Environmental Pollution Prediction of NOx by Process Analysis and Predictive Modelling in Natural Gas Turbine Power Plants

  • 2021-01-18 17:41:21
  • Alan Rezazadeh
  • 0

Abstract

The main objective of this paper is to propose K-Nearest-Neighbor (KNN)algorithm for predicting NOx emissions from natural gas electrical generationturbines. The process of producing electricity is dynamic and rapidly changingdue to many factors such as weather and electrical grid requirements. Gasturbine equipment are also a dynamic part of the electricity generation sincethe equipment characteristics and thermodynamics behavior change as theturbines age. Regular maintenance of turbines are also another dynamic part ofthe electrical generation process, affecting the performance of equipment. Thisanalysis discovered using KNN, trained on relatively small dataset produces themost accurate prediction rates. This statement can be logically explained asKNN finds the K nearest neighbor to the current input parameters and estimatesa rated average of historically similar observations as prediction. This paper incorporates ambient weather conditions, electrical output as wellas turbine performance factors to build a machine learning model to predict NOxemissions. The model can be used to optimize the operational processes forreduction in harmful emissions and increasing overall operational efficiency.Latent algorithms such as Principle Component Algorithms (PCA) have been usedfor monitoring the equipment performance behavior change which deeplyinfluences process paraments and consequently determines NOx emissions. Typicalstatistical methods of machine learning performance evaluations such asmultivariate analysis, clustering and residual analysis have been usedthroughout the paper.

 

Quick Read (beta)

loading the full paper ...