Sensor Fault Detection Research Articles

Diagnosis of single and multiple-source faults of chiller sensors using EWEEMD-ICKNN by time sequence denoising and non-Gaussian distribution feature extraction

Currently, fault detection of chiller sensors primarily focuses on single-source fault (SSF) sensors. These methods are typically ineffective in multiple source faults (MSF) cases. These methods assume a Gaussian distribution of the chiller data and ignore the coupling between the sensors. Therefore, we propose temporal feature denoising and non-Gaussian distribution-based methods in this study. First, the chiller sensor signals are denoised based on exponential weighted moving average ensemble empirical mode decomposition (EWEEMD). Subsequently, an independent component analysis k-nearest neighbour (ICKNN) classifier is proposed based on the non-Gaussian distribution features of the chiller sensor data. A residual vector combined with independent component analysis was used for feature extraction to propose the ICKNN classifier. Finally, according to their characteristics, different SSF and MSF were applied to the chiller sensors. Experimental comparisons with other methods were performed for four single-source faults and six multisource faults. The EWEEMD-ICKNN was experimentally confirmed to have excellent fault detection performance in the SSF and MSF chiller sensors. The experimental results demonstrate that the fault diagnosis rate of EWEEMD-ICKNN ranges from 99.90% to 86.80%. Additionally, the fault diagnosis efficiency of EWEEMD-ICKNN remains consistently stable across various fault states, exhibiting an average F1 score of 0.9589 with F1 score statistics of 0.038.

Kalman filter based sensor fault detection in wireless sensor network for smart irrigation

Ensuring the reliability of data obtained from Wireless Sensor Networks (WSNs) deployed in farmland to monitor soil parameters is crucial for optimizing smart irrigation. These distributed sensor nodes encounter various challenges and are vulnerable to sensor faults that can significantly degrade the network's service quality. In this article, we present an innovative approach for detecting sensor faults within WSNs for smart irrigation by combining an autoregressive model with a Kalman filter. Integrating the Kalman filter and the autoregressive model combines their strengths in a synergistic manner. The algorithm is developed with consideration for the resource constraints of the sensor nodes and addresses the challenge of lacking ground truth information for the monitored area. The primary advantage of this proposed technique lies in its simplicity of implementation, requiring minimal computational complexity while enhancing the application's reliability. Through experimentation and validation, we demonstrate the effectiveness of this combined approach in detecting sensor fault detection in real-world WSNs scenarios.

The impact of improved PCA method based on anomaly detection on chiller sensor fault detection

With the widespread use of building automation systems (BAS), a large amount of chiller operating data is often readily available, which provides a good basis for optimizing the control of unit sensors. However, these data are highly coupled and contain a large number of outliers which would significantly affect sensor fault detection. Compared with the traditional PCA method, the improved kernel PCA can effectively handle nonlinear data. Meanwhile, anomaly detection has shown excellent ability in rejecting outliers, so an anomaly detection step before fault detection is also necessary. Therefore, to address the problem of data presence coupling and outliers, this study attempts to investigate the impact of different anomaly detection methods on PCA and KPCA based chiller sensor fault detection. Based on the optimal combination of features screened by importance and correlation analysis to construct a fault detection model, the impact of two anomaly detection methods (Isolation Forest and K-means) on KPCA fault detection was analyzed and compared. To obtain general results, an experimental dataset of RP-1043 is used to validate the study. The results show that both anomaly detection methods can improve the overall quality of the original training data. Compared with the traditional PCA method, KPCA improves the detection efficiency by 1.7%–28.29% at all fault magnitudes. The IF-KPCA and Kmeans-KPCA improve the fault detection efficiency by 3.66%–14.63% and 3.17%-20% on top of KPCA. This study would provide reference for the optimal feature combination selection and anomaly detection method selection for practical engineering.

RST Switching Bi-controller Based Flatness for an Electronic Throttle Valve

In this paper, an RST switching bi-controller, based on flatness and on Luenberger observers, is designed to control the opening angle change of an Electronic Throttle Valve (ETV), to compensate unexpected external disturbances and to detect sensor faults. Two identified mathematical linear models are established to simulate the ETV for two different positions of the throttle plate. The use of robust RST switching bi-controller based-flatness approach by the development of closedloop control is proposed, in order to obtain a stable system tracking a desired flat trajectory. The switching between the two models using stateflow tool is based on residual values generated by using the Luenberger observers in order to detect and to localize sensor faults occurrence. The observer’s gains are determined using Linear Matrix Inequalities (LMIs) taking into account the stability of the system based on Lyapunov theory. The simulation results show the efficiency of the developed robust switching RST bi-controller based-flatness in terms of tracking the desired angle’s reference trajectory, rejecting disturbances and detecting sensor faults.

A thermodynamic-law-integrated deep learning method for high-dimensional sensor fault detection in diverse complex HVAC systems

In building Heating, Ventilation and Air Conditioning (HVAC) systems, sensor healthy operation is the foundation of the adopted control strategies to improve building energy efficiency and indoor thermal comfort. For large and complex HVAC systems where a large number of sensors are often installed, associated sensor fault detection is highly challenging due to the high dimensionality of the sensor data and complex multiple-fault scenarios. To address this challenging issue, this study proposes a novel method in which the thermodynamic laws (i.e., mass balance and energy conservation) are integrated with deep learning. By making use of the intelligence, flexibility, and efficiency of deep learning, the proposed method can easily handle high-dimensional sensor measurements. More importantly, the integration enables the thermodynamic laws (which govern the mass and heat transfer processes in HVAC systems) to be explicitly learned and thus can effectively reduce/eliminate unreasonable results (e.g., violations of mass balance or energy conservation) frequently observed from sole deep learning methods due to their pure data-driven nature. Reduction/elimination of such unreasonable results can improve associated high-dimensional sensor fault detection performance in terms of accuracy and reliability. In the case study, compared with a conventional sole deep learning method, the proposed method increased the fault detection rate by 27.2%, and significantly reduced the false alarm rate by 77.4% in the complex multi-fault scenarios. Associated analysis demonstrated that the integration of thermodynamic laws can substantially alleviate the adverse intercorrelation impacts induced by faulty measurements inside the deep neural network when multiple sensor faults occurred. The proposed method provides an effective and reliable means to ensure the sensor healthy operation in large and complex HVAC systems in particular as increasingly more sensors are installed nowadays.

Comparative Analysis of Centralized and Federated Learning Techniques for Sensor Diagnosis Applied to Cooperative Localization for Multi-Robot Systems.

Cooperation in multi-vehicle systems has gained great interest, as it has potential and requires proving safety conditions and integration. To localize themselves, vehicles observe the environment using sensors with various technologies, each prone to faults that can degrade the performance and reliability of the system. In this paper, we propose the coupling of model-based and data-driven techniques in diagnosis to produce a fault-tolerant cooperative localization solution. Consequently, prior knowledge can guide a discriminative model that learns from a labeled dataset of appropriately injected sensor faults to effectively identify and flag erroneous readings. Going further in security, we conduct a comparative study on learning techniques: centralized and federated. In centralized learning, fault indicators generated by model-based techniques from all vehicles are collected to train a single model, while federating the learning allows local models to be trained on each vehicle individually without sharing anything but the models to be aggregated. Logistic regression is used for learning where parameters are established prior to learning and contingent upon the input dimensionality. We evaluate the faults detection performance considering diverse fault scenarios, aiming to test the effectiveness of each and assess their performance in the context of sensor faults detection within a multi-vehicle system.

Fault detection based on adaptive interval observer and its application in aeroengine

This paper is concerned with the problem of sensor fault detection (FD) for discrete-time linear systems. First, considering the influence of disturbance, an adaptive law is designed based on the interval observer theory, which can effectively compress the error interval of output estimation, and improve the sensitivity to sudden faults. Then, based on adaptive interval observer (AIO), a flexible event-triggering fault detection mechanism is designed. Different from the existing methods, the event-triggering condition has greater flexibility. At the same time, the adaptive law can ensure that the zero signal belongs to the upper and lower bound of residual. Based on the robustness theory, the linear matrix inequality conditions (LMIs) satisfying the H∞ performance index are derived. Finally, the effectiveness of the proposed method is demonstrated by an aeroengine simulation example.

Recurrent neural network based sensor fault detection and isolation for nonlinear systems: Application in PWR

Real-time monitoring of sensors is crucial in attaining consistent process performance with strict safety and eco-friendly measures. This article presents a fault detection and isolation (FDI) technique that diagnoses sensor faults regardless of the utilized model structure. A Recurrent Neural Network (RNN) is utilized to estimate state variables with the inputs and outputs of the process. RNNs build predictive models of the process. Then, a bank of residuals of state variables is obtained such that each residual is responsive to a subsection of faults and unresponsive to others. Consequently, an exclusive fault signature is attained for each fault case. One of the merits of the proposed procedure is that it does not necessitate the fault history of the process or first principle models contrasting to other existing outcomes in the literature. The efficacy of the proposed FDI methodology is studied on a highly nonlinear pressurized water nuclear reactor model.

Intelligent wind turbine sensor fault detection on benchmark model using support vector machine

Intelligent wind turbine sensor fault detection on benchmark model using support vector machine

Sensor Fault Detection in A Satellite Launcher Using Adaptive Neuro-fuzzy Observer

Fault detection is performed by estimating the states and comparing them with measured values. A fault is signaled when the difference between the estimated and measured values crosses a threshold value, This paper presents the design of adaptive neuro-fuzzy observer based sensor fault detection and fault tolerant control under sensor failure conditions for a satellite launcher. In this Adaptive Neuro-Fuzzy Inference System (ANFIS), optimal shape and parameters for the membership functions and effective rule base for the fuzzy system are fixed by a neural network. Three such adaptive neuro-fuzzy inference systems act as reduced order observers that estimate the system states. Decision functions are built from estimation errors to detect the fault. If any failure is identified, the control law is modified accordingly using the estimated value replacing the failed sensor output to implement the fault tolerant control. In this work, such fault detector is designed and simulated. The individual failures of three sensors in the satellite launcher are considered and the results are discussed. The results show that the system is able to detect any sensor failure situations perfectly.

Multi-level autonomous integrity monitoring method for multi-source PNT resilient fusion navigation

For the integrity monitoring of a multi-source PNT (Positioning, Navigation, and Timing) resilient fusion navigation system, a theoretical framework of multi-level autonomous integrity monitoring is proposed. According to the mode of multi-source fusion navigation, the framework adopts the top-down logic structure and establishes the navigation source fault detection model based on the multi-combination separation residual method to detect and isolate the fault source at the system level and subsystem level. For isolated non-redundant navigation sources, the system level recovery verification model is used. For the isolated multi-redundant navigation sources, the sensor fault detection model optimized with the dimension-expanding matrix is used to detect and isolate the fault sensors, and the isolated fault sensors are verified in real-time. Finally, according to the fault detection and verification results at each level, the observed information in the fusion navigation solution is dynamically adjusted. On this basis, the integrity risk dynamic monitoring tree is established to calculate the Protection Level (PL) and evaluate the integrity of the multi-source integrated navigation system. The autonomous integrity monitoring method proposed in this paper is tested using a multi-source navigation system integrated with Inertial Navigation System (INS), Global Navigation Satellite System (GNSS), Long Baseline Location (LBL), and Ultra Short Baseline Location (USBL). The test results show that the proposed method can effectively isolate the fault source within 5 s, and can quickly detect multiple faulty sensors, ensuring that the positioning accuracy of the fusion navigation system is within 5 m, effectively improving the resilience and reliability of the multi-source fusion navigation system.

Enhancing Wind Turbine Performance: Statistical Detection of Sensor Faults Based on Improved Dynamic Independent Component Analysis

Efficient detection of sensor faults in wind turbines is essential to ensure the reliable operation and performance of these renewable energy systems. This paper presents a novel semi-supervised data-based monitoring technique for fault detection in wind turbines using SCADA (supervisory control and data acquisition) data. Unlike supervised methods, the proposed approach does not require labeled data, making it cost-effective and practical for wind turbine monitoring. The technique builds upon the Independent Component Analysis (ICA) approach, effectively capturing non-Gaussian features. Specifically, a dynamic ICA (DICA) model is employed to account for the temporal dynamics and dependencies in the observed signals affected by sensor faults. The fault detection process integrates fault indicators based on I2d, I2e, and squared prediction error (SPE), enabling the identification of different types of sensor faults. The fault indicators are combined with a Double Exponential Weighted Moving Average (DEWMA) chart, known for its superior performance in detecting faults with small magnitudes. Additionally, the approach incorporates kernel density estimation to establish nonparametric thresholds, increasing flexibility and adaptability to different data types. This study considers various types of sensor faults, including bias sensor faults, precision degradation faults, and freezing sensor faults, for evaluation. The results demonstrate that the proposed approach outperforms PCA and traditional ICA-based methods. It achieves a high detection rate, accurately identifying faults while reducing false alarms. It could be a promising technique for proactive maintenance, optimizing the performance and reliability of wind turbine systems.

A sensor fault detection, isolation, and estimation method for intelligent vehicles

Accurate and reliable sensor information is crucial for intelligent vehicles to ensure safety and reliability. It is significant to devise effective solutions to diagnose possible sensor faults. In this paper, a sensor fault detection, isolation, and estimation approach is developed for intelligent vehicle-integrated motion systems. The Luenberger observer is designed and embedded into the fault diagnosis framework to real-time estimate the system states. Interval observers are constructed for vehicle subsystems to estimate the interval bounds of the corresponding observation errors. The estimation errors enter into the corresponding intervals generated by the adaptive laws when the sensor fault is free. In the event of a sensor fault, it can be detected efficiently and the faulty sensor can be isolated. Moreover, occurred sensor fault can be estimated by modifying the interval observer structure. The simulation results in the standard J-Turn test scenario are provided to evaluate and verify the effectiveness of the proposed method.

Detection of Sensor Faults with or without Disturbance Using Analytical Redundancy Methods: An Application to Orifice Flowmeter.

Sensors and transducers play a vital role in the productivity of any industry. A sensor that is frequently used in industries to monitor flow is an orifice flowmeter. In certain instances, faults can occur in the flowmeter, hindering the operation of other dependent systems. Hence, the present study determines the occurrence of faults in the flowmeter with a model-based approach. To do this, the model of the system is developed from the transient data obtained from computational fluid dynamics. This second-order transfer function is further used for the development of linear-parameter-varying observers, which generates the residue for fault detection. With or without disturbance, the suggested method is capable of effectively isolating drift, open-circuit, and short-circuit defects in the orifice flowmeter. The outcomes of the LPV observer are compared with those of a neural network. The open- and short-circuit faults are traced within 1 s, whereas the minimum time duration for the detection of a drift fault is 5.2 s and the maximum time is 20 s for different combinations of threshold and slope.

Integrated Edge Deployable Fault Diagnostic Algorithm for the Internet of Things (IoT): A Methane Sensing Application.

The Internet of Things (IoT) is seen as the most viable solution for real-time monitoring applications. But the faults occurring at the perception layer are prone to misleading the data driven system and consume higher bandwidth and power. Thus, the goal of this effort is to provide an edge deployable sensor-fault detection and identification algorithm to reduce the detection, identification, and repair time, save network bandwidth and decrease the computational stress over the Cloud. Towards this, an integrated algorithm is formulated to detect fault at source and to identify the root cause element(s), based on Random Forest (RF) and Fault Tree Analysis (FTA). The RF classifier is employed to detect the fault, while the FTA is utilized to identify the source. A Methane (CH4) sensing application is used as a case-study to test the proposed system in practice. We used data from a healthy CH4 sensing node, which was injected with different forms of faults, such as sensor module faults, processor module faults and communication module faults, to assess the proposed model's performance. The proposed integrated algorithm provides better algorithm-complexity, execution time and accuracy when compared to FTA or standalone classifiers such as RF, Support Vector Machine (SVM) or K-nearest Neighbor (KNN). Metrics such as Accuracy, True Positive Rate (TPR), Matthews Correlation Coefficient (MCC), False Negative Rate (FNR), Precision and F1-score are used to rank the proposed methodology. From the field experiment, RF produced 97.27% accuracy and outperformed both SVM and KNN. Also, the suggested integrated methodology's experimental findings demonstrated a 27.73% reduced execution time with correct fault-source and less computational resource, compared to traditional FTA-detection methodology.

Sensor abnormality detection in multistage compressor units: A “white box” approach using tree-based genetic programming

Sensors are crucial in detecting equipment problems in various plant systems. In particular, detecting sensor abnormality is challenging in the case of utilizing the data which are acquired and stored offline (data logs). These data are normally acquired using Internet of Things (IoT) system and stored in a dedicated server. This situation presents both opportunities and challenges for exploration in sensor abnormality detection task. In this paper, we propose a multistage compressor sensor fault detection method using data logs. In the proposed method, the compressor sensor output is modeled as a function of other sensors using static approach. Subsequently, the model output is used for detecting abnormality by observing the residuals. It has been shown that the histogram of residuals offers rich information to predict abnormality of the targeted sensor. In particular, we explore the concept using Genetic Programming (GP) to generate regression model which offers more “white box” solution to the operators. There are various advantages in this approach. Firstly, the conventional “black box” approach lacks model transparency and, thus, is highly undesirable in critical systems. Secondly, equations are more easily applied in Programmable Logic Controller (PLC) if autonomous flagging is required. We also compare the proposed model with Multiple Linear Regression (MLR) and Neural Network Regression (ANN). Results show that the best generated models are comparable with the latter but with more crisp “white box” mathematical equations utilizing lesser feature inputs (four features only). This model yields R2 of 0.991 and RMSE of 2.1×10−2.

A data-driven approach for fault diagnosis in multi-zone HVAC systems: Deep neural bilinear Koopman parity

Sensor faults in heating, ventilation, and air conditioning (HVAC) systems are inevitable and result in significant energy waste. This paper presents an innovative data-driven approach for sensor fault detection and isolation in multi-zone HVAC systems. The proposed solution integrates bilinear Koopman model realization, deep learning, and bilinear parity-space. A deep neural network realizes a bilinear model, enabling bilinear parity-space sensor fault detection and isolation. This yields a reliable, accurate, and interpretable data-driven framework. The method requires no prior HVAC dynamics knowledge, relying solely on normal operation data. It diagnoses additive, multiplicative, and complete failure sensor faults while minimizing false alarms, even with severe faults. A four-zone HVAC system is simulated in TRNSYS as a case study to demonstrate the performance and efficacy of the proposed approach. The proposed bilinear deep Koopman model realization is utilized to develop a bilinear model for the four-zone HVAC system. The bilinear model is then used for designing the bilinear parity-space. Further, considering various failure scenarios, the proposed sensor fault detection and isolation framework demonstrates promising diagnosis performance. Finally, a comparison is conducted to showcase the advantages of the proposed method over earlier works based on PCA and neural networks.

Classification of Speed Sensor Faults Based on Shallow Neural Networks

This paper presents a novel speed sensor fault detection, classification, and compensation mechanism in a permanent magnet synchronous motor (PMSM) drive system. Application is based on state variable observers and shallow neural networks (NN). Classical fault detection mechanism based on state variable observers has been extended with neural networks. This enables improved detection efficiency and increases immunity to false alarms. In addition, the use of neural networks allowed for the classification of the failure type. Three types of failures are considered in the paper: signal loss, scaling error, and signal interference. The detection efficiency of the proposed solution is about 97%. On the other hand, the classification of the worst type of failure—signal loss—was achieved at the level of 100%. Other considered failure types are classified at the level of 80–90%. In addition, tests were carried out for two types of observers—model reference adaptive system and sliding mode observer—to compare the results. The work presents experimental results carried out for various operating conditions of the drive system. The failure classification times in the experimental tests were achieved at a level of less than 30 ms.

Wasserstein GAN-Based Digital Twin-Inspired Model for Early Drift Fault Detection in Wireless Sensor Networks

In this Internet of Things (IoT) era, the number of devices capable of sensing their surroundings is increasing day by day. Based on the data from these devices, numerous services and systems are now offered where critical decisions depend on the data collected by sensors. Therefore, error-free data are most desirable, but due to extreme operating environments, the possibility of faults occurring in sensors is high. So, detecting faults in data obtained by sensors is important. In this paper, a digital twin inspired detection approach is proposed, and its ability to detect a single type of fault in several sensor is analyzed. The digital equivalent of the sensor is developed using a Generative Adversarial Network (GAN). As GANs inherently performs well with images, Gramian Angular Field (GAF) encoding is used to convert timeseries data to image. The GAF encoding preserves the temporal relations of the timeseries data. The GAN is trained with the GAF images. The trained GAN model acts as the virtual representation of the sensor, and the discriminator network of the GAN model, once it has learned the pattern of normal data, is used as the fault detector. The performance of the virtual sensor is promising because it successfully generates data for normal conditions. The best fault detection accuracy achieved by the proposed model is 98.7%, which makes this GAN-based digital twin inspired approach a promising candidate for sensor fault detection.

Gain Fault Detection Method for 3-phase AC Current Sensors

This paper proposes a novel method for detecting the gain fault of 3-phase AC current sensors in power electronics systems. In conventional methods, the sensor gain fault is detected by comparing the RMS values of the detection current of each phase. However, for an unbalanced three-phase load, a higher current will flow in the phase with the lower impedance, which causes misdiagnoses. In contrast, the proposed method utilizes the product of the detection current in each phase and sum of the detection current as part of the abnormality calculation (gain error calculation) of each current sensor, which can reduce calculation errors caused by an imbalance in the three-phase load. Simulation results show that the method can calculate the abnormality of the sensors with small error even if there is an imbalance in the three-phase load. Moreover, simulation results under conditions of varying current amplitude and frequency indicates that the method can detect sensor faults under transient conditions.