Sensor Signal Research Articles

Sensor Fault Detection and Classification Using Multi-Step-Ahead Prediction with an Long Short-Term Memoery (LSTM) Autoencoder

The Internet of Things (IoT) is witnessing a surge in sensor-equipped devices. The data generated by these IoT devices serve as a critical foundation for informed decision-making, real-time insights, and innovative solutions across various applications in everyday life. However, data reliability is often compromised due to the vulnerability of sensors to faults arising from harsh operational conditions that can adversely affect the subsequent operations that depend on the collected data. Hence, the identification of anomalies within sensor-derived data holds significant importance in the IoT context. This article proposes a sensor fault detection method using a Long Short-Term Memory autoencoder (LSTM-AE). The AE, trained on normal sensor data, predicts a 20-step window, generating three statistical features via SHapley Additive exPlanations from the estimated steps. These features aid in determining potential faults in the predicted steps using a machine learning classifier. A secondary classifier identifies the type of fault in the sensor signal. Experimentation on two sensor datasets showcases the method’s functionality, achieving fault detection accuracies of approximately 93% and 97%. It is possible to attain a perfect fault classification performance by slightly modifying the feature calculation approach. In a univariate prediction scenario, our proposed approach demonstrates good fault detection and classification performance.

Performance enhancement of SAW based hydrogen sensor employing Pd/Ni alloy thin film

Abstract This study focuses on determining the optimized design parameters of palladium/nickel (Pd/Ni) thin film coated surface acoustic wave (SAW) hydrogen (H2) sensors with high sensitivity and fast response by modulating the ratio and thickness of Pd/Ni thin film. The Pd/Ni thin films deposited on the wave propagation path of delay-line patterned SAW sensors had Ni content ranging from 0 to 30 at% and thickness ranging from 20 to 400 nm. A phase discrimination circuit was used to collect the signal of SAW H2 sensor. The experimental results indicated that the response sensitivity of the sensor decreased with the increase of Ni content, while the response time gradually accelerated. Besides, the stability of the sensor deteriorated with increasing thickness, and there existed a non-linear relationship between sensitivity and thickness, and the response time slowed down as the thickness increased. To analyze the SAW sensing mechanism, a theoretical simulation model of the Pd/Ni thin film coated SAW H2 sensor was established using perturbation theory, which was well-validated by experimental results. Considering the requirements of fast response, high sensitivity, and stability of the H2 sensor, the optimal thin film structure parameters were determined as Pd9Ni1 thin film with the thickness of 40 nm.

An MXene/cellulose nanofiber-based multi-parametric tactile sensor for object material identification

Abstract We propose a novel multi-parametric tactile sensor that can realize force and slip sensing during the contact between a manipulator and a target. The pressure-sensitive unit of this sensor is prepared by homemade MXene and cellulose nanofiber (CNF) composite conductive film, which has high sensitivity of pressure-sensitive response. The triboelectric unit is made of a chemically stable PTFE film with high electron-acquiring ability as the triboelectric sensing layer together with metal electrodes. Experiments have demonstrated that the prepared sensors have high sensitivity, fast response, and excellent durability and stability over 5000 cycles. Combined with the established machine learning algorithm model, the pressure and triboelectric signals of this multi-parametric tactile sensor were used to recognize 15 types of materials with similar characteristics and with 94.33% identification accuracy. This work shows the great potential of multi-parametric tactile sensors for applications in item sorting, human-machine interaction, and other fields.

Signal recognition based on transfer learning for Φ-OTDR fiber optic distributed disturbance sensor

Abstract Traditional classification models for optical fiber vibration signals can achieve good recognition results when there are a large number of samples. However, it is difficult to obtain sufficient training samples in real scenes. A method of Φ-OTDR sensor signal recognition based on VGGish transfer learning is proposed in this paper. Firstly, a selection indicator is designed to filter out vibration signals from sparse signals. Then, the VGGish network pretrained for sound classification is used for transfer learning as a feature extractor for optical fiber vibration signals. This method was applied to classify 6 kinds of optical fiber vibration signals. When the number of training samples is reduced to 480 samples, the method can still achieve 84.17% classification accuracy. Compared to the method of training from scratch, this proposed method reduces training time by 73.9%.

Motor fault diagnosis based on multisensor-driven visual information fusion

To need of accurate motor fault diagnosis in industrial system, we propose a fault diagnosis framework that utilizes motor current and electromagnetic signals, combining them with a self-attention-enhanced capsule network for enhanced signal analysis and accuracy. Firstly, the original signal extracted by multiple sensors is constructed into a symmetric point mode (SDP) image, and the visual fault information of different sensors and fusion signals of different motion health states are obtained by the proposed multi-channel image fusion method. Then, the capsule network, combined with self-attention, extracts spatial features from the high-dimensional tensor of the multi-channel fused image for adaptive recognition and extraction. Subsequently, advanced feature vector information is obtained through softmax for diagnosis. Diagnosis results of several datasets indicate that the developed diagnosis framework with compressed image information can availably identify 8 kinds of motor fault states under various loads, and the fault diagnosis rate is as high as 99.95 %, it is helpful for low cost and high-speed diagnosis of motors. In addition, by learning multiple sensor signals in the same state, it obtains stronger robustness and effectiveness than a single signal model.

Physical constraint flow field reconstruction method for immersion scan-step flow

The lithography immersion flow field is highly sensitive to minor variations due to the precision demands of lithography. The scanning flow induces pressure fluctuations on the lens surface, subsequently impacting the quality of exposure. Previous measurement methods have encountered sparse distribution measurements on the lens surface due to limitations in sensor size and flow field dimensions. In this manuscript, the authors propose a modal fusion flow field reconstruction method, effectively integrating sensor-based experimental methods and equation-based computational methods. This method achieves high-resolution measurements of stress distribution and velocity distribution. Initially, a simulation of the flow field under operational conditions is conducted to acquire a dataset, followed by decomposing the dataset into modes to establish a modal library. Throughout the measurement process, the modal library is dynamically restructured based on sensor signal using a sparse representation method to forecast the flow field, and subsequently, the flow field is corrected in accordance with the Navier-Stokes equation.

Detection of Hot Cracking for 6000 Series Aluminum Alloys using a Multi-Sensor Based Deep Learning Model

This study developed a monitoring technology using a multi-sensor based deep learning model to diagnose hot cracking in aluminum alloy laser welding. Hot cracking that occurs during the laser welding process of aluminum alloys is difficult to diagnose accurately with a single sensor signal, necessitating multi-sensor based process monitoring technology. To monitor these hot cracks, laser-induced plasma, acoustic, and elastic wave signals were simultaneously measured using a spectrometer, non-contact acoustic sensor, and contact acoustic sensor during the overlap laser welding process of 6000 series aluminum alloys. The welded specimens were classified into normal and cracked specimens through bead analysis, and features related to hot cracking were extracted from each sensor signal to utilize the measured multi-sensor signals for monitoring. The extracted features from each signal were used as inputs for a Deep Neural Network (DNN) model capable of learning complex nonlinear relationships, and the hyperparameters of the DNN model were optimized using a genetic algorithm. The DNN model trained with multi-sensor data diagnosed hot cracking with an accuracy of 93.75%.

MitoBADY-based Raman sensing of neutrophil-like cells

The aim was to evaluate the potential of MitoBADY as a Raman sensor in monitoring the drug-induced differentiation of promyelocytic cells into neutrophils. Neutrophils are important immune system components, and their metabolic disorders lead to serious diseases. Primary neutrophils found in the bloodstream are short-lived cells, terminally differentiated, and unable to proliferate. Thus, the characteristics of induced differentiation of promyelocytic cells are valuable for better chemotherapy design, as it is difficult to conduct such studies in vivo or ex vivo in bone marrow environments. Alternatively, a stable in vitro model can be used to analyze the differentiation process, e.g., the HL-60 promyelocytic leukemia cell line. The standard method to confirm the promyelocytic differentiation is immunophenotyping, which is complex and laborious. Here, we propose a simple spectroscopic-based alternative that uses the Raman signal of the MitoBADY sensor to identify neutrophil-like cells. We found that the ratio of the intensity of three bands, i.e., A= (I750+I2220)/I750, B=(I2850+I2220)/I2850, and A/B, of the Raman spectra of cells incubated with the MitoBADY is a specific marker indicating the induction of neutrophil differentiation of HL-60 cells. Flow cytometry, measuring the level of the expression of the CD11b surface protein, was used as a reference method. Raman measurements allowed the classification of neutrophil-like cells with a sensitivity of 79.8 % and a specificity of 79.5 %. These values demonstrate the great potential of the proposed methodology as a rapid and reliable technique for the evaluation of the differentiation of promyelocytic cells into neutrophils.

MRCFN: A multi-sensor residual convolutional fusion network for intelligent fault diagnosis of bearings in noisy and small sample scenarios

Bearing fault diagnosis is of great importance to ensure the safe and stable operation of mechanical equipment. The actual collected bearing fault signals are susceptible to strong noise interference and bearing samples for each fault state may be insufficient, which increases the difficulty of capturing effective features. Most of the existing diagnostic methods extract features from a single sensor signal for pattern recognition and fault diagnosis. The fault information provided by a single sensor is limited and incomplete, which is usually very difficult to meet the demand for accurate and reliable fault diagnosis in complex scenarios. To solve these problems, this paper proposes a multi-sensor residual convolutional fusion network (MRCFN) for intelligent fault diagnosis of bearings. Firstly, a convolutional pooling module (CPM) is coupled with the designed double ring residual module (DRRM) to rough feature extraction and deep feature mining, which not only captures the discriminative fault features from multi-sensor signal, but also avoids the performance degradation of network. Secondly, a spatial channel reconstruction module (SCRM) is further introduced to eliminate redundant information in the features and improve the network training efficiency. Finally, the presented global interactive perception fusion module (GIPFM) is connected with a classification block (CB) to globally fuse the features extracted from the acoustic and vibration signals and conduct automatic fault identification, which both can realize the complementarity and calibration of multi-sensor feature information and high precision diagnosis. The experiments and a series of comparisons are implemented on two datasets to efficaciously verify the superiority of the proposed method over five existing representative multi-sensor fusion diagnosis methods (i.e., FAC-CNN, MB-CNN, MsACNN, MRSDF, and IMSFDFL) under strong noise and small samples.

Portable Sensors for Dynamic Exposure Assessments in Urban Environments: State of the Science.

This study presents a fit-for-purpose lab and field evaluation of commercially available portable sensor systems for PM, NO2, and/or BC. The main aim of the study is to identify portable sensor systems that are capable of reliably quantifying dynamic exposure gradients in urban environments. After an initial literature and market study resulting in 39 sensor systems, 10 sensor systems were ultimately purchased and benchmarked under laboratory and real-word conditions. We evaluated the comparability to reference analyzers, sensor precision, and sensitivity towards environmental confounders (temperature, humidity, and O3). Moreover, we evaluated if the sensor accuracy can be improved by applying a lab or field calibration. Because the targeted application of the sensor systems under evaluation is mobile monitoring, we conducted a mobile field test in an urban environment to evaluate the GPS accuracy and potential impacts from vibrations on the resulting sensor signals. Results of the considered sensor systems indicate that out-of-the-box performance is relatively good for PM (R2 = 0.68-0.9, Uexp = 16-66%, BSU = 0.1-0.7 µg/m3) and BC (R2 = 0.82-0.83), but maturity of the tested NO2 sensors is still low (R2 = 0.38-0.55, Uexp = 111-614%) and additional efforts are needed in terms of signal noise and calibration, as proven by the performance after multilinear calibration (R2 = 0.75-0.83, Uexp = 37-44%)). The horizontal accuracy of the built-in GPS was generally good, achieving <10 m accuracy for all sensor systems. More accurate and dynamic exposure assessments in contemporary urban environments are crucial to study real-world exposure of individuals and the resulting impacts on potential health endpoints. A greater availability of mobile monitoring systems capable of quantifying urban pollutant gradients will further boost this line of research.

Improving the technological process of balancing electric machine rotors on a balancing machine

The object of research is the process of dynamic balancing of rotors on a balancing machine in the process of restoring electric machines. The work is aimed at improving the quality of balancing rotors in the process of major repairs of traction electric motors for electric trains. The problem addressed was the quality of balancing the rotors of electric machines on stationary balancing machines. According to the conventional balancing technology, the rotor to be balanced is mounted on the supports of the balancing machine with support surfaces that usually have mechanical defects. These defects cannot be eliminated by machining due to the peculiarities of rotor repair technology. Theoretical and experimental studies of the effect of damage to the rotor support surfaces on the balancing parameters have been carried out. It has been proven that the properties of the bearing surfaces of the rotor during its balancing on a balancing machine significantly affect the results of determining the imbalance. At the same time, the difference in the mass values of balancing loads can reach 25 %. This is because damage to the bearing surfaces of the rotor generates false signals unrelated to the imbalance. In order to increase the accuracy of determining the mass of balancing loads during rotor balancing, it is proposed to improve the technological process of balancing. The improvement involves the inclusion of a frequency filter in the signal conversion chain of acceleration sensors. The filter is designed to separate signals with a frequency greater than the rotational frequency of the rotor. A condition for the practical application of the research results is the expediency of introducing a frequency filter of signals of acceleration sensors with a threshold frequency of filtering signals that exceeds the rotational frequency of the rotor into the schematic diagram of the balancing machine.

Current Advances of CO Sensing Based on Low Dimensional Materials.

Carbon monoxide (CO) is a harmful gas with significant impacts on human health and the environment. Its timely detection, especially in the event of thermal runaway in automotive lithium batteries, is crucial to prevent casualties. This paper reviews the progress in the development of efficient, sensitive, and reliable CO sensors, focusing on electrochemical, optical, and resistive sensing materials. Low-dimensional materials have a large specific surface area, providing an abundant number of active sites, which has drawn extensive attention from researchers. According to the different sensor signals, we categorized these sensors into electrical and optical signal sensors. We hope that by systematically introducing the sensing mechanism and sensing performance of these two kinds of sensors, appropriate CO sensors can be developed in different application scenarios so as to realize early warning and monitoring to the maximum extent, reduce industrial losses, and ensure the life and health of personnel.

Predictive Maintenance Based on Identity Resolution and Transformers in IIoT

The burgeoning development of next-generation technologies, especially the Industrial Internet of Things (IIoT), has heightened interest in predictive maintenance (PdM). Accurate failure forecasting and prompt responses to downtime are essential for improving the industrial efficiency. Traditional PdM methods often suffer from high false alarm rates and inefficiencies in complex environments. This paper introduces a predictive maintenance framework using identity resolution and a transformer model. Devices receive unique IDs via distributed identifiers (DIDs), followed by a state awareness model to assess device health from sensor signals. A sequence prediction model forecasts future signal sequences, which are then used with the state awareness model to determine future health statuses. Combining these predictions with unique IDs allows for the rapid identification of facilities needing maintenance. Experimental results show superior performance, with 99% accuracy for the state awareness model and a mean absolute error (MAE) of 0.062 for the sequence prediction model, underscoring the effectiveness of the framework.

Balance confidence classification in people with a lower limb amputation using six minute walk test smartphone sensor signals.

The activities-specific balance confidence scale (ABC) assesses balance confidence during common activities. While low balance confidence can result in activity avoidance, excess confidence can increase fall risk. People with lower limb amputations can present with inconsistent gait, adversely affecting their balance confidence. Previous research demonstrated that clinical outcomes in this population (e.g., stride parameters, fall risk) can be determined from smartphone signals collected during walk tests, but this has not been evaluated for balance confidence. Fifty-eight (58) individuals with lower limb amputation completed a six-minute walk test (6MWT) while a smartphone at the posterior pelvis was used for signal collection. Participant ABC scores were categorized as low confidence or high confidence. A random forest classified ABC groups using features from each step, calculated from smartphone signals. The random forest correctly classified the confidence level of 47 of 58 participants (accuracy 81.0%, sensitivity 63.2%, specificity 89.7%). This research demonstrated that smartphone signal data can classify people with lower limb amputations into balance confidence groups after completing a 6MWT. Integration of this model into the TOHRC Walk Test app would provide balance confidence classification, in addition to previously demonstrated clinical outcomes, after completing a single assessment and could inform individualized rehabilitation programs to improve confidence and prevent activity avoidance.

Novel GIL mechanical fault diagnosis method based on multi-sensor data feature fusion and TDEAVOA-ELM

The gas insulated transmission line (GIL), a crucial component of the power system, faces challenges in mechanical fault diagnosis including difficulties in quantifying sensor signal selection, balancing recognition accuracy, and model training speed. This paper proposes a GIL mechanical fault diagnosis method based on the feature fusion of multi-channel vibration sensor signals and an improved optimization extreme learning machine (ELM). Initially, three optimal signals are selected based on the correlation coefficients of multiple signals, and their time–frequency domain features are extracted. These features are then fused using neighborhood preserving embedding (NPE) and input into the improved optimized ELM model. The enhanced optimization algorithm reduces the likelihood of falling into local optima, thereby improving the model’s recognition performance. Experimental results demonstrate that this method effectively enhances the recognition accuracy of mechanical faults in GIL equipment and enables rapid diagnosis.

Phase-Engineered MoSe2/CeO2 Composites for Room-Temperature Gas Sensing with a Drastic Discrimination of NH3 and TEA Gases.

Detecting and distinguishing between hazardous gases with similar odors by using conventional sensor technology for safeguarding human health and ensuring food safety are significant challenges. Bulky, costly, and power-hungry devices, such as that used for gas chromatography-mass spectrometry (GC-MS), are widely employed for gas sensing. Using a single chemiresistive semiconductor or electric nose (e-nose) gas sensor to achieve this objective is difficult, mainly because of its selectivity issue. Thus, there is a need to develop new materials with tunable and versatile sensing characteristics. Phase engineering of two-dimensional materials to better utilize their physiochemical properties has attracted considerable attention. Here, we show that MoSe2 phase-transition/CeO2 composites can be effectively used to distinguish ammonia (NH3) and triethylamine (TEA) at room temperature. The phase transition of nanocomposite samples from semimetallic (1T) to semiconducting (2H) prepared at different synthesis temperatures is confirmed via X-ray photoelectron spectroscopy (XPS). A composite sensor in which the 2H phase of MoSe2 is predominant lacks discrimination capability and is less responsive to NH3 and TEA. An MoSe2/CeO2 composite sensor with a higher 1T phase content exhibits high selectivity for NH3, whereas one with a higher 2H phase content (2H > 1T) shows more selective behavior toward TEA. For example, for 50% relative humidity, the MoSe2/CeO2 sensor's signal changes from the baseline by 45% and 58% for 1 ppm of NH3 and TEA, respectively, indicating a low limit of detection (LOD) of 70 and 160 ppb, respectively. The composites' superior sensing characteristics are mainly attributed to their large specific surface area, their numerous active sites, presence of defects, and the n-n type heterojunction between MoSe2 and CeO2. The sensing mechanism is elucidated using Raman spectroscopy, XPS, and GC-MS results. Their phase-transition characteristics render MoSe2/CeO2 sensors promising for use in distributed, low-cost, and room-temperature sensor networks, and they offer new opportunities for the development of integrated advanced smart sensing technologies for environmental and healthcare.

Semi-supervised learning for real-time anomaly detection in pulsed transfer wire arc additive manufacturing

In Wire-Arc Additive Manufacturing (WAAM), ensuring the quality and integrity of components is of key importance and performing anomaly detection during production is an efficient strategy for preventing defects and maintaining an acceptable quality with lower costs compared to the development of complex supervised learning techniques. This paper presents an approach based on semi-supervised learning for real-time anomaly detection in the production of Inconel 718 components via the Pulsed Transfer WAAM process. The workflow is based on the use of training data obtained by employing established process parameters, similar to what is done in Welding Procedure Qualification Records (WPQRs), to develop a semi-supervised anomaly detection application. The proposed approach involves depositing material using the already established process parameters and collecting the corresponding welding data in terms of welding current and voltage sensor signals. The collected data are then employed to train an unsupervised algorithm by using solely good deposition data to learn hidden complex patterns of normality and hence detect anomalies based on deviations from these patterns. With this aim, global and local features are extracted from the welding current and voltage sensor signals in both time and time-frequency domains via Wavelet Transform technique and a deep learning approach based on a residual convolutional autoencoder. To showcase the effectiveness of the proposed approach, a case study involving wire arc additive manufacturing of Inconel 718 components was conducted. The proposed algorithm was tested and achieved an F-score of 0.895, representing a 30 % improvement over state-of-the-art methods that rely on time domain features for anomaly detection. This research work contributes to the improvement of anomaly detection methodologies, offering substantial advantages over alternative techniques for quality control and reliability of additive manufacturing processes such as WAAM.

Managing remaining useful life of cyber-aeroengine systems using a graph spatio-temporal attention recurrent network with phase-lag index

The decision-making process concerning the maintenance of equipment significantly impacts the safety and reliability of the aeroengine system. A cyber-aeroengine system (CAS) integrates intricate aeroengine components with a cyber decision-making space based on the communication network, thereby facilitating computational support for forecasting the health status of the aeroengine system. Within the cyber realm, we introduce a novel approach termed the graph spatio-temporal attention recurrent network with phase-lag index (PLI-GSTARN) to assess the degradation states of aeroengine systems. This method constructs a spatio-temporal graph based on sensor signal correlations and information reachability across distributed sensors within the physical equipment. Subsequently, the proposed framework employs a graph attention network to capture spatial features and a long short-term memory (LSTM) layer equipped with an attention mechanism to capture temporal features inherent in the constructed graph. Ultimately, this methodology is applied to cyber–physical systems within aircraft engines to manage remaining health life and diagnose faults, yielding commendable outcomes.

Fault Diagnosis of the Electric Multiple Unit Door System by Machine Learning Using Sensor Signal of the Simulator

Fault Diagnosis of the Electric Multiple Unit Door System by Machine Learning Using Sensor Signal of the Simulator

Integrated mental stress smartwatch based on sweat cortisol and HRV sensors

Mental stress, a human's common emotion that is difficult to recognize and describe, can give rise to serious psychological disorders. Skin and sweat are easily accessible sources of biomarkers and bio-signals that contain information about mental stress. It is challenging for current wearable devices to monitor psychological stress in real-time with a non-invasive manner. Therefore, we have developed a smartwatch integrated with a sweat cortisol sensor and a heart rate variation (HRV) sensor. This smartwatch can simultaneously record the cortisol levels in sweat and HRV index in real time over a long period. The cortisol sensors based on organic electrochemical transistor (OECT) are fabricated by utilizing the Prussian-blue (PB) doped molecular imprinting polymer (MIP) modified gate electrode. The sensor signal current will decrease following the combination of sweat cortisol, due to the blocking of the PBMIP conductive path, demonstrating good sensitivity, selectivity, and stability. The HRV sensor is manufactured by a photoplethysmography method. We have integrated the two sensors into a wearable smartwatch that can match well with the mobile phone APP and the upper computer software. Through the use of this smartwatch, we have observed a negative correlation between cortisol levels in sweat and the HRV index in short-term stressful environments. Our research presents a great progress in real-time and non-invasive monitoring human's stress levels, which promotes not only the stress management, but also better psychological research.