NUST Institutions Library Catalogue

This research discusses the development of surrogate models for aircraft and rotorcraft design
during the conceptual phase. The approach proposed in this study, on a crude scale, holds a
strong utilization potential among the design community and aero modelers during the
preliminary sizing. The fixed-wing aircraft database includes fighters, trainers, transport aircraft,
agricultural planes, amphibious aircraft, light utility planes, motor-gliders, sport planes, and
freighters. Similarly, the rotorcraft database comprises conventional, fenestron, counter-rotating,
and no tail rotor configurations was developed from commercially available data. Linear/NonLinear regression modeling using power laws is carried out to explore highly correlated design
trends among aircraft and rotorcraft weight, geometric, propulsion, and performance parameters.
These design trends identify interdependencies among design parameters and provide initial
design bounds for preliminary sizing. Aircraft and rotorcraft design is complex process, and a
single variable can never predict the response with adequate confidence. Surrogate models using
multiple linear regression techniques were developed to estimate aircraft and rotorcraft
performance parameters: range, rate of climb, service ceiling, and maximum velocity. Moreover,
these models are validated using a two-step process that includes verifying each model using
quantitative criteria and checking the prediction accuracy of each model.