A General Approach to Reducing Transfer Function Errors for Design of Experiments (DOE)  

Oleh:

Liu, Xiong ; Zhang, Li Hong

Jenis: Article from Proceeding
Dalam koleksi: 12th ANQ Congress in Singapore, 5-8 Agustus 2014, page 1-8.
Topik: Design of Experiments; DOE; transfer function; design for six sigma; optimization
Fulltext: QP2-4.3-P0145.pdf (263.19KB)

Isi artikel

Design of Experiments (DOE) can be used to build transfer functions to predict a system output from its input variables, or factors. The ‘best’ transfer functions are those having the predicted values close to the observed raw data, and the statistical uncertainty for each coefficient is small. Conventionally, the estimated coefficients are drawn in linear regressions using the maximum range of factors. For a non-linear process, this way of estimating coefficients may result in large errors in the predicted models. This paper discusses the effect of factor range selection on magnitude of error for predicted models using various probability distributions

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali