Planar electromagnetic sensors array for nitrate and sulphate detection

• Main Author: Mohamad Nor, Alif Syarafi
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utm.my/id/eprint/48921/25/AlifSyarafiMFKE2015.pdf; http://eprints.utm.my/id/eprint/48921/; http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:85030
• Institution: Universiti Teknologi Malaysia

This work expounds the development of three types of sensor arrays based on planar electromagnetic for environmental monitoring. Three types of sensor array are proposed: parallel, star, and delta. The modeling and simulation of all types of sensor array have been carried out to calculate the sensor’s impedance value. The contamination state has been simulated by altering the electrical property values of the environment at the model subdomain to represent water contamination. The simulation results agree with the experimental trends. The star array configuration shows the highest simulated inductance and capacitance responses with the best signal strength and sensitivity. Moreover, experiments have been conducted to determine the relationship between sensor’s impedance and water contamination due to nitrate and sulphate. The sensors have been tested with added distilled water with different concentrations of nitrate and sulphate to observe the system performance. Experimental results show that the best sensor is the star array planar electromagnetic sensor. Artificial Neural Networks (ANN) is used to classify different levels of nitrate and sulphate contaminations in water sources. The impedance of star array planar electromagnetic sensors was derived to decompose by Wavelet Transform (WT). Classification of WT has been applied to extract output signal features. These features are fed into ANN to classify different nitrate and sulphate concentration levels in water. The model is capable of distinguishing contaminants concentration level in the presence of other types of contaminants with a Root Mean Square Error (RMSE) of 0.0132 with 98.68% accuracy.