Optimizing the Controlling Parameters of a Biomass Boiler Based on Big Data

Abstract

This paper presents a comprehensive method for optimizing the controlling parameters of a biomass boiler. The historical data are preprocessed and classified into different conditions with the k-means clustering algorithm. The first-order derivative (FOD) method is used to compensate for the lag of controlling parameters, the backpropagation (BP) neural network is used to map the controlling parameters with the boiler efficiency and unit load, and the ant colony optimization (ACO) algorithm is used to search the opening of air dampers. The results of the FOD-BP-ACO model show an improvement in the boiler efficiency compared to the predicted values of FOD-BP and the data compared to the historical true values were observed. The results suggest that this FOD-BP-ACO method can also be used to search and optimize other controlling parameters.

Keywords

• biomass
• boiler
• big data
• machine learning