Upload
karthik-nunna
View
215
Download
0
Embed Size (px)
Citation preview
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
1/20
ANN APPROACH FOR THE INCIPIENT FAULTS DETECTION IN SINGLE PHASE
INDUCTION MOTOR.
ABSTRACT:
Emerging technology of artificial neural networks has successfully been applied in a variety
of areas such as , fault detection, control, signal processing, and many other applications. This
paper presents an artificial neural network(ANN)methodology for the detection of incipient
faults in induction motor. The ANN based detector avoids the problems associated with
traditional incipient faults detection schemes by employing readily available information,
such as motor intake current, rotor speed ,and stator winding temperature. The two types of
ANN based incipient fault detectors are developed , ie , conventional and higher order
ANN fault detector. The results of the evaluation indicates that the higher order ANN
based incipient faults detector provides more accurate results.
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
2/20
INTRODUCTION:
The use of small and medium size induction motors in industry and in home appliances is
extensive and continuous. These machines are exposed to wide variety of environments and
conditions and there by develop incipient faults. The turn-to turn insulation failure and bearing
wear are the most common types of faults to be investigated. The various causes for weaking of
insulation and bearing wear of the electrical machines are (i) aging,(ii)switching
surge,(iii)heat,(iv) contamination by oil (v)incipient faults, (vi)damage during installation, etc
.A turn-to turn short circuit of winding causes the decrease in equivalent turns of the winding of
the machine. It results in rise in input line current of winding . This causes the increased heating
in the core due to additional I2R losses and drop in speed. The increased heating will cause a
corresponding temperature rise of stator winding, there by decreasing the life expectancy of the
winding insulation. Stator winding insulation failure will cause additional shorted turns, further
rise in temperature, and this leads to increase in the rate of deterioration of winding insulation. Ifleft unchecked, this process will cause eventual destruction of the relative winding and render the
machine inoperative.
The motor bearings of induction motor are subject to
deterioration caused by inadequate or contaminated lubrication, misapplication, or misalignment
. The frictional losses within the motor bearings are used as a criterion to determine their
condition. The bearings losses that are negligible or within 5% of the overall losses of the motorand can be consider as good. However, bad bearings can cause permanent damage to them
selves as well as to the rotor or the stator winding and must be detected quickly.
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
3/20
In earlier works, two parameters have been considered for the fault detection, which are motor
intake current(I) and rotor speed(w). However , in this paper , stator winding temperature(Tw) is
included as an additional parameter for better fault detection. The proposed artificial neural
network (ANN) based detector avoids the problem associated with traditional schemes, such as
continuous on-line monitoring by experts, parameter estimation approach, etc. In this technique
more readily available informations , such as , motor intake current , rotor speed and stator
winding temperature have been used for decision making.
Neural network toolbox in MATLAB environment is used to optimize thenetwork. The data required to train the neural network is generated in the laboratory on specially
designed single phase, squirrel cage 1- hp induction motor. The winding turns of one pole in a
four pole machine is brought out for the shortening in steps. This gives the effect of inter-turn
short circuit of motor winding, and it causes the imbalance in air gap flux.Thus , this produced
the mechanical stresses over the shaft of motor. This corresponds to the additional eccentric
loading caused by a worn bearing in the ac induction motor. This allows nondestructive
emulation of worn or failed induction motor bearings.
In this paper, ANN is trained with three inputs , such as (i) motor intake
current, (ii) speed , and stator winding temperature as a simple conventional network. Further
the original input space is expanded to six dimentions ,which is a higher order network. In both
these causes the ANN is tested satisfactorily . Results of evaluation indicate that the higher order
neural network based incipient fault detector predicts the winding and bearing condition of the
motor to a more satisfactory level of accuracy.
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
4/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
5/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
6/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
7/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
8/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
9/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
10/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
11/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
12/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
13/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
14/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
15/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
16/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
17/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
18/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
19/20
8/7/2019 ANN approach for the incipient faults detection in single phase induction motor
20/20