This paper presents an analysis of the corrosion hazard in the burner belt area of waterwalls in pulverised fuel (PF) boilers that results from low-NOx combustion. Temperature distributions along the waterwall tubes in subcritical (denoted as SUB) and supercritical (SUP) boilers were calculated and compared. Two hypothetical distributions of CO concentrations were assumed in the near-wall layer of the flue gas in the boiler furnace, and the kinetics of the waterwall corrosion were analysed as a function of the local temperature of the tubes. The predicted rate of corrosion of the boiler furnace waterwalls in the supercritical boilers was compared with that of in the subcritical boilers.

In this paper a mathematical model enabling the analysis of the heat-flow phenomena occurring in the waterwalls of the combustion chambers of the boilers for supercritical parameters is proposed. It is a one-dimensional model with distributed parameters based on the solution of equations describing the conservation laws of mass, momentum, and energy. The purpose of the numerical calculations is to determine the distributions of the fluid enthalpy and the temperature of the waterwall pipes. This temperature should not exceed the calculation temperature for particular category of steel. The derived differential equations are solved using two methods: with the use of the implicit difference scheme, in which the mesh with regular nodes was applied, and using the Runge-Kutta method. The temperature distribution of the waterwall pipes is determined using the CFD. All thermophysical properties of the fluid and waterwall pipes are computed in real-time. The time-spatial heat transfer coefficient distribution is also computed in the on-line mode. The heat calculations for the combustion chamber are carried out with the use of the zone method, thus the thermal load distribution of the waterwalls is known. The time needed for the computations is of great importance when taking into consideration calculations carried out in the on-line mode. A correctly solved one-dimensional model ensures the appropriately short computational time.