Search results

The purpose of this paper is to propose a statistical approach, which enables the prediction of the risk of the appearance of discharge bearing currents in a given drive that can damage the motor bearings. Using such an approach, the drives for the decisive but expensive, long‐term destructive investigations can be selected. A comparative analysis is presented to show the applicability of the presented statistical model.

In the paper, a statistical approach to large data sets of measured bearing currents is presented. Weibull distribution has been chosen to describe the amplitudes of bearing currents, whereas exponential distribution is used for the approximation of the awaiting times to puncture. In order to check the hypothesis that these distributions are independent two tests has been conducted: F‐Snedecor, to check equality of mean values and Bartlett's to check variance equality. On this basis, joint distribution is expressed as the product of independent marginal distributions.

The comparative analyses show that special caution is required for a proper selection of EMI filters. The additional inductance of the filter in a common mode (CM) current path decreases the damping factor of the circuit and significantly increases the risk of bearing damage caused by electric discharge machining currents. The CM transformer is a cheap and effective solution that reduces CM current without increasing the rate and the amplitudes of bearing currents.

This paper presents a new approach to bearing currents that allows prediction of the risk of bearing damage in a given drive.

The purpose of this paper is to evaluate the conditions in which a saturation of the common mode (CM) choke might appear to be essential for proper design of the CM voltage filters. This paper presents a method for the determination of a CM choke flux density produced by multilevel inverters with carrier‐based modulations.

The proposed combination of secant and tangent methods allows efficient and high‐resolution determination of the CM voltage waveforms produced at the output of the multilevel inverters with commonly used carrier‐based modulations.

The presented results show that the application of a five‐level inverter with specific modulation causes a decrease of the maximum flux density, down to 15 per cent of the maximum level of the flux density reached in a two‐level inverter. The proposed, dedicated approximation method provides an accuracy of the root estimation better by about three orders for a comparable number of the function calls in comparison with Brent's method.

The presented theoretical evaluations make possible the determination of the maximum expected value of the flux density produced by multilevel inverters with various types of carrier‐based modulations, which allows a reduction in dimensions, weight and cost of CM chokes applied in CM voltage compensators.

In the paper, the new formulas that describe the placement of triangular carrier functions for commonly used multilevel inverters have been presented. In order to avoid accumulation of the estimation error, during determination of the CM voltage time integral, a dedicated, efficient method of roots approximation has been developed.