Here is the detailed information of how i have approached my experiment
I am studying the arsenic removal using electrocoagulation, I have chosen the initial 5 variables of interest which might be influencing the arsenic removal efficiency. The variables are Ph of the solution, current density, electrolysis time, phosphorus concentration and silica concentration. Next i designed the experiment setting 2 levels for each variable and used fractional factorial design of 2^5-1 with added 3 center points, taking the total number of experiments to 19. Now with the result of regression analysis out of 5 variables, i can know the significant variables to design the model. But how do i find out the optimum levels for the significant variables? what should be the next step after finding out the significant variables to find out the points where i should be performing an experiment resulting in highest yield, in this case arsenic removal?
Thank you!
Suggest you see the chapter on Response Surface Designs in Lawson(See Link). The standard work is Myers and Montgomery( See below). The final link gives a starting place on the web. Best Wishes, David
https://www.amazon.com/Analysis-Experiments-Chapman-Statistical-Science/dp/1439868131/ref=sr_1_4?ie=UTF8&qid=1488868425&sr=8-4&keywords=john+lawson
https://www.amazon.com/Response-Surface-Methodology-Optimization-Experiments/dp/1118916018/ref=sr_1_1?ie=UTF8&qid=1488868673&sr=8-1&keywords=myers+response+surface+methodology
http://www.bing.com/search?pc=COSP&ptag=D071616-A0A14E5293D1943C986F&form=CONBDF&conlogo=CT3330954&q=response+surface+methods
You are on the right track by using the design of experiments (DOE) methodology.
I recommend that your use a statistical software called JMP where you can easily find out the optimal values for your variables that maximize the arsenic removal efficiency.
Also, I recommend that you use three levels for each main factor because using two levels will give you straight lines but not curvatures (max & min).
Hope this helps answer your question.
Professor Yehia Khalil, Yale University, USA
Fellow of the University of Oxford, UK
Thank you everyone for your answers but I should have added earlier that only software available to use for me is R studio, also i have added center points and the replicates in the design including the 2 points, is that not enough points in knowing the direction of the curvature?
And as for if the regression model is linear or not, i have not conducted the experiments yet. i am in the designing phase, and can only start making experiments once i know what to do with the regression results that is to say once i know how to find the optimum points after having the regression analysis results.
Dear Nikunj,
It is better to understand the mathematical statistics i.e. regression analysis and knowledge of relevant software for Experimental Design or Design of Experiments.Regarding the arsenic removal and increasing treatment or removal efficiency, you have to consider several factors not limited to five. Composition, concentration, media, atmospheric condition, size and shape of the reactor and so on are being the significant factors.
Regards,
Er., Dr. Chet
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Climate Science
- Climate Change