A DSP based torque meter for induction motors

This work describes the steps to implement a torque meter for three phases induction motors, based on stator voltage and machine current measurement. The strategy is based on stator flux synthesis through Programmable Cascaded Low-Pass Filters (PCLPF). The electromagnetic torque estimation is processed by a DSP microprocessor in real time. The PCLPF filter outlines the problem of necessary numeric integration to calculate the stator flux starting from the samples of stator voltage and current. The Programmable Cascaded Low-Pass Filter is implemented using recurrent neural network (RNN-PCLPF) trained by an algorithm based on Kalman filter. The DSP based implementation of a torque meter results in an equipment with the same precision when comparing with torque meters based on torsion of metallic axes, with known elastic constant and strain gauges.