Optimization of broadband transducer designs by use of statistical design of experiments

Optimizing designs that involve several parameters is a multidimensional problem, and the optimum setting of one parameter may depend on the exact setting of other parameters. Statistical design of experiments is a powerful methodology to efficiently find the optimum operating points, especially when interactions between terms is important. Here, the methodology will be described and applied to a transducer design problem to optimize bandwidth and impulse response. The design problem involves four interactive factors, the acoustic impedance of the first and second matching layer, and the appropriate thicknesses of the first and second matching layer.