The Prosig sponsored Dalmeny Racing Formula Ford Team, whilst contesting the UK Formula Ford 1600cc championship, suffered several minor structural failures on a particular part of an exhaust pipe mount. During an important race staged at the Silverstone racetrack, the week before the British Grand Prix, Dalmeny Racing suffered the first serious and complete failure of this component.

Mechanical Engineer and Dalmeny Racing Team Manager Brian Wren commented, "We can see no reason for this normally sturdy component to suddenly fail. We have had no problems with this component for long periods of running, then almost for no reason a catastrophic failure." Initially there appeared to be no reason for the failure, the component simply fails at what seemed to be random intervals.Armed with the challenge of diagnosing possible causes and solutions to this problem, Prosig dispatched a team of engineers, armed with a Prosig P5650 and DATS software, to the Dalmeny Racing head quarters to investigate.After a brief survey of the damage the Prosig engineers made a preliminary assessment: "Our initial thoughts are that the exhaust itself might be resonating at particular engine speeds, causing some shear forces in the mount material leading to cracking and eventually failure. We need to understand how the entire structure is moving at different frequencies. That way we should be able to determine where stresses may be occurring."In order to prove the hypothesis a series of simple tests were instigated which would lead to frequency domain animations. This involved finding the frequency response at each measurement position on the structure so as to reveal the motion of that position at different frequencies. In effect, the team would be able to actually animate the exhaust pipe and analyse exactly the type of motions that were occurring at different frequencies and thus deduce which engine speeds which are contributing the most to problem.Initially a single axis accelerometer was attached to the rear of the exhaust, the position was chosen to be as near to the mount as possible. The position was further dictated by aligning one of the axes of the accelerometer to be parallel to the direction of load of the mount, with the final axis perpendicular to this. This is shown in Figure 1 & Figure 2.

Figure 1: Accelerometer mounted on exhaust	Figure 2 : Accelerometer mount from above

The next step required the excitation of the exhaust pipe in order that the response could be measured and recorded via the accelerometer. For clear repeatability and accuracy, each excitation has to be carried out three times at each point and then the average of the three readings is used.

All of the measurement, averaging and analysis is carried out by the DATS Hammer Impact Software. This greatly speeds up the time needed to test. A typical screenshot of the DATS Hammer software is shown in Figure 4.

Figure 3 : Using an instrumented hammer to measure vibration response	Figure 4: The DATS Hammer Impact software at work

The hammer used for excitation is a rubber-based hammer with a force meter inside the head. The force meter is quite simply a piece of piezo electric crystal. When under compression a small voltage is generated. The Prosig P5650 is used to measure this force. This can be used to measure the amount of force applied when hit against the exhaust. When used in tandem with the accelerometer, it is possible to compare the force and response in the frequency domain at different point on the structure.The first set of excitations were to be carried out perpendicular to the direction of load in the mount, these were marked with masking tape for accuracy and repeatability. To give good resolution to the animations twenty points along the exhaust were used. As shown in Figure 3.Once these measurements were complete, another set of results was collected by taking measurements perpendicular to the original set of readings, thus both the X-axis and the Y-axis can be deduced.In order to completely understand the exhaust and the mount it was necessary to complete the tests on the mount itself as well. The markings for these tests are shown in Figure 5. This would enable the animation software to display the exhaust and the mount both simultaneously and separately. This will allow the simulation of the two components in one model, thus the movements of the two can be visualised.

Figure 5 : Exhaust mount ready for testing

Whilst on site, and with the car instrumented, the engineers also took some other noise and vibration measurements. These, however, are not relevant to the current problem and are the subject of a later article.

With all the measurements done and reviewed quickly on site, the Prosig engineers were happy that they had collected a good set of readings. Now it was back to base for some analysis. A mobile system, such as the DATS/P5000 (or P8000) combination, can be used to take readings and analyse results "on site". However, once the data had been safely captured the car could be freed up for other test and development work.Now it was time to look at the data from the hammer tests and see what could be gleaned. The first step was to create a model using the DATS Structural Animation software. A detailed version was created first.This is ideal if you want to completely visualise the entire test piece. It can also be used to view visually stunning colour plots such as that in Figure 7.

Figure 6: Detailed model of exhaust

Other interesting views provided by the DATS Structural Animation software include one that shows arrows indicating motion or acceleration. Another animates the model, but leaves a persistent "trail" of previous motion. Examples of both are shown in Figure 9 and Figure 10.

Figure 7: Motion displayed using colour

Whilst this produces nice pictures, shows off the software and is good for reports to management, the best way to really see what was going on was to reduce the exhaust and its bracket to a simple model of lines.

Figure 8: Exhaust modelled as simple lines

The transfer functions collected from the hammer tests were then fed into the model. The transfer function graphs showed the largest vibration at around 325Hz. Animating the model at this frequency made it clear that the problem lay in the stiffness at the bottom of the bracket. Whilst the exhaust itself and most of the bracket was free to move, the bottom of the bracket was rigidly fixed to the gearbox. Thus all of the stresses were concentrated at this point.The dynamics of motor racing mean that a racing car spends the majority of running within a particular and often small engine speed range. In effect, this means that the frequencies that seem to be causing the problem don't occur very often or for extended periods. It is only under certain circumstances, like slowing down laps or in engine testing that these frequencies occur for longer periods. Therefore the effect of these frequencies seemed initially to have no pattern, but when the correct engine frequencies that cause this particular excitation had been found it was possible to diagnose how why and when the problem had been building and occurring.

Figure 9: Structural Animation display using motion/acceleration arrows


Figure 10: Structural Animation display using persistent display

Simply uprating the strength and weight of the component as Dalmeny Racing had been doing was having very little effect. The design of the component required some relatively minor changes. One option would have been to add one or more extra struts to the bracket. This would have had the effect of making the bracket more rigid and therefore reducing the stresses. However, it was felt that this would simply delay the inevitable. That is, the same problem would have occurred, but less often. There was also some concern that a modification of this sort might contravene the fairly strict regulations governing the formula. A better solution was proposed. By adding a small rubber bush to the lower mount, the whole exhaust assembly was free to move without subjecting the bracket to the same stress. The rubber bush could much more easily absorb the vibrations. With this simple modification the problem was fixed and weight saved on the original plan of adding a bigger heavier bracket.

Several different mount materials were selected and a series of tests planned. The results of these tests will be discussed in a future article.

James W Wren
January 2005

Some related links:

Dalmeny Racing - www.dalmeny-racing.co.uk

Circuits

Thruxton - www.barc.net/thruxton.htm
Brands Hatch - www.octagonmotorsports.com
Castle Combe - www.castlecombecircuit.co.uk
Silverstone - www.octagonmotorsports.com
Donington - www.donington-park.co.uk

Organisations

BRSCC - www.brscc.co.uk