Investigation on the effects of changing automotive vehicles leaf springs from steel to hybrid polymer composite leaf springs

Weight reduction is a critical factor in the vehicles and numerous researchers worked on better materials and design optimization to achieve that goal. This paper is focused to design a natural fibre reinforced hybrid polymer composite mono leaf parabolic spring to replace the existing steel multi-leaf semi-elliptical spring in an attempt of reducing the unsprung mass and to investigate its characteristics. The spring so designed is analyzed for its deflection and stiffness, by Virtual work method, whose complex integral was solved numerically by Gauss Quadrature. Along with this, SolidWorks was used to construct a model and ANSYS tests were used to compute the deflection, analyze the stress pattern and verify that the fail proof condition is met. Subsequently, three prototypes of the springs were fabricated by hand lay-up method. These springs were subjected to a static load test which results in deflection values similar to that of laminated steel springs with an appreciable 83 percent weight reduction. Modal Analysis conducted ensures that the fundamental frequencies of different modes are conveniently away from values of human discomfort and also from the nominal driving engine speed of 2000 rpm. The responses to harmonic disturbances which were studied experimentally, analytically and by ANSYS which indicates that vibration amplitudes due to average road undulations are under control, limiting the amplitude within 12 mm. The newly designed spring shall be recommended as a potential replacement, which can further be augmented by future works to be carried out on pot holes and studying the force transmitted on to the vehicle.


Introduction
In the current circumstance, the major emphasis of car manufacturers has been on reducing vehicle weight without compromising the performance.This is with the intention of preserving energy and thereby the fuel.By replacing the leaf spring with composite one, the overall weight of the vehicle may potentially be brought down.
Utilization of better materials, optimization of the design, and improvements in production processes are three key factors that may significantly contribute to weight reduction.One of the areas in which there is potential for weight reduction is the suspension leaf spring, which contributes to ten to twenty percent of the unsprung weight [1].The rear suspension, typically takes the shape of a simply supported semielliptical beam, where long and slender plates called leaves are fastened to the frame [2].They may be positioned either above or below the axle.Depending on the requirements of the application, either single leaf springs or multi leaf springs may be employed.
Leaf spring serves the primary purpose of providing support for the vertical load carried by an automobile.In addition they are also used for the installation of the axle, the regulation of the height at which the vehicle travels and support for maintaining the tyres' contact with the road.It should also provide a comfortable ride by negotiating the bumps and uneven terrain of the road.The leaf springs are generally fixed to the frame at both ends or with one of the ends to the frame and the front attached by a shackle to the frame [3].Because of their elasticity, leaf springs are able to absorb induced energy in the form of strain energy, store it and then gradually release it after exhibiting a noticeable degree of deflection whenever they are loaded or subjected to vertical vibrations.Therefore, increasing the amount of strain energy that leaf springs could store is the key to ensure a better performing suspension system in automobile engineering.Consequently, the form that the leaf spring takes and the kind of material that is used for the springs are two of the most important design parameters.A number of works were carried on to consider the suitability of using composite material as a replacement for steel springs.Since the spring material needs a high specific strain energy storage capacity, engineering materials with maximum strength and minimum modulus of elasticity both in the axial direction are appropriate materials for a leaf spring system.Composites with good storage modulus and their low weight qualifies as a substantial candidate as leaf spring material [4][5][6][7][8][9][10][11][12][13][14][15][16][17].
As reinforcements Carbon, Glass, Kevlar, and Graphite are the most often employed fibres with epoxy and polyester being the most frequently used matrices.A flexural composite member would ideally have epoxy resin as matrix which would also exhibit high tensile strength, adhesion strength and inter-laminar shear strength.It also shows good compatibility with most of the natural and synthetic fibres.It is observed that, the epoxy matrix accounts for 60% of the composite, while fibre reinforcements accounts for 40% of the total weight [18].The epoxy and the hardener are combined in a mixture with a ratio of 10:1, by weight respectively [18][19][20][21][22][23][24][25].The growing demand for the use of natural fibre in manufacturing owing to their innate ability like predominant behavioural properties, low weight, low cost, great mechanical properties along with being non abrasive, ecofriendly and biodegradable [26] recommends them as a suitable candidate of reinforcement for composite springs.Study of natural fibres shows that Sisal fibres have good tensile and flexural strength and bond well with Epoxy resin.They also exhibit good interfacial bonding strength [27] Glass fibre [28] would combine well with sisal to attain the required stiffness and strength.Sisal hybrid composite have satisfying strength and deflection with a stress value nearly half of that of GFRP composite and is a potential candidate for spring material [29].The dynamic mechanical analysis carried on sisal fibre reinforced composites shows that the addition of Sisal fibre improves the natural frequency which is considered as a positive trait.The damping ratio also increases with the addition of Sisal fibre.It was also seen that the storage modulus as well as damping ratio of composite at a particular frequency is directly proportional to the amount of fibre used [30].
Normally composite springs are designed as mono-leaf springs that would by themselves serve to replace a multi-leaf spring with adequate strength and stiffness [19].Altogether, a number of research works suggest that an attempt to design and analyse a hybrid composite spring made of Sisal and E-glass fibres in epoxy resin would explore the futuristic models.Research on the effect of the geometry of monoleaf composites springs over their mechanical properties suggests that the composite stiffness is very much sensitive to thickness than the width of the spring.The stress distribution along the length is better if the thickness varies parabolically from the centre to the axle along the length, permitting a higher strength to weight ratio while preserving the deflection obtained [31][32][33].The basic curvature is usually semi-elliptical where but the cross sectional variation would be the critical factor that establishes the dynamic behaviour for the given material combination.Here the variation of cross section has been governed by a profile of parabolic thickness with a constant width, a condition which gives a good overall ride quality [34] and better vehicle dynamics.Also the sides are kept as straight as possible to prevent the inducement of bending moment about the longitudinal axis [35].
The lay of fibres also play a role in the mechanical properties.Various studies show that laying of fibres in the longitudinal direction enabled the spring to absorb the flexural load effectively.SEM studies depicted that the bonding between the glass fibre and epoxy was appreciable with fewer voids and required hardness in this orientation [36].Similarly, leaf spring with long fibre reinforcements carry three times the load than short fibre reinforced leaf spring, as increase in fibre length increases the crystallinity of matrix structure, thermal characteristics and mechanical behaviour under static and dynamic load [37].The ply angle also influences stiffness of a parabolic composite leaf spring.It was observed that stiffness is minimum when 45°ply angle is used while it is maximum at 0°ply angle scheme.This indicates that 0°ply scheme utilizes the full elastic performance and is better suited for leaf springs [38].The spring can be modelled and analysed in ANSYS with the boundary conditions as one end of the spring free to travel in the horizontal direction along with rotation in transverse direction whereas other end is allowed only to rotate in the transverse direction [39].The modal and harmonic analysis on steel and CFRP composite springs in ANSYS showed that for the first five mode shapes the composite leaf spring had a higher natural frequency than steel springs establishing their suitability for suspension [39].
From the literature, it is observed that E-Glass fibre reinforced in epoxy matrix can replace steel leaf spring [4,7,9] but have low flexural strength.In order to improve the flexural strength, an attempt is made to study the suitability of Sisal fibre hybrid composite as an alternate material for leaf spring.Multi leaf springs can be replaced by monoleaf composite springs of varying thickness, with thickness being maximum at the centre and gradually tapering towards the axle [19].This variation in thickness is done parabolically, which exhibits better stress distribution.
Although various studies were conducted to study the suitability of Sisal fibre -Eglass hybrid composite, such studies have not been conducted on springs with parabolic profiles.Analysis of the spring by using Virtual work method and Gauss Quadrature is a novelty introduced here, which is not seen in the scope of literature survey.
The Tata ACE Gold Petrol Mini Truck was selected as the vehicle for this study, where the purpose is to suggest that a parabolic monoleaf spring made up of Sisal-E Glass composite materials can be used as a suitable alternative for the standard steel leaf spring by delivering improved weight, reduced Von Mises stress, and acceptable deflection.

Materials and methodology
The matrix content of 60% would be the most ideal one for correct bonding properties of the composite and the remaining 40% is shared by two reinforcing fibres that are Sisal and E-glass, which with different combinations yield different properties where we shall select the optimal one [18].In the present study, a combination of fibres containing 20% Sisal and 20% E-glass, that were selected through an optimization procedure involving experiments with reduced models obtained through applying the conditions of dynamic similarity.The properties of the constituents used for fabrication are tabulated in table 1 as supplied by the manufacturer.

Design of the spring
Mono-leaf springs that are used to replace the conventional springs are generally designed with one of these conditions: Varying width and constant thickness, varying thickness and constant width and varying width and thickness with constant cross sectional area.A non-varying cross section would require a critical minimum value of cross sectional area which would also remain the same at the unwanted areas giving a poor design with wastage of excess material, for which reason the option was rejected.In this work, leaf spring is designed with constant width and varying thickness.Assuming the same constant width used in the conventional spring, would not only be simple, but also be efficient and ideal.Hence, width 'b' is taken as 60 mm.The thickness of the mono leaf varies parabolically, with thickness at the center being the maximum, as the pay load, which is the main load share for each wheel happen to act at the center of the springs.The bending moment diagram clearly indicates maximum bending moment at the centre and would gradually reduce towards the ends.Section gradually tapers towards the end corresponding to the magnitude of overall load distribution.A shear force diagram though indicates constant shear force regions on either sides of the central line of symmetry, practicality suggest that shear force would be maximum at the ends.Preventing the failure at the ends would be by ensuring that a critical thickness is maintained at the axle end that would guarantee enough shear strength.The thickness at the center is determined with maximum deflection as the main criterion using the uniform strength equation that is usually used in the design of multi-leaf springs, which idea is extended over here.To estimate the tensile, compressive, flexural and shear properties of the composite material are important prerequisite for study of the behavioral aspects of the geometry of the spring.By conducting experiments on few small specimens fabricated for that purpose, the following values have been obtained for the required composite.The central thickness is given by Where, F is the design load, L the half span length, b the width and σ max is the bending stress.On substituting the values h max is found to be 24 mm at the center.As the shear force plays a crucial role at the end, the thickness h min is obtained based on shear strength considering the shear stress distribution, yielding a maximum shear stress of 3 V/2 A obtained as 7.32 mm, observed at the neutral axis the requirement of any value above 7.32 mm we nominally choose 10 mm at the ends which would make it convenient for attaching the eye end.
Though spring is a dynamic element involved in executing vibrations and also taking up dynamic loads, the very representative factor of a spring is its stiffness which could be understood here through its static deflection.Every of its dynamic behavior and the vibration characteristics would very much depend upon this static deflection value.The deflection based on the previous multi-leaf spring of TATA ACE would lead to similar desirable characteristics but this time with a lighter spring.As parabolic springs have better compression resistance with greater flexibility, the thickness of the spring varies parabolically along the span of the spring with the thickness being max at its centre and tapers with parabolic profile towards the axle end according to the relation, Where, h x is the thickness at the distance x from the axle end, l the span length and h min are the end thickness and h is the difference between h max and h min .

Profile generation
Profile of the spring baring the eye region is developed as a parabola mentioned above, that is developed over an elliptical base profile given by X 2 / a 2 + Y 2 / b 2 = 1.The table 2, indicates the values of Y and h x for the various values of X along the entire length of the spring.The figure 1, shows the Parabolic profile imposed over the elliptical camber Series 1 shows the base profile that is semi elliptical camber and Series 2 represents the curve when the thickness varies parabolically along the span.

Analytical estimation with numerical solution 2.3.1. Virtual work method
With a variable cross section spring, out of various methods to determine deflection, it is decided to select virtual work method in this case.Unit load method would be more convenient here that gives us the solution in the form of a definite integral indicated as Where M is the bending moment and m the unit moment load The equivalent Youngs' Modulus E of the composite is given by A load range of 0 to 4000 N was applied upon, corresponding to which deflections are to be determined.

Numerical solution by gauss quadrature
Here with variable moment of inertia evaluating the definite integral directly would be very complex and usually some numerical method is chosen.Gauss Quadrature would be a powerful and accurate method for solving this.According to Gauss Quadrature, we select Where I is the Area Moment of Inertia of the cross-section of the spring which is a variable and a function of x and the values are symmetrically varying with respect to the center.Thus,   The deflection values are calculated for a series of loads up to the maximum recommended loading range for the vehicle.Obviously the numerical solution values follow a linear pattern and this is to be verified with an actual spring for the similar behavior.

Modeling and development of leaf spring
In view of analyzing the spring by Finite element Analysis, a model is created using Solid works software.With the complex geometry being involved, the basic curve is generated using the ellipse equation and the parabolic profile for one-tenth of the thickness is imposed on the curve resulting in a layer which is further extruded to the spring width.Ten such layers combine to form the total geometry in which each layer can be assigned the properties corresponding to one of the three ingredients based on the composition corresponding to our requirement in every of the spring there would be five layers of epoxy and the number of Sisal and E-Glass layers would be two and three respectively.Finally, the model has been configured and is saved as IGES files, so as to facilitate its transfer to ANSYS, where they would be analyzed.To determine the mechanical properties experimentally, a mould is prepared to fabricate the spring.High Density High Moisture Resistance (HDHMR), an upgraded term of plywood is used to develop tough and robust wooden mould for usage.Figure 2 shows the layout of female die and male dies for the spring.
A semielliptical-shaped wooden mould for fabricating the fibre composite leaf spring with closed female and male die showing the spring cavity is shown in figure 3.
The two pieces of wood on each side were connected using a nut and bolt in the appropriate position.For every 100 g of resin, 10 g of hardener are added to the mixture.Using the weighing scale an accurate weight of the fibres and epoxy needed is measured.According to the dimensions of the leaf, the glass fibre was cut to size.In this case, a total of 2 sisal fibre layers in addition to 3 roving layers are laid in order to attain a thickness of 24 mm across the middle.In the traditional method of hand lay-up, the first step is to cover the mould with a polythene sheet on both dies for the purpose of separating the substance from the mould.Figure 4 shows the arrangements for fabrication of leaf spring before straightening of fibres.Polythene sheets are best suited for use in wooden moulds as an alternative to applying release agent, for it produces a nice surface finish.Figure 5 shows the arrangements of composites before the fabrication.Initially epoxy resin was applied with maintaining one roving e-glass layer, followed by applying resin on it again and then keeping one sisal fibre layer and continuing the process by alternating between sisal fibre layer and e-glass roving layer each time.The rolling procedure helps eliminate any air bubbles that may have been present.This method is used for each successive layering of the glass fibre strip.After the process of rolling, the mould is closed from the top followed by applying weight of around 2kN, so transforming it into a compression moulding.The epoxy should then be painted on with a brush as the first layer covering the cavity.During the process of individually laying down each layer of the spring, extreme caution was taken to straighten the fibres as the lay of fibres play a crucial role in tensile strength and stiffness.The set up is allowed to cure at room temperature for 24 h.By using the aforementioned fabrication method, mono composite leaf springs were produced.These are the leaf springs after having removed from the mould, still need polishing.The cutting and grinding procedure were used to complete the leaf springs in order to achieve the end product, for further characterization.

Static analysis by ansys and experiment
A universal testing machine was used to conduct a static load test on three springs that had been fabricated.The test included applying a central transverse force to the springs in the same manner that it would be applied in the vehicle.The resultant load versus deflection characteristics reveal a trend that is very much comparable to that suggested by ANSYS and the Virtual work approach, but with values that are somewhat higher than those obtained from both of those methods.The decrease of unsprung mass was the first key goal of this work and it can be said that this objective has been effectively accomplished here.The freshly produced springs each weigh 1.51 kg, 1.54 kg, and 1.43 kg, whereas the weight based on geometry is calculated from volume estimates and density considerations.Volume of the spring was obtained by straightening the elliptical base curve and placing the parabola at 10 mm above this line, resulting in a rectangle and a parabolic area.Parabola would contain 2 thirds the area of its enclosing rectangle.The volume of the spring is 594100 mm 3 calculated from equation (10) and the weight 1.42 kg calculated from equation ( 11)

Result and discussion of static analysis
The next essential need for this spring would be for it to sustain the utmost force that is applied upon it.The findings of the ANSYS stress distribution indicate that under the maximum loading condition of 4000 N, as shown in figures 6(a) & (b), the stress level near the eye end and the Von-Mises stress value that is seen to be at its highest is 151.6 MPa, which is lower than the bending strength.This ensures that there is no possibility of a spring failure.
The second factor taken into consideration is the spring's stiffness, which can be determined by calculating the average of the results by the three different methods for the spring's deflection as shown in table 3 above.The results of the virtual work approach calculated using Gauss Quadrature and ANSYS show linearity in stiffness with values of 101 N mm −1 and 83.66 N mm −1 respectively.The deflection values acquired experimentally from three identical prototypes indicate a very similar trend with slightly higher values yielding an average stiffness value of 81.63 N mm −1 .The linearity in stiffness as indicated in figure 7, suggests that the spring would operate admirably under dynamic loads, producing the required vibration characteristics; nevertheless, this will need to be verified experimentally.
The virtual work method assumes a homogeneous design with a single value of equivalent Young's modulus.ANSYS, on the other hand, assumes 10 layers of strips with three layers of glass fibre and two layers of sisal fibre in between five layers of matrix.The actual prototype is then filled by hand lay up to try to get an even distribution with the right proportion.The extra bending that was seen in the actual prototypes was caused by the lack of homogeneity and the short lengths of the Glass fibres and Sisal fibres compared to the length of the springs.

FEA modal analysis of composite spring
When a system is in a condition of free vibration, it is vibrating under the influence of its intrinsic forces after being disturbed from its equilibrium position and being free from the impact of any external forces.It will vibrate at the natural frequency, but due to damping, its amplitude progressively decreases.In free vibration, the natural frequency and the amplitude are the two factors that are of the most importance.When designing a structure to withstand dynamic loading circumstances, critical design criteria include the natural frequencies as well as the mode shapes.While designing a component, knowledge about the natural frequency of vibration forms the basis of further dynamic analysis.Hence for studying the vibration characteristics, modal analysis is carried out which serves as a basic step before carrying out harmonic analysis or a spectrum analysis.Solid Edge was used to construct the master leaf spring model, with the boundary condition (figure 8) for Eye 1 being fixed, serving as a guide with zero DOF and Eye 2 is restricted to moving just along the X-axis with one degree of freedom.Table 4 reveals the natural frequencies obtained by modal analysis.Figures 9(a) to (f) shows the various fundamental modes of vibration of the leaf spring.

Result and discussion of modal analysis
The natural frequencies or the fundamental modes of the spring with values starting from 50 Hz and going above, assures that they are above and away from frequencies of human discomfort that is 10 Hz and below.We  can also find that a comfortable driving speed of 2000 rpm which corresponds to 33.33 Hz is also below the first modal frequency, ensuring avoidance of any resonance in this range.

Harmonic analysis for vibration
Road disturbances are in many cases assumed to be sinusoidal waves.In an attempt to study the effect of these harmonic disturbances, a harmonic analysis is carried out on the springs.For carrying out the analysis, a test rig was designed and fabricated so that experiments with a frequency range up to 10 Hz could be conducted with various loads attached to the rig.

Design of test rig for harmonic analysis
The test rig consists of frame with roller support at the end for holding the eye ends.Vibration is given by an eccentric set up driven by a variable speed motor driven through a shaft with a spring, whose speed is controlled using an autotransformer.The ½ hp motor with a rated speed of 2800 rpm is easily able to cater to the needs of experiments conducted in the range of 400 to 600 rpm.The leaf spring held on the supports is facilitated to carry load over its periphery with which harmonic disturbances were applied while conducting experiments.
The road surface without considering large pot holes is approximated as a harmonic wave of amplitudes in the range of 4 to 10 mm which could be applied in the test rig as Force to stiffness ratio.The amplitude of force applied during the experiment is the net effect due to the eccentric as well as the additional load added over the spring.On conducting the experiment, the dynamic response obtained is tabulated as shown below in table 5. Also the theoretical deflection is calculated by using the formula   Where x is the amplitude of vibration, F 0 is the harmonic force applied and k is the stiffness of the spring, w is the forcing frequency, and n w is the natural frequency.

Results and discussion of harmonic analysis
The experimental and theoretical amplitude of vibration of the spring was plotted as shown in figures 10 & 11, respectively.The theoretical amplitude has been calculated using the equation (12).
A harmonic analysis covering the frequency range of 40 Hz to 1000 Hz has been carried out in ANSYS.It could be observed that the frequency range included in this study does not produce any peak and that in the simulated normal roads and normal driving speeds, there are no resonant vibrations that could have brought much discomfort to the travelers and disturbances or damages to luggage contents.From the above all Experimental Analysis, ANSYS Analysis, and Analytical Analysis of the Composite Mono Leaf Spring, into determining the amplitude of vibration (displacement of leaf from equilibrium position), for all of the different types of analysis lead to very similar inferences as suggested by ANSYS as shown in the table 6.The plot in figure 12. explains how ANSYS would cover a wide range of frequencies beyond our scope of study, pointing out deformation responses for various frequencies.
A comparison of the amplitudes of the vibrations of the master leaf spring with eye acquired by experimentation and those produced from analysis is shown in figure 13.The plot clearly demonstrates that the differences between the ANSYS, experimental and theoretical values of deformation responses are not significant and hence could be approximated as one and the same, considering the maximum value of the three at every frequency.The amplitude values within 12 mm in the entire disturbing force range, is considered to be under control, which leads to inferences that the spring can handle the vibrations due to road undulations successfully.

Conclusion
• The findings of the experiments indicate that epoxy's poor tensile strength as well as a low impact and flexural strength, when reinforced with sisal and e-glass shows significant increases in tensile, impact, and flexural strength ratings.
• In the case of springs, they bear the full load of the vehicle, by sharing one fourth of the load each and the need of strength and its efficacy cannot be overstated.Since of this, the materials may be made of epoxy reinforced with sisal and e-glass, which would make them more effective because they are able to sustain a large load with acceptable compliances.
• As a consequence, an effective design of a leaf spring together with the appropriate material is developed.
• Static Analysis shows the strength and stiffness characteristics being sufficient and practical.
• The modal Analysis displays that resonant frequencies are away from the vehicles operating speed ranges that were converted to frequencies.
• Harmonic Analysis results carried out experimentally, theoretically and by ANSYS portray the resultant deflections as much under the discomforting or damaging values.
• The work also leads towards future works such as transient analysis due to pot holes that which would show the capacity of the spring to handle impacts.Also the force transmitted and handling characteristics can be studied.
• The work so far herewith implies a strong candidature for the newly developed spring to replace the older one.

Figure 2 .Figure 3 .
Figure 2. (a) Layout of Female die for spring.All dimensions are in mm.(b) Layout of male die for spring.All dimensions are in mm.

Figure 4 .
Figure 4. Arrangement of materials in die mould before straightening of fibres.

Figure 5 .
Figure 5. (a) Fabricated composite leaf spring.(b) leaf spring with eye in fixture.

Figure 8 .
Figure 8. Boundary condition applied on the spring.

Figure 9 .
Figure 9. (a) Leaf spring.(b) First mode of vibration.(c) Second mode of vibration.(d) Third mode of vibration.(e) Fourth mode of vibration.(f) Fifth mode of vibration.

Figure 10 .
Figure 10.Experimental amplitude of vibration of spring.

Figure 11 .
Figure 11.Theoretical amplitude of vibration of spring.

Figure 13 .
Figure 13.Graphical representation of various methods.

Table 1 .
Properties of the constituents of composition.

Table 3 .
Load versus deflection by experiment, virtual work and ANSYS.

Table 4 .
Natural frequencies obtained by applying boundary conditions.