Fortis Saxonia – experimental stress analysis when driving

Environmental pressure, with increasingly scarce energy and raw materials, is proving a strong driver for finding innovative solutions. For automobile manufacture, this means aspiring to lightweight construction vehicles with efficient and economical drive systems. One group following this dream is the student research project Fortis Saxonia (Saxony the Brave) at TU Chemnitz.

The Fortis Saxonia project

Environmental pressure, with increasingly scarce energy and raw materials, is proving a strong driver for finding innovative solutions. For automobile manufacture, this means aspiring to lightweight construction vehicles with efficient and economical drive systems. One group following this dream is the student research project Fortis Saxonia (Saxony the Brave) at TU Chemnitz.

Around 20 students and graduates from the faculties of Mechanical Engineering, Electrical Engineering and IT, Computer Science and Humanities and Social Sciences are currently researching fuel cell technology, lightweight construction, simulation and control.

The aim is to develop a vehicle to achieve an extremely low fuel consumption under defined test conditions. Then, every year, the resulting vehicles compete in the Shell Eco-Marathon. This establishes which of the vehicles taking part can cover the longest distance, powered by just one liter of fuel. The record is currently 5,385 km. In May 2008, at the 24th Eco-Marathon, the record was 3,382 km with Fortis Saxonia doing a distance of more than 2,500 km at less than 0.04 L per 100 km ^[1].

Vehicle design

As a result of several years of intensive research work, a thoroughly competitive vehicle is now available to the Fortis Saxonia team. The so-called “Sax 3” (Figure 1), like its predecessor, was designed and built as a three-wheel vehicle.

The two front wheels are rigid, while the rear wheel combines the steering and the electrical drive. The energy for the electric motor is produced by a fuel cell with an output of 500 W. The actual chassis comprises an extremely lightweight aramid honeycomb structure, strengthened by several layers of laminated carbon fiber. As the driver is virtually lying between the two front wheels, the front suspension of the vehicle was designed as a sort of portal axle and is connected to the chassis by two aluminum supports.

Front wheel braking is via brake disks. Figure 2 shows the design of the front axle, together with the positions of the strain gage measuring points. See below for more detail.

Fig. 2: Side view of front axle, strain gage measuring points

The link to the chassis is made with five tension struts, to maintain rear wheel steering (Figure 3). The rear axle construction is tilted round a vertical axis with the aid of a mechatronic drive, to enable cornering. The drive required for acceleration was integrated in the rear wheel as a hub motor. A rim brake is used as the brake for the rear wheel. Without its driver, the total weight of the vehicle is a mere 45 kilograms.

Fig. 3: Isometric view of rear axle, strain gage measuring points for determining tensile stress (see Fig. 5)

Task

The test vehicle with which Fortis Saxonia participates in the annual Shell Eco-Marathon is permanently under development, so that lighter and more efficient components can be developed to continuously enhance performance. The aim of the current experimental stress analysis is to precisely determine the forces, moments and accelerations that act on the front and rear axle constructions when the vehicle is being driven. The measurement results are the basis for further optimization of the overall design, making possible an increasingly precise interpretation of the components, with a view to maximizing the level of performance.

Instrumentation, processing and recording measured values

The transducers required for the test run, as well as the cabling, amplifier and software, were made available by the Professor of Solid Mechanics at the Chemnitz University of Technology (TU Chemnitz). All the mechanical quantities to be measured are acquired with electrical strain gages. The amplifier used was a QuantumX MX840 from HBM. This has 8 channels for measurement acquisition and was mounted on the back of the driver’s seat (Figure 4). The power supply for the amplifier came from 12 V battery with a capacity of 10 Ah. This was located in the vehicle nose area. This ensured that there would be a stable energy supply for the whole measurement run period. A laptop (Windows Vista Home) loaded with version 2.24 of the catman^®Easy software was installed in the vehicle to make the amplifier settings and to acquire and store the measurement data. The link between the laptop and the amplifier was made by a network cable. The measurement frequency for data acquisition was 100 Hz.

During the test runs, the laptop was beneath the driver’s legs. At the completion of each test lap, the laptop was removed during a “pit stop”, to check and save the measured values.

Fig. 4: Rear end of vehicle with QuantumX MX840 amplifier (HBM)

The main component of the front axle is a carbon tube, which picks up the torsional and bending moments caused by acceleration, braking and cornering, as well as the static loads of vehicle plus driver. Because this stress is complex, the existing loading is analyzed here. To measure these external loadings, the carbon tube was replaced by an aluminum tube with defined mechanical parameters. The two supports for mounting the tube on the chassis were also moved 15 mm further towards the center of the vehicle, to allow the strain gages to be attached to the aluminum component. These are the details of the strain gages used on the front axle:

• Torsion: XY41-3/350 (HBM), full bridge circuit
• Bending: LY13-1,5/120 (HBM), half bridge circuit

For the detailed acquisition of the component loadings of the rear axle construction, the four tension struts were prepared with XY41-1.5/120 strain gages (HBM) in a full bridge circuit (Figure 5).

Fig. 5: Tension struts with prepared strain gages in a full bridge circuit

To carry out strain gage preparations, it was necessary to totally dismantle both the front and rear axles, as the positions to be prepared were difficult to access in the installed state.

The strain gages aligned on the roughened and cleaned surface with the aid of adhesive tape were attached to the front axle by cold-curing epoxy resin and during curing, loaded by a band at 100 N. The strain gages were attached to the tension struts by hot-curing epoxy resin at 180 °C. All the strain gages were then covered with polyurethane paint (PU120) or silicone rubber (SG250).

In addition, three acceleration sensors were installed; the transducers for longitudinal and transverse acceleration each had a measuring range of 10 m/s² and a sensor for vertical acceleration of the vehicle had a range of 100 m/s². They were mounted in a bracket especially made for the purpose in the center of the vehicle, immediately in front of the front axle (Figure 6).

Fig. 6: Strain gage acceleration transducers

Test runs

The test runs were made on a section of the “Powerhall” go-kart track in Chemnitz. Defined loading conditions were initially specified to determine the vehicle’s maximum loadings. Three aspects were of particular interest. The vehicle was driven around a bend at its maximum speed to measure the transverse accelerations on the vehicle. An emergency stop from maximum speed was executed to measure maximum acceleration along the direction of travel. Finally an impact-type loading was created by running over a 5 mm diameter semicircular obstacle. The strain gages connected to half and full bridges measure the strains that occur during these driving conditions, by first recording the bridge unbalance and then calculating the loading from this. Once preparations were completed, the tension struts were calibrated by loading by weight and the acceleration transducer (statically) by gravitational acceleration. DAQ jobs previously set up in HBM’s catman^®Easy software, Version 2.24, were used for this. Firstly, this made it possible to check that the transducers were functioning properly, and secondly, the calibrating quantities could be used to convert the measured values into force and acceleration quantities.

Fig. 7: Diagrammatic view of the test track on the go-kart track

The opportunity to use the go-kart hall for experimental purposes meant that conditions could be optimized, giving perfect road and weather conditions. To meet the loading conditions, the vehicle first accelerated along the straight, so that the bend could be driven at high speed. This was followed by a turning loop, to return to the straight through the same bend, to drive over the obstacle at maximum speed and finally to do an emergency stop. The vehicle was then turned. Figure 7 shows the test track in diagrammatic form.

A total of six measurement runs were made. The reasoning behind this was that the driver should be allowed to make several passes, in order to reach the limits of the vehicle’s loading capacity. This also meant a larger number of comparable measured values, lending greater credibility to the determination of a maximum value. After each run, the measurement data were stored, the top speeds in the bend and on the straight before running over the obstacle were recorded and the software was reset for a new run. Figure 8 shows the test vehicle at full speed and Figure 9 shows it during a pit stop, when reading out and storing the measured values.

Fig. 8: Test vehicle during a test run

Fig. 9: Pit stop between two measurement runs, the author and test driver C. Gerlach

Analysis and conclusions

HBM’s catman^®Easysoftware and subsequently Microsoft^® Excel were used to process the test data. The static components of the forces were determined, as were the maximum loading values at dynamic stress. At the start of the analysis, the measured values were exported from catman^®Easy to an ASCII format, so that they could then be imported into Excel. Before further processing, the added data sets were first prepared. Both before and after each run, there were a certain number of corrupted measured values in the vehicle raw data, caused by human intervention (modification, working on the laptop, etc.), and these were removed. In the result, each series of tests contained a recording of the load data during a measurement run, to determine the dynamic characteristic values and in also for five seconds at the start and at the end, in order to define the static force values. The results from the preceding calibration measurements were used to work out the resultant force and acceleration quantities from the recorded transducer voltage ratios. The bending and torsional moments at the aluminum tube were defined in accordance with equations (1) to (4), similar to the way specified in ^[2], among others.

The bending moment is produced by a half bridge circuit from

with the axial geometrical moment of inertia following from

To calculate the torsional moment,

with

applies as the polar geometrical moment of inertia.

The quantities occurring in (1) to (4), are summarized below.

The four graphs shown below are examples taken from the vast amount of available measurement data. It is easy to relate to the track shown in Figure 7 from the curves of the vehicle’s transverse acceleration (Figure 10) and of the forces (Figures 12 and 13) on the long tension struts. Negative acceleration values represent left-hand bends, whereas positive values correspond to right-hand bends. A comparison of the stress curves of the two long tension rods shows not only passages with an opposite response (taking the bend, seconds 40-57) but also passages with the same response (driving straight ahead, seconds 57-65). Figure 11 clearly shows that torsional moments are only initiated in the front axle during any kind of braking maneuver. Overall, these curves not only demonstrate the high quality of the preparations, but also the proper functioning of the transducers in the course of the tests.

 

Fig. 10: Transverse acceleration curve

Fig. 11: Torsional moment curve

Fig. 12: Force curve at long, left-hand tension rod

Fig. 13: Force curve at long, right-hand tension rod

With the aid of these results, it is now possible to make an accurate statement about the efficiency of component performance. For example, there is potential for improvement in the design of the struts for the rear axle suspension. Because the forces here are relatively low, the rod cross-sections can be considerably reduced. In the front axle area, the best solution is currently to use the carbon tube, as this component ensures maximum rigidity for the combination of torsion and bending around several axes. However, the tube is not required for the area between the two aluminum supports. No torsional moments are initiated here and with regard purely to bending, narrow I-beams have considerably more rigidity with the same crosssection. Further weight-savings are possible by re-designing the brake system because good deceleration is of secondary interest in this project.

Literatur

[1] www.fortis-saxonia.de
[2] Hoffmann, K.: An Introduction to Measurements
Using Strain Gages, Darmstadt,
Hottinger Baldwin Messtechnik GmbH (1987).