Normal State Data Research Articles

Field Validation of a Universally Applicable Condition-Based Maintenance System for Mud Pumps

Summary Although mud pumps are considered to be critical rig equipment, their health monitoring currently still relies on infrequent human observation and monitoring. This approach often fails to detect pump damage at an early stage, resulting in nonproductive time (NPT) and increased well construction costs when initial damage progresses and pumps go down unexpectedly and catastrophically. Automated approaches to condition-based maintenance (CBM) of mud pumps to date have failed due to the lack of a generalized solution applicable to any pump type and/or operating conditions. This paper presents a field-validated universally applicable solution to mud pump CBM. The system uses a sensor package that includes acoustic emission sensors and accelerometers in combination with anomaly detection deep learning data analysis to pinpoint any abnormal behavior of the pump and its components. The deep learning models are trained with undamaged normal state data only, and a damage score characterizing the extent of damage to the mud pump is calculated to identify the earliest signs of damage. The system can then generate alerts to notify the rig crew of the damage level of key mud pump components, prompting proactive maintenance actions. Field tests were conducted while drilling an unconventional shale well in west Texas, USA, and a geothermal well in Japan (i.e., two very different drilling operations) to verify the feasibility and general applicability of the developed pump CBM solution. Sensors were attached to pump modules, and data were collected and analyzed using the deep learning models during drilling operations. During the field tests, different hyperparameters and features were compared to select the most effective ones for identifying damage while at the same time delivering low false positive rates (i.e., false alarms during normal state pump operation). The system required only several hours of normal state data for training with no prior pump information. Moreover, it correctly identified the degradation of the pump, swabs, and valves and produced early alerts several hours (in the range of 0.5–17 hours) before actual pump maintenance action was taken by the rig crew. This generally applicable pump CBM system eliminates the environmental, health, and safety concerns that can occur during human-based observations of mud pump health and avoids unnecessary NPT associated with catastrophic pump failures. The final version of this system will be a fully self-contained magnetically attachable box containing sensors and a processor, generating simple indicators for recommending proactive pump maintenance tasks when needed.

Dual fault warning method for coal mill based on Autoformer WaveBound

The coal mill is one of the important auxiliary equipment of thermal power units. Power plant performance and reliability are greatly influenced by the coal mill. To avoid abnormal operating conditions of coal mills in time and effectively, a dual fault warning method for coal mill is proposed. Three typical faults of coal mill plugging, coal breakage and deflagration are warned by this method. Firstly, the maximum information coefficient method and double complex wavelet packet transform are employed to approximate the historical data of the coal mill. Secondly, WaveBound regularization is incorporated to improve the generalization capabilities of Autoformer. Normal state data is used in the construction of unsupervised learning models. Then, combined with a multi-step rolling prediction method, vital monitoring parameters are supervised by setting the corresponding thresholds. Furthermore, an abnormal detection method based on Wasserstein distance is developed for coal mills overall state monitoring. A dual alarm method considering univariate abnormal detection and multivariate coupling thresholds is formed. Finally, the proposed method is validated with real data from a medium-speed coal mill. The results illustrate that early warning is effectively carried out using the method.

Intelligent Complementary Multi-Modal Fusion for Anomaly Surveillance and Security System.

Recently, security monitoring facilities have mainly adopted artificial intelligence (AI) technology to provide both increased security and improved performance. However, there are technical challenges in the pursuit of elevating system performance, automation, and security efficiency. In this paper, we proposed intelligent anomaly detection and classification based on deep learning (DL) using multi-modal fusion. To verify the method, we combined two DL-based schemes, such as (i) the 3D Convolutional AutoEncoder (3D-AE) for anomaly detection and (ii) the SlowFast neural network for anomaly classification. The 3D-AE can detect occurrence points of abnormal events and generate regions of interest (ROI) by the points. The SlowFast model can classify abnormal events using the ROI. These multi-modal approaches can complement weaknesses and leverage strengths in the existing security system. To enhance anomaly learning effectiveness, we also attempted to create a new dataset using the virtual environment in Grand Theft Auto 5 (GTA5). The dataset consists of 400 abnormal-state data and 78 normal-state data with clip sizes in the 8-20 s range. Virtual data collection can also supplement the original dataset, as replicating abnormal states in the real world is challenging. Consequently, the proposed method can achieve a classification accuracy of 85%, which is higher compared to the 77.5% accuracy achieved when only employing the single classification model. Furthermore, we validated the trained model with the GTA dataset by using a real-world assault class dataset, consisting of 1300 instances that we reproduced. As a result, 1100 data as the assault were classified and achieved 83.5% accuracy. This also shows that the proposed method can provide high performance in real-world environments.

Harmony better than uniformity: A new pre-training anomaly detection method with tensor domain adaptation for early fault evaluation

Aiming to identify incipient fault as earlier as possible, early fault detection (EFD) becomes the key step to trigger predictive maintenance. To lower deployment burden, pre-training techniques are employed to provide well-initialized features before fine-tuning different EFD tasks. Current large-scale pre-training models focus on seeking uniformity features to align every task’s characteristic. But in practical engineering, degradation data is usually small-scale and under irregular noise interference. The pre-trained uniformity features are easily biased and less discriminative to early fault. To address this concern, this paper proposes a pre-training anomaly detection method with tensor domain adaptation. Running on the classical contrastive learning architecture, the proposed method first builds hypersphere-formed representation of anomaly detection rule. With the merit of tensor decomposition in extracting the intrinsic information from raw signals, the proposed method second designs a new tensorized rule adaptation mechanism to learn the task-invariant detection rule from normal state data that are used for pre-training. The pre-trained feature representation is then obtained from the feature extractor in contrastive learning via back-propagation. This paper takes rolling bearing as the validation object. Experimental results on the IEEE PHM Challenge 2012 and XJTU-SY datasets show that the proposed method can effectively improve the convergence speed and detection accuracy on different EFD tasks with shallow model, deep model and transfer learning, and also prove that the harmony on task-invariant detection rule and task-specific information facilitates the universal applicability of EFD pre-trained features in small scale. Harmony is then believed better than uniformity for the EFD pre-training.

Few-Shot Fault Diagnosis Method of Rotating Machinery Using Novel MCGM Based CNN

The existing fault diagnosis methods can achieve good results when various status fault data are available. However, the construction of the diagnosis model is often unachievable in the actual application because only normal data are available, which is actually a few-shot fault diagnosis problem. Therefore, a novel intelligent few-shot fault diagnosis method of rotating machinery based on the convolutional neural network (CNN) using virtual samples generated by the mechanism character generative model (MCGM) integrating the generative adversarial network (GAN) is proposed. The distribution pattern of common parameters that reflect the fault category is learned using the GAN and source domain fault data. Then, the normal state data of the target domain is combined with the distribution common parameters to generate virtual samples in target domain based on the MCGM. Moreover, the fault diagnosis model is trained by virtual samples based on the CNN. Finally, the proposed fault diagnosis method is validated using the laboratory bearing data, the industrial data and the public data of the rotating machinery, respectively. The results show that the proposed method achieves an average accuracy of 93.38% in the diagnostic task, exhibiting at least 4.56% better performance than other comparison methods.

Performance simulation and diagnosis of faulty states in air-cycle refrigeration systems in civil aircrafts

In the present study, performance simulation models and diagnosis method of faulty states in air-cycle refrigeration systems in civil aircrafts are developed, and the effects of the training data set on the diagnosis accuracy were analyzed. A base model for normal state simulation of air-cycle refrigeration system (ACS) was established based on the existing method with several modifications to improve accuracy and performance. Dedicated models for faulty states were developed according to the causes of fault. Based on the established models, a data set was generated. Preliminary training of multiple two-layer neural network models is performed with the existing data set, and the trained model achieved a validation accuracy of 99.00 % which confirms the feasibility of using double-layer neural network for fault identification and classification based on simulation data. The effects of the training data set to the model are analyzed, and it was concluded that: a data set of about 50,000 entries can be used to train a satisfactory diagnostic model; the proportion of normal state data needs to be reduced appropriately to improve the sensitivity of the model to faulty states; it is possible to largely reduce the number of sensors with smaller effect to the diagnosis accuracy.

Analysis of operator’s activity in control rooms of underground coal mines

The analysis for estimating operator’s activity and quantity of information in control rooms of underground coal mines is presented in this paper. Monitoring of operators in control rooms was conducted in mines Rembas, Soko, Ibar, Jasenovac and Stavalj. The network models are used during the analysis of operator’s activity in control rooms. The result of the analysis is proposed critical path network model. The application of network models improves the quality and level of research, making it more efficient. Most of the analysed information are related to normal and alarm state data received from the mine shaft and operator’s response to this data. Results are given according to actual scope of work depending on the work dynamics in underground coal mines, via histogram of changes in average number of activities per 1 h for 24-h period, then number of operator’s activities per work hour and insight into the global view of technological process in the coal mine with focus on the work scope of mine workers. Analysis of operator’s activities presented in this paper determines the relations between devices in control rooms and operators’ capabilities in order to get better ergonomic design of control rooms.

1D convolutional neural network-based adaptive algorithm structure with system fault diagnosis and signal feature extraction for noise and vibration enhancement in mechanical systems

A variety of adaptive algorithms are utilized to control the vibration and noise for mechanical systems in the industrial field, since their characteristics usually changes due to unwanted disturbance and malfunctioning. A reference signal plays an important role in signal tracking through adaptive algorithms, and well-characterized information about the tracked signal should be employed. However, it is typically difficult to determine the signal feature because operating systems generate complex signals. Moreover, when the system state changes due to malfunctions or disturbances, signal feature can also be changed, affecting the performance of adaptive algorithms. This study proposes a novel strategy based on a 1D convolutional neural network (1D CNN) for improving the signal tracking performance. 1D CNN has the ability to deal with consecutive data that is directly measured by an accelerometer attached to mechanical systems, as well as to extract the signal feature through convolutional layers. The development of proposed algorithm can be explained with following three parts: 1) Data measurement opportunities are limited, particularly because fault data is more difficult to measure than normal state data. Therefore, when performing learning, there is a problem of a lack of training data. Thus, a data generator is proposed that is based on a generative adversarial network (GAN) and a variational auto encoder (VAE). 2) In order to extract the frequency and phase and consider the system state to automatically define the reference signal and increase tracking robustness, the methodology of signal feature extraction and diagnosis is proposed based on a 1D CNN. In order to validate the tracking performance, a neural network-based signal tracking algorithm is applied to the developed structure. Furthermore, bearing and linear motion (LM) guide data from experiments is employed to confirm the versatility of the developed algorithm. Through above process, it is demonstrated that based on defined reference signal through the extracted feature, the signal tracking performance can be enhanced.

Mechanical Equipment Performance Degradation Assessment Method Based on Improved Degradation Feature Extraction and Sample Completeness

Performance degradation assessment (PDA), as an important part of health management, is playing a crucial role in evaluating the degenerate state of mechanical equipment. To evaluate the performance degradation state more effectively by using monitoring data of mechanical equipment, a PDA method based on improved degradation feature extraction technology and considering sample completeness is proposed in this paper. Firstly, multiple degradation features, including statistical features and intrinsic energy features, are extracted to construct a high-dimensional feature set calculated by time-domain analysis and the EMD method. Then, a sensitive feature set, which has significant robustness, correlation, and monotonicity in the degenerate process, is reduced and processed by the improved PCA method to obtain the final health index. Secondly, the degradation assessment model is established by considering the completeness of the samples. As for the complete sample dataset, which has both normal and failure state data, a logistic regression model (LRM) is built to assess the performance degradation status. In addition, an SVDD model is established for the incomplete sample dataset, which only has normal state data. Finally, the reliability of the proposed method is verified by the public XJTU-SY dataset, and the comparative experiment further verifies the superiority of this method.

데이터 불균형 대응 믹스업 데이터를 이용한 전이 학습 전동기 고장 진단 시스템의 성능 향상

Most of the data obtained for motor fault diagnosis is the data of normal motors. Therefore, the model is trained using normal data collected in the field and fault data collected in experiment. In addtion, transfer learning is used for fast learning convergence. Fault diagnosis performance may be reduced by the number of unbalanced data in each state during transfer learning. In this study, a transfer learning method using mix-up data is proposed to solve this data imbalance. A mix-up is a technique of adding insufficient data to the data with clear decision boundaries, and can be used to balance a large number of normal state data with a small number of fault-state data. The proposed method is compared with GAN, another technique of data augmentation. As a result of the experiment, it was confirmed that the proposed mix-up technique is a method to improve the performance of fault diagnosis.

High-fidelity positive-unlabeled deep learning for semi-supervised fault detection of chemical processes

With the rapid development of the modern chemical process industry, process monitoring techniques have been investigated to enhance the loss prevention capability. Fault detection which attempts to detect whether an abnormal condition has happened is an essential step of process monitoring. Supervised learning based fault detection methods usually are not feasible in practical industry situations, because their most important prerequisite that adequate data labels are accessible is extremely hard to meet. In real industrial situations, a common scenario is that a few normal state data are labeled while no fault state data are labeled, and this is a natural fit for the positive-unlabeled (PU) learning. In this work, a three-step high-fidelity PU (THPU) approach based on deep learning is proposed for semi-supervised fault detection of chemical processes. A self-training nearest neighbors (STNN) algorithm is designed in Step 1 to conservatively generate a small reliable positive dataset by extracting data from the unlabeled training dataset. A positive and negative data recognition (PNDR) algorithm is designed in Step 2 to augment the reliable positive dataset and generate a reliable negative dataset based on the remaining unlabeled dataset. A convolutional neural network is trained in Step 3 based on the above reliable datasets as the binary classifier for online fault detection. The benchmark simulated Tennessee Eastman process dataset and a real industrial hydro-cracking process dataset are utilized to illustrate the effectiveness and robustness of the proposed THPU approach. The experimental results demonstrate the superiority of the proposed approach compared to the other competing PU learning approaches and supervised fault detection models.

Low-Temperature Thermal Conductivity of the Two-Phase Superconductor CeRh2As2

CeRh2As2is a rare unconventional superconductor (Tc= 0.26 K) characterized by two adjacent superconducting phases for a magnetic fieldH‖c-axis of the tetragonal crystal structure. Antiferromagnetic order, quadrupole-density-wave order (T0= 0.4 K) and the proximity of this material to a quantum-critical point have also been reported: The coexistence of these phenomena with superconductivity is currently under discussion. Here, we present thermal conductivity and electrical resistivity measurements on a single crystal of CeRh2As2between 60 mK and 200 K and in magnetic fields (H‖c) up to 8 T. Extrapolation of our normal-state data to zero temperature validates the Wiedemann-Franz law within the error bars. TheTdependence of the thermal conductivityκ(T) shows a pronounced drop belowTcwhich is also field dependent and thus interpreted as the signature of superconductivity. However, the large residual resistivity and the lack of sharp anomalies inκ(T) at the expected transition temperatures clearly indicate that samples of much higher purity are required to gain more information about the superconducting gap structure.

Anomaly detection model of mooring system based on LSTM PCA method

Aiming at the difficulty of damage detection of the mooring system, this paper proposes an anomaly detection framework combined with statistical analysis and neural network methods. Specifically, based on the real design parameters of a semi-submersible platform, a hydrodynamic model was established. Taking the damage degree of 5%, 10%, 15%, and 20% as examples, the analysis of the responses was carried out when the anchor chain was damaged in different positions. And then, based on the Long Short-Term Memory (LSTM) network and normal state data, a joint prediction model of the platform's six-degree-of-freedom (6DoF) motion was established considering the marine environment loads. The motion prediction error was only 0.0078. The Principle Component Analysis (PCA) together with the prediction residual sequences was tailored and employed to build the anomaly detection model. The Upper Confident Level (UCL) of the model was also given. The recognition accuracy reached 100% which can guide the safe service of the platform.

A Study on Failure Prediction Using Time Series Data of Hydraulic Excavator

Since unexpected machine failures are huge losses for users, maintenance activities are essential. If the failures can be predicted in advance using a supervised learning, the machines can be maintained before they break down and some failures can be prevented. However, although a large number of failure data are required to predict failures using a supervised learning, failures rarely occur in the actual field. In this study, we propose to detect the failure of a hydraulic excavator using an autoencoder, which is an unsupervised learning. By using the autoencoder to model normal state data, the failure can be predicted in advance. This paper shows the results of evaluating failure predictions using the LSTM (Long Short-Term Memory) autoencoder model for actual failure of hydraulic excavators.

An early warning method for mechanical fault detection based on adversarial auto-encoders

The vibration signal of mechanical equipment is non-linear and non-stationary, and it is difficult to fully reflect the operation state of equipment through traditional fixed threshold alarm method. While the early warning method based on multi-feature parameter fusion relies on manual experience to extract features, which is difficult to ensure the accuracy of the extracted features and cannot achieved good early warning effect. To solve this problem, a feature self-learning method based on adversarial auto-encoders is proposed in this paper, which encodes high-dimensional monitoring data in normal state into low-dimensional vectors with certain statistical laws and uses it as a benchmark to detect abnormalities in the operating state of the equipment in time by measuring the difference between the encoded features of real-time monitoring data and the benchmark. The actual application cases of reciprocating compressors show that the proposed method can detect the weak signs of equipment fault at the early stage, and realize early warning. At the same time, by comparing with the Auto-Encoders network-based warning method and the Dirichlet process mixture model-based warning method, It is verified that the method in this paper has more advantages in terms of warning accuracy and warning time.

Novel K-Medoids Based SMOTE Integrated With Locality Preserving Projections for Fault Diagnosis

In the field of the fault diagnosis of industrial processes, there are many problems in process data, such as missing critical fault data, high repeatability of normal state data, and poor representation of faults data, which may reduce the accuracy of fault diagnosis. In this paper, a novel K-medoids based synthetic minority over-sampling technique that combines locality preserving projections (KMS-LPP), is proposed for fault diagnosis. First, the synthetic minority sampling technology (SMOTE) is designed based on K-medoid to generate minority fault samples to address the imbalanced problem of data. Second, to extract the key fault-relevant features and reserve the local structure information at the same time, the manifold learning (ML) approach locality preserving projections (LPP) is performed to reduce the dimensionality of data. Finally, the Adaboost. M2, as an ensemble classifier, is conducted for fault classification. Simulations of the Tennessee Eastman process (TEP) are performed for validation of the performance of the presented KMS-LPP method. The obtained results show that KMS-LPP has enhanced the performance of fault diagnosis due to its higher accuracy compared with traditional oversampling and feature extraction methods, which indicates the effectiveness of KMS-LPP.

A Deep Autoencoder-Based Convolution Neural Network Framework for Bearing Fault Classification in Induction Motors.

Fault diagnosis and classification for machines are integral to condition monitoring in the industrial sector. However, in recent times, as sensor technology and artificial intelligence have developed, data-driven fault diagnosis and classification have been more widely investigated. The data-driven approach requires good-quality features to attain good fault classification accuracy, yet domain expertise and a fair amount of labeled data are important for better features. This paper proposes a deep auto-encoder (DAE) and convolutional neural network (CNN)-based bearing fault classification model using motor current signals of an induction motor (IM). Motor current signals can be easily and non-invasively collected from the motor. However, the current signal collected from industrial sources is highly contaminated with noise; feature calculation thus becomes very challenging. The DAE is utilized for estimating the nonlinear function of the system with the normal state data, and later, the residual signal is obtained. The subsequent CNN model then successfully classified the types of faults from the residual signals. Our proposed semi-supervised approach achieved very high classification accuracy (more than 99%). The inclusion of DAE was found to not only improve the accuracy significantly but also to be potentially useful when the amount of labeled data is small. The experimental outcomes are compared with some existing works on the same dataset, and the performance of this proposed combined approach is found to be comparable with them. In terms of the classification accuracy and other evaluation parameters, the overall method can be considered as an effective approach for bearing fault classification using the motor current signal.

The Effect of Partial Substitution of Non-Magnetic Impurity Zn on the Magnetic Moments of Eu<sub>1.88</sub>Ce<sub>0.12</sub>Cu<sub>1-y</sub>Zn<sub>y</sub>O<sub>4+α-δ</sub>

Structure and magnetic properties of electron-doped superconducting cuprates have been investigated in order to study the effect of magnetic impurity to its physical properties. Here, we reported structure and magnetic properties of Eu1.88Ce0.12Cu1-yZnyO4+α-δ (ECCZO) with y = 0 and 0.03. The properties of ECCZO have been studied from X-ray diffraction data and temperature dependence of magnetic susceptibility data, to elucidate the effect of partial substitution of non-magnetic impurity Zn for Cu to its structure, Tc and the value of magnetic moments per unit volume extracted from susceptibility data in normal state. Magnetic-susceptibility measurements were carried out down to 2 K on-field cooling at 5 Oe for Eu1.88Ce0.12Cu1-yZnyO4+a-d with y = 0 and 0.03. For ECCZO sample with y = 0 and d = 0.0669 indicated the change of magnetic behavior from paramagnetic to diamagnetic below 12 K which is addressed to the Tc onset of this samples. Diamagnetic behavior is observed starting from about 12 K. Above 12 K, all samples show paramagnetic behavior with the values of the magnetic moment in every volume unit increased with increasing Zn.

Wind Turbine Anomaly Detection Efficient Deployment Using Representation Learning of Normal State Data

Wind Turbine Anomaly Detection Efficient Deployment Using Representation Learning of Normal State Data

Optical and hidden transport properties of BaFe1.91Ni0.09As2 film

Optical spectroscopy was used to study the electrodynamics and hidden transport properties of a BaFe1.91Ni0.09As2 thin superconducting (SC) film. We analyzed the normal state data using a Drude–Lorentz model with two Drude components: one narrow (D1) and another broad one (D2). In the SC state, two gaps with –2.0 and –4.3 are formed from the narrow component D1 while the broad component D2 remains ungapped. The calculated total DC resistivity of the film and the low-temperature scattering rate for the narrow Drude component show a hidden Fermi-liquid behavior. The change of total electron–boson coupling (λ tot) and representative energy (Ω0) in the normal state with respect to the SC state is typical of other iron-based materials as well as high-temperature superconducting (HTSC) cuprates.