A sparse auto-encoder-based deep neural network approach for induction motor faults classification

This paper offers a review of the artificial intelligence (AI) algorithms and applications presently being used for smart machine tools. These AI methods can be classified as learning algorithms (deep, meta-, unsupervised, supervised, and reinforcement learning) for diagnosis and detection of faults in mechanical components and AI technique applications in smart machine tools including intelligent manufacturing, cyber-physical systems, mechanical components prognosis, and smart sensors. A diagram of the architecture of AI schemes used for smart machine tools has been included. The respective strengths and weaknesses of the methods, as well as the challenges and future trends in AI schemes, are discussed. In the future, we will propose several AI approaches to tackle mechanical components as well as addressing different AI algorithms to deal with smart machine tools and the acquisition of accurate results.Keywords: artificial intelligence; smart machine tools; learning algorithms; intelligent manufacturing; fault diagnosis and prognosis Brief IntroductionWe believe that a new epoch of the "Industrial Internet of Things (IIoT) plus artificial intelligence (AI)", characterized by big machinery data, data-driven techniques, ubiquitous networks, mass innovation, automatic intelligence, cross-border integration, and shared services, has arrived [1-3]. The fast development and combination of new AI and energy technologies, for materials, bioscience, the Internet, and new-generation information exchange, is a fundamental part of this new epoch. This will, in turn, permit game-changing transformation of models, ecosystems, and means in the light of their application to national security, well-being, and the economy. The main objective is a review and summary of recent achievement in data-based techniques, especially for complicated industrial applications, offering reference for further study from both an academic and practical point of view. Yin et al.[1] describes a brief evolutionary overview of data-based techniques over the last two decades. Recent development of modern industrial applications is presented mainly from the perspectives of monitoring and control. Their methodology, based on process measurements and model-data integrated techniques, will be introduced in the next study. Jeschke et al. [2] developed the core system science needed to enable the development of complex IIoT/manufacturing cyber-physical systems (CPS). Moreover, readers can learn the current state of IIoT and the concept of cybermanufacturing from this book. In 2014, Lund et al. [3] described the central issues contributing to, and characterizing, the worldwide and regional growth of the IoT. Besides, researchers can utilize the trend analysis of IoT their region markets in the future.There are many AI algorithms for machine health monitoring and other machine tool applications: The second-order recurrent neural networks (RNN) method for the learning and extraction of finite

show abstract

“…Sun et al [107] proposed a DNN approach to Induction Motor fault diagnosis. The scheme employed SAE to study features.…”

Section: Other Mechanical Componentsmentioning

confidence: 99%

A Review of Artificial Intelligence Algorithms Used for Smart Machine Tools

2018

View full text Add to dashboard Cite

show abstract

“…As one of many famous unsupervised feature learning methods, SAE combines with DL to realize its effectiveness. The aim of this process is to reconstruct input data at the output layer by a sparse penalty term β [13].…”

Section: B Deep Learning Based Chf Detection Algorithmmentioning

confidence: 99%

A CHF detection method based on deep learning with RR intervals

Chen

Liu

et al. 2017

2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

View full text Add to dashboard Cite

Abstract-There are extensive studies investigating congestive heart failure (CHF) detection based on heart rate variability. Although a high level of accuracy has been achieved, its robustness under different conditions is not guaranteed. To improve the robustness, we applied sparse auto-encoder-based deep learning algorithm in CHF detection with RR intervals. A total data size of 30,592 (5-min RR interval) was obtained from 72 healthy persons and 44 CHF patients. The deep learning algorithm first extracts unsupervised features using a sparse auto-encoder from raw RR intervals, then constructs a deep neural network model with various hidden nodes combinations. Results showed that the model achieved 72.41% accuracy. This demonstrated that RR intervals have potential in CHF detection but cannot fully reflect dynamic change in 24-h. I. INTRODUCTIONAs a common chronic cardiovascular syndrome, congestive heart failure (CHF) has attracted many researchers' attention. CHF is linked to several complex problems, including end stage heart suffering [1], that cause chambers of the heart to fail. When symptoms suggest CHF, the diagnosis is usually confirmed by physical examination, patient history and various tests to detect abnormal function of the left ventricle and/or heart valves [2]. This diagnosis process often results in time delay and high costs.Electrocardiogram (ECG) is a commonly used physiological signal in monitoring cardiovascular condition, as it's non-invasive, portable and easy to use. ECG shows both heart rate and rhythm (steady or irregular) [1].For CHF detection, many researchers focused on ECG analysis, including QRS waves [3], QT intervals [4], RR variability [5], and so on. RR interval is one of the important components of ECG and has attracted great attention as its variability has great potential in CHF prognosis [5]. This variability is either described by RR variability (i.e. variation of RR intervals) or heart rate variability (HRV). HRV based CHF detection has achieved over 95% accuracy with linear/non-linear measures of short-term (5-min)/long-term (24-h) RR intervals data [5]. However, due to its sensitivity along with clinical condition changes [2], the robustness of the HRV-based approaches is still an issue to be addressed. In our previous work [5], we proposed dynamic HRV to describe W. Chen is with the engineering faculty, Sun Yat

show abstract

“…With this idea in mind, 8) are going to be directly discarded from the in-control sample. χ 2 α ;p stands for the (1 − α) percentile of the chi-squared distribution with p degrees of freedom.…”

Section: Robust Quality Control Chartsmentioning

confidence: 99%

Robust Detection of Incipient Faults in VSI-Fed Induction Motors Using Quality Control Charts

García-Escudero

Duque-Pérez

Fernandez-Temprano

et al. 2017

IEEE Trans. on Ind. Applicat.

View full text Add to dashboard Cite

Abstract-A considerable amount of papers has been published in recent years proposing supervised classifiers to diagnose the health of a machine. The usual procedure with these classifiers is to train them using data acquired through controlled experiments, expecting them to perform well on new data, classifying correctly the condition of a motor. But, obviously, the new motor to be diagnosed cannot be the same that has been used during the training process; it may be a motor with different characteristics and fed from a completely different source. These different conditions between the training process and the testing one can deeply influence the diagnosis. To avoid these drawbacks, in this paper, a new method is proposed, which is based on robust statistical techniques applied in quality control applications. The proposed method is based on the online diagnosis of the operating motor and can detect deviations from the normal operational conditions. A robust approach has been implemented using high-breakdown statistical techniques, which can reliably detect anomalous data that often cause an unexpected overestimation of the data variability, reducing the ability of standard procedures to detect faulty conditions in earlier stages. A case study is presented to prove the validity of the proposed approach. Motors of different characteristics, fed from the power line and several different inverters, are tested. Three different fault conditions are provoked: a broken bar, a faulty bearing, and mixed eccentricity. Experimental results prove that the proposed approach can detect incipient faults.

show abstract

A sparse auto-encoder-based deep neural network approach for induction motor faults classification

Cited by 621 publications

References 27 publications

A Review of Artificial Intelligence Algorithms Used for Smart Machine Tools

A Review of Artificial Intelligence Algorithms Used for Smart Machine Tools

A CHF detection method based on deep learning with RR intervals

Robust Detection of Incipient Faults in VSI-Fed Induction Motors Using Quality Control Charts

Contact Info

Product

Resources

About