MIRA Coursework Brief

Background

Development of models for the use of predicting powertrain performance is an important process in the design of a modern vehicle. Once a model has been developed it is essential to validate it against experimental data.

Over the last few weeks you have developed a MATLAB Simulink vehicle and powertrain simulation model that can be used to estimate the straight-line performance of a passenger vehicle, this model will be used in the coursework. Experimental data will be collected from test track experiments at MIRA and this will be used to both tune and assess the validity of the model as well as illustrate factors affecting the vehicle’s acceleration performance.

Coursework assignment

The objective of the coursework assignment is to obtain real-world experience of straight-line vehicle performance characteristics and to use a simulation model to predict the vehicle motion under such conditions. This will involve;

• Measurement of empirical data from the vehicle CANBUS.
• Graphical examination and discussion of the measured data.
• Tuning of key simulation model parameters.
• Comparing the model predictions and measured data.

Test vehicle

The test vehicle is a Ford Focus ST. Information about the speed of the vehicle is obtained from the vehicle’s CANBUS and GPS data. The driver’s inputs of clutch and throttle pedal position are measured and these are used as inputs to the powertrain model. All variables are recorded on a laptop.

Test procedure

The coastdown tests will be conducted on the parallel straight test tracks, these consist of two dual carriageways, one mile long straights, linked with banked turns. The straight sections have zero gradients.

You are responsible for devising a test plan and communicating this on the day of the test to the driver. A suggested partial plan for the coastdown test is shown below, this does not include all of the tests required and serves only as an example. Two test runs would normally be collected in a single result set, with the vehicle moving in opposite directions.This allows you to average out the effect of wind.

• Prior to testing the vehicle is weighed.
• The vehicle and tyres are warmed up by driving a couple of laps and the testing is then started.
• A speed of approximately 100kph is reached on the turn and the vehicle quickly accelerated up to approximately 115kph once on the straight.
• The gearbox is put into neutral and the vehicle coasts.
• Immediately after neutral is selected the data collection is started.
• At the end of the straight, the data collection is stopped and the test is then repeated in the opposite direction.
• The vehicle should be weighed on completion of all tests.

After completion of the coastdown tests, standing start acceleration tests should also be carried out as well as additional tests that you specify to identify other parameters in the model that you are interested in.

After completing all of the tests, the recorded data should be transferred to the analysis PC/laptop for viewing and subsequent model validation. It is essential that the different test runs are compared and the most suitable test runs selected for further analysis. To help you get used to the format of the data that will be obtained, an example data set is provided here.

In devising your test plan think carefully about which model parameters you will be using the data to tune. For example higher speed coastdown tests can be used to tune the aerodynamic coefficients (as these will be dominant at higher speeds). At lower speeds the rolling resistance terms will be dominant.

There are many more tests that can be done in the time available (with some careful planning). Good data collection will enable many opportunities for tuning the model response, for example you may wish to consider collecting data to support tuning of some of the following;

• Clutch biting point (as a function of pedal position)
• Drivetrain oscillation (frequency response)
• Engine torque production
• Tyre adhesion characteristics
• Clutch logic
• Engine inertia
• Traction control on launch

Model tuning

Coastdown tests. These act as a precursor to the acceleration tests, allowing appropriate tuning of the drag parameters. You should compare simulated and measured vehicle speeds, and tune the $A_d$, $B_d$ and/or $C_d$ values to get the best match. It is also possible to change drivetrain friction parameters and/or the vehicle frontal area (provided it remains physically realistic).

Acceleration tests. You should aim to get the best correlation between vehicle/wheel speed in the simulation, and measured wheel speed from the MIRA tests, also aim for good correlation between measured and simulated engine speeds. The most likely discrepancies between the results will probably relate to the tyre adhesion characteristics and clutch biting point. Adjust these parameters in the simulation model to match the measured data as accurately as possible. Other sources of error for example could be;

• Non-linearity of the clutch and/or throttle pedal inputs.
• Engine torque scaling.

These parameters can also be tuned to achieve a better match, but the tuned parameters must remain meaningful. Always try to understand and be prepared to explain the reason for changes.

Preparation for MIRA week

Before the test week, you should attempt a model tuning dry run (for acceleration tests). Tune the model parameters to obtain a single model that gives best performance on average for the calibrated test files provided for the lab exercise. Note that the actual MIRA tests results may vary significantly from the sample data and as such the tuned values will differ (i.e. you will need to tune the model to the actual measured data). The main purpose of the lab exercise is to gain familiarity with the software and parameter names.

In groups, you should decide how to instruct the driver to drive during the acceleration tests. Within reason they can be requested to perform any safe straight line manoeuvre involving the accelerator and clutch. Please note however, that the condition of the clutch and tyres will be monitored and it may be necessary to restrict manoeuvres involving excessive wear (excessive clutch slip manoeuvres and wheel spin are not allowed). All tests to be undertaken should be written down clearly and handed to the driver before starting the test. It is unlikely that amendments to the test plan can be made during the testing. Note, the driver of the vehicle is unlikely to know anything about the model you have created and will not be in a position to help you collect ‘correct’ data, this is your responsibility.

Report

The report must be written in a formal, technical style. The maximum length of the report is 10 sides of A4 including the title page and appendices. You should include the following sections;

1. Introduction
2. Vehicle Test Plan
3. Results
4. Model Tuning
5. Simulation Results and Discussion
6. Improving the Results
7. Conclusions
8. Bibliography

Your report will be assessed as follows;

If you would like more guidance on report writing please have a look at this advice sheet from the library.

Collaboration

The testing and data post-processing for this work will be done in groups. The primary tuning of the model can also be carried out (and shared) within your group. Please note however, that fine tuning of the model and report writing must be carried out individually.

---