The problems in modeling of consequences of thermocouples electrodes degradation processes are considered in this paper. It is showed that application of the least squares method (LSM) leads to bias in estimated model coefficients a when free member is equal to zero. Estimations of model coefficients made by LSM when a free member is not equal to zero have all properties of traditional LSM-estimations but haven’t got any physical sense. The new method which has a physical sense is proposed in this paper. Its estimations of the coefficients are unbiased. Method is called the conditional least squares method.
the least squares method, LSM-estimations of model coefficients, thermocouples, error of drift, acquired during operation time inhomogeneity of thermoelectrodes
"Metod umovnykh naimenshykh kvadrativ dlia pobudovy modeli pokhybky termopar" [The conditional least squares method for thermocouples error modelling],
Information Processing Systems,