Circuit and Safety Analysis System (CSAS)

Predicting the trajectory and velocity of a racing car when it is driven at the limit within the confines of a racing track, is now the subject of a great deal of analytical work by almost all teams involved in racing at all levels. However, predicting the trajectory and velocity of a car once the driver has lost control of it has not been something the teams have devoted a great deal of time to. This can now also be analyzed though in the same sort of detail, to assess the safety features of the circuits on which it is raced. The two tasks are very different, and the FIA had to start almost from scratch when it set out to develop software for its Circuit and Safety Analysis System (CSAS). The system is up and running and being applied to new and existing circuits. The development and a description of the software were the subjects of a paper at the SAE Motor Sport Engineering Conference, in Dearborn last November (SAE Paper: 2000-01-3573 Circuit and Safety Analysis System (CSAS). F. Corcelle and P.G. Wright, F?d?ration Internationale de l'Automobile).

Fig.1. Examples of straight trajectories. © Inside F1, Inc.

Fig.2. Examples of all possible trajectories. © Inside F1, Inc.

Fig.3. Stopping distances in the run-off area, highlighting points where the run-off is inadequate to stop the car. © Inside F1, Inc.

Fig.4. Residual velocity, perpendicular to the boundary of the run-off area. © Inside F1, Inc.

Fig.5. Residual velocity, perpendicular to a 2-row tyre barrier, after impact with it. © Inside F1, Inc.

The last two decades have seen a steady build up of the R&D effort going into vehicle dynamics modeling, particularly by those teams that design and develop cars as well as race them. The pace of development has been set by the availability of powerful PC's, the generation of vehicle and component data, and the supply of suitably qualified graduates to carry out the work. Their task is to be able to model and predict the effects of every nuance of aerodynamic, tire, engine, damper etc., characteristic on the speed of their car at every point on a given circuit. The detail in the model will only be limited by available dynamic characteristics and track data, and will require a driver model to complete the picture. However, they are only interested in the performance of the car while the tires are in contact with the tarmac, and the driver is operating them at or below their peaks.

The FIA, on the other hand, starts to be interested in what happens when the driver exceeds the limit and is unable to recover control of the car, or when something breaks and the computer model almost literally falls apart. Knowledge of the speed of the car all around a circuit is needed, but the precise speed differences due to small improvements in some car characteristic have little affect on the outcome of this analysis. Major changes in lap speeds, due for instance to the effects of tire competition or regulation changes are relevant, and so CSAS has a lap simulation as its core, to generate speed profiles for any circuit and any class of racing car. It is a fairly elementary simulation compared to those in use for performance prediction by teams, but is regularly updated with engine power curves, Pacejka tire coefficients, typical aerodynamic characteristics, and weight changes. Checks that the speed predictions are sufficiently accurate can be made by comparison against speed data supplied from a typical car.

Circuit details are supplied in AutoCAD. This software was chosen because of the ease of adding modules to perform the CSAS-specific operations, and also because the majority of circuit maps are supplied by the circuit designers in this format. CSAS is run via the AutoCAD interface, with additional tool bars corresponding to the CSAS-specific applications. Circuit information is in multiple layers, e.g., left side of track, right side of track, curbs, run-off areas, access roads, removable barriers, permanent barriers, being the most relevant. The track edges can be modified using the AutoCAD drawing tools - the addition of a chicane is simply a few click-and-drag operations of the mouse! The operator draws the racing line on the track (an automatic routine for doing this is being investigated, but the manual approach is currently preferred as knowledge of whether drivers clip curbs or avoids a bumpy section of track, provides a better match of speed profiles) and selects the calculation of the speed profile. Generally, the speed is calculated every 3 meters around the track, which provides adequate resolution, at each of these points a prediction of the trajectory of an out of control car is made.

A driver's natural reaction, once he realizes that he has no further hope of regaining control, is to stamp on the brakes and bring the car to a halt before hitting anything. A car with its wheels locked up, whether it is travelling forwards, backwards or sideways, or spinning, will tend to travel in a straight line unless it hits something (Fig.1). Thus, the most likely trajectory is a straight line, tangential to the racing line at the point control is lost; all circuit safety criteria are currently based on this trajectory assumption. However, if the driver does not give up and tries to catch the car while it spins, or to influence which way it goes, or if a component failure substantially takes over the steering of the car, there is a possibility that some lateral forces will be generated by the tires (they could be up to 4g on a Formula 1 car), in which case the trajectory will be curved, just as if the car was cornering. However, the curved trajectory will probably not follow the curve of the track (Fig.2).

These "unpredictable" trajectories are the hardest to plan for, without lining the whole circuit with run-off areas and barriers. In many cases e.g., if a wing fails on the straight which causes the car to turn into the wall lining the straight, the car cannot accelerate to a high speed perpendicular to the wall, and the speed is scrubbed off by sliding along it. Spectacular though this may be, this sort of accident tends not to lead to high impact decelerations or injuries to the driver. However, in a high-speed corner, the car can end up going off in a direction that until then, has not been predicted and so is not protected. Zonta, in the accident in Brazil in which he received leg injuries, tried to collect his BAR after he lost it on a bump in the 4th-gear Ferradura and struck a section of Armco instead of the tire barrier erected to protect cars in that corner. He was not meant to hit the barrier at that location. CSAS is being developed to be able to predict the impact velocity for any possible trajectory.

Another example of unpredictable trajectories occurred on the Circuit de Catalunya, during the Spanish GP in 1997. Morbidelli accelerated his Minardi out of the pit lane and lost control of it as he joined the track, possibly due to the speed limiter cutting out suddenly. He accelerated across the full width of the Start/Finish straight into the concrete wall, fortunately without collecting anyone else travelling at top speed on the straight. He hit the wall head-on at just under 50kph, performing a near perfect FIA frontal crash test!

Having established the speed at any point on the track, CSAS calculates the trajectory of a car leaving the racing line and the distance travelled along it. The path of the car is initially on the track, subsequently on a run-off area, if one exists, and may finally impact a barrier. The boundaries of all these features are set up from the circuit plans, in AutoCAD. The circuit criteria guidelines have been established such that under normal or average conditions, the car will stop before it reaches a barrier. Under abnormal conditions this may not happen, and in certain locations on circuits it may not be possible to provide adequate run-off - Monaco, or indeed any street circuit, is the classic case of this - hence the need for barriers. The deceleration characteristics for an out of control car on the track and on any type of run-off area are set in CSAS and may be quite complex relationships based on speed. One of the purposes of fitting Accident Data Recorders (ADR) to Formula 1 and Formula 3000 cars is to gain real deceleration data. With data gained over the last four years, it has been possible to analyze it statistically and derive "normal" characteristics for wet and dry tracks and for gravel beds. These characteristics are used in CSAS to determine how large the gravel beds need to be and to establish the likely impact velocity with a barrier, where it is not possible to install an adequate run-off area. CSAS plots the trajectories, and the ends of these lines form the desired limits of the run-off areas, which can be compared with existing or planned boundaries. Discrepancies show up immediately on the screen (Fig.3).

Faced with sections of run-off areas that do not stop a car before it reaches the edge of the area, the circuit designer has a number of options. If he cannot extend the run-off, one option is to modify the corner to reduce the speed, however, the critical trajectories are often those of a car that loses control under braking, when it maybe necessary to reduce the top speed on the preceding straight - the result is often the unpopular chicane. Alternatively, barriers can be placed along the critical edges of the run-off area. CSAS calculates the impact velocity, perpendicular to the boundary, in the absence of a barrier (Fig.4). Barrier characteristics have been measured for a number of barrier configurations, particularly for a variety of tire barrier arrangements. Conservative characteristics based on the test results are used in CSAS to calculate the resultant velocity of the car after it has penetrated the barrier i.e. the velocity the car will impact the solid boundary behind the barrier (Fig.5). This velocity or, to be more precise, the residual energy in the car, is what the crushable structures on the car will have to absorb without injuring the driver.

One issue that CSAS addresses is whether the critical case for stopping a car is under wet or dry conditions. In the dry, initial speeds are higher but on-track deceleration is greater than in wet conditions. Wet or dry, the gravel beds perform pretty well the same. Based on the data available to date, the indication is that the critical case is under dry conditions.

The worst scenario for any safety engineer is when a car "flies". Whether it is a big sports or GT car, with excessively pitch sensitive aerodynamics, or an open-wheeled car touching wheels with one ahead of it, if a car leaves the ground it is almost impossible to provide a means of decelerating it. It will decelerate due to aerodynamic drag, and CSAS can assess this case provided the drag characteristics are known as the car tumbles through the air. Gravel beds that cause light cars with wide tires to skip through them (a sort of "ducks and drakes" effect) do not seem to exhibit very different overall deceleration rates from beds where the car stays in contact. Although the deceleration is reduced while the car is in the air, it is much higher when it lands and digs in, and the average deceleration is very much the same.

CSAS has facilitated the synthesis of the results from a number of safety R&D programs that are gradually putting motorsport safety on a sound scientific basis. It uses the actual speed of the cars at any point on a circuit, representative deceleration rates on- and off-track, and tested barrier performance to size and specify circuit safety features. Changes to the specification of the cars, particularly those that increase top speed or cornering speed, and changes to the layout of tracks can be monitored for their effect on the size of run-off areas and barrier specifications. Any class of car can be evaluated by inputting its performance parameters to the lap simulation and obtaining a speed profile, such that the grading of circuits and their suitability for particular classes of racing can be studied.

The development of CSAS is ongoing. Routines to facilitate and speed up the application are being studied and the database for the performance of the various circuit safety features is continuously updated and added to, to ensure that any variations in the deceleration parameters, e.g. due to an extra tire groove, are taken into account. It is an invaluable tool at the design stage of new circuits, avoiding much of the need to revise either track or run-off areas after the circuit has been built, and is providing detailed insights into how existing circuits can be upgraded in the continual quest for greater safety.

Follow grandprixdotcom on Twitter
Print Feature