Sensor Fault Diagnosis Research Articles

An efficient sensor and thermal coupling fault diagnosis methodology for building energy systems

An efficient sensor and thermal coupling fault diagnosis methodology for building energy systems

Gas Sensor Array Fault Diagnosis Based on Multi-Dimensional Fusion, an Attention Mechanism, and Multi-Task Learning.

With the development of gas sensor arrays and computational technology, machine olfactory systems have been widely used in environmental monitoring, medical diagnosis, and other fields. The reliable and stable operation of gas sensing systems depends heavily on the accuracy of the sensors outputs. Therefore, the realization of accurate gas sensor array fault diagnosis is essential to monitor the working status of sensor arrays and ensure the normal operation of the whole system. The existing methods extract features from a single dimension and require the separate training of models for multiple diagnosis tasks, which limits diagnostic accuracy and efficiency. To address these limitations, for this study, a novel fault diagnosis network based on multi-dimensional feature fusion, an attention mechanism, and multi-task learning, MAM-Net, was developed and applied to gas sensor arrays. First, feature fusion models were applied to extract deep and comprehensive features from the original data in multiple dimensions. A residual network equipped with convolutional block attention modules and a Bi-LSTM network were designed for two-dimensional and one-dimensional signals to capture spatial and temporal features simultaneously. Subsequently, a concatenation layer was constructed using feature stitching to integrate the fault details of different dimensions and avoid ignoring useful information. Finally, a multi-task learning module was designed for the parallel learning of the sensor fault diagnosis to effectively improve the diagnosis capability. The experimental results derived from using the proposed framework on gas sensor datasets across different amounts of data, balanced and unbalanced datasets, and different experimental settings show that the proposed framework outperforms the other available methods and demonstrates good recognition accuracy and robustness.

Subspace-Aided Sensor Fault Diagnosis and Compensation for Industrial Systems

Sensors are ubiquitous in automatized industrial systems. To ensure the safety of the process control, the fault diagnosis and fault-tolerant control of sensors are necessary. This paper proposes subspace-aided sensor fault diagnosis and compensation control approaches based on the data-driven stable kernel representation (SKR) and stable image representation (SIR) identified by the process data decompositions. First, this paper obtains data-driven SKR and SIR through the mapping relationship of the subspaces of signals and proposes a series of fault diagnosis and compensation approaches. Furthermore, considering the accuracy and timeliness, the paper presents an accurate online fault diagnosis and compensation approach by the online updating LQ decomposition. These approaches can perform fault diagnosis, fault estimation, and fault compensation for the multiple and different types of additive sensor faults. The effectiveness of the strategies has been verified by the numerical study and the three-tank experimental system, which has a specific engineering significance.

All-Pass NGD FIR Original Study for Sensor Failure Detection Application

This paper deals with original design, implementation, test, and potential application of finite impulse response (FIR) with uncommon all-pass negative group delay (AP-NGD) circuit. The counterintuitive AP-NGD numerical circuit theory is established. Design method of first order AP-NGD circuit is introduced with respect to the specified time-advance. The first order difference equation was synthesized, designed, and implemented in STM32 Microcontroller Unit, and experimented to validate the AP-NGD characteristics. The AP-NGD FIR prototype was tested in the time-domain with Gaussian and arbitrary waveform sensored signals. Calculated and experimented results in very good agreement with the most significant time-advance -10 s ever experimented before are obtained. The potential industrial application of the NGD predictor for sensor fault diagnosis in real-time is described.

Simultaneous sensor fault diagnosis and reconstruction for intelligent monitoring in wastewater treatment plants: An explainable deep multi-task learning model

Sensor malfunctions in wastewater treatment plants (WWTPs) significantly disrupt process control and energy usage, highlighting the critical need for effective sensor fault diagnosis and reconstruction. This study aims to introduce a novel application of a multi-task learning network in WWTPs to address the challenge of sensor malfunction by enabling simultaneous fault diagnosis and reconstruction. The proposed approach introduces an explainable deep multi-task learning autoencoder network (DMTL-UNet), which effectively allows sharing information among tasks through attention gates and residual connections. The effectiveness of the DMTL-UNet model is validated using a real-world dataset from a WWTP in South Korea. The results demonstrate the remarkable capability of the DMTL-UNet model in accurately diagnosing multiple faults (F1-score = 99.08 %) and achieving superior reconstruction performance (RMSE = 31.1175 mg/L) for faulty WWTP sensors. Moreover, implementing the DMTL-UNet model offers significant energy savings of 37.44 %, corresponding to a reduction in the aeriation cost by 1154.91 USD. Therefore, implementing the DMTL-UNet model for calibrating faulty sensors can enhance sensor reliability, improve maintenance practices, and contribute to the sustainable operation of WWTPs.

Physics-Guided Neural Network Model for Aeroengine Control System Sensor Fault Diagnosis under Dynamic Conditions

Sensor health assessments are of great importance for accurately understanding the health of an aeroengine, supporting maintenance decisions, and ensuring flight safety. This study proposes an intelligent framework based on a physically guided neural network (PGNN) and convolutional neural network (CNN) to diagnose sensor faults under dynamic conditions. The strength of the approach is that it integrates information from physics-based performance models and deep learning models. In addition, it has the structure of prediction–residual–generation-fault classification that effectively decouples the interaction between sensor faults and system state changes. First, a PGNN generates the engine’s non-linear dynamic prediction output because the PGNN has the advantage of being able to handle temporal information from the long short-term memory (LSTM) network. We use a cross-physics–data fusion scheme as the prediction strategy to explore the hidden information of the physical model output and sensor measurement data. A novel loss function that considers physical discipline is also proposed to overcome the performance limitations of traditional data-driven models because of their physically inconsistent representations. Then, the predicted values of the PGNN are compared with the sensor measurements to obtain a residual signal. Finally, a convolutional neural network (CNN) is used to classify faults for residual signals and deliver diagnostic results. Furthermore, the feasibility of the proposed framework is demonstrated on an engine sensor fault dataset. The experimental results show that the proposed method outperforms the pure data-driven approach, with the predicted RMSE being reduced from 1.6731 to 0.9897 and the diagnostic accuracy reaching 95.9048%, thereby confirming its superior performance.

Sensor fault analysis of aero-engine using ensemble SCNN and Bayesian interval estimation

Sensors of an aircraft engine have the problems of multiple kinds of faults and difficulty to diagnose. Moreover, the diagnosis performance cannot meet practical demands. Therefore, this paper proposes a novel approach for sensor fault diagnosis integrating the advantages of selective ensemble learning, siamese convolutional neural network (SCNN), and Bayesian interval estimation (BIE). The Bayesian interval estimation with hypothesis testing is developed with real-time data update to reduce the confidence interval. The attention mechanism is introduced to enhance fault diagnosis performance. Moreover, ensemble selection learning with Akaike information criteria is proposed to obtain optimal SCNNs. The theorems of convergence and optimal confidence data in ensemble learning are proven. Furthermore, this paper develops a novel improved learning algorithm based on AdaMod (Adaptive and momental bound method) and SGD (stochastic gradient descent), and achieves a new fault diagnosis algorithm to improve the diagnosis performance. The sensor data of aircraft engines from China Eastern Airlines is adopted for validation and experimental results show that proposed approach reaches an excellent tradeoff between sensor diagnosis performance, confidence interval, and runtime. The ablation studies are performed to analyze the role of leaner’s components. The proposed method improves training accuracy by 6.2%, fault diagnosis performance by 5.1%, fault prediction mean result by 4.6%, and fault prediction variance is reduced by 38.1% than the state-of-the-art methods.

A Novel Set-Valued Sensor Fault Diagnosis Method for Lithium-Ion Battery Packs in Electric Vehicles

Sensor fault diagnosis is of great significance to ensure safe battery operation. This paper proposes a novel sensor fault diagnosis method that achieves the simultaneous fault detection, fault source and type identification, and fault estimation in a comprehensive way. Specifically, a set-valued observer, featuring a state predictor and a state estimator, is first constructed and designed to guarantee the inclusion of the unavailable actual battery state due to unknown modeling errors and noises at every instant of time. Compared with the traditional observers, a distinct feature of the proposed one lies in that the calculated state predictions and estimations of the battery system at each time step are ellipsoidal sets in state space rather than single vectors. The boundedness of state prediction and estimation errors is formally proved, and the tractable design criteria for determining the real-time optimal prediction and estimation ellipsoids are also derived. As for diagnosis algorithm, fault detection is implemented based on the intersection between the prediction and estimation ellipsoids. Then, a two-layer Pearson correlation coefficient analysis mechanism is developed to identify the source and type of sensor faults. Another set-valued observer based on an augmented battery model is further designed to estimate the fault level. Finally, experimental studies of a battery cell under different sensor fault sources, types and values are elaborated to verify the effectiveness of the proposed method.

A Reduced-Order Observer-Based Method for Simultaneous Diagnosis of Open-Switch and Current Sensor Faults of a Grid-Tied NPC Inverter

This paper presents a reduced-order observer-based simultaneous diagnosis strategy for grid-tied neutral point clamped (NPC) inverters subjected to open-switch and current sensor faults. First, the augmented descriptor system is constructed for the NPC inverter to transfer the current sensor fault into a generalized state vector. Then, the matrix transformations are applied to decouple the open-switch fault from the inverter system state and the current sensor fault. Subsequently, a reduced-order observer is developed for the transformed augmented descriptor system to achieve simultaneous precise estimation of the phase current and the current sensor fault. Finally, using the estimation results, a diagnosis algorithm with an adaptive threshold is proposed, which can not only distinguish between current sensor faults and open-switch faults, but also locate the faulty power switch and estimate different types of current sensor faults. Experimental results and comparisons are provided to verify the robustness and effectiveness of the proposed fault diagnosis algorithm.

SNRepair: Systematically Addressing Sensor Faults and Self-Calibration in IoT Networks

An accurate and robust technique for sensor fault diagnosis proves useful for an uninterrupted supply of correct monitoring data across the Internet of Things (IoT) network. The manual checking and calibration of thousands of sensors deployed in IoT network is a challenging task. Further, most calibration techniques require additional hardware support for calibration. To address these issues, a unique IoT-based framework, SNRepair, has been proposed that uses modified Deep Reinforcement Learning (DRL) for detecting different types of sensor faults, such as bias, drift, complete failure and precision degradation. This technique is also capable of self-calibrating the faulty sensors automatically using the historical healthy data of the same sensor. The fault detection module identifies four types of sensor faults with 96.17% accuracy. The self-calibration module can calibrate the faulty sensors within a few seconds, thereby ensuring the uninterrupted availability of accurate monitoring information. The proposed model is robust as it can work efficiently even in noisy environments with accurate results.

An Online Metric Learning-Based Open-Switch and Current Sensor Fault Diagnosis for MSPMSM Systems

Rapid and accurate removal of faulty sectors is the conventional fault tolerance method for highly secure multi-sector permanent magnet synchronous motors (MSPMSMs). In this paper, a new fault diagnosis method for MSPMSMs based on metric learning (ML) is proposed. Its advantage is that it can complete fault detection and faulty sector location within 0.1 current fundamental cycle, which is faster than previous methods. Meanwhile, for the faulty sector, 28 types of open switch faults and current sensor faults can be accurately identified. In addition, since the fault diagnosis is based on structural characteristics, the proposed method is independent of system parameters and naturally robust to the speed and current variation. Particularly, using the consistent structural characteristics of MSPMSM sectors, an amplitude-trend distance (ATD) is presented to obviously reflect the difference of the same phase currents between healthy and faulty sectors. Subsequently, to ensure the speed of fault detection and high false alarm immunity, the ATD value under different fault states is theoretically analyzed, and a reasonable threshold is set. Compared to traditional ML methods, which requires complex computational classification methods, four characteristic parameters sector, polar, close to zero (CTZ), and sensor flag (SF) are developed to locate the faulty sector and the faulty switch or current sensor. The experiment results on four-sector PMSM verify the effectiveness of the proposed fault diagnosis method.

Structural Damage Identification Based on Convolutional Neural Network Group Considering the Sensor Fault

This article proposes a structural damage identification method based on one-dimensional convolutional neural network group considering sensor faults. The method aims to reduce the damage misjudgment caused by sensor faults. In the proposed method, according to the sensor layout, some convolutional neural network sub-models are established to extract the features from raw vibration data for sensor fault diagnosis and structural damage identification; then two convolutional neural networks groups, namely the sensor fault diagnosis group and the damage identification group are designed on the basis of the functions of each sub-model. The sensor fault diagnosis group determines whether the sensor data is abnormal and truncates the abnormal signal. The remaining normal signal are entered into the damage identification group and the final damage identification results are calculated according to the statistical decision module. The effectiveness of the devised method is verified by the IASC-ASCE benchmark structure and laboratory experiments. The results demonstrate that the sensor fault diagnosis and damage identification accuracy of each sub-model ranges from 98.54% to 99.77% and from 87.21% to 91.74% respectively at different noise levels; the damage identification group can reduce the impact of sub-model misjudgment on the structural damage identification. The accuracy of the final damage identification results is 100%. The identification time of all samples in the test set is 53.09 s and 22.93 s, respectively, for SHM benchmark and Laboratory experiment cases. And the average judgment time of each submodel in the sensor fault diagnosis group was 278 and 94 ms, and that of each submodel in the damage identification group was 294 and 105 ms, respectively, for a single test sample, which fulfills the requirements of online damage identification for structural health monitoring.

Pressure sensor fault-tolerant control for the filling phase of wet clutches

The filling pressure control is one of the key technologies of wet clutches. Pressure sensor is a key component to realize the closed-loop control of the filling pressure. However, with the working hours increasing, the pressure sensor is prone to be faulty. To ensure the filling performance of wet clutches, this article investigates a pressure sensor fault diagnosis and fault-tolerant control strategy for the filling phase of wet clutches. First, the dynamic model of a hydraulic clutch control system is established where the clutch pressure sensor fault is considered. Second, the optimal filling pressure reference trajectory is constructed by Pontryagin’s minimum principle. In the fault-free case, a feedback controller is designed to track the desire pressure trajectory with an observer estimating the unmeasurable state of the system. Third, an adaptive observer is constructed to estimate the pressure sensor fault. Based on the estimated sensor fault information, a pressure sensor fault-tolerant control strategy is proposed for the filling phase of wet clutches. Simulation results show that the proposed fault-tolerant controller can realize smooth and precise clutch-filling control with pressure sensor fault.

Multivariate statistical analysis based cross voltage correlation method for internal short-circuit and sensor faults diagnosis of lithium-ion battery system

Multivariate statistical analysis based cross voltage correlation method for internal short-circuit and sensor faults diagnosis of lithium-ion battery system

Reliability evaluation of the hybrid-redundancy sensor fault tolerate system in the dredging perception system

The accurate evaluation of equipment reliability is critical in engineering applications and especially challenging for equipment with complex structures and high reliability requirements. In this paper, a reliability evaluation method based on semi-Markov process is proposed and applied to the hybrid-redundancy sensor fault tolerant (HRSFT) system within the perception system of a cutter suction dredger (CSD). The method takes into account the harsh operating environment of the CSD sensor, and considers the temporal occurrence of sensor faults as well as the imperfection of the sensor fault diagnosis and isolation (SFDI). The outcomes of the model applied to a case study indicate that three key factors, namely, the failure rate of the sensors, the false alarm rate (FAR) of the SFDI system and the self-correction rate (SCR) of the HRSFT system, have a significant impact on the system reliability. Additionally, we performed a comparative analysis between the HRSFT and Double Redundancy Sensor Fault Tolerant (DRSFT) systems, from which the superiority of the HRSFT system in terms of reliability is demonstrated. A series of experiments were conducted with varying parameters values, to show the high adaptability of the model to the HRSFT system.

From Online Systems Modeling to Fault Detection for a Class of Unknown High-Dimensional Distributed Parameter Systems

Fault detection for distributed parameter systems reported so far is model-based in general and the performance heavily relies on the prior known model information. This restricts the usability of these methods in industrial applications. In this paper, we make the first attempt to establish a brand-new framework that contains both on-line systems modeling and fault detection of unknown high-dimensional DPSs. These two parts interact with each other in the sense that the systems modeling error is transformed into the residual signal for fault detection while the on-line modeling switches to off-line mode depending on the fault detection results. The high-dimensional DPSs are first decomposed into spatial features and temporal sequences. Then a receding-horizon scheme is applied for the temporal dynamics learning and the residual signal is converted by the temporal validation error. Experiments on sensor faults diagnosis for the thermal process of a 2-D battery cell are provided for method validation.

Sensor fault diagnosis modeling of lithium-ion batteries for electric vehicles

Electric Vehicles (EVs) are considered the replacement for gas-emitting and environmentally polluting fuel vehicles. The development of EVs relies on electronic devices and communication circuits for ease of access. The heart of EVs is the battery that requires post-run-time charging and precise maintenance. Electronic sensors attached to the batteries monitor their health, discharging, and charging rate for user notification and prolonged functioning. Therefore, the operation of the sensors is to be monitored promptly for effective battery maintenance. This article introduces a Fault Detection Method (FDM) using Operational Data (OD) accumulated from the EV. This data is first split into active and passive that references the run-time and idle time of the vehicle. The sensor operations during the idle time are considered similar due to minimal or no sensing function. Contrarily the run-time observation shows up variations that are not huge compared to the previous history. In this data verification process, recurrent learning is employed for identifying variations in the active and passive states of the EVs. If variations are continuous regardless of the active and passive states, then the sensor is identified as faulty post the manual battery life assessment. The learning is trained using the current and previous sensor observations and batter health for identifying the variations.

Bioinspired membrane learnable spiking neural network for autonomous vehicle sensors fault diagnosis under open environments

Bioinspired membrane learnable spiking neural network for autonomous vehicle sensors fault diagnosis under open environments

Robust fault detection and diagnosis of primary air data sensors in the presence of atmospheric turbulence

Abstract This paper presents a fault detection and diagnosis (FDD) algorithm for various faults in the primary air data sensors (PADS) of an aircraft in the presence of external disturbances such as atmospheric turbulence. Rapid wind variations due to turbulence induce excessive error in the externally fitted air data probe measurements, which may lead to loss of control and misinterpretations by the flight crew. In adverse environmental conditions, the FDD of air data prefers robust and adaptive air data estimates that use an analytical redundancy approach with fewer computations. The proposed method considers the kinematics of the aircraft instead of the dynamics used in the state-of-the-art algorithms. The advantage of using kinematics is that it can reduce modeling errors significantly, avoiding high false alarm rates in the FDD process. For the estimation of stable and accurate air data under external disturbance, the inertial navigation system and global positioning system (INS/GPS) output are considered instead of actual air data probe or sensor measurements. The proposed algorithm uses estimates of air data using an exponentially weighted adaptive extended Kalman filter (EW-AEKF) to detect and diagnose PADS faults, which can perform well even in the presence of uncertain noise due to atmospheric turbulence experienced during flight. The simulation was carried out to validate the algorithm with flight data obtained from the X-Plane flight simulator under moderate atmospheric turbulence. The simulation experiments were carried out using the MATLAB programming platform. The results show that the proposed method achieves satisfactory FDD performance with lower root mean square error (RMSE) and computation time than traditional EKF-based algorithms.

Fault Diagnosis and Its Applications to Fault Tolerant Control of a Turbojet Engine

This paper presents a comprehensive study of model-based fault diagnosis (FD) and a fault-tolerant control (FTC) scheme for sensor and actuator faults of turbojet engines. For actuator FD, an unbiased estimation scheme with a modified Kalman filter (KF) was developed. For sensor FD, two approaches, the generalized likelihood ratio with robust KF and the pseudo actuator model with modified KF, were investigated in a comparative study. For fault detection and isolation, test statistics are commonly employed to detect fault behavior. For FTC, integral-type sliding mode control using control reconfiguration and the reconstruction of the sensor signal was adopted with the FD schemes. The effectiveness of the employed methods was demonstrated in this study and discussed with numerical simulations.