Basic tools for vibration analysis with applications to predictive maintenance of rotating machines: an overview

Popescu, Theodor D.; Aiordăchioaie, Dorel; Culea-Florescu, Anisia

doi:10.1007/s00170-021-07703-1

Cited by 32 publications

(25 citation statements)

References 85 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We remark that the statistics employed in QuantTree [9] estimate the bin probability p j by its target value π j since it is assumed that a large training set is provided, and πj → π j as N → ∞ by definition (5). Here we also consider cases where N is small, thus in the QT-EWMA statistics ( 6) and (7) we employ πj , which is a more accurate estimate of p j . Impact of the choice of K. The number of bins K of the QuantTree histogram is a fundamental parameter that influences the change-detection performance of QT-EWMA.…”

Section: The Qt-ewma Algorithmmentioning

confidence: 99%

“…. , y K,t )} 5000 t=1 to compute the QT-EWMA statistics {T t } 5000 t=1 by ( 6)- (7). To compute the thresholds {h t } t yielding the desired ARL 0 , we repeat the procedure above 1,000,000 times, and define h 1 as the empirical (1 − α)-quantile of all the values of T 1 , where α = 1/ARL 0 as in (15).…”

Section: Computing Thresholds To Control the Arlmentioning

confidence: 99%

“…A relevant example is industrial process monitoring, where production machinery is equipped with multiple sensors that measure vibration frequency, flow, temperature, pressure, etc. These observations form datastreams that must be monitored since any distribution change might indicate failures or ongoing deterioration of specific components such as bearings and gears, thus change detection can be key for predictive maintenance [7]. Change detection is also studied in the machine learning literature since classification and recommendation systems often operate on streaming data.…”

Section: Introductionmentioning

confidence: 99%

“…In particular, any change-detection algorithm is characterized by a trade-off between the frequency of false alarms and the detection power. In many applications, including our previous example of industrial process monitoring [7], promptly detecting changes is crucial (e.g., to avoid a failure). Still, any false alarm might trigger a costly intervention (e.g., the replacement of functioning machinery).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Nonparametric and Online Change Detection in Multivariate Datastreams Using QuantTree

Frittoli¹,

Carrera²,

Boracchi³

2022

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

We address the problem of online change detection in multivariate datastreams, and we introduce QuantTree Exponentially Weighted Moving Average (QT-EWMA), a nonparametric change-detection algorithm that can control the expected time before a false alarm, yielding a desired Average Run Length (ARL 0 ). Controlling false alarms is crucial in many applications and is rarely guaranteed by online change-detection algorithms that can monitor multivariate datastreams without knowing the data distribution. Like many change-detection algorithms, QT-EWMA builds a model of the data distribution, in our case a QuantTree histogram, from a stationary training set. To monitor datastreams even when the training set is extremely small, we propose QT-EWMA-update, which incrementally updates the QuantTree histogram during monitoring, always keeping the ARL 0 under control. Our experiments, performed on synthetic and real-world datastreams, demonstrate that QT-EWMA and QT-EWMA-update control the ARL 0 and the false alarm rate better than state-of-the-art methods operating in similar conditions, achieving lower or comparable detection delays.

show abstract

Section: The Qt-ewma Algorithmmentioning

confidence: 99%

Section: Computing Thresholds To Control the Arlmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Nonparametric and Online Change Detection in Multivariate Datastreams Using QuantTree

Frittoli¹,

Carrera²,

Boracchi³

2022

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

show abstract

“…Therefore, to avoid affecting the control performance of the servo system, multi-parameter identification of the nonlinear system is very important, and relevant personnel have made some meaningful attempts and explorations. The existing multi-parameter identification methods of servo systems mainly include the least-squares method [5] and the transfer function identification method based on step response [6][7][8], such as the two-point and three-point methods.…”

Section: Introductionmentioning

confidence: 99%

Reverse Nonlinear Sparrow Search Algorithm Based on the Penalty Mechanism for Multi-Parameter Identification Model Method of an Electro-Hydraulic Servo System

et al. 2022

View full text Add to dashboard Cite

Aiming at the multi-parameter identification problem of an electro-hydraulic servo system, a multi-parameter identification method based on a penalty mechanism reverse nonlinear sparrow search algorithm (PRN-SSA) is proposed, which transforms the identification problem of a non-linear system into an optimization problem in a high-dimensional parameter space. In the initial stage of the sparrow search algorithm (SSA), the population distribution is not uniform, and the optimization process is easily disturbed by the local optimal solution. First, adopting a reverse learning strategy increases the exploratory nature of individuals in a population, improves population diversity, and prevents premature maturity. Subsequently, a flexible strain mechanism is provided through the nonlinear convergence factor, adaptive weight factor, and golden sine and cosine factor. The introduction of a nonlinear factor fully balances the global search and local development abilities of the algorithm. Finally, a punishment processing mechanism is developed for vigilantes while retaining the population, providing a suitable search scheme for individuals beyond the boundary, and making full use of the value of each sparrow individual. The effectiveness of each improved strategy is verified through simulation experiments with 23 benchmark functions, and the improved algorithm exhibits better robustness. The results of the model parameter identification of the electro-hydraulic servo system show that the method has a high fitting accuracy between the identification model data and the experimental data, and the fitting degree of the identification model exceeds 97.54%, which further verifies the superiority of the improved algorithm and the effectiveness of the proposed identification strategy.

show abstract

Automatic Anomaly Detection in Vibration Analysis Based on Machine Learning Algorithms

Torres,

Ramalho,

Correia

2022

Innovations in Mechatronics Engineering II

View full text Add to dashboard Cite

This paper presents an approach for automatic anomaly detection through vibration analysis based on machine learning algorithms. The study focuses on induction motors in a predictive maintenance context, but can be applied to other domains. Vibration analysis is an important diagnostic tool in industrial data analysis to predict anomalies caused by equipment defects or in its use, allowing to increase its lifetime. It is not a new technique and is widely used in the industry, however with the Industry 4.0 paradigm and the need to digitize any process, it gains relevance to automatic fault detection. The Isolation Forest algorithm is implemented to detect anomalies in vibration datasets measured in an experimental apparatus composed of an induction motor and a coupling system with shaft alignment/misalignment capabilities. The results show that it is possible to detect anomalies automatically with a high level of precision and accuracy.

show abstract

Basic tools for vibration analysis with applications to predictive maintenance of rotating machines: an overview

Cited by 32 publications

References 85 publications

Nonparametric and Online Change Detection in Multivariate Datastreams Using QuantTree

Nonparametric and Online Change Detection in Multivariate Datastreams Using QuantTree

Reverse Nonlinear Sparrow Search Algorithm Based on the Penalty Mechanism for Multi-Parameter Identification Model Method of an Electro-Hydraulic Servo System

Automatic Anomaly Detection in Vibration Analysis Based on Machine Learning Algorithms

Contact Info

Product

Resources

About