Structural Cost Optimisation of Warships

This thesis describes the development of a structural optimisation model for warships designed to the MoD (N) NES 110 structural design code based on total relative fabrication cost. Total relative fabrication cost is evaluated for a representative portion of the midship section of a typical Royal Navy frigate and attempts to take into account the costs associated with material purchase, subassembly, assembly and erection of ship's structural components. The inherent work content associated with modern day warship-building techniques are estimated by generating construction task algorithms. Each construction task algorithm can be regarded as a sequential activity list of elemental tasks which must be undertaken to effect the completion of the overall task. Every individual elemental task has an associated manhour value, this value having been derived by work study methods. Thus incorporated in the program SHBPCOST is an appropriate database of cost elements representing warship-building fabrication techniques, for the subassembly and assembly of the major structural components, namely orthogonally stiffened flat and curved panels. The formulations used for longitudinal structural design embody the current MoD (N) structural design code for surface steel ships while those for the transverse structure are based on DnV Classification Society Rules. These "first principles" and "Rules based" design methods are combined within the program FRIGATE to offer the designer an opportunity to investigate the possibilities of optimising both longitudinal and transverse warship structure with respect to total steel relative fabrication cost. Three individual studies were undertaken to investigate a basis model structure for least relative fabrication costs. Two of these studies investigated orthogonally stiffened flat panels deck structures while the third dealt with a typical frigate's double bottom structure. The results of the flat panel studies, constrained to have constant transverse panel section area together with either constant or varying transverse structure and spacing, indicate that labour costs generally increase with a corresponding increase in longitudinal stiffener numbers and decrease with transverse spacing. In addition, when the Tee bar stiffeners of the basis model flat deck panel are replaced by commercially available OBP and flat bar sections, savings of 9.0% and 10.2 % respectively, are predicted. The results from the double bottom study, when also constrained by constant sectional area, indicate that savings can be made on the total relative fabrication cost of the basis model by varying the plate thicknesses in relation to both section type and numbers. It is also demonstrated that the fabrication sequence of elemental tasks adopted in the construction of a double bottom has an important bearing on the manhours needed to complete this particular structural component. These results further demonstrate that labour cost dominates in the total relative fabrication cost relationship. This is highlighted by one option that indicates a 40% saving in material cost but only achieves a 14% saving in total relative fabrication cost. As with the flat deck panel studies, savings on the relative fabrication cost can be achieved by replacing the basis model Tee bar stiffeners by OBP and flat bar sections. This thesis describes a basic working package of two independent computer programs developed for the evaluation of alternative structural variants to a general frigate arrangement. A limiting factor applying to the studies described has been the fixed position of midship section neutral axis by virtue of a simplifying constant sectional area constraint. It is reasonable to assume, that on removal of this constraint, different structural optima would be obtained. Further investigation is required both to demonstrate this and explore the full range of possible structural cost savings resulting from variations in the vertical position of the neutral axis of the midship section even though the neutral axis is not usually a free variable in structural design. A present limitation of SHIPCOST is an inability to allow variations in the construction task algorithms applied to the fabrication of any of the structural components. Complete flexibility in this area would render SHIPCOST unwieldy, cumbersome and time consuming - unwanted attributes of a preliminary design tool. The basis of a useful preliminary design evaluation tool has been developed and demonstrated. However, further effort is required to fully generalise the models to suit any warship structural configuration