正交实验设计（Orthogonal experimental design）.docVIP

正交实验设计（Orthogonal experimental design） What is orthogonal experimental design? The orthogonal experimental design (Orthogonal experimental design) is a kind of design method of multi factors and multi levels, it is based on the orthogonality selected from comprehensive test of some representative points test, these representative points have evenly dispersed, neat features than the orthogonal test design is fractional factorial design of the main method. It is an efficient, fast and economical experimental design method. Taguchi Xuanichi, a famous Japanese statistician, made a list of the horizontal combinations chosen by orthogonal tests, called orthogonal lists. For example, for a three factor, three level experiment, according to the requirements of a comprehensive experiment, 3^3 = 27 combinations of experiments should be carried out, and the number of repetitions of each combination has not been considered. If the experiment is arranged according to the L9 (3) orthogonal table, only 9 times, and the 18 experiment by L18 (3) orthogonal table will obviously reduce the workload greatly. Therefore, orthogonal experimental design has been widely used in many fields. The orthogonal table is a set of rules of the design table, using L as the orthogonal table of the code name, n for the number of tests, t is the level, C is the column, that is, the number of factors that may be arranged most. For example, L9 (3), which indicates that 9 experiments are required, with the maximum observation of 4 factors, each of which is 3 levels. The number of levels in an orthogonal table can also be unequal. We call it a mixed orthogonal table, such as L8 (4 * 2). In the 5 column of this table, there are 1 for 4 and 4 for the 2. Edit the orthogonal test design of this section Table 1 Orthogonal test design [1] Orthogonal test, factor level table, orthogonal test design scheme and test result range analysis table (or index and factor relation chart) variance analysis table (when simpl