Neural observer-based small fault detection and isolation for uncertain nonlinear systems

Small faults (some weak faults with a tiny magnitude) are difficult to detect and may cause severe problems leading to degrading the system performance. This paper proposes an approach to estimate, detect, and isolate small faults in uncertain nonlinear systems subjected to model uncertainties, disturbances, and measurement noise. A robust observer is developed to alleviate the lack of full state measurement. Using the estimated state, a dynamical radial basis function neural networks observer is designed in form of LMI problem to accurately learn the function of the inseparable mixture between modeling uncertainty and the small fault. By exploiting the knowledge obtained by the learning phase, a bank of observers is constructed for both normal and fault modes. A set of residues is achieved by filtering the differences between the outputs of the bank of observers and the monitored system output. Due to the noise dampening characteristics of the filters and according to the smallest residual principle, the small faults can be detected and isolated successfully. Finally, rigorous analysis is performed to characterize the detection and isolation capabilities of the proposed scheme. Simulation results are used to prove the efficacy and merits of the proposed approach.     

Rolling element bearing fault diagnosis for rotating machinery using vibration spectrum imaging and convolutional neural networks

The International Journal of Advanced Manufacturing Technology - In this paper, we propose a novel method for the classification of bearing faults using a convolutional neural network (CNN) and...     

Effect of tempering on mechanical proprieties and corrosion behavior of X70 HSLA steel weldments

The aim of this work is to study the effect of tempering on the corrosion behavior of X70 HSLA (high-strength low alloy) steel weldments in aerated 1      

A new growing pruning deep learning neural network algorithm (GP-DLNN)

Neural Computing and Applications - During the last decade, a significant research progress has been drawn in both the theoretical aspects and the applications of Deep Learning Neural Networks....     

Cotton fabric graft copolymerization using microwave plasma. I. Universal attenuated total reflectance–FTIR study

The effect of microwave plasma on lightweight cotton fabric was investigated. N₂‐plasma, O₂‐plasma, and Ar‐plasma were obtained using a microwave generator at 2.45 GHz under vacuum. The universal attenuated total reflectance–Fourier transform infrared (UATR–FTIR) instrument was used to monitor the changes created after N₂‐, O₂‐, and Ar‐plasma treatments. The exposure of cotton fabrics to the plasma for 240 s with a microwave power of 500 W was sufficient to create active carbonyl groups, as shown by the presence of a peak around 1725 cm^?1 in the FTIR spectra of the treated cotton fabrics. Ar‐plasma was found to generate more active groups than N₂‐ and O₂‐plasmas. The active centers created within the cellulose chains were used to initiate copolymerization reactions with vinyl monomers to impart hydrophobic character to lightweight cotton fabric. The efficiency of the grafting process and the presence of grafted monomers on fabric surface were confirmed using UATR–FTIR. Testing of treated fabric revealed that excellent water repellency was obtained. © 2004 Wiley Periodicals, Inc. J Appl Polym Sci 93: 145–154, 2004     

Prediction of the hardness profile of an AISI 4340 steel cylinder heat-treated by laser - 3D and artificial neural networks modelling and experimental validation

This paper presents a comprehensive approach developed to design an effective prediction model for hardness profile in laser surface transformation hardening process. Based on finite element method and Artificial neural networks, the proposed approach is built progressively by (i) examining the laser hardening parameters and conditions known to have an influence on the hardened surface attributes through a structured experimental investigation, (ii) investigating the laser hardening parameters effects on the hardness profile through extensive 3D modeling and simulation efforts and (ii) integrating the hardening process parameters via neural network model for hardness profile prediction. The experimental validation conducted on AISI4340 steel using a commercial 3 kW Nd:Yag laser, confirm the feasibility and efficiency of the proposed approach leading to an accurate and reliable hardness profile prediction model. With a maximum relative error of about 10 % under various practical conditions, the predictive model can be considered as effective especially in the case of a relatively complex system such as laser surface transformation hardening process.     

Thermal buckling analysis of nanoplates based on nonlocal elasticity theory with four-unknown shear deformation theory resting on Winkler–Pasternak elastic foundation

The thermal buckling analysis of nanoplates is based on nonlocal elasticity theory with four-unknown shear deformation theory resting on Winkler–Pasternak elastic foundation. The nanoplate is assumed to be under three types of thermal loadings, namely uniform temperature rise, linear temperature rise, and nonlinear temperature rise through the thickness. The theory involves four unknown variables with small-scale effects, as against five in the case of other higher-order theories and first-order shear deformation theory. Closed-form solution for theory was also presented. Results are presented to discuss the influences of the nonlocal parameter, aspect ratio, side-to-thickness ratio, and elastic foundation parameters on the thermal buckling characteristics of analytical rectangular nanoplates.     

Active fault tolerant control based on a neuro fuzzy inference system applied to a two shafts gas turbine

The main aim of the present work is development of an active fault tolerant control for two shafts gas turbine fault detection and isolation based on a neuro fuzzy inference system adaptive approach. This approach combines the advantages of the neural networks with the fuzzy inference systems. The reconfiguration mechanism of the proposed active fault tolerant control is performed by detecting the malfunction of the studied gas turbine in an automatic manner. The obtained experimental results are presented to illustrate the great interest of the developed active fault tolerant control approach and to demonstrate its effectiveness in maintaining the stability with acceptable performance under the presence of defects in the presented gas turbine.     

FTIR analysis of crosslinked cotton fabric using a ZnSe–universal attenuated total reflectance

Two cotton fabrics were treated with increasing amounts of a textile‐finishing agent (1,3‐dimethyl‐4,5‐dihydroxy‐2‐imidazolidinone) to impart durable press properties. The Universal Attenuated Total Reflectance Fourier Transform Infrared (UATR–FTIR) with a ZnSe–Diamond composite crystal was used to determine the amount of the crosslinking agent effectively linked to the cellulose after the required laundering cycles. Textile performance testing conducted on treated and untreated fabrics demonstrated the effectiveness of the treatment applied. The results obtained showed very good correlation between AATCC grading, automatic image analysis of fabric smoothness, textile performance testing, and the amount of finish as evaluated by the UATR–FTIR. The ZnSe–Diamond composite FTIR accessory was proven to be a fast and precise nondestructive technique to evaluate the amount of the crosslinking agent linked to the cellulose macromolecules. © 2005 Wiley Periodicals, Inc. J Appl Polym Sci 96: 392–399, 2005     

New oligomeric containing aliphatic moiety phthalonitrile resins: their mechanical and thermal properties in presence of silane surface-modified zirconia nanoparticles

This study deals with the influences of both the length of the aliphatic spacer within the phthalonitrile monomers backbones, and the amount of the silane surface modified zirconia nanoparticles on the mechanical and thermal properties of the so-called second generation phthalonitrile resins. Investigation on the curing behavior under differential scanning calorimeter outlined an important gain in the processability as the aliphatic spacer became longer. Results from the mechanical tests revealed that changing the length of the aliphatic spacer affects the mechanical properties in different ways. For instance, as the aliphatic spacer became longer, the toughness state was enhanced. At the same time, the tensile modulus and stress as well as the microhardness values were slightly reduced. It was also noticed that the introduction of the reinforcing phase caused an increase in all the tested mechanical properties. Furthermore, results from the thermogravimetric analysis and dynamic mechanical analysis revealed that reducing the length of the aliphatic spacer and adding nanofillers caused an increase in the thermal stability, storage modulus, and glass transition temperature. Moreover, a morphological study has been conducted under scanning electron microscope and transmission electron microscope to put in light the mechanisms of enhancements. Finally, this study demonstrated that the excellent properties of the phthalonitrile resins can be tailored by two ways either by monomers design or by inorganic nanoparticles reinforcement.