EE 153 Power Electronics:Fast Fourier Transform Spectrum Analyzer

o perform the vibrations and identify the problems in the electrical machinery at the plant. The curves of vibration are taken frequently for a Squirrel cage Induction Motors from a single machine with the rating of 3 phase, 5 HP, 415V, 50Hz, 1440rpm, 28 bar rotors, 36 stator slots (Silva, 2011, p. 111).

The vibration curves which were recorded include healthy motor FFT in the first month for the conditions of a half, full, and no load conditions, and FFT took after 5 months during one rotor bar failure for the conditions of a half, no, and full load. FFT took after 10 months during two rotor bar failure conditions of full, no, and half load. The following shows the analysis and identifications of the FFT spectrum together with their correct label:

Healthy motor no load FFT spectrum

Healthy motor half load FFT spectrum

Healthy motor full load FFT spectrum

When the supply of three phase of balanced applied voltage to the motor induction that is healthy, the constant current flow in entire phases occurred. This makes a constant field of a magnet in rotor and stator. The MMF interaction in the rotor circuit with the flux of the magnet stator may yield a positively charged torque of motor that is steady which drives the rotor in a direction of forward giving out mechanical output which is useful. An acceptable vibration radial will be present within the stator. The evaluation of spectral signal of vibration produces a dominate 49.81Hz of a frequency with the magnitude of 68.21dB and without harmonic band frequencies as shown above. The frequency of 41.81 Hz is close to 50Hz components of frequency and is a positively charged sequence that is good for a rotor driving the system in the inward path and this element indicates a healthy motor (Goldman, 2012, p. 219).

FFT spectrum of the motor one bars of broken rotor with a condition of no load

FFT spectrum of the motor two bars of broken rotor under the condition of no load

Table 1 showing analysis of no load one broken and 2 broken rotor bars conditions

 
Due to the broken rotor bar, the quantity of current flowing in the rotor is not equal and will make an asymmetrical field of a magnet. The interaction between the field of stator and MMF of asymmetrical rotor gives a torque pulsating, leading to the production of the vibration in the motor. The slip increased and the speed reduced. The spectral analysis of vibration of motor indicates the components of a frequency of 49.81Hz with the amplitude of 64.01dB for a single rotor broke bar and 41.81Hz with the amplitude of 61.21dB amplitude of two bar broken rotor. The existence of left-hand fs (1-2ks) and fs (1+2ks) on the right side frequency components of the bands are presents due to the asymmetry of rotor and fluctuation of speed. Also, the figures above show the amplitude and the faults of frequency for two and one bar broken rotor under the condition of half load (Silva, 2011, p. 310).   

FFT spectrum of the motor one bars broken rotor under condition of half load

FFT spectrum of the motor two bars broken rotor under condition of half load

Table 2 showing analysis of one and two bars broken rotor under the condition of half load 

 
When the load of the motor verified improved to 50 percent of the value evaluated, there will be the increase in current and slip but a decrease in speed. The magnetic field effect asymmetrical will be higher related to the condition of no load. Therefore, oscillations of speed and the torque pulsating in rotor setup will produce vibration and noise buzzing in the motor. The evaluation of spectrum of vibration of signal illustrates that the side band's modular of frequency f_b = (1+ or - 2ks) fs increases compared to the condition of no-load which is shown in the figures above where frequency and magnitude of the side harmonic band components of frequency for two and one bar broken through half load conditions (Goldman, 2012, p. 66).

FFT spectrum of the motor one bars broken rotor under condition of full load

FFT spectrum of the motor two bars broken rotor under condition of full load

Table 3 below shows the analysis of the one and two bars broken under condition of full load 

 

When the load on the fault motor increases to the peak value of the load, the speed and terminal voltage reduces while the slip increases. The asymmetry of the magnet in the oscillation speed and rotor escalates, resulting in extra vibration in the setup motor (Goldman, 2012, p. 97). The side band of harmonic magnitude components f_b = (1+ or – 2ksfs) increase when compared to the condition of half and no load. This is shown in the figures above where magnitudes and frequencies (1+ or – 2ks) fs harmonic side band components under a condition of the full load. In every situation, the results of the experiment are analyzed and validated using equations and the final results gotten are matching closely.   

Problems Faced During the Experiment of Vibration Analysis

Analysis and experimental results of the broken bar detection grouped on the motor operation vibration under half, full and no load condition by the use of the presented accelerometer. The outcome of the experiment shows that the presence of two frequency slip (2sfs) because of torque pulsation in the range of low frequency and the side band of harmonic frequency component fs (1+ or – 2ks) nearby the essential components of frequency may show clearly that the rotor fault bar was damaged (Goldman, 2012, p. 248).

The more the number bars that have been broken, the greater harmonic side bands fs (1+ or – 2ks) monitored by extra buzzing noise vibration and this, in turn, reduce the torque of the motor (Silva, 2011, p. 134). Also as there is variation between load and broken motor bar, the side band of harmonic amplitude (1+ or – 2ks) fs increases. This severity of fault becomes high causing the premature failure.

How to Rectify the Problem

In ensuring the consistency, repeatability and effectiveness of the proposed method of getting % factor of severity as in another comparison, it should be useful to simulation results plus experiments with different ratings of the motor induction under conditions of different loading (Goldman, 2012, p. 82). The person performing the experiment should also be in a position to detect the abnormalities in the induction motor to avoid the damages mentioned above. The fault of premature failure during the FFT spectrum experiment can be mitigated through reducing the number of broken bars which will enable the reduction of harmonic side bands fs (1+ or – 2ks) monitored by extra buzzing noise vibration. This step will prevent the reduction in the torque of the motor hence preventing the fault of premature failure.

Conclusion

This experiment presents the method of rotor bar detection. The use of FFT allows for the classical detection of the bar which has been broken based on the transform of Fourier applications to the current stator of the motor induction running in a stable state. The discovery of faults is attained by evaluating the components of two side band that occur around the frequency components of supply. This approach has its important advantages like being simple in the acquisition of data system and software required, and also toughness and provides results of satisfaction. Even though in some situations, induction motors that are unloaded, the sidebands usually overlap the frequency of supply and the slip is normally low. This makes the detection of their presence difficult and also their use.

References

Goldman, S. (2012). Vibration Spectrum Analysis: A Practical Approach. London: Industrial Press Inc.

Silva, C. W. (2011). Vibration: Fundamentals and Practice, Second Edition. Michigan: CRC Press.