Show simple item record

AdvisorFadlalla, Adam
AuthorShaheen, Mohammad Ahmad
Available date2019-10-14T09:32:19Z
Publication Date2019-06
URIhttp://hdl.handle.net/10576/12119
AbstractNatural gas industry becomes one of the largest sectors in the today’s world economy. With the growth of production, pollutants from gas processing plants continue to grow, forcing government and legislative authorities to compel stringent restrictions on the release of emissions. These restrictions become among top vital factors influencing plant operations nowadays in the state of Qatar. One of the most important tasks for gas processing plants is not only to ensure a continuous gas supply to customers, but also to monitor, control and address concerns about the environmental impact of the operations. The use of the gas fired steam boilers in the gas operations has led to a significant increase in the emissions of toxic substances such as NOx. Therefore, significant worries have been raised about the environmental effect on climate and air quality of noxious emissions associated with the steam generation process of the boilers. Data mining is a powerful tool that has been used for decades for advanced process analytics of large quantities of plant data in order to extract useful information and to reach a better understanding of the process. In this research, some of data mining techniques such as artificial neural networks and decision trees are applied on real plant data representing 27 industrial process parameters from a gas fired steam boiler of one gas processing plant at Ras-Laffan Industrial City in order to predict the most important factors impacting the NOx formation in the combustion process of the boiler. Closer attention to those factors can be given promptly in order to enhance the plant environmental performance. The results obtained by the artificial neural network and decision tree models showed that the NOx emissions are directly related to the air flow parameters such as O2 concentration, the excess air level in the boiler and the amount of the flue gas at the boiler outlet. It was also shown that the company is following a traditional technique of lowering the amount of oxygen in the boiler aiming to reduce the NOx emission levels. However, this technique was proven to be limited under certain threshold of boiler load due to other important factors such as the adiabatic flame temperature (AFT) which induces the formation of NOx emissions.
Languageen
SubjectData Mining
SubjectArtificial Neural Network
SubjectDecision Tree
SubjectModeling
SubjectSteam Boiler
SubjectNOx Emissions
TitleUsing Data Mining Techniques to Predict NOx Emissions Levels in Gas Fired Boilers
TypeProfessional Masters Project
DepartmentBusiness Administration


Files in this item

Icon

This item appears in the following Collection(s)

Show simple item record