Prediction of gaseous emissions from a chain grate stoker boiler using neural networks of ARX structure

Abstract
The authors present the application of feedforward multi-layered perceptron networks as a simplistic means to model the gaseous emissions emanating from the combustion of lump coal on a chain-grate stoker-fired boiler. The resultant ‘black-box’ models of the oxygen concentration, nitrogen oxides and carbon monoxide in the exhaust flue gas were able to represent the dynamics of the process and delivered accurate one-step-ahead predictions over a wide range of unseen data. This system identification approach is an alternative to the mathematical modelling of the physical process, which although lacking in model transparency and elegance, is able to produce accurate one-step- ahead predictions of the derivatives of combustion. This has been demonstrated not only with data sets that were obtained from the same series of experiments (which also demonstrated the repeatability of the model) but also for data with a temporal separation of almost eight months from the training data set.