Mechanical parameter identification of two

Alternating current (AC) motor drive systems are widely used and their performances are greatly affected by motor parameters and load disturbances, so it is necessary to identify and observe their moment of inertia, the viscous friction coefficient and load torque. This paper presents an identification method combined a Luenberger observer and the variable forgetting factor recursive least squares method (FFRLSM), which does not require the torque sensor but a position encoder and can identify the moment of inertia of the drive system, the viscous friction coefficient, and the load torque, simultaneously. The Luenberger observer takes the position and the actual current to estimate speed and angular position. This information, together with variable FFRLSM is used to identify the moment of inertia, viscous friction coefficient and load torque. The identified moment of inertia of the system is substituted into the Luenberger observer model to correct the error of the system model. According to a probability density function of a normal distribution, the proposed method estimates the identification target in a short time, with stable and high identification precision. Finally, a comparison is made between the proposed algorithm and the FFRLSM. Results show that the proposed identification method achieves intended purpose and promise application values with its short consuming-time and high identification accuracy