I am new to the field of Response Surface Methodology and want to use it for optimisation of cogging torque.I have 6 factors to be investigated. What are the statistical tools that can be used in framing RSM model of cogging torque? Please help.
To frame a Response Surface Model, you require the Design Expert (several versions available) software to ably investigate the interaction of all 6 factors you are interested in. With this software, you will be in position to design the experiments for the 6 factors, using the Response Surface option (choose the Central Composite Design; since its multifactorial) to generate a series of random samples to be tested to give responses you are interested in. This's the start of it all, but optimisation comes afterwards. I hope this was helpful.
Cogging torque is a response from several interacting factors; upon analysis of the data from the factors in consideration, RSM generates an ideal polynomial model which also will indicate the significant factors. Attached is review paper, that will be helpful. Thank you.
Are you sure all six factors are important for your experiment? I would use Design Expert software and generate an 8 or 9-factor Definitive Screening Design. These are a new type of design that allow the user t test linear and quadratic effects. If you add some extra runs to the experiment, you can also test interactions. The 9 factors DSD will use 19 runs. You will have 6 extra terms for looking at interactions and finding smaller effects. Once this experiment is done, you can use it to make a better experiment or you can add to this experiment to gather all the data you want/need.
You may use MiniTab software for Response surface designs.
It provides two types of designs CCD (2 to 10 factors); Box-Behnken (3-10 factors), further full, half, quarter and eighth factorial designs in blocked and unlocked mode. I would use these design combinations for various chemical reactions as well as diesel engine applications in combination with evolutionary algorithms.