Parameter Estimation for IPMSM Using Coupled Adaline Neural Networks without Signal Injection

In this paper, an innovative method for the parameter estimation of internal-mounted permanent magnet synchronous motors (IPMSMs) is proposed. Five coupled Adaline Neural Networks (AdNNs) are adopted to estimate the parameters using data collected during two different steady states without signal injection or dedicated tests. The distortion of the stator voltages caused by the inverter nonlinearity is compensated to improve the estimation accuracy. A detailed theoretical analysis of the convergence properties of the proposed algorithm is provided to design the learning rates of the AdNNs. Data obtained by means of numerical simulations of an IPMSM drive are exploited to validate the proposed method.