of airbag force graph and a prediction of acting parameters on the system.
2.3 Test Rig Specifications and Calculations
To get an understanding of what force that is loading the steering column, VMCC pro- vided a graph from a crash simulation, see figure 2.3. From this graph the mass and velocity can be derived by integration. By simplifying the graph with four straight lines it is easy to make the integration, see figure 2.3. The result of this calculation is shown in table 1, and as can be seen the mass of the falling weight should be 40 kg and have a velocity of 4.4 m/s. The impact velocity can be calculated by the energy equations 2.1- 2.2. But when the weight (m1) hits the lower part (m2), the mass changes (m3), thus equation 2.4 gives the new velocity (v3). There will be some losses in energy due to friction when the weight falls down, as well as when the two masses collide, which will lower v3. From table 1, the velocity (v3) at the chosen total mass (m3) was calculated to 4.4 m/s. To achieve this the moment after the impact, the starting velocity (v1) must be calculated. Equation 2.1 says if the mass is dropped from 2.1 m, the velocity without any losses in energy or friction would be 4.7 m/s. This is assumed to lead to a velocity of about 4.4 m/s with all losses included.
mgh = mv2
2 (2.1)
v =p
2gh (2.2)
m1v1+ m2v1 = m3v3 (2.3)
v3 = m1v1+ m2v1
m3 (2.4)
(a) The graph illustrates how the force is acting on a steering column during a crash situation.
(b) Simplified airbag-force graph used for integration. From the left the lines are enumerated one to four. Each one of them has an own equation called f(x), after integration it is called F(x), see table one.
Figure 2.3: The force a human body experiance when thrown in to an airbag.
2.4 Cutting Insert Geometry
In cutting operations, the geometry of the cutting insert has a certain influence on the stability of the force. The first cutting insert made in the Sirius course for this application was made to be simple and manage the load it was being exposed for, see figure 2.4.
2.4. Cutting Insert Geometry 11
Figure 2.4: Left: First design of the cutting insert made in the Sirius course. Right: Resulting grooves from cutting performed with the insert shown in the left figure.
Cutting insert Cutting angle Cutting angle Chip breaking geometry
1 0o Neutral No
2 0o Neutral Yes
3 4o Positive No
4 4o Positive Yes
5 4o Negative No
6 4o Negative Yes
Table 2: Properties of cutting insert from 1-6
Problem with this geometry was discovered while it was being tested. As can be seen in figure 2.4, the grooves are very rough. The reason for the roughness is thought to be caused by three things; the energy absorbing material, the cutting insert did not have loosing angles on all sides as well as it did not have any chip breaking geometry. If the insert is without loosing angle on the sides, material might get sticky and influence force stability. A desired geometry of the cutting insert is one that allows the chip to roll off continuously while it is being cut.
To find out which geometry that reduces the force fluctuations the most, six cutting inserts have been developed and tested. The difference in geometry between the six cutting inserts is described and shown in table 2 and in figure 2.5.
All cutting inserts have a width of four millimetres. The width has been increased with one millimetre compared to the first cutting insert developed in the Sirius project of 2004/2005 [1]. The reason was to get a stronger cutting insert as well as make it is
Figure 2.5: Illustration of the cutting inserts that were developed for testing.
easier to manufacture more complex geometries. The new cutting inserts all have loosing angles of three degrees on all sides that is in contact with the cutting material, see figure 2.5. Three different cutting angles were tested, each with and without a chip breaking geometry. The chip breaking geometry was added to investigate how much it influences the force stability.
2.5 Materials to be Tested
Materials were chosen according to their believed ability to produce long chips, as well as not having to high Kc (Kc »1000 MPa). Therefore, four metals and one plastic were chosen. Each of the materials is presented below (More detailed properties can be found in Appendix B).
• Aluminium - SS 4212, EN AW-6082-T6 (Heat treated precipitation hardened and hot aged). Magnesium and silicon based allow with high strength, good corrosion resistance and high toughness.
• Brass - CW602N-00 (untreated state). Ingots with a small percentage of lead.
• Copper - SS 5011-04, CW008A. Drawn rod of oxygen free copper.
• Bronze(Tin) SS 5465-15, JM 3-15. High strength bronze with excellent corrosion resistance. Often used in bearing applications.
2.5. Materials to be Tested 13
• PTFE (Poly Tetra Flour Eten). Plastic with excellent temperature and aging durability.
C HAPTER 3 Test Setup 1
3.1 Test Series 1
In order to try to find the most suitable cutting geometry, it was necessary to perform several tests. The optimal geometry could not be decided without alternating the material to be cut as well. For example, the most suitable geometry in copper might not be the best in another material. Therefore, the most suitable geometry of the cutting insert had to be decided together with the different materials.
All tests in this series were performed at a cutting depth of 0.8 mm (except PTFE, which had a cutting depth of 2 mm), a falling weight of 40 kg and an impact velocity of about 6.1 m/s. The energy absorbing structure had a mass of 12 kg, which, theoretically, would give a cutting velocity of about 4.4 m/s at the start (for calculations, see equation 2.1- 2.4). Three drop tests were done in each setup to get a measurement of reliability. The complete list of tests performed in the first evaluation series can be viewed in appendix D. Graphs of all drop tests can be viewed in appendix G- L.
15
Figure 3.1: The result from a drop test with an aluminium test plate. Notice how the chip is not broken into several small pieces but is compressed and kept together throughout the whole cutting distance.
3.2 Results
The photo in figure 3.1 is taken from one of the first drop tests, and shows the typical result of a test with aluminium as energy absorbing material. In figure 3.2, graphs are shown from a test series using a cutting insert with no chip breaking groove and a neutral cutting angle. The graphs show the typical behaviour for each of the test materials.
After the test series, it was realized that the test setup was not stiff enough to reduce the vibration that occurred, which resulted in a variation of the cutting depth. Therefore, a new setup was prepared to decide the most favourable geometry of the cutting insert.
The surface of the cutting groove, however, made it clear which material that was suitable for this type of cutting. PTFE, bronze and brass all had a visible pattern of repeatable notches, where the chips have been torn away, see figure 3.3 (the close-ups are magnified seven times). This discontinuous chip breaking, where the chips are torn off along the entire cutting distance, results in a dip in force every time it happens, and then an increase as the cutting inserts goes into the material again. The chips that were torn off from these materials ranged from 0-2 mm lengthwise, and as seen in the graphs (figure 3.2), the forces vary several thousands Newton at worse.
Aluminium and copper, on the other hand, both had smooth surfaces and a continuous chip breakage. As predicted, this behaviour clearly gave a more favourable distribution of forces. Figure 3.4 show typical cutting grooves from these materials as well as the corresponding chips. The wave-like pattern that can be seen in figure 3.4(a) and figure 3.4(c) are the variation of cutting depth. If these variations could be removed, the stability of the levels of force would be increased.
3.2. Results 17
(a) Aluminium. (b) Copper.
(c) Brass. (d) Bronze.
(e) PTFE.
Figure 3.2: Graphs of measured forces versus time, counted from the impact
(a) Bronze. (b) Brass. (c) PTFE.
Figure 3.3: Notice the magnified notches in the cutting surfaces in each material.
(a) Aluminium. (b) Continuous aluminium chip.
(c) Copper. (d) Continuous copper chip.
Figure 3.4: The figure shows that the chip breaking is continuous and a smoth surface is the result from the cutting process.
3.3 Conclusions
The results from test series one showed that the aluminium alloy and the copper roughly gave the same result regarding force fluctations. However, the aluminium alloy is lighter, cheaper and more environmental friendly compared to the copper. These facts made the difference and it was thereby decided that the aluminium alloy was to be tested in further tests.
To illustrate how the difference in cutting depth influenced the fluctuations in force, the cutting depths in the aluminium test plates were measured for each geometry that was used. The force was given every 0.5 ms from the sampling program, which will give the deceleration according to equation 3.1:
a = F
m (3.1)
3.4. Measurement Stability 19 When the deceleration is known at each point, the new velocity after each time step can be calculated from equation ( 3.2). This requires the assumption that the deceleration is constant under each time step, as well as that the friction in the test rig is neglected.
v = v0− at (3.2)
Finally, when the velocity at each time step is known, the length (s) can be calculated by equation ( 3.3):
s = vt (3.3)
Cutting depth was measured at each 2 mm, and then be plotted side by side with the cutting force. The mean cutting depth and mean force from each test were then used to calculate the mean specific cutting force by using equation 1.1. This enabled the force to be calculated at every 2 mm and then plotted next to the measured values. If these two graphs would be similar, it basically tells that if only the cutting depth is kept at a constant level, the force should be much more stable during the stroke. As seen in the graphs 3.5(a)- 3.5(b), the two curves follow each other, and this was the behaviour in all of the graphs (see appendix M). A new test setup was therefore developed that should make sure that the cutting depth would be kept at a constant value.
3.4 Measurement Stability
The Kistler plate gives an analogue output signal which is not filtered. In the tests the sample rate was set to 2 kHz. A sample rate of 5 kHz was tested to see if the sampling rate was adequate, but the resulting graph showed the same behavior, which means that the 2 kHz was enough. To verify that the output data is not a sub-sampling of a noisy signal, the theoretical calculations could be used. By putting the theoretic force in the same graph as the measured force, a great similarity could be seen; peaks can be observed at the same length in both cases, see figure 3.5(b). From this observation it was clear that the signal was not a result from a sub-sampling phenomenon.
(a) The measured force plotted together with the measured cutting depth for a test with an aluminium test plate, cut with an insert with a chip breaking groove and negative geometry.
(b) Same test as in figure 3.5(a), but now with the measured force plotted together with the calculated force.
Figure 3.5: Graphs of measured forces plotted together with cutting depth and calculated force.
C HAPTER 4 Test Setup 2
4.1 Test Series 2
Figure 4.1: The second test setup viewed from above.
The second test setup, that was designed to further stabilize the cutting process, differed from the first in basically two ways, see figure 4.1- 4.2. First, the test plates were positioned with the cutting surface 90 degrees different. Secondly, two cutting inserts were used instead of one for each test. The cutting inserts were pointing towards each other, and the test plates were being pressed between them. This design was based on the assumption that even though one cutting insert may cut deeper then the other, the total depth will be the same, and therefore the cutting force will be more stable.
21
Figure 4.2: To the left, the first test setup. To the right, the new test setup with two cutting inserts.
Both cutting inserts were fastened in to a holder that is allowed to float relative the force measuring device in the axis of the cutting depth, see figure 4.3. This way, the cutting insert holder was thought to reposition itself if one cutting insert got a deeper cutting depth than the other.
(a) Left position. (b) Right position.
Figure 4.3: The cutting insert holder is able to move freely in the horizontal direction because of the two axles.
However, the repositioning idea proved not to work as intentioned. The sliding function was therefore disabled so the cutting insert holder was not able to move, as this would simplify the design of the steering column dramatically. A series of tests (see table 1 and figure 4.4- 4.5) was performed to determine the most suitable cutting geometry.
4.1. Test Series 2 23 Three tests where carried out to get a sign of repeatability. Also see appendix F for magnification of the graphs.
Chip Breaking Cutting Angle Aluminum (tests)
No Neutral 1,2,3
No Negative 1,2,3
No Positive 1,2,3
Yes Neutral 1,2,3
Yes Negative 1,2,3
Yes Positive 1,2,3
Table 1: Test Series 2 - Variation of Cutting Geometry.
(a) No Groove, Neutral Geometry (b) No Groove, Negative Geometry
(c) No Groove, Positive Geometry (d) Groove, Neutral Geometry Figure 4.4: Tests performed to decide optimal geometry.
(a) Groove, Negative Geometry (b) Groove, Positive Geometry Figure 4.5: Tests performed to decide optimal geometry.
4.2 Statistical Analysis of the Cutting Geometries
When the test series had been carried out, the results needed to be analysed in a way so they could be validated and compared. The result from each measurement was as- sumed to be a normal distribution. From the normal distribution the standard deviation (percentage) could be calculated, see table 2.
Type of cutting insert: Standard deviation (percentage) from mean value
No Groove, Neutral Geometry +/- 17.6%
No Groove, Negative Geometry +/- 16.8%
No Groove, Positive Geometry +/- 15.2%
Groove, Neutral Geometry +/- 15.9%
Groove, Negative Geometry +/- 15.9%
Groove, Positive Geometry +/- 15.7%
Table 2: The standard deviation (percentage) from mean value of each cutting geometry.
The standard deviation (percentage) could then be compared for each cutting insert.
As can be seen in table 2, the cutting insert with no groove and a positive cutting angle (see figure 4.6) had the lowest standard deviation (percentage).
4.3 Other Parameters
To further examine the energy absorption behaviour, a number of other parameters were interesting to test. These include, but are not limited to: cutting depth, mass of weight, temperature extremes and impact velocity.
4.3. Other Parameters 25
Figure 4.6: Cutting insert geometry that resulted in the lowest standard deviation (percentage).
Mass [kg] Velocity [m/s] Force [N] Cutting Depth, Total [mm]
40 4.4 2000 0,64
40 4.4 4000 1,21
55 4.7 6000 1,93
70 4.9 8000 2,57
85 5.0 10000 3,20
Table 3: Parmeter values for different cutting depths.
4.3.1 Variation of Cutting Depth
To be able to cut all the way through the test plates at all levels of force, the mass of the falling weight had to be increased. By using the same equations as when the cutting depth was calculated, the masses that would be decelerated to zero velocity after full cutting distance were calculated for the different levels of forces above 4 kN (equations 2.1- 2.4 and 3.1- 3.3 where used). Although the weight was dropped from the same height, because of the increase in mass, the velocity of the two bodies would also be different, see appendix E. Finally, the expected cutting depth was calculated by using the average specific cutting force from the previous test series. A complete list of the parameters used can be seen in table 3. Due to the limitations of the force measuring device, forces above 10.019 kN could not be registered. The results can be viewed in figure 4.7 a) - 4.7 d). The large fluctuations between 25-30 ms in figure 4.7 b) were thought
(a) Tests performed at 0.32 mm cutting depth per cutting insert. Average force was 2145 N.
(b) Tests performed at 0.61 mm cutting depth per cutting insert. Average force was 3920 N.
(c) Tests performed at 0.96 mm cutting depth per cutting insert. Average force was 5574 N.
(d) Tests performed at 1.28 mm cutting depth per cutting insert. Average force was 7447 N.
Figure 4.7: Tests with different cutting depths.
to depend on the testing material. When magnified, the reason could be observed, see figure 4.8. In figure 4.8(a), a clear vertical line can be seen, which most likely have influenced the otherwise continuous cutting process. Aluminum has a tendency to build up edges at the cutting insert, and when the cutting process got close to the defect, the built up material probably loosened and got stuck in the cutting groove. Figure 4.8(b) show a picture taken from a random location on the same test plate as a reference.
An interesting observation can also be seen in these graphs. The cutting velocity does not seem to influence the cutting force. If this would not have been the case, the force would have risen towards the end. To see if higher cutting depths gave a different variation in force, statistical analysis was again used to verify the fluctuations. As seen in table 4, all but the 4 kN level gave a variation of about +/-15 %.
4.3. Other Parameters 27
(a) The defect that can be seen after 93 mm of cutting from test 1.
(b) Random part of the cutting area from test 1.
Figure 4.8: 45 times magnification of the aluminum test plate from test 1.
Force Level Mean Percentage Deviation from Mean Force Level
2000 N +/- 14.57%
4000 N +/- 17.37%
6000 N +/- 15.27 %
8000 N +/- 12.99%
Mean Deviation for all Tests: +/- 15.05%
Table 4: Variation in force for different cutting depths.
4.3.2 Temperature Extremes
Depending on geographic location, vehicles are used in different temperatures. A hy- draulic damping system has the problem that the viscosity of the damping medium is strongly influenced by the surrounding temperature. But it is also well known that metals gets more brittle at temperatures below zero. And on the other hand, as the temperature of the material to be cut increases, so does the edge build-up edge in the cutting process.
In order to test the cutting process at the different temperatures, both the cutting insert and the test plates were frozen respectively heated. Due to limitations of the cooling medium that was used, the lowest temperature that could be tested was around -15 ◦C.
The test equipment were frozen at -17.5 ◦C for three hours, and then tested within five minutes. After one test, the equipment was put in the freezer for another 30 minutes before being tested again. In the other case, the equipment was heated at 60 ◦C for one hour before a test was made.
In this test series, only one level of force was tested. As mentioned earlier, the force measuring device could not properly measure forces above 10 kN. Therefore, the 8 kN
Figure 4.9: Tests performed at -15 ◦C (Average Force 7371N).
Figure 4.10: Tests performed at 55◦C (Average Force 7056N).
These results should be compared to the tests with the same parameters but performed at room temperature. The average force for the test at room temperature was 7447 N,
4.3. Other Parameters 29 which means that the temperature difference between -15 ◦C and 55 ◦C has little to almost none influence on the cutting force.
This was further shown when a statistical analysis was made. For the test at the low temperature, the standard deviation was +/- 13.4 %, and for the high temperature the respective value was +/- 14.5 %. These values should be compared to the mean value from the cutting depths tests, which was +/- 15%. This means that the temperature does not have a big influence to the force fluctuations.
4.3.3 Variation of Impact Velocity
Normally in cutting operations, the specific cutting force (Kc) is close to independent of the velocity above a certain value [7]. Because Kc is directly proportional to the cutting force, it is important to find if and where the velocity starts to influence Kc. A test was performed with a tensile strength measuring machine to find the static load that would initiate the cutting. At the chosen cutting depth (0.5 mm), Kc was only slightly higher then the average Kc from a test performed at 5 m/s. That leads to the somewhat surprising conclusion that Kc, and with that also the force, are not depending on the velocity at the range from 0 to 5 m/s.
C HAPTER 5 Results
5.1 The Correlation between Cutting Depth and Cutting Force
One of the main issues with the idea described in this thesis, and one of the major advantages with this principle of absorbing energy, is the assumption that the cutting force is linear with respect to the cutting depth. The graph shown in figure 5.1 shows how the average cutting force from the tests presented in chapter 4.3.1 varies with the cutting depth.
In normal cutting operations like turning and milling, the specific cutting force has a higher value at low cutting depths, and then decreases and flattens out to a close to constant value after a certain cutting depth [7]. In these tests, a similar pattern could be shown, see figure 5.2.
These graphs, together with the standard behaviour of cutting processes, make it pos- sible to estimate which cutting parameters are required to reach the different levels of force that are not possible to measure with the present test equipment. The width of the cutting insert should, theoretically, not have any significant importance other than making the cutting inserts stronger; i.e. the width of the cutting insert is thought to be proportional to the cutting force.
The graphs also show another important thing. The design with two cutting inserts pointing towards each other was made with the assumption that the cutting force would be linear to the cutting depth. This is true for cutting depths above ca 1.3 mm, where the graph can be approximated linear. This behaviour can be used to explain some of the variations that the tests show, for if one of the cutting inserts has a cutting depth lower then 1.3, then that will result in a higher total force because the Kc is increased.
31
Figure 5.1: Average force versus cutting depth from the tests presented in 4.3.1 Variation of Cutting Depth.
Figure 5.2: The dependency of cutting depth on the specific cutting force.
C HAPTER 6 Design Propositions
This chapter will present design propositions for the steering column, the ride down seats and the shoulder load limiter. Focus has been put on the steering column, which will be discussed in more detail than the other two. An alternative concept for the steering column is also presented in appendix C.
6.1 Steering Column
Already in the Sirius course it was realised that the length of the steering column and energy absorbing device needed to be minimized. Therefore, focus has been put on decreasing the overall length of the system.
The two test setups that have been used to evaluate the cutting principle proved that two cutting inserts with cutting edges facing each other gave a relative stable energy absorption. This meant that it was important to place the cutting insert this way in the steering column.
The proposed steering column works with a telescopic function in a crash, see figure 6.1. The numbers inside the rectangles do not tell the length of the part, they just symbolize the two phases in a crash.
One pair of cutting inserts will be active during the first 80 mm of translation, and a second pair for the remaining distance. Designing the system this way, and still be able to change the second cutting inserts during the translation of the column, requires some kind of locking mechanism. Of course, the second pair of cutting inserts could be used to keep that part still, because those cutting inserts will always have a deeper cutting depth then the first pair. This would be a risky design though, as there would be a force acting on the tip of the cutting inserts which would drastically increase the energy needed to change the cutting depth.
33
(a) Driving position. (b) 80mm translation. (c) 200 mm translation.
Figure 6.1: Tthe telescopic function in a crash.
Figure 6.2: The final design of the steering column.
6.1.1 Design
Several design concepts were developed with the above conditions in mind. After con- sultation with supervisor Kolita Mendis, the following design was chosen, see figure 6.2.
The proposed concept consists of a total of four cutting inserts, divided into two pairs, but where only three are able to be adjusted and one would be fixed, see figure 6.3.
Each movable cutting insert has a holder that holds the insert in place and enables it to be adjusted by the pyrotechnical charge, see figure 6.4. Because a cutting insert with
6.1. Steering Column 35
Figure 6.3: Arrows point out the location of the cutting inserts. The dotted arrow points at the fixed cutting insert.
Figure 6.4: The cutting insert holder that can adjust the insert with a pyrotechnical charge.
has been changed.
(a) Level adjuster and cutting insert. (b) Level adjuster and cutting insert after the py- rotechnical charge has been fired.
Figure 6.5: Illustration of level adjuster and it’s function.
The locking mechanism that would keep the second part of the system in place while the first is translating basically consists of a piston and a cylinder, see figure 6.6. At the end of the first stage, the translating part will hit the piston, which will cause two sprints to be pushed outwards, and thereby releasing the cylinder from the part attached to the chassis, see figure 6.7.
The telescopic parts are supposed to be kept together at normal driving with a glue nail, similar to the design of the bottom part of the old Volvo S40 steering column.
6.1. Steering Column 37
Figure 6.6: Locking mechanism for the second part of the steering column.
Figure 6.7: Locking device for the telescopic function. For simplicity, the piston is removed in this picture.
been increased to compensate for this. Two FE-simulations were made, with different angle of the applied force; one horizontal and one vertical relative the driver, see figure 6.8 and figure 6.9 respectively.
Figure 6.8: Force applied from the side. The two arrows mark the area where the total force where applied, trying to resemble the pressure generated from the steering wheel.
6.2. Ride Down Seats 39
Figure 6.9: Force applied from the steering column to the axis that holds the ergonomic lever.
The somewhat high stress is thought not to be of any concern, since no exposed part show any high stresses.
6.2 Ride Down Seats
If the graph from figure 5.2 is extrapolated until a total cutting depth of 4mm, (2mm on each cutting insert) the specific cutting force would be close to 700 MPa. The ride down seats are supposed to be subjected to forces up to 62 kN [4]. A suitable combination of cutting insert and cutting parameters would therefore be: 8 Cutting inserts, with a width of 5.5 mm and a maximum cutting depth of 2 mm on each. The theoretical force from each cutting insert would then be 7.7 kN.
A design suggestion is presented in figure 6.10- 6.11.
It is strongly recommended that the cutting inserts are placed in pairs of two, like in the test setup. It is also a good idea to place them horizontally relative to the car, so no cutting insert is affected by the weight of the driver.
With eight cutting inserts, there will be a maximum of 256 possible levels of force.
This is of course a little too much, so one suggestion is to make all but two fixed, and let these two be able to be adjusted like in the steering column, see figure 6.12. Instead of just repositioning the cutting insert to a decreased cutting depth, they can be completely removed from the cutting process. This will have a greater affect on the levels of force.
With these suggestions, there will be three different energy absorption levels available, see table 1.
Figure 6.10: The ride down seat, positioned for driving.
Total Force (kN) Cutting Insert Position
46 Both Removed
54 One Removed
62 None Removed
Table 1: An example of available levels of force with two movable cutting inserts (out of 8).
6.2. Ride Down Seats 41
(a) Seat positioned for ingress/egress. (b) Seat positioned for driving.
(c) Seat after a crash.
Figure 6.11: The position of the ride down seat before and after crash.
Figure 6.12: Front view of one ride down seat with eight cutting inserts and where two are able to be adjusted.
to a sliding track, see figure 6.13. Force levels can be chosen between 2, 3, 4 and 5 kN, but the design can be modified to either have fewer or more available force levels. This is accomplished by either making one cutting insert fixed, or by adding more cutting inserts.
By having one of the cutting inserts rigid, the concept will get two force levels. The length of energy absorption could also easily be modified to suit different cars. However, the concept should not be modified in the following way; one of the cutting insert is removed completely. This would most certainly result in big fluctuations in the energy absorption.
In a crash situation the concept will react as follow:
1. Dependency of the severity of the crash, the cutting inserts might be repositioned.
The cutting inserts are always positioned to absorb the highest amount of energy before a crash; this is if complications occur and the computer fails.
2. When the cutting inserts are in position, the system will start to absorb energy from the belt and decelerate the person, see figure 6.14.
6.3. Shoulder Belt Load Limiter 43
Figure 6.13: Visualization of the shoulder belt load limiter with the two cutting insert, energy absorbing material and the belt compartment.
Figure 6.14: The full stroke of the shoulder belt load limiter start position from the left.
C HAPTER 7 Conclusions
7.1 Conclusions
When the thesis work first started, the results from the Sirius course had to be improved in terms of decreasing the force fluctuations in the tests and reworking the design of the steering column. From having a cutting force that varied several thousand Newton, the results are now down to around +/-500 N. If these results can be implemented in the proposed steering column (or in another design) and kept throughout a crash, this method of absorbing energy is believed to both be cost effective and robust. As no major temperature influence has been seen in the range from -15◦C to 55◦C, this principle has an advantage over hydraulic systems. The velocity also seems to be independent, which is another advantage over a hydraulic system, which suffers high losses at high flows.
7.2 Future Work
There are some areas that need further investigation. Some of these areas have been looked at but either proved to be to time consuming or expensive, or they simply need to be examined in more detail. The proposed future work can be categorized in either test rig issues or prototype test issues.
7.2.1 Test Rig
The lowest temperature that was tested was somewhere around -15oC. The resulting lev- els of force did not differ much from the ones performed at room temperature. Aluminium alloys, as opposed to most types of steel alloys, have an extremely low ductile-to-brittle transition [8]. No data of the ductile-to-brittle transition temperature have been found
45
Figure 7.1: The cutting geometry that gave the best results.
of the specific steel used in the cutting inserts. But even if that temperature would have been known, temperatures down to around -50oC would be necessary to test if a system like this would be implemented in a car.
In all turning and milling operations, the specific cutting force varies with the feed and also a little with the cutting velocity. In this case, the feed and cutting velocity are integrated in each other. But in the tests that have been carried out, almost no difference in force was found at the end of the cutting distance, where the velocity had been decreased. If this really is the case, or if some error was present in the equipment or method of measurement, is uncertain and should be examined.
It has already been shown that Kc depends on the cutting depth (see figure 5.2).
Theoretically, the area of the chip that is being cut decides the cutting force at a certain Kc. The width of the cutting insert is thought to be proportional to the cutting force, but his is yet to be proved.
Since the geometry that proved to be the most suitable in this test setup had and angle, it is likely that there might exist a more optimal angle, see figure 7.1, that both can withstand the impact force as well as not creating to high forces in the other directions.
A sharper angle usually gives a cleaner cut, with less tendency of building up edges on the cutting edge.
Only one type of aluminium alloy was tested, and therefore it can be useful to test additional qualities or aluminium alloys. Materials that have a low variation of Kc with cutting depth is sought, as this probably would reduce the force fluctuations further.
47
7.2.2 Steering Column Prototype
The test rig was built with focus on being stiff and have small looses. To be able to use the results from this work it is important to test if the stiffness can be kept in the proposed design as well as throughout a crash. If a prototype was to be build, more detailed tests could be carried out; for example how the system acts when the force is applied off centre, if the levels of force are changed when an air bag is between the mass (human body) and the column, as well as to make sure that the telescopic function and the locking device work as intended.
R EFERENCES
[1] Sirius students of 2004/2005. CHALMERS UNIVERSITY OF TECHNOLOGY and LULEÅ UNIVERSITY OF TECHNOLOGY, 2005.
[2] SAAB. www.saab.com.
[3] Volvocars. http://www.volvocars.se/Volvoownership/volvosavedmylife.
[4] Sirius students of 2003/2004. CHALMERS UNIVERSITY OF TECHNOLOGY and LULEÅ UNIVERSITY OF TECHNOLOGY, 2004.
[5] Ohlins. http://www.ohlins.com/ces.shtml.
[6] Kistler. www.kistler.com.
[7] Anders Tollstén and Gunnar Ruding. Författarna och Industrilitteratur AB, 1998.
[8] William D Callister Jr John Wiley & Sons,. John Wiley and Sons, Inc, 2000.
49
A PPENDIX A KISTLER, 3-Component
Dynamometer
51
53
55
A PPENDIX B Tested Materials
Standard Notation
Material Composition
Yield Strength [MPa]
Tensile Strength [MPa]
Hardness [Brinell, HB]
SS 4212 EN AW-6082-T6*
AlSi1MgMn 260 310 94
Table B.1: Aluminum, *T6 Heat treated precipitation hardened and hot aged.
Standard Notation
Material Composition
Yield Strength [MPa]
Tensile Strength [MPa]
Hardness [Brinell, HB]
CW602N-00*
(SS N/A)
CuZn36PbAs 120 330 70
Table B.2: Brass, *Untreated state
57
