Abstract: Out of a selection of 24 parameters from numerous factors which correlate to plane manhours in manufacture, six key parameters, including maximum takeoff weight, maximum speed, climbing rate, maximum overload, ratio of titanium/aluminum alloy and ratio of composite/aluminum alloy, are chosen via statistical and theoretical analysis based on existing plane samples. In view of the advantages of partial least-squares regression (PLS) method in analyzing multivariate data with small samples, it is utilized to build an estimation model of plane manhours in manufacture. The results show that the average error is about 10%. With this model, very little information, such as performance index, material composition, etc., is needed for the estimation of plane manufacturing manhours before the start of manufacturing activities. The PLS estimation model exhibits many advantages such as less parameters requirement, perspicuous physical meaning, and simple operation. It can be used to predict the manhours for plan optimization, design optimization, cost analysis and control, capital asset investment and production arrangement in the preliminary development stage of a plane with the absence of detailed design and manufacture information.

Key words: plane manhour, characteristic parameter, estimation model, partial least-squares regression method, multivariate data with small samples