Stability Analysis of the Natural Field Orientation Controlled Induction Machine Drive G. Mirzaeva † and A. Rojas ∗ School of Electrical Engineering and Computer Science The University of Newcastle, Callaghan, 2308 NSW, Australia Phone: +61 (2) 49215963 † ; +61 (2) 49216023 ∗ Email: Galina.Mirzaeva@newcastle.edu.au † ; Alejandro.Rojas@newcastle.edu.au ∗ KEYWORDS Vector control, Variable speed drive, Control of drive, Induction motor. ABSTRACT Natural Field Orientation (NFO) Control proposed in the 1980’s is an alternative low cost control method for induction machines. The NFO Control has shown to have some stability issues in regeneration which have been partially addressed in the previous papers. In the present paper a linearised stability analysis is presented for the motor/drive system based on both the non-augmented and the augmented versions of the NFO control. Based on the results of this analysis, the limitations of the augmentation strategy are explored and conditions are identified when it does not ensure stability. Theoretical results are confirmed by simulations and experiments. Directions of further improvement of the NFO stability are outlined. I NTRODUCTION Natural Field Orientation (NFO) is a patented control strategy proposed in 1980’s ( [1], [2]). It belongs to the Stator Flux Orientation (SFO) control family and is primarily distinguished by that the magnitude of the stator flux vector is not obtained by integration but is assumed to be equal to its reference value. Consider the stator voltage equation in the stator flux oriented reference frame (denoted by the index “ψ s ”): e sψs = u sψs − R s i sψs = d|ψ s | dt + jω ms |ψ s | (1) where |ψ s | is the magnitude of the stator flux vector ψ s ; R s is stator resistance; ω ms is angular velocity of the stator flux vector ψ s . Splitting equation (1) into real and imaginary parts yields: e sx = u sx − R s i sx = d|ψ s | dt = L m d|i ms | dt (2) e sy = u sy − R s i sy = ω ms |ψ s | = ω ms L m |i ms | (3) where x-axis is aligned with and y-axis is in quadrature to the vector ψ s ; L m is magnetising inductance; |i ms | is the magnitude of the stator magnetising current. The magnitude of the stator flux vector can be obtained from (2) by integrating the estimated value of e sx . The angular velocity of the stator flux vector can then be found from (3). The NFO control technique suggests to exclude the flux integrator and to assume that |ψ s | equals to its reference value |ψ ∗ s |. This assumption results in elimination of the problems commonly associated with the flux integrator (see, for example, [3]), and in acquisition of some new properties as compared to the traditional SFO control (hence a separate name for the NFO control technique). The property of “natural” frame orientation was explained in [5]. The true position of the stator flux vector and the asso- ciated (x, y) reference frame is not known to the algorithm. It uses the estimated stator flux angular position and associates with it a reference frame that we will call (d, q), or the control frame. If the control frame deviates from the true frame than an inplicit mechanism “naturally” present in the NFO algorithm would act to compensate for such misalignment. It was also shown in [5] that for the original version of the NFO algorithm the natural self-alignment mechanism is limited to the motoring mode of operation only. Under certain conditions in regeneration, in the case of a small initial misalignment, the control frame would drift further away from the true frame and the stability of the algorithm would be lost. An augmentation to the original algorithm was proposed in [5] and [6] that overcomes this limitation (to a certain extent) and ensures stable operation (in the steady state sense only) in regeneration with limited torque. Another desirable property of the NFO algorithm is its high tolerance of the parameter errors. The only two machine parameters that it needs to know are the stator resistance R s and the magnetising inductance L m . It was shown in [4] that the original and, especially, the augmented NFO versions are very little sensitive to the accurate knowledge of both parameters. The NFO algorithm compensates for the parameter errors by introducing a small angular misalignment between the control and the true reference frames, which is usually very small and does not compromise the stable operation. The purpose of this paper is to present a more compre- hensive stability analysis of the NFO-controlled induction machine including the dynamic effects. Moreover, the analysis will include not only the frame angular error but the close-loop 1155 978-1-4244-1742-1/08/$25.00 c  2008 IEEE Authorized licensed use limited to: University of Newcastle. Downloaded on November 17, 2008 at 20:06 from IEEE Xplore. Restrictions apply.