Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
A SYSTEM FOR ASSESSMENT OF MULTIPLE FAULTS IN INDUCTION MOTORS
Document Type and Number:
WIPO Patent Application WO/2019/167086
Kind Code:
A1
Abstract:
A system for assessment of multiple faults in induction motors, said motor motor being supplied from a 3-phase supply (100) using a variable frequency drive (VFD), said system comprising: a current sensor (102) for sensing a single phase stator current; an anti-aliasing filter (105) for subsequent conversion of sensed current into digital format using an analog to digital converter (106), said acquired signal consisting of a high magnitude fundamental component, fault frequency components, and noise and amplitude estimation block (107) for estimation of fundamental signal further conditioned with input signal to motor in order to obtain a conditioned input signal for estimation of slip and speed of motor to determine a fault search band correlative to fault frequency components, said fault frequency components estimated by means of a spectral estimator to determine if said motor comprises rotor faults, eccentric specific faults, bearing faults, and / or stator faults.

Inventors:
ROUTRAY, Aurobinda (Department Of Electrical Engineering Indian Institute Of Technology, Kharagpur Kharagpur 2, West Bengal, 721302, IN)
NAHA, Arunava (Department Of Electrical Engineering Indian Institute Of Technology, Kharagpur Kharagpur 2, West Bengal, 721302, IN)
SAMANTA, Anik Kumar (Department Of Electrical Engineering Indian Institute Of Technology, Kharagpur Kharagpur 2, West Bengal, 721302, IN)
PAWAR, Amey (Department Of Electrical Engineering Indian Institute Of Technology, Kharagpur Kharagpur 2, West Bengal, 721302, IN)
SAKPAL, Chandrasekhar (1C/172, Kalpataru Aura Lbs Marg Ghatkopar West,,Mumbai 6, Maharashtra, 400086, IN)
Application Number:
IN2019/050181
Publication Date:
September 06, 2019
Filing Date:
March 01, 2019
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
ROUTRAY, Aurobinda (Department Of Electrical Engineering Indian Institute Of Technology, Kharagpur Kharagpur 2, West Bengal, 721302, IN)
International Classes:
H02K16/04; G01R31/00
Foreign References:
US4761703A1988-08-02
US20130049733A12013-02-28
Attorney, Agent or Firm:
TANNA, Chirag (Ink Idee, B-72, 62, 73 Pereira Nagar No. 7,Khopat, Thane 1, Maharashtra, 400 601, IN)
Download PDF:
Claims:
CLAIMS,

1. A system for assessment of multiple faults in induction motors, said motor motor being supplied from a 3-phase supply (100) using a variable frequency drive (VFD), said system comprising:

- a current sensor (102) for sensing a single phase stator current;

- an anti-aliasing filter (105) for subsequent conversion of sensed current into digital format using an analog to digital converter (106), said acquired signal consisting of a high magnitude fundamental component, fault frequency components, and noise and

- amplitude estimation block (107) for estimation of fundamental signal further conditioned with respect to input signal to said motor in order to obtain a conditioned input signal for estimation of slip and speed of said motor to further determine a fault search band correlative to said fault frequency components, said fault frequency components being estimated by means of a spectral estimator in order to determine if said motor comprises rotor faults, eccentric specific faults, bearing faults, and / or stator faults.

2. The system for assessment of multiple faults in induction motors as claimed in claim 1 wherein, said spectral estimator (110, 111) estimates spectrum of the conditioned stator current in the specified spectral band and sends it to a peak detector, said peak detector (112) finds the fault peak information form the spectrum, and an amplitude estimator (107, 113) estimates the amplitude of the fault peaks.

3. The system for assessment of multiple faults in induction motors as claimed in claim 1 wherein, said spectral estimator is configured to determine of rotor faults, in that, a broken rotor bar (BRB) and a broken end ring, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency (/) of the motor obtained by sensing motor current by means of a motor current sensor, slip (5) of the motor estimated by a slip estimator, and rotational frequency of the said motor by using the slip and fundamental frequency derived from the stator current.

4. The system for assessment of multiple faults in induction motors as claimed in claim 1 wherein, said spectral estimator is configured to determine eccentric specific faults, in that, a static eccentricity fault, a dynamic eccentricity fault, and a mixed eccentricity fault, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency (/) of the motor obtained by sensing motor current by means of a motor current sensor, sensing rotor slots ( R ) using an input mechanism associated with this system, rotational frequency (k = 1, 2, 3, ... , . fr) of the motor by means of a rotational frequency sensor, slip (5) of the motor estimated by a slip estimator, number of bars of the motor input using an input mechanism associated with this system, number of pole-pairs (p ) of the motor using an input mechanism associated with this system.

5. The system for assessment of multiple faults in induction motors as claimed in claim 1 wherein, said spectral estimator is configured to determine bearing faults, in that bearing faults, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency if) of the motor by sensing motor current by means of a motor current sensor, sensing rotational frequency (k) by means of deployed rotational frequency sensor, and using an input mechanism configured to facilitate input of number of balls (AO of the motor, input of ball diameter (bd ), input of ball pitch diameter ( dp ), and contact angle (b ) between the ball and the races.

6. The system for assessment of multiple faults in induction motors as claimed in claim 1 wherein, said spectral estimator is configured to determine of stator faults, in that stator faults, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency if) of the motor, sensing slip (5) of the motor (estimated by a slip estimator), and using an input mechanism configured to input number of pole-pairs ip) of the motor and number of bars of the motor.

Description:
A SYSTEM FOR ASSESSMENT OF MULTIPLE FAULTS IN

INDUCTION MOTORS

FIELD OF THE INVENTION:

This invention relates to the field of electrical engineering.

Particularly, this invention relates to a system for non-invasive assessment of squirrel cage induction motor faults and running information using only a single phase stator current as input.

Specifically, this invention relates to a system for assessment of multiple faults in induction motors.

BACKGROUND OF THE INVENTION:

Squirrel Cage Induction Motors (SCIMs) are the primary motive force provider for any industry and railway transportation. Induction motors may become faulty because of the loading transients, manufacturing defects, installation issues, degradation of insulation etc. Faults developed initially are incipient in nature, so even if there is a fault in the motor, the motor may operate as a normal system with subtle deviation in its states. If faults are not detected and proper maintenance has not been taken, the faulty motor may leads to complete failure resulting in loss of productivity.

As mentioned in the non-patent literature by S. Nandi, H. A. Toliyat, and X. Li,“Condition Monitoring and Fault Diagnosis of Electrical Motors— A Review,” IEEE Trans. Energy Convers., vol. 20, no. 4, pp. 719-729, Dec. 2005, among the various induction motor faults around 40%-50% are bearing faults, 5%-10% are severe rotor faults, and 30%-40% are stator related faults.

Across all prior art, the more fundamental issues related to the accurate spectral analysis of the armature current and detectability of the faults remain unaddressed. Some of the issues are supply frequency estimation when the motor is running with variable frequency drives, ill conditioning of the spectral estimation algorithms, etc.

OBJECTS OF THE INVENTION:

An object of the invention is to provide a non-invasive assessment of 3- phase squirrel cage induction motor faults.

Another object of the invention is to provide a system and apparatus for remote assessment of faults of a 3 -phase squirrel cage induction motor using a cloud-based server. Data from the induction motor can be assessed directly by the system, or it can be uploaded into a cloud-based server for processing.

Another object of the invention is to provide a non-invasive assessment of squirrel cage induction motor faults and running information using only a single -phase current as input.

Yet another object of the invention is to develop an early warning system, apparatus, and method for 3-phase squirrel cage induction motor faults in their inception with a single current sensor. Still another object of the invention is to provide a sub- system and method for estimating supply frequency and rotational speed of a 3 -phase induction motor without using any additional sensors other than the single current sensor. These parameters are used for classification of specific faults.

An additional object of the invention is to provide a system and method that can efficiently condition the armature current signal by tracking and removing the fundamental supply frequency component from the armature current signal. Conditioning of the input signal is quintessential for detection of weak failure modes of the induction motor.

SUMMARY OF THE INVENTION:

This invention describes an online condition monitoring system developed to detect incipient induction motor faults.

Spectral analysis of the stator current is central to fault detection. The acquired current signal is preconditioned by removing the high-amplitude fundamental component using Extended Kalman Filter (EKF) and a second- order signal model. Pre-conditioning of the input signal improves the detectability of the low amplitude fault frequency component under adverse loading conditions. An elegant Rayleigh quotient-based spectral estimator is designed specifically to detect and quantify weak SCIM faults. The accuracy of the proposed estimator was found to be comprehensive when compared with Fourier analyses and Multiple Signal Classifier (MUSIC) with its time complexity lower than MUSIC. Accurate relative amplitude estimation capability and low complexity make this estimator appropriate for embedded applications. The slip and the supply frequency are found non-invasively from the stator current for forming adaptive fault frequency search bands required for the spectral estimation. Spectral estimation over small multiple bands reduced the computational burden extensively. The peak magnitude of the spectrum is utilized to quantify the fault severity. As all the methods require only the single -phase stator current, the reliability is increased, and the complexity and the cost of using additional sensors is reduced. The faults under consideration comprise broken rotor bar (BRB), broken end ring (BER), inter-tum short circuit (ITSC), eccentricity related faults, and bearing faults. The bearing faults are classified as outer raceway fault, inner raceway fault, and rolling element fault.

The features of the fault detection unit can be summarized as:

1. Detection of BRB, BER, ITSC, Eccentricity, and all the bearing faults.

2. The software is capable of generating an automated report in document formats having all the motor fault information and graphs in it.

3. A provision is made to store fault specific signature history of each motor so that the fault progression can be monitored over time.

4. Estimation of motor slip and supply frequency with input signal conditioning are the added features.

In at least an embodiment, an estimator is configured to determine of rotor faults, in that, a broken rotor bar (BRB) and a broken end ring, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency (f) of the motor obtained by sensing motor current, rotational frequency ( k = 1,2,3,...,. f r ), and slip (5) of the motor.

Specifically, Broken rotor bar; (BRB) f brb = (1 ± 2 ks)f, (eq. la)

• Broken end ring; f ber = (1 ± 2 ks)f, (eq. lb)

Using the spectral estimator, a band of frequency is searched around the f brb component. If a peak is found in the band around the theoritical f brb , then the motor is said to have broken rotor bar. The searchband of 2 Hz has been fixed for finding the fault component around the vicinity of the f brb . If the f ber is found to be within the 2 Hz band around the theoritical f ber , then the motor is said to have broken end ring. For this, a motor current sensor is deployed to sense motor current.

In at least an embodiment, an estimator is configured to determine eccentric specific faults, in that, a static eccentricity fault, a dynamic eccentricity fault, and a mixed eccentricity fault, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency ( / ) of the motor, rotor slots (R ), ( k = 1,2,3,...,.), slip ( .v ) of the motor, number of pole-pairs ( p ) of the motor, and the order of the stator time harmonics that are present in the power supply driving the motor is given by (v = ±l,±2,...)· . For this, a motor current sensor is deployed to sense motor current. Additionally, an input mechanism facilitates input of number of rotor slots of the motor, number of bars of the motor, and number of pole-pairs of the motor.

Eccentricity Faults: f ecc (kR ± n d )——— ± v f, (eq. lc)

P

Static Eccentricity n d = 0,

Dynamic Eccentricity n d = 1, 2, 3, ... , Mixed Eccentricity (eq. Id)

In at least an embodiment, an estimator is configured to determine of bearing faults, in that bearing faults, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency (/) of the motor, the rotational frequency ( / r ) from the slip and fundamental frequency information, and using an input mechanism the system is configured to facilitate input of number of balls (A of the bearing, input of ball diameter (b ), input of ball pitch diameter ( d ), and contact angle ( b ) between the ball and the races. The rotational frequency is given by f = (l- s)f / p.

Bearing Faults f = f ± t (eq. le) o Inner raceway;

o Outer raceway;

o Rolling element;

In at least an embodiment, an estimator is configured to determine the stator faults, in that stator faults, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency ( / ) of the motor, slip ( s ) of the motor, and using an input mechanism configured to input number of pole-pairs ( p ) of the motor and number of bars of the motor n . • Stator Faults

( n L

o Inter turn short circuit (ITSC); f i sc l±-*-(l- s) f (eq. If)

V P J

In the above paragraphs, s, f, p,R,n b are the slip, fundamental supply frequency, and number of pole-pairs, rotor slots, number of bars, respectively.

The order of the stator time harmonics that are present in the power supply driving the motor is given by (v = ±l,±2,...), and k = 1,2,3,...,. f r is the rotational frequency,

N is the number of balls,

b d is the ball diameter,

d p is the ball pitch diameter, and

b is the contact angle between the ball and the races.

This invention presents methods, technologies, and apparatus required for online condition monitoring of induction motors. Induction motors’ are widely used in different industries and hence led to wide-scale research to detect faults. Detecting faults at an early stage can avoid unscheduled cessation of production. An unsupervised condition monitoring system for the detection of weak and incipient faults in induction motors is the objective of this patent. The faults under consideration comprise of broken rotor bar, broken end ring, eccentricity related faults, inter-turn short circuit, and bearing faults. Spectral analysis of the stator current is carried out for fault detection. A Rayleigh quotient-based spectral estimator is used for this purpose. This method has a lower computational complexity and can also find the amplitude of constituent frequency components with high accuracy. This information is used for quantifying the severity of fault present.

Pre-conditioning of the stator current is carried out by an EKF-based signal conditioner. The spectral estimator thus was able to detect sinusoids that are close to the fundamental and have negligible magnitude compared to the fundamental. It was found that the component of the mixed eccentricity was present inherently in all the motors. Rotational speed is estimated from the slip obtained from this mixed eccentricity component. The slip along with the value of the fundamental frequency is used for creating multiple search bands for detecting multiple faults efficiently. The central frequency of the bands are selected from the fault signatures given in (eq.la - eq.lf) using the motor nameplate parameters, estimated value of slip, and the fundamental frequency. The Rayleigh-quotient-based spectral estimator is then used to find whether a fault component is present the in the band. In case the fault frequency component is present, the amplitude of the fault component decides the severity of the fault. The above-mentioned procedure is valid for detecting all the faults. The spectral estimator can reliably quantify the degree of damage. The normalized peak magnitude of sideband frequency components with respect to the fundamental is considered as the discriminating feature for fault detection and quantification. An embedded hardware platform is developed for online fault diagnosis and RT fault simulation.

The systems developed in the present invention finds useful applications not only in areas in fault detection but also other areas of science and engineering in the near future. For example, the Rayleigh quotient-based spectral estimator can be modified to be used in the direction of arrival applications with the added advantage of exact peak magnitude which is not available in subspace-based methods.

The system, of this invention, is capable of storing the history of the fault components. As a result, a user can find out the progression of each fault using this feature. The threshold of the faults are found using statistics of faulty and healthy data. Further, reliability is achieved by pairing the threshold with historical data of the motor once the system starts monitoring the motors.

Typically, the system provides an apparatus for continuous or an intermittent monitoring of squirrel cage induction motor in their inception with only a single -phase stator current.

Typically, the system provides for estimating the rotational speed of the motor using only the single-phase stator current input.

Typically, the system provides for estimating the fundamental frequency supplied to the motor using only the single-phase stator current input.

Typically, the system provides for EKF-based input signal conditioning for detection of weak failure modes of the induction motor.

Typically, the system provides for finding a search space consisting of multiple small frequency bands. This search space is formed using the supply frequency and slip information Typically, the system provides for spectral estimation for detection induction motor faults.

Typically, the system provides for detection and estimation of sinusoidal amplitude of closely spaced frequency components.

Typically, the system provides for finding spectral information in multiple specific small search bands for faster execution.

Typically, the system provides for normalization of the fault magnitude using the magnitude of the fundamental component.

According to this invention, there is provided a system for assessment of multiple faults in induction motors, said motor motor being supplied from a 3-phase supply (100) using a variable frequency drive (VFD), said system comprising:

a current sensor (102) for sensing a single phase stator current;

an anti-aliasing filter (105) for subsequent conversion of sensed current into digital format using an analog to digital converter (106), said acquired signal consisting of a high magnitude fundamental component, fault frequency components, and noise and

amplitude estimation block (107) for estimation of fundamental signal further conditioned with respect to input signal to said motor in order to obtain a conditioned input signal for estimation of slip and speed of said motor to further determine a fault search band correlative to said fault frequency components, said fault frequency components being estimated by means of a spectral estimator in order to determine if said motor comprises rotor faults, eccentric specific faults, bearing faults, and / or stator faults.

In at least an embodiment, said spectral estimator (110, 111) estimates spectrum of the conditioned stator current in the specified spectral band and sends it to a peak detector, said peak detector (112) finds the fault peak information form the spectrum, and an amplitude estimator (107, 113) estimates the amplitude of the fault peaks.

In at least an embodiment, said spectral estimator is configured to determine of rotor faults, in that, a broken rotor bar (BRB) and a broken end ring, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency (f) of the motor obtained by sensing motor current by means of a motor current sensor, slip (s) of the motor estimated by a slip estimator, and rotational frequency of the said motor by using the slip and fundamental frequency derived from the stator current.

In at least an embodiment, said spectral estimator is configured to determine eccentric specific faults, in that, a static eccentricity fault, a dynamic eccentricity fault, and a mixed eccentricity fault, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency (f) of the motor obtained by sensing motor current by means of a motor current sensor, sensing rotor slots (R) using an input mechanism associated with this system, rotational frequency ( ) of the motor by means of a rotational frequency sensor, slip (s) of the motor estimated by a slip estimator, number of bars of the motor input using an input mechanism associated with this system, number of pole-pairs (p) of the motor using an input mechanism associated with this system.

In at least an embodiment, said spectral estimator is configured to determine bearing faults, in that bearing faults, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency (f) of the motor by sensing motor current by means of a motor current sensor, sensing rotational frequency (k) by means of deployed rotational frequency sensor, and using an input mechanism configured to facilitate input of number of balls (N) of the motor, input of ball diameter ( ), input of ball pitch diameter ( ), and contact angle ( ) between the ball and the races.

In at least an embodiment, said spectral estimator is configured to determine of stator faults, in that stator faults, in terms of spectral signature information of the motor, is determined by computing fundamental supply frequency (f) of the motor, sensing slip (s) of the motor (estimated by a slip estimator), and using an input mechanism configured to input number of pole-pairs (p) of the motor and number of bars of the motor.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS

The invention will now be described in relation to the accompanying drawings, in which:

Figure 1 illustrates a schematic diagram of the steps involved in designing the fault detection system; Figure 2 illustrates a spectral estimation algorithm and its implementation; and

Figure 3 illustrates the implementation of the overall system on a cloud based server.

DETAILED DESCRIPTION OF THE ACCOMPANYING DRAWINGS:

According to this invention, there is provided a system for assessment of multiple faults in induction motors.

Figure 1 illustrates steps involved in operation of the fault detection system.

Figure 1 describes a non-invasive system for assessment of faults for a three phase squirrel cage induction motor (103). The motor is supplied from a 3- phase supply (100) using a variable frequency drive (VFD) (101). For assessment of faults, only a single phase stator current is sensed using a current sensor (102). For this purpose, a current-transformer type or a hall- effect type sensor can be used having an effective bandwidth of 0 Hz to 2000 Hz. The motor is assumed to be working against a load represented by

(104). The designed system is well-suited for handling motors running with different quantity of constant loads.

Once the current signal is sensed, it is passed through an anti-aliasing filter

(105) for subsequent conversion into digital format using an analog to digital converter (106). The acquired signal consists of a high magnitude fundamental component, fault frequency components, and noise. Presence of the high- amplitude fundamental component makes it difficult for the system to detect low-magnitude fault components.

Hence, the acquired data samples are sent to the fundamental frequency and amplitude estimation block (107) for estimation of the fundamental signal using extended- Kalman filter based method using a third order signal model, as shown in the non-patent document by A. Routray, A. K. Pradhan, and K. P. Rao,“A novel Kalman filter for frequency estimation of distorted signals in power systems,” IEEE Trans. Instrum. Meas., vol. 51, no. 3, pp. 469-479, Jun 2002.

Once the fundamental signal is estimated, it is deducted from the input signal using (108). The signal that remains after the fundamental cancellation is the conditioned input signal and is suitable for fault detection. The conditioned signal is used for estimation of slip and speed. Slip is estimated (109) by finding the mixed eccentricity fault component which is inherently present in all motors. Using the slip information, a fault search band is generated from (eq la) to (eq If) for the spectral estimator (110). This band is essential for efficient implementation of the spectral estimator (111). The spectral estimator estimates the spectrum of the conditioned stator current in the specified spectral band and sends it to the peak detector. Description of the spectral estimator is provided in Figure (2). The peak detector (112) finds the fault peak information form the spectrum, and the amplitude estimator (113) estimates the amplitude of the fault peaks. Severity of the faults can be assessed from the peak-amplitude information. A decision module (114) decides whether a particular fault has occurred and also quantifies the severity of the fault. The information thus gathered is conveyed through the display unit (115) along with additional information like speed and supply frequency.

A single search-band is used to find the peak-frequency of the mixed eccentricty fault component. The slip and speed are evaluated by (eq. 1)

Where s is the slip, p is the number of poles, f 0 is the fundamental frequency, and f mixeci is the peak frequency of the mixed eccentricity component. Using the slip information, a fault frequency search band is generated for the spectral estimator (110) For generating the fault search band, a 4 Hz band around the theoritical fault component obtained using the non-patent document S. Nandi, H. A. Toliyat, and X. Li, “Condition Monitoring and Fault Diagnosis of Electrical Motors A Review,” IEEE Trans. Energy Corners., vol. 20, no. 4, pp. 719-729, 2005 is used. This band is essential for efficient Implementation of the spectral estimator (111). The spectral estimator estimates the spectrum of the conditioned stator current in the specified spectral band and sends it to the peak detector. Description of the spectral estimator is provided in Figure (2). The peak detector (112) finds the fault peak locations form the spectrum, and the amplitude estimator

(113) estimates the amplitudes of the fault peaks. Severity of the faults can be assessed from the peak-amplitude information. The decision module

(114) compares the peak magnitudes with thresholds and their positions to decide whether a particular fault has occurred and also quantifies the severity of the fault. The information thus gathered is conveyed through the display unit (115) along with additional information like speed and supply frequency.

Figure 2 illustrates steps involved in operation of the spectral estimator.

Figure 2 describes the implementation of the spectral estimator that was used in Figure 1. The conditioned stator current signal is given as input (200) to the spectrum estimator. A frequency search space (201) is fixed and provided as input to the spectral estimator. From the input signal, a data matrix (202) is created and using the frequency search space a search manifold matrix is created. The structure of both these matrices are shown below:

er limit of the frequency search space. From the data matric the autocorrelation matrix (204) is formed using the relation: The autocorrelation matrix is pre-multiplied by the Hermitian transpose (205) of the manifold matrix followed by post-multiplication by the manifold matrix. The diagonal elements of the resultant matrix (206) are then extracted by (207). The extracted diagonal elements give the spectrum of the input signal in the specified search band.

Figure 3 illustrates implementation schematics of the cloud based 3 -phase induction motor incipient fault detection and failure prognosis system.

Figure 3 shows the overall implementation of the scheme for a cloud based assessment of faults for multiple motors. Here multiple induction motors (300) can transmit data using current transformer connected with a WiFi module (301). The transmitted data are received by a WiFi hub (302) and sent to the cloud server (303) for data processing. The output of the fault diagnostic module can be accessed using cell phone apps in the shop floor for further action. The TECHNICAL ADVANCEMENT of this invention lies in providing a system which uses only a single stator current as input in order to detect multiple faults. Additionally, the system is configured to detect faults under adverse running conditions like low loading. Additionally, in the system, the spectral estimator can detect closely spaced sinusoids; as a result, closely spaced fault components can be resolved. Furthermore, the spectral estimator, of this invention, can estimate the amplitude of the fault component accurately and hence the severity. Still additionally, the detectability of the closely spaced fault components is further improved by using the signal conditioning unit.

The TECHNICAL ADVANCEMENT of this invention further lies in providing a system which detect multiple weak faults in induction motor with conditions that are adverse to the detection of faults. The testing of faults have been carried out with different severity of faults wherever applicable. Moreover, an online system has been developed to establish the importance of such a system in industrial perspective. The system uses only a single phase stator current from the motor as input which can be acquired using a low-cost current sensor. The system can additionally estimate the speed, slip, and fundamental supply frequency provided to the motor. The system can also assess the severity of the fault that has occurred.

The ADVANTAGES of the system are enumerated below:

The invention provides a system that can monitor and estimate vital running information of a motor as follows: • The system can monitor the condition of induction motors in any industry and help in preventive maintenance. This reduces downtime and can help in planning in suitable maintenance.

• Use of single-phase stator current as input.

• Detect weak faults under different loading conditions

• Estimation of vital parameters of the motors.

According to a non-limiting exemplary embodiment, experiments were carried out with a 22 kW, four-pole, and three-phase delta connected induction motor from ABB. Power to the motor is supplied with an ABB make variable frequency drive. Experiments were carried out for detection of weak faults. Variable loading is achieved through rheostatic loads, with a 24 kW separately-excited DC generator coupled to the motor. Clamp-type Hall-effect sensors from Fluke (model: ilOlOs) were used to sense the current signal. Yokogawa DL850v oscilloscope was used for data acquisition, and the analysis was carried out with MATLAB. Data was acquired with a sampling rate of 20 kSamples/s. A higher sampling rate was chosen for creation of a standard database. Broken rotor bar and different types of bearing faults were seeded in the motor for the experiment, whereas, the eccentricity fault was found to be inherently present in one of the sample motors.

While this detailed description has disclosed certain specific embodiments for illustrative purposes, various modifications will be apparent to those skilled in the art which do not constitute departures from the spirit and scope of the invention as defined in the following claims, and it is to be distinctly understood that the foregoing descriptive matter is to be interpreted merely as illustrative of the invention and not as a limitation.