263 © Institution of Engineers Australia, 2012 * Paper E11-922 submitted 5/03/11; accepted for publication after review and revision 25/07/11. † Corresponding author Boyang Hu can be contacted at boyang. hu@sydney.edu.au. A novel sensorless algorithm of the trapezoidal back-electromagnetic force brushless DC motors from near-zero to high speeds * B Hu † , J Lee, S Sathiakumar and Y Shrivastava School of Electrical and Information Engineering, The University of Sydney, New South Wales ABSTRACT: This paper proposes a novel sensorless drive method for the brushless DC (BLDC) motors. It provides an accurate and robust position and speed estimation from the near-zero speed to the rated speed. The proposed sensorless algorithm uses a trapezoidal waveform of the back- electromotive force (EMF) signal with speed information. The back-EMF signals are estimated based on the three-phase current and terminal voltage signals. The validity of the proposed method is veriied through both simulation and experimental results. KEYWORDS: BLDC motor; sensorless; near-zero speed. REFERENCE: Hu, B., Lee, J., Sathiakumar, S. & Shrivastava, Y. 2012, “A novel sensorless algorithm of the trapezoidal back-electromagnetic force brushless DC motors from near- zero to high speeds”, Australian Journal of Electrical & Electronics Engineering, Vol. 9, No. 3, pp. 263-274, http://dx.doi.org/10.7158/E11-922.2012.9.3. 1 INTRODUCTION Permanent magnet brushless DC (BLDC) motors have been widely used in motion control, consumer products and automotive applications due to high eficiency, high power density and high torque-to- current ratio, as discussed in pervious literature by Jahns & Soong (1996), Holtz & Springob (1996), Jang & Kim (2005), Shao et al (2003) and Consoli et al (1994). A BLDC motor is typically driven by an electronic controller that switches the DC power supply voltage between the stator windings while the rotor is rotating. To properly energise the BLDC motor, the position information should be known accurately. One electrical period is 360°, which is conventionally divided into six sectors of 60° each. Based on the position information, it can be determined which 60° sector the rotor is in and thus activate the correct pulsed signal pair. A BLDC motor typically requires a rotor position sensor to obtain the rotor position information. However, the position sensor has several drawbacks, such as machine size, cost and reliability. Sensorless BLDC motor control has been a motivated ield in achieving small size, low cost and high reliability applications. During the last two decades, in order to eliminate the unreliable and costly position sensors, different methods for BLDC motor sensorless control have been suggested. Back-electromotive force (EMF) based sensorless methods were categorised into the following four groups by Johnson et al (1999): (i) back-EMF sensing method; (ii) third harmonic back-EMF sensing method; (iii) back-EMF integration method; and (iv) adaptive sliding mode observer. Back-EMF sensing methods are the most popular, and were proposed by Iizuka et al (1985) and Damodharan & Vasudevan (2010). In each conduction interval, only two out of three phases conduct current and the other phase floats without current. The zero-crossing points of back-EMF can be directly measured from the loating phase. These methods work reasonably well at medium and high speeds. However, their performance deteriorates drastically at low speeds due to the back-EMF signal being too small to be detected. As the rotor position detection error consists of the error deviation and group delay, Australian Journal of Electrical & Electronics Engineering, Vol 9 No 3