Design of experiments for fitting regression models on the tubular SOFC CHP100kWe: Screening test, response surface analysis and optimization

The SOFC CHP 100kWe Field Unit built by Siemens is operating at the Gas Turbine Technologies in Torino. The generator started up on June 2005, and is delivering around 100kWe of AC electrical energy to the grid and 55kWt of thermal energy, providing heating and cooling for the facility's buildings.An experimental plan has been designed by applying the factorial analysis. Before performing the experiments on the plant, this approach has been applied to a computer model simulation: a 2323 factorial analysis has been applied to the generator model (independent variables: fuel utilization factor, air utilization factor and air temperature at the generator inlet). First-order regression models are obtained for the investigated dependent variables (i.e. DC power, recovered heat, combustion zone temperature, etc.). The data are analyzed by applying the response surface methodology. Finally, constrained optimization methods are applied to investigate optimal operation points in terms of DC generated power or exhausts’ recovered heat. A comparison with the actual operation point of the SOFC generator is shown.