Identification aspect of methodology design of control system time-variant process

Specificity of a food manufacture demands perfection of automatic control systems of processes in devices, units and installations. Creation of an adaptive control system by technological process of a food on the basis of model of control object it is necessary to carry out the additional analysis for choice algorithm of identification on real enough to representative sample of input data and output signal/data. In article on the basis of simulation it is analyzed over 53 algorithms of recurrent identification plus the basic modifications of these algorithms by 47 criteria for time-varying multivariable linear dynamic objects. On the basis of this analysis for engineering practice for a considered class of objects some algorithms are recommended. Possibilities of the software suite having for today the fullest set of parametrical identification algorithms are discussed. For given specific conditions of comparison in the package identification algorithms for identification of stationary coefficients in the equation object of the most effective were: Yzerman-1, Kaczmarz, Nagumo-Noda, Rastrigin, Kalman filter, the forgetting factor, Zipkin. When pointwise object - Kaczmarz, Nagumo-Noda, Kalman filter; showed the best result identification algorithm-Nagumo Noda.