The wear of cutting tools, cutting force determination, surface roughness variations and other machining responses are of keen interest to latest researchers. The variations of these machining ...responses results in change in dimensional accuracy and productivity upto great extent. In addition, an excessive increase in wear leads to catastrophic consequences, exceeding the tool breakage. Therefore, this article discusses the online trend of modern approaches in tool condition monitoring while different machining operations. For this purpose, the effective use of new sensors and artificial intelligence (AI) is considered and followed during this holistic review work. The sensor systems used for monitoring tool wear are dynamometers, accelerometers, acoustic emission sensors, current and power sensors, image sensors, other sensors. These systems allow to solve the problem of automation and modeling of technological parameters of the main types of cutting, such as turning, milling, drilling and grinding. The modern artificial intelligence methods are considered, such as: Neural networks, Image recognition, Fuzzy logic, Adaptive neuro-fuzzy inference systems, Bayesian Networks, Support vector machine, Ensembles, Decision and regression trees, k-nearest neighbors, Artificial Neural Network, Markov model, Singular Spectrum Analysis, Genetic algorithms. Discussions also includes the main advantages, disadvantages and prospects of using various AI methods for tool wear monitoring. Moreover, the problems and future directions of the main processing methods using AI models are also highlighted.
•Data driven condition monitoring is limited to cases where in-service data exists.•Simulation models are proposed as a data source for machine learning training data.•Simulation-trained classifiers ...achieve up to 94% accuracy on four exp. datasets.
Increasing the accuracy of mechanical fault detection has the potential to improve system safety and economic performance by minimizing scheduled maintenance and the probability of unexpected system failure. Advances in computational performance have enabled the application of machine learning algorithms across numerous applications including condition monitoring and failure detection. Past applications of machine learning to physical failure have relied explicitly on historical data, which limits the feasibility of this approach to in-service components with extended service histories. Furthermore, recorded failure data is often only valid for the specific circumstances and components for which it was collected. This work directly addresses these challenges for roller bearings with race faults by generating training data using information gained from high resolution simulations of roller bearing dynamics, which is used to train machine learning algorithms that are then validated against four experimental datasets. Several different machine learning methodologies are compared starting from well-established statistical feature-based methods to convolutional neural networks, and a novel application of dynamic time warping (DTW) to bearing fault classification is proposed as a robust, parameter free method for race fault detection.
This letter utilizes the turning point of preheating stage of case temperature T c as an efficient indicator for condition monitoring of IGBT modules, which could be extracted from dynamic response ...of T c . Theoretical analysis shows that the turning point will increase with aging process, and its effectiveness is experimentally verified. This method only requires temperature sensor to measure T c without strict measurement requirements. Thus it brings negligible hardware burden, which is low cost and has more flexibility in online application. Besides, the installation of case temperature sensor does not cause intrusive change to the original system, and the measurement does not affect the normal operation of converter. These features make it more practical for non-invasive and real-time monitoring.
Condition-based monitoring of rotating machines requires robust features for accurate fault diagnosis, which is indeed directly linked to the quality of the features extracted from the signals. This ...is especially true for vibration data, whose quasi-stationary nature implies that the quality of frequency domain extracted features depends on the signal-to-noise ratio (SNR) condition, operating condition variations, and data segmentation. This paper presents a novel statistical spectral analysis, which leads to highly robust fault diagnosis with poor SNR conditions, different time-window segmentation, and different operating conditions. The amplitudes of spectral contents of the quasi-stationary time vibration signals are sorted and transformed into statistical spectral images. The sort operation leads to the knowledge of the empirical cumulative distribution function (ECDF) of the amplitudes of each frequency band. The ECDF provides a robust statistical information of the distribution of the amplitude under different SNR and operating conditions. Statistical metrics have been adopted for fault classification, by using the ECDFs obtained from the spectral images as fault features. By applying simple statistical metrics, it is possible to achieve fault diagnosis without classifier training, saving both time and computational costs. The proposed algorithm has been tested using a vibration data benchmark: comparison with state-of-the-art fault diagnosis algorithms shows promising results.
For intelligent condition monitoring (ICM) tasks of machine tools (MTs), physics-based and data-driven models typically suffer from two major challenges that constraint their applicability: first, ...the complex machining parameters set up along with the incompleteness of physics-based models and second, the limited representation ability of small-scale dataset for data-driven models. Considering that it is impractical for the cases of MTs to obtain sufficient scale and well-balanced dataset due to unaffordable specimen cost and strict manufacturing schedule. Accordingly, this paper proposes a new physics-informed scaling evolutionary transformer network, abbreviated as PIS-ETN, to incorporate prior knowledge into the ICM model. Specifically, it mainly includes three parts. First, a texture digital twin (TDT) model is designed to exploit prior knowledge from machining parameters and semi-observable sensor information. Secondly, a texture knowledge embedding module is designed to enhance representation capability. Thirdly, the Pareto-optimal solution space is adopted for further architecture optimization. The experiments indicate that the designed TDT model can effectively provide rich prior empirical knowledge for the designed scaling lightweight model and accelerate model convergence. The proposed lightweight architecture with its Pareto optimal training strategy shows promising fine-grained texture representation ability.
•AE mechanisms in RE systems were firstly classified.•AE sensors, signal processing and fault diagnosis and leading advances across representative RE fields were summarized.•Challenges and ...development trends of AE in RE systems were proposed.
Renewable energy (RE) does not pollute environment at the point of energy generation, and generally has a much lower pollution footprint than traditional energy from installing to decommissioning, and can diversify the power generation technology. Because of the high operation and maintenance (O&M) costs, it is necessary to build remote, online, credible monitoring and inspection techniques. Acoustic emission (AE) technology is effective and efficient to monitor and detect mechanical damage, deterioration, and failure, etc. Over the recent years, a remarkable number of scientific papers demonstrate the capability of AE in nondestructive testing (NDT), structure health monitoring (SHM), condition monitoring (CM) and fault diagnosis for RE generation, transmission, transformation and storage systems. Most of work focusing on detection principle, sensor design, signal processing and diagnosis has provided a lot of valuable contributions for academic and industrial fields. Nevertheless, all this valuable information is scattered over many sub-fields of literature, and the knowledge is not systematic. This paper is dedicated to analyze the different AE principles in RE systems, and to comprehensively summarize and clearly highlight the advanced methods and challenges. Development trends in research, application and standard are also discussed and suggested.
•A novel method for determining the reconstructed signal based on the relative change rate of singular value kurtosis (SVK) is proposed in order to determine the optimal reconstructed order of ...SVD.•The FBE method was improved based on the maximum kurtosis principle, which is named as OFBE.•The combination of SVD-SVK and OFBE is applied to the fault diagnosis of rolling bearing, so a new method of fault feature extraction, named SVD-SVK-OFBE, is proposed.•The search domain of bandwidth parameter of OFBE is discussed. In conclusion, the method proposed in this paper has a good effect, and has a good practical value.•The advantages of the proposed method over other methods, in terms of feature extraction, are verified.
Singular value decomposition (SVD) is widely used in condition monitoring of modern machine for its unique advantages. A novel relative change rate of singular value kurtosis (SVK) is proposed in order to determine the reconstructed order of singular values effectively. Since the bandwidth parameter of the band-pass filter designed by FBE need to be determined based on experience, obviously, there are significant deficiencies. Then, a optimized frequency band entropy (OFBE) method based on the principle of maximum kurtosis is proposed to optimize the bandwidth parameters. In addition, because the fault signal of the rolling bearing at the initial stage is very weak and submerged by ambient noise, SVD cannot extract fault features clearly, a new method for fault feature extraction of rolling bearing based on SVD and OFBE, named SVD-SVK-OFBE, is proposed. Firstly, the Hankel matrix is reconstructed from the original vibration signal in the phase space and the noise reduction is performed using SVD. Here, the relative change rate of singular value kurtosis is performed on the Hankel matrix to determine the reconstructed order. Secondly, the OFBE analysis is performed on the reconstructed signal to determine the center frequency and the bandwidth of the band-pass filter adaptively. The bandwidth of the designed band-pass filter is optimized by the kurtosis maximum principle. Thirdly, the reconstructed signal of SVD is filtered by the optimized filter, and the envelope demodulation analysis is performed on the filtered signal. Finally, the fault feature frequency is extracted and compared with the theoretical fault feature frequency to identify the fault type of the rolling bearing. The effectiveness and advantages of the method described in this paper are verified by the simulation analysis and experimental data analysis of the rolling bearing.
•Anomaly detection in condition monitoring using healthy vibrations for training.•Design of data-driven model architectures that integrate a physical/statistical understanding of different ...faults.•LSTM/SVM capable of detecting new deterministic components introduced by gear faults.•A two-step LSTM regression + one-class SVM for detecting new random components caused by bearing faults.•SVM capability in fault detection when vibrations collected using different sensors.
Fault detection is a critical step for machine condition monitoring and maintenance. With advances in machine learning technologies, automated faulty condition identification can be achieved by training an artificial intelligent (AI) model if sufficient data is available. In most practical applications, it is unlikely that enough data from faulty cases are available for supervised training of AI models for anomaly detection. This work is based on training data that are exclusively constituted of healthy signals (i.e., semi-supervised) and aims to develop an automated algorithm capable of identifying any abnormal mechanical behaviour captured by vibration measurements. A deep learning method with long short-term memory (LSTM) architectures combined with a one-class support vector machine (SVM) is used to separate abnormal data from normal vibration signals collected during an endurance test of a reduction gearbox and helicopter test flight data measured by multiple sensors situated at different locations of the aircraft. For the gearbox dataset, a detailed understanding of the physical mechanisms by which different types of faults (gear wear and bearing faults) affect the vibration signal led to the design of two anomaly detection architectures: i) an LSTM regression + one-class SVM for detecting new deterministic components introduced by gear faults; and ii) a two-step LSTM regression + one-class SVM for detecting new random components caused by bearing failures. For the helicopter dataset, which does not contain consecutive time-series, we show that the LSTM regression is not advantageous, and a better performance can be achieved by a simpler one-class SVM outlier detection based on statistical features. This work contributes to the field of machine condition monitoring by introducing a novel two-step LSTM configuration for removing deterministic components associated with dominant gear signals in the first step and removing the ‘residual deterministic’ components associated with varying gear signals in the second step.
Advanced manufacturing depends on the timely acquisition, distribution, and utilization of information from machines and processes across spatial boundaries. These activities can improve accuracy and ...reliability in predicting resource needs and allocation, maintenance scheduling, and remaining service life of equipment. As an emerging infrastructure, cloud computing provides new opportunities to achieve the goals of advanced manufacturing. This paper reviews the historical development of prognosis theories and techniques and projects their future growth enabled by the emerging cloud infrastructure. Techniques for cloud computing are highlighted, as well as the influence of these techniques on the paradigm of cloud-enabled prognosis for manufacturing. Finally, this paper discusses the envisioned architecture and associated challenges of cloud-enabled prognosis for manufacturing.
Machine condition monitoring is an emerging research domain to use monitoring data to monitor machine conditions and prevent unexpected machine failures. In our previous study, the sum of weighted ...normalized square envelope was proposed as a generalized framework of some well-known sparsity measures including kurtosis, negative entropy, smoothness index, and Gini index. This framework revealed that a main difference among these sparsity measures is that they use different weights. Consequently, the design of new weights can generate new sparsity measures. Our previous study also showed that a convex-optimization problem could be formulated to automatically design weights by a data-driven way. One attractive experimental finding was that solving the sum of weighted normalized Fourier spectrum/envelope spectrum results in informative frequency bands/fault characteristic frequencies. However, this finding was only experimentally observed based on positive optimized weights and omitted the importance of negative optimized weights. In this short communication, we revisit this work and provide insightful investigations for signs of optimized weights and mathematically prove that both positive and negative optimized weights are extremely important to distinguish fundamental frequency components and fault-generated frequency components. Three new propositions are proposed to show the ability of optimized weights for determining informative frequency bands/fault characteristic frequencies. Experimental results are provided to verify the effectiveness of our new propositions given in this short communication. The significance of this short communication is that it is helpful for engineers and scholars to quickly and effectively identify all vibration contributions from fundamental frequency components and fault-generated frequency components using positive and negative signs of optimized weights in our previous framework. Moreover, fully interpretable weights would be helpful for designing physics-informed machine learning methods for machine condition monitoring.
PDF