Feasibility of a second-order strip theory for the longitudinal strength of ships