Osman Bilgin

Automatic detection and classification of rotor cage faults in squirrel cage induction motor

The detection of broken rotor bars and broken end-ring in three-phase squirrel cage induction motors by means of improved decision structure. The structure consists of current signal analysis (CSA), Artificial Neural Network (ANN) and diagnosis algorithm. Effects of broken bars and end-ring on current signal and feature extraction are in the CSA. The rotor cage faults are classified by using ANN. And result matrixes of ANN are considered two different ways for diagnosis. Then the diagnoses are compared with each other. In this study six different rotor faults, which are one, two, three broken bars, bar with high resistance, broken end-ring and healthy rotor, are investigated. The effects of different rotor faults on current spectrum, in comparison with other fault conditions, are investigated by analyzing side-bands in current spectrum. To reduce bad effects of changing of distance between the side-band and main component on the detection and classification of the faults, the spectrum is achieved with low definition. Thus, the improved decision structure diagnoses faulted rotors with 100% accuracy and classified rotor faults 98.33% accuracy.

Design optimization of PMSM by particle swarm optimization and genetic algorithm

One of the electric power researches is the design optimization studies of permanent magnet synchronous motors. The main advantages of design optimizations of permanent magnet synchronous motors are to contribute to comfort, cost, and especially to energy savings. Although absence of rotor windings affects efficiencies of permanent magnet synchronous motors, stringent selection of values of geometrical design parameters affects the efficiency. Artificial intelligence techniques are satisfactory in choosing of design parameters of electric motors. This study aims to provide the design optimization of surface mounted permanent magnet synchronous motor thus. First of all geometrical design parameters of the motor were identified and then preliminary analytical design and design optimization by using genetic algorithm and particle swarm algorithm were studied. The obtained efficiency results were compared with each others and the results is satisfactory.

Effects of rotor faults in squirrel-cage induction motors on the torque-speed curve

This study presents analysis and effects of rotor faults in squirrel cage induction motor on the torque-speed curve. In the study, the used experimental data were obtained from 10 HP, 25 HP, 30 HP and 50HP submersible induction motors. The experiments were made during the motors start up under no-load condition. Torque and speed of the motors were sampled and saved for every starting. The motor speed was computed from these data by using Fast Fourier Transform (FFT) and Short Time Fourier Transform (STFT). The motor torque values were computed by using average of the torque data window which represents every STFT window. So, the speed- torque curve was occurred. In the experiments; (i) a rotor with one broken rotor bar, (ii) a rotor with two adjacent broken rotor bars, (iii) a rotor with three adjacent broken rotor bars, (iv) a rotor with a highly resistance rotor bar, (v) a rotor with a broken end-ring faults and (vi) healthy motor were used. Then the results were analyzed. The effects of rotor faults on the speed- torque curve were ranked according to size of the rotor faults. The study differentiates in two points from the literature. The first point is the method of obtaining of the torque-speed curve and the second point is rank of the faults effects according to size of the faults. Test results revealed that the size of the rotor fault is directly proportional to its effects on the curve. And broken end-ring fault was found to procedure similar effects to broken bar faults.

The Detection of Rotor Faults By Using Short Time Fourier Transform

In this paper an experimental study detecting of rotor faults in three-phase squirrel cage induction motors by means of short time Fourier transform (STFT) is presented. The frequency spectrum of motor line current is exploited for the detection. By obtaining a number of frequency spectrums from a current data with STFT and averaging these spectrums, faults are diagnosed instead of fast Fourier transform frequently applied at the detection of broken rotor faults in the literature. Five different faulted rotors are investigated. These faults are one bar with high resistance of the rotor, one broken bar of the rotor, two broken bars of the rotor, three broken bar of the rotor and broken end ring of the rotor. Artificial neural network is used for classification of faults. Test results show that this method increase the accuracy of the fault diagnose.

Efficiency Analysis of Submersible Induction Motor with Broken Rotor Bar

This study analyzes effects of squirrel cage faults on submersible induction motors efficiency at steady-state condition. There are a lot of studies about effects of the cage faults on motor performance. Especially, the effects of the cage faults on the motor parameters such as current, torque and speed are well known. Unlike the literature, cage fault effects on efficiency are analyzed in this study. Furthermore, fluctuations and mean value changes resulting from the rotor faults are ranked according to size of these faults. Healthy and five different faults were investigated by using 10, 25, 30 and 50 HP submersible induction motors in both simulations and experiments. Time stepping finite element method solution was used to compute motor quantities in the simulation. Good agreement was achieved between simulation and experimental results. The effects of rotor faults on motor efficiency were clearly ranked according to size of faults.

Comparison of stochastic optimization methods for design optimization of permanent magnet synchronous motor

This study presents design optimization of permanent magnet synchronous motor by using different artificial intelligence methods. For this purpose, three stochastic optimization methods—genetic algorithm, simulated annealing, and differential evolution—were used. Motor design parameters and efficiency results obtained by the artificial intelligence methods were compared with each other. The results were later checked by finite element analysis. Consequently, the motor efficiencies obtained from the algorithms have high accuracy. Approaches strategies of the artificial intelligence algorithms are quite sufficient and remarkable for design optimization of permanent magnet synchronous motor. The differential evolution is better and more reliable optimization method nevertheless.

Parameter determination of induction machines by hybrid genetic algorithms

In general, a genetic algorithm combined with other algorithms (e.g. tabu search, simulated annealing, etc.) is well known to be a powerful approach. In this paper, an efficient hybrid approach containing local search and genetic algorithms is presented. The purpose of the using local search mechanisms is to provide better the solution quality and to increase the convergence speed. It is demonstrated that the performance of the proposed algorithms is significantly better than the conventional genetic algorithm methods.

The effect of magnet temperature on speed, current and torque in PMSMs

In this study, the effect of magnet temperature on speed, current and torque in permanent magnet synchronous motor was investigated. First, the demagnetization curves of Neomag S 28VC class magnet related to different temperatures, which was used for MCS06C41 coded motor of Lenze Company, were approached. A mathematical expression was achieved to calculate the remanence value for any temperature using these curves and curve fitting method. Then, a simulation was made according to the field oriented vector control method in MATLAB/Simulink using real motor and driver parameters. The mathematical expression that gives the change of magnetic flux related to the temperature influence was included in the modelling. In this way, the change of the speed, current and torque for different magnet temperatures were made observable. Six simulations were made with 25°C increase from 25°C to 150°C to observe the effect of the temperature on the motor much better. The simulation results have clearly demonstrated the effects of the increase of the magnet temperature especially on the current and torque.

Detection of induction motors rotor faults by using negative selection algorithm based on Park's vector approach

The stator current signals of induction motors are often used to detect the broken rotor bar faults. However, the detection work becomes rather tedious owing to the fact that the fundamental supply frequency is much greater and quite close to the fault frequency. To get rid of this difficulty, the use of filters is often employed. Besides the fundamental supply frequency, the filter methods used tend to press on the fault frequency and necessitate a lot of calculations. Several studies on methods detecting rotor faults exist in the literature. Unlike literature, this study applies the Negative Selection Algorithm on training and testing schemes of the a, b, c stator phase currents subjected to Park transformation after changing them into the id and iq magnitudes. The use of Park transformation eliminates the need of filter designs and lessens calculation load for the algorithm whereas application of the Negative Selection Algorithm helps with detection of both normal and fault values in the system. Backed with the experimental results, the method proposed in this study has shown the effects, accuracy and applicability of the algorithm in detecting broken rotor bar faults in induction motors.

The detection of rotor faults in the manufacturing of submersible induction motor

In this study, rotor faults detection in submersible induction motors which is used at deep well submersible pumps is presented by analyzing stator current. In some production squirrel cage rotor bars are welded to end rings by argon welding. While the welding sometimes some bars are not connected to end rings ore bad connection have been occurred. This affects the motor performance. For not preventing the production speed motor tests should be made quickly. In this study practical results are taken from POLMOT factory which produce submersible induction motors. When the motor construction is finished its robustness is tested with no load test. Their stator current time frequency domain is made and its current spectrum is investigated. According to current spectrum analysis its fault and robustness is determined. For classification artificial neural network (ANN) is used. A decision mechanism that uses ANN result matrixes is occurred to detect faulted rotors.