Industrial Sensor Research Articles

Abstract Diamond multi-wire sawing machines are essential in semiconductor manufacturing, especially for slicing hard and brittle third-generation materials such as silicon carbide (SiC) and gallium nitride (GaN). The increased difficulty in processing these materials has highlighted the urgent need for reliable machine health monitoring and anomaly detection systems. While Predictive Maintenance and Prognostics and Health Management (PHM) frameworks have been widely applied across various industries, little research has specifically addressed semiconductor cutting equipment, where operational dynamics and data confidentiality present unique challenges. This study, in collaboration with an industry partner, develops two anomaly detection models tailored for diamond multi-wire sawing machines. The first model is a rule-based approach that utilizes sliding window techniques to extract statistical features and establish dynamic thresholds for anomaly detection. The second model employs a data-driven Univariate Autoencoder (UAE) to perform unsupervised anomaly detection by learning reconstruction errors from normal operating data. Both models are trained and validated using confidential industrial sensor datasets. Experimental results demonstrate that the UAE-based model achieves high detection accuracy with no observed false positives, providing an effective solution for enhancing operational reliability and production efficiency in semiconductor wafer slicing processes.

The quality of input data is critical to the performance of time-series classification models, particularly in the domain for industrial sensor data where noise and anomalies are frequent. This study investigates how various filtering-based preprocessing techniques impact the accuracy and robustness of a Transformer model that predicts power efficiency states (Normal, Caution, Warning) from minute-level IIoT sensor data. We evaluated five techniques: a baseline, Simple Moving Average, Median filter, Hampel filter, and Kalman filter. For each technique, we conducted systematic experiments across time windows (360 and 720 min) that reflect real-world industrial inspection cycles, along with five prediction offsets (up to 2880 min). To ensure statistical robustness, we repeated each experiment 20 times with different random seeds. The results show that the Simple Moving Average filter, combined with a 360 min window and a short-term prediction offset, yielded the best overall performance and stability. While other techniques such as the Kalman and Median filters showed situational strengths, methods focused on outlier removal, like the Hampel filter, adversely affected performance. This study provides empirical evidence that a simple and efficient filtering strategy such as Simple Moving Average, can significantly and reliably enhance model performance for power efficiency prediction tasks.

The widespread integration of Internet-connected devices into industrial environments has enhanced connectivity and automation but has also increased the exposure of industrial cyber–physical systems to security threats. Detecting anomalies is essential for ensuring operational continuity and safeguarding critical assets, yet the dynamic, real-time nature of such data poses challenges for developing effective defenses. This paper introduces DataSense, a comprehensive dataset designed to advance security research in industrial networked environments. DataSense contains synchronized sensor and network stream data, capturing interactions among diverse industrial sensors, commonly used connected devices, and network equipment, enabling vulnerability studies across heterogeneous industrial setups. The dataset was generated through the controlled execution of 50 realistic attacks spanning seven major categories: reconnaissance, denial of service, distributed denial of service, web exploitation, man-in-the-middle, brute force, and malware. This process produced a balanced mix of benign and malicious traffic that reflects real-world conditions. To enhance its utility, we introduce an original feature selection approach that identifies features most relevant to improving detection rates while minimizing resource usage. Comprehensive experiments with a broad spectrum of machine learning and deep learning models validate the dataset’s applicability, making DataSense a valuable resource for developing robust systems for detecting anomalies and preventing intrusions in real time within industrial environments.

Enhancing non-tobacco related materials control and improving the purity of tobacco leaves have emerged as pivotal quality indicators for raw material processing in both domestic and foreign industrial enterprises. In order to accurately detect non-tobacco related materials, this paper introduces an enhanced variant of the YOLOv8(You Only Look Once version 8) model, termed NTRM-YOLO. NTRM-YOLO use deep learning methods to detect non-tobacco related materials. The attention mechanism module is integrated into the backbone network of NTRM-YOLO, aimed at enhancing the delineation of non-tobacco related materials features, thereby bolstering the detection efficacy of the model. In order to reduce the number of model parameters, this paper integrates GhostConv(Ghost Convolution) module within the neck network, alongside the design integration of a GhostConv-C2F module. This strategic substitution serves to diminish the model’s parameters while concurrently enhancing its capacity for feature expression. Within the Head network, capitalizing fully on the merits of multiple attention mechanisms, Dyhead(Dynamic Head) is introduced with the aim of markedly enhancing the detection accuracy of the model. This study also optimized the loss function by using the vector angle. Moreover, this paper uses industrial camera sensors to collect images containing non-tobacco related materials and constructed of an NTRM dataset after preprocessing. Subsequently, a meticulously series of experiments was conducted on the NTRM dataset to showcase the efficacy of NTRM-YOLO model in applications pertaining to non-tobacco related materials detection. The experimental findings reveal that in contrast to the baseline model, NTRM-YOLO attained a detection performance of 95.6%, marking a notable improvement of 2% over the baseline model. Additionally, it exhibited a parameters of 10.0 MB, reflecting a 10% reduction compared to the baseline model. These experiments furnishes a theoretical foundation and technical substantiation for the subsequent advancement of more refined industrial impurity removal instruments and equipment.

High-precision automated detection of metal welding defects is critical to ensuring structural safety and reliability in modern manufacturing. However, existing methods often struggle with insufficient fine-grained feature retention, low efficiency in multi-scale information fusion, and vulnerability to complex background interference, resulting in low detection accuracy. This work addresses the limitations by introducing the DEIM-SFA, a novel detection framework designed for automated visual inspection in industrial machine vision sensors. The model introduces a novel structure-aware dynamic convolution (SPD-Conv), effectively focusing on the fine-grained structure of defects while suppressing background noise interference; an innovative multi-scale dynamic fusion pyramid (FTPN) architecture is designed to achieve efficient and adaptive aggregation of feature information from different receptive fields, ensuring consistent detection of multi-scale targets; combined with a lightweight and efficient multi-scale attention module (EMA), this further enhances the model’s ability to locate salient regions in complex scenarios. The network is evaluated on a self-built welding defect dataset. Experimental results show that DEIM-SFA achieves a 3.9% improvement in mAP50 compared to the baseline model, mAP75 by 4.3%, mAP50–95 by 3.7%, and Recall by 1.4%. The model exhibits consistently significant superiority in detection accuracy across targets of various sizes, while maintaining well-balanced model complexity and inference efficiency, comprehensively surpassing existing state-of-the-art (SOTA) methods.

In recent years, predictive maintenance has emerged as a critical component for improving the efficiency and reliability of industrial systems. However, much of the existing research has primarily emphasized model development, often overlooking the fundamental role of data quality and class distribution in shaping predictive performance. To address this gap, this study proposes an integrated preprocessing framework that ensures high-quality data readiness across all stages. A case study was conducted on a dataset of industrial sensors for fault prediction. The preprocessing pipeline involved handling missing values using K-Nearest Neighbors (KNN), detecting outliers with Isolation Forest (IF), and correcting abnormal values through the Clipping method. To address data imbalance, synthetic data were generated using Generative Adversarial Networks (GAN), Variational Autoencoders (VAE), and a hybrid GAN-VAE model that leverages the strengths of both approaches. The hybrid GAN-VAE demonstrated superior data generation performance, yielding the highest Pearson correlation and best Kernel Density Estimation (KDE) fit, thereby ensuring dataset reliability for training. The effectiveness of the preprocessing framework was validated using a 1-Dimensional Convolutional Neural Network (1D-CNN) classifier, which achieved a high accuracy of 98.83%.

49. The development and application of on-animal sensors in the extensive grazing industries with reference to disease detection and the implications for production, welfare and biosecurity

Summary Predictive maintenance such as remaining useful life (RUL), a cornerstone of modern industrial practices, faces the challenge of maintaining high prediction accuracy as sensors approach the end of their operational life. The nonlinear and complex nature of sensor degradation becomes more pronounced over time, leading to a decline in model performance just when precise predictions are most critical. In response to this challenge, synthetic data have proven to be a valuable tool in training machine learning models for predicting the RUL of downhole drilling sensors. By simulating realistic run-to-failure scenarios, synthetic data enable the modeling of diverse wear patterns and degradation processes, helping improve the accuracy of predictions, especially in the later stages of sensor life. This approach is essential in supporting Industry 4.0-driven maintenance strategies that rely on accurate and timely insights to optimize equipment reliability and minimize unplanned downtime. Leveraging a long short-term memory (LSTM) model has proved to effectively address this issue, demonstrating the lowest prediction error with a normalized root mean square error (nRMSE) of 7.60% and a mean absolute percentage error (MAPE) of 14.80%, outperforming extreme gradient boosting (XGBoost), which achieves an nRMSE of 8.28% and a MAPE of 17.30%. The LSTM’s ability to manage nonlinear degradation processes over time makes it particularly well-suited for predicting RUL in complex sensor systems. Furthermore, the inclusion of temperature-based features has been shown to significantly enhance predictive accuracy, particularly in modeling the intricate patterns that emerge as sensor health deteriorates. This approach aligns with Industry 4.0’s focus on smart and connected systems, where predictive maintenance ensures that machinery runs efficiently while minimizing downtime. The use of synthetic run-to-failure data opens the door to modeling various wear and failure scenarios, making it possible to generate more accurate predictions for different industrial tools and sensors.

Detailed fault detection of industrial sensor based on semantic segmentation models

Quantum-enhanced LSTM for predictive maintenance in industrial IoT systems

Industry 4.0 has transformed modern manufacturing by incorporating internet-connected devices, artificial intelligence, cyber-physical platforms, and advanced data analysis. These innovations have raised automation levels, improved operational performance, and enabled predictive maintenance. Industry 5.0 builds on these foundations by emphasizing collaboration between humans and intelligent machines, promoting environmentally responsible practices, and supporting personalized production models. In parallel, affordable sensor technologies are gaining traction due to their low cost and practical ability to support essential monitoring across diverse applications. This study offers a comparative evaluation of sensor technologies in Industry 4.0, Industry 5.0, and budget-conscious systems, outlining their advantages, limitations, and areas of application. The findings suggest that sensor integration remains central to industrial progress, particularly when linked with IoT, artificial intelligence, and edge-computing solutions that enhance adaptability and real-time decision-making.

Empowering small-sample industrial soft sensor modeling through cross-silo knowledge fusion and incremental learning approaches

Abstract Machine Learning (ML) and Deep Learning (DL) have been used for anomaly detection in Industrial Internet of Things (IIoT) environments. The presence of imbalanced data, high noise levels, missing values, and high dimensionality poses an enormous challenge for existing methods, leading to inconsistent reliability in detecting anomalies in real-world industrial environments. Current anomaly detection solutions suffer from high false-negative rates due to class imbalance and noisy sensor data, limiting their practical applicability. This paper proposes the Ensemble Wasserstein Generative Adversarial Network for IIoT (EWAD-IIoT) framework, which is uniquely designed to address these challenges. The aim is to build a robust anomaly detection model with high recall (94.7%) and precision (93.6%) while minimizing miss rates in complex IIoT settings. Evaluations on two benchmark datasets, SECOM (industrial sensor data) and MNIST (image data), demonstrate EWAD-IIoT’s superiority over traditional methods like standalone WGAN and WGAN-GP. To highlight its efficacy, we compare results against these benchmarks, showcasing improvements in F1-score (95.8%) and noise robustness. The framework leverages advanced preprocessing (Z-score filtering, Min-Max scaling), SMOTE-based balancing, and WGAN-generated synthetic samples to handle data imbalance and dimensionality. The results validate EWAD-IIoT’s capability to detect rare anomalies in IIoT environments, with a balanced trade-off between recall and precision, making it a scalable solution for predictive maintenance and fault diagnosis.

Variational masking progressive learning method for multi-rate industrial processes soft sensor

A missing data imputation method for industrial soft sensor modeling

AutoFactory Dataset to Support AI in Manufacturing Systems

In the context of intelligent coal mine safety monitoring, an integrated prediction and early-warning method named MTGNN-Bayesian-IF-DBSCAN (Multi-Task Graph Neural Network-Bayesian Optimization-Isolation Forest-Density-Based Spatial Clustering of Applications with Noise) is proposed to address the challenges of gas concentration prediction and anomaly detection in coal mining faces. The MTGNN (Multi-Task Graph Neural Network) is first employed to model the spatiotemporal coupling characteristics of gas concentration and wind speed data. By constructing a graph structure based on sensor spatial dependencies and utilizing temporal convolutional layers to capture long short-term time-series features, the high-precision dynamic prediction of gas concentrations is achieved via the MTGNN. Experimental results indicate that the MTGNN outperforms comparative algorithms, such as CrossGNN and FourierGNN, in prediction accuracy, with the mean absolute error (MAE) being as low as 0.00237 and the root mean square error (RMSE) maintained below 0.0203 across different sensor locations (T0, T1, T2). For anomaly detection, a Bayesian optimization framework is introduced to adaptively optimize the fusion weights of IF (Isolation Forest) and DBSCAN (Density-Based Spatial Clustering of Applications with Noise). Through defining the objective function as the F1 score and employing Gaussian process surrogate models, the optimal weight combination (w_if = 0.43, w_dbscan = 0.52) is determined, achieving an F1 score of 1.0. By integrating original concentration data and residual features, gas anomalies are effectively identified by the proposed method, with the detection rate reaching a range of 93-96% and the false alarm rate controlled below 5%. Multidimensional analysis diagrams (e.g., residual distribution, 45° diagonal error plot, and boxplots) further validate the model's robustness in different spatial locations, particularly in capturing abrupt changes and low-concentration anomalies. This study provides a new technical pathway for intelligent gas warning in coal mines, integrating spatiotemporal modeling, multi-algorithm fusion, and statistical optimization. The proposed framework not only enhances the accuracy and reliability of gas prediction and anomaly detection but also demonstrates potential for generalization to other industrial sensor networks.

Purpose This study explores the interaction between operations management and information systems by applying the Design Science Research (DSR) methodology for intelligent early fault management. Prior research primarily addressed post-fault identification and classification but has struggled with catastrophic forgetting. Thus, this work proposes an innovative data-driven artifact that leverages a deep learning (DL)-based approach for early fault detection and future fault forecasting. Design/methodology/approach Following the DSR methodology, the work proposes an innovative data-driven artifact for early fault management. The proposed artifact extracts key features from industrial sensor data in real time using a Deep Sparse Autoencoder with a sparsity penalty. These features are then processed using an Exponentially Weighted Moving Average method for monitoring process variations, while a Transformer-based Neural Network forecasts potential faults. To mitigate catastrophic forgetting, the Elastic Weight Consolidation technique is applied during offline training to preserve previous patterns when new information becomes available. Findings The artifact enhances operational decision-making by generating early warning alerts and delivering actionable insights. Experimental evaluation using real-world sensor data validates that the proposed approach outperforms existing DL methods. Originality/value Unlike traditional approaches that are limited to fixed fault distributions, this work introduces novel design propositions for industrial fault management systems, enabling dynamic learning and continuous improvement with new data.

With the increasing scale of industrial devices and the growing complexity of multi-source heterogeneous sensor data, traditional methods struggle to address challenges in fault detection, data security, and trustworthiness. Ensuring tamper-proof data storage and improving prediction accuracy for imbalanced anomaly detection for potential deployment in the Industrial Internet of Things (IIoT) remain critical issues. This study proposes a blockchain-powered Long Short-Term Memory Network (LSTM)-Attention hybrid model: an LSTM-based Encoder-Attention-Decoder (LEAD) for industrial device anomaly detection. The model utilizes an encoder-attention-decoder architecture for processing multivariate time series data generated by industrial sensors and smart contracts for automated on-chain data verification and tampering alerts. Experiments on real-world datasets demonstrate that the LEAD achieves an F0.1 score of 0.96, outperforming baseline models (Recurrent Neural Network (RNN): 0.90; LSTM: 0.94; and Bi-directional LSTM (Bi-LSTM, 0.94)). We simulate the system using a private FISCO-BCOS network with a multi-node setup to demonstrate contract execution, anomaly data upload, and tamper alert triggering. The blockchain system successfully detects unauthorized access and data tampering, offering a scalable solution for device monitoring.

Clustered Temporal Patch Transformer Network via Deep Local Feature Learning for Industrial Soft Sensor Applications