Academia.eduAcademia.edu

Figure 2 – uploaded by MUHAMMAD Awais

See full PDFdownloadDownload figure
Table 2. Comparison with other published work. Table 3. Summary of classification results.

Table 2 Comparison with other published work. Table 3. Summary of classification results.

Related Figures (15)

Figure 1. Block diagram of gases—data collection and classification. The block diagram for the flow of gas sensing and classification is shown in Figure 1. The MOS sensor (sensor array) collects the gas mixture produced by an industrial plant, and then, further processes the mixture via a preprocessing technique using PCA.
Figure 1. Block diagram of gases—data collection and classification. The block diagram for the flow of gas sensing and classification is shown in Figure 1. The MOS sensor (sensor array) collects the gas mixture produced by an industrial plant, and then, further processes the mixture via a preprocessing technique using PCA.
Figure 3. Time response of sensors based on metal oxide semiconductor (MOS) techniques for three various samples of volatile organic compounds (VOCs), each having three statistical records for the database (D). the same manner, points 7 to 12 show the curves for gas E: point 7 shows the response of sensor TGS 20, point 8 represents the sensor MICS 5135, point 9 depicts the response of MICS 5521, point 1C ows the response of TGS 2602, point 1 shows the sensor response of TGS 2611, point 12 represents 2 sensor response of TGS 2600. In a similar way, points 13 to 18 show the curves for gas P, point 12 ows the response of TGS 2620, point 4 shows the response of sensor MICS 5521, point 15 shows 2 response of TGS 2600, point 16 shows the response of the sensor TGS 2602, point 17 shows the sponse of MICS 5135, point 18 shows t > six sensors show different responses he response of TGS 2611. From this figure, we can see that all for different gases.
Figure 3. Time response of sensors based on metal oxide semiconductor (MOS) techniques for three various samples of volatile organic compounds (VOCs), each having three statistical records for the database (D). the same manner, points 7 to 12 show the curves for gas E: point 7 shows the response of sensor TGS 20, point 8 represents the sensor MICS 5135, point 9 depicts the response of MICS 5521, point 1C ows the response of TGS 2602, point 1 shows the sensor response of TGS 2611, point 12 represents 2 sensor response of TGS 2600. In a similar way, points 13 to 18 show the curves for gas P, point 12 ows the response of TGS 2620, point 4 shows the response of sensor MICS 5521, point 15 shows 2 response of TGS 2600, point 16 shows the response of the sensor TGS 2602, point 17 shows the sponse of MICS 5135, point 18 shows t > six sensors show different responses he response of TGS 2611. From this figure, we can see that all for different gases.
Figure 4. Multivariate technique on organic gases for different feature loading designs. readings (Figure 4). Herein, we present one to six different irregular points in this design. In Figure 4, point one shows TGS 2600, point two shows TGS 2602, point three shows TGS 2611, point four shows TGS 2620, point five shows MICS 5135, and point six shows MICS 5521.
Figure 4. Multivariate technique on organic gases for different feature loading designs. readings (Figure 4). Herein, we present one to six different irregular points in this design. In Figure 4, point one shows TGS 2600, point two shows TGS 2602, point three shows TGS 2611, point four shows TGS 2620, point five shows MICS 5135, and point six shows MICS 5521.
Figure 5. Multivariate technique on organic gases for different feature notch design. Figure 4 shows the results of the data difference design. It had no sensor set response correlated with the other. Figure 5 shows that the notch design for multivariate techniques examines the relationships among data difference considerations. Here, a multivariate method PCA was used to discriminate all three clusters of A, E, and P, represented in one cluster each, using a separate boundary. Figure 5 also shows that PCA can discriminate all these three clusters by the separate boundary. In Figure 6, both variable scatter models and observations are presented together. On the left part of Figure 6, the principal component of MOS sensors is presented, while observations of gas samples are presented on the right side.
Figure 5. Multivariate technique on organic gases for different feature notch design. Figure 4 shows the results of the data difference design. It had no sensor set response correlated with the other. Figure 5 shows that the notch design for multivariate techniques examines the relationships among data difference considerations. Here, a multivariate method PCA was used to discriminate all three clusters of A, E, and P, represented in one cluster each, using a separate boundary. Figure 5 also shows that PCA can discriminate all these three clusters by the separate boundary. In Figure 6, both variable scatter models and observations are presented together. On the left part of Figure 6, the principal component of MOS sensors is presented, while observations of gas samples are presented on the right side.
Figure 6. Multivariate technique on organic gases for different features in the Bi design.
Figure 6. Multivariate technique on organic gases for different features in the Bi design.
Figure 7. Multivariate technique on organic gases fractional feature loading design. Figure 7 indicates the results of the data fractional design. It shows that the MOS sensor set {5,6} responses were correlated with each other or had the same properties. In Figure 7, point one shows TGS 2600, point two shows TGS 2602, point three shows TGS 2611, point four shows TGS 2620, point five shows MICS 5135, and point six shows MICS 5521. The responses were opposite to each other for the other sets of MOS sensors. The notch design shown in Figure 8 inspects the coordination among the study of data difference. Figure 8 further shows that the multivariate technique (PCA) can discriminate against all three clusters of A, E, and P by the separate boundary. Hence, PCA helps classification in incomplete data preprocessing methods, which can ultimately help differentiate the three gases from the mixture. Figure 9 summarizes the results of the loading and notch designs, known together as the Bi design. Both variable scatter models and observations are shown in Figure 9. On the left part of Figure 9, the principal MOS sensor components are presented, while the observation of gas samples are presented on the right side of Figure 9.
Figure 7. Multivariate technique on organic gases fractional feature loading design. Figure 7 indicates the results of the data fractional design. It shows that the MOS sensor set {5,6} responses were correlated with each other or had the same properties. In Figure 7, point one shows TGS 2600, point two shows TGS 2602, point three shows TGS 2611, point four shows TGS 2620, point five shows MICS 5135, and point six shows MICS 5521. The responses were opposite to each other for the other sets of MOS sensors. The notch design shown in Figure 8 inspects the coordination among the study of data difference. Figure 8 further shows that the multivariate technique (PCA) can discriminate against all three clusters of A, E, and P by the separate boundary. Hence, PCA helps classification in incomplete data preprocessing methods, which can ultimately help differentiate the three gases from the mixture. Figure 9 summarizes the results of the loading and notch designs, known together as the Bi design. Both variable scatter models and observations are shown in Figure 9. On the left part of Figure 9, the principal MOS sensor components are presented, while the observation of gas samples are presented on the right side of Figure 9.
Figure 8. Multivariate technique on organic gases fractional feature notch design.
Figure 8. Multivariate technique on organic gases fractional feature notch design.
Figure 9. Multivariate technique on organic gases fractional feature Bi design.
Figure 9. Multivariate technique on organic gases fractional feature Bi design.
Figure 10. Multivariate technique on organic gases relative feature loading design. presented on the left part of Figure 12 while observation of gas samples are seen on the right part of Figure 12. The primary sensor component was usually distributed and did not show the same act for preprocessing done by relevant data. The information’s point was too crowded because it became difficult to distinguish the smell by the relative method.
Figure 10. Multivariate technique on organic gases relative feature loading design. presented on the left part of Figure 12 while observation of gas samples are seen on the right part of Figure 12. The primary sensor component was usually distributed and did not show the same act for preprocessing done by relevant data. The information’s point was too crowded because it became difficult to distinguish the smell by the relative method.
Figure 11. Multivariate technique on organic gases relative feature notch design.
Figure 11. Multivariate technique on organic gases relative feature notch design.
Table 1. Summary of the performance of the classifier for data sensors sets. Table 2 shows a comparison between this paper’s proposed method with past studies. Prior research has implemented MIR-PLS, GC-MS, and PLS techniques for differentiation and classification of A, E, and P. Each study presented in Table 2 used either A, E, or P, but did not consider the mixture of these three gases. Additionally, the scope of their studies was focused on detecting human diseases by the gases coming out from petrochemical storages. The proposed method presented in this paper overcame the disadvantage of previous studies by designing multivariate techniques to differentiate and classify the A, E, and P mixture. The three preprocessing methods used here easily organize the gas mixture by making a cluster boundary between the gases. This classification will enable people working in the industry to receive alerts regarding the presence of these gases in the environment and to take preventative measures.
Table 1. Summary of the performance of the classifier for data sensors sets. Table 2 shows a comparison between this paper’s proposed method with past studies. Prior research has implemented MIR-PLS, GC-MS, and PLS techniques for differentiation and classification of A, E, and P. Each study presented in Table 2 used either A, E, or P, but did not consider the mixture of these three gases. Additionally, the scope of their studies was focused on detecting human diseases by the gases coming out from petrochemical storages. The proposed method presented in this paper overcame the disadvantage of previous studies by designing multivariate techniques to differentiate and classify the A, E, and P mixture. The three preprocessing methods used here easily organize the gas mixture by making a cluster boundary between the gases. This classification will enable people working in the industry to receive alerts regarding the presence of these gases in the environment and to take preventative measures.
Figure 12. Multivariate technique on organic gases relative feature Bi design.
Figure 12. Multivariate technique on organic gases relative feature Bi design.
5. Conclusions
5. Conclusions
Table A1. Summary of the literature review. Appendix C. Sensor Array
Table A1. Summary of the literature review. Appendix C. Sensor Array
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved