Current location - Education and Training Encyclopedia - Graduation thesis - How to do orthogonal experiment?
How to do orthogonal experiment?
My experiment is three factors and three levels. If I do orthogonal experiments, it is nine experiments (I get a set of data from one experiment a month). I think many papers are divided into two steps: the first step, the orthogonal design of 1, the orthogonal design, and the subsequent experimental scheme. 2. Carry out the experiment according to the experimental scheme of orthogonal design (there is no process and chart, because the collocation of each single factor is random, so it is impossible to draw a graphic analysis), and only give the results. 3. According to the experimental results, the influence degree of each single factor is obtained by orthogonal analysis. I'm going to do nine experiments in orthogonal design. ) Second, experiments supplemented by single factor experiments verify that the analysis results of orthogonal design are comparable, and the role of single factor can be analyzed by graphics. In this way, I did my experiment twice (although the control factors of these two experiments may be different). After so many experiments, when will you graduate? My question is: some people say that it is more convenient to determine the factor level of orthogonal design by single factor first. I don't want to do orthogonal design. I want to get the best experimental scheme directly through simple comparison method, but the paper made by this method is not dazzling enough, and I feel that there is no technical content, which is far from other orthogonal designs. I have a headache and can't decide the plan. I hope friends passing by can help me.