Raw Monitoring Data Research Articles

Multi-Head Self-Attention-Based Fully Convolutional Network for RUL Prediction of Turbofan Engines

Remaining useful life (RUL) prediction is widely applied in prognostic and health management (PHM) of turbofan engines. Although some of the existing deep learning-based models for RUL prediction of turbofan engines have achieved satisfactory results, there are still some challenges. For example, the spatial features and importance differences hidden in the raw monitoring data are not sufficiently addressed or highlighted. In this paper, a novel multi-head self-Attention fully convolutional network (MSA-FCN) is proposed for predicting the RUL of turbofan engines. MSA-FCN combines a fully convolutional network and multi-head structure, focusing on the degradation correlation among various components of the engine and extracting spatially characteristic degradation representations. Furthermore, by introducing dual multi-head self-attention modules, MSA-FCN can capture the differential contributions of sensor data and extracted degradation representations to RUL prediction, emphasizing key data and representations. The experimental results on the C-MAPSS dataset demonstrate that, under various operating conditions and failure modes, MSA-FCN can effectively predict the RUL of turbofan engines. Compared with 11 mainstream deep neural networks, MSA-FCN achieves competitive advantages in terms of both accuracy and timeliness for RUL prediction, delivering more accurate and reliable forecasts.

A novel method for settlement imputation and monitoring of earth-rockfill dams subjected to large-scale missing data

Various missing values inevitably exist in the settlement monitoring data of prolonged operation of earth-rockfill dams, which delays or even interrupts the dam structure analytical procedure. This study aims to develop a reliable method for addressing the missing data problem and monitoring earth-rockfill dam settlement behavior. First, an imputation model, support vector regression based on the finite element method and time input (FEMTSVR), is proposed and offers superior performance when capturing complex nonlinear mappings from environmental variables to settlement on small sample data. Herein, an improved particle swarm optimization algorithm (IPSO) is developed, realizing the nonlinear adjustment of weights and parameter reduction during hyperparameter optimization. Then, this study establishes a sequential prediction model based on gate recurrent unit (GRU) networks to monitor dam behavior on the imputed and complete dataset. Eventually, the proposed method is evaluated using real-world earth-rockfill dam monitoring data with the help of statistical indicators, demonstrating its efficiency in monitoring settlement data subjected to large-scale missingness. This study provides a robust database for dam structural health monitoring, while also providing a promising framework for the safety assessment of other civil or hydraulic engineering based on raw monitoring data.

Aircraft Engine Fault Diagnosis Model Based on 1DCNN-BiLSTM with CBAM

As the operational status of aircraft engines evolves, their fault modes also undergo changes. In response to the operational degradation trend of aircraft engines, this paper proposes an aircraft engine fault diagnosis model based on 1DCNN-BiLSTM with CBAM. The model can be directly applied to raw monitoring data without the need for additional algorithms to extract fault degradation features. It fully leverages the advantages of 1DCNN in extracting local features along the spatial dimension and incorporates CBAM, a channel and spatial attention mechanism. CBAM could assign higher weights to features relevant to fault categories and make the model pay more attention to them. Subsequently, it utilizes BiLSTM to handle nonlinear time feature sequences and bidirectional contextual feature information. Finally, experimental validation is conducted on the publicly available CMAPSS dataset from NASA, categorizing fault modes into three types: faultless, HPC fault (the single fault), and HPC&Fan fault (the mixed fault). Comparative analysis with other models reveals that the proposed model has a higher classification accuracy, which is of practical significance in improving the reliability of aircraft engine operations and for Remaining Useful Life (RUL) prediction.

Moisture content online detection system based on multi-sensor fusion and convolutional neural network.

To monitor the moisture content of agricultural products in the drying process in real time, this study applied a model combining multi-sensor fusion and convolutional neural network (CNN) to moisture content online detection. This study built a multi-sensor data acquisition platform and established a CNN prediction model with the raw monitoring data of load sensor, air velocity sensor, temperature sensor, and the tray position as input and the weight of the material as output. The model's predictive performance was compared with that of the linear partial least squares regression (PLSR) and nonlinear support vector machine (SVM) models. A moisture content online detection system was established based on this model. Results of the model performance comparison showed that the CNN prediction model had the optimal prediction effect, with the determination coefficient (R2) and root mean square error (RMSE) of 0.9989 and 6.9, respectively, which were significantly better than those of the other two models. Results of validation experiments showed that the detection system met the requirements of moisture content online detection in the drying process of agricultural products. The R2 and RMSE were 0.9901 and 1.47, respectively, indicating the good performance of the model combining multi-sensor fusion and CNN in moisture content online detection for agricultural products in the drying process. The moisture content online detection system established in this study is of great significance for researching new drying processes and realizing the intelligent development of drying equipment. It also provides a reference for online detection of other indexes in the drying process of agricultural products.

Automated data labeling of building automation systems using time series data and conditional probabilities

As energy efficiency demands increase, buildings get smarter, and the amount of data to analyze grows, where each building device may generate multiple data streams. These extensive quantities of monitoring data serve as a great opportunity for detecting anomalies in building automation systems and for optimizing their control. However, each building usually uses a custom format for data labels, therefore requiring an individual data label analysis per building. This makes the conceptually manageable task of detecting energy systems from the raw data increasingly complex and error-prone, which is a further hurdle that any building operation optimizer must resolve. This paper presents a methodology for automatically categorizing and labeling raw monitoring data from building automation systems. Using statistical features of the data, the method checks which data streams follow which known building operation rules and patterns. Therefore, an initial labeling of the data streams takes place. Furthermore, examining the correlation between the data streams indicates possible related system components using the concept of conditional probability. As a use case for the methodology, unlabeled data from a real building automation system are examined. The results show that, using unlabeled time series, data types from certain sensors and actuators can be reliably identified. The proposed methodology could therefore simplify the implementation of energy applications such as operation optimization and fault detection of building automation systems

Monitoring and controlling of the micro-environment in a special exhibition in the Shanghai Museum

Monitoring and controlling of the showcase environment in a special exhibition in Shanghai Museum are described in this work. The long show time started on March 2022 and ended on January 2023. Various items were shown in this exhibition with different combination modes. Air conditioners, humidifiers and three kinds of passive humidity controlling methods were applied to this exhibition. Massive raw monitoring data are given and some series of data are transformed into specific values to conduct various analyses. Factors including outdoor climate, exhibition hall structure, attendance, showcase structure, tightness and volume, monitor position and object combination modes are discussed in order to assess their effect on monitoring and controlling of temperature and relative humidity. The relation between the average value of temperature (or relative humidity) fluctuation in the whole exhibition period and the corresponding range of maximum temperature (or relative humidity) fluctuation in any 24 h is analyzed in order to obtain the probability of abnormally high fluctuation. The results show that the temperature control target is reached in general, while some unreasonable showcase structures and settings are responsible for the difficulty in monitoring and controlling of relative humidity. By conducting the analysis of big continuous historical data and learning the average values and the ranges reached, it is possible to assess the current environmental risk and predict the future risk in a simple way.

An efficient astronomical seeing forecasting method by random convolutional Kernel transformation

Astronomical seeing is the primary indicator of the observation quality of large-scale optical telescopes. Forecasting astronomical seeing is an essential and challenging task in the operation of large optical telescopes, as it can help observers more efficiently arrange observation tasks as well as providing hints for mechanism optimization of optical telescopes. In this paper, we propose an efficient data-driven astronomical seeing forecasting method by leveraging the environment monitoring data from multiple sensors deployed with optical telescopes. The basic idea is to use a large number of random yet fixed convolution kernels to extract features from the original data, and then train a simple tree-based predictor based on these features. In particular, the raw monitoring data is transformed by differential operations, enabling a more explicit representation of spatial and temporal features; then an elaborate convolution mechanism, including the convolution kernel and the pooling operator, is designed to extract a variety set of features of the input time series, which serve as the foundation for constructing the time series forecasting model. In contrast to existing deep learning models, there is no need to calculate and back-propagate the gradients during the training procedure, making the proposed method much more efficient. We collect real-world monitoring data from Large Sky Area Multi-Object Fiber Spectroscopy Telescope (LAMOST), one of the world’s largest optical reflecting telescopes, and conduct extensive empirical evaluations. The experimental results demonstrate that the proposed method can achieve state-of-the-art prediction accuracy, i.e., 0.1125 in MAE, 0.0252 in MSE, and 0.0322 in MAPE, outperforming most deep-learning-based forecasting models. Moreover, it surpasses both machine learning and deep learning approaches in terms of training efficiency with a significant margin.

Real-Time Malfunction Detection of Maglev Suspension Controllers

This study aims to develop a track-side online monitoring system for malfunction detection in the suspension controllers of maglev trains during their in-service operation. The hardware module of the system includes two arrays of accelerometers deployed on an F-type rail and a data acquisition unit. The software module of the system consists of codes for three functions: (i) the identification of time intervals in relation to the passage of each suspension controller via synchrosqueezing transform; (ii) the extraction of a feature index (FI) sequence synthesized by modulating the response amplitude, frequency, and running speed; and (iii) the formulation of a Bayesian dynamic linear model for real-time malfunction detection in maglev suspension controllers. For verification of the proposed monitoring system and malfunction detection algorithm, full-scale tests have been conducted on an maglev test line using the devised system, where a maglev train was run at different speeds with malfunction occurring in the suspension controllers. The malfunction detection results of the proposed approach are exemplified via comparison with the recorded suspension gaps after the trial run of the maglev train. The fidelity of the results obtained using the extracted FI sequence and using the raw monitoring data are compared. The superiority of the proposed malfunction detection algorithm is also discussed via comparison with the results of the different train speeds.

Real-time prediction of mechanical behaviors of underwater shield tunnel structure using machine learning method based on structural health monitoring data

Predicting the mechanical behaviors of structure and perceiving the anomalies in advance are essential to ensuring the safe operation of infrastructures in the long run. In addition to the incomplete consideration of influencing factors, the prediction time scale of existing studies is rough. Therefore, this study focuses on the development of a real-time prediction model by coupling the spatio-temporal correlation with external load through autoencoder network (ATENet) based on structural health monitoring (SHM) data. An autoencoder mechanism is performed to acquire the high-level representation of raw monitoring data at different spatial positions, and the recurrent neural network is applied to understanding the temporal correlation from the time series. Then, the obtained temporal-spatial information is coupled with dynamic loads through a fully connected layer to predict structural performance in next 12 h. As a case study, the proposed model is formulated on the SHM data collected from a representative underwater shield tunnel. The robustness study is carried out to verify the reliability and the prediction capability of the proposed model. Finally, the ATENet model is compared with some typical models, and the results indicate that it has the best performance. ATENet model is of great value to predict the real-time evolution trend of tunnel structure. • A more credible prediction model is presented to solve practical engineering problem. • Timely prediction for the subsequent state and precise perception for anomaly in advance. • Interdisciplinary research of machine learning in the analysis of structural mechanical behaviors.

Core loss: Mining core samples efficiently for robust machine anomaly detection against data pollution

Machine anomaly detection is the task of detecting machine abnormal condition via the collected monitoring data. Recently, increasing attention has focused on autoencoder (AE) based unsupervised anomaly detection (UAD) for mechanical equipment. Typically, the UAD is based on the assumption that all of the training data are normal. In real scenarios, however, the raw monitoring data may be polluted by abnormal data due to machine failures, environmental noise, sensor failures, etc. Without effective regularization, AE-based methods would overfit these polluted data. To address this issue, we design an effective loss, called core loss, which can perform AE-based UAD in a both model-agnostic and end-to-end manner under data pollution. First, we experimentally observe that the AE shows the self-clean characteristic (SCC) under data pollution, i.e., the network will prioritize learning the normal data from the polluted training data. Next, we focus on the core samples, which make up a clean and representative subset of the unlabeled data. Based on the SCC, we propose the core prior that reconstruction errors in the middle with high density correspond to the core samples. Finally, to enhance the SCC, the core loss exploits the core prior to effectively mine and learn the core samples. The experimental results show that the core loss can effectively improve the performance of different network structures on both clean and polluted data. The corresponding Python codes are available at https://github.com/albertszg/Coreloss.

Data-driven Seeing Prediction for Optics Telescope: from Statistical Modeling, Machine Learning to Deep Learning Techniques

Predicting seeing of astronomical observations can provide hints of the quality of optical imaging in the near future, and facilitate flexible scheduling of observation tasks to maximize the use of astronomical observatories. Traditional approaches to seeing prediction mostly rely on regional weather models to capture the in-dome optical turbulence patterns. Thanks to the developing of data gathering and aggregation facilities of astronomical observatories in recent years, data-driven approaches are becoming increasingly feasible and attractive to predict astronomical seeing. This paper systematically investigates data-driven approaches to seeing prediction by leveraging various big data techniques, from traditional statistical modeling, machine learning to new emerging deep learning methods, on the monitoring data of the Large sky Area Multi-Object fiber Spectroscopic Telescope (LAMOST). The raw monitoring data are preprocessed to allow for big data modeling. Then we formulate the seeing prediction task under each type of modeling framework and develop seeing prediction models through using representative big data techniques, including ARIMA and Prophet for statistical modeling, MLP and XGBoost for machine learning, and LSTM, GRU and Transformer for deep learning. We perform empirical studies on the developed models with a variety of feature configurations, yielding notable insights into the applicability of big data techniques to the seeing prediction task.

Data-driven dynamic failure assessment of subsea gas pipeline using process monitoring data

Monitoring data of process variables can reflect the system condition, which can be used to perform real-time failure assessment of process operations. This paper presents a data-driven methodology for dynamic failure assessment of process operations using process monitoring data, and it is comprised of Variational Mode Decomposition (VMD) implemented with Hierarchical Bayesian Model (HBM). The monitoring data of process variables tend to be non-stationary and complex. Firstly, VMD is used to separate the noises from monitoring data to obtain the true signals. Subsequently, the failure of process operations is estimated using HBM with the true signals as the inputs. A case study of a subsea gas pipeline system is used to illustrate the methodology. It is observed that the methodology can capture the failure time of subsea gas pipeline by using raw monitoring data of process variables. It can be a useful tool to support the online failure assessment of process operations.

Automated location of steel truss bridge damage using machine learning and raw strain sensor data

Strategic major infrastructure ageing requires structural health monitoring usage to avoid critical safety issues and disasters. Machine Learning can be a valuable tool to automate the process of analysing raw monitoring data. Usually, frequency domain damage-sensitive features are extracted with data pre-processing procedures; thus these features are used as input for classification or regression problems.This paper describes a method of locating damage in steel truss railway bridges through machine learning classification tools, enabling automatic analysis of raw strain sensors signals without any pre-processing or preliminary feature extraction. Data were generated by simulating different damage scenarios with a finite element software, and then were processed by two machine learning classification tools: (a) the K-nearest Neighbours was adopted with the Dynamic Time Warping algorithm metric to select the most informative features; (b) a model suitable for high-dimensional data analysis, known as the Convolutional Neural Network, was then trained to classify strain sensors time series. The results indicate that the method applied can detect damages with an accuracy of 93% and is suitable for structural health monitoring.

Signal response of the Swiss plate geophone monitoring system impacted by bedload particles with different transport modes

Abstract. Controlled experiments were performed to investigate the acoustic signal response of the Swiss plate geophone (SPG) system impacted by bedload particles varying in size, impact angle, and transport mode. The impacts of bedload particles moving by saltation, rolling, and sliding were determined by analyzing the experimental videos and corresponding vibration signals. The finite element method (FEM) was utilized to construct a numerical model of the SPG system and to simulate the signals triggered by a quartz sphere hitting the plate at impact angles ranging from 0∘ to 90∘. For a particle impact on the bed or on the geophone plates, the signature of the generated signal in terms of maximum amplitude, number of impulses, and centroid frequency was extracted from the raw monitoring data. So-called signal packets were determined by performing a Hilbert transform of the raw signal. The number of packets was calculated for each transport mode and for each particle size class, with sizes ranging from 28.1 to 171.5 mm. The results show how the number of signal impulses per particle mass, the amplitude of the signal envelope, and the centroid frequency change with increasing particle size, and they also demonstrate the effect of bedload transport mode on the signal response of the SPG system. We found that there is a general increase in the strength of the signal response or in the centroid frequency when the transport mode changes from sliding to rolling to saltation. The findings of this study help us to better understand the signal responses of the SPG system for different bedload transport modes, and may also contribute to an improvement of the procedure to determine bedload particle size from the SPG signal.

Reproducible determination of transpulmonary pressures

Oesophageal pressures, as measured in an oesophageal balloon catheter, are a validated substitute for pleural pressures.Transpulmonary pressures, indispensable to improve our understanding of ventilatory physiology, are therefore typically calculated as the difference between airway and oesophageal pressures.The oesophageal pressure signal, however, features a superimposed oscillation due to cardiac motion, not representative for pleural pressure. Additionally, oesophageal contractions or surgical manipulation can alter the signal. In practice, transpulmonary pressures are therefore manually determined from the pressure-time graphic by visual inspection of the waves and averaging a limited number of samples.We suggest an approach to extract the end-expiratory transpulmonary pressure from the raw monitoring data.•Our approach reproducibly determines end-expiratory transpulmonary pressures at a given level of set positive end-expiratory pressure at the ventilator.•Our approach ignores surgical disturbance and cardiac oscillations in the oesophageal pressure signal.

DRLSTM: A dual-stage deep learning approach driven by raw monitoring data for dam displacement prediction

Dam displacement is an important indicator of the overall dam health status. Numerical prediction of such displacement based on real-world monitoring data is a common practice for dam safety assessment. However, the existing methods are mainly based on statistical models or shallow machine learning models. Although they can capture the timing of the dam displacement sequence, it is difficult to characterize the complex coupling relationship between displacement and multiple influencing factors (e.g., water level, temperature, and time). In addition, input factors of most dam displacement prediction models are artificially constructed based on modelers’ personal experience, which lead to a loss of valuable information, thus prediction power, provided by the full set of raw monitoring data. To address these problems, this paper proposes a novel dual-stage deep learning approach based on one-Dimensional Residual network and Long Short-Term Memory (LSTM) unit, referred to herein as the DRLSTM model. In the first stage, the raw monitoring sequence is processed and spliced with convolution to form a combined sequence. After the timing information is extracted, the convolution direction is switched to learn the complex relationship between displacement and its influencing factors. LSTM is used to extract this relationship to obtain Stage I prediction. The second stage takes the difference between the actual measurement and the Stage I prediction as inputs, and LSTM extracts the stochastic features of the monitoring system to obtain Stage II prediction. The sum of two stage predictions forms the final prediction. The DRLSTM model only requires raw monitoring data of water level and temperature to accurately predict displacement. Through a real-world comparative study against four commonly used shallow learning models and three deep learning models, the root mean square error and mean absolute error of our proposed method are the smallest, being 0.198 mm and 0.149 mm respectively, while the correlation coefficient is the largest at 0.962. It is concluded that the DRLSTM model performance well for evaluating dam health status.

Two-Step Approach to Processing Raw Strain Monitoring Data for Damage Detection of Structures under Operational Conditions.

Strain data of structural health monitoring is a prospective to be made full use of, because it reflects the stress peak and fatigue, especially sensitive to local stress redistribution, which is the probably damage in the vicinity of the sensor. For decoupling structural damage and masking effects caused by operational conditions to eliminate the adverse impacts on strain-based damage detection, small time-scale structural events, i.e., the short-term dynamic strain responses, are analyzed in this paper by employing unsupervised modeling. A two-step approach to successively processing the raw strain monitoring data in the sliding time window is presented, consisting of the wavelet-based initial feature extraction step and the decoupling step to draw damage indicators. The principal component analysis and a low-rank property-based subspace projection method are adopted as two alternative decoupling methodologies. The approach’s feasibility and robustness are substantiated by analyzing the strain monitoring data from a customized truss experiment to successfully remove the masking effects of operating loads and identify local damages even concerning accommodating situations of missing data and limited measuring points. This work also sheds light on the merit of a low-rank property to separate structural damages from masking effects by comparing the performances of the two optional decoupling methods of the distinct rationales.

Entropy-Driven Stochastic Federated Learning in Non-IID 6G Edge-RAN

Scalable and sustainable AI-driven analytics are necessary to enable large-scale and heterogeneous service deployment in sixth-generation (6G) ultra-dense networks. This implies that the exchange of raw monitoring data should be minimized across the network by bringing the analysis functions closer to the data collection points. While federated learning (FL) is an efficient tool to implement such a decentralized strategy, real networks are generally characterized by time- and space-varying traffic patterns and channel conditions, making thereby the data collected in different points non independent and identically distributed (non-IID), which is challenging for FL. To sidestep this issue, we first introduce a new a priori metric that we call dataset entropy, whose role is to capture the distribution, the quantity of information, the unbalanced structure and the “non-IIDness” of a dataset independently of the models. This a priori entropy is calculated using a multi-dimensional spectral clustering scheme over both the features and the supervised output spaces, and is suitable for classification as well as regression tasks. The FL aggregation operations support system (OSS) server then uses the reported dataset entropies to devise 1) an entropy-based federated averaging scheme, and 2) a stochastic participant selection policy to significantly stabilize the training, minimize the convergence time, and reduce the corresponding computation cost. Numerical results are provided to show the superiority of these novel approaches.

InfoGAN-MSF: a data augmentation approach for correlative bridge monitoring factors

Bridge health evaluation has been a challenging issue due to high assessment errors with imbalanced and insufficient monitoring data of correlative bridge monitoring factors. We propose a data augmentation model as the preprocessing of bridge health evaluation to expand existing monitoring data of bridge monitoring factors on the basis of generative adversarial nets (GAN), named as information-GAN (InfoGAN)-multi-scale-filtering. In this model, new data of bridge monitoring factors are produced by the InfoGAN-based model with regards to learning coupling relations among bridge monitoring factors. To resolve generalization issues from the parameter matrix in traditional InfoGAN, we improve the discriminator with spectral normalization to optimize the weight training process. To deal with the instability of InfoGAN performance, which creates defective samples, a multi-scale filtering scheme is designed to obtain effective samples from the InfoGAN-based model. The scheme picks credible samples from both quantitative and qualitative aspects with the multiple scale filtering procedure. Additionally, inherent properties of bridge monitoring factors (e.g. distributions) are discovered within the generation process. Finally, filtered data are mixed into raw monitoring data to train classifiers. Simulation results imply that the proposed model performs effectively in data generation of real-world bridge monitoring factors and improves the performance of bridge health evaluation.

Estimation of Extreme Cable Forces of Cable-Stayed Bridges Based on Monitoring Data and Random Vehicle Models

For long-span cable-stayed bridges, cables serve as one of the most important components to guarantee structural integrity. Forces of stay cables indicate not only the performance of cables themselves but also the overall condition of bridges. In order to help stakeholders to make maintenance decisions, an extreme cable force estimation method was proposed based on cable force measurements and traffic data from the weighing system. First, raw monitoring data were preprocessed based on a median filtering to obtain usable cable force signals. The multiresolution wavelet method was used to extract traffic-induced force component from mixed signals. Then, a Monte Carlo-based random vehicle model was developed using traffic data from the weighing system. Based on field temperature measurements and simulation of traffic-induced effects, extreme cable forces with respect to vehicle loads and temperature effects were predicted by extreme value theory. The Generalized Pareto Distribution (GPD) was adopted to establish the probability distribution models of the daily maximum cable force. Then, the extreme value within a return period of 100 years was determined and compared with the design loading demand. Finally, the effectiveness of the proposed method was validated through a cable-stayed bridge in China. As a result, the low-frequency varying component of cable force response had positive correlation with environmental temperatures, and the extreme value of the predicted cable force under prospective traffic volumes was within limit interval value according to the design code. The conclusions can be utilized by bridge owners to make maintenance decisions.