Development of early order stage cost estimation models for sheet metal component

Timely submission of an accurate cost estimate is vital to the survival of manufacturing support group companies. Late submission of quotation and uncompetitive cost estimate result in lost of customers while an underestimate quotation may reduce company’s profit and usually leads to problems later in the production process. Short response time and lack of necessary technical information are the main difficulties frequently faced during the early product cost estimation. This study develops mathematical models to estimate process cost of sheet metal components at early ordering stage from collection of intuitive cost data and machine operation time. Multivariate statistical analysis is carried out to examine the relative importance of part feature representation extracted from engineering part drawing and to establish the significance relationship between the part features and the unit part process cost. This study introduces a systematic statistical framework for modeling the development of cost estimation for sheet metal components. An analytical classification approach in identifying critical factors is also established. The results of analysis illustrate that number of standard holes, bounding box perimeter, inner curved cut-out length, number of sharp edges, number of bend lines and inner straight cut-out length are among the most significant factors that influence the unit part processing cost of sheet metal components. The stability and reliability of the developed models are tested with an independent dataset using Mean Absolute Deviation and Mean Percent Absolute Deviation as main performance indicators. Cost model developed from machine operation time shows the ability to provide an estimation of part process cost within an acceptable range of 12% accuracy while cost models developed from quoted price shows a more narrowed results within 9.0% accuracy. Both the statistical time-based and price-based models provide a better estimation results compared to estimation cost produced by the company’s estimators generated based on their intuitive knowledge and experience. The statistical quality of the established mathematical models proves that the methodology developed in the study is capable to generate acceptable mathematical models for estimating early part process cost within certain degree of accuracy. The models which utilize parametric techniques provide an acceptable alternative solution to estimate process cost for manufacturing sheet metal components at early order stage in the absence of complete technical part information.

Preview PDF FK 2010 60R.pdf Download (730kB) | Preview

Actions (login required)

View Item