Inertia identification method based on the product of the integral of torque reference input and motor speed

This paper presents a novel approach to the inertia identification problem of mechatronic servo systems with limited strokes. In the proposed approach, the product of the integral of torque reference input and motor speed of mechatronic servo system in periodic motion is calculated. Then its time average over one period is used to calculate inertia. The proposed approach can identify inertia with infinitesimal motions within short time interval, and suits online and off-line identification of mechatronic servo systems with limited strokes, and off-line inertia identification of industrial machineries with interfering degrees of freedom especially in the preparatory period after they are turned on and before their normal operation starts. The proposed approach utilizes an integral of torque reference input, and can identify inertia when torque reference input contains large amount of noises. Simulations and experiments on an electric motor and a linear slider suggest the feasibility of the proposed approach on single-degree-of-freedom (SDOF) systems and two-degree-of-freedom (2DOF) systems with limited strokes.