Assessment of Seat Comfort by Virtual Simulation

Next to an increase in the number of car users, comfort has become a major aspect by which car manufacturers can distinguish their products from their competitors. At today, the assessment of seating comfort is largely based on subjective measures. A disadvantage of such measures is that the relationship with design parameters is often unclear. Furthermore, prototype development and testing are both time consuming and costly. In this respect, virtual testing tools represent promising solutions to support early automotive tests and, ultimately, reduce development times and associated costs. Moreover, virtual testing tools allow to investigate parameters that are hard to measure, such as the intervertebral disc pressure or the pressure distribution in the human soft tissues of the buttocks, and put them in relation with (dis)comfort and physical complaints.For prediction of seating comfort by virtual testing, in addition to a detailed seat model, a highly sophisticated model of the occupant and objective parameters are necessary. In the simulation work of this thesis, the human model CASIMIR, presented by Pankoke (2003), was used. CASIMIR represents a dynamic, anatomical predictive finite element model of a man in a sitting posture (Siefert et al., 2006) and it is currently available for the ABAQUS code. The most significant characteristics of the CASIMIR model in seat comfort analysis are: a detailed model of the lumbar spine, including frequency-dependent damping properties of the intervertebral discs, a detailed model of the relevant abdominal and dorsal musculature, together with a detailed skeletal model. Regarding the objective parameters, the mean and maximum pressure, the load percentage and contact area are the parameters, which are mostly reported in the literature to relate to (dis)comfort, and are therefore used in this thesis. The comfort factor that relate to the automobile seat is generally explored in the literature by the following indicators: height, adjustment, position, properties of upholstered seat such as width, length and shape, foam softness and aesthetics. Some works have been conducted to show association between seat dimensions, seat shape, seat material and interface pressure but little is known about the influence of the properties of the seat and their interactions on the (dis)comfort perception of drivers. Moreover, a strong optimization of seat properties in order to fulfill comfort criteria is not established in literature.This thesis tries to enlighten the interactions of different parameters in order to achieve optimum seat design based on the comfort criterion proposed by Mergl in 2006. In the first step, the validity of the model was investigated through the level of correlation between experimental pressure maps and numerical results. The test group consisted of six male subjects with anthropometric data close to 50th percentile (a man with average anthropometric values). The pressure maps were obtained through the Xsensor pressure mat during static seating of the subjects. The comparison between experimental objective parameters related to (dis)comfort and numerical output data showed a good correlation. In the second phase of the thesis, the numerical model was then used to evaluate the sensitivity of static seat comfort to some input variables, i.e. the stiffness of cushion and backrest foams and some geometry parameters of the seat cushion and backrest. In order to perform the sensitivity analysis, a reduced full factorial technique was used. Sensitivity analysis of the DOE technique showed a direct relation between maximum interaction pressure and shape of seat cushion, as well as stiffening of materials in soft parts of the seat. Thereafter, the results of DOE run were used to create an approximation of the response variable over the design space, in order to implement the multi-objective optimization based on comfort criteria.