|dc.description.abstract||The design of combustion devices is very important to society today. They need to be highly efficient, while reducing emissions in order to meet strict environmental standards. These devices, however, are currently not being designed effectively. The most common method of improving them is through parametric studies, where the design parameters are altered one at a time to try and find the best operating point. While this method does work, it is not very enlightening as it neglects the non-linear interactions between the design parameters, requires a large amount of time, and does not guarantee that the best operating point is found. As the environmental standards continue to become stricter, a more robust method of optimizing combustion devices will be required.
In this work a robust design optimization algorithm is presented that is capable of mathematically accounting for all of the interactions between the parameters and can find the best operating point of a combustion device. The algorithm uses response surface modeling to model the objective function, thereby reducing computational expense and time as compared to traditional optimization algorithms.
The algorithm is tested on three case studies, with the goal of improving the radiant efficiency of a two stage porous radiant burner. The first case studied was one dimensional and involved adjusting the pore diameter of the second stage of the burner. The second case, also one dimensional, involved altering the second stage porosity. The third, and final, case study required that both of the above parameters be altered to improve the radiant efficiency. All three case studies resulted in statistically significantly changes in the efficiency of the burner.||en