After 9 months of demanding work, we had the pleasure of showing our racecar to friends, family, sponsors and everyone else interested in our accomplishments at our unveiling event. Some of you were, mistakenly, thinking that the work was mainly finished, and we were ready to go off to competition, competing against the finest engineering students Europe has to offer. In reality we still had a long way to go. In eight weeks, a non-driving composition of nice-looking materials had to be turned into the fine-tuned machinery seen at competition. Today you will get some insight into the process ensuring this happens.
One of the main elements of the test season is taking the racecar out on track and performing the same events that will take place during competition. Systems and parts on the car, which previously only have been tested in simulations and controlled conditions, suddenly interact with all other elements of the car, as well as less than optimal track conditions. Even though best judgement design decisions are made at the start of the year using the available data and experience from previous years, we still experience failures every year without exception. This period is therefore crucial to uncover potential problem areas that must be fixed before competition.
Considering we have hundreds of parts on our racecar, several being electrical, we have no chance of uncovering problems using only the naked eye. We instead use our self-developed telemetry software, Revolve Analyze, combined with Radionor’s communication equipment to receive, and visualize data from hundreds of sensors monitoring the vehicle status live. This ensures we can monitor safety-critical systems, minimizing the chance of critical failures happening unnoticed. When errors occur, we can look through the data to figure out why it happened and what changes must be made to prevent it from happening again.
To give an example there are around 300 sensors inside our self-developed accumulator, which measure cell voltage and temperature for each individual cell. With this data we can monitor the accumulator’s condition live. If we observe large deviations between cells, spikes or high/low temperature and voltage values, we can alert the driver and electrical safety officer to ensure we stop before something critical happens. The cell voltage is in turn used in our newly developed kinetic energy recovery system, KERS, state machine, which dictates how much energy we recover during braking dependent on state of charge in our accumulator. From telemetry the behavior of the state machine can be observed and validated, ensuring we do not recover more energy than the cells can handle.
Not only do we get safety-critical data through telemetry, we also get performance-based data. The main area we want to monitor in this regard is torque vectoring. Torque vectoring is the control system distributing torque on our four independent motors and acts as a controller for the rotational dynamics around the vertical axis of our racecar. From it we can gather force setpoints on the motors, yaw rate reference vs yaw rate, state estimations etc. which we can use to validate the control algorithm. Using this data combined with driver feedback we can tune the car to better traction, cornering, control, stability and balance during certain areas on the racetrack. With Revolve Analyze we have the ability to change parameters online to achieve these changes instantaneously, making tuning of our algorithms very unproblematic.
With our new database system, SKN, every log from every test and competition is stored at the same location. This data may then be used in a multitude of diverse ways. During test season this is brilliant, as there is a free flow of data allowing test crew participants, members and alumni to look at data independently from each other. Thereby giving members responsible for certain systems the ability to analyze data relevant for their system and tune their systems/fix problems with the help of alumni.
The data has proven to be of huge help to future teams as well, since they can analyze data from previous years and find problems that can be addressed in their concepts. As a modeling and simulation member, I have also found that data from previous years has been crucial to validate and tune the vehicle model used in our simulators. The same is true for other systems, included but not limited to, torque vectoring, autonomous systems and kinematics.
The telemetry system has thereby proved imminent to our continued success, allowing us an insight into the dynamics taken place hidden from the naked eye. As well as given us a platform to test, validate and tune the cutting-edge control algorithms guiding the dynamic behavior off our high-performance racecar.