Estimating thermal boundary conditions of boiler membrane water-wall using decentralized fuzzy inference with sensitivity weighting

• Fuzzy inference with sensitivity weighting is used to solve an inverse problem.
• Boundary conditions of membrane water-wall are estimated using this method.
• This method performs better than the conjugate gradient method.

The thermal boundary conditions including the fireside heat flux, temperature of water–steam mixture in water-wall tubes and heat transfer coefficient on the inner surface of membrane water-wall, are estimated at the same time using the decentralized fuzzy inference (DFI) method with the sensitivity weighting based on the local measured temperature on the back of membrane water-wall. The influences of different initial guessed thermal boundary conditions, the measurement points, and the temperature measurement errors on the estimated results are discussed, and comparisons with the conjugate gradient method (CGM) are conducted. The study shows that the thermal boundary conditions of membrane water wall, the temperature distribution on the fireside and the positions of the dangerous points can be determined more accurately by the proposed method. It is concluded that DFI method with the sensitivity weighting is of better anti ill-posed characteristic. It would provide necessary basis for analyzing and monitoring the operation state of power plant boilers.