Multi-Objective Optimization and Robust Design of Brake By Wire System Components

摘要

A Brake By Wire (BBW) system is generally composed of electro-mechanical calipers at each wheel, a pedal simulator and a central controller. The brake demand is processed by the pedal and the central controller commands the brake distribution for each brake actuator. The highly responsive and independent brake actuators lead to enhanced controllability which should result in not only better basic braking performance, but also improvements in various active braking functions such as integrated chassis control, driver assistance systems, or cooperative regenerative braking. Although the BBW system has the potential for numerous advantages and innovations in braking, it has yet to be successfully introduced in series production mainly due to safety and cost concerns. Recent studies have been made to investigate the functional safety aspects and additional mechanical backup measures in this regard. Another area that needs to be considered is the optimization of key BBW system components to increase product competitiveness and cost effectiveness. In this study, optimization and robust design was carried out for the Electric Wedge Brake (EWB) caliper and the pedal simulator. First, the EWB actuator has a very uncommon wedge mechanism that needs to be designed in the caliper housing. This can lead to various designs of the caliper that can easily grow in size and weight. An optimal design is needed in order for the caliper to have a package competitive size while at the same time maintaining required stiffness. Second, the pedal simulator is the direct connection to the driver that provides the actual braking pedal feel. Thus it is imperative that the pedal simulator is designed to give driver friendly pedal feel without sacrificing its structural integrity. The multi-objective robust design concept based on the Non-dominated Sorting Genetic Algorithm-ii (NSGA-ii) was used as the optimization method in this investigation. This method considers multiple objective functions and external noise factors, and uses self-learning weight factors. The objective functions for the EWB caliper and pedal simulator were designated as caliper weight and stiffness, and pedal feel and rubber stiffness, respectively. And variations in operation direction were considered as the noise factors for both designs. Finally, Finite Element Analysis (FEA) based on the Design of Experiments (DOE) procedure was used to obtain data to define the objective functions and the resulting robust optimal designs were verified using FEA analyses and experiments.