MODELING AND OPTIMIZATION OF VEHICLE SUSPENSION BEHAVIOR FOR ENHANCED COMFORT

Abstract

The Ride Comfort (RC), is an important parameter of Vehicle Performance & Occupant Safety that is directly influenced by the effectiveness of a vehicle's suspension system, Road Excitations and Bio-Dynamic Response of Human Body. The major sources of discomfort in vehicles are prolonged or Excessive Whole-Body Vibration transmitted to the Driver/Passenger via various road irregularities causing Fatigue, Back Pain & eventually Musculo-Skeletal and Neurological Disorders. The current research focuses on developing a Computational Model to simulate, analyze and control the transmission of vibrations from the Vehicle to the Occupants to improve the Ride Comfort through various Simulations, Controls and Optimizations. This study was developed through the development of a computational bond graph modeling for a 3 DOF quarter car model combined with a (4-DOF) lumped human biodynamic model which represents the head, upper torso, lower torso and pelvis (thigh) when seated on a vehicle in a static posture with no backrest. The models have the ability to provide a realistic representation of the transmission of vibrations from the road surface to the vehicle seat and then to the different parts of the human body. A simulation is created by applying deterministic and random excitations to the model that represent the variability of road surfaces (ISO 8608) classified as A, B, C and D with sine, trapezoidal and square waveforms. This type of excitation allows for a wide range of variability in the road conditions from very smooth highways to very uneven terrain driving environments. Time and frequency domain biodynamic response values are measured for each part of the human body and the vehicle body in response to various road excitations. Results demonstrate how the characteristics of the road surface affect the biodynamic response, providing important information regarding road roughness and its effect on human comfort. For enhanced comfort, a novel Bond Graph-based PID controller was developed to reduce vibrations generated from bumps. The Bond Graph is an effective modeling tool to represent multi-domain energy systems and their control hardware. The SYMBOL SHAKTI software is utilized to simulate an active vehicle suspension system, which is theoretically modeled as a combination of a 4-DOF iv Human Bio-Dynamic Model and a 3-DOF Quarter Automobile System. A comparison of the proposed PID controller's performance versus that of a passive model was performed. Time domain and frequency domain analyses of the systems' performance criteria were evaluated to determine the efficiency of the proposed PID controller suspension system. The results indicate that a substantial decrease in the vibration transmitted to the vehicle and the human body parts enhances ride comfort by the proposed Bond Graph PID controller model of the active vehicle suspension. In addition, the study introduced a Fractional Order PID (FOPID) controller with gain parameters optimized by Particle Swarm Optimization (PSO). The results indicated that both the Bond Graph-based PID controller and the PSO-optimized FOPID controller improve ride comfort compared to a passive vehicle system. The results also indicate that the PSO-optimized FOPID controller exhibited superior performance and achieved smoother dynamic responses over a wide range of frequencies. The research is further broadened to include an 11 DOF full car model linked to a 4 DOF human biodynamic model. A new integrated computer model is created using the bond graph method to evaluate the dynamic performance of a large road vehicle. First, 3-D models of all the individual vehicle bases, suspension units, tires, seats, and joints will be generated as bond graph elements with power ports that allow them to physically connect. This study displays the dynamic performance of a 4 DOF lumped human biomechanical model bonded to an 11 DOF full-car vehicle model utilizing a bond graph simulation method. The dynamic response of the bio dynamic human linked to a full car has been validated by ride comfort criteria based on ISO 2631:1997. The dynamic responses to various bumps in the time domain were simulated at different speeds (40 km/h) when the vehicle passed over identical sized (1m, 2m and 3.7m wide x 1m high) bumps. In addition, RMS responses were analyzed for varying vehicle speeds. The results depicted in the paper illustrate the extraordinary behavior of full-car vehicles. In the last stage of this study; the research will focus on developing an empirical mathematical model with Response Surface Methodology (RSM); that can be used to estimate ride comfort (RC) and road holding (RH) from the experimental data collected. In addition, a three-factor, three-level full factorial design was v implemented with twenty total experimental runs. The developed model was then evaluated to determine if it was adequate to provide accurate predictions by conducting an Analysis of Variance (ANOVA) with a 95% Confidence Level. The variables of interest included: Suspension stiffness coefficient (Ks), suspension damping coefficient (Cs), and tire stiffness (Kt); whereas the response variables were Ride Comfort (RC), and Road Holding (RH). The data showed that all three of the independent variable parameters (i.e., Ks, Cs, and Kt) significantly affect the two dependent variable parameters (i.e., RC and RH). Therefore, an optimal parameter combination for improved performance was identified as the best combination for improving ride comfort and better road holding characteristics.