A carregar...
Publicação
Modelo de Comportamento Dinâmico do Veículo Formula Student para otimização do Sistema de Suspensão
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 4.55 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
O presente trabalho descreve o desenvolvimento de um simulador de dinâmica de veículos com 12 graus de liberdade em MATLAB Simulink para um protótipo de Fórmula Student, motivado pela ausência de ferramentas de simulação, tanto próprias como comerciais, durante o período de participação na equipa entre 2021 e 2023, e pela ambição de aprofundar competências em dinâmica de veículos e em MATLAB e Simulink com vista a ingressar no mercado de trabalho da competição automóvel.
O modelo integra subsistemas de dinâmica longitudinal e lateral, subsistema de movimento da suspensão e modelo de pneus baseado na Fórmula Mágica de Pacejka, entre outros, sendo validado com dados de telemetria reais obtidos nas provas da FSPT 2025 e da FSG 2025. A simulação da prova de aceleração demonstrou que o modelo replica com boa aproximação o comportamento longitudinal do veículo, obtendo-se um RMSE de 1,805 km/h. A simulação em curva revelou-se inviável devido à instabilidade do controlador de direção em condições transientes, tendo o simulador sido reorientado como ferramenta de análise cinemática, recebendo diretamente dados de telemetria para calcular grandezas tipicamente não medidas em pista, como os movimentos de pitch, roll e heave da carroçaria e o deslocamento individual da suspensão em cada roda. A utilidade prática da ferramenta foi demonstrada através da análise do efeito da introdução de uma barra estabilizadora, onde se verificou uma redução de 25% no ângulo de roll. A ferramenta desenvolvida representa um contributo concreto para a equipa, colmatando uma lacuna que persistiu ao longo de vários anos.
This work describes the development of a 12 degrees-of-freedom vehicle dynamics simulator in MATLAB Simulink for a Formula Student prototype, motivated by the absence of simulation tools, both in-house and commercial, during the period of participation in the team between 2021 and 2023, and by the ambition to deepen expertise in vehicle dynamics and MATLAB and Simulink with the goal of pursuing a career in motorsport engineering. The model integrates subsystems for longitudinal and lateral dynamics, suspension motion and a tyre model based on Pacejka's Magic Formula, among others, and was validated against real telemetry data obtained during the FSPT 2025 and FSG 2025 competitions. The acceleration event simulation demonstrated that the model replicates the longitudinal behaviour of the vehicle with good accuracy, achieving an RMSE of 1.805 km/h. The cornering simulation proved unfeasible due to instability of the steering controller under transient conditions, leading to the simulator being repurposed as a kinematic analysis tool, receiving telemetry data directly to compute quantities typically not measured during events, such as pitch, roll and heave body motions and individual suspension displacement at each wheel. The practical utility of the tool was demonstrated through the analysis of the effect of introducing an anti-roll bar, where a 25% reduction in roll angle was observed. The developed tool represents a concrete contribution to the team, addressing a gap that persisted over several years.
This work describes the development of a 12 degrees-of-freedom vehicle dynamics simulator in MATLAB Simulink for a Formula Student prototype, motivated by the absence of simulation tools, both in-house and commercial, during the period of participation in the team between 2021 and 2023, and by the ambition to deepen expertise in vehicle dynamics and MATLAB and Simulink with the goal of pursuing a career in motorsport engineering. The model integrates subsystems for longitudinal and lateral dynamics, suspension motion and a tyre model based on Pacejka's Magic Formula, among others, and was validated against real telemetry data obtained during the FSPT 2025 and FSG 2025 competitions. The acceleration event simulation demonstrated that the model replicates the longitudinal behaviour of the vehicle with good accuracy, achieving an RMSE of 1.805 km/h. The cornering simulation proved unfeasible due to instability of the steering controller under transient conditions, leading to the simulator being repurposed as a kinematic analysis tool, receiving telemetry data directly to compute quantities typically not measured during events, such as pitch, roll and heave body motions and individual suspension displacement at each wheel. The practical utility of the tool was demonstrated through the analysis of the effect of introducing an anti-roll bar, where a 25% reduction in roll angle was observed. The developed tool represents a concrete contribution to the team, addressing a gap that persisted over several years.
Descrição
Palavras-chave
Formula Student Suspensão Matlab Desenvolvimento