The thermal conductivity of refractory bricks with a high thermal conductivity was measured by the hot wire method. So far a sufficiently large specimen was used, a linear relation was obtained between the heat source temperature (θ) and the logarithm of time (logt) for high conductive materials such as a dense magnesia brick. The thermal conductivity values which were obtained from the linear relation were accurate within ±4% from the mean value.
However, the length of the linear portion in θ-logt plot decreased with decreasing specimen size, and it gave a large error in the result. That is, it became more than ±17% when a cylindrical specimen of a diameter about 70mm was used.
By adjusting the nonlinear portion in θ-logt at the beginning of the measurement with the time t0 which was obtained by ∂t/∂θ-t plot, the length of the linear portion could be made longer and the error was decreased below the permitted limit in the practical measurement.
With application of this adjustment, thermal conductivities of magnesia, carbon and silicon carbide bricks were measured up to 1200°C. The thermal conductivities of these bricks decreased as the temperature rises.