Please elaborate, how to design an experiment for better yield optimization with minimum experiments. I have four variables, mole ratio, temp, time, catalyst loading, and sometimes instead of a cat. loading I use microwave watt power
Yogeshsing N. Rajput
Regarding your question, "how to design an experiment for better yield optimization with minimum experiments". Taguchi Method can help to achieve an optimized result with very few sample sizes. But, the prediction might lead to a false hypothesis. Depending on the design factors and level of those factors, you can come to a prediction/optimization of the experiment. The trade-off will be, the full factorial design might give better optimization (Quaradatic model, Interaction of factors, etc.), where a partial factorial design might not be as accurate. Whereas you can save time and cost doing a small number of experiments.
You can also follow the link regarding in-depth analysis from the design perspective.
Link: (https://eng.libretexts.org/Bookshelves/Industrial_and_Systems_Engineering/Book%3A_Chemical_Process_Dynamics_and_Controls_(Woolf)/14%3A_Design_of_Experiments/14.01%3A_Design_of_Experiments_via_Taguchi_Methods_-_Orthogonal_Arrays#:~:text=The%20experimental%20design%20proposed%20by,method%20tests%20pairs%20of%20combinations.)
Dear Yogeshsing N. Rajput
In fact, the Taguchi method uses the concept of orthogonal arrays to investigate a large number of controllable factors with a small number of experiments. Taguchi method finds the optimal level of controllable factors by minimizing the effect of noise.
Best regards,
Mahmoud
Dear all thank you for your answers.
but can Taguchi be used for the optimization of chemical synthesis?
Yogeshsing N. Rajput
It depends on your end result. If you want to find the intersection/influence of variables on the result. Then you can use Taguchi in finding the optimum combination of parameters for the solution.
I'm not sure how you will use the method in synthesis, but if your goal is to find the optimum parameters, sure you can use Taguchi.
Hello, If you give me more detail, I can create a model for you and send it back you.
Dear Yogeshsing N. Rajput ,
Yes, you can use L9 (34). But you should notice that you should use Pulled Error for response of variable parameter.
I draw your attention to following article:
Article Mathematical Modeling and Optimization of Fused Filament Fab...
Good luck!
- Badges
- Science topic