Tests on large circulating fluidized bed boilers are extremely expensive and computer modeling is not yet reliable. The paper tries to answer the question: Can tests on laboratory-scale equipment be scaled to satisfy the need for information? Criteria for combustion scaling are defined in the form of two Damkohler numbers: one for processes in the vertical direction and one for the horizontal direction. Examples are given to show the outcome of combustion modeling. It is concluded that it is possible to reasonably well model the combustion process in the vertical direction by providing sufficient residence time in the riser, but in a laboratory device it is not possible to model the horizontal combustion processes, except in special cases.