Methods of detecting a winding fault in an electrical machine operated as a motor or operated as a generator

Abstract

A method of detecting a winding fault in an electrical machine comprises: applying a three-phase current to windings of the electrical machine; sampling 102 the three-phase instantaneous current of the electrical machine; calculating 105 a combined sequence current by multiplying the three phase instantaneous current with a permutation of the three phase instantaneous current, wherein each of the phase currents of the three-phase instantaneous current is multiplied with one of the permutated phase currents of the three-phase instantaneous current; extracting 106 a DC component from the combined sequence current; comparing 107 the DC component with a threshold value; and notifying 108 a winding fault if the DC component is larger than the threshold value. An alternative method applies a three-phase voltage rather than current and extracts a DC component of voltage to detect a fault (figure 6). The electrical machine may be a motor or a generator.

Description

Field of the disclosure The present disclosure relates to methods of detecting a winding fault in an electrical machine operated as a motor or operated as a generator. Background There is a growing trend to deliver power and propulsion products to emerging electric aircraft markets such as Urban Air Mobility (UAM) and Commuter Aircraft (CAP). An electrical machine is a critical component in a power and propulsion system which can be used as a propulsion motor or power generator. Unforeseen failures in the electrical machine may lead to loss of propulsion or power generation onboard and jeopardize the safety. Hence, it is important to improve the reliability and availability of electrical machines by an early detection of faults. One of the most common electrical faults in an electrical machine is a stator winding fault. A stator winding fault means that a fault with the insulation or integrity of the stator winding occurs. An example is a so called turn-to-turn fault which involves a short circuit between two turns of the same phase within a stator winding. Traditionally, a stator winding fault is detected by analysing the magnitude of stator winding fault characteristics frequency components in a three phase current, and by analysing the sequential components in three phase voltages and currents. The key disadvantages of such methods are the dependency on frequency spectrum analysers and a sluggishness in taking decisions owing to a delay introduced by a spectrum analyser. Therefore, the prior art is not suitable for online condition monitoring of a stator winding in mission critical applications in which a fast response on faults is required. There is a need to provide a method and apparatus of detecting winding faults in an electrical machine in a simple and efficient manner. Summary of the disclosure According to a first aspect, a method of detecting a winding fault in an electrical machine operated as a motor is provided. The method comprising the steps of: applying a three-phase current to windings of the electrical machine; sampling the three-phase instantaneous current of the electrical machine; calculating a combined sequence current by multiplying the three phase instantaneous current with a permutation of the three phase instantaneous current, wherein each of the phase currents of the three-phase instantaneous current is multiplied with one of the permutated phase currents of the three-phase instantaneous current; extracting a DC component from the combined sequence current; comparing the DC component with a threshold value; and notifying a winding fault if the DC component is larger than the threshold value. Aspects of the invention are thus based on the idea of detecting an unbalance in a three-phase current by implementing a novel calculation method referred to a combined sequence current calculation. The combined sequence current is calculated in that each of the phase currents of the three-phase instantaneous current is multiplied with one of the permutated phase currents of the three-phase instantaneous current. For example, when the three-phase instantaneous current has first, second and third phase currents ia, ib, ic, a permutation of that sequence is ib, ia, ic, wherein the first phase current ia is multiplied with ib, the second phase current ibis multiplied with ia, and the third phase current icis multiplied with ic. Under a healthy winding condition the calculated combined sequence current has only AC components. When a stator winding fault occurs in the electrical machine, the three phase current become unbalanced which introduces a DC component in the calculated combined sequence current, which can be extracted. This DC component is used as an indicator to detect the stator winding fault. Such detection may trigger a protective relay to protect the electrical machine from the winding failure. The method is able to detect a winding fault at a very early stage at which conventional approaches are unable to detect such fault. A further advantage associated with the present invention lies in that the method is simple and does not require any frequency analysis or a large memory. It can be easily implemented in hardware or software and is able to detect faults without noticeable delay. For example, the method may be implemented completely in hardware without the need of software, whereas conventional approaches require a processor, i.e., both software and hardware for execution. A still further advantage associated with the present invention lies in that the method can be implemented without requiring an additional sensor. The existing sensors in electrical drives or power generation system, i.e. voltage and current sensors, are sufficient to implement the method. Generally, the method can be used to detect imbalance in any three phase electrical system. In an embodiment, when multiplying each of the phase currents of the three-phase instantaneous current with one of the permutated phase currents of the three-phase instantaneous current a sample by sample multiplication is carried out, wherein for each phase current the samples are multiplied with the samples of the respective permutated phase current. In this respect, it is to be noted that when sampling a time-dependent signal, a number of samples of the signal is provided, wherein this number is dependent on the sampling frequency. Carrying out a sample by sample multiplication means that the corresponding samples of the different phase currents are multiplied, such as the first sample in a considered time interval of phase current ia is multiplied with the first sample of phase current ib, the second sample of phase current ia is multiplied with the second sample of phase current ib, etc. Further, in an embodiment, the three products of the sample by sample multiplication when multiplying each of the phase currents of the three-phase instantaneous current with one of the permutated phase current of the three-phase instantaneous current are added. For example, the products of the currents ia* ib, ib*ic, and ic*ic are added. Accordingly, the combined sequence current is calculated by a summation of each of the respective products. It is pointed out with a three phase current, there exists a total of six permutations. However, as the products are summed up, the number of permutations which lead to different results is reduced to three. The permutation used to calculate the combined sequence current may be any of the three possible permutations. Accordingly, one of three possible permutations of the three-phase instantaneous current may be used for multiplication, thereby arriving at calculating the combined sequence current by one of the following formulas: CSC (nT) = ia(nT) * ib(nT) + ib(nT) * ia(nT) + i^nT) * ic(nT) (1) CSC (nT) = ia(nT) * ia(nT) + ib(nT) * ic(nT) + ic(nT) * ib(nT) (2) CSC (nT) = ia(nT) * ic(nT) + ib(nT) * ib(nT) + ic(nT) * ia(nT) (3) wherein CSC is the combined sequence current, n is the sample number, T is the sample period, ia(nT) is the first phase current of the three-phase instantaneous current, ib(nT) is the second phase current of the three-phase instantaneous current, and ic(nT) is the third phase current of the three-phase instantaneous current. It is pointed out that the combined sequence current CSC is a discrete signal dependent on time, wherein each sample of the CSC signal is formed by the products and sums is a defined in formulas (1), (2) or (3). In an embodiment, the DC component is extracted from the combined sequence current by low pass filtering the combined sequence current. In another embodiment, the DC component is extracted from the combined sequence current by a discrete frequency transform (such as Fast Fourier, cosine, etc.) of the combined sequence current and by determining the magnitude of the DC component of the transform. In a further embodiment, the phase currents are normalized before calculating the combined sequence current to remove load dependent components. Such normalization may be implemented in that for each phase current the peak value is detected and the sample values are divided by the peak value. In another embodiment, normalization is implemented in that for each phase current a frequency transform is carried out and the fundamental magnitude is extracted, wherein the sample values for each phase current are divided by the respective fundamental magnitude. As mentioned before, the present invention may be implemented both in an electric motor and in an electrical generator. In motor operation, the electrical machine is driven by a three-phase power supply, wherein the voltage and current are controlled to control the motor speed and torque. In generator operation, the electrical machine is driven by a prime mover such as a gas turbine engine, wherein the windings generate voltage. Therefore, to detect a stator winding imbalance in a motor operation, a three-phase current is evaluated. When detecting a stator winding imbalance in a generator operation, on the other hand, a three-phase voltage is generated. Accordingly, when considering an electrical machine operated as a generator, the method discussed above is applied not to currents but to voltages. Therefore, in a second aspect of the present invention, a method is provided of detecting a winding fault in an electrical machine, the method comprising the steps of: applying a three-phase voltage to windings of the electrical machine; sampling the three-phase instantaneous voltage of the electrical machine; calculating a combined sequence voltage by multiplying the three phase instantaneous voltage with a permutation of the three phase instantaneous voltage, wherein each of the phase voltages of the three-phase instantaneous voltage is multiplied with one of the permutated phase voltages of the three-phase instantaneous voltage; extracting a DC component from the combined sequence voltage; comparing the DC component with a threshold value; and notifying a winding fault if the DC component is larger than the threshold value. As to the advantages of such method, reference is made to the comments above which similarly apply. Also, corresponding embodiments apply when considering voltage signals rather than current signals. Therefore, the embodiments discussed with respect to current signals apply similarly to voltage signals. For example, calculating the combined sequence voltage may be by one of the following formulas: CSV (nT) = Va(nT)* Vb(nT)+ Vb(nT)* Va(nT)+ Vc(nT)* Vc(nT) (4) CSV (nT) = Va(nT)* Va(nT)+ Vb(nT)* Vc(nT)+ Vc(nT)* Vb(nT) (5) CSV(nT) = Va(nT)* Vc(nT)+ Vb(nT)* Vb(nT)+ Vc(nT)* Va(nT) (6) wherein CSV is the combined sequence voltage, n is the sample number, T is the sample period, Va(nT) is the first phase voltage of the three-phase instantaneous voltage, Vb(nT) is the second phase voltage of the three-phase instantaneous voltage, and Vc(nT) is the third phase voltage of the three-phase instantaneous voltage. In a still further aspect of the present an apparatus for detecting a winding fault in an electrical machine operated as a motor is provided, wherein the apparatus comprises: a sampler configured to sample the three-phase instantaneous current of the electrical machine that comprises a winding; a calculation module configured to calculate a combined sequence current by multiplying the three phase instantaneous current with a permutation of the three phase instantaneous current, wherein each of the phase currents of the three-phase instantaneous current is multiplied with one of the permutated phase currents of the three-phase instantaneous current; a filtering module configured to extract a DC component from the combined sequence current; and a diagnosis module configured to compare the DC component with a threshold value, and notify a winding fault if the DC component is larger than the threshold value. In case of voltage signals, an apparatus for detecting a winding fault in an electrical machine is provided, wherein the apparatus comprises: a sampler configured to sample the three-phase instantaneous voltage of the electrical machine that comprises a winding; a calculation module configured to calculate a combined sequence voltage by multiplying the three phase instantaneous voltage with a permutation of the three phase instantaneous voltage, wherein each of the phase voltages of the three-phase instantaneous voltage is multiplied with one of the permutated phase voltages of the three-phase instantaneous voltage; a filtering module configured to extract a DC component from the combined sequence voltage; and a diagnosis module configured to compare the DC component with a threshold value, and notify a winding fault if the DC component is larger than the threshold value. The skilled person will appreciate that except where mutually exclusive, a feature or parameter described in relation to any one of the above aspects may be applied to any other aspect. Furthermore, except where mutually exclusive, any feature or parameter described herein may be applied to any aspect and / or combined with any other feature or parameter described herein. Brief description of the drawings Embodiments will now be described by way of example only, with reference to the Figures, in which: FIG. 1 is an embodiment of a method of detecting a winding fault in an electrical machine operated as a motor; FIG. 2 is a schematic depiction of a three-phase alternating current; FIG. 3 is a schematic depiction of a sampled phase current; FIG. 4 is a schematic depiction of a combined sequence current as provided by the method of FIG. 1; FIG. 5 is an example apparatus configured to detect a winding fault in an electrical machine operated as a motor; FIG. 6 is an embodiment of a method of detecting a winding fault in an electrical machine operated as a generator; FIG. 7 is an example apparatus configured to detect a winding fault in an electrical machine operated as a generator; and FIG. 8 is an example coil winding of a stator of an electrical machine. Detailed description There is a need to detect and diagnose winding faults or any winding imbalance in electrical machines at an early stage. Winding faults in electrical machine are typically initiated as inter-turn faults or turn-to-turn faults. FIG. 8 illustrates an example coil assembly in the form of a single-tooth coil 1 wound on a bobbin. Such single-tooth coils are used, for example, in stators of a permanent magnet synchronous motor, but may be used in electric generators as well. A winding wire 10 forms a plurality of windings 11, which together form the coil 1 and which each form straight longitudinal sections 12 as well as further sections 13 at their ends, which are wrapped around a winding head support. The coil 1 is contacted via contact ends 14, 15. The winding wire 10 can be formed as a stranded wire or as a solid wire. A turn-to-turn fault may occur when the insulation between adjacent windings 11 is compromised or damaged. As a result, the effective number of windings 11 decreases due to the short-circuiting of these adjacent turns. Such fault can expand to phase-to-phase or phase-to-ground faults. If the inter-turn or turn-to-turn fault is detected at early stage then the more severe faults such as phase-to-phase or phase-to-ground faults can be avoided. Of course, the stator winding shown in FIG. 8 is to be understood as an example only in order to illustrate potential winding faults. Winding fault may occur in any possible winding configuration, both in a stator and in a rotor. An inter-turn or turn-to-turn fault in stator or rotor winding causes imbalance in the three phase windings. By detecting the imbalance, the winding fault can be detected. The electrical machine in electric or hybrid electric aircraft power and propulsion systems are used as motor or generator. In motor operation, the electrical machine is driven by three phase power supply, and the voltage and current are controlled to control the motor speed and torque. In generator operation, the electrical machine is driven by a prime mover such as an engine and the windings generate voltage. Hence, it is proposed to detect a stator winding imbalance in motor operation by analysing three phase currents as will be discussed with respect to FIGS. 1 to 5 and in generator operation by analysing three phase voltages will be discussed with respect to FIGS. 6 and 7. A first embodiment of a method for detecting a winding fault is shown in FIG. 1. The situation is considered that a three-phase current is applied to the windings of an electrical machine which is operated as a motor. This situation is present at start 101. FIG. 2 shows schematically a three-phase alternating current. The three-phase alternating current comprises three phase currents ia(t), ib(t) and ic(t). The three-phase current may be supplied to the electric machine, e.g., by means of an inverter. In 102, the three-phase instantaneous electrical machine currents ia(t), ib(t) and ic(t) are each sampled. A standard analog-to-digital converter may be used to implement that task. FIG. 3 depicts schematically one of the phase currents, namely, phase current ia(t) after having been sampled. The sampling frequency is fs. The sample period is T = 1 / fs. For example, fora sampling frequency of 100 kHz the sample period T is equal to 10 ps. After sampling, the phase signal ia(t) is a discrete signal with samples at time intervals nT, wherein n is the sample number which is an integer increasing in time. The sampled phase current for phase “a” is ia(nT). In an analogous manner, the sampled phase currents “b” and “c” are represented as ib(t) = ib(nT), and ic(t) = ic(nT), respectively. Next, in step 103, the discrete phase currents are normalized. By normalizing the phase currents load component in the signals can be removed. Normalization leads to a peak magnitude of “1”. Normalization of the phase currents may be implemented in a plurality of manners. In an embodiment, normalization is conducted by peak detection. For every complete cycle of the phase “a” current sine waveform, the peak value is detected. Subsequently, each sample in the cycle is divided by the peak value. The same procedure is repeated for phases “b” and “c”. In another embodiment, normalization is conducted by a Fast Fourier Transform (FFT) method. FFT is performed of phase “a” current. Then, the fundamental magnitude is extracted and each sample in the waveform given as input to the FFT is divided by the fundamental magnitude. The procedure is repeated for phase “b” and “c” currents. The sequence of the three discrete, normalized phase currents is then permutated in step 104. With three phase currents, there total number of permutations is six, wherein any permutation can be chosen. In the embodiment of FIG. 1, the three discrete, normalized phase currents ia(t), ib(t) and ic(t) are permutated to arrive at the sequence ib(t), ia(t) and ic(t). Next, in step 105, a combined sequence current (CSC) is calculated by multiplying each of phase current with a permutated phase current and adding the three products. Thereby, the six possible permutations of step 103 are reduced to three (as pairs of two are identical). The calculation is carried out sample by sample. Depending on the permutation chosen in step 103, the CSC may be calculated by one of the following formulas: CSC (nT) = ia(nT)* ib(nT)+ ib(nT)* ia(nT)+ ic(nT)* ic(nT) (1) CSC (nT) = ia(nT)*ia(nT)+ ib(nT)* ic(nT) + ic(nT)* ib(nT) (2) CSC (nT) = ia(nT) * iJnT) + ib(nT) * ib(nT) + ic(nT) * iJnT) (3) While calculating the CSC, the calculation is carried out sample by sample as illustrated in FIG. 4. Accordingly, for each sample n a CSC value is calculated by one of the formulas (1), (2) and (3). FIG. 4 is an example of a CSC curve after a sample-by-sample calculation of the CSC. As can be seen, the positive areas are larger than the negative areas in the depicted embodiment, such that a DC component is present. Steps 104 and 105 may be implemented as one step. In step 106, the DC component of the CSC is extracted. Extracted the DC component of the CSC may be implemented by either by using a low pass filter or by taking a FFT of the CSC and then pick-up the DC component magnitude. In step 107, it is determined if the DC component extracted in step 106 has a value that is larger than a threshold value which indicates that the machine runs healthy. The closer to zero the value is, the healthier is the electrical machine. The DC component starts to increase when the imbalance in the winding increases due to winding fault. The magnitude of the DC component thus represents a fault indicator to detect winding faults such as turn-to-turn or inter-turn short circuits in stator windings. If the DC component is larger than the threshold value, a stator winding fault is notified. Such notification may be to a controller of the electric machine, or be signalled to a cockpit. Such notification may also indicate the severity of the fault, wherein the severity is high when the DC component is high. If the DC component is not larger than the threshold value, it is determined that the electric machine is healthy in step 109. Possibly, a fault is at the rotor side of the electrical machine. This may be detected and diagnosed in step 110. FIG. 5 shows an apparatus for detecting a winding fault in an electrical machine operated as a motor which is configured to implement the method of FIG. 1. However, it is pointed out that such implementation represents an example only. The method of FIG. 1 may be implemented in a plurality of manners, including solely as a hardware implementation or as a software implementation. FIG. 5 depicts an electric motor 31 which receives a three-phase alternating current from a power supply / variable frequency drive such as a DC battery in combination with a power inverter. The current phases ia, ib and ic are provided to the electrical motor. For example, the three-phase alternating current supplies three stator windings of an electrical motor. There are provided sensors 41-43 in the form of current transformers which measure the instantaneous three-phase currents of the electrical motor 3. The respective phase current values are provided to a sampler 51 of an apparatus 5. The sampler 51 may be implemented as an analog-to-digital converter. In sampler 51, the three signals are sampled with a predefined sampling rate. In addition, sampler 51 may normalize the three sampled phase currents in the manner discussed above. The apparatus 5 further comprises a calculation module 52 which is configured to receive the sampled and normalized three-phase instantaneous currents from sampler 51 and which is further configured to calculate a combined sequence current by multiplying the three phase instantaneous current with a permutation of the three phase instantaneous current, wherein each of the phase currents of the three-phase instantaneous current is multiplied with one of the permutated phase currents of the three-phase instantaneous current. This is implemented in the manner discussed with respect to the step 105 of FIG. 1. After the combined sequence current CSC has been calculated in calculation module 52, the signal is provided to a filtering module 53 in which the DC component of the CSC is extracted. The filtering module 53 may comprise a low-pass filter of an FFT module. The extracted DC component is then provided to a diagnosis module 54 which implements fault detection, diagnosis, and decision-making. In particular, diagnosis module 54 is configured to compare the DC component with a threshold value, and notify a stator winding fault if the DC component is larger than the threshold value. The output Fl of the diagnosis module 54 is a fault index. FIG. 6 shows a method for detecting a winding fault in the windings of an electrical machine which is operated as a generator. Therefore, different to the embodiments of FIGS. 1 to 5, voltage signals are now analysed. However, the above remarks and discussions regarding current signals similarly apply to voltage signals. For example, FIG. 2 could describe in the same manner a three-phase alternating voltage signal. The three-phase alternating voltage comprises three phase voltages Va(t), Vb(t) and Vc(t). In step 602, which corresponds to step 102 of FIG. 1, the three-phase instantaneous electrical machine phase voltages Va(t), Vb(t) and Vc(t) are each sampled. The sampled phase voltages are Va(nT), Vb(nT), and Vc(nT). In step 603, which corresponds to step 102 of FIG. 1, the discrete phase voltages are normalized. Normalization leads to a peak magnitude of “1”. The sequence of the three discrete, normalized voltages is then permutated in step 604, which corresponds to step 104 of FIG. 1. With three voltage currents, there total number of permutations is six, wherein any permutation can be chosen. In the embodiment of FIG. 1, the three discrete, normalized phase voltages Va(t), Vb(t) and Vc(t) are permutated to arrive at the sequence Vb(t), Va(t) and Vc(t). Next, in step 605, which corresponds to step 105 of FIG. 1, a combined sequence voltage (CSV) is calculated by multiplying each of phase voltages with a permutated phase voltage and adding the three products. Thereby, the six possible permutations of step 603 are reduced to three (as pairs of two are identical). The calculation is carried out sample by sample. Depending on the permutation chosen in step 603, the CSV may be calculated by one of the following formulas: CSV (nT) = Va(nT) * Vb(nT) + Vb(nT) * Va(nT) + Vc(nT) * Vc(nT) (4) CSV(nT) = Va(nT)* Va(nT)+ Vb(nT)* Vc(nT)+ Vc(nT)* Vb(nT) (5) CSV (nT) = Va(nT) * V^nT) + Vb(nT) * Vb(nT) + Vc(nT) * Va(nT) (6) While calculating the CSV, the calculation is carried out sample by sample similar as in FIG. 4. Accordingly, for each sample n a CSV value is calculated by one of the formulas (4), (5) and (6). In step 606, which corresponds to step 105 of FIG. 1, the DC component of the CSV is extracted. Extracted the DC component of the CSV may be implemented by either by using a low pass filter or by taking a FFT of the CSV and then pick-up the DC component magnitude. In step 607, which corresponds to step 107 of FIG. 1, it is determined if the DC component extracted in step 606 has a value that is larger than a threshold value which indicates that the machine runs healthy. The closer to zero the value is, the healthier is the electrical machine. The DC component starts to increase when the imbalance in the winding increases due to winding fault. The magnitude of the DC component thus represents a fault indicator to detect winding faults such as turn-to-turn or inter-turn short circuits in stator windings. If the DC component is larger than the threshold value, a stator winding fault is notified. Such notification may also indicate the severity of the fault, wherein the severity is high when the DC component is high. If the DC component is not larger than the threshold value, it is determined that the electric machine is healthy in step 609. Possibly, a fault is at the rotor side of the electrical machine. This may be detected and diagnosed in step 610. FIG. 7 shows an apparatus for detecting a winding fault in an electrical machine operated as a generator and configured to implement the method of FIG. 6. FIG. 7 depicts an electric generator 32 which provides a three-phase alternating voltage to a device 2. The current voltages Va, Vb and Vc are provided by the electrical generator 32. For example, the three-phase alternating voltage is provided by three stator windings of an electrical generator. There are provided voltage sensors 61-63 which measure the instantaneous three-phase voltages. The respective phase voltage values are provided to a sampler 55 of an apparatus 5. The sampler 55 may be implemented as an analog-to-digital converter. In sampler 55, the three signals are sampled with a predefined sampling rate. In addition, sampler 55 may normalize the three sampled phase voltages in the manner discussed above. The apparatus 5 further comprises a calculation module 52, a filtering module 53, and a diagnosis module 54, which correspond to the respective modules of FIG. 5, such reference is made in this respect to the description of FIG. 5. The output Fl of the diagnosis module 54 is a fault index. It should be understood that the above description is intended for illustrative purposes only, and is not intended to limit the scope of the present disclosure in any way. Also, those skilled in the art will appreciate that other aspects of the disclosure can be obtained from a study of the drawings, the disclosure and the appended claims. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. Various features of the various embodiments disclosed herein can be combined in different combinations to create new embodiments within the scope of the present disclosure. In particular, the disclosure extends to and includes all combinations and sub-combinations of one or more features described herein. Any ranges given herein include any and all specific values within the range and any and all sub-ranges within the given range.

Claims

1. A method of detecting a winding fault in an electrical machine operated as a motor, the method comprising the steps of:applying a three-phase current to windings of the electrical machine;sampling (102) the three-phase instantaneous current of the electrical machine;calculating (105) a combined sequence current by multiplying the three phase instantaneous current with a permutation of the three phase instantaneous current, wherein each of the phase currents of the three-phase instantaneous current is multiplied with one of the permutated phase currents of the three-phase instantaneous current;extracting (106) a DC component from the combined sequence current;comparing (107) the DC component with a threshold value; andnotifying (108) a winding fault if the DC component is larger than the threshold value.

2. The method of claim 1, wherein when multiplying each of the phase currents of the three-phase instantaneous current with one of the permutated phase currents of the three-phase instantaneous current a sample by sample multiplication is carried out, wherein for each phase current the samples are multiplied with the samples of the respective permutated phase current.

3. The method of claim 2, wherein the three products of the sample by sample multiplication when multiplying each of the phase currents of the three-phase instantaneous current with one of the permutated phase current of the three-phase instantaneous current are added.

4. The method of claim 3, wherein the one of three possible permutations of the three-phase instantaneous current is used for multiplication, thereby arriving at calculating the combined sequence current by one of the following formulas:CSC (nT) = ia(nT)* ib(nT)+ ib(nT)* ia(nT) + ic(nT)* ic(nl)CSC (nT) = ia(nT)* ia(nT)+ h(nT)* ic(nT) + ic(nT)* ib(nT)CSC (nT) = ia(nT)* ic(nT)+ ib(nT)* ib(nT)+ ic(nT)* ia(nT)whereinCSC is the combined sequence current,n is the sample number,T is the sample period,ia(nT) is the first phase current of the three-phase instantaneous current, tb(nT) is the second phase current of the three-phase instantaneous current, and i / nT) is the third phase current of the three-phase instantaneous current.

5. The method of any preceding claim, wherein a DC component is extracted from the combined sequence current by low pass filtering the combined sequence current.

6. The method of any one of claims 1 to 4, wherein a DC component is extracted from the combined sequence current by a discrete frequency transform of the combined sequence current and by determining the magnitude of the DC component of the transform.

7. The method of any preceding claim, wherein before calculating the combined sequence current the phase currents are normalized (103).

8. The method of claim 7, wherein the phase currents are normalized (103) in that for each phase current a peak value is detected and the sample values are divided by the peak value.

9. The method of claim 7, wherein the phase currents are normalized (103) in that for each phase current a frequency transform is carried out and the fundamental magnitude is extracted, wherein the sample values for each phase current are divided by the respective fundamental magnitude.

10. A method of detecting a winding fault in an electrical machine, the method comprising the steps of:applying a three-phase voltage to windings of the electrical machine;sampling (602) the three-phase instantaneous voltage of the electrical machine;calculating (605) a combined sequence voltage by multiplying the three phase instantaneous voltage with a permutation of the three phase instantaneous voltage, wherein each of the phase voltages of the three-phase instantaneous voltage is multiplied with one of the permutated phase voltages of the three-phase instantaneous voltage;extracting (606) a DC component from the combined sequence voltage;comparing (607) the DC component with a threshold value; andnotifying (608) a winding fault if the DC component is larger than the threshold value.

11. The method of claim 10, wherein when multiplying each of the phase voltages of the three-phase instantaneous voltage with one of the permutated phase voltages of the three-phase instantaneous voltage a sample by sample multiplication is carried out, wherein for each phase voltage the samples are multiplied with the samples of the respective permutated phase voltage.

12. The method of claim 11, wherein the three products of the sample by sample multiplication when multiplying each of the phase voltages of the three-phase instantaneous voltage with one of the permutated phase voltage of the three-phase instantaneous voltage are added.

13. The method of claim 12, wherein the one of three possible permutations of the three-phase instantaneous voltage is used for multiplication, thereby arriving at calculating the combined sequence voltage by one of the following formulas:CSV (nT) = Va(nT)* Vb(nT)+ Vb(nT)* Va(nT)+ Vc(nT)* Vc(nT)CSV(nT) = Va(nT)* Va(nT)+ Vb(nT)* Vc(nT)+ V^nT)* Vb(nT)CSV(nT) = Va(nT)* Vc(nT)+ Vb(nT)* Vb(nT)+ Vc(nT)* Va(nT)whereinCSV is the combined sequence voltage,n is the sample number,T is the sample period,Va(nT) is the first phase voltage of the three-phase instantaneous voltage,Vb(nT) is the second phase voltage of the three-phase instantaneous voltage, and Vc(nT) is the third phase voltage of the three-phase instantaneous voltage.

14. The method of any one of claims 10 to 13, wherein a DC component is extracted from the combined sequence voltage by low pass filtering the combined sequence voltage.

15. The method of any one of claims 10 to 13, wherein a DC component is extracted from the combined sequence voltage by a discrete frequency transform of the combined sequence voltage and by determining the magnitude of the DC component of the transform.

16. The method of any one of claims 10 to 15, wherein before calculating the combined sequence voltage the phase voltages are normalized (603), wherein the phase voltages are normalized in that: for each phase voltage a peak value is detected andthe sample values are divided by the peak value; or for each phase voltage a frequency transform is carried out and the fundamental magnitude is extracted, wherein the sample values for each phase voltage are divided by the respective fundamental magnitude.

17. An apparatus for detecting a winding fault in an electrical machine operated as a motor, the apparatus comprising:a sampler (51) configured to sample a three-phase instantaneous current of the electrical machine that comprises a winding;a calculation module (52) configured to calculate a combined sequence current by multiplying the three phase instantaneous current with a permutation of the three phase instantaneous current, wherein each of the phase currents of the three-phase instantaneous current is multiplied with one of the permutated phase currents of the three-phase instantaneous current;a filtering module (53) configured to extract a DC component from the combined sequence current; anda diagnosis module (54) configured to compare the DC component with a threshold value, and notify a winding fault if the DC component is larger than the threshold value.

18. An apparatus for detecting a winding fault in an electrical machine, the apparatus comprising:a sampler (55) configured to sample the three-phase instantaneous voltage of the electrical machine that comprises a winding;a calculation module (52) configured to calculate a combined sequence voltage by multiplying the three phase instantaneous voltage with a permutation of the three phase instantaneous voltage, wherein each of the phase voltages of the three-phase instantaneous voltage is multiplied with one of the permutated phase voltages of the three-phase instantaneous voltage;a filtering module (53) configured to extract a DC component from the combined sequence voltage; anda diagnosis module (54) configured to compare the DC component with a threshold value, and notify a winding fault if the DC component is larger than the threshold value.

19. The apparatus of claim 17 or 18, wherein the apparatus further comprises a normalizing module (51) normalizing the three-phase signal.The apparatus of any one of claims 17 to 19, wherein the filtering module (53) comprises a low pass filter.16

Images