Embedded Files
Background
Background
Formula SAE is a yearly intercollegiate competition consisting of static and dynamic events in which students defend and test their design of a formula-style racecar. UC San Diego’s Formula SAE team, Triton Racing, has been improving the power output of their engine year after year, making it difficult for student drivers to control the car. Implementing a traction control system will assist an amateur driver in maximizing the car’s performance by electronically controlling the engine output to ensure tire wheel grip remains in the optimum range throughout various driving scenarios.
Figure 1: TR-19 Car
Objectives
Objectives
The objectives of this project are to design, implement, and tune a traction control system which produces measurable improvements in car accelerations and lap times. The traction control system will work to increase traction at the wheels by attenuating the engine power output when excessive slipping occurs. The system will consist of a custom designed fuel control circuit to prevent combustion and reduce the torque produced by the engine. The system's on-board sensors will activate this fuel control circuit when they detect excessive slip at the rear, driven wheels. The system will be designed and installed on Triton Racing's most recent car, TR-19.
Design
Design
The Traction Control System design consists of four wheel speed sensors, a triple-axis accelerometer, and one steering angle sensor. All six of these sensors continuously send data to an Arduino Mega microcontroller, as seen in Figure 2.
Figure 2: TCS Schematic
The microcontroller processes data from these sensors to compute the current state of wheel slip. Vehicle simulation results and tire data are used to determine the reference optimal slip condition for the tires. If the current slip exceeds the optimal amount, the microcontroller will then actuate a fuel control circuit seen in Figure 3 to attenuate engine power and reduce wheel slip.
Figure 3: Fuel Control Circuit
The four wheel speed sensors are mounted to the car via machined aluminum brackets and vendor provided hardware. These brackets are then bolted to the car's bearing retaining rings located within the wheel assembly. The rear wheel speed sensor assembly is shown in Figure 4.
Figure 4: Rear Wheel Speed Sensor CAD Design
The DC-DC converter, Arduino, and fuel control circuit are all housed within 3D printed PLA casings shown in Figure 5. These casings provide protection from rocks and debris and also provide strain relief for wires. The three casings are attached to the floor panel under the driver seat with Velcro strips.
Figure 5: Electrical Housing CAD Design
Final Product
Final Product
Figure 6 shows the front and rear wheel speed sensors mounted to the car, and Figure 7 provides images of the electrical component setup in the car's cockpit under the seat.
Figure 6: Wheel Speed Sensors Installed on Car
Figure 7: System Installed On Car
Testing Results
Testing Results
Before installing the traction control system, baseline data was collected to demonstrate the need for a traction control system and provide a quantifiable point of comparison. Preliminary data showed that excessive slip occurred for over 30% of time spent accelerating, showing significant room for improvement.
Testing of the first iteration traction controller showed that the fuel cut power attenuation concept was effective but needed tuning in order to maximize performance. Cutting a fuel supply to a single cylinder was shown to reduce excessive wheel speeds during a standing-start launch by around 25% as compared to an unassisted launch. Double cylinder fuel cuts reduced excessive wheel speeds by 40%, providing a significant improvement to the initial portion of vehicle acceleration. Due to issues with sensor noise and controller parameters, however, the first iteration controller actually resulted in slower acceleration times than when the car runs unassisted. Unassisted acceleration times averaged 3.789 seconds, and the single and double cut control modes produced average acceleration times of 3.980 seconds and 4.500 seconds.
Testing the second iteration controller yielded favorable results. After examining data for all controller settings, it was found that the wheel speed limit of 5 Hz, or about 14 mph, produced the fastest accelerations. Unassisted acceleration runs from that testing session from produced average 0-60 mph acceleration times of 4.177 seconds with a standard deviation of 0.200 seconds. This traction controller enabled the driver to achieve faster and more consistent times of 3.896 seconds with a standard deviation of 0.030 seconds.
Project Video
Project Video
156 Vid3.mp4Page updated
Report abuse