Optimisation of liquid tank geometry for enhancement of static roll stability of partially-filled tank vehicles

An optimisation method, based on a genetic algorithm, is applied to derive ail optimal tank geometry for enhancement of the roll stability limit of a tank-vehicle combination, specifically under partial fill conditions. The genetic algorithm is applied to minimise the primary overturning moment imposed Oil the vehicle due to lateral acceleration and cargo load shift expressed in terms of lateral and vertical movement of the liquid bulk within the partly filled tank. The periphery of an arbitrarily shaped tank is generated using B-spline curves to achieve the necessary flexibility to arrive at an optimal shape, where each control point has local influence on the curve. The genetic algorithm considers a combination of the control points as an individual, while each population is formed with a certain number of individuals. The permutations of the individuals are evaluated to derive the best solution that would minimise the objective function based upon the primary overturning moment due to the liquid bulk. The minimisation process is subjected to a number of limit constraints on the resulting total cross-sectional area, overall height and overall width. A static roll plane model of a commercial tank vehicle comprising optimal tank geometry and quasi-static fluid slosh is utilised to study the benefits of the optimal cross-section under various fill conditions in terms of rollover threshold, lateral acceleration and sprung mass roll angle. The potential performance benefits of the proposed optimal tank are discussed in relation to those of the currently used tank cross-sections.