<p><i>The transfer function, defined as the ratio between the Laplacian transform of the time-domain output and the Laplacian transform of the time-domain input, is a common tool to analyze the relationships between inputs and outputs of a system. When the frequency response is sufficient to describe the system dynamics, the frequency response function (FRF) can be used as a special case of the transfer function in which Fourier transforms are used. Learning the FRF for mechanical systems can facilitate system identification, condition-based health monitoring, and improve performance metrics. Existing FRF identification assumes there is a one-to-one mapping between each input frequency component and output frequency component. However, during dynamic operations, the FRF can present complex dependencies with frequency cross-correlations due to modulation effects, nonlinearities, and mechanical noise. This paper proposes a vector regression method based on deep neural networks for the learning of generalized FRFs from dynamic system data. More specifically, a neural network based on an encoder-decoder structure is proposed. The deep encoder-decoder network features partially-tied weights and simultaneously learns the embedding space for the inputs and outputs as well as the mapping relationship between input embeddings and output embeddings. The learning model is validated using experimental data from a high-pressure hydraulic test rig. The results show that the proposed model can effectively learn the FRF between different sensor measurements with high accuracy and denoising capability.</i></p>

A Deep Neural Network Model for Learning Generalized Frequency Response Function Using Sensor Measurements

Paper Proceedings: https://asmedigitalcollection.asme.org/MSEC/MSEC2021/volume/85062

2021 International Manufacturing Science and Engineering Conference (MSEC2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.