Modelling of a post-combustion CO<inf>2</inf> capture process using extreme learning machine

This paper presents modelling of a post-combustion CO2capture process using bootstrap aggregated extreme learning machine. Extreme learning machine (ELM) randomly assigns the weights between input and hidden layers and obtains the weights between the hidden layer and output layer using regression type approach in one step. This paper proposes using principal component regression to obtain the weights between the hidden and output layers. Due to the weights between input and hidden layers are randomly assigned, ELM could have variations in performance. This paper proposes combining multiple ELMs to enhance model prediction accuracy and reliability. To predict the CO2 production rate and CO2 capture level, seven parameters in the process were regarded as input variables: inlet gas flow rate, CO2 concentration in inlet flow gas, inlet gas temperature, inlet gas pressure, lean solvent flowrate, lean solvent temperature, lean loading and reboiler duty. The bootstrap re-sampling of training data was applied for each single ELM and then the individual ELMs are stacked, thereby enhancing the model accuracy and reliability. The bootstrap aggregated extreme learning machine (BA-ELM) can provide fast learning speed and good generalization performance, which will be used to optimize the CO2 capture process.