Feature-Based Multi-Class Classification and Novelty Detection for Fault Diagnosis of Industrial Machinery

Calabrese, Francesca; Regattieri, Alberto; Bortolini, Marco; Galizia, Francesco Gabriele; Visentini, Lorenzo

doi:10.3390/app11209580

Cited by 16 publications

(13 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MLTs applied to fault identification and classification consist of various methods, including support vector machines [11], fuzzy logic [12], artificial neural networks [13], and decision trees [14]. Additionally, the research results of equipment fault identification and classification based on MLT have been widely used in electronics [15,16], machinery [17][18][19][20], photovoltaics [21,22], and other fields [23,24], and outstanding results have been achieved. These studies indicate that MLT can effectively improve the accuracy and reliability of fault classification and diagnostic models.…”

Section: Of 20mentioning

confidence: 99%

Identification Model of Fault-Influencing Factors for Dam Concrete Production System Based on Grey Correlation Analysis

Zhou,

Mi,

Zhao

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

A concrete production system (CPS) fault in dam engineering is one of the important factors influencing dam construction quality, which may directly affect the concrete-pouring construction progress and construction efficiency of the dam, and can even cause construction quality defects in the dam body. Reasonable classification and identification are of great significance to ensure the construction progress and quality of concrete dams. In this study, based on the concrete production logs of multiple concrete dams and literature reviews, a fault classification system for a CPS is proposed by comprehensively considering its mechanical structure characteristics and operating characteristics. The faults of the CPS are divided into 4 large categories and 22 subcategories. Additionally, the causes of CPS faults are summarized as human factors, environmental factors, mechanical component service life factors, and other factors. Based on the grey correlation analysis (GCA) method, a fault identification model of the CPS is established. With the actual production system fault statistical data of Shatuo hydropower station, the correlation coefficients for the four types of faults and the four influencing factors are calculated to determine the key faults of the CPS. The research results of the case study show that the service life factors of mechanical components have the greatest impact on batching metering system faults and mixer faults, with high grey correlation degrees of 84.66% and 76.85%, respectively. Environmental factors have the greatest impact on material delivery system faults and pneumatic system faults, with high grey correlation degrees of 90.81% and 94.9%, respectively. This paper provides theoretical support for the realization of fault pattern recognition of CPSs and provides a guiding reference for targeted fault handling.

show abstract

Section: Of 20mentioning

confidence: 99%

Identification Model of Fault-Influencing Factors for Dam Concrete Production System Based on Grey Correlation Analysis

Zhou,

Mi,

Zhao

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…On the other hand, a threshold value that is too low might also lead to removing significant features. A common trade-off choice of this threshold is 0.9, as in [54]- [58], which is also the Th value used in this paper. This step of eliminating redundant features will be implemented for each of the studied FS methods.…”

Section: The Feature Selection Scheme Implemented In This Papermentioning

confidence: 99%

On the Stability and Homogeneous Ensemble of Feature Selection for Predictive Maintenance: A Classification Application to Tool Condition Monitoring in Milling

Assafo¹,

Städter²,

Meisel³

et al. 2023

Preprint

View full text Add to dashboard Cite

<p>Feature selection (FS) represents an essential step for many machine learning-based predictive maintenance (PdM) applications, including various industrial processes, components, and monitoring tasks. The selected features do not only serve as inputs to the learning models, but also they can influence further decisions and analysis, e.g., sensor selection, understandability of the PdM system. Hence, before deploying the PdM system, it is crucial to examine the reproducibility and robustness of the selected features under variations in the input data. This is particularly critical for the real-world datasets with a low sample-to-dimension ratio (SDR). However, to the best of our knowledge, stability of the FS methods under data variations has not been considered yet in the field of PdM. This paper addresses this issue with an application to tool condition monitoring in milling, where classifiers based on support vector machine and random forest were employed. We used a 5-fold cross-validation to evaluate three popular filter-based FS methods, namely Fisher score, maximum relevance minimum redundancy (mRMR), and ReliefF, in terms of both stability and macro-F1. Further, for each method, we investigated the impact of the homogeneous FS ensemble on both performance indicators. To gain broad insights, we used four (2:2) milling datasets generated from our experiments and NASA’s repository, which differ in the operating conditions, sensors, SDR, number of classes, etc. Among the findings: 1) Different FS methods can yield comparable macro-F1, yet considerably different FS stability values. 2) Fisher score (single and/or ensemble) is the superior in most of the cases. 3) mRMR’s stability is overall the lowest, the most variable over different settings, e.g., sensor(s), subset cardinality, and the one most benefiting from the ensemble. </p>

show abstract

“…This will result in some form of novelty score, which is then compared with a decision threshold, where new unseen inputs are classified as novel if the threshold is exceeded. Novelty detection has gained much research attention, especially in diagnostic and monitoring systems [ 10 , 11 , 12 ]. An overview of the existing approaches is provided in [ 13 ].…”

Section: Related Workmentioning

confidence: 99%

Incremental Learning of Human Activities in Smart Homes

Chua

Foo

Guesgen

et al. 2022

Sensors

View full text Add to dashboard Cite

Sensor-based human activity recognition has been extensively studied. Systems learn from a set of training samples to classify actions into a pre-defined set of ground truth activities. However, human behaviours vary over time, and so a recognition system should ideally be able to continuously learn and adapt, while retaining the knowledge of previously learned activities, and without failing to highlight novel, and therefore potentially risky, behaviours. In this paper, we propose a method based on compression that can incrementally learn new behaviours, while retaining prior knowledge. Evaluation was conducted on three publicly available smart home datasets.

show abstract

Feature-Based Multi-Class Classification and Novelty Detection for Fault Diagnosis of Industrial Machinery

Cited by 16 publications

References 54 publications

Identification Model of Fault-Influencing Factors for Dam Concrete Production System Based on Grey Correlation Analysis

Identification Model of Fault-Influencing Factors for Dam Concrete Production System Based on Grey Correlation Analysis

On the Stability and Homogeneous Ensemble of Feature Selection for Predictive Maintenance: A Classification Application to Tool Condition Monitoring in Milling

Incremental Learning of Human Activities in Smart Homes

Contact Info

Product

Resources

About