Uncertainty Quantification of Vibroacoustics with Deep Neural Networks and Catmull–Clark Subdivision Surfaces

Table 6

MCs results of simulation and DNN.

Frequency (Hz)	Variables	Expectation		Standard deviation
		Simulation	DNN	Simulation	DNN
50	1-D radius	4.54E − 02	4.67E − 02	1.60E − 02	1.72E − 02
	1-D thickness	3.23E − 02	3.23E − 02	1.04E − 03	1.11E − 03
	1-D density	4.51E − 02	4.52E − 02	3.76E − 03	4.00E − 03
	1-D Young’s modulus	4.44E − 02	4.46E − 02	3.05E − 03	2.87E − 03
	1-D Poisson ratio	4.49E − 02	4.49E − 02	1.11E − 03	1.16E − 03
	5-D variables	4.57E − 02	4.64E − 02	1.83E − 02	1.90E − 02
300	1-D radius	8.71E − 01	8.81E − 01	1.75E − 01	1.85E − 01
	1-D thickness	7.47E − 01	7.47E − 01	1.74E − 02	1.78E − 02
	1-D density	8.84E − 01	8.85E − 01	3.00E − 02	3.05E − 02
	1-D Young’s modulus	8.80E − 01	8.83E − 01	4.56E − 02	4.44E − 02
	1-D Poisson ratio	8.85E − 01	8.85E − 01	4.02E − 03	4.18E − 03
	5-D variables	8.88E − 01	8.92E − 01	1.78E − 01	1.74E − 01