Online corrosion probes enable the online determination of high-temperature corrosion rates in biomass combustion plants. These probes are exposed to the flue gas and require a layer of corrosion products and deposits on their surface to be able to perform the corrosion measurements. The formation of this layer takes time and depends on the build-up rate of the deposits as well as on the ability of the deposits to form a conductive layer. During this initial phase, which lasts about 300 h based on experiences with forest wood chip combustion, the measurements are inaccurate. This inaccuracy can be neglected during long-term measurements performed at real-scale plants which last more than several 1000 h. Since high-temperature corrosion in biomass combustion plants is known to generally follow a paralinear trend, high corrosion rates prevail in the beginning. Following, if short-term measurements shall reflect or come close to real-life conditions, this "start-up" effect cannot be neglected. Consequently, to be able to correct this "start-up" effect, a methodology has been developed which combines results from parallel measurements of an online corrosion probe and a mass loss probe. The mass loss probe consists of multiple test rings and is exposed to the flue gas next to the corrosion probe with similar surface temperatures. With the mass loss probe the trend of the corrosion rate within the first 300 h is determined by gravimetric mass loss measurements of each ring. For evaluation this trend of the corrosion rate from the mass loss probe is coupled with the signal from the corrosion probe. Hereby, the inaccuracy at the beginning of the online corrosion probe measurements can be eliminated and reliable short-term measurements can be performed. A 50 kW biomass grate combustion system coupled with a drop tube was used for the development of the mass loss probe and the evaluation methodology. Such a test rig provides multiple advantages for high-temperature corrosion measurements compared to a real-scale plant. Specific and homogenous fuels can be tested, specific influencing parameters can be adjusted and varying operating conditions occurring in a real-scale plant are excluded. (C) 2015 Elsevier B.V. All rights reserved.