Statistical approaches for semi-supervised anomaly detection in machining

Denkena, Berend; Dittrich, Marc-André; Nöske, H. D.; Witt, Matthias

doi:10.1007/s11740-020-00958-9

Cited by 12 publications

(4 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2A. Besides sensor information for tool condition monitoring like in [1] or additionally in [3,4,28], data of the NC can be used as additional information for anomaly detection [29]. These kinds of evaluation are often based on key figures and fixed rule engines to provide a real-time statement for online and inline analyses.…”

Section: Utilizing Sensor Datamentioning

confidence: 99%

Tooling systems with integrated sensors enabling data based process optimization

Bleicher¹,

Ramsauer²,

Leonhartsberger³

et al. 2021

Journal of Machine Engineering

View full text Add to dashboard Cite

Sensor integration into machining equipment has become an important factor for gaining deep process insights mainly driven by increasingly smaller and cheaper sensors and transmitters. Due to advances in microelectronics and communication technology, a broader field of applications in production processes and machine tools can be addressed using sensing devices and their implementation potentials. Ensuring a sensitive but robust data stream from close to the actual process allows not only reliable monitoring but also process and quality control based on sensor information. This paper provides an overview of the utilization of sensor data for the purpose of condition monitoring, model fitting and real-time control coping with stochastic effects. Examples of sensor integration in fields of injection molding, roll forming and heavy-duty milling comprise the state of the art of sensor implementation, data evaluation and possible feedback loops in the respective application scenarios.

show abstract

Section: Utilizing Sensor Datamentioning

confidence: 99%

Tooling systems with integrated sensors enabling data based process optimization

Bleicher¹,

Ramsauer²,

Leonhartsberger³

et al. 2021

Journal of Machine Engineering

View full text Add to dashboard Cite

show abstract

“…The variability of the cavity is the result of many factors. Among the most important are: variability of the nominal depth of cut, which depends on the kinematic characteristics of the method [ 1 ]; the irregularities of the workpiece surface in the machining zone and formation of the side pile-ups in the abrasive grain—workpiece contact zone [ 2 , 3 ]; local susceptibility of the workpiece material and abrasive grains [ 1 , 4 ]; local variations of the grinding wheel active surface [ 5 , 6 ]; vibration of the tool and abrasive grains [ 7 ]; significant local variation (in the grain interaction zone) of temperature rise [ 8 , 9 , 10 , 11 ]; especially when machining at very high speeds, materials with low thermal conductivity produce variation in the properties of the workpiece material in micro volumes compared with the volumes of the cut layers [ 12 , 13 ]; macro- and micro-discontinuity of chip and pile-ups formation [ 14 ]; micro chipping of grains and changes in the grinding wheel active surface [ 15 , 16 , 17 ]. …”

Section: Introduction: Selected Features Of Grinding Processes Useful...mentioning

confidence: 99%

Technological Aspects of Variation in Process Characteristics and Tool Condition in Grinding Process Diagnostics

et al. 2023

View full text Add to dashboard Cite

The article presents the technological aspects of the diagnostics of grinding processes. The main features of the grinding process and their importance in diagnostic issues are discussed. Selected issues of research and assessment of the condition of the active surface of grinding wheels are presented. The authors pointed out that the parameters used to assess the topography of the ground surfaces do not have sufficient possibilities to differentiate the surface condition of the grinding wheels. This publication draws attention to the possibility of using new dedicated parameters to assess the properties of the grinding wheel surface. These parameters have a high ability to differentiate changes occurring as a result of the abrasion of grain vertices, their chipping or loading of the grinding wheel surface. The methodology of assessing the processes of abrasive grain wear and changes in the shape and dimensions of the grinding wheel, taking into account the probabilistic features of the grinding process, was formulated. The directions for the development of abrasive tools are presented, pointing to hybrid tools with a multi-phase structure, modified by additions of abrasive aggregates. A new research direction has also been formulated on the use of additive technology to produce specialised abrasive tools, including those with built-in process sensors.

show abstract

“…Anomaly detection is a critical task in several applications, such as fraud detection [19], video surveillance [25], industrial defects [4], and medical image analysis [24], among others. In addition, it can be used as a preprocessing step in a machine learning system.…”

Section: Introductionmentioning

confidence: 99%

LEAN-DMKDE: Quantum Latent Density Estimation for Anomaly Detection

Gallego-Mejia¹,

Bustos-Brinez²,

González³

2022

Preprint

View full text Add to dashboard Cite

This paper presents an anomaly detection model that combines the strong statistical foundation of density-estimation-based anomaly detection methods with the representation-learning ability of deeplearning models. The method combines an autoencoder, for learning a low-dimensional representation of the data, with a density-estimation model based on random Fourier features and density matrices in an end-to-end architecture that can be trained using gradient-based optimization techniques. The method predicts a degree of normality for new samples based on the estimated density. A systematic experimental evaluation was performed on different benchmark datasets. The experimental results show that the method performs on par with or outperforms other state-of-the-art methods. Keywords anomaly detection • density matrix • random Fourier features • kernel density estimation • approximations of kernel density estimation • quantum machine learning • deep learning

show abstract

Statistical approaches for semi-supervised anomaly detection in machining

Cited by 12 publications

References 15 publications

Tooling systems with integrated sensors enabling data based process optimization

Tooling systems with integrated sensors enabling data based process optimization

Technological Aspects of Variation in Process Characteristics and Tool Condition in Grinding Process Diagnostics

LEAN-DMKDE: Quantum Latent Density Estimation for Anomaly Detection

Contact Info

Product

Resources

About