Multi-Model Switching Based Fault Detection for the Suspension System of Maglev Train

Wang, Ping; Li, Zhiqiang; Dang, Ning

doi:10.1109/access.2018.2889733

Cited by 10 publications

(13 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To collect the real data of the suspension system, the monitoring suspension unit (MSU) (Type: CSNR-MSU0302), as shown in Fig.9, is developed to collect the current, acceleration, gap, temperature [28]. The frequency of the MSU is 10 Hz.…”

Section: Results and Analysis A Data Acquisitionmentioning

confidence: 99%

Fault Detection Method for Suspension Systems of Maglev Train Based on Optimized Random Matrix Theory

Zhou

et al. 2020

IEEE Access

Self Cite

View full text Add to dashboard Cite

The fault detection of the suspension system in a maglev train is of great importance for its operational safety and reliability. However, in random matrix theory (RMT), the size of the random matrix direct impacts the result of the mean spectral radius (MSR). In this paper, a state-of-the-art fault detection method for suspension systems is proposed using optimized RMT. The random matrix with the largest number of eigenvalues is obtained by reshaping the original data, with the help of the auto-correlation length from the correlation analysis. Finally, the optimized MSR is applied to detect the fault. The results of the experiment illustrate that the proposed method is applicable and effective. INDEX TERMS Fault detection; auto-correlation length; random matrix theory; mean spectral radius; suspension system.

show abstract

Section: Results and Analysis A Data Acquisitionmentioning

confidence: 99%

Fault Detection Method for Suspension Systems of Maglev Train Based on Optimized Random Matrix Theory

Zhou

et al. 2020

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

“…Due to the different operating conditions of the systems in actual operation, the data generated by the systems under different operating conditions is greatly different. In this case, when the same model is used to detect the system, it is likely to cause two problems [24]. One problem is false positives.…”

Section: Real-time Switching Of Multiple Operating Conditionsmentioning

confidence: 99%

“…To sum up, it is necessary to divide the data of the train into different operating conditions. According to the actual operation, the conditions of suspension node can be divided into three operating conditions: suspension static, positive line operation, and inbound and outbound operation [24]. The status word of the train, which can reflect the real-time operating status of the train, can be obtained by MSU so that switching the model by switching rules [24].…”

Section: Multi-model Switching Based On Logical Judgmentmentioning

confidence: 99%

See 1 more Smart Citation

Fault Detection for Non-Gaussian Processes Using Multiple Canonical Correlation Analysis Models and Box-Cox Transformation

Wang

Lv³

et al. 2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

Although the canonical correlation analysis (CCA)-based method has been successfully applied to Gaussian processes, the methods cannot obtain good results for the non-Gaussian processes. Meanwhile, due to the different operating conditions of the systems in actual operation, the data generated under different operating conditions is greatly different, which makes the fault detection more difficult. To deal with these problems, a fault detection method based on multiple canonical correlation analysis models and Box-Cox transformation is proposed. The method divides all conditions into some categories of conditions and builds the switching rules. Then the current condition is determined and switched in real time through real-time status data and switching rules. Since the systems are the non-Gaussian processes, the quadratic statistic established by the CCA for each condition is non-negative and non-Gaussian distributed variable. So, Box-Cox transformation is used to convert quadratic statistic into Gaussian distributed variables, such that the properties of Gaussian distribution can be employed to determine the ranges of the quadratic statistic corresponding to different health conditions. Finally, the fault detecting thresholds of the models are established based on a 99.7% confidence interval of the Gaussian distribution. The proposed method is applied to the process of monitoring of the suspension system. The application results indicate that the proposed approach can give higher fault detection rate and lower false positive rate and improves the detection ability. INDEX TERMS Switching rules, non-Gaussian processes, canonical correlation analysis, Box-Cox transformation, Gaussian distribution.

show abstract

“…Therefore, this paper mainly focuses on the operating and static suspension. Moreover, the switching rules of the two operation conditions are established correspondingly [24].…”

Section: Data Introduction and Operation Condition Divisionmentioning

confidence: 99%

Fault Detection for Complex System under Multi-Operation Conditions Based on Correlation Analysis and Improved Similarity

Shi

Zeng

2020

Symmetry

View full text Add to dashboard Cite

During actual engineering, due to the influence of complex operation conditions, the data of complex systems are distinct, and the range of similarity differs under complex operation conditions. Simultaneously, the length of the data used to calculate the similarity will also impact the result of the fault detection. According to these, this paper proposes a fault detection method based on correlation analysis and improved similarity. In the first place, the complex operation conditions are divided into several simple operation conditions via the existing historical data. In the next place, the length of the data used to calculate the similarity is determined by correlation analysis. Then, an improved similarity calculation method is proposed to make the range of the similarity under multi-operation conditions identical. Finally, this method is applied to the suspension system of the maglev train. The experiment results indicate that the method proposed in this paper can not only detect the fault or abnormal state of the suspension system but also observe the health index (HI) changes of the system at distinct times under multi-operation conditions.

show abstract

Multi-Model Switching Based Fault Detection for the Suspension System of Maglev Train

Cited by 10 publications

References 25 publications

Fault Detection Method for Suspension Systems of Maglev Train Based on Optimized Random Matrix Theory

Fault Detection Method for Suspension Systems of Maglev Train Based on Optimized Random Matrix Theory

Fault Detection for Non-Gaussian Processes Using Multiple Canonical Correlation Analysis Models and Box-Cox Transformation

Fault Detection for Complex System under Multi-Operation Conditions Based on Correlation Analysis and Improved Similarity

Contact Info

Product

Resources

About