A method is proposed based on the transmissibility function and the OnlineSequence Extreme Learning Machine (OS-ELM) algorithm, which is applied to theimpact damage of composite materials. First of all, the transmissi...A method is proposed based on the transmissibility function and the OnlineSequence Extreme Learning Machine (OS-ELM) algorithm, which is applied to theimpact damage of composite materials. First of all, the transmissibility functions of theundamaged signals and the damage signals at different points are calculated. Secondly,the difference between them is taken as the damage index. Finally, principal componentanalysis (PCA) is used to reduce the noise feature. And then, input to the online sequencelimit learning neural network classification to identify damage and confirm the damagelocation. Taking the amplitude of the transmissibility function instead of the accelerationresponse as the signal analysis for structural damage identification cannot be influencedby the excitation amplitude. The OS-ELM algorithm is based on the ELM (ExtremeLearning Machine) algorithm, in-creased training speed also increases the recognitionaccuracy. Experiment in the epoxy board shows that the method can effectively identifythe structural damage accurately.展开更多
Online sequential extreme learning machine(OS-ELM)for single-hidden layer feedforward networks(SLFNs)is an effective machine learning algorithm.But OS-ELM has some underlying weaknesses of neglecting time series timel...Online sequential extreme learning machine(OS-ELM)for single-hidden layer feedforward networks(SLFNs)is an effective machine learning algorithm.But OS-ELM has some underlying weaknesses of neglecting time series timeliness and being incapable to provide quantitative uncertainty for prediction.To overcome these shortcomings,a time series prediction method is presented based on the combination of OS-ELM with adaptive forgetting factor(AFF-OS-ELM)and bootstrap(B-AFF-OS-ELM).Firstly,adaptive forgetting factor is added into OS-ELM for adjusting the effective window length of training data during OS-ELM sequential learning phase.Secondly,the current bootstrap is developed to fit time series prediction online.Then associated with improved bootstrap,the proposed method can compute prediction interval as uncertainty information,meanwhile the improved bootstrap enhances prediction accuracy and stability of AFF-OS-ELM.Performances of B-AFF-OS-ELM are benchmarked with other traditional and improved OS-ELM on simulation and practical time series data.Results indicate the significant performances achieved by B-AFF-OS-ELM.展开更多
文摘A method is proposed based on the transmissibility function and the OnlineSequence Extreme Learning Machine (OS-ELM) algorithm, which is applied to theimpact damage of composite materials. First of all, the transmissibility functions of theundamaged signals and the damage signals at different points are calculated. Secondly,the difference between them is taken as the damage index. Finally, principal componentanalysis (PCA) is used to reduce the noise feature. And then, input to the online sequencelimit learning neural network classification to identify damage and confirm the damagelocation. Taking the amplitude of the transmissibility function instead of the accelerationresponse as the signal analysis for structural damage identification cannot be influencedby the excitation amplitude. The OS-ELM algorithm is based on the ELM (ExtremeLearning Machine) algorithm, in-creased training speed also increases the recognitionaccuracy. Experiment in the epoxy board shows that the method can effectively identifythe structural damage accurately.
文摘Online sequential extreme learning machine(OS-ELM)for single-hidden layer feedforward networks(SLFNs)is an effective machine learning algorithm.But OS-ELM has some underlying weaknesses of neglecting time series timeliness and being incapable to provide quantitative uncertainty for prediction.To overcome these shortcomings,a time series prediction method is presented based on the combination of OS-ELM with adaptive forgetting factor(AFF-OS-ELM)and bootstrap(B-AFF-OS-ELM).Firstly,adaptive forgetting factor is added into OS-ELM for adjusting the effective window length of training data during OS-ELM sequential learning phase.Secondly,the current bootstrap is developed to fit time series prediction online.Then associated with improved bootstrap,the proposed method can compute prediction interval as uncertainty information,meanwhile the improved bootstrap enhances prediction accuracy and stability of AFF-OS-ELM.Performances of B-AFF-OS-ELM are benchmarked with other traditional and improved OS-ELM on simulation and practical time series data.Results indicate the significant performances achieved by B-AFF-OS-ELM.
