Sensor Fault Diagnostics Research Articles

Sensor Fault Diagnostics Using Physics-Informed Transfer Learning Framework

The field of smart health monitoring, intelligent fault detection and diagnosis is expanding dramatically in order to maintain successful operation in many engineering applications. Considering possible fault scenarios that can occur in a system, indicating the type of fault in a sensor is one of the most important and challenging problems in the area of intelligent sensor fault diagnostics. Within this frame of reference, we extended the physics-informed transfer learning framework, first presented previously for a fault cause assignment, to the level of sensor fault diagnostics for a range of different fault scenarios. Hence, the framework is utilized to perform intelligent sensor fault diagnostics for the first time. The underlying dynamics of the reference system are extracted using a completely data-driven methodology and dynamic mode decomposition with control (DMDc) in order to generate time-frequency illustrations of each sample with continuous wavelet transform (CWT). Then, sensor fault diagnostics for bias, drift over time, sine disturbance and increased noise sensor fault scenarios are achieved using the idea of transfer learning with a pre-trained image classification algorithm. The classification results yields a good performance on sensor fault diagnostics with training and test accuracy along with a fair robustness level with a set of reference benchmark system parameters.

Sensor Fault-Tolerant Control of Microgrid Using Robust Sliding-Mode Observer.

This work investigates sensor fault diagnostics and fault-tolerant control for a voltage source converter based microgrid (model) using a sliding-mode observer. It aims to provide a diagnosis of multiple faults (i.e., magnitude, phase, and harmonics) occurring simultaneously or individually in current/potential transformers. A modified algorithm based on convex optimization is used to determine the gains of the sliding-mode observer, which utilizes the feasibility optimization or trace minimization of a Ricatti equation-based modification of H-Infinity () constrained linear matrix inequalities. The fault and disturbance estimation method is modified and improved with some corrections in previous works. The stability and finite-time reachability of the observers are also presented for the considered faulty and perturbed microgrid system. A proportional-integral (PI) based control is utilized for the conventional regulations required for frequency and voltage sags occurring in a microgrid. However, the same control block features fault-tolerant control (FTC) functionality. It is attained by incorporating a sliding-mode observer to reconstruct the faults of sensors (transformers), which are fed to the control block after correction. Simulation-based analysis is performed by presenting the results of state/output estimation, state/output estimation errors, fault reconstruction, estimated disturbances, and fault-tolerant control performance. Simulations are performed for sinusoidal, constant, linearly increasing, intermittent, sawtooth, and random sort of often occurring sensor faults. However, this paper includes results for the sinusoidal nature voltage/current sensor (transformer) fault and a linearly increasing type of fault, whereas the remaining results are part of the supplementary data file. The comparison analysis is performed in terms of observer gains being estimated by previously used techniques as compared to the proposed modified approach. It also includes the comparison of the voltage-frequency control implemented with and without the incorporation of the used observer based fault estimation and corrections, in the control block. The faults here are considered for voltage/current sensor transformers, but the approach works for a wide range of sensors.

A Self-Reconstructing Algorithm for Single and Multiple-Sensor Fault Isolation Based on Auto-Associative Neural Networks

Recently different approaches have been developed in the field of sensor fault diagnostics based on Auto-Associative Neural Network (AANN). In this paper we present a novel algorithm called Self reconstructing Auto-Associative Neural Network (S-AANN) which is able to detect and isolate single faulty sensor via reconstruction. We have also extended the algorithm to be applicable in multiple fault conditions. The algorithm uses a calibration model based on AANN. AANN can reconstruct the faulty sensor using non-faulty sensors due to correlation between the process variables, and mean of the difference between reconstructed and original data determines which sensors are faulty. The algorithms are tested on a Dimerization process. The simulation results show that the S-AANN can isolate multiple faulty sensors with low computational time that make the algorithm appropriate candidate for online applications.

An integrated nonlinear model-based approach to gas turbine engine sensor fault diagnostics

Aircraft engine sensor fault diagnosis is closely related technology that assists operators in managing the health of gas turbine engine assets. As all gas turbine engines will exhibit performance changes due to usage, the on-board engine model built up initially will no longer track the engine over the course of the engine’s life, and then the model-based method for sensor fault diagnosis tends to be failure. This necessitates the study of the sensor fault diagnosis techniques due to usage over its operating life. Based on our recent results, an integrated approach based on nonlinear on-board model is developed for the gas turbine engine sensor fault diagnostics in this paper. The architecture is mainly composed of dual nonlinear engine models; one is a nonlinear real-time adaptive performance model and the other a nonlinear on-board baseline model. The extended Kalman filter estimator in the nonlinear real-time adaptive performance model is used to obtain the real-time estimates of component performance, and the nonlinear on-board baseline model with performance periodically update to provide the nominal reference in flight. The novel update strategy to sensor fault threshold based on the model errors and noise level is also presented. Important results are obtained on step fault and pulse fault behavior of the engine sensor. The proposed approach is easy to design and tune with long-term engine health degradation. Finally, experiment studies are provided to validate the benefit of the engine sensor fault diagnostics.

Fault Diagnostics for Turbo-Shaft Engine Sensors Based on a Simplified On-Board Model

Combining a simplified on-board turbo-shaft model with sensor fault diagnostic logic, a model-based sensor fault diagnosis method is proposed. The existing fault diagnosis method for turbo-shaft engine key sensors is mainly based on a double redundancies technique, and this can't be satisfied in some occasions as lack of judgment. The simplified on-board model provides the analytical third channel against which the dual channel measurements are compared, while the hardware redundancy will increase the structure complexity and weight. The simplified turbo-shaft model contains the gas generator model and the power turbine model with loads, this is built up via dynamic parameters method. Sensor fault detection, diagnosis (FDD) logic is designed, and two types of sensor failures, such as the step faults and the drift faults, are simulated. When the discrepancy among the triplex channels exceeds a tolerance level, the fault diagnosis logic determines the cause of the difference. Through this approach, the sensor fault diagnosis system achieves the objectives of anomaly detection, sensor fault diagnosis and redundancy recovery. Finally, experiments on this method are carried out on a turbo-shaft engine, and two types of faults under different channel combinations are presented. The experimental results show that the proposed method for sensor fault diagnostics is efficient.

Aircraft Engine Sensor Fault Diagnostics Based on Estimation of Engine's Health Degradation

A duty in development of an on-line fault detection algorithm is to make it associate with estimation of engine's health degradation. For this purpose, an on-line diagnostic algorithm is put forward. Using a tracking filter to estimate the engine's health condition over its lifetime, can be reconstructed an onboard model, which is then made to match a real aircraft gas turbine engine. Finally, a bank of Kalman filters is applied in fault detection and isolation (FDI) of sensors for the engine. Through the bank, the real faults that have occurred can be detected and isolated. The on-line fault detection algorithm has the ability of maintaining the effectiveness over the engine's lifetime and is verified by simulation using a nonlinear engine model.

Sensor fault diagnostics for a class of non-linear systems using linear matrix inequalities

This paper develops a systematic approach to fault diagnostic system design for sensor health monitoring in Lipschitz non-linear systems. The methodology applies to non-linear systems with three or more sensors in which the state is observable through any one of the sensor measurements. Two major issues are addressed in the paper—observer design for the non-linear system to ensure directional growth of residues for failure identification and use of linear matrix inequalities for explicit design of the observer gain. The use of the methodology is demonstrated through an illustrative example.

Sensor Fault Diagnostics of Magnetically Suspended Sliders

While magnetically suspended sliders controlled under five degrees of freedom have been used as a linear guide system for wafer transportation in clean atmosphere, three degrees of freedom controlled sliders may be applied to suspensory equipments which can move on a ceiling board without an effect of friction. Hence, using this slider with a propeller, a transportation system without a track can be constructed. In this paper, sensor fault diagnostics using an observer for magnetically suspended sliders is discussed and the design method for the dead beat observer which can maintain stability in replacement of outputs caused by isolation of a faulty sensor is proposed. In experiments, disturbance signals are added to a displacement sensor and the responses are measured. Sensor faults are successfully detected and isolated by the designed diagnostic system for various disturbances.

Sensor Fault Detection and Diagnostics of Nonlinear Systems with Application to Paper Machines

This paper presents an efficient real-time method of sensor fault detection, isolation and estimation for a class of MIMO nonlinear time-varying stochastic systems. The sensor faults may be a jump or slow drift type. The presented approach is based on the combination of the separate-bias estimation algorithm to randomly time-varying bias of nonlinear systems proposed by the authors and the modified Bayes' decision algorithm. The proposed technique is used to design a sensor fault diagnostics system for the headbox of an actual paper machine. Actual headbox data have demonstrated the effectiveness of the proposed method