Time filter

Source Type

Berten O.,Free University of Colombia | Hendrick P.,Free University of Colombia | Villar A.,Intelligent Information Systems Unit | Seveno D.,University of Mons
Proceedings of the ASME Turbo Expo | Year: 2013

The Université Libre de Bruxelles/Aero-Thermo-Mechanics Department (ULB/ATM) has developed a flow bench used to mimic the complete behaviour of an aircraft gas turbine engine lubrication system. This test bench has been improved to be fully instrumented and is really multi-purposes in order to test different lubrication devices in the supply circuit (pure oil) or in the scavenge part (two-phase flow) simulating real flight conditions (oil flow rate, oil temperature and pressure⋯). The paper will first present the characteristics of the lubrication test bench and its capabilities. In a second part, it will be presented the integration and the test results of two different types of sensors into the lubrication test bench. One sensor is based on measuring the change in frequency of a quartz crystal resonator. The other one is composed of two optical sensors, which are able to monitor different properties of the oil and also to detect particles in the oil. Copyright © 2013 by ASME.

Villar A.,Intelligent Information Systems Unit | Fernandez S.,Intelligent Information Systems Unit | Gorritxategi E.,Atten2 Advanced Monitoring Technologies | Ciria J.I.,IK4 Tekniker | Fernandez L.A.,University of the Basque Country
Chemometrics and Intelligent Laboratory Systems | Year: 2014

This paper deals with the description of the optimization by variable selection methods of the multivariate calibration process of a low-cost Visible-Near Infrared (400-1100. nm) sensor, developed for the on-line monitoring of the insoluble content in diesel marine engine lubricating oil. The performance of the calibration model developed for the Vis-NIR sensor was compared with the performance of the calibration model developed with spectra obtained with a UV/Vis-NIR laboratory spectrometer. The calibration results obtained with the two devices were compared to determine the limitations of the sensor system with respect to the laboratory equipment. First, the spectra were correlated with the insoluble content analyzed in Wearcheckiberica's oil laboratories obtaining a calibration model based on Partial Least Squares-regression (PLSR). Once the pre-processing strategy had been defined, the most significant predictor variables were chosen with the help of Martens uncertainty test, interval Partial Least Squares (iPLS) and Genetic Algorithms (GA) variable selection techniques. Finally, the two models were compared based on the number of latent variables of each model of the values of the Root Mean Square Error of the Cross Validation (RMSECV), the Standard Error of Performance (SECV) and the Ratio of Prediction to Deviation (RPD). © 2013 Elsevier B.V.

Villar A.,Intelligent Information Systems Unit | Gorritxategi E.,Intelligent Information Systems Unit | Aranzabe E.,Intelligent Information Systems Unit | Fernandez S.,Intelligent Information Systems Unit | And 2 more authors.
Food Chemistry | Year: 2012

This paper describes the calibration, validation and testing process of a low-cost on-line visible-near infrared (400-1100 nm) sensor for the monitoring of fat and fatty acids content in milk during the manufacturing process of milk. The optical, mechanical and electronic designs of the sensor have been developed in Tekniker IK4 research centre just as its manufacturing process. The measurement range of the sensor is 400-1100 nm thus it covers the visible range (400-780 nm) and the short-wave near infrared (780-1100 nm). Chemometric techniques were applied with the purpose of obtaining a predictive model for each parameter correlating the spectra obtained by the sensor with the parameters analysed in the laboratory. The models were developed by Partial Least Squares Regression (PLS) obtaining one model for each parameter. The raw milk samples used in this work were provided by CAPSA (Asturias, Spain). © 2012 Elsevier Ltd. All rights reserved.

Bravo-Imaz I.,Intelligent Information Systems Unit | Garcia-Arribas A.,University of the Basque Country | Gorritxategi E.,Technical Office | Arnaiz A.,Intelligent Information Systems Unit | Barandiaran J.M.,University of the Basque Country
IEEE Transactions on Magnetics | Year: 2013

The actual trend in the field of maintenance in mechanical and electrical machinery, points towards the implementation of the so-called proactive maintenance. In this framework, different parameters of the system are monitored so that the current health state of the machinery can be precisely known. The health state of the lubricating oil is within the set of parameters that are to be monitored. Nearly 40% of the total reported malfunctions in heavy machinery are due to failures in lubrication. Among the different parameters that define the state of lubricant oil, viscosity is one of the most important. Lubricant oil prevents moving parts to get into direct contact so, inadequate viscosity may cause malfunctioning and even a fatal breakdown. Besides, the monitoring of the viscosity can help establishing the state of degradation of the lubricating oil. Magnetoelastic sensors based on the magnetoelastic resonance phenomena have been shown to be useful in the determination of viscosity. In this work we describe an experimental prototype to determine the viscosity of lubricant oils using the magnetoelastic resonance. The measurements are done using different commercial oils with viscosities ranging from 32 to 326 cSt. The amplitude of the resonance and the value of the resonance frequency are shown to correlate satisfactorily with the viscosity of the measured oil, demonstrating the possibility of developing a working real-time, on-line viscosity sensor based on this principle. We also show how the temperature effects on viscosity can be taken into account. © 1965-2012 IEEE.

Discover hidden collaborations