WO2011126462A1

WO2011126462A1 - Speed and rotor position estimation of electrical machines using rotor slot harmonics or higher order rotor slot harmonics

Info

Publication number: WO2011126462A1
Application number: PCT/TR2010/000070
Authority: WO
Inventors: Hulusi Bulent Ertan; Ozan Keysan
Original assignee: Hulusi Bulent Ertan; Ozan Keysan
Priority date: 2010-04-05
Filing date: 2010-04-05
Publication date: 2011-10-13
Also published as: EP2556381A1; EP2556381B1

Abstract

The present invention is a method and a system developed for speed and rotor position estimation of electrical machines. In this invention, rotor slot harmonics are detected either from the motor current or voltage or voltage induced on an external search coil or by similar means. The excitation frequency component of the signal used is eliminated; rotor slot harmonics are extracted from the resulting signal via demodulation. Rotor speed and position is obtained by processing the rotor slot harmonic variation in real time; and this process does not employ any spectral analysis technique.

Description

SPEED AND ROTOR POSITION ESTIMATION OF ELECTRICAL MACHINES USING ROTOR SLOT HARMONICS OR HIGHER ORDER ROTOR SLOT HARMONICS

Related Field of the Invention:

The present invention is a method and a system developed for speed and rotor position estimation of electrical machines that comprises rotor speed and position detection by processing higher order rotor slot harmonics in real time that does not need any spectral analysis technique or sensor in some variations of the application.

Background of the Invention (Prior Art):

The speed and instantaneous rotor position determination is essential in electrical machine control. The speed and position is typically measured by using speed sensors. However, a speed sensor introduces extra cost on the system and reduces the system reliability. Thus, measuring rotor speed without an extra sensor is essential on induction motor control. In the following rotor position detection without a direct position component (sensorless) voltage, current and flux measurement sensors may be utilized.

One common approach to sensorless rotor speed and position estimation is to employ a field orientation technique which is generally called in the literature as "vector control". In sensorless vector control, rotor speed is estimated by using voltage or current model of the induction motor. Briefly, rotor flux vector is obtained indirectly by integrating the induced voltage which can be measured by sense coils or phase voltage. However, the accuracy of rotor speed estimation is mainly dependent on the accuracy of integration and usually fails at low speeds.

An alternative way to improve the measurement accuracy at low speed is achieved by injecting a high frequency carrier signal with frequency which is generally higher and independent of the fundamental excitation frequency. For example, Jansen et al., U.S. Pat. Nos. 5,565,752 and 5,585,709 injects a carrier signal to improve rotor speed estimation. On the other hand, these methods either need a modification in the applied pulse-width modulation signals or more generally need an inclusive redesign of the inverter circuitry hardware. Another disadvantage of signal injection has been to generate cogging torques especially at small sized machines. A different way to measure rotor speed can be achieved by using rotor slot harmonics. The variation of the air gap reluctance due to rotor slot openings modulates the fundamental air-gap flux. This modulation can be observed either from the flux of the machine via a sensing coil, from phase voltages of a current controlled machine or from phase currents of a voltage controlled machines. The harmonics caused by rotor slots are generally called as rotor slot harmonics (RSH) in the literature.

RSH are relatively small in magnitude compared to the magnitude of fundamental excitation and its harmonics. The frequency and the magnitude of the RSH are altered by mechanical faults such as broken rotor bars, eccentric rotor or a worn bearing which is described briefly in Negrea M. D., "Electromagnetic flux monitoring for detecting faults in electrical machines", Helsinki University of Technology, Ph.D. thesis, 2006. Thus, RSH are generally utilized for condition monitoring of induction machines for example, in commonly- assigned patents of Kliman et al., U.S. Patent No. 4,761,703 and Casada U.S. Patent No. 5,519,337.

The frequency of the RSH is directly related with the number of rotor slots and the rotor speed. Thus, knowing the number of rotor slots and the frequency of RSH, the rotor speed can be estimated.

It is observed that previous studies always use spectral estimation techniques to determine the frequency of RSH, for example in U.S. Pat No. 5,530,343 and 6,208,132. The spectral estimation technique is mainly based on Fast Fourier Algorithm (FFT). For spectral estimation techniques long sampling periods and complex calculations are inevitable. The sampling and calculation period is typically range of a few seconds to tens of seconds. These estimation periods are suitable for offline condition monitoring methods. But the spectral estimation techniques become useless for instantaneous speed estimation since they take very long time. For cases where speed transients occur, these algorithms cannot estimate the rotor speed due to long data sampling periods.

There are also some other techniques aiming to reduce the sampling periods and calculation time such as U.S. Pat No. 6,789,028, but they are still classified as a variation of spectral analysis method. These approaches obtain the frequency of rotor slot harmonics explicitly then, calculate the rotor speed from this data. Since all mentioned studies use fundamental rotor slot harmonics the spectral analysis is almost inevitable because they exist in a spectrum band where other irrelevant harmonics such as, line harmonics, inverter harmonics or other electromagnetic noise harmonics present. Some patents such as; Jansen et al., U.S. Patent No. 6,388,420 and Takasi et al. DE Patent No. 102006008048 use phase locked loops or adaptive filtering but they still use the fundamental rotor slot harmonics and suffer from the listed problems above.

The invention here approaches the problem in a different way. In this case, instantaneous current and voltage data is sampled and higher order rotor slot harmonics are isolated. They exist in higher frequencies than the irrelevant harmonics which in this case they can be eliminated with some basic filtering blocks and the desired harmonics can be extracted. Then, without the need of determining the frequency of rotor slot harmonics explicitly, by only using the extracted signal as an output of an incremental encoder the rotor position and thus rotor speed can be determined.

K.D. Hurst, et. al. in "A Comparison of Spectrum Estimation Techniques or Sensorless Speed Detection in Induction Machines", (IEE Trans. On Industry Applications, vol. 33, No. 4, 1997, pp 898-905") compares different rotor speed estimation techniques, using rotor slot harmonics. But all the techniques listed use fundamental (dominant) rotor slot harmonics and most of them utilize some kind of spectral estimation technique, which increases complexity and cost of the system.

- Nandi S, et al. in "Detection of Rotor Slot and Other Eccentricity Related Harmonics in a Three Phase Induction Motor with Different Rotor Cages", IEEE Energy Conversion, Vol.16, No.3, 2001, pp 253-260 introduces some techniques to detect rotor slot harmonics in induction machines. But the proposed method needs a specially designed or modified rotor structure, which is either impossible or very costly for most of the current induction motors and drives. Again, the analysis of these harmonics has been performed by spectral estimation techniques.

Brief Description of the Invention:

In high performance motor drive applications, instantaneous detection of rotor position is often a necessity. Rotor position may be detected by using a position sensor of some kind. Alternatively, sensorless position methods exist which are limited at the low speed range and have lower accuracy. It is very desirable to improve the accuracy of such systems simply using an inexpensive sensor which can be externally employed or can be implemented using the motor current or voltage which is already measured in most drives. . c

Furthermore, it is envisaged that in future motor and motor drive may be treated as a component and the drive may be manufactured to fit on the motor frame. Also in some applications, speed control needs to be implemented without modification of the existing mechanical arrangement. Therefore, a position sensing system which does not require any modification on the motor and the mechanical arrangement, yet gives accurate position information is very valuable. The invention here provides a solution to these requirements.

The invention here is; a method and system facilitating rotor speed and position detection, by means of detecting rotor slot harmonics or higher order rotor slot harmonics and processing this information in real time to achieve the purpose.

The current, voltage or air-gap flux of nearly all electrical machines contains harmonics caused by rotor slots. Filtering out the harmonics of the fundamental excitation harmonics and extracting the rotor slot harmonics, the rotor speed and position can be estimated, provided that the number of rotor slots is known. The present invention differs from others, as it does not need any spectral analysis techniques or complex calculations in implementation. The computation required in the present invention can be implemented so that the whole speed or position detection process can be executed within the control cycle of a vector controlled drive.

Definition of the Figures:

The features of the invention believed to be novel are set forth in the appended claims. Further object, features and advantages of position and speed estimation with the method here, especially using higher order rotor slot harmonics will be better understood from the following description, relative to an example of a preferred and explanatory, non limiting schematic drawings, which:

FIG.l is a flowchart showing the overall method for determining the rotor position

in accordance with an embodiment of the invention.

FIG.2 is the spectrum data of the measured variable showing the higher rotor slot

harmonics during demodulation and filtering processes.

FIG. 3 is a block diagram showing demodulation process, filtering and the obtained

discrete signal.

FIG.4 is a spectrum data of the measured variable showing line harmonics,

fundamental rotor slot harmonics and higher order rotor slot harmonics.

FIG. 5 is the spectrum data of the measured variable of processed data and experimental results during the proposed estimation method

FIG. 6 is representation of demodulated and filtered discrete signal used for position estimation.

FIG. 7-8 are the experimental results obtained using the proposed rotor position estimation algorithm.

Definition of the Elements (Features/Components/Parts) on the Figures:

The definition of the features/components/parts which are covered in the figures, that are prepared in order to explain the present invention better, are separately numbered and given below.

1. Sampling process: Digitalization of the analogue measurements from the electrical machine (current, voltage, air gap flux, fringing flux etc.) using an analogue to digital converter with a sampling frequency.

2. Data demodulation: Multiplication of the sampled data with a virtually generated sinusoidal signal with fundamental frequency 3.

3. Fundamental frequency: Frequency equal to machine fundamental excitation voltage or current frequency.

4. Filtering: Filtering process of frequencies that are irrelevant to desired rotor slot harmonics.

5. Center frequency: Center frequency of the band-pass filter used in filtering process 4.

6. Position Counter: Algorithm that counts the filtered data to estimate position and speed of the rotor.

7. Identify Machine Speed & Position: Algorithm that outputs the speed and position information using the position counter 6. It requires number of rotor slots and number of poles as input.

8. Calculate Center Frequency: Algorithm that calculates the necessary center frequency 5 for filtering process 4 by using current speed and position data.

9. Non-available desired rotor slot harmonic: Non-available frequency component in the sampled data which becomes available after demodulation 2 process. 10. Side band harmonics: Lower and higher side band harmonics of the rotor slot harmonics available, before demodulation 2 process.

11. Spectral analysis visualization graph before demodulation 2 process; amplitude versus frequency

12. Desired rotor slot harmonic: Frequency component in the sampled data that becomes available after demodulation 2 process.

13. Modified side band harmonics: Lower and higher side band harmonics of the demodulated data, which deviate ±2f_s from desired slot harmonic, after demodulation 2 process.

14. Spectral analysis visualization graph after demodulation 2 process; amplitude versus frequency

15. Isolated rotor slot harmonic: desired and isolated frequency component in the demodulated data after filtering 4 process.

16. Spectral analysis visualization graph of filtering 4 process: amplitude versus frequency

17. Sampling frequency: frequency component used in sampling 1 process.

18. Electrical measurement: Real time machine electrical measurements such as voltage, current, air-gap flux, fringing flux etc.

19. ADC: Analogue to digital conversion process using the electrical measurements 18 and sampling frequency 17.

20. Sample[n]: Digitized form of sampled data waveform.

21. Demodulation: Demodulation process as described in data demodulation 2.

22. Sin (fs[n]): Digitized form of fundamental excitation frequency waveform as described in 3.

23. Demodulated Signal: Digital signal after the demodulation process 21.

24. Band-pass filter: Digital band-pass filter algorithm to eliminate irrelevant line harmonics and side-band 10 components as described in filtering process 4.

25. Filtered Signal: Digital form of the signal after eliminating the side-band harmonic components 10. 26. Digital visualization of the filtered signal: The visualization of the signal after the filtering process 4 that will be used for rotor speed and position estimation.

27. Fundamental rotor slot harmonics: Dominant rotor slot harmonics that are used in most of the previous studies and patents.

28. Rotor slot harmonics k=2: Higher order rotor slot harmonics with harmonic order of two.

29. Rotor slot harmonics k=3: Higher order rotor slot harmonics with harmonic order of three.

30. Lower side band component: Lower side band component of the rotor slot harmonic before demodulation 2 process.

31. Higher side band component: Higher side band component of the rotor slot harmonic before demodulation 2 process.

32. Spectral analysis visualization graph of sampled data 1 before demodulation 2 process; amplitude versus frequency.

33. Lower side band component: Lower side band component of the rotor slot harmonic after demodulation 2 process.

34. Central component: Central component of the rotor slot harmonic after demodulation 2 process that is used for rotor speed and position estimation.

35. Higher side band component: Higher side band component of the rotor slot harmonic after demodulation 2 process.

36. Spectral analysis visualization graph of sampled data 1 after demodulation 2 process;

amplitude versus frequency.

37. Filtered central component: Central component of the rotor slot harmonic after filtering process 4 that is used for rotor speed and position estimation.

38. Spectral analysis visualization graph of sampled data 1 after filtering process 4;

amplitude versus frequency.

39. Displacement ΔΘ: Visualization of the signal variation of the demodulated & filtered data over a single rotor slot displacement of rotor.

40. Displacement in a single sampling data A6_g: The displacement in a single sampled data during the motion of rotor. 41. Experimental results of sampled data 1 after filtering process 4; amplitude versus time.

42. Experimental results and comparison of measured and estimated rotor position;

degree versus time.

43. Zero crossing points of the demodulated and filtered signal 41.

44. Actual rotor position without interpolation: graph of the actual rotor position measured using an incremental encoder.

45. Estimated rotor position without interpolation; graph of the rotor position estimated by the aforesaid algorithm and method.

46. Experimental results of sampled data 1 after filtering process 4; amplitude versus time, for 47

47. Experimental result and comparison of measured and estimated rotor position for improved interpolation algorithm; degree versus time.

48. Actual rotor position with interpolation: graph of the actual rotor position measured using a coupled incremental encoder.

49. Estimated rotor position with interpolation; graph of the rotor position estimated by the aforesaid improved interpolation algorithm.

Detailed Description Of The Invention:

Magnetic structure of electrical machines (rotor and stator slotting) cause various harmonics in the air gap flux, phase current and phase voltage. These harmonics are combined in a generalized formula by Nandi S, et al.

In this embodiment of the invention rotor slot harmonics or its higher order harmonics may be utilized.

The frequency of higher order rotor slot harmonics (f^n) can be expressed as given in the following equations:

fRSH = k.(Z/p)xf_r ± f_s (1)

or

fRSH = k x f_R ± f_s (2) where:

k represents an integer related with the order of rotor slot harmonic

Z represents the number of rotor slots

p represents the number of pole pairs in the electrical machine

f_r represents the mechanical frequency of rotor (Hz)

f_s represents the applied stator frequency (Hz)

f_R represents the rotor slot permeance variation frequency (number of rotor slots per pole multiplied with mechanical frequency of rotor)

In one embodiment of the present invention, as illustrated in Fig. 1, the speed and position estimation algorithms start by sampling 1 of the data. The sampled data may be induced voltage from an external search coil, the phase voltage measurement for a current controlled machine or phase current measurement for a voltage controlled machine. One advantage of the present invention is the sampling 1 process does not need a sequential data set taken in a sampling period and storage medium as illustrated in Fig 3. As the data sampling continues, each captured data is demodulated 2 via multiplication with a sinusoidal signal at the applied fundamental voltage or current frequency 3. Note that in this process no information from electrical machine parameters is needed. In the demodulation process only an artificially generated sinusoidal signal with unit magnitude is used 22. This arbitrary signal may be easily generated since; applied excitation frequency is readily available in most drive applications.

Then, the demodulated signal is filtered 4 with a band-pass filter. This operation needs a reference signal f_c 5. The process, until 4 in the flow chart of Fig. 1, is explained once again in detail in Fig. 2, where identification of the signal at frequency f_c (kF_R) is also explained.

The frequency of demodulated and filtered signal is directly related with RSH frequency. Instead of determining the signal at frequency RSH explicitly, a signal at higher order rotor slot harmonic frequency can be isolated from the other signals and used in the process. Due to the very high frequency of higher order harmonics utilizing these in the process may be more advantageous. Each cycle of the extracted signal corresponds to one slot pitch of the rotor. The position-counter 6 block of the algorithm is responsible for identifying the of rotor harmonic cycles. In this manner the number of cycles (hence the distance travelled) in a known time interval can be found and position of the rotor and the speed of the rotor can be accurately estimated 7.

Also, from the estimated rotor speed the center frequency of the filter for next estimation cycle is calculated in block 8 of the speed estimation algorithm. The algorithm continues by sampling new data.

Although, no spectrum analysis is necessary in the present invention, for ease of illustration in Fig. 2, the spectrum analysis of the sampled data during the demodulation 2 and filtering 4 processes are shown. As given in equation 2 and presented also in Fig .2 lower and higher side band harmonics 10 exist in sampled data. These two side band harmonics 10 exist apart from each other with twice the frequency of applied fundamental excitation. This helps identification of harmonics due to due to rotor slotting.

Demodulation process 14 is simply the multiplication of each sampled data with a unit magnitude sinusoidal signal with applied voltage frequency. In one embodiment of the present invention, the process can be applied as given in Fig. 3. For sampling of the data a computer medium or any analog to digital converter (ADC) 19 with a sampling frequency of (fsamp) 17. Demodulation 21 is a real-time process which is performed continuously as the each data is being sampled. In the demodulated signal of higher order rotor slot harmonics 14 there are three components; higher and lower side band harmonics 13. These are placed with a frequency difference of twice the supply frequency, from rotor frequency kf_R 12. Obtaining the center component (kf_R) 15 is vital for the present invention. Therefore, the sideband harmonics should be eliminated by means of some filtering 16 and the center component (kf_R) 15 must be extracted. In one embodiment of the invention, a digital bandpass filter 24 is used which eliminates all harmonics but the center component (kf_R). The output of the filter 26 is a digital signal with frequency of kf_R.

Each cycle of the filtered signal 26 correspond to small change in rotor position. This displacement angle (ΔΘ) is equal to;

2.π

Αθ = (3)

k.Z

In one embodiment of the present invention, the period of the signal can be detected by measuring the zero crossings of the signal. In that embodiment, the rotor position (Θ_Γ) can be calculated as a sum of displacement angles. Thus, rotor position (Θ_Γ) becomes; Where, n represents the number of zero crossings.

In another embodiment of the present invention, the accuracy of the rotor position estimation can be improved by some means of interpolation. In other words, instead of updating position data at each cycle, it can be updated by smaller steps at each sampled data. Also, instead of zero crossing detection of the filtered data 26, peak detection or any other method may be used as needed.

The method has been tested on several induction machines. In one of these tests, the fringing flux is measured with an external sensing coil while the induction machine is running on a voltage controlled inverter at 40 Hz. The machine has two poles with a number of 18 rotor slots and rotating at 2225 rpm (in other words; f_r = 37.08 Hz) as the measurements are captured.

Fig 4. is an example of this embodiment of the invention. In the spectral analysis, the fundamental rotor slot harmonics 27, and second 28 and third 29 order rotor slot harmonics are shown. Higher order rotor slot harmonics 28, 29 are the harmonics that are preferred to be used in the invention.

Using equation 1, the third order (k=3) rotor slot harmonics 29 can be verified to occur at 1962 Hz and 2042 Hz. Note that, fundamental rotor slot harmonics are mixed with many line harmonics, which is the main reason for preferring the use of higher order rotor slot harmonics.

In Fig.5, the detailed view of the spectral analysis of the data presented in Fig. 4, during position estimation steps is given. In demodulation step 36, the multiplication of the sampled data with a signal with frequency of applied voltage (40 Hz) output three components; lower side-band component at 1922 Hz 33, center component at 2002 Hz 34 and higher side-band component at 2082 Hz 35. The center frequency component frequency is directly related with rotor slot permeance variation frequency (f_R), since in this embodiment the rotor slot harmonic order three (k=3) is used; it can be concluded from equation 2 that, 2082 Hz 35 is equals to k.f_R. Also note the frequency difference between side band harmonics 33, 35 doubles as a result of the demodulation process.

The filtering process 38 aims to extract only the central component 37, thus real-time position estimation can be performed. Once the central higher order rotor slot harmonic component is extracted, the rotor position can be estimated in many simple ways.

In Fig.6, the demodulated and filtered data for one embodiment of the present invention is given. This final data is a discrete signal with a sampling frequency of f_samp- As given in equation (3), one full cycle of the signal 39 corresponds to a small displacement angle (ΔΘ). The full displacement can be detected either by zero-crossing detection, peak detection or any other methods that is not listed. Once, the full cycle of the signal is detected, rotor position should be updated as follows 7;

Θ = Θ' + ΑΘ (5) In Eq. 5, Θ is the updated rotor position and θ' is the previous rotor position.

Rotor speed may be easily calculated by measuring the elapsed time period for a certain rotor displacement. Then the center frequency for the filter is updated if there is any change in the rotor speed for next estimation cycle 8.

In Fig.7 the demodulated and filtered higher order rotor slot harmonics obtained from air gap flux measurement is given 41 while the machine is driven with an inverter at 20Hz. Also, the estimated rotor position 45 which is updated on zero crossings 43 of the rotor slot harmonics is given. The estimated rotor position is compared in Fig. 7 with the actual rotor position 44 which is measured with an accurate incremental encoder coupled to machine shaft. The results show that the algorithm predicts the rotor position very accurately. However, numerous improvements are possible for rotor position algorithm using some interpolation techniques. In Fig.8, the same data is evaluated using on of the interpolation technique. The rotor position is updated on each zero crossing 47 as previously mentioned; but instead of assuming rotor position constant between each zero crossing instants it is updated at each cycle using a simple interpolation technique. As it can be seen, the estimated rotor position matches very well with the actual rotor position 48.

The described embodiments of the present invention have many advantages,. The technique envisaged above is very efficient in computational manner because the present invention does not need any spectral analysis. The rotor position can be estimated by simple multiplication/ filtering and counting blocks. Therefore, the estimation can be easily performed within a typical control cycle of a vector controlled drive. Hence the dynamic performance of so called sensorless vector control techniques can be greatly improved. Specific options have been described above for carrying out the more generic embodiments of the present invention. It is contemplated, and will be apparent to those skilled in the art from the foregoing specification, drawings, and example that modifications and/or changes may be made in the embodiments of the invention. Accordingly, it is expressly intended that the foregoing description and example are illustrative of one embodiment only, not limiting its possible implementations, and that the true spirit and scope of the present invention be determined by reference to the appended claims.

Claims

1. A method comprising the steps of:

(a) Sampling a data that includes rotor slot harmonics or higher order rotor slot harmonics

(b) Demodulation of the sampled data

(c) Filtering the data to eliminate side band rotor slot harmonics

(d) Applying counting algorithm to count the number of cycles of rotor slot harmonics within a time interval in the extracted signal.

(e) Updating rotor position, speed and filter characteristic for next cycle.

2. The method of claim 1 wherein sampled data is obtained from

(a) Phase voltage

(b) Phase current

(c) Induced voltage from_, a sensing coil that measures air gap flux or fringing leakage flux

3. The method of claim 1 wherein sampling process comprises an analog to digital conversion with sampling frequency (fsamp).

4. The method of claim 3 comprises either a stand alone circuitry or a computer medium.

5. The method of claim 1 wherein higher order rotor slot harmonics comprises the fundamental as well as the second, third or higher (k=2, 3 ...) frequency multiples of rotor slot harmonics.

6. The demodulation process of claim 1 comprises the multiplication of the sampled data with an artificially generated sinusoidal signal.

7. The sinusoidal signal of claim 6 comprises a sinusoidal signal with following properties;

(a) With some magnitude

(b) Frequency equals to the applied fundamental excitation frequency of the machine.

8. The method of claim 7 comprises generation of a discrete sinusoidal signal with sampling frequency equals to the analog to digital conversion sampling frequency.

9. The method of claim 6 comprises the multiplication of sampled data and the generated signal on each cycle simultaneously without any buffer.

10. The method of claim 6 output three components;

(a) Central component with frequency k.f_R

(b) The lower sideband which has a frequency smaller than twice the fundamental excitation frequency than the central frequency component (k.f_R - 2.fs)

(c) The higher sideband which has a frequency larger than twice the fundamental excitation frequency than the central frequency component (k.f + 2.fs)

11. The method of claim 1 wherein filtering process comprises the elimination of sideband harmonics and other irrelevant harmonics except the central component which is comprised by claim 10.

12. The method of claim 10 where f_R represents the rotor slot permeance variation frequency which corresponds to number of rotor slots per pole multiplied with mechanical frequency of rotor in Hertz.

13.The method of claim 1 wherein filtering comprises a band pass filter.

14. The method of claim 1 wherein counting algorithm comprises the counting the number full cycles of the filtered signal.

15. The method of claim 14 comprises detecting

(a) Zero crossings of the signal.

(b) Peaks of the signal.

(c) Or any other property of the signal

16. The method of claim 1 wherein counting algorithm comprises an algorithm that counts the number of points sampled between each cycle of the filtered signal.

17. The method of claim 1 wherein updating rotor position comprises increasing the previous rotor position by adding the displacement angle detected.

18. The method of claim 17 displacement angle compromises change in the rotor position which equals in radians to 2.ΊΙ divided by order of harmonics (k) and number of rotor slots (Z).

19.The method of claim 17 wherein displacement angle comprises higher resolution detection with interpolation.

20. In one method of claim 19 wherein interpolation comprises dividing the displacement- angle method of claim 18, to number of points sampled between each cycle of the filtered signal.

21.1n the method of claim 1 wherein updating rotor position speed calculation comprises determining the elapsing the time period for a certain amount of rotor displacement and dividing displacement angle to elapsed time.

22. The method of claim 1 wherein updating rotor position speed comprises calculating rotor mechanical speed in Hertz and in revolutions per minute.

23. The method of claim 1 wherein updating rotor filter characteristic comprises two properties of a band pass filter;

(a) Centre frequency

(b) Bandwidth

24.The method of claim 23 wherein center frequency property of band pass filter comprises the present higher order rotor slot harmonic which is found by multiplication of rotor speed in Hertz(f_r) with higher order rotor slot harmonic order (k) and number of rotor slots per pole pair (Z/p).

25. The method of claim 23 wherein band with property of band pass filter comprises update of filter bandwidth according to applied frequency.

26. The method of claim 23 wherein updating the filter characteristics comprises updating within each algorithm cycle or updating in more cycles, if the rotor mechanical time constant is longer compared to update frequency.

27. The method of claim 26 wherein updating the filter characteristics comprises redefining the filter constants, within a computer medium or analog circuitry.

28. The method of claim 26 wherein updating the filter characteristics comprises a stand alone varying center frequency band-pass filter circuitry.