Concept Design for Transmission Systems Colin A. BELL, Cristinel MARES and Romeo P. GLOVNEA School of Engineering and Design, Brunel University Kingston lane, Uxbridge, UB8 3PH United Kingdom
ABSTRACT The intense competition on the global markets faced by manufacturing companies is reflected by critical issues like decreasing the product cycle-time, lowering costs and increasing quality. The quest to meet the market requirements can be viewed as a multi-criteria decision-making problem set within the space of engineering characteristics for the designed system. In this paper one particular design method: Quality Function Deployment (QFD) is used to derive different design concepts for transmission systems and asses them according to specific criteria linked with requirements. The engineering design process is based on a structured design methodology and multiple aspects of the design process are presented from the perspective of a set of criteria and objectives like: power density, ratio range, efficiency, drivability, response time, smoothness, noise and costs. This study will try to explore the intimate relation between the most important features for the system definition in specific applications and their relation to different concepts. Keywords: design methods, QFD, transmission systems, CVT, design optimisation INTRODUCTION During the product development process, in order to answer to the rapid changes stimulated by technological innovations and changing customer demands, the product design incorporates many facets, such as requirement analysis, conceptual design, engineering design, product manufacturing, analysis of distribution and service support, and dispose/recycle aspects. As discussed by the two main schools of thought regarding the product development technologies: the scientific, Dixon [1], and the engineering, Koen [2], the design process has been optimised for different specific targets: design and assembly, design for disassembly, design for
recycle, design for cost, design for reliability, design for serviceability, and design for environment. These design approaches focus on a stage in the product life cycle and procedures for integration of all these aspects have become an important aim for the engineering community. The Quality Function Deployment (QFD) method, proposed initially by Akao in Japan in 1966 [3] is a mechanism to translate into engineering language, the voice of customers, relying on the Pugh’s concept selection method [4], and by using a translation chart, ‘House of Quality’ [5]. This includes the desired attributes and the engineering characteristics together with their weights, with the aim of prioritising the characteristics by using the information stored in the House of Quality. The House of Quality is produced by the following specific steps: • specify ‘whats’: determine important requirements by design experience and market research and collect it into the HQ; • translation of the requirements into ‘hows’, or technical characteristics; • construct the relationship matrix which will show the connection between ‘whats’ and ‘hows’; • analysis of the correlation matrix will show which requirements are more important by their weighting and how for a product each of the requirements are ranked; • based on the analysis, design target are set and will be prioritised during the design process. During the QFD process the product design team must make a selection of the design features. The complexity of the design process relates to the experience of the designer to lead to an unforced evaluation of the tables involved in the QFD.
The present paper aims at using the prioritising methodology, specific to the QFD process, for the analysis and design of a new transmission system using the traction drive concept. The specific requirements (rated by importance) are used for developing a new concept, the attributes of which address a specific market sector in an integrated approach [6, 7]. The discussion of the design attributes is centred on their interaction and effects on the final product. The design approach included in this study shows only the initial application of the HQ methodology, focusing on the characteristics of a mechanical transmission device influenced by a number of quality requirements. BACKGROUND Continuously variable transmissions (CVT) have been a very attractive alternative to automatic gearboxes for many years; however their potential has never been fully uncovered and applied routinely into automobiles’ drive-train. The advancement in the application of CVTs to automobiles has been slow. The main reasons have traditionally been considered to be excessive weight, limited torque capability in dynamic conditions, large inertia of rotating elements, limited durability and high cost in comparison to manual transmissions. In relatively recent years intense R&D work has seen important advances in the design and materials for such devices, which improved their durability, efficiency and power/weight ratio. There are two main types of CVTs currently used in the transmission of automobiles: belt/chain type and toroidal traction drives. The former type uses belts of various constructions which transmit the power from a set of conical pulleys located on the driving shaft to another set on the driven shaft. By changing the axial position of one pulley, the radius of contact between the belt and the cones changes, varying the transmission ratio accordingly. The main problems involved with this type of CVT are the relatively low power transmitted, limited by the friction coefficient between belt and pulleys, the tensile strength of belt, and wear of the elements due to high dry or boundary friction forces present. Toroidal traction drives transmit power by shearing a film of lubricant entrained between contacting elements, which are typically toroidal rollers. In this
case the abrasive wear is eliminated as the metallic surfaces are always separated by the continuous fluid film, the durability being limited by contact fatigue. Careful design of the device in order to limit the magnitude of the Hertzian pressure, correct choice of the surface finish and hardness and the selection of the materials for the contacting bodies, can make the working life of these elements beyond that of the whole assembly. The fluid used in these applications is specially designed for high traction coefficient and good behaviour at relatively large temperatures. In any case the design of the CVT is made for a traction coefficient well below the maximum value characteristic for the fluid. For example the limit traction coefficient is typically 0.1 at 40°C and 0.08 at 100°C, [8] however the value used in calculations is usually only 0.045. This allows for overloading of the contacts or unexpected temperature rise, without the failure of the film. Current toroidal CVTs use full or half toroidal input/output discs, with toroidal intermediary elements, the main differences between different manufacturers being in the loading and synchronising mechanisms of the intermediary elements. Some also employ a planetary gear set to achieve infinite transmission ratios, and various systems for selecting neutral position and the reverse gear. The reported efficiency of toroidal CVTs is between 90 and 95 percent with a transmission efficiency of 85-90 percent [9, 10] depending on the particular construction, which is superior to an automatic gearbox, but lower than that of a manual gearbox. The main advantage of this type of transmission thus relies on its ability to allow the engine to work at its maximum efficiency, at any given load and speed, which will have an effect on improving overall fuel efficiency of the vehicle. A novel type of toroidal CVT, capable of automatically adjusting the transmission ratio as a function of the resistive torque, has been reported in [11, 12]. A schematic of the device, which reveals its functioning principle, is shown in Figure 1. The device consists of two input discs, one conical, fixed to the shaft and the other toroidal, which has axial but not torsional mobility relative to the shaft. The inverted conical output disc is connected to the output shaft through a mechanism which is able to
convert torque to axial force, e.g. a screw. Between the input and output discs there are placed a convenient number of spherical elements, which do not have a materialised axis of rotation.
ratios when the load behind the toroidal disc is generated by a spring, ensuring a linear relationship between force and displacement [11]. The present paper shows the initial design methodology aimed at optimising the CVT in order to fulfil certain functional criteria.
R
THE QFD PROCESS
Input toroidal disc Ball screw coupling
Separator Output conical disc
Figure 2. Design solution for a constant-power CVT This design offers a relatively compact solution for a certain power with a good range of transmission
Physical Practical
Output shaft
Usable
Input shaft
Ratio Range
Spring
Ratio Range
Efficiency
Input conical disc
3
Total
Intermediary spheres
Ride Smoothness
Based on this principle a design solution for a device can keep the power output constant, or quasi-constant, is shown in Figure 2 [12].
The importance of each requirement is subjective; a driver for example might favour a smoother ride and quicker response time, whilst an engineer might favour efficiency or an improved power-weight ratio. In an attempt to quantify the relative importance of each criterion a simple house of quality matrix has been created, as shown in Figure 3, with the intended use of a typical family sized automobile.
Response Time
A resistive toque applied to the exit shaft causes the output disc to move axially, forcing the balls to change their position and thus the radius of contacts, relative to the axis of the input shaft. This will cause a change of the transmission ratio, the degree of this change depending of the geometry of the elements and the force loading the contacting elements.
Noise
Figure 1. Functional principle of CVT
The QFD process can be adapted for different types of applications, such as passenger cars, performance automobiles, heavy goods vehicles, etc. The most important considerations will reflect the specific requirements of each particular use. Each requirement can then be analysed and improved individually, however the interrelation of the requirements means that all must be considered at each stage to ensure that improving one area doesn’t significantly worsen another.
Maintenance
Axis of input shaft
Cost
A
Durability
T
F
The general requirements of automotive transmission systems are well known and documented, however to truly optimise a device to meet specific requirements, each criterion must be rated relative to one another.
Power-Weight Ratio
r
3
9
1
9
9
3
9 46
3
3
9
9
9
9
9 54
3
3 9
9 9
9 3
3 3
3 36 3 34
3
9
3
3 32
1
1
1 10
1
1 18
Efficiency
3
Power-Weight Ratio Durability
3 1
3 3
3
Cost
9
1
3
1
Maintenance
1
1
1
1
3
Noise
1
1
1
3
1
9
Response Time
3
1
3
3
3
9
9
Ride Smoothness
1
1
3
3
3
9
9
Figure 3. Matrix of Importance
3 34 3
32
Within the matrix each requirements has been divided into three categories, physical, practical, and usable. Each requirement is given a score of 1, 3 or 9 relative to another where 3 indicates that the requirements are perceived to be of equal importance, and 9 indicates a priority of a column requirement.
The response time of this design will largely depend on the transmission ratio’s function of torque. This can be controlled independently by altering the loading system applied to the toroidal disc, hence response time will not be considered within this paper.
Physical properties are those specifically related to measurable parameters. Power-to-Weight ratio for example covers the size, mass, and power capacity, whilst the efficiency incorporates specific power losses such as frictional and contact. Practical properties relate to ‘real-world’ issues such as the feasibility of actually implementing the product in a particular type of automobile. Durability encompasses life span, tendency to wear and reduction in performance over the product’s life. Usable properties look solely at the interaction with the user (driver). Whilst noise is clearly measurable (although hard to predict at this stage), response time and ride smoothness are somewhat subjective. CVTs offer near perfect ride smoothness due to the lack of a discrete gear change; however some drivers see this as a negative feature due to the lack of perceived acceleration that occurs in a CVT-fitted automobile.
To observe the effect of altering dimensions and features, a computer program has been created that quickly displays the quantifiable requirements relative to the current dimensions of key elements.
The importance of reduced weight in an automobile is highlighted by the highly placed ‘power-to-weight ratio’. Excess weight in automobiles can severely reduce fuel economy, performance and cornering.
The addition of a planetary gear system will also negate the need for an additional ‘clutch’ since there can be a net speed output of zero even when ‘in gear’, hence reducing the total weight of the transmission system. The ratio range can also be changed by altering the physical dimensions of the key elements. Minimising the angle of the conical output disc will increase the ratio range as seen in Figure 4, however it will also increase the movement required from the disc itself, thus significantly increasing the overall length of the device. 0.3
4
0.25
3
0.2
2
0.15
1
0.1
0 0
20
40
Transmission Ratio
As derived from this matrix the five most important considerations in automotive transmission design are efficiency, ratio range, power-weight ratio, durability and response time. With the current industry emphasis on reduced fuel consumption in automobiles, it is not unsurprising that efficiency is viewed as being of highest importance. The improvements offered in fuel economy are often touted as the reason that CVTs are considered the future of automotive transmission. Ratio range likewise should also feature highly on any device designed to provide a variable ratio.
Ratio Range To compete with current transmissions, the CVT must offer at least an equivalent range of ratios. Previous dimensions [11] have offered a range of approximately 1:1.2 to 1:2.7, which is comparable, but not as good as most manual transmissions. The ratio range can be substantially improved by the addition of a planetary gear system. This system effectively deducts the engine speed from the transmission output speed, thus allowing a theoretical infinite selection of ratios up to a maximum limit. The total ratio range of the transmission can be increased to whatever value is required by altering the final gear ratio.
Length [m]
Although physical properties are fairly easy to predict and calculate, practical and usable properties generally require testing or simulation, which is difficult to achieve at the design stage. Certain parameters can be predicted based on previous experience. Durability for example is often governed by the Hertzian contact pressure in traction devices. Similarly costs are increased substantially for higher tolerance requirements, or more complex designs.
DISCUSSION
60
Output Disc Angle [deg]
Figure 4. Change in ratio range (▲) and overall length (□) function of output disc cone angle
Power Output [KW]
Power-Weight Ratio The power capacity of the transmission system is limited by the traction coefficient of the transmission fluid. Assuming this value cannot reasonably be increased beyond 0.045 without reducing the life of the device there are a limited number of changes that can be made to the CVT without increasing the size and weight. Removing the need for a clutch reduces the total weight of the transmission; furthermore it is possible to remove the need for an additional flywheel assuming the input discs have sufficient mass/inertia. Further improvements to the power-to-weight ratio can be made by altering the dimensions of the elements. Elementary calculations show that the maximum power capacity for any particular overall size can be achieved by keeping the surfaces of the conical discs as close to parallel as possible, as seen in Figure 5.
100.0 100 90.0 80.0 80 70.0 60.0 50.0 60 40.0 30.0 20.0 40
20 25 30 Conic al Inp 35 40 45 ut Dis 50 c Ang le [de g]
17.5 12.5 7.5 2.5 55
tD tpu u O
is
le ng cA
g] [de
Figure 5. Variation of average power output This is not necessarily practical due to the large force that would be required to alter the position of the intermediary elements. An additional corollary of this is a similar increase in Hertzian pressure, as shown in Figure 6. Increasing the Hertzian pressure will decrease durability, which is rated as equally important to the power-weight ratio, so there is little point in improving one at the expense of another. An alternative solution is to increase the number of contact points by including additional intermediary elements. The additional elements will have a beneficial effect upon the magnitude of the Hertzian pressure in the elastohydrodynamic contacts, by
dividing the force on each element by the number of elements.
Hertzian Pressure [GPa]
Since the ratio range can be theoretically infinite with the addition of a planetary gear system, there is little advantage in considering the range when optimising the dimensions of the elements.
2.1 2.2 2.0 1.8 1.5 1.6 1.4 1.2 0.9 1.0
17.5
0.8 0.6 0.4 0.4
20
12.5
25
Conic a
30
l Inpu
35
t Disc
40
45
50 Angle [deg]
gl e 7.5 An c s i 2.5 tD tpu 55 u O
g] [de
Figure 6. Variation in maximum Hertzian pressure Calculations show that a variator of this design could handle a typical amount of torque produced by a large engine (approximately 300Nm) whilst still keeping the Hertzian pressure below 1GPa, assuming there are at least six intermediary elements. The overall diameter of this device would be about 0.25m, with a similar length. This size could be reduced almost linearly for lower powered engines. Efficiency Efficiency calculations have shown that the greatest losses within the traction device occur between the intermediary ball elements and the separator designed to force them to rotate and prevent adjacent elements from coming into contact. Reducing the friction coefficient between the two surfaces would reduce these losses; however the greatest reduction could be achieved by ensuring a rolling contact. This is made more difficult by the variation in axis of rotation of the intermediary elements. A fixed-angle rolling separator would only provide a rolling contact for a small variation in the angle of the axis of rotation, at other times the rotation would be almost perpendicular to the separator. A variable angle separator would solve this but would incur significant additional costs and potential areas of failure (reduced durability), both of which are rated as relatively important requirements. Additional losses occur due to parasitic, relative motions on the contact area (spin and side-slip) [13, 14] together with standard bearing and fluidic churning losses. A solution that reduced these losses without incurring significant penalties in other requirements would be perhaps more attractive, however despite these losses the efficiency of the
device could still theoretically be above 90%. As stated previously, whilst this is relatively low in comparison to manual transmissions, the overall specific fuel consumption of the vehicle could still be significantly lower since the engine could operate more often within a higher efficiency envelope.
CONCLUSIONS This paper presents an application of the Quality Function Deployment methodology to the design of a transmission system, with the intention of improving the design in response to certain rated criteria. These criteria were rated according a single specified application, and certain solutions were offered in an attempt to meet these requirements. The design method highlights the most important features, but these must not be considered in isolation. Often the best solution for a particular requirement might not offer the best overall benefit due to significant reductions in other important areas. The advantage of this design method is that it is not only the ranking, but the relative scores that can be taken into account when weighing up alternative solutions. A further step would be to apply more advance design methodologies to make the process less subjective to take into account uncertainties in the assignment of relative scores. Although only the first stage is covered here, the method has been successfully applied to a product resulting in significant design improvements.
REFERENCES [1] Dixon, J., “On research methodology towards a scientific theory of engineering design”, Design Theory ‘88, Newsome S. L., Spillers W. L. and Finger S. (Eds.), Springer-Verlag, 1988. [2] Koen, B., The definition of the engineering method, The American Society of Engineering Education, 1985. [3] Akao, Y., Quality function development, Productivity Press, Cambridge, 1990. [4] Pugh, S., “Concept selection: a method that works”, Conference proceedings ICED, 1981. [5] Houser, J. and Clausing, D. P., The House of Quality, Harvard Business Review, May-June, 1988.
[6] Barkay, J., “Design for Serviceability”, Machina Design, 2005, 77, 134. [7] Butarrs, S. and Roland, R., “Design for Manufacture”, Electron, 2006, 23, 24-26. [8] Anghel, V., Glovnea, R.P., and Spikes, H.A., “Friction and film-forming behaviour of five traction fluids”, Journal of Synthetic Lubrication –2004, 21-1, 13-32 [9] Newall, J.P., Cowperthwaite, S., Hough, M. and Lee, A.P., “Efficiency modelling in the full toroidal variator: Investigation into optimization of EHL contact conditions to maximize contact efficiency”, 2004 International Continuously Variable and Hybrid Transmission Congress, September 23-25, 2004, September, 2004, UC Davis [10] Hirohisa Tanaka, Nozomi Toyoda, Hisashi Machida, Takashi Imanishi, “Development of a 6 Power-Roller Half-Toroidal CVTMechanism and Efficiency”, 2004 International Continuously Variable and Hybrid Transmission Congress, September 23-25, 2004, September, 2004, UC Davis [11] Cretu, O.S., and Glovnea, R.P., “Constant power continuously variable transmission (CP-CVT) – Operating principle and analysis”, ASME Transaction, Journal of Mechanical Design, 2005, Vol. 127, 1, 114-119 [12] Cretu, O.S., and Glovnea, R.P., “Geometrical and dimensional optimisation of a constant-power continuously variable transmission”, 2006, Buletinul Intsitutului Politehnic Iasi (Bulletin of the Technical University Iasi), Vol LII, Part 6A, Construction of machines section, 149-161 [13] Cretu, O.S., and Glovnea, R.P., “Traction Drive with Reduced Spin Losses”, ASME Trans. J. Trib., 2003, 125, 3, 507-512 [14] Zhang, Y., Zhang, X., and Tobler, W., “A Systematic Model for the Analysis of Contact, Side Slip and Traction of Toroidal Drives”, Journal of Mechanical Design., 2000, 122, 4, 523-528
