A study on the development of a compensation method for humidity effect in QCM sensor responses

Abstract In this study, humidity influence on quartz crystal microbalance (QCM) sensor responses during gas applications is investigated. The goal is to search a compensation method to remove humidity influence over the QCM sensor array at the data processing stage. Some experiments have been conducted to observe sensors’ behavior with the industrial gas of toluene under changing humidity conditions. The observed drifts due to humidity in the sensor responses point out that a compensation model is possible. The system consists of a sensor cell, an electronic circuitry and a data processing unit. The data processing unit has a humidity compensation module which is built using an artificial neural network (ANN). The dimensionality of the inputs to the compensation module is reduced to two using principal component analysis (PCA). The data processing unit uses the sensors’ stable state frequency shift values, and outputs the gas concentration independent from ambient humidity. The system does not need a humidity sensor. The ANN is trained using 80% of the data, and remaining is used for validation and testing. The average absolute relative error in predicting the toluene concentrations is 1.15%.

[1]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[2]  A. Šetkus,et al.  Response time based output of metal oxide gas sensors applied to evaluation of meat freshness with neural signal analysis , 2000 .

[3]  V. V. Malyshev,et al.  Investigation of gas-sensitivity of sensor structures to carbon monoxide in a wide range of temperature, concentration and humidity of gas medium , 2007 .

[4]  Jia Zhou,et al.  A highly sensitive QCM sensor coated with Ag+-ZSM-5 film for medical diagnosis , 2004 .

[5]  T. Sundic,et al.  Fuzzy inference system for sensor array calibration: prediction of CO and CH4 levels in variable humidity conditions , 2002 .

[6]  Mohammad Bagher Menhaj,et al.  Training feedforward networks with the Marquardt algorithm , 1994, IEEE Trans. Neural Networks.

[7]  Jae Ho Sohn,et al.  Non-specific conducting polymer-based array capable of monitoring odour emissions from a biofiltration system in a piggery building , 2009 .

[8]  L. Zeller,et al.  Characterisation of humidity dependence of a metal oxide semiconductor sensor array using partial least squares , 2008 .

[9]  I. Jolliffe.  Principal Component Analysis , 2002 .

[10]  Ray Sinnott,et al.  Chemical Engineering Design , 2007 .

[11]  K. R. Kashwan,et al.  Tea quality prediction using a tin oxide-based electronic nose: an artificial intelligence approach , 2003 .

[12]  Francesco Gianfelici.  Review of Random signals and systems by B. Pincibono Englewood Cliffs, NJ, Prentice-Hall, 2003 , 2009 .

[13]  F. Hossein-Babaei,et al.  Compensation for the drift-like terms caused by environmental fluctuations in the responses of chemoresistive gas sensors , 2010 .

[14]  Cihat Tasaltin,et al.  Mesomorphic phthalocyanine as chemically sensitive coatings for chemical sensors , 2003 .

[15]  Cihat Tasaltin,et al.  Predicting gas concentrations of ternary gas mixtures for a predefined 3D sample space , 2008 .

[16]  José Pedro Santos,et al.  Analysis of neural networks and analysis of feature selection with genetic algorithm to discriminate among pollutant gas , 2004 .

[17]  Cihat Tasaltin,et al.  Finding the composition of gas mixtures by a phthalocyanine-coated QCM sensor array and an artificial neural network , 2006 .

[18]  Martin T. Hagan,et al.  Neural network design , 1995 .