The prediction of the oxygen content of the flue gas in a gas-fired boiler system using neural networks and random forest

<p><span lang="EN-US">The oxygen content of the gas-fired boiler flue gas is used to monitor combustion efficiency. Conventionally, this measured using an sensor. However, because it operates in extreme conditions, sensor tends have disadvantage high maintenance costs. In addition, absence other sensors as element redundancy and when there damage causes manual handling by workers. It dangerous for these workers, considering environmental conditions with high-risk hazards. We propose artificial neural network (ANN) random forest-based soft predict overcome problems. The prediction made utilizing data on power plant’s boiler, consisting 19 process variables from a distributed control system. research has proved that proposed successfully predicts content. Research forest shows better performance results than ANN. errors are mean absolute error (MAE) 0.0486, squared (MSE) 0.0052, root-mean-square (RMSE) 0.0718, Std Error 0.0719. While ANN MAE 0.0715, MSE 0.0087, RMSE 0.0935, 0.0935.</span></p>