Multi-step ahead direct prediction for the machine condition prognosis using regression trees and neuro-fuzzy systems

Tran, Van Tung; Yang, Bo‐Suk; Tan, Andy

doi:10.1016/j.eswa.2009.01.007

Cited by 101 publications

(51 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After attaining the survival function, the process of training and forecasting by using SVM in association with time-series forecasting techniques is carried out. The multi-step ahead direct prediction method [24] of time-series forecasting techniques is applied for this study. The values of survival function from the 151th to 291th point are used to train SVM model in which the Gaussian kernel 2 2 ( , ) exp( | | /(2 ) ) K x y x y σ = − − is employed.…”

Section: Resultsmentioning

confidence: 99%

Machine Performance Degradation Assessment and Remaining Useful Life Prediction Using Proportional Hazard Model and SVM

Tran

Pham

Yang

et al. 2012

Engineering Asset Management and Infrastructure Sustainability

Self Cite

View full text Add to dashboard Cite

This paper proposes a three-stage method involved system identification techniques, proportional hazard model, and support vector machine for assessing the machine health degradation and forecasting the machine remaining useful life (RUL). In the first stage, only the normal operating condition of machine is used to create identification model to mimic the dynamic system behaviour. The machine degradation is indicated by degradation index which is the root mean square of residual errors. These errors are the difference between identification model and behaviour of system. In the second stage, the Cox's proportional hazard model is generated to estimate the survival function of the system. Finally, support vector machine, one of the remarkable machine learning techniques, in association with direct prediction method of time-series techniques is utilized to forecast the RUL. The data of low methane compressor acquired from condition monitoring routine are used for appraising the proposed method. The results indicate that the proposed method could be used as a potential tool to machine prognostics.

show abstract

Section: Resultsmentioning

confidence: 99%

Machine Performance Degradation Assessment and Remaining Useful Life Prediction Using Proportional Hazard Model and SVM

Tran

Pham

Yang

et al. 2012

Engineering Asset Management and Infrastructure Sustainability

Self Cite

View full text Add to dashboard Cite

show abstract

“…The Iterative approach is the simplest to implement [22], [26]. However, this approach suffers from propagation error: the accuracy decreases as the length of the prediction horizon increases [12], [22], [24], [29]. Moreover, this approach does not take into account the temporal behavior [24].…”

Section: B Iterative Approachmentioning

confidence: 99%

“…According to [12], the DirRec approach has the same disadvantage as the Iterative approach with respect to the propagation of the error, although the new model is created after each step of the prediction process.…”

Section: Dirrec Approachmentioning

confidence: 99%

See 1 more Smart Citation

Connexionist-Systems-Based Long Term Prediction Approaches for Prognostics

Gouriveau

Zerhouni

2012

IEEE Trans. Rel.

View full text Add to dashboard Cite

Abstract-Prognostics and Health Management aims at estimating the remaining useful life of a system (RU L), i.e. the remaining time before a failure occurs. It benefits thereby from an increasing interest: prognostic estimates (and related decisionmaking processes) enable increasing availability and safety of industrial equipment while reducing costs. However, prognostics is generally based on a prediction step which, in the context of data-driven approaches as considered in this paper, can be hard to achieve because future outcomes are in essence difficult to estimate. Also, a prognostic system must perform sufficient long term estimates, whereas many works focus on short term predictions. Following that, the aim of this paper is to formalize and discuss the connexionist-systems-based approaches to ensure multi-step ahead predictions for prognostics. Five approaches are pointed out: the Iterative, Direct, DirRec, Parallel, and MISMO approaches. Conclusions of the paper are based, on one side, on a literature review; and on the other side, on simulations among 111 time series prediction problems, and among a real engine fault prognostics application. These experiments are performed using the exTS (evolving extended Takagi-Sugeno system). As for comparison purpose, three types of performances measures are used: prediction accuracy, complexity (computational time), and implementation requirements. Results show that all three criteria are never optimized at the same time (same experiment), and best practices for prognostics application are finally pointed out.Index Terms-Prognostics and health management, multistep ahead predictions, connexionist system, evolving extended Takagi-Sugeno system. ACRONYMS

show abstract

“…According to Sorjamaa et al [26], long-term prediction is divided into three frequently used strategies that involve recursive prediction, direct prediction and DirRec prediction. The detailed information of these strategies could be found in reference [27]. In this section, the direct prediction strategy applying for p-CART model is specifically presented.…”

Section: Long-term Direct Prediction Strategy For P-cart Modelmentioning

confidence: 99%

A Classification and Regression Trees (CART) Model of Parallel Structure and Long-term Prediction Prognosis of Machine Condition

Tran

Yang²,

Tan

2009

Structural Health Monitoring

Self Cite

View full text Add to dashboard Cite

This paper presents a combined prediction model involving the parallel of classification and regression trees (CART) model, namely p-CART, and a long-term direct prediction methodology of time series techniques to predict the future stages of the machine's operating conditions. p-CART model consists of multiple CART models which are connected in parallel. Each submodel in the p-CART is trained independently. Based on the observations, these sub-models are subsequently used to predict the future values of the machine's operating conditions separately with the same embedding dimension but the different observations' indices. Finally, the predicted results of sub-models are combined to produce the final results of the predicting process. Real trending data acquired from condition monitoring routine of low methane compressor are employed for evaluating the proposed method. A comparative study of the predicted results obtained from traditional CART and p-CART models is also carried out to appraise the prediction capability of proposed model. In addition, a further improvement in predicting capability of p-CART is proposed to ameliorate the accuracy and efficiency of this method.

show abstract

Multi-step ahead direct prediction for the machine condition prognosis using regression trees and neuro-fuzzy systems

Cited by 101 publications

References 25 publications

Machine Performance Degradation Assessment and Remaining Useful Life Prediction Using Proportional Hazard Model and SVM

Machine Performance Degradation Assessment and Remaining Useful Life Prediction Using Proportional Hazard Model and SVM

Connexionist-Systems-Based Long Term Prediction Approaches for Prognostics

A Classification and Regression Trees (CART) Model of Parallel Structure and Long-term Prediction Prognosis of Machine Condition

Contact Info

Product

Resources

About