Relationship between odour intensity assessed by human assessor and TGS sensor array response

Sensor array measurements and odour intensity were performed for single and two component gas mixtures. Toluene and ethyl acetate were chosen for the experiment as gas mixture components. TGS sensors were used for sensor array measurements. Odour intensity measurements were performed by human assessor, with labelled magnitude scale (LMS). Based on results of both measurements, models transforming sensor responses into human assessment of odour intensity were developed. In case of single component gas mixture, analytical, power function model was proposed. The independent variable was sensor response to toluene or ethyl acetate and the dependent variable was adequate odour intensity. In case of two component gas mixture the intensity of odour was successfully modelled with neural model, generalised regression network (GRN). The network input was the response of sensor array to toluene–ethyl acetate mixture and network output was the intensity of odour.