Model-based Sensor Fault Detection Research Articles

Real-Time Simulation and Experimental Implementation of Luenberger Observer-Based Speed Sensor Fault Detection of BLDC Motors

Abstract Modern control systems’ dependability, safety and efficiency have all been improved by studying fault-tolerant control systems (FTCS). FTCS techniques can typically be active or passive controls. The fault detection and diagnosis (FDD) method is used in this study’s active control branch to identify probable faults that could develop in the speed Hall sensors of brushless DC motors (BLDC). FDD methodologies can be categorised into two types, depending on the available data and the process involved: model-based methods and data-based methods. The proposed approach in this study explores the implementation of the Luenberger observer methodology as part of the model-based approach. The chosen methodology was practically implemented and subjected to experimental evaluation. The proposed observer relies on the residual signal, which displays the difference between the plant’s observed and estimated speed signals and serves as a failure alert for the entire system. Given the increasing demand for BLDC motors in various industrial control applications, including medical fields, automation and robotics, this particular motor was selected as a benchmark to thoroughly evaluate and validate the proposed method. The primary contribution of this paper lies in the real-time application of model-based sensor fault detection methods to BLDC motors. The efficiency of the suggested method is showcased through extensive MATLAB simulations, where the obtained results confirm the successful detection of faults with a high level of responsiveness. As a result, the project was successfully implemented in real-time, and the experimental results exhibited a close correlation with the simulated outcomes. This consistency between simulation and practical implementation validates the accuracy and reliability of the proposed methodology for detecting faults in the BLDC motor speed sensor. The results underscore the heightened reliability and safety attained by promptly and accurately detecting sensor faults during the operation of the motor.

Model-Based Sensor Fault Detection and Diagnosis in Closed-Loop Power Converters for Electric Vehicles

Electric Vehicle (EV) involves electric motors equipped with closed loop powered conversion systems which use the energy from storage batteries for vehicle propulsion and rely on sensor measurements as feedback to the controller. Therefore, sensor fault detection and diagnosis are an essential process that ensures performance and safety of EVs. From the National Renewable Energy Laboratory (NREL) study it is confirmed that electrical vehicle’s fault percentages vary, however, sensor faults constitute approximately 20-30% of total faults in them. Thus, sensor fault detection is critical for effective EV operation and requires the utmost focus from EV manufacturers and operators. Model-based techniques and Kalman Filter (KF) make it possible to detect faults on sensors in closed-loop EV power converter. It involves the use of mathematical model, sensor data and KF to identify fault before it becomes a threat and guarantees proper function of the vehicle. In this article, a state-space model of a CUK converter is developed, and an LQR closed-loop control system is designed for the power converter. A sensor fault detection method was developed for the system which is then simulated and tested under sensor bias fault conditions for both open and closed-loops control for determining the difference between them. The results showed a distinctive feature extracted between these two systems in the presence of system disturbances and noise applied to the sensor readings. The second section of simulation results in a closed-loop configuration validated the mathematical equations used for sensor fault detection, demonstrating their resistance to false alarms caused by system or component faults in addition to its ability to diagnose faults.

Model-based sensor fault detection, isolation and tolerant control for a mine hoist

Encoder is essential for speed control and very valuable in condition monitoring for a mine hoist and the failure in sensor measurements can lead to serious accidents. In this paper, a novel encoder fault detection, isolation and tolerant control strategy based on finite time observers and constrained fault-tolerant controller is proposed for a mine hoist. The hybrid nonlinear observers which can converge to the origin in finite time are employed to detect and isolate faulty sensors, a residuals evaluation unit is then used to provide reconstructed signals. The constrained fault tolerant controller is presented to guarantee steady and safe running of the mine hoist when sensor fails. This approach is feasible and practical because it does not require a complicated update process, and the fault tolerance controller also makes the hoist run more reliably. Compared with traditional ways, the proposed method has superior performance and can be more effective, which are verified by experimental results.

Development of L1-norm sliding mode observer for sensor fault diagnosis of an industrial gas turbine

In this article, a novel approach for model-based sensor fault detection and estimation of gas turbine is presented. The proposed method includes driving a state-space model of gas turbine, designing a novel L1-norm Lyapunov-based observer, and a decision logic which is based on bank of observers. The novel observer is designed using multiple Lyapunov functions based on L1-norm, reducing the estimation noise while increasing the accuracy. The L1-norm observer is similar to sliding mode observer in switching time. The proposed observer also acts as a low-pass filter, subsequently reducing estimation chattering. Since a bank of observers is required in model-based sensor fault detection, a bank of L1-norm observers is designed in this article. Corresponding to the use of the bank of observers, a two-step fault detection decision logic is developed. Furthermore, the proposed state-space model is a hybrid data-driven model which is divided into two models for steady-state and transient conditions, according to the nature of the gas turbine. The model is developed by applying a subspace algorithm to the real field data of SGT-600 (an industrial gas turbine). The proposed model was validated by applying to two other similar gas turbines with different ambient and operational conditions. The results of the proposed approach implementation demonstrate precise gas turbine sensor fault detection and estimation.

Threshold Tuning-Based Wearable Sensor Fault Detection for Reliable Medical Monitoring Using Bayesian Network Model

As the medical body sensor network (BSN) is usually resource limited and vulnerable to environmental effects and malicious attacks, faulty sensor data arise inevitably which may result in false alarms, faulty medical diagnosis, and even serious misjudgment. Thus, faulty sensory data should be detected and removed as much as possible before being utilized for medical diagnosis-making. Most available works directly employed fault detection schemes developed in traditional wireless sensor network (WSN) for body sensor fault detection. However, BSNs adopt a very limited number of sensors for vital information collection, lacking the information redundancy provided by densely deployed sensor nodes in traditional WSNs. In light of this, a Bayesian network model-based sensor fault detection scheme is proposed in this paper, which relies on historical training data for establishing the conditional probability distribution of body sensor readings, rather than the redundant information collected from a large number of sensors. Furthermore, the Bayesian network-based scheme enables us to minimize the inaccuracy rate by optimally tuning the threshold for fault detection. Extensive online dataset has been adopted to evaluate the performance of our fault detection scheme, which shows that our scheme possesses a good fault detection accuracy and a low false alarm rate.

모델 기반 연료전지 스택 온도 센서 고장 감지 및 판별

본 연구에서는 PEM 연료전지 온도 센서의 고장을 감지 및 판별할 수 있는 모델 기반 센서 고장 감지 방법이 적용된다. 연료전지 차량이 작동하는 과정에서 스택 온도는 연료전지의 내구성에 영향을 미친다. 따라서 고장 진단 알고리즘이 고장 신호를 감지하는 것은 중요하다. 센서 고장 감지의 주요 목적은 연료전지 시스템의 안정적인 작동을 보장하여 고온과 저온으로부터 스택을 보호하는 것이다. 상태 공간에 기반한 패러티 방정식이 스택 온도와 냉각수 입구 온도와 같은 센서 고장을 감지하는데 적용되며, 잔차는 정상적인 온도 신호와 비교된다. 그리고 잔차는 현재의 센서 고장을 감지하는 다양한 고장 시나리오에 의해 평가된다. 결론적으로, 본 연구에서 설계된 고장 알고리즘이 고장 신호를 감지할 수 있다.

Design and optimisation of a distributive model-based sensor fault detection method for automated in-network execution in a wireless sensor network

In this paper, a distributed model-based sensor fault detection method is presented for detecting and identifying spike faults without the requirement of the existence of reference sensors. This method partitions the sensor network into sensor pairs and carries out fault diagnosis within these sensor pairs based on autoregressive with exogenous input time series analysis. The performance of the proposed method is evaluated by implementing the algorithm in a 16-node wireless sensor network deployed to monitor the traffic induced accelerations of the Grove Street Bridge (Ypsilanti, Michigan). Spike faults are generated on-site and superimposed on the acceleration measurement before being acquired by some of the monitoring system wireless sensors. In addition to accuracy evaluation, this study focuses on the relationship between the detection accuracy and three different network partition methods. Based on this relationship, a communication energy saving partition method is presented. The proposed algorithm achieved a detection rate of over 85% yet reduced communication energy by more than 54% when compared to a centralised fault detection method implemented in the monitoring system base station.

Design and optimisation of a distributive model-based sensor fault detection method for automated in-network execution in a wireless sensor network

In this paper, a distributed model-based sensor fault detection method is presented for detecting and identifying spike faults without the requirement of the existence of reference sensors. This method partitions the sensor network into sensor pairs and carries out fault diagnosis within these sensor pairs based on autoregressive with exogenous input time series analysis. The performance of the proposed method is evaluated by implementing the algorithm in a 16-node wireless sensor network deployed to monitor the traffic induced accelerations of the Grove Street Bridge (Ypsilanti, Michigan). Spike faults are generated on-site and superimposed on the acceleration measurement before being acquired by some of the monitoring system wireless sensors. In addition to accuracy evaluation, this study focuses on the relationship between the detection accuracy and three different network partition methods. Based on this relationship, a communication energy saving partition method is presented. The proposed algorithm achieved a detection rate of over 85% yet reduced communication energy by more than 54% when compared to a centralised fault detection method implemented in the monitoring system base station.

Model-based sensor fault detection and isolation method for a vehicle dynamics control system

Conventional vehicle electronic stability control requires one steering-wheel angle sensor, one lateral acceleration sensor and one yaw rate sensor to obtain a good control performance. The control system stops working when a sensor fault is detected, which means that the vehicle runs in an unprotected state. Thus, various sensor fault diagnosis algorithms have been designed to detect and isolate the faulty sensor, but these algorithms also can be used for fault-tolerant control to preserve the safety of the vehicle. However, determining which of the different sensors is faulty is very difficult as the conventional residual comparison algorithm can only find the existence of a sensor fault but cannot locate the faulty sensor, and very few research studies have focused on this problem. In this paper, an ingenious sensor fault diagnosis algorithm is proposed. The sensor fault is detected, located and isolated by cross-checking with three different yaw rate estimates. The steering-wheel angle observer and the lateral acceleration observer are designed to provide corresponding estimated sensor signals which are employed to estimate the different yaw rates by using an extended Kalman filter. A novel decision-making process is carefully designed to locate the faulty sensor based on the different yaw rate residuals. Electronic stability control is not interrupted as the faulty sensor signal is reconfigured by the estimated signal. Experimental tests on a real car show that the proposed algorithm is efficient for detecting the sensor fault and identifying which sensor is faulty. Simulations show that the vehicle stability control strategy based on the proposed sensor fault-tolerant control algorithm has a better performance than the traditional control strategy does.

Sensor fault detection and isolation for a lithium-ion battery pack in electric vehicles using adaptive extended Kalman filter

This paper presents an effective model-based sensor fault detection and isolation (FDI) scheme for a series battery pack with low computational effort. The large number of current and voltage sensors in the battery pack, make it of high computational complexity. The major purpose of sensor FDI is to guarantee the healthy operations of the battery management system (BMS), and thus to prevent the battery from over-charge and over-discharge. In the voltage sensors fault scenarios, the most possibly being over-charged and over-discharged cells are these two cells with the maximum and minimum voltage respectively. Within the proposed scheme, these two cells are monitored in real time to diagnose the pack current sensor fault, or a voltage sensor fault of these two cells, while the rest cells are monitored offline with a long time interval, guaranteeing other voltage sensors working normally. For the scheme implementation, adaptive extended Kalman filter (AEKF) is used to estimate the battery states of each individual cell, and the estimated output voltage is compared with the measured voltage to generate a residual. Then the residuals are evaluated by a statistical inference method that determines the presence of the fault. Finally, the effectiveness of the proposed sensor FDI scheme is experimentally validated with a series battery pack under the UDDS driving cycles.

Robustness studies of sensor faults and noises for semi-active control strategies using large-scale magnetorheological dampers

This paper presents comparative studies for various sensor faults and noise effects on the performance of several recently proposed semi-active control algorithms for the control of large-scale magnetorheological dampers. Sensor faults or noises due to various environmental factors and long-term deteriorations may cause degradations in the performance of the control system. In this paper, the authors have carried out an in-depth literature review of the sensor fault or malfunction and noise problems, diagnostic methods and compensation approaches applicable to semi-active control algorithms. For three recently developed semi-active controllers, namely clipped optimal control, decentralized output feedback polynomial control and Lyapunov controller, an extensive study has been carried out on the robustness of these controllers during sensor faults and noise effects. Based on this research, a new real-time qualitative model-based sensor fault detection and diagnosis (RMSFDD) method has been proposed. This novel RMSFDD method shows a good performance during diverse sensor fault types in semi-active control with a real-time application without any training and heavy computational cost.

Model-Based and Signal-Based Gearbox Sensor Fault Detection, Identification and Accommodation

Model-Based and Signal-Based Gearbox Sensor Fault Detection, Identification and Accommodation

Real-Time Data Fusion and MEMS Sensors Fault Detection in an Aircraft Emergency Attitude Unit Based on Kalman Filtering

The design, realization, and experimental validation of an original avionic attitude estimation unit are presented. The core of the system is a nine-state extended Kalman filter that optimally blends complementary kinematic data provided by orthogonal triads of inertial micro-electro-mechanical systems sensors: rate gyros (short-term fast dynamics) and accelerometers (long-term static reference). The unit is embedded in a novel aircraft emergency guidance system based on miniaturized solid-state sensors. While achieving the required extreme compactness, state-of-the-art performance is preserved: 50 Hz update rate, 0.1 <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex Notation="TeX">$^{\circ}$</tex></formula> angular resolution, 0.5 <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$^{\circ}$</tex></formula> static accuracy, and 2 <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$^{\circ}$</tex></formula> dynamic accuracy (400 <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$^{\circ}/{\rm s}$</tex></formula> max. angular rate, 10 g max. acceleration), all experimentally verified and granted over the extended thermal range. The selection of the state variables has been carefully trimmed in order to maximize the performance/speed tradeoff for real-time running in an embedded processor. The adoption of the Kalman observer also enables the implementation of model-based sensor fault detection with no extra computational cost.

Survey and application of sensor fault detection and isolation schemes

Model-based sensor fault detection, isolation and accommodation (SFDIA) is a direction of development in particular with UAVs where sensor redundancy may not be an option due to weight, cost and space implications. SFDIA via neural networks (NNs) have been proposed over the years due to their nonlinear structures and online learning capabilities. The majority of papers tend to consider single sensor faults. While useful, this assumption can limit application to real systems where sensor faults can occur simultaneously or consecutively. In this paper we consider the latter scenario, where it is assumed that a 1 s time gap is present between consecutive faults. Furthermore few applications have considered fixed-wing UAVs where full autonomy is most needed. In this paper an EMRAN RBF NN is chosen for modelling purposes due to its ability to adapt well to nonlinear environments while maintaining high computational speeds. A nonlinear UAV model is used for demonstration, where decoupled longitudinal motion is considered. System and measurement noise is also included in the UAV model as wind gust disturbances on the angle of attack and sensor noise, respectively. The UAV is assumed to operate at an initial trimmed condition of speed, 32 m/s and altitude, 1000 m. After 30 separate SFDIA tests implemented on a 1.6 GHz Pentium processor, the NN-SFDIA scheme detected all but 2 faults and the NN processing time was 97% lower than the flight data sampling time.

SFDIA of consecutive sensor faults using neural networks – demonstrated on a UAV

Neural network based sensor fault detection, isolation and accommodation (NN-SFDIA) is becoming a popular alternative to traditional linear time-invariant model-based sensor fault detection, isolation and accommodation (SFDIA) schemes, such as observer-based methods. Their online training capabilities and ability to model complex nonlinear systems have attracted much research interest in the applications area of neural networks. In this article, we design an NN-SFDIA scheme to detect multiple sensor faults in an unmanned air vehicle (UAV). Model-based SFDIA is a direction of development in particular with UAVs where sensor redundancy may not be an option due to weight, cost and space implications. In this article, a maximum of three consecutive faults are assumed in the pitch gyro, normal accelerometer and angle of attack sensor of a nonlinear UAV model. Furthermore, a novel residual generator which is designed to minimise the false alarm rates and missed faults, is implemented. After 33 separate SFDIA tests implemented on a 1.6 GHz Pentium processor, the NN-SFDIA scheme detected all but three faults with a fast execution time of 0.55 ms per flight data sample.

Real-Time Model-Based Fault Detection and Isolation for UGVs

The paper presents a model-based sensor fault detection and isolation system applied in real-time to unmanned ground vehicles. Structural analysis is applied on the nonlinear model of the vehicle for building the residual generation module, followed by an ad-hoc residual evaluation module for detecting single and multiple sensor faults. The overall proposed diagnosis scheme has been tested in real-time on a real mobile robot in an outdoors environment and for different tasks. The obtained experimental results are satisfactory in terms of diagnosis performance and real-time implementation.

Model-based sensor fault detection and isolation in balanced three-phase signals

The inherent redundancy of balanced three-phase currents (voltages) is exploited to design a fault detection and isolation system for current (voltage) sensors. To this end, a model for balanced three-phase signals (either sinusoidal or non sinusoidal) is presented. It is used to design a residual generator based on the so-called Generalized Observer Scheme (GOS). Next, a decision system made of a combination of vector CUSUM (Cumulative Sum) algorithms is used in order to process the residual vector and achieve detection and isolation of incipient additive faults. The approach is validated on experimental data. Fault detection and isolation (FDI) is achieved for a large range of fault magnitudes and an accurate estimation of the fault occurrence time is also determined.

Model based sensor fault detection for an active vehicle suspension

At the beginning, a short introduction into the topic of active vehicle suspensions and the LOLIMOT neuronal network, which is used to identify accurate models of the active suspension system, is presented. Subsequently, physical mathematical models of the considered active vehicle suspension implemented in a test rig are derived. The structure of these nonlinear physical models is then used to define the inputs and outputs of the local linear neuronal network LOLIMOT. Hence, a semi-physical modelling approach results. This network is then trained with test rig data, in order to obtain accurate parity equations for model based fault detection. Various faults are simulated with test rig data and it is shown, how these faults are detected.