Fractional-factorial design of a porous-carbon fuel-cell electrode

Nanostructured Gas Diffusion and Catalyst Layers for Proton Exchange Membrane Fuel Cells

“…Platinum catalyst deposition on MWCNTs was carried out as follows. 27,28 MWCNTs were uniformly dispersed in deionized water through ultrasonication and magnetic stirring. The dispersion was heated to about 70°C with continuous stirring operation.…”

Inference for Effect-Saturated Fractional Factorials

“…e2 alpha= 0.3, k= 15 Critical values are given in Table 2 for use with this pro- This relationship has not been verified for other sample Figure 5. Active Contrast Plots of the Kannan et al (1988) sizes, however. A more resistant version of the PSE-based Example With Priors of (a) a = .…”

“…Pareto Plot (a) and Half-Normal Plot (b) for the example fromKannan et al (1988). The vertical lines are the estimated scales multiplied by the a = .05 empirical critical values for the following methods: , PSE(q = .5, b = 2.5); --, ASE(q = .5,b = 2.5); ---, PSE(q = .467, b = 1.25); --, TSE(q = .6), where q is the proportion of effects included in the initial scale estimator and b is the multiplier used in trimming large effects prior to calculating the final scale estimate.…”

A 27-3 fractional factorial optimization of polybenzimidazole based membrane electrode assemblies for H2/O2 fuel cells

“…However, one big drawback with factorial design is that the total number of experiments goes up sharply as the number of factors, thus requiring too much time and resources for optimization. The use of fractional factorial experiments can greatly reduce this number of experiments facilitating a quicker realization through the path of steepest ascent [61][62][63]. The principle of fractional factorial experiment is that when the higher order effects of two or more combination of experiments are the same, they cannot be distinguished from one another, and hence they can be aliased, thus reducing the total number of experiments.…”

“…Factorial experiments are one of the most efficient designs when multiple parameters interact significantly among themselves and when they have a complementary impact on each other [59][60][61][62][63]. Further, factorial experiments can point the way to the choice of conditions outside those originally selected by means of which a greater response of the required parameter can be achieved using the ''method of steepest ascent''.…”