Test device and method for testing rotating electrical machines
The test device and method using N-phase waveform generators efficiently diagnose faults in rotating electrical machines by stimulating and measuring electrical responses, reducing complexity and cost while effectively identifying component issues.
Patent Information
- Application Number
- PCT/EP2026/063786
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-12
- Filing Date
- 2026-05-12
- Publication Date
- 2026-11-19
AI Technical Summary
Existing test methods for rotating electrical machines are inefficient and complex, failing to effectively diagnose faults in components like the stator and rotor, leading to increased power losses and decreased efficiency.
A test device and method using N-phase waveform signal generators and measurement devices to stimulate electrical responses in rotating electrical machines, allowing for efficient fault diagnosis by measuring electrical parameters and generating indications of machine health.
The solution reduces the size and cost of testing equipment while enabling effective fault detection in rotating electrical machines, including motors and generators, by simulating both rotating and stationary magnetic fields without physical rotation of components.
Smart Images

Figure EP2026063786_19112026_PF_FP_ABST
Abstract
Description
[0001] TEST DEVICE AND METHOD FOR TESTING ROTATING ELECTRICAL MACHINES
[0002] Technical Field
[0003] The present invention relates to a test device and method for testing rotating electrical machines. Specifically, but not exclusively, the present invention relates to performing angular field tests of rotating electrical machines such as a motors or generators.
[0004] Background
[0005] Rotating electrical machines are devices that convert electrical energy to mechanical energy or mechanical energy to electrical energy through rotational motion. They are called "rotating" because their operation involves a rotating part, or “rotor”, and a stationary part, or “stator”. There are mainly two types of rotating electrical machines, electric motors and electric generators. Electric motors convert electrical energy into mechanical energy (rotation). Examples of electric motors include motors in fans, pumps, and electric cars. Electric Generators convert mechanical energy (rotation) into electrical energy. Examples of electric generators include power plant generators, wind turbines.
[0006] Rotating electrical machines also typically include magnetic cores, windings, air gaps, and bearings. The windings are coils of wire through which electrical energy is supplied or induced. The air gap provides a space between the stator and the rotor allowing rotation. The bearings are mechanical components that support the rotating shaft.
[0007] The internal components, such as the stator, the rotor, or other moving components of rotating electrical machines can develop faults and weaknesses. Considering the stator of a rotating electrical machine as one example, the stator core, is generally manufactured from thin insulated silicon-steel laminations. In larger machines these laminations may be interlocked with axial frame bars and clamped. Due to thermal, electrical, mechanical, or environmental stress the interlaminar insulation material may be subject to deterioration and damage, causing shorts between laminations.If laminations are shorted, an abnormal circulating interlaminar eddy current can be induced. The increase in eddy current can result in additional losses and decrease the efficiency of the machine. Similarly faults in windings of the stator or the rotor, particularly short circuiting, can cause excessive heating and reduced torque.
[0008] It is desirable to provide test equipment and testing methods which are able to efficiently test rotating electrical machines to diagnose faults and identify when repair or maintenance is required.
[0009] Summary
[0010] According to a first aspect of the present invention, there is provided a test device comprising: a set of N-l waveform signal generators; a set of measurement devices configured to measure electrical parameters; a set of terminals for connecting the set of waveform signal generators and the set of measurement devices to a respective winding of an N-phase rotating electrical machine, the N-phase rotating electrical machine being configured to be operated using N-phase power supply signals supplied to N windings; circuitry configured to: connect each of the set of N-l waveform signal generators to a respective one of the set of terminals; and connect the set of measurement devices to the set of terminals; and a controller configured to: operate the set of N-l waveform signal generators to drive N-l test signals in N-l of the N windings, the N-l test signals each being driven at a respective phase and amplitude such that the N-l test signals are adapted to stimulate an electrical response in the N-phase rotating machine, the stimulated response representing a response of the N-phase rotating machine to being driven by N-phase power supply signals; operate each of the set of measurement devices to measure electrical parameters across respective terminals, the electrical parameters being representative of a response to the N-l test signals; process the measured electrical parameters to determine one or more properties of the response to the N-l test signals; and generate an indication of the one or more properties of the response to the N-l test signals, wherein N is three or more.
[0011] In this way it becomes possible to test for faults in an N-phase rotating electrical machine such as a motor or generator using N-l waveform signal generators. This reduces the size, cost, and technical complexity of the test device while still enabling the efficient testing of the electrical response of the N-phase rotating machine. The testdevice may use the N-l waveform signal generators to perform various test including rotating magnetic field test and stationary magnetic field tests. While the examples described herein relate to the application of test signals to stator windings of an N-phase rotating electrical machine, it is to be appreciated that similar testing may be performed by driving test signals through rotor windings for rotating electrical machines that include wound rotors.
[0012] In some examples, the set of measurement devices comprise N-l voltage measuring devices and N-l current measuring devices, and wherein the circuitry is configured to connect a voltage measuring device to each of the terminals that are connected to a waveform signal generator and to enable a current measuring device to measure currents for each of the N-l windings that are connected to a waveform signal generator.
[0013] The voltage drop across the two terminals connected to waveform signal generators can be measured and used to determine properties of a response of the N-phase power supply to the test signals. In this way, it is possible to reduce the number of measurement devices needed in the test device.
[0014] In some examples, the set of measurement devices comprise N voltage measuring devices and N-l current measuring devices, and wherein the circuitry is configured to connect a voltage measuring device to each of the terminals, and to enable a current measuring device to measure currents for each of the N-l windings that are connected to a waveform signal generator.
[0015] In certain testing operations the parasitic impedances of test leads connecting the test device to the motor should be accounted for. In these cases, an additional voltmeter may be provided on the third terminal such that the voltage drop across the other two terminals can be determined relative to the voltage drop across the third terminal.
[0016] In some examples, the N-l test signals are configured to simulate three waveform signals, each simulated waveform signal being out of phase with the other simulated waveform signals whereby to generate a rotating magnetic field in a three-phase rotating electrical machine, and wherein the N-l test signals comprise: a first test signal representing a difference between a first simulated waveform signal and a secondsimulated waveform signal; and a second test signal representing a difference between a third simulated waveform signal and the second simulated waveform signal.
[0017] In this way, the two test signals are able to stimulate a response to a rotating magnetic field in the three-phase rotating electrical machine that would have been generated had three-phase power supply been signals been provided across all three sets of windings in the three-phase rotating electrical machine. This enables a rotating magnetic field test to be performed in which the response of the three-phase rotating electrical machine can be measured at various magnetic field angles as the magnetic field rotates in the three-phase rotating electrical machine.
[0018] In some examples, three simulated waveform signals have a common amplitude and wherein the first test signal and the second test signal are out of phase whereby to generate the rotating magnetic field in the three-phase rotating electrical machine.
[0019] In this way, the rotating magnetic field may be a circular or have a uniform magnitude throughout its rotation in the three-phase rotating electrical machine.
[0020] In some examples, the three simulated waveform signals are uniformly phase shifted whereby the first simulated waveform signal is out of phase by 120 degrees with respect to the second simulated waveform signal and the third simulated waveform signal, and the second simulated waveform signal is out of phase by 120 degrees with respect to the first simulated waveform signal and the third simulated waveform signal.
[0021] In some examples, the measured electrical parameters comprise current and voltage and wherein operating each of the set of measurement devices to measure electrical parameters across respective terminals comprises: obtaining voltage data representing voltage measurements over a first predetermined time period; and obtaining current data representing current measurements over a second predetermined time period, wherein processing the measured electrical parameters to determine one or more properties of the response to the N-l test signals comprises processing the voltage data and the current data to generate property data representing the response to the N-l test signals.
[0022] In this way, it becomes possible to measure the response of the three-phase rotating electrical machine as the rotating electrical magnetic field rotates. The current and voltage may be measured over different time periods. For example, the current may be measured over a longer time period than the voltage, and the current data may benormalised or averaged such that it can be compared over the same time period or phase angle as the voltage data. Similarly, the voltage may be measured over a different time period than the current and averaging or normalization may be applied.
[0023] In some examples, generating an indication of the one or more properties of the response to the N-l test signals comprises: processing the property data to identify if an anomaly in the one or more properties of the response over the predetermined time period; and generating an indication of either: an identified anomaly; or an absence of an anomaly.
[0024] In this way, the test device is able to provide an indication of the health or status of the N-phase rotating electrical machine under test. For example, the test device may generate an indication that an anomaly has been detected in the N-phase rotating electrical machine. Such as result may be represented as a simple pass / fail indication. In other examples, the indication may be a more complex representation of the properties of the response of the N-phase rotating electrical machine. For example, an indication of a specific phase angle in the rotating electrical machine at which anomalous behaviour is detected.
[0025] In some examples, the property data represents one or more varying properties in a stationary reference frame of the N-phase rotating electrical machine, and wherein generating the indication of the one or more properties of the response to the N-l test signals comprises transforming the property data to a rotating reference frame corresponding to the rotating magnetic field.
[0026] In this way, abnormal behaviour may be more easily detected by the test device. For example, in the stationary reference frame certain properties of the response of the N-phase rotating electrical machine may vary considerably over time and / or at different phase angles, making small deviations difficult to detect. In the rotating reference frame, the one or more properties may be expected to be substantially constant over time or at different phase angles, thereby making small deviations easier to detect.
[0027] In some examples, the one or more properties of the response comprise any one or more of: resistance; reactance; impedance; power; and a decay time associated with the response. In this way it is possible to evaluate the various properties of the response of the N-phase rotating electrical machine to the N-l test signals. For example, impedance at different phase angles of the magnetic field can be analysed to identifyfluctuations resulting from faults. Similarly, non-uniform power responses can be identified which may represent a fault causing power losses at one or more phase angles in the N-phase rotating electrical machine.
[0028] In some examples, the N-l test signals are configured to simulate three waveform signals having respective amplitudes, the three simulated waveform signals being in-phase whereby to generate a stationary magnetic field in a three-phase rotating electrical machine, and wherein the N-l test signals comprise: a first test signal representing a difference between a first simulated waveform signal and a second simulated waveform signal; and a second test signal representing a difference between a third simulated waveform signal and the second simulated waveform signal.
[0029] In this way, a stationary magnetic field may be generated in the motor at a specific phase angle in the N-phase rotating electrical machine to test the response of the N-phase rotating electrical machine at specific angular positions of the magnetic field.
[0030] In some examples, the N-l test signals are a first set of N-l test signals and the stationary magnetic field is generated at a first angle in the three-phase rotating machine and the controller is configured to: operate the set of N-l waveform signal generators to drive a second set of N-l test signals that are configured to simulate N-phase power supply signals whereby to generate a further stationary magnetic field at a second angle in the three-phase rotating machine; operate each of the set of measurement devices to measure further electrical parameters across respective terminals, the electrical parameters being representative of a further response to the second set of N-l test signals; process the measured electrical parameters to determine one or more properties of the further response to the N-l test signals; and generate an indication of the one or more properties of the further response to the second set of N-l test signals.
[0031] In this way, a scan of the response of the N-phase rotating electrical machine may be determined by obtaining measured electrical parameters for a plurality of different angular positions of the magnetic field.
[0032] In some examples, processing the measured electrical parameters to determine one or more properties of the response to the N-l test signals comprises: processing electrical parameters measured during a first time period to determine a first set of properties; processing electrical parameters measured during a second time period, thesecond time period being after the first time period, to determine a second set of properties; and determining the one or more properties of the response to the N-l test signals by correcting the first set of properties based on the second set of properties.
[0033] This may enable a correct to be applied to mitigate for resistive effects in the biphase rotating electrical machine.
[0034] In some examples, the set of N-l waveform signal generators are configured to drive the N-l test signals at predetermined voltages.
[0035] In some examples, the set of N-l waveform signal generators are configured to drive the N-l test signals at predetermined currents.
[0036] In some examples, the N-l test signals comprise alternating current signals. In some examples, the N-l test signals comprise direct current signals.
[0037] In some examples, the N-l test signals are waveform test signals and wherein a frequency of each of the N-l test signals corresponds to a nominal operating slip frequency of N-phase rotating machine.
[0038] According to a second aspect of the present disclosure there is provided, a method for a test device, the method comprising: providing N-l waveform signal generators; providing a set of measurement devices configured to measure electrical parameters; connecting each of the N-l waveform signal generators to a respective one of N windings of an N-phase rotating machine, the N-phase rotating electrical machine being configured to be operated using N-phase power supply signals provided to the N windings; operating the N-l waveform signal generators to drive N-l test signals in N-1 of the windings, the N-l test signals each being driven at a respective phase and amplitude such that the N-l test signals are adapted to stimulate an electrical response in the N-phase rotating machine, the stimulated response representing a response of the N-phase rotating machine to being driven by N-phase power supply signals; operating each of the set of measurement devices to measure electrical parameters representing a response of the N-phase rotating machine to the N-l test signals; processing the measured electrical parameters to determine one or more properties of the response to the N-l test signals; and generating an indication of the one or more properties of the response to the N-l test signals, wherein N is three or more.
[0039] In some examples, the method comprises connecting one of the N windings to a reference point.In some examples, the N-l test signals are configured to simulate three waveform signals, each simulated waveform signal being out of phase with the other simulated waveform signals whereby to generate a rotating magnetic field in the biphase rotating electrical machine, and wherein the N-l test signals comprise: a first test signal representing a difference between a first simulated waveform signal and a second simulated waveform signal; and a second test signal representing a difference between a third simulated waveform signal and the second simulated waveform signal.
[0040] In some examples, the three simulated waveform signals have a common amplitude and wherein the first test signal and the second test signal are out of phase whereby to generate the rotating magnetic field in the three-phase rotating electrical machine.
[0041] In some examples, the measured electrical parameters comprise current and voltage and wherein operating each of the set of measurement devices to measure electrical parameters comprises: obtaining voltage data representing voltage measurements over a predetermined time period; and obtaining current data representing current measurements over the predetermined time period, wherein processing the measured electrical parameters to determine one or more properties of the response to the N-l test signals comprises processing the voltage data and the current data representing the response to the N-l test signals over the predetermined time period, and wherein generating an indication of the one or more properties of the response to the N-l test signals comprises: processing the property data to identify an anomaly in one or more properties of the response over the predetermined time period; and if an anomaly is identified, generating an indication of the anomaly.
[0042] In some examples, the N-l test signals are configured to simulate three waveform signals having respective amplitudes, the three simulated waveform signals being in-phase whereby to generate a stationary magnetic field in a three-phase rotating electrical machine, and wherein the N-l test signals comprise: a first test signal representing a difference between a first simulated waveform signal and a second simulated waveform signal; and a second test signal representing a difference between a third simulated waveform signal and the second simulated waveform signal.
[0043] In some examples, the N-l test signals are a first set of N-l test signals and the stationary magnetic field is generated at a first angle in the N-phase rotating machineand the method comprises: operating the set of N-l waveform signal generators to drive a second set of N-l test signals that are configured to simulate N-phase power supply signals whereby to generate a further stationary magnetic field at a second angle in the three-phase rotating machine; operating each of the set of measurement devices to measure further electrical parameters across respective terminals, the electrical parameters being representative of a further response to the second set of N-l test signals; processing the measured electrical parameters to determine one or more properties of the further response to the N-l test signals; and generating an indication of the one or more properties of the further response to the second set of N-l test signals.
[0044] Further features and advantages of the invention will be apparent from the following description of exemplary embodiments of the invention, which are given by way of example only.
[0045] Brief Description of the Drawings
[0046] Figure 1 is a schematic diagram showing a three-phase rotating electrical machine according to examples.
[0047] Figure 2 is a schematic diagram showing test apparatus for testing a three-phase rotating electrical machine, the test apparatus comprising three waveform signal generators for stimulating respective phases in the three-phased rotating electrical machine.
[0048] Figure 3 is a schematic diagram showing a test device according to examples. Figure 4 is a flow chart of a test operation performed by the test device according to examples.
[0049] Figure 5A is a graph showing typical three-phase power supply signals configured to generate a rotating magnetic field in a three-phase rotating electrical machine according to examples.
[0050] Figure 5B is a schematic diagram showing a rotating magnetic field generated in a three-phase rotating electrical machine according to examples.
[0051] Figure 5C is a schematic diagram showing a position of a rotating magnetic field at three different times corresponding to the three-phase power supply signals.Figure 6 is a schematic diagram showing how two test signals may be configured to stimulate a three-phase rotating electrical machine as if driven by three power supply signals according to examples.
[0052] Figure 7 is a schematic diagram showing how three power supply signals may be configured to generate a stationary magnetic field in a three-phase rotating electrical machine according to examples.
[0053] Figure 8 is a schematic diagram showing how two test signals may be configured to stimulate a three-phase rotating electrical machine as if driven by three power supply signals to generate a stationary magnetic field according to examples.
[0054] Figure 9 is a flow chart showing how processing of measured electrical parameters may be performed according to examples.
[0055] Figure 10 is a schematic diagram showing a power response of a three-phase rotating electrical machine according to examples.
[0056] Figure 11 is a schematic diagram showing an expected power response of a three-phase rotating electrical machine represented in a stationary reference frame according to examples.
[0057] Figure 12 is a schematic diagram showing an expected power response of a three-phased rotating electrical machine represented in a rotating reference frame according to examples.
[0058] Figure 13 is a schematic diagram of a time decaying response to a stationary magnetic field in a three-phase rotating electrical machine according to examples.
[0059] Figure 14 is a schematic diagram of a test device configured to perform four wire sensing across different phases of a three-phase rotating machine according to examples.
[0060] Figure 15 is a flow chart showing a method for a test device according to examples.
[0061] Detailed Description
[0062] Faults in rotating electrical machines can arise due to a variety of different causes. For example, stator-rotor contact due to mechanical defects in the rotating machine, core overheating, defects in manufacturing, damage, or foreign materials introduced during manufacturing or maintenance, and abrasion due to vibration. Whileexamples of faults in a stator core are described above, it is to be appreciated that faults may be introduced in other components as well, for example, defects or abrasive wear in a rotor may also be the source of a fault leading to decreased efficiency in the rotating electrical machine.
[0063] In general, defects or damage in a stator or rotor increase power losses, thereby decreasing efficiency. Since increases in power losses can be a symptom of such faults, monitoring those power losses can be used to identify and diagnose certain faults.
[0064] In some cases, a rotating electrical machine can be tested by applying excitations to each of the set of windings in the electrical machine. The response to the excitations is determined, for example, by measuring, and the response is analysed to identify anomalous behaviour in the rotating electrical machine. This may involve inducing a magnetic field at a specific angle relative to the rotor, and measuring the power response in the rotating electrical machine. The test can be repeated at different angles with respect to the rotor to determine whether the power response in the rotating electrical machine varies at different angles. If the power response of the rotating electrical machine is different at specific angles this may be an indication of non-uniform behaviour due to, for example, a fault at a specific position of the stator.
[0065] Certain examples described herein aim to provide a method and test device for testing a rotating electrical machine. According to the examples described herein, an biphase rotating electrical machine can be tested by providing excitations to N-l of the windings. The response to the N-l excitations in the rotating electrical machine can be used to identify and / or diagnose faults. An N-phase rotating electrical machine is generally a rotating electrical machine having N number of phases in its electrical system. A common number of phases in a rotating electrical machine is three. Three-phase rotating electrical machines, such as motors, typically consume high-power electricity in three-phase circuits which generate rotating magnetic fields. In a three-phase rotating electrical machine, three sets of electrical windings are provided in the stator. During normal operations substantially sinusoidal electrical signals are excited in each of the three sets of windings, with the electrical signals being phase shifted at 120° apart. In some cases, each set of windings of a rotating electrical machine may be referred to as a “phase” of the rotating electrical machine. For a perfectly balanced, orideal, motor this creates a rotating magnetic field having a constant amplitude and rotating at a constant speed.
[0066] It should be noted that asynchronous motors (such as induction motors) have a rotating speed of the rotor smaller than the synchronous speed of the rotating magnetic field, referred to as the “slip”. For example, a 55 kW motor can have synchronous speed of 1500 rpm, but at the nominal load the rotor spins at 1480 rpm. This gives a difference (slip) of 20 rpm, which divided by 50 Hz means just 0.4 Hz - and this is the frequency of currents flowing in the rotor under nominal operating conditions of full power or full load. In induction-type generators the rotating speed of the rotor is typically greater than the synchronous speed of the rotating magnetic field. In such cases, the slip refers to the difference between the slower rotating speed of the magnetic field and the faster rotating speed of the rotor. For example, where the synchronous speed of the rotating magnetic field is 1500 rpm, the nominal load would be 1520 rpm, resulting in a slip of 20 rpm and therefore a similar nominal frequency slip as perceived by the rotor but in this case the slip refers to the difference between the slower rotating speed of the magnetic field speed compared to the faster rotating speed of the rotor.
[0067] Various testing protocols are described herein including rotating magnetic field tests and stationary field tests. A rotating magnetic field test, referred to simply as a “rotating field test”, may involve inducing a rotating magnetic field in the rotating electrical machine and measuring a response of the rotating electrical machine. By measuring the response of the rotating electrical machine over the rotation of the magnetic field it is possible to measure and analyse the response to the rotating electrical machine at various magnetic field angles and without having to repeat the test operation at various angles. It is to be appreciated that while the term “rotating field test” is used herein, during such a test protocol the rotor itself may be stationary and not rotate. This may be achieved, for example, by using test signals which supply a power that it is below a normal operating power of the rotating electrical machine.
[0068] A stationary field test, referred to simply as a “stationary test”, may involve inducing a stationary magnetic field in the rotating electrical machine at a specific angle. This process may then repeated by inducing the stationary magnetic field at a variety of different angles in the rotating electrical machine. The response of therotating electrical machine may be measured at each of these angles and analysed to identify anomalous behaviour in the rotating electrical machine.
[0069] Figure 1 shows an example of a three-phase motor 100 to which the methods described herein may be applied. The three-phase motor 100 comprises a stator core 102, a rotor 104, and three sets of stator windings < >x, (p2, and (p3. Each of the sets of stator windings < >x, (p2, and (p3are equally spaced apart around the stator core 102. The motor 100 is configured to be operated, or driven, using three-phase power supply signals supplied to the three stator windings
[0070]
[0071] (p2, and (p3. The excitations are typically phase shifted by 120° with respect to the other such that a rotating magnetic field is induced in the rotor 104 at a constant angular velocity causing the rotor 104 to rotate. It is to be appreciated that where the phase difference between the excitations provided to each of the set of stator windings
[0072]
[0073] (p2, and (p3is not uniform, the magnetic field may have an elliptical shape and / or a varying angular velocity in the motor 100.
[0074] Figure 2 shows an example of components that may be used to test a three-phase motor 100 having three stator windings
[0075]
[0076] (p2, and (p3. Figure 2 shows three sinusoidal voltage sources 202A, 202B, and 202C, three ammeters 204A, 204B, and 204C, and three voltmeters 206 A, 206B, and 206C. Each voltage source 202 A to 202C, ammeter 204A to 204C, and voltmeter 206A to 206C, are connected to a respective stator winding
[0077]
[0078] (p2, and (p3or “phase” of the motor 100 with a cable 208A to 208C. Each of the three stator windings
[0079]
[0080] (p2, and (p3is associated with a respective impedance Z15Z2, and Z3and each cable is similarly associated with a respective impedance ZC1, ZC2, and ZC3. The stator windings
[0081]
[0082] (p2, and (p3are connected in a Star configuration with Star point voltage V0. In practice the Star point voltage V0 may not be accessible or measurable and in some cases may be floating. It is to be appreciated that in other examples the motor 100 may be configured with a Delta configuration or other types of configurations.
[0083] A reference point 210 is also included in the circuit connecting the voltage sources 202 A to 202B with the stator windings
[0084]
[0085] (p2, and (p3. The reference point 210 represents local or internal ground of a test device that includes the voltage sources 202 A to 202C and the measurement devices 204 A to 206C. It has been found that this reference point can be moved to anywhere in the circuit as no current flows through it.When testing the three-phase motor 100 test signals may be driven by each of the voltage sources 202A to 202C to induce a response from the motor 100. The response of the motor 100 is measured using the ammeters 204 A to 204C and voltmeters 206 A to 206C. The measurements from the ammeters 204 A to 204C and the voltmeters 206A to 206C may then be used to determine whether the behaviour of the motor 100 is anomalous which may occur when one or more faults are present in the motor 100.
[0086] It has surprisingly been found that it is possible to produce the same phase-to-phase voltages and the same currents from the viewpoint of the load under test, in this case the motor 100, with two test signals rather than three as shown in Figure 2. Due to the closed circuit nature of the configuration of Figure 2, it is also noted that the currents Al to A3 must produce net zero, such that A3 = - Al - A2, as there is no alternative current path and because V0 is floating. As such, measuring A3 is not needed as the current A3 can be determined based on the measurements of Al and A2. Therefore, the ammeter 204C can be removed.
[0087] If the ammeter 204C is removed and the reference point 210 moved to its position, then the potential V3 will now be connected directly to the local ground and so its voltage will be expected to be zero. Hence that voltage V3 also does not need to be measured and the voltmeter 206C can be removed.
[0088] The third voltage source 202C can then also be removed by combining its function with the voltages provided by the other two voltage sources 202 A and 202B. We will now consider a situation in which the first voltage source 202A produces a voltage Va, the second voltage source 202B produces a voltage Vb, and the third voltage source 202C produces a voltage Vc. The amplitudes of each of the voltages Va, Vb, and Vc is equal to Vamp. The first voltage source 202A and the third voltage source 202C can be combined into a single voltage source producing a voltage V’ a with V’a = Va - Vc = V’amp* 3*sin(mt + 30°) where V’amp is greater than the amplitude Vamp. The second voltage source 202B may also be combined with the third voltage source into a single voltage source producing a voltage V’b with V’b = Vb - Vc = V’amp* 3*sin(mt + 90°). It is to be appreciated that cosine rather than sine may be used with a corresponding angular correction.By combining the function of the third voltage source 202C with the first voltage source 202A and the second voltage source 202B it becomes possible to stimulate the same response from the motor 100 as if three separate voltage sources 202A to 202C were used but while only stimulating two voltage signals in two of the stator windings <p and (p2.
[0089] Figure 3 shows a test device 300 according to examples of the present disclosure. The test device 300 comprises a set of two wave form signal generators 302A and 302B. In this example, the waveform signal generators 302A and 302B are alternating current voltage sources configured to produce a voltage that periodically reverses its polarity and changes magnitude over time. The test device 300 has a set of measurement devices 304 A, 304B, and 306 A, to 306C, including ammeters 304 A and 304B for measuring current, and voltmeters 306 A and 306B for measuring voltage. In some cases a third voltmeter 306C is also provided as will be described further below.
[0090] The test device 300 also includes a set of terminals 308A to 308C. The terminals 308 A to 308C provide paths to connect the test device 300 to the motor 100 for testing. Specifically, the terminals 308A and 308B enable the waveform signal generators 302A and 302B to be connected to respective stator windings
[0091]
[0092] and (p2in the motor 100. The terminals 308A and 308B also enable an ammeter 304A and 304B and voltmeter 306 A and 306B to be connected in the path with each of the waveform signal generators 302A and 302B and the respective stator windings <p and (p2. The third terminal 308C provides a path for connection the third stator winding3in the circuit with the two waveform signal generators 302 A and 302B and the local reference point or local ground in the testing device 300.
[0093] Circuitry 312 in the test device 300 is configured to connect the waveform signal generators 302A and 302B to the respective terminals 308A and 308B, and the measurement devices 304A to 306C to the respective terminals 308A and 308B. The circuitry 312 enables the waveform signal generators to be connected to the motor 100, when the motor 100 is connected to the terminals 308A to 308C. The circuitry also connects the third terminal 308C to the local ground 310 and with the waveform signal generators 302 A and 302B.The test device 300 also includes a controller 314 that is configured to perform one or more methods of testing the motor 100 which will be described further below with respect to Figures 4 to 14.
[0094] Figure 4 shows a general method of operation of the controller 314. According to this example, the controller 314 is configured to operate 402 the set of waveform signal generators 302 A and 302B to drive test signals in two of the stator windings <p and (p2- The two test signals are each driven at a respective phase and amplitude such that the two test signals are adapted to stimulate an electrical response in the motor 100, the stimulated response representing a response of the motor 100 to being driven by a three-phase power supply signal. The specific amplitudes and phases that are used in each of the test signals will depend on the type of testing method which is being performed. Various examples of the test signals will be described further below with respect to Figures 5 to 8.
[0095] The set of measurement devices 304 A to 306B are operated 404 to measure electrical parameters across respective terminals 308A and 308B. In this example, the electrical parameters include voltages and currents and are representative of the stimulated response of the motor 100 to the two test signals.
[0096] The measured electrical parameters are then processed 406 to determine one or more properties of the stimulated response to the two test signals. An indication of the one or more properties of the response is then generated 408. For example, if the one or more properties of the stimulated response indicate that the function of the motor is normal, or as expected, then the indication may reflect that the motor 100 is operating as normal. If the one or more properties of the response are representative of anomalous behaviour in the motor 100, then the indication may be representative of the anomalous behaviour. For example, the indication may specify that maintenance is required and / or specify a position in the motor 100 that is a likely to have developed a fault.
[0097] Rotating Field Test
[0098] In a first example, the test performed by the test device 300 is a rotating field test in which a rotating magnetic field is induced in the motor 100. Figure 5 A shows an example of three waveform signals 502A, 502B, and 502C which the two test signalsgenerated by the waveform signal generators 302A and 302B are adapted to simulate. The three simulated waveform signals 502A to 502C are each out of phase with the other emulated waveform signals 502A to 502C and thereby would generate a rotating magnetic field in the motor 100 if each of these simulated waveform signals were provided to the three stator windings < >i, (p2, and (p3. In this example, the first simulated waveform signal 502A would be provided to first stator winding
[0099]
[0100] the second simulated waveform signal 502B would be provided to the second stator winding (p2, and the third simulated waveform signal 502C would be provided to the third stator winding (p3.
[0101] In the example shown, the three simulated waveform signals 502A to 502C are uniformly phase shifted with respect to each other such that the phase difference between each waveform signal is the same. In this example, there are three simulated waveform signals and so the phase difference between each is 120° or 2TI / 3 to provide a uniform phase shift. It is to be appreciated that a uniform phase shift of such power supply signals can be generally expressed as 360° / N, where N is equal to the total number of power supply signals or phases in the motor 100.
[0102] Figure 5B shows a vector 504 indicating the direction of the magnetic field and an indication of its rotation 506 in the motor 100 with respect to time. Figure 5C illustrates several successive angles at which the magnetic field may be applied at different points in time t0, t1(t2as it rotates. The corresponding amplitudes of the three simulated waveform signals can also be seen in the graph of Figure 5A. In the example shown, the frequency of the emulated wave form signals is 1 Hz, but it is to be appreciated that the speed of rotation of the magnetic field is dictated by the frequency of the three emulated waveform signals 502A to 502C.
[0103] Figure 6 shows an example of the two test signals 602A and 602B that are used to simulate the three waveform signals 502A to 502C. The two test signals 602A and 602B are determined by subtracting a common one of the simulated waveform signals 502C from a respective one of the other waveform signals 502 A and 502B. For example, the first test signal 602A represents a difference between the first simulated waveform signal 502A and the third simulated waveform signal 502C. The second test signal 602B represents a difference between the second simulated waveform signal 502B and the third simulated waveform signal 502C. While the labels “first”, “second”,and “third”, are used to refer to the different waveform signals 502A to 502C herein, it is to be appreciated that the numbering of these waveform signals 502A to 502C may be different.
[0104] As shown in Figure 6, the test signals 602 A and 602B have a common amplitude and are out of phase by 60° such that a rotating magnetic field is generated in the motor 100. It is to be noted that the resulting test signals 602A and 602B have a higher amplitude of 1.732V compared to the amplitude of the simulated waveform signals 502 A to 502C, and have a different phase offset. The amplitude of the first test signal 602A may be determined as V’a = Va- Vc = V’amp* 3*sin(o>t + 30°), where V’ amp = 1.732V. The second test signal 602B may be determined as V’b = Vb - Vc = V’amp* 3*sin(mt + 90°). In this example the first test signal 602A will be connected to and drive the first stator winding
[0105]
[0106] and the second test signal 602B will be connected to and drive the second stator winding (p2. The third stator winding3will be connected to the third terminal 308C in the test device 300 and thereby connected to the local reference point or local ground 310.
[0107] Stationary Test
[0108] In another example, the test device 300 is configured to perform a stationary test on the motor 100 in which a stationary magnetic field is induced in the motor 100. Figure 7 shows an example of three waveform signals 702A to 702C which may be used to generate a stationary magnetic field 704 in the motor 100. The three waveform signals 702A to 702C each have respective amplitudes and are in-phase. The amplitudes of the three waveform signals 702A to 702C can be derived by taking the magnitudes of the respective waveform signals at a desired angle or offset from the rotating field case. As shown in Figure 7, the magnitudes 706A, 706C, and 706C of the three waveform signals 502A to 502C at a 45° position can be determined from the rotating field case. The three waveform signals for generating a stationary magnetic field are then produced with an amplitude corresponding to the magnitudes at the 45° position.
[0109] This may be expressed according to:
[0110] Vs,i = Vn * sin(a + 0)Where Vs,i represents the amplitude of the i-th power supply signals for the stationary magnetic field, Vn represents the magnitude of the n-th phase power supply signals from the rotating field case of Figure 5 ( e.g. Va, Vb, or Vc) and also shown in Figure 7, a represents the angular offset between the n-th phase power supply signal from the rotating field case, and 0 represents the angular position of the stationary magnetic field to be generated. As such the amplitude of the first power supply signal 702A in the stationary case can be determined using:
[0111] Vs, I = Va * sin(45) = 0.707
[0112] The amplitude of the second power supply signal 702B in the stationary case can be determined using:
[0113] Vs, 2 = Vb * sin(120 + 45) = -0.966
[0114] The amplitude of the third power supply signal 702C in the stationary case can be determined using:
[0115] Vs, 3 = Vc * sin(-120 + 45) = 0.259
[0116] This results, in the stationary case, as three phases having the same electrical angle, all being sinusoids with phase shift of 0°, but with amplitudes determined as above.
[0117] As in the rotating field test case, two test signals adapted to stimulate a response from the motor 100 as if the motor 100 was being driven using the three waveform signals 702A to 702C can be determined based on the differences between the three waveform 702A to 702C. Figure 8 shows an example in which two stationary field test signals 802A and 802B having amplitudes Vs’a and Vs’b are used to drive the first <p and second (p2stator windings respectively. In this example, the first test signal 802A represents a difference between the first simulated waveform signal 702A and the third simulated waveform signal 702C and the second test signal 802B represents a difference between the second simulated waveform signal 702B and the third simulated waveform signal 702C. Which is to say that Vs’a can be determined as:
[0118] Vs ’a = Vs, 1 - Vs, 3And Vs’b can be determined as:
[0119] Vs’b = Vs, 2 - Vs, 3
[0120] By stimulating the first stator winding <p using the first test signal 802A and the stimulating the second stator winding (p2using the second test signal 802B, while the third stator winding3is connected in the circuit but not directly stimulated using a waveform signal generator, it is possible to stimulate a response in the motor 100 as if the motor 100 was being driven using the three waveform signals 702A to 702C.
[0121] In the examples shown, magnetic field angle of 45° is generated in the motor 100. It is to be appreciated that other angles are possible, and the test signals 802A and 802B by using a different angular position 0 in the derivation of Vs,l, Vs, 2, and Vs, 3 illustrated above. In this way it is possible to measure the response of the motor 100 at various magnetic angles and without needing to rotate any physical components in the motor 100, such as the rotor 104. In order to produce information for a full 360° rotation the stationary test may be repeated at multiple times at a sufficiently small angular steps to identify faults presenting in specific angular ranges.
[0122] In the examples described above, the waveform signal generators 302 A and 302B are alternating current voltage sources. However, it is to be appreciated that alternative waveform signal generators are also possible. For example, the waveform signal generators 302 A and 302B may generate test signals using an inverter. In such a case, the test signals may be substantially alternating, having the general shape of the envelope being sinusoidal but also potentially including harmonics arising from a switch-mode power supply of an inverter rather than a perfectly smooth waveform. For example, the waveform signal generators 302 A and 302B may be configured to drive test signals at predetermined currents rather than voltage. The test signals may be alternative current signals or direct signals. For example, direct current signals may be used for the stationary test. In this case the power supplies for the waveform signal generators 302 A and 302B may be unipolar and combined with a reversing or change-over relay to achieve bipolar output.Data Analysis
[0123] In certain examples, the measured electrical parameters representing the stimulate response of the motor 100 include voltage and current. Figure 9 shows an example in which operating the set of measurement devices 304 A to 306C to measure electrical parameters across respective terminals comprises obtaining 902 voltage data 904 representing voltage measurements over a first predetermined time period. Current data 906 is also obtained 908 over a second predetermined time period. The voltage data 904 and the current data 906 are then processed 910 to generate property data 912 representing the stimulated response to the test signals 602A and 602B or 802A and 802B.
[0124] In the rotating field test, the voltage data 904 and the current data 906 may be generated based on a continuous measurement of the current and voltage across the terminals 308A and 308B while the magnetic field rotates in the motor 100.
[0125] In a first example, the first time period and the second time period are the same and may correspond to a frequency of the test signals 602 A and 602B. For example, where measurements are taken for an integer number of rotations of the magnetic field in the motor 100 the first and second time periods may correspond to M / f, where AT is equal to a number of full rotations of the magnetic field in the motor 100 and f is equal to the frequency of the test signals 602A and 602B.
[0126] In a second example, the first time period and the second time period may be different. For example, the first time period and the second time period may have a common duration but measured over different periods, the different periods being mutually exclusive, or partially overlapping. In other examples, the first time period and the second time periods may have a different duration. For example, one of the current or voltage may be measured over a longer duration and the resulting data may be normalized or averaged for comparison with the other data.
[0127] The property data 912 may represent various properties of the stimulated response of the motor 100. Examples of one or more properties that may be represented in the property data 912 include voltage, current, resistance, reactance, impedance, power, including real power, reactive power, apparent power, phase angle, a decay time associated with the response, and others.In an example shown in Figure 10, the one or more properties of the response comprise a power response in the motor 100. Figure 10 shows an example of a power response plotted over different magnetic field angles in the motor 100. Figure 10 shows a first power response 1002 representing the power response of a healthy, that is non-faulty, motor at a different magnetic field angles from 0 to IT in the motor 100. As shown in Figure 10, the power response of a healthy motor 100, when a uniform magnetic field is induced is constant at all angles of rotation. A second power response 1004 of a faulty motor 100 is also shown. When there is a fault in the motor 100 the power response may not be uniform at all angles of the rotating magnetic field. It is to be appreciated that while an example of a power response is shown here, such illustrations may similarly be generated for various quantities in the response of the motor 100 including resistance, reactance, and impedance. It may also be possible for the one or more properties to include three or more properties analysed in combination, for example, a three-dimensional representation of a power response against magnetic field angle and test signal phase could be generated to identify inconsistencies in the response to specific phase positions of the test signals 602A and 602B.
[0128] In the stationary test, the stimulation of the first winding <p and the second winding (p2, may be repeated in which additional test signals configured to generate a stationary magnetic field in the motor 100 at different angles are used for each iteration. In this way, it is possible to generate the current data 906 and the voltage data 904 by programmatically performing the testing at various magnetic field angles and accumulating the results.
[0129] Processing the measured electrical parameters to identify one or more properties of the response to the test signals 602A and 602B or 802A and 802B may involve any suitable data analysis techniques. For example, Fourier transformation may be applied to enable data analysis in different domains for example by applying a transformation of the measured electrical parameters from a time-domain to a frequency domain.
[0130] Other examples of processing techniques include Prony methods. The Prony method is a technique used to analyse and model signals, particularly those that can be represented as a sum of damped sinusoids or complex exponentials.
[0131] In some examples, the property data 912 represents one or more varying properties in a stationary reference frame of the motor 100. The one or more propertiesmay vary over a time period and / or over a magnetic field angle in the motor 100. Figure 11 shows an example of a power response 1102 represented by the property data 912 that is represented in a stationary reference frame. The power response 1102 represents a measurement of the power response at a specific position 1104 in the stator core 102. As shown in Figure 11, when represented in a stationary reference frame with respect to the motor 100, the power response has a large variation over time or magnetic field angle. This may make it difficult to identify anomalous behaviours exhibited as deflections or changes to the amplitude of the power response.
[0132] In the ordinary three-phase approach, the frame of reference is stationary, and the signals of the measured electrical parameters revolve around this frame, so there are substantially 3 sinusoids, shifted with respect to each other by 120°. These can be described by using the sin or cos functions, and with an arbitrary starting angle a as:
[0133] VI = Vamp * sin (9 + a)
[0134] V2 = Vamp * sin (9 - a + a)
[0135] V3 = Vamp * sin (9 + a + a)
[0136] where: VI, V2, V3 - phase voltages, Vamp - amplitude of voltage, 9 = co*t -instantaneous angular position, a = 129° - angular phase shift, a - angular starting point. Analogous equations can be written for purely sinusoidal signals (no distortion), for quantities such as current, power, and so forth.
[0137] In this example, generating 498 the indication of the one or more properties of the response to the test signals 692A and 692B or 892A and 892B, may involve transforming the property data 912 to a rotating reference frame corresponding to the rotation of the magnetic field. It is to be appreciated that any suitable transform may be used when applying a transform. For example, a rotating frame transformation, direct-quadrature-zero, Park-Clarke transformations, positive-negative sequence and so forth.
[0138] In the rotating frame of reference the axis pointing to the maximum of the positive sequence is typically known as the direct axis (d) and thus the apparent voltage associated with it is Vd, and the axis orthogonal to it is the quadrature axis (q) with the apparent voltage Vq. The zero-sequence voltage Vz quantifies fixed offset point arising from unbalanced signals, and in a three-wire three-phase system (with ungrounded star point) is typically zero, or it is intended to be zero (if the supply sinusoidal waveformshave no DC offset). For a perfectly balanced three -phase system the quadrature value is also zero (or infinity, as it might be the case for impedance).
[0139] The transformation from the stationary-frame signals VI, V2, V3 to the rotating-frame signals Vd,Vq,Vz involves additional trigonometric calculations, which can be summarised in the matrix notation as:
[0140] vdcos (0) cos (0 - 120°) cos (0 + 120°)' F^il = 2 / 3 sin (0) sin (0 - 120°) sin (0 + 120°)2
[0141] V . 0.5 0.5 0.53J
[0142] or equivalently, if the input waveforms were based on sin rather than cos:
[0143] VAsin (0) sin (0 - 120°) sin (0 + 120°) ’ F^il = 2 / 3 cos (0) cos (0 - 120°) cos (0 + 120°) ^2
[0144] Vz. 0.5 0.5 0.5
[0145]
[0146] in which the angular starting point & was omitted for brevity, but it can be added with an arbitrary value, as necessary to define the real starting point. Equivalent calculations are performed for the currents. And once the V and I quantities are known, then the one or more properties can be represented in the rotating reference frame. For example, for impedances:
[0147] Zd = Vd / Id
[0148] Zq = Vq / Iq
[0149] Zz = Vz / Iz
[0150] And for power:
[0151] Pd Vd Id
[0152] Pq Vq Iq
[0153] Pz = Vz X Iz
[0154] Figure 12 shows an example in which the effective sensor position 1204 is in a rotating reference frame. In this case, the resulting power response 1202, for a balanced motor 100 would be expected to be uniform throughout the various phase angles of the magnetic field generated in the motor 100.
[0155] In an ideal case the test signal 602A and 602B or 802A and 802B generation and acquisition of the measured electrical parameters acquisition should be synchronous, i.e. at each instant of synchronous time the quantities should be known precisely. This can be achieved by having several dedicated channels which perform the analogue-to-digital and digital-to-analogue conversion, so that they operate from the same synchronous clock, or at least with synchronised clocks. The data samplingwould be also synchronised between the different channels, by the virtue of employing several synchronised channels.
[0156] However, it is possible to perform non-synchronous data generation and acquisition, for example by using a multiplexer circuit, so that a single channel can mimic several channels. During such operation there will be an additional artificial delay between the signals, especially if the signals are varying in a sinusoidal manner, because a sample acquired after a small delay needed for multiplexing will have a slightly different value. Such operation is perfectly possible, and the delay in generation of the test signals 602A and 602B or 802A and 802B can be easily compensated out by using negative delay in the digital format, so that after generation the negative numerical delay and the actual physical delay will cancel each other out. Similarly for the acquisition of the measured electrical parameters, the delay between the multiplexed channels can be quantified or known, and the signals can be corrected by the virtue of point-to-point linear or non-linear interpolation so that the apparent delay is removed from the data, or at least minimised. Techniques such as Fourier transform can be applied to achieve higher-quality information about the global phase of the signal.
[0157] When the stationary testing is performed, additional processing may be used to correct for imbalanced resistive characteristics in the motor 100. For example, processing the measured electrical parameters to determine one or more properties of the response to may involve processing electrical parameters measured during a first time period to determine a first set of properties and processing electrical parameters measured during a second time period, the second time period being after the first time period, to determine a second set of properties. The one or more properties of the response to the test signals may then be determined by correcting the first set of properties based on the second set of properties.
[0158] MCSA-Like Testing
[0159] Motor Current Signature Analysis, or “MCSA”, is a live on-line test, in which motor 100 is energised from the normal grid connection. The motor 100 rotates and drives its normally connected mechanical load. The voltages and currents are measured (e.g. via current transformers, or other safe sensors), and the harmonics spectrum is used to detect small-level harmonics. If a rotor bar is cracked then side-band harmonicsare generated due to intermodulation, and their amplitude increases with the level of damage. The change in dB (decibels) is typically used for such detection.
[0160] The MCSA method is more sensitive the more the load on the motor 100 is closer to nominal. At lighter load the modulation is reduced, and at no load there is no detectable modulation, because the slip frequency reduces almost to zero and the harmonics are too small to detect reliably. Because of this, the MCSA test has been found to not work very well with light load, with operation below 10 % of load becoming impossible to discern the harmonics, simply due to the limits of resolution and noise floor, but also because of the harmonic resolution (due to the slip tending to zero, so there is no harmonic separation).
[0161] Performing the rotating and / or stationary tests as described above has been found to have a surprising benefit in that MCSA-like testing can be performed without requiring the motor 100 to be driven at its nominal operating load. These tests enable the driving of the stator windings at the nominal operating slip frequency (e.g. below 1 Hz), such that MCSA-like testing can be performed. For example, with the slip frequency being at its nominal (e.g. 0.4 Hz under full load, or in general < 2 Hz for most motors, and < 1 Hz for larger motors). Therefore, it is possible to achieve accurate MCSA-like testing using test procedures in which the rotor 104 itself is kept stationary. This is because the motor can be switched off (offline test), the rotor can be stationary, yet it is possible to simulate a stable power supply at the nominal frequency as perceived by the rotor during normal operating conditions.
[0162] This approach is likely to be less sensitive to the damage in laminations (increased power loss happens only at higher excitation level, especially for test at higher frequencies, e.g. 250 Hz), but it is still able to detect motor problems such as cracked bars, which is valuable in real-world motor testing. This is because it can be done without disassembling the motor 100, or even disconnecting it from its mechanical load, which is a costly and difficult operation). In such an example, test signals 602A to 602B or 802A and 802B are waveform test signals having a frequency that corresponds to a nominal operating slip frequency of the motor 100.
[0163] The MCSA method presents some challenges, because the additional current undulations caused by the cyclically varying the load (e.g. blades of a pump comprising the motor 100) will add to the current signature of the motor itself, and telling theseapart is not always possible. Additionally, if the rotor load is not constant, then MCSA cannot be applied in a straightforward manner, because the slip frequency will vary and so will the sideband frequency positions, again making the analysis difficult or impossible. Variation in grid frequency and harmonics introduce further complications.
[0164] Applying an MCSA-like test using the rotating or stationary field test described above mitigates these problems because the rotor is stationary and so there is no influence of the mechanical load at all. The test is performed offline, and it is possible to exert full control over the shape of the waveforms of the test signals 602 A and 602B or 802A and 802B, and so the mains harmonics can be ignored. By definition, there is also no noise from the grid or from the load, and thus it is possible to scan the motor 100 both in temporal sense (rotating field) as well as in the angular sense (stationary field at multiple angles).
[0165] It is to be appreciated that MCSA-like testing may involve applying Fourier transformations to the measured electrical parameters to determine current spectrum against frequency measurements of the response of the N-phase rotating electrical machine 100.
[0166] Dual Imbalance Testing
[0167] Conventional tests of resistive imbalance between conduction paths, e.g. stator windings, of rotating electrical machines such as motors are carried out by applying a test current to each conduction path in turn, and determining resistance by measuring average voltage and current values for the period that the test current is applied. The results have been found to be useful for detecting faults in the windings of the rotating electrical machine, but the accuracy of the measurements of resistance may be affected by the presence of the rotor.
[0168] It has been found that currents and voltages measured in the conduction paths through the stator of an electrical machine may be affected by voltages induced in the rotor, which are due to a change in the test current applied to the stator. As a result, conventional measurements of resistive balance, carried out with the rotor in place, are influenced by a combination of resistive and induced voltage effects. This makes it difficult to diagnose problems with the electrical machine, because the test cannotdistinguish between purely resistive and electromagnetic effects, each of which may be caused by a different fault.
[0169] Providing a delay period from the start of providing the test current allows the measurements to be made without being influenced by a combination of resistive and induced voltage effects. After the delay period, the effects of resistive and / or induced voltage effects may be measured separately. A dual imbalance (DI) test may be performed, in a first case, in which the test current is provided during the time window for the measurements, a measure of imbalance of resistance may be made after the delay period with reduced effects from induced voltages, which were initially induced when the test current was turned on. In a second case, in which the test current is stopped before the time window for measurements, a measure imbalance of induced voltage may be made with reduced resistive effects, because the test current is not flowing during the time window, and so a voltage is not generated by the test current flowing through the conductive path.
[0170] It has been found that the results of resistive imbalance measured after the delay period may be used to correct measurements of impedance imbalance measured in an initial period after the application of the test current. This correction may remove or reduce the effects of resistive imbalance, leaving predominantly the effects of imbalance caused by differences in induced voltage in the rotor. As a result, two measures of imbalance may be generated, which can be used to diagnose different faults in the electrical machine. The resistive imbalance, which may be referred to as static imbalance, may be used to diagnose resistive faults in the stators or connecting cables, such as corroded connectors, loose bolts or damaged conductors. The corrected impedance imbalance measured in the initial period after application of the test current, which may be referred to as dynamic imbalance, may be used to diagnose electromagnetic faults, such as faulty gaps in the magnetic circuit between the stator and the rotor, as well as faults in the rotor.
[0171] However, this type of dual imbalance test does not allow for high-resolution angular scanning through 360° of a rotor 104 in the motor 100. because it performs just 3 tests, so it gives a very limited view of one point every 120° (electrically). In theory, it could be possible also to short-circuit two terminals A and B together, and thus perform the test as A-B||C, thus provide 60° resolution. However, the outcome of suchtest would be uncertain, because it is not possible to guarantee that the two phases would share the current precisely equally, due to imperfect, and in general unequal, contact resistance or lead impedance. In any case, measurement A-B (with C open) could return a completely different number than A-B||C and therefore would not be directly comparable to the former case.
[0172] It has been found that the use of two waveform signal generators 302 A and 302B, specifically current sources, could be used to facilitate dual imbalance testing. Their amplitudes could be specified by similar means as illustrated in Figure 7, where the trigonometric calculations are used to derive the required amplitudes to obtain the required effective angle of the axis - but of course using the two-phase waveforms 802A and 802B as shown Figure 8.
[0173] The output of the DI sub-test is a “trajectory” as shown in Figure 13, which is obtained with a DC “step” of the current source, applied quickly and held at a constant amplitude for the whole duration of this sub-test. The trajectories of Figure 13 are recorded whilst energised with a single current source. But if two current sources were combined in a two-phase system, similar to that shown in Figure 3, then the measured phasic voltages can be also processed in a vectorial manner, and thus it would be possible to derive the equivalent dual imbalance trajectory at an arbitrary angle, allowing calculation of the static and dynamic imbalance, for each angular position, but otherwise in a similar way as in the dual imbalance method.
[0174] However, for accurate measurement of motor response with the dual imbalance test, it will be actually required in practice to utilise Kelvin sensing, which will mean measuring 3 voltages (VI, V2, V3), shown in Figure 14. In the example of Figure 14, additional wires are required between these voltage measuring points and the tester to measure these remote voltages, not shown in Fig. 14. This is because there are considerable parasitic impedances of the test leads Zll, Z12, Z13, which are unavoidable between the tester and the load. The current does not need to be measured in the third lead, but the three measured potentials / voltages (VI, V2, V3) will allow bypassing the voltage drops across the test leads and thus precisely measuring the response of the load only, whether connected by some additional cable or measured directly at the motor terminals.Based on the implementation of the test device 300 shown in Figure 14, a further testing operation is possible. Rather than performing stationary test at each separate angular position discussed above with respect to Figures 7 and 8, the motor 100 can be tested by applying a slowly-varying AC signal (sinusoidal) and thus deriving the response for the full 360° rotation from a single test. However, in the rotating mode it is not possible to obtain the static imbalance value, because the individual currents will change sinusoidally, and thus never reach proper DC steady state. It is therefore envisaged that a single DC test could be performed first for deriving the static resistive imbalance, and the resistive “offsets” contributed to by DC resistance of each cable / winding. And later the rotating field test would be applied, for which the static correction could be added / subtracted, and thus used to derive dynamic imbalances neatly in one rotation through 360°.
[0175] It is to be noted that in implementations of the test device according to Figure 14, the voltages provided by the waveform signal generators 302 A and 302B are defined differently, because they are measured with respect to the local ground 310 of the tester 300, which is at a different, and varying, potential than the hypothetical neutral point of the load. Therefore, the equation needs to be modified to comprise the actually measured signals, if expressed by sines, as:
[0176] Vdcos (0) cos (0 — 120°) cos (0 + 120°)l 1% - V2‘ Vq= 2 / 3 sin (0) sin (0 - 120°) sin (0 + 120°) V2- V3z 0.5 0.5 0.5
[0177]
[0178] Lk3-vj
[0179] And as an example for the actually measured current values:
[0180] 'Id sin (0) sin (0 - 120°) sin (0 + 120°) ’ r / i - / 2iIq = 2 / 3 cos (0) cos (0 - 120°) cos (0 + 120°) I2 ~ I3 JZ. . 0.5 0.5 0.5 -13 ~ Ii-
[0181] It is to be appreciated that the testing methods described herein, including for example, the rotating field tests, the MCSA-like testing, and the dual imbalance applications may also be implemented using the apparatus shown in Figure 2. Specifically, these testing operations may be performed using three waveform signal generators 202A to 202C used to stimulate three test signals in the rotating electrical machine 100. In such a case, each of the waveform signal generators 202 A to 202Cwould be configured to provide a respective test signal to a respective one of the stator windings < >x, (p2, and (p3.
[0182] While the above description presents examples of a three-phase motor 100, it is to be appreciated that the techniques described herein may be generalised to N-phase rotating electrical machines including motors or generators, where N is equal to three or more.
[0183] Additionally, the waveform signals that are simulated by the test signals 602A and 602B or 802A and 802B are shown to be uniformly phase shifted and having a common amplitude. This is the case where a uniform magnetic field is to be generated in the motor 100. It is to be appreciated that in other examples, the magnetic field to be generated in the motor 100 may not be uniform at all angular positions and / or may rotate with a varying angular velocity. In such cases, the simulated waveform signals may have different amplitudes and / or be non-uniformly phase shifted. In such cases, the processing of the measured electrical parameters may be modified to correct for changes the magnitude and / or angular velocity of the magnetic field.
[0184] Figure 15 shows a method 1500 for the test device 300. The method 1500 comprises providing 1502 N-l waveform signal generators 302A and 302B. A set of measurement devices 304A to 306B configured to measure electrical parameters are also provided 1504. Each of the N-l waveform signal generators 302A and 302B are connected 1506 to a respective one of N stator windings of an N-phase rotating machine, the N-phase rotating electrical machine being configured to be operated using N-phase power supply signals provided to the N stator windings.
[0185] The N-l waveform signal generators 302 A and 302B are then operated 1508 to drive N-l test signals in N-l of the stator windings, the N-l test signals each being driven at a respective phase and amplitude such that the N-l test signals are adapted to stimulate an electrical response in the N-phase rotating machine, the stimulated response representing a response of the N-phase rotating machine to being by driven by N-phase power supply signals. Each of the set of measurement devices 304A to 306C are then operated 1510 to measure electrical parameters representing a response of the N-phase rotating machine to the N-l test signals. The measured electrical parameters are then processed 1512 to determine one or more properties of the response to the N-1 test signals and an indication of the one or more properties of the response to the N-l test signals is then generated 1514.
[0186] In some examples, an Nth stator winding3may be connected to the third terminal 308C of the test device 300 to provide a connection to the local reference or local ground 310 in the test device 300. It is to be appreciated that the variations and combinations of features described above with respect to the test device 300 will similarly apply to the method 1500.
[0187] The above embodiments are to be understood as illustrative examples of the invention. Further embodiments of the invention are envisaged. For example, while the preceding description provides examples of testing in which the waveform signal generators are connected to stator windings of a motor, it is to be appreciated that similar testing may also be performed by driving test signals through rotor windings for N-phase rotating electrical machines that include wound rotors. It is to be understood that any feature described in relation to any one embodiment may be used alone, or in combination with other features described, and may also be used in combination with one or more features of any other of the embodiments, or any combination of any other of the embodiments. Furthermore, equivalents and modifications not described above may also be employed without departing from the scope of the invention, which is defined in the accompanying claims.
Claims
CLAIMS1. A test device comprising:a set of N-l waveform signal generators;a set of measurement devices configured to measure electrical parameters; a set of terminals for connecting the set of waveform signal generators and the set of measurement devices to a respective winding of an N-phase rotating electrical machine, the N-phase rotating electrical machine being configured to be operated using N-phase power supply signals supplied to N windings;circuitry configured to:connect each of the set of N-l waveform signal generators to a respective one of the set of terminals; andconnect the set of measurement devices to the set of terminals; anda controller configured to:operate the set of N-l waveform signal generators to drive N-l test signals in N-l of the N windings, the N-l test signals each being driven at a respective phase and amplitude such that the N-l test signals are adapted to stimulate an electrical response in the N-phase rotating machine, the stimulated response representing a response of the N-phase rotating machine to being driven by N-phase power supply signals; operate each of the set of measurement devices to measure electrical parameters across respective terminals, the electrical parameters being representative of a response to the N-l test signals;process the measured electrical parameters to determine one or more properties of the response to the N-l test signals; and generate an indication of the one or more properties of the response to the N-l test signals,wherein N is three or more.
2. The test device of claim 1, wherein the set of measurement devices comprise N-1 voltage measuring devices and N-l current measuring devices, and wherein thecircuitry is configured to connect a voltage measuring device to each of the terminals that are connected to a waveform signal generator, and to enable the current measuring devices to measure currents for each of the N-l windings that are connected to a waveform signal generator.
3. The test device of claim 1 or claim 2, wherein the set of measurement devices comprise N voltage measuring devices and N-l current measuring devices, and wherein the circuitry is configured to connect a voltage measuring device to each of the terminals, and to enable a current measuring device to measure currents for each of the N-l windings that are connected to a waveform signal generator.
4. The test device of any preceding claim, wherein the N-l test signals are configured to simulate three waveform signals, each simulated waveform signal being out of phase with the other simulated waveform signals whereby to generate a rotating magnetic field in a three-phase rotating electrical machine, and wherein the N-l test signals comprise:a first test signal representing a difference between a first simulated waveform signal and a second simulated waveform signal; anda second test signal representing a difference between a third simulated waveform signal and the second simulated waveform signal.
5. The test device of claim 4, wherein the three simulated waveform signals have a common amplitude and wherein the first test signal and the second test signal are out of phase whereby to generate the rotating magnetic field in the three-phase rotating electrical machine.
6. The test device of claim 4 or claim 5, wherein the three simulated waveform signals are uniformly phase shifted whereby the first simulated waveform signal is out of phase by 120 degrees with respect to the second simulated waveform signal and the third simulated waveform signal, and the second simulated waveform signal is out of phase by 120 degrees with respect to the first simulated waveform signal and the third simulated waveform signal.
7. The test device of any preceding claim, wherein the measured electrical parameters comprise current and voltage and wherein operating each of the set of measurement devices to measure electrical parameters across respective terminals comprises:obtaining voltage data representing voltage measurements over a first predetermined time period; andobtaining current data representing current measurements over a second predetermined time period,wherein processing the measured electrical parameters to determine one or more properties of the response to the N-l test signals comprises processing the voltage data and the current data to generate property data representing the response to the N-l test signals.
8. The test device of 7, wherein generating an indication of the one or more properties of the response to the N-l test signals comprises:processing the property data to identify if an anomaly in the one or more properties of the response over the predetermined time period; andgenerating an indication of either:an identified anomaly; oran absence of an anomaly.
9. The test device of claim 7 or claim 8, wherein the property data represents one or more varying properties in a stationary reference frame of the N-phase rotating electrical machine, and wherein generating the indication of the one or more properties of the response to the N-l test signals comprises transforming the property data to a rotating reference frame corresponding to the rotating magnetic field.
10. The test device of any preceding claim, wherein the one or more properties of the response comprise any one or more of:resistance;reactance;impedance;power; anda decay time associated with the response.
11. The test device of any one of claims 1 to 3, wherein the N-l test signals are configured to simulate three waveform signals having respective amplitudes, the three simulated waveform signals being in-phase whereby to generate a stationary magnetic field in a three-phase rotating electrical machine, and wherein the N-l test signals comprise:a first test signal representing a difference between a first simulated waveform signal and a second simulated waveform signal; anda second test signal representing a difference between a third simulated waveform signal and the second simulated waveform signal.
12. The test device of claim 11, wherein the N-l test signals are a first set of N-l test signals and the stationary magnetic field is generated at a first angle in the three-phase rotating machine and the controller is configured to:operate the set of N-l waveform signal generators to drive a second set of N-l test signals that are configured to simulate N-phase power supply signals whereby to generate a further stationary magnetic field at a second angle in the three-phase rotating machine;operate each of the set of measurement devices to measure further electrical parameters across respective terminals, the electrical parameters being representative of a further response to the second set of N-l test signals;process the measured electrical parameters to determine one or more properties of the further response to the N-l test signals; andgenerate an indication of the one or more properties of the further response to the second set of N-l test signals.
13. The test device of claim 11, wherein processing the measured electrical parameters to determine one or more properties of the response to the N-l test signals comprises:processing electrical parameters measured during a first time period to determine a first set of properties;processing electrical parameters measured during a second time period, the second time period being after the first time period, to determine a second set of properties; anddetermining the one or more properties of the response to the N-l test signals by correcting the first set of properties based on the second set of properties.
14. The test device of any preceding claim, wherein the set of N-l waveform signal generators are configured to drive the N-l test signals at predetermined voltages.
15. The test device of any one of claims 1 to 15, wherein the set of N-l waveform signal generators are configured to drive the N-l test signals at predetermined currents.
16. The test of device of any preceding claim wherein the N-l test signals comprise alternating current signals.
17. The test device of any preceding claim wherein the N-l test signals comprise direct current signals.
18. The test device of any preceding claim, wherein the N-l test signals are waveform test signals and wherein a frequency of each of the N-l test signals corresponds to a nominal operating slip frequency of N-phase rotating machine.
19. A method for a test device, the method comprising:providing N-l waveform signal generators;providing a set of measurement devices configured to measure electrical parameters;connecting each of the N-l waveform signal generators to a respective one of N windings of an N-phase rotating machine, the N-phase rotating electrical machine being configured to be operated using N-phase power supply signals provided to the N windings;operating the N-l waveform signal generators to drive N-l test signals in N-l of the windings, the N-l test signals each being driven at a respective phase and amplitude such that the N-l test signals are adapted to stimulate an electrical response in the N-phase rotating machine, the stimulated response representing a response of the N-phase rotating machine to being by driven by N-phase power supply signals;operating each of the set of measurement devices to measure electrical parameters representing a response of the N-phase rotating machine to the N-l test signals;processing the measured electrical parameters to determine one or more properties of the response to the N-l test signals; andgenerating an indication of the one or more properties of the response to the N-1 test signals,wherein N is three or more.
20. The method of claim 19, wherein the method comprises connecting one of the N windings to a reference point.
21. The method of claim 19 or claim 20, wherein the N-l test signals are configured to simulate three waveform signals, each simulated waveform signal being out of phase with the other simulated waveform signals whereby to generate a rotating magnetic field in the N-phase rotating electrical machine, and wherein the N-l test signals comprise:a first test signal representing a difference between a first simulated waveform signal and a second simulated waveform signal; anda second test signal representing a difference between a third simulated waveform signal and the second simulated waveform signal.
22. The method of claim 21, wherein the three simulated waveform signals have a common amplitude and wherein the first test signal and the second test signal are out of phase whereby to generate the rotating magnetic field in the three-phase rotating electrical machine.
23. The method of any one of claims 19 to 22, wherein the measured electrical parameters comprise current and voltage and wherein operating each of the set of measurement devices to measure electrical parameters comprises:obtaining voltage data representing voltage measurements over a predetermined time period; andobtaining current data representing current measurements over the predetermined time period,wherein processing the measured electrical parameters to determine one or more properties of the response to the N-l test signals comprises processing the voltage data and the current data representing the response to the N-l test signals over the predetermined time period, andwherein generating an indication of the one or more properties of the response to the N-l test signals comprises:processing the property data to identify an anomaly in one or more properties of the response over the predetermined time period; andif an anomaly is identified, generating an indication of the anomaly.
24. The method of claim 19, wherein the N-l test signals are configured to simulate three waveform signals having respective amplitudes, the three simulated waveform signals being in-phase whereby to generate a stationary magnetic field in a three-phase rotating electrical machine, and wherein the N-l test signals comprise:a first test signal representing a difference between a first simulated waveform signal and a second simulated waveform signal; anda second test signal representing a difference between a third simulated waveform signal and the second simulated waveform signal.
25. The method of claim 23, wherein the N-l test signals are a first set of N-l test signals and the stationary magnetic field is generated at a first angle in the N-phase rotating machine and the method comprises:operating the set of N-l waveform signal generators to drive a second set of N-1 test signals that are configured to simulate N-phase power supply signals whereby togenerate a further stationary magnetic field at a second angle in the three-phase rotating machine;operating each of the set of measurement devices to measure further electrical parameters across respective terminals, the electrical parameters being representative of a further response to the second set of N-l test signals;processing the measured electrical parameters to determine one or more properties of the further response to the N-l test signals; andgenerating an indication of the one or more properties of the further response to the second set of N-l test signals.
