Control of particulate matter (PM) emissions from industrial plant using ANFIS based controller

• Main Authors: Sambo, Aliyu Umar, Babawuro, Adamu Y., Salami, Momoh Jimoh Eyiomika, Khan, Md. Raisuddin
• Format: Article
• Language: English; English
• Published: Asian Research Publishing Network (ARPN) 2016
• Subjects: QD Chemistry; T Technology (General); TD Environmental technology. Sanitary engineering
• Online Access: http://irep.iium.edu.my/49347/1/49347_Control%20of%20particulate%20matter.pdf; http://irep.iium.edu.my/49347/2/49347_Control%20of%20particulate%20matter_SCOPUS.pdf; http://irep.iium.edu.my/49347/; http://www.arpnjournals.org/jeas/research_papers/rp_2016/jeas_0116_3352.pdf

In recent times, the negative effect of air pollution such as particulate matter (PM) emitted from industrial plants has compelled researchers in finding efficient control system to control such pollutants in order to keep the environment safe. The aim of this study is to develop a reliable method of controlling the emissions of PM using wet scrubber system as a control device. The process of a wet scrubber is nonlinear in nature. Due to difficulty in selecting optimum scrubbing liquid droplet size in wet scrubbing process, the system becomes complex. Thus, Adaptive Neuro Fuzzy Inference System (ANFIS) based control technique is employed in this paper to handle the nonlinearities. ANFIS control technique has the advantage to integrate fuzzy logic systems and learning ability of neural network, thus able to handle nonlinear systems better. The controller is developed using data of PM emission from cement kiln. The system is simulated using triangular and trapezoidal membership function (MF) with 2 and 3 input MF in each case. The performance of the controller is evaluated based on settling time. The results indicated that the developed controller was able to maintain the PM emission below a set point of 20µg/m3 which is the maximum allowable PM emission limit recommended by world health organization (WHO). The controller with 2 input triangular membership functions indicated a better performance with a settling time of 5.2 seconds.