Permanent Magnet DC Motor Friction Measurement and Analysis of Friction's Impact

摘要

Accurate modeling of permanent magnet (PM) DC motors is a prerequisite for expeditious feedback design in many practical implementations such as electric drive powered mobility vehicles: scooters, wheelchairs, mobility vans, etc. The existing theoretical model, which is often used in the literature and by researchers, does not take into account all the existing frictions. Besides, the friction values are not generally known in advance. So, in order to accurately model a DC motor it is necessary either to measure the frictions values when the classical modeling approach can be used, or apply more sophisticated modeling techniques that are based on system identification. In this paper, the values of unknown static friction, Coulomb friction and viscous friction coefficient were obtained by measuring, using direct armature current control. For the viscous friction coefficient the temperature dependence and thus time dependence has been verified. In order to have an accurate model of a PM DC motor, the recursive least squares (RLS) method is used. Speed of a discrete Simulink model whose parameters were obtained using RLS identification is compared with measured system speed response. The responses are also compared with simulation velocity output when all frictions are neglected and with the simulation output attained using the classical approach when only viscous friction is taken into account. The results confirm the quality of the RLS modeling approach and show that for accurate modeling of a PM DC motor using the classical approach (when this approach can be used), all frictions should be considered. Experiments have been performed on a PM DC motor with a real-time data acquisition NI DAQ 6009 card.