A neural network based method for on-line calculation of equivalent stresses in steam turbine rotors

Abstract

An on-line calculation method, based on a neural network, for calculating the equivalent stress of steam turbine rotors, is being proposed. The method of expressing the equivalent stress in an steam turbine rotors, in terms of formulas applicable for solid cylindrical models of the same diameter, is being explained; together with the numerical finite element calculation method. A M ethod and calculation steps of on-line calculation by neural network are presented, accompanied by a practical example of calculating the equivalent stress in the high and intermediate pressure rotor (HP-IP rotor) of a 600MW steam turbine. The application of neural network technique, for obtaining the equivalent stress in on-line monitoring systems of steam turbine rotors, raises the calculation precision of the equivalent stress of steam turbine rotors; providing herewith a better basis for on-line monitoring the equivalent stress and life consumption management of steam turbine rotor's.