Intelligent classification of ammonia concentration based on odor profile

• Main Author: Muhamad Faruqi, Zahari
• Format: Thesis
• Language: English
• Published: 2016
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://umpir.ump.edu.my/id/eprint/17671/16/Intelligent%20classification%20of%20ammonia%20concentration%20based%20on%20odor%20profile.pdf; http://umpir.ump.edu.my/id/eprint/17671/
• Institution: Universiti Malaysia Pahang

This thesis presents the intelligent classification of ammonia concentration based on the standard of oil and gas industries wastewater discharge. The intelligent classification using signal processing is a well-known technique in many applications and as well in the oil and gas industry. The intelligent classification technique for ammonia concentration classification is a demanding technique especially in the environmental sector. Ammonia solution properties and ammonia solution preparations were studied in this thesis which commonly used in industry. The objectives of this thesis are to develop an intelligence classification of ammonia concentration based on the oil and gas industry wastewater discharge schedule and to analyze performance of the intelligent classification of ammonia concentration based on the oil and gas industry wastewater discharge schedule. In this thesis the ammonia odor profile has been pre-identified by chemist using four sensor array. The ammonia concentration was validated using a commercialized gas sensor and spectrophotometer to cross-validated e-nose instrument. The odor profile from two different samples; high (20 ppm and 25 ppm) and low (5 ppm, 10 ppm and 1 5ppm) concentration that have been normalized and visualized in a 2D plot to extract the unique patterns. The variance of the low and high concentration of ammonia odor profile has been identified as different group samples. This group samples have been analyzed statistically using Boxplot, calibration curve and proximity matrix, The thesis describes the statistical techniques to visualize the pattern and using mean features to classify between the low and high concentration. Two intelligent classification techniques have been used which are Artificial Neural Network (ANN) using the back-propagation approaches and then, the result of ANN model was cross-validated.using CBR. Both ANN model and CBR classifier have been measured using several performance measures. From the results, it is observed that ANN model and CBR classifier are capable of classifying 100% of ammonia concentration odor profile from the water. The results can also significantly reduce the cost and time, and improve product reliability and customer confidence.