Method for monitoring a high voltage switching device

Abstract

There is disclosed herein a system and a (computer-implemented) method for condition monitoring of a high voltage switching device, the method comprising: obtaining a vibration signal (400) over time comprising vibration information for a high voltage switching device at least during a switching event of said high voltage switching device; segmenting the vibration signal into a plurality of segments (402, 404, 406); scaling at least one segment based on at least one other segment to thereby obtain a scaled segment; and determining a condition of the high voltage switching device based on the scaled segment.

Description

METHOD FOR MONITORING A HIGH VOLTAGE SWITCHING DEVICE Technical Field

[0001] The present disclosure relates to high voltage power grid devices, and maintaining the reliable operation thereof. More particularly, the present disclosure relates to a method for monitoring the condition of a high voltage switching device (HVSD). The present disclosure further relates to a high voltage switching system.Background

[0002] High voltage switching devices are used in electrical power networks or inside a power network station, to control, protect, and isolate different sections of the electrical power network system, ensuring safe and reliable power transmission and distribution. These devices, such as circuit breakers, disconnectors, earthing switches, and load switches, may allow operators to manage power flow or perform maintenance safely, and can quickly respond to faults like short circuits or overloads. By rapidly interrupting high currents and preventing equipment damage, or isolating and grounding components, they enhance system stability and minimize downtime.

[0003] Mechanical failures constitute a significant portion of failures for high voltage switching devices. Traditional approaches for diagnosing the health of a high voltage switching device are based on the measurement of specific features, such as closing I opening time or contact speed and subsequently comparing them with the manufacturer specification. These features are typically extracted from a measured travel curve, which describes how the position of the electrical contacts of the high voltage switching device varies with time during a switching operation. Many mechanical malfunctions or degradations, which constitute some of the major failure modes in a high voltage switching device, may manifest in an abnormal vibration of the high voltage switching device.

[0004] Known methods for vibration signal analysis were primarily developedon rotating machinery, such as pumps and motors, which operate continuously for at least several seconds at a time. However, for high voltage switching devices, the vibration and shock signals from some operations (e.g., switching) are very short (e.g., tens of milliseconds) and are also noisy.

[0005] Therefore, improved monitoring methods are required for high voltage switching devices.Summary

[0006] Aspects of the present disclosure deal with condition assessment from vibration signals of HVSDs during switching, including opening, closing, and combinations thereof.

[0007] It is realized as a part of the present disclosure that there is a problem when attempting to accurately monitor the vibration signal of an HVSD. That is, the variance of the vibration signals and features has numerous sources. Conventional vibration monitoring techniques may rely on a ‘good’ or ‘normal’ HVSD signal as reference, or a catalog of signals indicative of fault states, but how these reference signals are defined is non-trivial. There are numerous sources of variance that influence a ‘good’ or ‘normal’ signal.

[0008] As one example, there may be a natural variance in the vibration signal and features of a HVSD from one switching operation to another, i.e. , vibration features are not constant over time due to aging, e.g., from Tun-in’ (early life) to ‘wear-out’ (end of life), and other effects, including from unknown or random causes. Differences in mechanical tolerances, age and operating speed also influence the variances for individual HVSDs.

[0009] As another example, the specific placement and application of vibration sensors can introduce variations in the vibration signal measured, for example, from different attachment methods (adhesive wax, bolt, magnet etc.), temperature, slight sensor misplacements, or misorientations.

[0010] These example sources of variances may be present even for ahealthy or ‘good’ HVSD. In a non-healthy state, certain vibration features may be affected, but these differences must be distinguished from the other sources of variance.

[0011] Because of these factors that give rise to variance in the vibration signal, defining a ‘good’ vibration signal is, as mentioned before, non-trivial. Moreover, the raw vibration signal, and especially the root-mean-square (RMS) and peak acceleration, might be of limited diagnostic value. Therefore, a new method to analyze vibration signals is required to better compare measurements for diagnostic and prognostic purposes.

[0012] Aspects of the present disclosure are therefore directed at providing such an improved monitoring method for a high voltage switching device (HVSD). Specifically, it is an object of the present disclosure to improve the monitoring of vibration of a HVSD.

[0013] According to a first aspect of the present disclosure, a (computer-implemented) method for condition monitoring of a high voltage switching device is provided. The method comprises obtaining a vibration signal over time comprising vibration information for a high voltage switching device (during operation, such as at least during a switching event of said high voltage switching device), and segmenting the vibration signal into a plurality of segments. The method further comprises scaling at least one segment based on at least one other segment to thereby obtain a scaled segment, and determining a condition of the high voltage switching device based on the scaled segment.

[0014] Obtaining a vibration signal may comprise obtaining a measurement from a vibration sensor arranged on or near the HVSD. Measuring vibrations in an HVSD is typically done using an accelerometer to continuously record vibration patterns during opening and closing operations of an HVSD.Accelerometers can be placed in various locations in, on, or near a HVSD. Placing an accelerometer close to the operating mechanism, e.g., close to the tripping / opening coils or the electrical contacts, can result in vibrationsmeasured during the release of stored energy and the movement of the contacts. Placing the accelerometer close to the housing, on the other hand, may result in measured vibrations that are transmitted through the structure, indicating overall mechanical health and structural integrity. Accelerometers placed along the operating mechanism shaft or the bearing assembly can in turn detect vibrations related to bearing conditions that can be indicative of wear and alignment issues. Mounting positions can be chosen such that resonance effects, which can distort the vibration signal, are minimized.

[0015] The vibration signals are preferably measured at a measurement frequency higher than the expected vibration frequency. For example, if it is expected that a switching event may generate vibrations up to 5 kHz, then a measurement frequency of approximately 10 kHz or more may be selected.

[0016] It will be appreciated that additional information may also be obtained, as well as the vibration signal. Obtaining such additional information about the performance of the HVSD during a switching event can for example be done by measuring the position of the electrical contacts, monitoring the temperature in the HVSD, measuring the response time of the HVSD during opening and closing operations.

[0017] A switching event may be triggered automatically or manually, and occur due to load switching, overloads, short circuits, ground faults, or manual / remote operator commands. In case of load switching, the HVSD is used to connect and disconnect parts of the electrical grid to direct the flow of electricity. In case of an electrical load exceeding an HVSD’s rated capacity, the HVSD may switch, or ‘trip’ or ‘trigger’, to damage to the HVSD and other connected equipment, such as a power transformer. In case of a short-circuit, the HVSD may trip to prevent damage to the system and other connected equipment from high current flow. When a ground fault occurs, an HVSD may trip to prevent electric shock and connected equipment damage due to voltage surges. Electrical equipment can be vulnerable to variations in voltage and current fluctuations, which is why HVSDs may trip in order to protect thesystem. The (re-)closing command of a switched, or tripped, HVSD may be done manually by an operator after a switching event has occurred or may be done through remote operation via remote control systems. Some HVSDs may have automatic reclosing features to restore power after a temporary fault. The HVSDs having automatic reclosing features may, for example, close after a preset time.

[0018] The obtaining of the vibration signal may be initiated by a switching command for causing the high voltage switching device to enact the switching event. A switching event can be characterized from the time the trip mechanism is activated until the contacts reach their end travel position, including potential damping or rebounding movement. A switching event may also be characterized from the time the trip mechanism is activated until the HVSD is restored to normal operation. It will be appreciated that a switching event may also include the period(s) of time before and / or after the movement of the switch contacts.

[0019] The switching command may be given by another protection device configured to detect faults such as a fault to ground or a short circuit upon which the high voltage switching device acts. Obtaining the vibration signal only during switching events may be beneficial as it may allow for data collection only under a relevant period of time. Therefore, triggering the obtaining of the signal only when a switching event occurs may save computational and memory resources.

[0020] In one example, a vibration sensor may substantially constantly record, monitor, or otherwise collect a vibration signal for a HVSD. The vibration signal may thus take the form of a datastream signal or the like. The vibration signal may be stored on a temporary memory such as a flash memory, and then purged therefrom in a batch or windowed manner if no switching event has occurred. Then, the occurrence of a switching event (e.g., indicated by a command for the HVSD to open) may cause the vibration signal stored in the temporary memory to be sent for further processing. Inthis way, it can be ensured that the pre-opening or opening of the HVSD is properly represented in the vibration signal.

[0021] Vibration measurements from the vibration sensor collected in a period of time after the start of switching event may be sent directly for further processing, either from the temporary memory or bypassing the temporary memory. The end of the relevant period of the vibration signal may be identified as the expiry of predetermined time period (e.g., 30 or 50 ms), as a period of consistently low-amplitude vibration, or using the ‘travel curve’, i.e. , the monitored position of the electrical contacts of the HVSD, if this additional information is available.

[0022] Segmenting the vibration signal into a plurality of segments is advantageous as it allows for a relative comparison of different parts of the vibration signal which represents a switching event. For example, the vibration occurring during the closing of the switch may be compared with the vibration occurring during subsequent closings of the switch. Comparing segments may thus remove the need to define or identify any reference signal, thereby addressing at least some of the problems related to variance as discussed before.

[0023] Segments in the sense of the present disclosure may relate to portions, partitions, subsets, bands or windows in the vibration signal. The vibration signal may be in the time-, frequency, or time-frequency domain. Therefore, the segments may divide the signal by ‘cutting’ the signal at one or more points along the time or frequency dimension. One example of a way to segment the signal is by dividing the signal into arbitrary time segments (e.g., every 10ms), or arbitrary frequency segments.

[0024] It may be preferred to capture or isolate respective moments or landmarks of the switching event in respective segments — for example, a segment containing a respective part of the vibration signal captured during the initial opening of the switch, the damping of the switch contacts (e.g., to slow their motion before arriving at the extreme open position), and the finalfull opening of the switch. This way of segmenting the vibration signal may be advantageous because a mechanical failure of a HVSD may manifest, at least to start with, in one of these motions of the switch. Indeed, each landmark (sub-event) of a switching event may be associated with a different mechanical component, such as a damping being carried out by a damper, or a spring releasing energy. Hence, a degradation of the component (e.g., the damper) may lead to a change in the vibration signal during its associated landmark (e.g., damping) but may not affect other landmarks (e.g., initial contact separation).

[0025] Therefore, according to example implementations of the present disclosure, there are provided techniques for a dynamic and intelligent segmentation.

[0026] According to one such example, the segmenting of the vibration signal may comprise determining, in the vibration signal, one or more events having relatively higher vibration intensity and segmenting the vibration signal such that each segment comprises a respective higher intensity event. According to this example, a number of amplitude peaks may be identified in the signal, such as by identifying a predetermined number of strongest peaks or by identifying peaks which exceed a particular threshold. Each peak may be assumed to correspond to a different landmark event, such as initial contact separation or the final fully-open state of the contacts. Thus, segments may be defined such that each segment contains only one of the peaks. A peak value is also a valuable characterizing value for a segment, and thus isolating the main peaks in the vibration signal in respective segments allows any characteristic values derived therefrom to convey more complete information about the vibrations signal as a whole. Further according to this example, a (running) average across the signal may be obtained to identify areas of relatively high intensity.

[0027] According to another example, the segmenting of the vibration signal may further comprise processing the vibration signal to thereby obtain asmoothed vibration signal and determining, in the smoothed vibration signal, one or more minima. After determining the minima, the vibration signal may be segmented based on said one or more minima, e.g., by placing segment boundaries at each minimum. It will be appreciated that a raw vibration signal may be a high frequency and noisy signal. Hence, smoothing the vibration signal (e.g., by a running average or the like) may be beneficial because it may reduce the impact of noise and fluctuations in the signal. In such a smoothed signal, the areas of high intensity vibration (e.g., during opening, damping, and closing of the switch) may be readily identified as peaks.Hence, defining segments based on minima between these peaks advantageously simplifies the capture of each part of the switching event (i.e. , each landmark) within its own respective segment. It will be appreciated that the same process could be applied for maxima (or any inflection point, of which a minimum and a maximum are examples).

[0028] It is appreciated, however, that the segmenting can be done according to different principles, depending for example on the amount of additional information about the monitored asset is available. With additional information it is here meant for example a travel curve, coil current, or contact separation. These features may allow the segmentation to be done according to key events such as ‘coil start’, ‘coil end’, ‘travel start’, ‘contact make’, ‘contact break’, ‘damping’ and / or ‘travel end’. Other key events might be additionally identified. The segmentation may be done automatically or manually.

[0029] The segmenting of the vibration signal may, according to yet a further example, comprise determining, in the vibration signal, one or more characteristic features corresponding to respective landmarks of the switching event and segmenting the vibration signal based on the one or more characteristic features such that each of the plurality of segments substantially isolates each landmark of the switching event. The landmark may be established from domain knowledge of the HVSD, for example the type of high voltage switching device. Knowledge of the type of switching device is advantageous because it may facilitate identification of a landmarkand, consequently, a characteristic value. The characteristic features may also be derived from additional, associated, information such as a travel curve (i.e. , the position of the electrical contacts of the switch over time).

[0030] Segmenting the vibration signal based on the one or more characteristic features such that each of the plurality of segments substantially isolates each landmark of the switching event may facilitate a more accurate detection of a potential fault. If a fault is identified in one of the segments, it is possible to match the relevant segment to a respective landmark of the switching event. Knowing during which phase of the operation of an HVSD a fault has occurred means that more targeted maintenance actions may be taken that correspond to the specific phase. A maintenance action for a fault during opening may, for example, be different compared to a maintenance action for a fault during closing.

[0031] The landmarks of the switching event may comprise, e.g., an opening of the high voltage switching device, a pre-close damping of the high voltage switching device, and a closing of the high voltage switching device. The mechanical movement of the high voltage switching device may be dampened in order to manage the mechanical energy generated during the opening and closing operations, ensuring smooth and controlled movement. High voltage switching devices may implement different damping mechanisms such as oil damping or piston and cylinder damping. Damping requirements are generally more critical during opening operations but damping the closing operation may ensure a smooth and controlled closure.

[0032] Landmarks comprising the mentioned different stages of a switching event are advantageous as the vibration signal may be segmented into segments with distinct vibration patterns. During the initial opening of a high voltage switching device, an arc forms, causing a sudden release of energy, generating significant vibrations from movement of the electrical contact system. When the contacts of an HVSD continue to open, damping mechanisms may come into play to slow and control the movement.Vibrations may still be present, but are moderated. In the final stage, when the contacts are fully separated, the vibrations may be reduced significantly. Residual vibrations from the mechanical components may still be detected, but they may be less intense compared to the vibrations of the initial opening stage. During closing, vibrations may be generated again due to the mechanical impact and re-establishment of the electrical connection where an arc may form briefly, and damping during closing.

[0033] The method as provided by the first aspect of the present disclosure may further comprise obtaining a characteristic value for each of the plurality of segments. A characteristic value may be a value that that refers to a statistical or spectral measure that concisely characterizes the properties of the signal within each segment, preferably as a single numerical value.Defining the segments by their characteristic value is advantageous as it may allow for a more concise or compressed representation of the signal, thereby reducing the need for computational power and data storage, and simplifying the scaling of segments relative to each other. In other words, the characteristic value may allow for simplification of the complex vibration signal into manageable and interpretable metrics.

[0034] The characteristic value may be, for example, a root mean square (RMS) or a peak intensity. The RMS provides a measure of the average power of energy content of the vibration signal, which may contribute to overall understanding of the intensity of the vibrations. RMS is sensitive to both the amplitude and the duration or length of a segment, making it effective in capturing the true magnitude of the vibrations. By averaging the squared values, RMS may help to reduce the impact of random noise, providing a more stable and reliable measure. Furthermore, identifying a peak intensity does not require complicated algorithms or high computational power. Therefore, using a peak intensity as a characteristic value does not increase the complexity of the method, while it may contribute to the method being more robust to for example noise or other disturbances. Both RMS and peak intensity may have the benefit of mathematical simplicity and being astandard metric. RMS and peak intensity are mere examples of characteristics values. Others include vibration energy, entropy, number of peaks, time between peaks, or frequency between peaks. Characteristic values may be derived from the raw signal segments or treated (e.g., smoothed or Fourier-transformed) signal segments.

[0035] Scaling at least one segment based on at least one other segment to thereby obtain a scaled segment is advantageous as it may facilitate analysis and relative comparison of the segments between each other. In other terms, the scaling can be described as normalizing each segment to a common range. This can be done using various methods. Scaling the segments may also allow for noise reduction and may readily allow for the identification of ‘outliers’, ‘anomalies’, or ‘trends’ in the scaled information. In the analysis of high voltage switching devices, outliers may indicate a failure or degradation and may therefore be important to be identified.

[0036] The scaling the at least one segment based on at least one other segment may comprise scaling the at least one segment or the characteristic value of the at least one segment based on the characteristic value of the at least one other segment. By scaling based on a characteristic value, the scaling operation may become simplified and may require less computational power. This is advantageous as it means that the calculation may be performed at a higher speed, which may be required in for example real time applications.

[0037] The at least one segment may be scaled relative to the at least one other segment based on said at least one other segment being determined as substantially invariant. For example, a degraded or faulty HVSD may experience more vibration during the damping of the switch during operation as compared to a non-degraded HVSD. Therefore, it is appreciated that in the case the segmenting has been done according to, for example, key events, the segment corresponding to the initial contact opening of the HVSD may be determined to be the substantially invariant segment. Upon scaling of the atleast one segment, which may for example be the damping segment, a fault may be readily detected if the segment has a characteristic value that is higher relative to the characteristic value of the invariant segment compared to a previous operation of the same switch or a contemporaneous operation of a similar switch (e.g., in the same bank, or bay, or one or more substations of switches). According to some example implementations, the operation of an HVSD may be simulated so as to identify operations that result in substantially invariant vibration signals. The simulated vibration signals may be used for further comparison.

[0038] Determining a condition of the high voltage switching device based on the scaled segment is advantageous as accurately determining the health status of a high voltage switching device may contribute to the remaining useful lifetime estimation of the device and to the safety of the system. Based on the condition of the monitored high voltage switching device, appropriate maintenance actions may be determined to prevent failure of the device. The increased lifespan lowers the need for replacement of the HVSDs.Replacement is highly costly and complicated, hence reducing the need for replacement may improve the efficiency, both economically and functionally, of the systems the HVSDs are operating in.

[0039] Determining a condition or state of health of the HVSD based on the one or more scaled segments may comprise comparing the one or more scaled segments to a previously obtained scaled segment for the same HVSD or comparing the one or more scaled segments to a scaled segment obtained for another HVSD of the same or similar type. If measurements and scaled segments are available for the same HVSD over time, comparing segments may provide insights regarding performance at various stages of its life. The segments can be compared to identify faults or degradation in the HVSD. If, on the other hand, measurements and scaled segments are not available overtime, an instantaneous comparison, or in other words a comparison at a single point in time, can be made between HVSDs. These comparisons can be made to identify anomalous HVSDs within a bay or fleetof HVSDs. When an anomalous HVSD is detected, appropriate maintenance actions may be taken.

[0040] The monitoring method as described above may preferably be deployed in a high voltage switching device controller, or a grid-level controller. The method is computer implemented and may be carried out by a data processing system comprising means for doing so. It will be appreciated that any such data processing system may be situated in any suitable infrastructure, for example in a switching system local to the HVSD or in a cloud-based solution.

[0041] Thus, according to a further aspect of the present disclosure, there is provided a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the monitoring method as described above. According to yet a further aspect of the present disclosure, there is provided a computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the monitoring method as described above.

[0042] According to another aspect of the present disclosure, a power grid system is provided. The power grid system comprises a high voltage switching device, a data processing system according to the present disclosure, a vibration sensor adapted to collect a vibration signal for the high voltage switching device and provide the vibration signal to the data processing system.

[0043] Although some advantages have been explicitly laid out above, those skilled in the art will immediately recognize, from the disclosure herein, further possible advantages provided by aspects of the present disclosure.Additionally, advantages provided by the monitoring method as such will inevitably precipitate into advantageous aspects of any data processing system executing such a method, or any high voltage switching device being monitored according to such a method. Indeed, it will be appreciated that the method disclosed herein could be generalized to any system which producestransient (non-steady state, short) vibration signals, such as actuators (Thomson coil, hydraulic actuator, solenoid, stepper motor, and so forth). Further advantages may become apparent through the consideration of specific examples, such as those laid out below.Brief Description of the Drawings

[0044] One or more example implementations of the present disclosure will be described below, with reference to the following drawings (or ‘figures’), in which:

[0045] Figure 1A schematically shows a power grid system comprising according to aspects of the present disclosure;

[0046] Figure 1B schematically shows a high voltage switching device in various stages of an opening operation;

[0047] Figures 2 illustrates a travel curve and a vibration signal for a high voltage switching device during a switching event;

[0048] Figures 3A illustrates a segmenting of a vibration signal based on landmarks identified from an associated travel curve;

[0049] Figure 3B illustrates segmenting of a vibration signal based on identified high intensity areas;

[0050] Figure 3C illustrates segmenting of a vibration signal based on the smoothed signal;

[0051] Figure 4 illustrates a process for condition monitoring a high voltage switching device according to aspects of the present disclosure;

[0052] Figure 5 illustrates a method for condition monitoring a high voltage switching device according to aspects of the present disclosure.Detailed Description

[0053] The present disclosure is described in the following by way of a number of illustrative examples. It will be appreciated that these examples are provided for illustration and explanation only and are not intended to be limiting on the scope of the present disclosure. Instead, the scope of the present disclosure is defined by the appended claims.

[0054] Furthermore, although embodiments be presented individually for the sake of focused discussion of particular features, it will be recognized that the present disclosure also encompasses combinations of the embodiments described herein.

[0055] Figure 1 schematically shows a power grid system 100 comprising a high voltage switching device 110, a data processing system 120 and a vibration sensor 130 adapted to collect a vibration signal for the high voltage switching device 110 and provide the vibration signal to the data processing system 120. The power grid system 100 is installed a power grid in a manner understood by those skilled in the art but the full details of the surrounding system(s) are outside the scope of the present disclosure. The power grid system 100 may be positioned at an electrical substation or at other locations where the possibility to interrupt the current flow is desired.

[0056] The HVSD 110 comprises two electrical contacts 102a and 102b. The electrical contacts 102 are used for closing and opening the HVSD 110 such that the HVSD 110 has an open position and a closed position defined by the relative positions of the electrical contacts 102. In the closed position, power is allowed to flow through the HVSD 110, and in the open position power is prevented from flowing through the HVSD 110. The contacts 102 can switch between the open and closed position, triggered by for example an electrical grid fault or the removal of a fault. A transition of the electrical contacts 102, i.e., a switching operation of the HVSD 110, may for example be a closing operation where the electrical contacts 102 move from open to closed, an opening operation where the electrical contacts 102 move from closed to open, or a close-open operation where the electrical contacts 102 move fromopen to closed, then back to open. Further, an open-close-open operation may also be performed in a short time. In this operation the electrical contacts 102 first move from closed to open for example when a fault is detected, and after a while the HVSD 110 may try to close the electrical contacts 102 to establish whether the fault is removed. If the fault is not removed, a closeopen operation will be performed.

[0057] In this example, the vibration sensor 130 is an accelerometer configured to measure vibrations in the housing of the HVSD 110. The vibration sensor 130 can continuously measure vibrations in the HVSD or can be triggered to start measuring when a switching event occurs and to stop measuring when the switching event is deemed to be finished. Such a trigger can, for example, come from another protection device on the power grid configured to detect a fault and communicatively connected to the vibration senor, or from a device configured to mechanically or electrically detect a switching event.

[0058] The data processing system 120 is connected to the vibration sensor 130 at least so that the vibration sensor 130 can provide a vibration signal to the data processing system 120. This connection may be through a wired or wireless communication. The data processing system 120 can, for example, comprise internal or external data storage means where the obtained vibration signal can be stored. In the case of the vibration sensor 120 continuously measuring, the measured signals may be stored temporarily before they are being disregarded when no switching event has occurred as to not over overload the data processing system 120 by accumulating an excessive amount of data. When a switching event is detected, on the other hand, data before, during and after the event may be stored for a longer period to allow for analysis of the signal.

[0059] The data processing system 120 can comprise a computer program 122 comprising instructions which, when executed by a computer on the data processing system 120, cause the computer to carry out the monitoringmethod as disclosed herein.

[0060] Figure 1 B schematically shows a high voltage switching device 110 in operation. Figure 1 B(a) shows an HVSD 110 with its electrical contacts 102a, 102b in closed position, similar to the position shown in Figure 1A. In closed position power is allowed to flow through the HVSD. Upon triggering of the switching mechanism, the electrical contacts 102a and 102b start moving apart. Figures 1B(b) and 1B(c) show how the electrical contacts 102a and 102b move away from each other. In the embodiment shown in the drawings, electrical contact 102b moves away from electrical contact 102a, but it is appreciated that electrical contact 102a can move away from electrical contact 102b, or that both electrical contacts 102a and 102b move away from each other (a so-called double motion operation principle).

[0061] Movement of the electrical contacts 102a and 102b is typically facilitated by a spring-loaded or hydraulic operating mechanism or the like that rapidly pulls the contacts 102a and 102b apart. The opening sequence may take only 10s of milliseconds, and the accelerations may be as high as 1000g or higher, hence the substantial risk of wear, degradation, or damage to the HVSD 110.

[0062] As the contacts 102a and 102b start to separate, such as shown in Figure 1 B(b), an electrical arc forms between them. Different types of HVSDs use various methods to cool and extinguish this arc. Air (or other types of gas such as SFe) HVSDs, for example, use compressed air to blow out the arc. Oil HVSDs, on the other hand, use arch quenching or electrical insulation oil to extinguish the arc. It is appreciated that there are different types of HVSDs suitable for different applications based on the requirements of for example the location on the power grid they are positioned on. Details about which type of HVSD to use and how they quench the electrical arc that arises upon separation of the electrical contacts are known to those skilled in the art and are outside the scope of the present disclosure.

[0063] As the electrical contact 102b approaches its extreme of movement(i.e. , the extreme of its downward movement as shown in the figure), it may have a very high velocity and may need to be decelerated before it reaches the end of the motion. This deceleration may comprise damping. In some examples, the opening of the contact 102b may also be damped.

[0064] The electrical contacts 102a and 102b are fully moved apart in Figure 1 B(c), and no power is allowed to flow through the HVSD 110, i.e., the HVSD 110 has reached its fully open position. In some switching events, the contact 102b may rebound from the fully open position, i.e., it may bounce as a result of reaching the extreme end of the motion. This rebound may be visible in the travel curve and / or the vibration signal.

[0065] Figure 2 illustrates a travel curve 200 of the electrical contacts, for example those shown in Figure 1 A and 1 B, of an HVSD and a vibration signal 220 of an HVSD. Starting with the travel curve 200, the electrical contacts are in a closed position at time 202. The descent of the curve from 202 to bump 204 indicates the opening of the electrical contacts, which may be damped towards the end, wherein the bump 204 may be caused by a rebound of the movement. At 206, the electrical contacts have been fully opened, and at 208 the electrical contacts are practically stationary, waiting to either be automatically or manually closed.

[0066] The vibration signal 220 shows the measured vibration during the operation of the HVSD during the switching event shown in the travel curve 200. Three distinct areas of higher vibrations, 222, 224 and 226 can be seen in the vibration signal 220, each corresponding to a key event, or landmark, of the switching operation. The first vibration 222 is caused by the triggering of the switching mechanism followed by initial separation of the electrical contacts. The second vibration 224 is caused by the damping of the moving contacts that ensures a smooth and safe opening of the electrical contacts. The last vibration as seen in the signal, 226, is caused by the contact finally reaching the extreme fully open position.

[0067] As seen in the vibration signal 220, even after full separation of theelectrical contacts, some vibrations may still be measured. This may be caused by vibrations propagating through the HVSD and / or the surrounding system, and those vibrations being picked up in the vibration signal because of a particular positioning of the vibration sensor.

[0068] Figures 3A, 3B and 3C show examples of ways to segment an obtained vibration signal 340 according to different example implementations of the present disclosure.

[0069] Figure 3A illustrates a travel curve 300 and a vibration signal 340 of an HVSD divided into segments 310, 320 and 330. The travel curve 300 may be obtained as additional information and is used to identify key events, or landmarks, of the switching event, based upon which the vibration signal 340 is segmented. In the embodiment as shown in the drawings the segments correspond to the key events or landmarks as described in Figure 2, e.g., the tripping of the switching mechanism and initial opening of the electrical contacts 310, the damping of the contacts 320, and the full opening of the contacts 330. With knowledge of these key events, or landmarks derivable from the travel curve 300, the corresponding vibration signal 340 can be segmented accordingly.

[0070] Figure 3B shows a different way of segmenting the vibration signal 340 in case there is no additional information available such as the travel curve 300 as shown in Figure 3A. Instead, the vibration signal is divided into major high intensity segments 350, 360, 370 and 380. It is appreciated that segmenting the vibration signal 340 into major high intensity segments corresponds to segmenting the vibration signal 340 according to key events, as shown in Figure 3B, as most key events lead to a peak in vibration intensity. It is noted, however, that in this particular embodiment the segmenting the vibration signal 340 into major high intensity segments leads to more segments compared to the segmentation in Figure 3A. For example, the initial tripping and opening of the electrical contacts were segmented into the same segment 310 in the segmenting as disclosed in Figure 3A, while thiskey event has been divided into two high intensity segments 350 and 360 in the segmenting as disclosed in Figure 3B. Segmenting the initial tripping and opening of the electrical contacts can be beneficial as it can facilitate a more accurate monitoring of this specific key events. A more detailed segmentation can lead to more precisely capturing invariant features, as will be discussed later on, which in turn may result in more accurate scaling and comparison.

[0071] Figure 3C schematically shows a vibration signal 340 together with a smoothed signal 342 divided into segments 352, 362, 372 and 382. The smoothed signal 342 can be, for example, a running average of the vibration signal 340. It is appreciated that the vibration signal 340 can be smoothed according to a variety of methods known to those skilled in the art, such as high / low pass filters, gaussian filters, median filters and the like.

[0072] The smoothed signal 342 as shown in the figure follows the frequency intensity of the vibration signal 340 and comprises minima in between the highest intensity points. The vibration signal 340 can be segmented according to these identified minima. For example, the center point between two adjacent minima can be identified as the border of two segments. It will be appreciated from this example that smoothing the signal can be beneficial as it may reduce noise and simplify the vibration signal 340, allowing for a more straightforward segmentation.

[0073] Figure 4 illustrates a process for condition monitoring an HVSD which is an example implementation of aspects of the present disclosure. A vibration signal 400 is obtained and segmented into, in this particular embodiment, three segments 402, 404 and 406. The segmenting of the vibration signal can be done according to any of the segmenting methods as described above, or according to a different method.

[0074] In this illustrated example, once the vibration signal 400 is obtained and segmented into segments 402, 404 and 406, characteristic values CV1 , CV2, CV3 are obtained for each segment. A characteristic value is a value that contains part of, or all of, the information of a segment in a single value.Examples of a characteristic values are root mean square (RMS) value, peak intensity, energy, entropy and peak to peak time or frequency.

[0075] Compressing each segment of the segments 402, 404 and 406 into single characteristic values CV1 , CV2 and CV3, allows the obtained data, e.g., the vibration signal, to require less data storage and thus be computationally efficient when it is stored, processed and / or analyzed on a data processing system. It is also more computationally simple to compare characteristic values with each other, which may be simply individual numbers, rather than comparing segments of the signal with many datapoints.

[0076] Once the characteristic values CV1 , CV2 and CV3 are obtained, it is determined whether or not an invariant segment is present (yes, Y, or no, N). At least in this example, an invariant segment is a segment that will not, or not substantially, change even when a fault may be present. Alternatively, or additionally, the characteristic value of such a segment does not, or not substantially, change when a fault may be present.

[0077] When an invariant segment is present and identified, each segment 402, 404 and 406 is scaled only by the invariant segment. This means, for example, dividing each characteristic value of each segment by the characteristic value of the invariant segment, thereby obtaining normalized characteristic values for each segment with respect to the characteristic value of the invariant segment. In the present embodiment this may include identifying segment 402 as the invariant segment. Then, CV1, CV2 and CV3 are divided by CV1 to obtain scaled segments in the form of scaled characteristic values. The scaled characteristic value C1 of the invariant segment 402 will, naturally, be 1.

[0078] When no invariant segment can be identified, each segment 402, 404 and 406 is scaled by each other segment 402, 404 and 406. This means, for example, dividing each characteristic value CV1 , CV2 and CV3 of each segment 402, 404 and 406 by the characteristic values CV1 , CV2 and CV3 of each of the other segments 402, 404 and 406. In the present embodiment thismay include performing the following divisions: CV1 / CV2, CV1 / CV3, CV2 / CV1, CV2 / CV3, CV3 / CV1 and CV3 / CV2. In other words, ratios of the characteristic values are determined, and in a later stage compared.

[0079] Figure 5 illustrates a method 500 for condition monitoring an HVSD according to aspects of the present disclosure. The first step comprises obtaining a vibration signal 501. This can be achieved through a sensor, such as for example an accelerometer, that measures vibrations in the HVSD as discussed before.

[0080] The next step is segmenting the vibration signal into segments 520 and can be done using additional information, or arbitrarily. After the segmenting 520, a segment is scaled based on another segment 530.Depending on whether an invariant segment has been identified or not, the segment is either scaled based on this invariant segment only, or else it is scaled based on all other segments.

[0081] Finally, a condition of the HVSD based on the scaled segment is determined 540. If an invariant segment has been identified in during the scaling 530, the scaled segment can be compared to this invariant segment. Otherwise, the scaled segment can be compared to, for example, a scaled segment of the same HVSD during another point in time. Alternatively, or additionally, the scaled segment can be compared to a scaled segment of another HVSD in, for example, the same substation or the same fleet in order to detect an anomalous HVSD.

[0082] Purely to illustrate the use of an invariant segment, the following example is given: it is determined that the invariant segment corresponds to the key-event or landmark of the contact switches of an HVSD being fully open. It may be further known that, for example, during normal operation, the beginning of a switching event typically has a highest peak intensity (CV2) three times higher than the highest peak intensity of the invariant segment (CV1). Put another way, CV2 / CV1 = 3 for a healthy switch. Here, the highest peak intensity is chosen as a characteristic value (CV). Next, a vibrationsignal is obtained and segmented and the segment corresponding to the beginning of the switching event is scaled by the characteristic value of the invariant segment CV1. The scaled characteristic value of the segment allows a condition of the HVSD to be determined. If the results deviate from the known ratio (e.g., CV2 / CV1 » 3), the assumption can be made that a fault may be present.

[0083] While the present disclosure is susceptible to various modifications and alternative forms, specific embodiments are shown and described above by way of example in relation to the drawings, with a view to clearly explain the various advantageous aspects of the present disclosure. It should be understood, however, that the detailed description herein and the drawings attached hereto are not intended to limit the disclosure to the particular form disclosed. Rather, the intention is to cover all modifications, equivalents, and alternatives falling within the scope of the following claims.

Claims

24Claims1. A computer-implemented method for condition monitoring of a high voltage switching device, the method comprising:obtaining a vibration signal overtime comprising vibration information for a high voltage switching device;segmenting the vibration signal into a plurality of segments; scaling at least one segment based on at least one other segment to thereby obtain a scaled segment; anddetermining a condition of the high voltage switching device based on the scaled segment.

2. The method according to claim 1 , further comprising obtaining a characteristic value for each of the plurality of segments.

3. The method according to claim 2, wherein scaling the at least one segment based on at least one other segment comprises scaling the at least one segment or the characteristic value of the at least one segment based on the characteristic value of the at least one other segment.

4. The method according to claim 2 or claim 3, wherein the characteristic value is a root mean square (RMS) or a peak intensity.

5. The method according to any preceding claim, wherein the obtaining of the vibration signal is initiated by a switching command for causing the high voltage switching device to enact the switching event.

6. The method according to any preceding claim, wherein segmenting the vibration signal comprises:determining, in the vibration signal, one or more characteristic features corresponding to respective landmarks of the switching event; and segmenting the vibration signal based on the one or morecharacteristic features such that each of the plurality of segments substantially isolates each landmark of the switching event.

7. The method according to any preceding claim, wherein landmarks of the switching event comprise an opening of the high voltage switching device, a damping of the high voltage switching device, and a closing of the high voltage switching device.

8. The method according to any of claims 1 to 5, wherein segmenting the vibration signal comprises:determining, in the vibration signal, one or more events having relatively higher vibration intensity; andsegmenting the vibration signal such that each segment comprises a respective higher intensity event.

9. The method according to any of claims 1 to 5, wherein segmenting the vibration signal comprises:processing the vibration signal to thereby obtain a smoothed vibration signal;determining, in the smoothed vibration signal, one or more minima; and segmenting the vibration signal based on said one or more minima.

10. The method according to any preceding claim, wherein the at least one segment is scaled relative to the at least one other segment based on said at least one other segment being determined as substantially invariant.

11. The method according to any preceding claim, wherein determining a condition or state of health of the high voltage switching device based on the one or more scaled segments comprises:comparing the one or more scaled segments to a previously obtained scaled segment for the same high voltage switching device; or comparing the one or more scaled segments to a scaled segmentobtained for another high voltage switching device of the same type.

12. A data processing system comprising means for carrying out the method of any preceding claim.

13. A high voltage switching system comprising:a high voltage switching device;the data processing system according to claim 12; anda vibration sensor adapted to collect a vibration signal for the high voltage switching device and provide the vibration signal to the data processing system.

14. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of any of claims 1 to 11.

15. A computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the method of any of claims 1 to 11.

Images