Using Multi Dimensional Grid to identify a Permanent Magnet Synchronous Machine Model

This paper is the second presented in this conference dealing with permanent magnet synchronous machine model identification. Main purpose of the research introduced in this paper, besides to introduce a new parameter estimation method, was to investigate the cost function properties. Understanding shape and behaviour of the cost function helps the authors to select the appropriate identification technique for identifying a permanent magnet synchronous machine model. Understanding the cost function properties is also useful to know how these identification algorithms can be improved and adopted to this special problem. In this paper, a brute force technique is used which evaluates the cost function of the model parameter identification using a four dimensional grid of parameters and then, depending on the location of the cost function minimal value, adjusting the grid resolution and moving the center position of the grid in every iteration step.