Fullscreen HTML5Med Res (??MB)HTML4

One of the Larrousse-Lamborghini cars in action at the Adelaide Grand Prix. A sophisticated traction control system was used to allow greater acceleration and cornering speeds and to improve the start-line performance. Traction Control Last month, we examined the traction control systems now used in some road vehicles. This time we look at how the technology has been used in motor racing. During 1993, electronic aids were permitted in the highest form of motor sport: Formula 1. This meant that, together with electronically-controlled gearboxes and active suspension, electronic traction control was used. In addition to preventing unwanted wheel-spin during normal acceleration, the system was also used during Grand Prix starts to give the best possible results. All electronic driver aids were banned from the 1994 season onwards and so the technology was seen largely for just the one year. PART 2: By JULIAN EDGAR 14  Silicon Chip The system examined here was fitted to the Lamborghini V12 engine of the Larrousse-Lamborghini cars, driven by Philippe Alliot and Erik Comas. It was developed by Bosch Motorsport in conjunction with Lam­ bor­ghini Engineering. System requirements The requirements of the traction control system were to control slip with precision; capable of subtle levels of control, yet able to be quickly recalibrated. It also needed to be easy to use, allowing driver interaction, yet not being driver dependent. Engine power was controlled in such a way that drive wheel slip was limited to a value which ensured maximum straight-line acceleration and cornering stability. Unlike normal Fig.1: the appropriate goal value of wheel slip was dependent on car speed and throttle position, the gear being used and the lateral (cornering) acceleration. road-vehicle traction control systems, the system did not use braking to control wheel-spin but relied entirely on engine torque control. This was achieved by progressive injector cut-off. System details A closed loop PID (proportional, integrating, differentiat­ing) controller was chosen to minimise racetrack setup of the traction control algorithm. In addition, fuzzy logic control elements from racing ABS systems were added. This control ap­proach gave the following set up advantages which were independ­ent of tyre wear characteristics and independent of the slip-goal target value. Only a simple ‘wet/dry’ driver-selectable goal-offset switch was required. The digital control process was handled by one of the existing engine management microcomputers which, as well as using engine sensor information, Fig.2: a PID controller was used to calculate the desired was fed with speed data from each wheel. per­centage reduction in engine torque output to reduce The procedure taken for the calculation of the wheel slip to an optimal value. rear wheel slip is shown in Fig.1. The basic goal value was derived from a map using the functions of car speed and throttle position, with an offset provided by the cockpit wet/dry switch. The value derived from a gear-dependent curve was added and this in engine torque, compensated by the current gear ratio is multi­plied by a factor based on the lateral acceleromand the differential ratio. eter input. The calculation of wheel slip was made by Should the driver have sensed that slip was occurring comparing the speed of each of the rear wheels with the and had lifted his foot during traction-controlled slip, reference speed of the car, which was derived from the problems could have occurred. To counteract this, a drivfront (non-driven) wheels. er-initiated torque reduction was also compensated for as The deviation between the desired slip and the actual a function of engine RPM and throttle position. slip values was fed to the PID controller, as shown in Fig.2. The calculated engine torque reduction was convertThe gain and time delay factors of each of the P, I and D ed to a corresponding injector cut-off pattern by dyna­ components were stored in maps as functions of the car mometer-derived data held in a 24-point curve. The speed/throttle position operating points. The output of encoded steps of injector shut-off ranged from “half” a the PID controller was the per­centage reduction required cylinder (one every other 720° cycle) to a full 12-cylinder April 1996  15 10 5 40 30 20 OSITIO N 70 60 50 THRO TTLE P CYLINDER CUT-OFF NUMBER 90 80 10 0 0 12000 10000 8000 6000 4000 0 2000 Fig.3: the maximum number of injectors which could be cut off was dependent on throttle position and engine RPM. This provided safety against engine stalling should the PID controller be programmed incorrectly or if part of the system failed. cut-off. An absolute limit calibration was incorporated, fixing the maximum number of cylin­ders which could be cut off at a given RPM and throttle position. This acted as a safeguard against engine die-outs at low RPM and also allowed rapid recalibration of the PID controller without upsetting overall vehicle dynamics. Fig.3 shows this overall cut-off limiting calibration. A completely separate algorithm was used during the stand­ing starts which occur in this form of racing. It used two dis­tinct control strategies. In part 1, the system allowed the driver to maintain full throttle prior to clutch engagement, with the ECU holding the engine RPM at the desired level. Once the clutch was engaged by the driver and the car exceeded a certain speed, part 2 of the system was enabled. This modulated the continued full throttle by means of injector cut-off, allowing control of wheel slip to the desired level. Normal PID control was activated once the car had reached a second, higher speed threshold. Fig.4 shows the telemetry record from a Grand Prix start. Note that the throttle is held fully open for the majority of the time and the rear wheel speed increase as the clutch is engaged in part 1. In part 2, a constant slip ratio is maintained, as indicated by the difference in the front and rear wheel speeds. Testing & development Calibrating the system to give the optimal level of slip proved very difficult. This was firstly because only limited traction control testing was undertaken, with the testing com­pleted only during normal chassis set-up procedures. Second, the preferences of the THR RPM CUTOFF PATTERN PART ONE REAR SPEED PART TWO FRONT SPEED Fig.4: the Grand Prix ‘start’ strategy, as shown by the telemetry data from an actual race. Note the small amount of wheel spin achieved, even though the throttle is being held fully-open most of the time! 16  Silicon Chip THR GOAL SLIP ACTUAL SLIP REAR WHEEL SPEED CUTOFF PATTERN Fig.5: the telemetry record from a wet track, with the system programmed to be very responsive to wheel slip. two drivers using the system varied: the amount of slip which suited one driver did not always suit the other! Extensive testing on a smooth, dry track revealed that 4-6% slip gave the best results but the engineers were unsure whether this would apply to all racing circuits. But while 4-6% longitudinal slip gave good acceleration, this amount of slip during cornering slowed the car. Although a lateral accelerometer input was available, it was found that a driver would not exceed a certain throttle threshold unless the car was within his ‘comfort’ yaw zone and so throttle position was able to be used to predict when more or less system intervention was required. However, driver comment and track side observation revealed that the optimal slip level wasn’t the test-derived 4-6%. In fact, the slip level which gave the best results varied from 1215% at low speeds, to less than 2% at very high speed. Rather than the percentage slip being the relevant factor, it was concluded that a slip which corresponded to a difference in wheel speed of 4-5km/h between the front and rear wheels at 90km/h was the critical value. This relative difference in rotational speed gave the car its characteristic feel in yaw and was what the driver was actu­ally feeling and describing. Once this was understood a spread­ sheet program was created to allow the new calibration of delta speed to be converted into percentage slip, FRONT WHEEL SPEED allowing the continued use of the existing software. Results Fig.5 shows the system, programm­ ed to be very responsive to slip, in action on a wet track. The car speed is shown by the “front wheel speed”, with the difference between front and rear wheel speeds being the amount of slippage, highlighted by the “actual slip” line. It can also be seen that when the throttle is closed briefly, slip ceases to occur and so momentarily drops below the “goal slip”. Acknowledgment: thanks to the Society of Automotive En­ g ineers for permission to use material from the “SAE Australasia” journals of September/October and November/ SC December 1995. Especially For Model Railway Enthusiasts Includes 14 projects for model railway layouts, including throttle controllers, sound simulators (diesel & steam) & a level crossing detector. Price: $7.95 plus $3 for postage. Order today by phoning (02) 9979 5644 & quoting your credit card number; or fax the details to (02) 9979 6503; or send cheque, money order or credit card details to: Silicon Chip Publications, PO Box 139, Collaroy, NSW 2097. April 1996  17