Response Simulation, Data Cleansing and Restoration of Dynamic and Static Measurements Based on Deep Learning Algorithms

Abstract

Abstract In this study, an output-based neuro controller was built based on the idea of the adaptive neuro-fuzzy inference system (ANFIS) and its capabilities in response simulation, data cleansing and restoration capability were verified using measurement data from actual structural testing. The ANFIS is a family of the deep learning algorithm, which incorporates the benefits of adaptive control technique, artificial neural network, and the fuzzy inference system. Thus, it is expected to produce very accurate predictions even for the highly nonlinear system. Forced vibration responses of a five-story steel building were simulated by ANFIS and its accuracy was compared with the results of Recurrent Neural Network (RNN), which is a type of traditional artificial neural networks. Simulations by ANFIS were very accurate with a much lower root means square error (RMSE) than RNN. Simulated data by ANFIS showed an almost perfect match with the original. Even the small ripples in the power spectrum plot outside the dominant frequency were successfully reproduced. In addition, the ANFIS was used to increase the sampling rate of dynamic data. It was shown that missing high-frequency contents could be successfully reproduced when the ANFIS was properly trained. Lastly, The ANFIS was applied to remove the noise in the measured data from RC column cyclic load tests. The outliers were corrected effectively, but the tendency of flattening the peak values was observed.

Keywords

• artificial intelligence
• deep learning
• adaptive neuro-fuzzy inference system
• response simulation
• data cleansing
• noise reduction