Vehicles are designed making a compromise between comfort, cost and handling. The frequency response to steering angle inputs allows evaluating the handling properties of the car.
There are two different ways to evaluate the frequency response of the vehicle: by using a mathematical model or by testing.
The first approach is usually used in the first steps of the design process, allowing predicting performances.
If a more comprehensive analysis is needed, the testing becomes necessary. For the shake of this assignment both analyses were considered and finally compared.
There are several methods to test the frequency response of a vehicle. Some of them are explained by the International Standard Organisation (ISO) in the ISO 704:1995(E):’Road vehicles – Lateral transient response test methods’.
– Step input
– Random inpu
– Pulse input
– Continuous sinusoidal input.
These test objectives is to determine the transient response behaviour of a vehicle. Further information about some of these tests can be found in SAE papers 930828, 930829 and 930830.
Also it is possible to perform the test with a steering wheel torque as input. This option was rejected due to two considerations:
– Four wheeled vehicles are principally controlled through steering wheel displacement.
– The steering wheel torque couldn’t be measured accurately.
For this assignment a sinusoidal sweep steering angle is used as input with constant speed test. Ford Motor Company under the name “Random Steer Test” uses similar test. In the test, time response signals (lateral acceleration, yaw rate and roll angle) to input signals (steer angle) are collected. Applying Fast Fourier Transform (FFT) techniques, the time domain system is transformed into the frequency domain. Then the output of this test is compared with a model generated in Autosim.
This final comparison between the linear model and the measured data can be possible due to the lateral acceleration being less than 0.35g. Up to this range, the vehicle has linear behaviour, even though, it has many sources of non linearity.
It is necessary to highlight this point because the model is linear and the frequency domain studies assume linear response.
The assignment was divided into four groups (see project management chapter): Theoretical, parameters measurement, instrumentation and mounting in car and signal processing.