Monitoring of transformers with tap changers

Abstract

A method for monitoring a transformer (10) having a tap changer (16), the method comprising: extracting information about a circulating current amplitude (A) covering at least a portion of a tap changing operation represented by at least one current difference waveform; C ) and / or circulating current time (t C ), and / or a transition current time (t R ).

Description

[Technical field]

[0001] FIELD OF THEINVENTION The present invention relates to a monitoring device, a method and a computer program product for monitoring a transformer comprising a tap changer. [Background technology]

[0002] background Transformers equipped with tap changers are frequently used in various types of power transmission environments, such as 10 kV and above. Transformers with tap changers can change the turns ratio between windings, thereby changing the voltage level. This capability is used in many systems to control the delivery of power.

[0003] Transformers are generally highly reliable. The probability that they will fail is low, say around 1%. However, of that 1% that do fail, typically 20-40% are due to tap changer failures.

[0004] The reason for this is that the tap changer is the only part of the transformer that has mechanically moving elements, therefore this part of the transformer is more likely to experience a fault than the rest of the transformer.

[0005] Therefore, when monitoring a transformer, it is important to monitor the tap-changing operation as the turns ratio is changed. If such monitoring is carried out, it may be possible to identify faults and worn parts of the tap changer at an early stage. This allows maintenance to be planned more easily, which is advantageous in terms of both reliability and economy.

[0006] US Patent No. 10473707 (Figure 3) discloses a method for extracting the operation duration time (ODT) associated with a tap-changing operation, i.e. the circulating current time and amplitude (AMP), using a monitoring device comprising a waveform recorder and a waveform analyzer. The waveform analyzer comprises a power loss handling block, a current difference handling block and a frequency domain processing block.

[0007] Transformers are monitored for tap changer operation, for example, by analyzing changes over time in the operation duration and amplitude of tap change "pulses." Using this information, it is possible to determine the health of the tap changer, and thus the health of the transformer.

[0008] Moreover, these common measurements are not always easy to use. As can be seen from Fig. 3, it is possible to obtain information that can be used for tap-changer monitoring purposes, for example based on the power loss. However, Fig. 3 shows the power loss at low load. If the transformer is subjected to high load, as Fig. 3 seems to suggest, it is not so easy to extract information indicative of the tap-changer performance from the available measured waveforms as shown. There should be side "pulses" in the waveforms, which are related to the load current passing through the transition impedance. The amplitude of such side "pulses" depends on the load current and may therefore be smaller, equal or larger than the amplitude of the main "pulse". Extraction of information on the main "pulse" and / or the side "pulses" is therefore not always possible by the methods disclosed in the prior art.

[0009] It is of interest to monitor transformers equipped with tap changers so that information indicative of the performance of the tap changer function can be extracted using waveforms that are available for condition monitoring purposes with the same reliability at all loads.

[0010] The present invention is directed to such improved transformer monitoring. Summary of the Invention [Problem to be solved by the invention]

[0011] Summary of the Invention The present invention is therefore directed to improved monitoring of transformers with respect to tap changer operation. [Means for solving the problem]

[0012] This object is achieved according to a first aspect of the present disclosure by a method for monitoring a transformer comprising a tap changer, the transformer having at least two magnetically coupled windings, the tap changer comprising at least one impedance element and a switch configured to progressively pass a load current through the at least one impedance element when switching between two tap changer positions during a tap changing operation. The method is implemented in a monitoring device and includes obtaining waveforms of measured currents recorded at a first transformer side and at a second transformer side; processing the recorded waveforms to obtain at least one current difference waveform representative of the tap changing operation, the processing of the recorded waveforms including obtaining a first current deviation waveform and a second current deviation waveform; the method further includes adding the first current deviation waveform to the second current deviation waveform and refining the current difference waveform by removing a steady-state component based on the load current, and extracting information indicative of the health of the tap changer from the at least one waveform representative of the tap changing operation. The extracted information may include a circulating current time covering at least a portion of a tap-changing operation determined from a circulating current start time and a circulating current end time of the tap-changing operation. The extracted information may also include a circulating current amplitude of the at least one waveform representative of the tap-changing operation.

[0013] The circulating current may appear as a main "pulse" in the waveform. Refining the current difference waveform by adding the first current deviation waveform to the second current deviation waveform and removing the steady-state component based on the load current is used to extract both the circulating current time and the circulating current amplitude. The refined circulating current analysis method is an improvement over the prior art because it provides only one "pulse" associated with the circulating current with clear rising and falling edges that are independent of the load current. Other methods such as power loss monitoring may not be able to identify the "pulse" associated with the circulating current during high loads. The observed amplitude of the circulating current may be compared to nominal values ​​obtained from impedance values ​​and voltage tap steps. Improved parameters such as circulating current time and circulating current amplitude may be used to better plan transformer maintenance and may give more reliable information about specific components of the tap changer. The improved accuracy of the parameters results in fewer false alarms when monitoring the tap changer.

[0014] According to a second aspect of the disclosure, this object is achieved by a method for monitoring a transformer comprising a tap changer, the transformer having at least two magnetically coupled windings, the tap changer comprising at least one impedance element and a switch configured to progressively pass a load current through the at least one impedance element when switching between two tap changer positions during a tap changing operation. The method is implemented in a protection device and includes obtaining measured current and voltage waveforms recorded at a first transformer side and at a second transformer side, and processing the recorded waveforms to obtain at least one current difference waveform and at least one power loss waveform, each representative of a tap changing operation. The processing of the recorded waveforms further includes extracting a circulating current start time and a circulating current end time from the at least one current difference waveform representative of the tap changing operation, and extracting a tap changing start time and a tap changing end time from the at least one power loss waveform representative of the tap changing operation. The method further comprises determining at least one transition current time and at least one circulating current time using the circulating current start and end times, and the tap changing start and end times, and extracting information indicative of tap changer health from the at least one current difference waveform and the at least one power loss waveform, each representing a tap changing operation, The extracted information comprises a circulating current amplitude and / or a circulating current time covering at least a portion of a tap changing operation represented by the at least one current difference waveform, and / or a transition current time covering at least a portion of a tap changing operation represented by the at least one power loss waveform.

[0015] A second embodiment combines the known power loss difference analysis with the known current difference analysis. This combination allows for the detection and extraction of the transition current times, which appear as side "pulses" before and / or after the circulating current "pulse". The transition current times provide timing information about the portion of the tap changing operation when at least one impedance element conducts current. For example, in the vacuum type tap changer example mentioned above, a significantly shorter transition current time may indicate that the main side vacuum bottle has not been opened. By examining the acquired waveforms, arcing at the transition current times may also be detected.

[0016] Optionally, processing the recorded waveforms further includes adding the first current deviation waveform to the second current deviation waveform and refining the current difference waveform by removing the steady-state component based on the load current.

[0017] Thereby, known power loss analysis is combined with the improved current difference method of the first embodiment, as outlined above, to enable monitoring of previously unavailable characteristics of the tap changer, such as the health of impedance elements and the occurrence of arcing during tap changing operations.

[0018] According to a third aspect of the present disclosure, this object is achieved by a monitoring device for monitoring a transformer comprising a tap changer. The transformer having at least two magnetically coupled windings arranged on a first transformer side and a second transformer side and a tap changer comprises at least one impedance element and a switch configured to cause a load current to flow progressively through the at least one impedance element when switching between two tap changer positions during a tap changing operation. The monitoring device comprises a waveform analyzer operative to monitor the transformer according to any of the steps of the method described above.

[0019] According to a fourth aspect of the present disclosure, this object is achieved by a computer program product comprising program code for causing a waveform analyzer to perform any of the steps of the methods outlined above.

[0020] According to a fifth aspect of the present disclosure, this object is achieved by a computer readable medium storing a computer program product as defined above.

[0021] BRIEF DESCRIPTION OF THE DRAWINGS The invention will now be described with reference to the accompanying drawings. [Brief description of the drawings]

[0022] [Figure 1] FIG. 1 is a schematic diagram of a transformer equipped with a tap changer. [Diagram 2] FIG. 2 is a schematic diagram of the tap changer and one winding of the transformer of FIG. [Diagram 3] FIG. 1 is a prior art diagram of transformer power losses occurring during tap-changing operations at very low or no load operation. [Figure 4] FIG. 1 shows a schematic diagram of a waveform recorder together with a waveform analyzer for monitoring a transformer. [Diagram 5] FIG. 2 is a block schematic diagram of a power loss handling block in a waveform analyzer. [Figure 6] 1 is a flow chart of prior art method steps of a method for monitoring a transformer as implemented by a power loss handling block in a waveform analyzer. [Figure 7] FIG. 2 is a block schematic diagram of a current difference handling block in a waveform analyzer. [Figure 8] 2 shows a flowchart of method steps in a first embodiment of a method for monitoring a transformer, as implemented by a current difference handling block in a waveform analyzer. [Figure 9]1 shows a flowchart of method steps in a second and optionally a third embodiment of a method for monitoring a transformer, as implemented by a power loss handling block and a current difference handling block in a waveform analyzer; [Figure 10] FIG. 13 is a diagram of parameter extraction according to the present disclosure for transformer power losses occurring during a tap-changing operation involving two impedance elements. [Figure 11] FIG. 13 is a diagram of parameter extraction according to the present disclosure for a transformer current difference occurring during a tap-changing operation involving two impedance elements. [Figure 12] FIG. 13 is a diagram of parameter extraction according to the present disclosure for transformer power losses occurring during a tap-changing operation involving one impedance element. [Figure 13] FIG. 13 is a diagram of parameter extraction according to the present disclosure for transformer current difference occurring during a tap-changing operation involving one impedance element. [Figure 14] FIG. 13 is an illustration of further feature extraction according to an embodiment of the present disclosure. [Figure 15] FIG. 13 is an illustration of further feature extraction according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0023] Detailed Description of the Invention Below is provided a detailed description of preferred embodiments of the present invention.

[0024] FIG. 1 shows a transformer 10 having a first winding 12 and a second winding 14 magnetically coupled to each other. These windings therefore form a pair, often referred to as a primary winding and a secondary winding. The first winding 12 is further connected to a tap changer 16. The figure also shows some transmission characteristics of the transformer that can be measured during operation. There is an input current In supplied to the first winding 12 and an input voltage Uin applied to the first winding 12. There is also an output current Iout delivered from the second winding 14 and an output voltage Uout provided by the second winding 14. The input current In and the input voltage Uin are measured quantities that are measured at a first measurement terminal MT1 and a second measurement terminal MT2 of the first winding 12, the first measurement terminal MT1 being provided at a first end of the first winding 12 and the second measurement terminal MT2 being provided at a second end of the first winding 12. The output current Iout and the output voltage Uout are measured quantities at the third and fourth measurement terminals MT3 and MT4 of the second winding 14, the third and fourth measurement terminals MT3 and MT4 being provided at the first and second ends of the second winding 14. It can thus be seen that power enters and leaves the transformer at the measurement terminals. The above example assumes a power transfer from the first winding through the transformer to the second winding. The first winding 12 thereby forms the input side of the transformer 10, and the second winding forms the output side. However, it should be understood that the power may also be transferred in the opposite direction, in which case the input current and input voltage are measured at the third and fourth measurement terminals MT3 and MT4, while the output current and output voltage are measured at the first and second measurement terminals MT1 and MT2. The second winding 14 then forms the input side, and the first winding 12 forms the output side. The measured quantity is therefore also a transmission characteristic that is measured at a measurement terminal.

[0025] The illustrated transformer is represented diagrammatically. It is understood that it may also comprise an iron core. It should also be understood that the transformer is often a three-phase transformer. This means that there are three pairs of windings. The transformer may also include two or more secondary windings magnetically coupled to the same primary winding. A tap changer may also be connected to the second winding as an alternative. The transformer may also be a three-phase transformer configured as a system of multiple single-phase transformers, each connected to a corresponding tap changer.

[0026] 2 shows the first winding 12 diagrammatically together with a tap changer comprising a regulating winding 19. The first winding has a first end and a second end, the first end being connected to a first measurement terminal MT1 and the second end being connectable to the regulating winding 19. The regulating winding 19 also has a first end and a second end. The tap changer further comprises a diverter 20 and a selector 18 for selecting the number of turns of the regulating winding 19 to be connected to the primary winding 12.

[0027] The selector 18 has a selector switch 24 used to reverse the orientation of the regulating winding 19, and thus has a first end connected to the second end of the first winding 12 and a second end movable between two positions: a first position at the first end of the regulating winding 19 and a second position at the second end of the regulating winding 19. Each winding 12 and 19 comprises a conductor wound multiple times. Furthermore, the regulating winding 19 is provided with several tap points, six of which 1, 2, 3, 4, 5 and 6 are shown as examples. The tap points are used to determine the number of turns of the regulating winding 19 that are connected to the first winding 12 by a first selector arm connected to the first diverter terminal DT1 and a second selector arm connected to the second diverter terminal DT2.

[0028] The diverter 20 then comprises a diverter switch 22 having a first end connected to the second measurement terminal MT2 and a second end connectable between four contact positions, the first contact position P1 communicating with the first diverter terminal DT1 via a first diverter arm, the second contact position communicating with the first diverter terminal DT1 via an impedance element in the form of a first resistor R1, the third contact position communicating with the second diverter terminal DT2 via an impedance element in the form of a second resistor R2, and the fourth contact position P4 communicating directly with the second diverter terminal DT2 via a second diverter arm. The diverter 20 is provided for commutating the load between the two selector arms.

[0029] It should be understood here that this is only one realization of the selector and diverter tap changer. There are several other types of selector and diverter tap changers. There are also other types of tap changers. Another type is, for example, a selector-switch tap changer that combines selection and commutation in one motion, but has a contact sequence similar to the selector and diverter tap changers. Yet another example is a vacuum tap changer, where the contact sequence is often modified but still includes a time when power is lost in the impedance elements. The illustrated tap changer also includes resistors as impedance elements. However, it is also known to use other types of impedance elements, such as inductors. Furthermore, the illustrated tap changer has two impedance elements. It should be understood that it is also possible to have fewer impedance elements, such as one, or more impedance elements, such as three or four.

[0030] What is common to all these tap changers is that during tap changing, the switch moves progressively between two positions, in the case of the tap changer of FIG. 2, the diverter switch 22 moves between the first position P1 and the fourth position P4. When a tap change is desired, one of the selector arms is originally connected between the tap position and the corresponding diverter terminal. As an example, the first selector arm may be connected between the second tap position 2 shown in FIG. 2 and the first diverter terminal DT1, while at the same time the diverter switch 22 is in the first contact position P1. This causes the first selector arm to be loaded, i.e. the load current passes through the first diverter arm and enters the first selector arm. The tap position that obtains the desired turns ratio change is then selected for the unloaded selector arm, which in this example is the second selector arm. The second diverter terminal DT2 thus connects the unloaded second selector arm to the selected tap position, which in the example of FIG. 2 is the first tap position 1. To divert the load, i.e. to perform a turns ratio switch, the diverter switch 22 is then moved progressively from the first contact position to the fourth contact position. In this progressive movement, the diverter switch 22 first establishes contact with the first contact position P1 and at the same time with the first resistor R1. The diverter switch 22 then releases contact with the first contact position P1 and the first diverter arm. The load current now flows only through the first resistor R1 to the first selector arm. The diverter switch then connects to the second resistor R2. At this point, the load current flows through both resistors R1 and R2 to both the first selector arm and the second selector arm. There is also a circulating current that is generated through the voltage difference between the diverter terminals DT1 and DT2. Following this, the diverter switch 22 disconnects the first resistor R1, so that the load current flows only through the second resistor R2 to the second selector arm. Finally, the diverter switch reaches the fourth contact position P4, after which the load current flows through the second diverter arm to the second selector arm. This completes the tap-changing operation.It can be seen that during such movement, energy is loaded or stored in at least one impedance element, in the example of Fig. 2, two resistors R1 and R2. This energy is dissipated in the case of a resistor, but is only temporarily stored in the case of an inductor. This causes the movement of the switch 22 to pass the load current progressively through the first resistor R1 and the second resistor R2.

[0031] By the above-described operation of the tap changer of Figure 2, i.e. during the gradual switching between positions P1 and P4, the impedances R1 and R2 are connected between the measurement terminals MT3 and MT4 and thus energy is stored in them. Since in this example the impedances are resistive, real power is lost. If the impedances were inductive, reactive power would be stored instead.

[0032] The instantaneous power loss can be determined, for example, according to the following equation (1):

[0033]

number

[0034] Thus, the formula provided for three-phase systems defines the power losses of a three-phase transformer. The instantaneous power loss Ploss is thus calculated as input voltage Uin × input current In - output voltage Uout × output current Iout based on the transmission characteristic measurements taken at the first measurement terminal MT1, the second measurement terminal MT2, the third measurement terminal MT3 and the fourth measurement terminal MT4, with three such differences being obtained, one for each phase.

[0035] From equation (1), we can determine real power loss as the average of instantaneous power loss over a period of time, and reactive power loss as oscillations with a mean of zero.

[0036] What the power losses look like at low load for such a three-phase transformer equipped with a tap changer can be seen in figure 3. In the figure, the power losses of each phase due to the coordinated tap-changing operation in all three phases are shown. In the first phase there is a power Ploss_A, in the second phase there is a power loss Ploss_B and in the third phase there is a power loss Ploss_C. These three power losses are further summed up to a total power loss Ploss, which has the shape of a "pulse". In the figure, it can be seen that the tap-changing operation is performed between the time 0.00 s and 0.02 s. It can be seen that the "pulse" or this interval as an example has a width of 20 ms, this width being the peak duration of the energy deposition or peak duration energy deposition time, in which interval each phase experiences a power loss. In the prior art method shown in figure 3, this peak duration is called the operation duration ODT between the rising and falling edges of the "pulse", since it corresponds to the duration of at least some steps in the operation of the tap changer. It therefore covers at least a part of the tap-changing operation. More specifically, it can be seen that the operation duration ODT corresponds to the time that the diverter switch 22 is solely connected to resistors R1 and R2. It can also be seen that the power loss during tap changer operation has an amplitude AMP. The amplitude and operation duration ODT are examples of information indicative of the tap changer performance or tap changer health, which can be obtained via the recorded waveforms and used for monitoring purposes. In the present disclosure, the operation duration ODT is defined as the time when the circulating current time t C In this disclosure, the amplitude AMP corresponds to the circulating current amplitude A C Corresponds to.

[0037] Transformers are generally highly reliable. The probability of them failing is low, say around 1% per year. But of that 1%, typically 20-40% are due to tap changer failures.

[0038] As mentioned above, the only moving element in the transformer is contained in the tap changer, which makes it relevant to reliability.

[0039] If the transformer is monitored with respect to the operation of the tap changer, for example by analyzing the change over time of the circulating current time and the amplitude of the tap-changing "pulse", it is possible to determine the health of the tap changer and therefore also of the transformer. This can be used to better plan the maintenance of the transformer. Very short commutation times can be dangerous due to the fact that a possible arc is most reliably extinguished at the current zero crossing. This means that the circulating current time must be longer than a half period. If the arc is allowed to withstand the operation, a short circuit of the regulating winding will occur, which will generate a lot of energy and may cause the transformer to fail. Thus, a circulating current time that is too short may correspond to a failure of the tap changer. On the other hand, a commutation time that is too long indicates that there is friction in the system, but there is no sign of a failure. Thus, a long circulating current time indicates that service is needed, since long commutation times generate more heat for resistive tap changers that can affect the general performance of the transformer, for example, a drop in power. This may be an indicator that more cooling is needed. Thereby, by monitoring the change in circulating current time it is possible to predict when a fault will occur and to plan the service of the transformer. The amplitude then indicates the size of the resistance. As the amplitude changes, the resistance changes. Also, a high amplitude corresponds to a low resistance and a low amplitude corresponds to a high resistance. The amplitude can be significant when no "pulse" is visible at all, which means that the tap changer is not moving or the resistance is close to infinity, i.e. the resistor is broken. Both of these situations are significant fault cases, the latter of which can cause an arc that shorts the regulating winding.

[0040] Furthermore, the load current passing through one or more of the involved transition impedance elements (resistors) may generate a power loss proportional to the square of the load current, which appears as an additional loss "pulse" immediately preceding or following the power loss "pulse" associated with the circulating current. Such adjacent "pulses" indicate one or more time information of the operating sequence. With prior art methods, these transition impedance losses may be difficult to identify separately at significantly higher loads.

[0041] Another problem is that tap changers cannot be monitored directly: transformers typically do not have any measurements directly related to tap-changing operations. Therefore, monitoring of the transformer health with respect to the tap changer must be done using typical transformer measurements such as input / output currents and voltages.

[0042] Moreover, these common measurements are not always very easy to use. As can be seen from figure 3, for example based on the power loss it is possible to obtain information that can be used for tap changer monitoring purposes.

[0043] It is an object of the present disclosure to improve upon the prior art and provide a method for using a monitoring device to extract more information based on current difference monitoring and power loss monitoring.

[0044] One way of realising the monitoring device 25 is shown in figure 4. The monitoring device 25 comprises a waveform recorder 26 comprising a waveform recording block WFR 28 which receives the measured quantities of the transformer in the form of the measured transmission characteristics In, Uin, Iout and Oout, as well as a tap changer control signal TC_CTRL, which is a control signal sent to the tap changer to perform or initiate a tap changing operation. Such a control signal is typically sent from a tap changer control unit which may be associated, as an example, with the power control of the power transmission system. The waveform recorder 26 also comprises a waveform memory WFM 30 in which the recorded waveforms are stored. Once the tap changer control signal TC_CTRL is obtained by the waveform recorder 26, the waveform recording block 28 records several waveforms of the input and output currents and the input and output voltages (if available) and stores these in the waveform memory 30. Thus, obtaining or receiving the tap changer control signal TC_CTRL triggers the recording of a waveform, which then continues for a predefined time, such as 10 seconds. This has the advantage of providing a waveform that is clearly consistent with the tap changer operation. Furthermore, this length allows for safe waveform prediction, and the control signal typically precedes the actual tap-changing operation so that a reliable prediction of the steady-state operating waveform can be made.

[0045] The monitoring device 25 also comprises a waveform analyzer 34 comprising a power loss handling block PLH 36, a current difference handling block CDH 38 and, optionally, a frequency domain processing block FDP 40. Each of these blocks provides a separate way of obtaining at least one waveform from which information indicative of the tap change can be extracted. Finally, the waveform analyzer 34 comprises an information analyzing block IA 43 and a health data memory HDM 42 in which (time-stamped) waveforms and / or (time-stamped) tap changer performance indicator information are stored for condition monitoring purposes.

[0046] The waveform analyzer 34 may be implemented via a computer or processor having an associated program memory containing computer instructions implementing the above-mentioned blocks. It may also be realized via one or more dedicated components, such as an Application Specific Integrated Circuit (ASIC) or Field Programmable Gate Array (FPGA), that implement the blocks. It should also be understood that the waveform analyzer may sometimes only need to comprise one of the power loss handling block 36, the current difference handling block 38, and the frequency domain processing block 40. It should also be understood that the information analysis block 43 may be omitted and its functionality provided in another entity, such as a separate monitoring computer.

[0047] The implementation of the waveform recorder 26 is known as such, but in some aspects may be implemented in the same way as the waveform analyzer 34. However, it may also comprise an Analog / Digital (A / D) converter and possibly a scaling unit for scaling at least a portion of the received signal.

[0048] One method of operating the waveform analyzer 26 to monitor the transformer 10 will now be described with reference to FIG. 5, which shows a block schematic diagram of the power loss handling block 36 and FIG. 6, which shows a flow chart of some method steps in a prior art method of monitoring the transformer 10 implemented within the power loss handling block 36. The prior art method of FIG. 6 is part of the inventive concept of the present disclosure and is shown in FIG. 6 as an isolated method only to facilitate understanding of the improved method of the present disclosure shown in FIG.

[0049] The power loss handling block 36 comprises a power loss determining element PLD (power loss determining) 44, a power loss predicting element PLP (power loss predicting) 46, a power loss waveform forming element PLWF (power loss waveforming) 48 and a waveform analyzing element WFA (waveform analyzing) 49.

[0050] As mentioned above, the waveform recorder 26 records the waveform appearing at the measurement terminals when there is a tap changing action triggered by the tap changer control signal TC_CTRL, and this waveform is stored in the waveform memory 30.

[0051] The recorded waveform also has a start point and an end point. To obtain waveforms indicative of the tap changing performance, the power loss determination element 44 of the power loss handling block 36 first obtains the recorded waveforms In, Uin, Iout, and Uout from the waveform memory 30 of the waveform recorder 26 (step 50), and thus the waveforms are the transmission characteristic waveforms recorded at the first transformer side and the second transformer side.

[0052] The power loss handling block then processes the recorded waveforms to obtain at least one waveform representative of the tap-changing operation.

[0053] The process here involves processing the difference between the transmission characteristics of both sides of the transformer using the measured transmission characteristics. The transmission characteristics in question here are the derived characteristics of the power, and the difference between the two sides is the power loss through the transformer, represented as a power loss waveform. The process further involves predicting at least one waveform of the same type as the difference waveform, and obtaining each waveform representing a tap-changing operation as the difference between the difference waveform and a corresponding predicted waveform, so that there is one predicted waveform and one difference waveform.

[0054] Thus, the decision element 44 uses the input and output current In, Uin, Iout and Oout waveforms to determine the power loss of the transformer (step 52). In this case, the power loss may be determined in the manner shown in equation (1) above.

[0055] At the same time, the power loss prediction element 46 predicts the same power loss (step 54). The prediction may include prediction of the stationary waveforms of the currents and voltages on the first and second sides of the transformer 10. Each such signal Uin, In, Uout, Iout may be predicted from previous observations by generating a spectrum from the vibration signal and extracting the significant frequency amplitudes. From these amplitudes, the waveforms at future times may be predicted. The prediction method is described in more detail in U.S. Pat. No. 8,095,326, which is incorporated herein by reference.

[0056] Thus, the predicted waveforms may correspond to stationary waveforms of the currents and voltages on the first and second sides of the transformer, i.e., waveforms without any tap-changing information embedded in them. These predicted waveforms may then be used to form the predicted power losses (step 54).

[0057] The power loss waveform shaping element 48 then determines the difference between the determined power loss and the predicted power loss (step 56) to obtain a power deviation waveform that defines the tap-changing operation. The difference is then used as the waveform that defines the tap-changing operation. As mentioned above, the power loss prediction represents the steady state power loss without tap-changing operation. Therefore, in such a difference waveform, essentially only the "pulses" shown previously in FIG. 3 should be present, such that information indicative of the tap change performance can be easily extracted from the waveform. The power loss difference waveform derived from the power loss method is also shown in FIG. 10 and FIG. 12.

[0058] The difference can be obtained, for example, as follows.

[0059]

number

[0060] One way this can be done is by the following.

[0061]

number

[0062] Equation (3) shows the determination of the input or output power change, where i encodes the input or output side, and ΔI and ΔV are the difference between the predicted and actual current waveforms and the difference between the predicted and actual voltage waveforms.

[0063] The power loss difference between the actual power loss and the predicted power loss is:

[0064]

number

[0065] This removes the steady state component and the result is a waveform that represents the tap-changing operation and essentially only contains the power loss due to the tap-changing operation. TS and tap switching end time t TE , thereby determining, for example, the time between the rising and falling edges of the waveform (step 58) and the amplitude A C By identifying the amplitude A of the tap-changing operation (step 60), it is easy to determine the overall width of the tap-changing operation, which can be performed by the waveform analysis element 49. CThe tap change start and end times may then be stored (step 62), possibly time-stamped, along with the power loss difference waveform in the health data memory 42 for use in transformer condition monitoring. In this regard, other information may also be stored, such as initial and final tap positions, load currents, etc.

[0066] The health determination and prediction can then be performed by the information analysis block 43. The information analysis block 43 can thus analyze the information indicative of the tap-changing operations stored in the memory 42 by the waveform analysis element 49 to estimate the wear of the transformer and determine when to perform maintenance or rather in which service window the transformer is replaced or serviced. As an alternative, the power loss analysis element 49 may be omitted and the power loss waveform formation element 48 may store the power loss difference waveform in the health data memory 42. In this case, the information analysis block 43 can instead retrieve the power loss difference waveform from the health data memory 42 and extract therefrom the information indicative of the tap-changing operations when determining the health of the transformer.

[0067] The above-described method of obtaining the start and end times of tap-changing operations does not by itself allow for the extraction of other parameters of interest such as transition current times that can be used to monitor impedance elements. A first embodiment of the invention is an improved current difference method that uses a measured current as input and compares the measured current to a predicted current.

[0068] This first embodiment will now be described with reference to Figures 7 and 8. Figure 7 shows a block schematic diagram of the current difference handling block 38, and Figure 8 shows a flow chart of some method steps in a method for monitoring a transformer, which are implemented by the current difference handling block 38.

[0069] The current difference handling block 38 comprises a current difference determining element CDD 64 , a current difference predicting element CDP 66 , a current deviation waveform determining element CDWF 68 and a current difference analysis element 69 .

[0070] Similar to the power loss difference of the prior art described above, in a first embodiment there is a processing of the difference between the transmission characteristics of both sides of the transformer using a measured transmission characteristic. In this first embodiment, the transmission characteristic in question is the measured transmission characteristic of the current, and the difference between the two sides is the current difference between the two sides. This processing includes predicting a waveform of the same type as the difference waveform, i.e. predicting a current difference waveform. The processing also includes obtaining each waveform representing a tap-changing operation as the difference between the difference waveform and a corresponding predicted waveform. In this first embodiment, two waveforms are predicted, one forward in time from the difference waveform and one backward in time, and obtaining at least one waveform includes obtaining two deviation waveforms, one being the difference between the difference waveform and a first predicted waveform of the predicted waveforms, and the other being the difference between the difference waveform and a second predicted waveform of the predicted waveforms. According to this first embodiment, the two deviation waveforms are added and the steady-state component based on the load current is removed. See FIG. 11 and FIG. 13. FIG. 11 shows the current deviation waveforms B, F and the resulting waveform S, while FIG. 13 shows only the resulting waveform S. The steady-state component is at a level of about 4 amperes in FIG. 11. From the resulting waveform S, the circulating current time t C and the amplitude can be determined. The stationary component can be written as:

[0071]

number

[0072] where Δn is the ratio change between the two taps, e.g., 0.0167. To obtain at least one waveform indicative of the performance of the tap changer, the current difference determination element 66 of the current difference handling block 38 first obtains the recorded waveforms In and Iout of all phases from the waveform memory 30 of the waveform recorder 26 (step 70), and the measured transmission characteristic waveforms in this embodiment only include the current waveforms recorded on the first transformer side and the second transformer side.

[0073] Thereafter, the current difference determining element 64 determines the current difference Idiff between the currents on either side of the transformer (step 72).

[0074] The current difference may be based on determining partial current differences of different phases. For phase p, the corresponding partial current difference can be obtained as follows:

[0075]

number

[0076] The total current difference of all phases can then be determined as follows:

[0077]

number

[0078] The same prediction techniques described above can be used to make a prediction of the current difference. However, it should be noted that if the transformer includes, for example, three connected single-phase transformers, each provided with a tap changer, or a three-phase transformer with more than one tap changer, equation (6) cannot be used to determine the total current difference of all phases. Instead, the current difference waveforms are analyzed individually.

[0079] In this embodiment, the current difference prediction element 66 predicts the first difference current. The prediction may be based on the same equations (5) and (6) above, modified according to equations (2) and (3). Furthermore, in this case, the prediction is made in a forward direction from the beginning of the waveform to the end of the recording (step 74).

[0080] The current deviation waveform determiner 68 then obtains or determines (step 76) a first current deviation waveform as the difference between the actual current difference waveform and the first predicted current difference waveform. In this waveform, the current difference analyzer 69 can then detect (step 78) the start of a tap-changing operation, for example by detecting a rising edge or a high positive time derivative of the waveform. It is also possible for it to detect the amplitude. However, this first embodiment provides an improved way of detecting the rising edge and the amplitude, as explained below.

[0081] However, the end of the tap-changing operation is not easily detected using the first current deviation waveform because the steady-state waveform has changed after the turns ratio change due to the tapping operation. Therefore, to detect the end of the tap-changing operation, the current difference prediction element 66 predicts a second current difference in the reverse direction, i.e., starting at the end of the waveform recording and progressing backwards in time (step 80).

[0082] Following this, the current deviation waveform determiner 68 obtains or determines a second current deviation waveform as the difference between the actual current difference waveform and the second predicted current difference waveform (step 82), in which the current difference analyzer 69 can then detect the end of the tap-changing operation (step 84), for example by detecting a falling edge or a large negative time derivative of the waveform. Again, the amplitude can be estimated.

[0083] In the first embodiment, the pulse width, i.e., the circulating current time t CBy adding (85) the first current deviation waveform to the second current deviation waveform and removing (87) the steady-state component based on the load current, the current difference analysis element 69 determines the circulating current "pulse" width, i.e., the circulating current time t C Moreover, by removing the steady-state component based on the load current and dividing by two, only one clearly defined "pulse" remains after the improvement, so the amplitude can be obtained more accurately compared to the prior art.

[0084] FIG. 11 shows an example of a flag-cycle tap changer having two resistors R1 and R2 and illustrates how a first current deviation waveform F and a second current deviation waveform B are summed and the steady-state component based on the load current is removed to result in a waveform S from which a circulating current time (t C ) can be extracted. This is because the circulating current start time (t CS ) and the end time of the circulating current (t CE As is clear from FIG. 11, the first current deviation waveform F and the second current deviation waveform B are both determined from the current amplitude (A C ) and is not even able to extract the amplitude from the first current deviation waveform F. This is a significant improvement over the prior art.

[0085] FIG. 13 shows an example of a pennant cycle tap changer with one resistor R1 monitored by the current difference method of the first embodiment.

[0086] The current difference analysis element 69 can determine the "pulse" width or circulating current time from the start and end, i.e., rising and falling edges, respectively, of the circulating current "pulse" (step 86). The amplitude can be determined or obtained by looking at the height of the circulating current "pulse" (88).

[0087] Information indicative of tap changing performance may then be extracted, possibly time-stamped, and stored (90) in health data memory 42, where other information may also be stored, such as initial and final tap positions, load currents, etc. The current deviation waveform determiner 66 may also store the first current deviation waveform and the second current deviation waveform in memory 42.

[0088] The information analysis block 43 can then analyze the stored information and / or waveforms for condition monitoring purposes. Again, if not already done, the information analysis block 43 can extract tap change performance index information from the sum of current deviation waveforms.

[0089] The first embodiment is therefore an improvement over the prior art current difference method, providing more accurate and detailed results.

[0090] The described embodiment relies on coordinated tap-changing operations on three phases. However, for some tap-changer installations, most notably where there are three separate tap-changers, one for each phase, the tap-changing operations on different phases may not be coordinated. This may result in tap-changing operations being performed independently of each other on different phases. It may therefore be necessary to monitor the transformer for tap-changing operations separately for the three phases. Thus, as mentioned above, equation (6) is not applicable for such transformers.

[0091] In a second embodiment, the power loss difference method described above is combined with either the conventional current difference method or the improved current difference method of the first embodiment described above to result in an improved method of monitoring a tap changer. The combined method is shown in Figure 9 and involves power loss difference and current difference steps performed in parallel, with the results being combined to obtain additional monitoring information, in particular information regarding at least one impedance element R.

[0092] In Fig. 9, the steps of the first embodiment are shown in dashed lines as optional steps. In the conventional current difference method, the two current difference waveforms are not summed and the steady-state component of the current is not removed. Instead, the pulse width is determined by, for example, the start time t CS and the end time t CE The amplitude can also be determined or obtained (step 88) by looking at either of the current deviation waveforms.

[0093] Thus, the method according to the second embodiment relates to monitoring a transformer 10 comprising a tap changer 16. The transformer has at least two magnetically coupled windings 12, 14. The tap changer 16 further comprises at least one impedance element R and a switch 22 configured to progressively pass a load current through the at least one impedance element R when switching between two tap changer positions P1, P4 during a tap changing operation. The method is implemented in a monitoring device 25 and comprises obtaining 50, 70 waveforms of measured currents and voltages Iin, Uin, Iout, Uout recorded at the first transformer side and at the second transformer side and processing 52, 54, 56, 72, 74, 76, 78 to obtain at least one current difference waveform and at least one power loss waveform each representative of the tap changing operation.

[0094] The processing of the recorded waveforms involves extracting the circulating current start time (t CS ), the end time of the circulating current (t CE ) and the circulating current amplitude A C , and determining (86, 88) a tap-switching initiation time (t TS ) and the end time of tap switching (t TE ) and Circulating current start point (t CS ) and the end time of the circulating current (t CE), tap switching start time (t TS ) and the end time of tap switching (t TE ) to determine at least one transition current time (t R ) and at least one circulating current time (t C ) determining 100; and determining from the at least one current difference waveform and the at least one power loss waveform, each of which represents a tap-changing operation, information indicative of the health of the tap-changing (t C , A C , t R ) is extracted 102 and Further comprising: The extracted information includes a circulating current amplitude A that covers at least a portion of the tap-changing activity represented by the at least one current difference waveform. C and / or circulating current time (t C ), and / or at least one transition current time t that covers at least a portion of the tap-changing operation represented by said at least one power loss waveform. R Includes.

[0095] Therefore, from the power loss deviation waveform, the waveform analysis element 49 can determine, for example, the tap switching start time t TS Identify the tap switching end time t from the falling edge of the waveform. TE By identifying the total tap-changing operation time through the width of the power loss deviation waveform, the total tap-changing operation time can be determined (58).

[0096] By combining the data extracted from the power loss and current loss methods, one or more transition current times t R and the circulating current time can be determined as the interval. The transition current time t, which appears as a "pulse" to the left of the circulating current "pulse", R is the tap switching start time t TS and the circulating current start time t CS The transition current time, which appears as a "pulse" to the right of the circulating current "pulse", can be determined as the interval between the circulating current end time t CE and tap switching end time tTE Furthermore, the circulating current time t C is the time when the circulating current starts, t CS and the end time of the circulating current t CE is determined as the interval between

[0097] Information indicating the health or performance of the tap changer C , A C , t R may be extracted 102 from the at least one current difference waveform and the at least one power loss waveform.

[0098] Information indicating the performance of the tap-changing operation, i.e., the circulating current amplitude A C , circulating current time t C and / or at least one transition current time t R may then be stored 102, possibly with a time stamp, along with the power loss difference waveform in the health data memory 42 for use in transformer condition monitoring. In this regard, other information such as initial and final tap positions, load current, etc. may also be stored.

[0099] Health determination and prediction can then be performed by the information analysis block 43. Thus, the information analysis block 43 uses the information t indicative of tap-switching operations stored in the memory 42 by the waveform analysis elements 49, 69. C , A C , t R can be analyzed to estimate the wear on the transformer / tap changer and decide when to perform maintenance or rather in what service time frame the transformer will be replaced or serviced.

[0100] In this way, it can be seen that it is possible to identify tap changing operations regardless of the size of the load and extract new tap changing performance indicator information relating to the transition current times extracted from the side "pulses" of the power loss waveform and the circulating current times extracted from the current difference waveform, allowing for improved maintenance planning.

[0101] In short, improved monitoring of transformers equipped with tap changers is obtained and it is found possible to evaluate the performance of the tap changers. This allows for more reliable maintenance planning. This is also done without the need for dedicated tap changer monitoring sensors, using only the measurement quantities conventionally available for transformers. In this way, improved monitoring is obtained using a minimum of additional hardware and without the introduction of additional sensors and modifications to the waveform recorder.

[0102] The method according to the second embodiment may optionally be refined by the current difference method according to the first embodiment, where the processing of the recorded waveform further comprises refining the current difference waveform by adding (85) the first current deviation waveform to the second current deviation waveform and removing (87) the steady-state component based on the load current.

[0103] Although not detailed in the description of the second embodiment above, the determination of the two current deviation waveforms and the power loss deviation waveform follows the same steps as described above in connection with the description of the power loss method shown in Figures 5 and 6 and the description of the current difference method of the first embodiment shown in Figures 7 and 8.

[0104] Several variations are possible apart from those already described. For example, the monitoring device could comprise only a waveform analyzer set up to communicate with a waveform recorder. It is also possible that the monitoring device is provided as part of an Intelligent Electronic Device (IED) provided for the transformer.

[0105] FIG. 10 shows the first transition current time t R1 and the second transition current time t R2 To determine the circulating current start time t CS and the end time of the circulating current t CE is the tap switching start time t TS and tap switching end time t TEThe transition currents are shown as "pulses" on the left and right of the circulating current "pulse."

[0106] FIG. 12 shows the method used on a pennant cycle tap changer with a single transition impedance R1.

[0107] Figures 14 and 15 show problematic tap-changing behavior as seen in the resulting power loss difference waveform. Figure 14 shows how a deviation X is detected in a pennant cycle tap changer as monitored in Figures 12 and 13. In Figure 14, the main side vacuum bottle of one phase does not open. t C In Figure 15, the transition current time t is much shorter than predicted because the main side vacuum bottles of all phases did not open, resulting in an earlier start compared to normal operation. R1 and a longer than expected circulating current time t C This occurs.

[0108] The waveform analyzer may be realized in the form of separate components. However, it may also be implemented in the form of a processor with an associated program memory containing computer program code which, when executed on the processor, performs the desired control functions. A computer program product carrying this code may be provided as a data carrier, such as one or more CD ROM disks, hard disk storage, or one or more memory sticks, which carry the computer program code which, when loaded into the waveform analyzer, performs the waveform analyzer functions described above.

[0109] From the above description it will be apparent that the invention can be varied in many ways. Consequently, it will be understood that the invention should be limited only by the scope of the appended claims.

Claims

1. A method for monitoring a transformer (10) comprising a tap changer (16), the transformer having at least two magnetically coupled windings (12, 14), the tap changer (16) comprising at least one impedance element (R) and a switch (22) configured to progressively pass a load current through the at least one impedance element (R) when switching between two tap changer positions (P1, P4) during a tap changing operation, the method being implemented in a monitoring device (25), - obtaining (70) waveforms of the measured currents (Iin, Iout) recorded on the first transformer side and on the second transformer side, - processing (72, 74, 76, 78) the recorded waveforms to obtain at least one current difference waveform representative of tap changing operations, the processing of the recorded waveforms comprising obtaining (76, 82) a first current deviation waveform and a second current deviation waveform; - refining the current difference waveform by adding (85) the first current deviation waveform to the second current deviation waveform and removing (87) a steady-state component based on a load current; - deriving information indicative of the health of the tap changer (t C , A C ) and storing it (90); Including, The extracted information covers at least a portion of the tap-changing operations and includes a circulating current start time (t CS ) and the end time of the circulating current (t CE Circulating current time (t C ), and / or the circulating current amplitude (A C ).

2. Obtaining the first current deviation waveform and the second current deviation waveform includes: - determining (72) a current difference waveform representative of the difference between the currents on either side of the transformer based on the measured currents and predicting (74, 80) from said current difference waveform a first predicted current waveform forward in time and a second predicted current waveform backward in time; - determining the first current deviation waveform as the difference between the difference waveform and the first predicted current waveform, and determining the second current deviation waveform as the difference between the difference waveform and the second predicted current waveform (76, 82); The method of claim 1 further comprising:

3. A method for monitoring a transformer (10) comprising a tap changer (16), the transformer having at least two magnetically coupled windings (12, 14), the tap changer (16) comprising at least one impedance element (R) and a switch (22) configured to progressively pass a load current through the at least one impedance element (R) when switching between two tap changer positions (P1, P4) during a tap changing operation, the method being implemented in a monitoring device (25), - obtaining (50, 70) waveforms of measured currents and voltages (Iin, Uin, Iout, Uout) recorded at the first transformer side and at the second transformer side; - processing the recorded waveforms (52, 54, 56, 72, 74, 76, 78) to obtain at least one current difference waveform and at least one power loss waveform, each representative of a tap-changing operation; Including, Processing the recorded waveform includes: - determining the circulating current start time (t CS ), the end time of the circulating current (t CE ), and the circulating current amplitude (A C ) (86, 88); and - determining a tap-switching start time (t TS ) and the end time of tap switching (t TE ) (58); and - the starting point of the circulating current (t CS ) and the end time of the circulating current (t CE ), the tap switching start time (t TS ) and the end time of tap switching (t TE ) to determine at least one transition current time (t R ) and at least one circulating current time (t C ) (100); and - deriving information indicative of the health of the tap-changing from the at least one current difference waveform and the at least one power loss waveform, each of which represents a tap-changing operation (t C , A C , t R ) (102); and Further comprising: The extracted information includes a circulating current amplitude (A) covering at least a portion of the tap-changing activity represented by the at least one current difference waveform. C ) and / or circulating current time (t C ), and / or a transition current time (t R ).

4. The circulating current start time (t CS ) and the end time of the operation (t CE said extracting information including said time course of events (i.e., the time course of events) comprises processing said recorded waveform; - obtaining a first current deviation waveform and a second current deviation waveform from said current difference waveform (76, 82); Including, The starting point of the circulating current (t CS ) is extracted as the start of the first current deviation waveform, and the circulating current end time (t CE 4. The method of claim 3, wherein: ##EQU1## is extracted as the end of the second current deviation waveform.

5. 5. The method of claim 4, wherein processing the recorded waveforms further comprises refining the current difference waveform by adding (85) the first current deviation waveform to the second current deviation waveform and removing (87) a steady-state component based on a load current.

6. Obtaining the first current deviation waveform and the second current deviation waveform includes: - determining a current difference waveform representative of the difference between the currents on either side of the transformer based on the measured currents and predicting from said current difference waveform a first predicted current waveform forward in time and a second predicted current waveform backward in time (74, 80); - determining the first current deviation waveform as the difference between the difference waveform and the first predicted current waveform, and determining the second current deviation waveform as the difference between the difference waveform and the second predicted current waveform (76, 82); The method of claim 4 or 5, further comprising:

7. Processing the recorded waveforms to obtain the at least one power loss waveform indicative of tap-changing activity comprises: - determining (52) a power loss difference waveform representing the difference between the power on both sides of the transformer based on the measured currents and voltages; - predicting a predicted power waveform (54); determining (56) a power deviation waveform as the difference between said power loss difference waveform and said predicted power waveform; The method of any one of claims 3 to 6, further comprising:

8. 8. A monitoring device (25) for monitoring a transformer (10) comprising a tap changer (16), the transformer (10) having at least two magnetically coupled windings (12, 14) arranged on a first transformer side and on a second transformer side, the tap changer (16) comprising at least one impedance element (R) and a switch (22) configured to progressively pass a load current through the at least one impedance element (R) when switching between two tap changer positions (P1, P4) during a tap changing operation, the monitoring device comprising a waveform analyzer (34) operative to monitor the transformer (10) according to the steps of any one of claims 1 to 7.

9. A computer program product comprising program code for causing the waveform analyzer (34) of claim 8 to carry out the steps of any one of claims 1 to 7.

10. 10. A computer readable medium having stored thereon the computer program product of claim 9.

Images