Gas-insulated switchgear

Abstract

The invention relates to an operating method for a gas-insulated switchgear (1) having a gas-tight housing portion (8), which is filled with an insulating gas, and to such a switchgear, wherein • measurement data of a sensor (10) for detecting the pressure in the housing portion (8) is detected, • a first value (tl), which is a measure of the remaining service life, is determined using sensor (10) measurement data which extends over a first period of time (ZI), • a second value (t2), which is a measure of the remaining service life, is determined using sensor (10) measurement data which extends over a second period of time (Z2), the second period of time (Z2) being at least ten times as long as the first period of time (ZI), and • a common evaluation variable (BG) for the remaining service life is determined from the first and second value (tl, t2), the first value (tl) being compared with a first time corridor (ZK1) for this purpose, the first time corridor (ZK1) being at most ten times as long as the first period of time (ZI).

Description

[0001] Description

[0002] Gas-insulated switchgear

[0003] The invention relates to a gas-insulated switchgear with a device for detecting when a defined gas density inside is undershot, e.g. a pressure sensor for detecting the internal pressure.

[0004] Gas-insulated switchgear (GIS) is electrical switchgear in which switching and protection functions are achieved through gas-tight partitions and insulating gases such as sulfur hexafluoride (SF6). It is used to switch, control, and protect electrical energy in high-voltage networks. Gas-insulated switchgear offers a compact design, high operational reliability, and reduced maintenance requirements. It is primarily used in urban areas where space is limited but a reliable power supply is essential.

[0005] Gas-insulated switchgear is valued for its maintenance-free operation and long service life. However, this is only true if sufficient insulating gas is present in the sealed container. Accordingly, a density or pressure monitoring system is integrated into these systems. Since the pressure of a given quantity of insulating gas in a sealed volume depends on the temperature, a temperature-compensated value of the absolute pressure referenced to zero is relevant for assessing insulating gas loss. The term "pressure" will be used in the following discussion. Devices and methods for determining temperature-compensated absolute pressure values ​​are state of the art.

[0006] For the operator, it is important to know at what point a defined gas density is undershot and how much time remains for a reaction. In the following text, the term pressure is also used in relation to gas density. In this context, the terms are used synonymously. Determining the point in time is complicated by the fact that the loss of insulating gas density is a complex function of many fluctuating physical parameters such as ambient temperature, operating current, aging behavior of the sealing materials, and the cause of the pressure drop. Causes include, for example, permeation loss through seals, micro-leaks or micro-cracks in welds, or even large leaks. Since the signal can be very unstable due to this complex behavior, appropriate data processing is advantageous. One method for stabilizing the signal using a calculated standard deviation is described, for example, in EP 3 790 032.

[0007] A1 is known.

[0008] To generate an early warning, measuring systems with multiple alarm contacts are used. These contacts are activated even before the minimum fill pressure is reached. However, it is not apparent to the user how serious the problem is or what the specific courses of action are. Therefore, it is very difficult to make a generally applicable statement about the remaining service life. Each cause has different required response times.

[0009] The object of the invention is to provide a gas-insulated switchgear in which an improved prediction of remaining service life is made possible.

[0010] This problem is solved by an operating method with the features of claim 1. Another solution consists of a gas-insulated switchgear with the features specified in claim 12.

[0011] The operating method according to the invention is used in a gas-insulated switchgear with a gas-tight housing section filled with an insulating gas.

[0012] In this operating procedure, measurement data from a sensor for detecting pressure or gas density in the housing section is recorded. Using measurement data from the pressure sensor, which extends over an initial period, a preliminary value is determined, which serves as a measure of the remaining service life. For simplicity, the sensor will subsequently be referred to as the pressure sensor.

[0013] Furthermore, using measurement data from the pressure sensor, which extends over a second period, a second value is determined that is a measure of the remaining service life. This second period is at least ten times longer than the first.

[0014] A common rating for the remaining service life is determined from the first and second values ​​by comparing the first value with a first time corridor, where the first time corridor is at most ten times as long as the first period. The gas-insulated switchgear according to the invention comprises a gas-tight housing section filled with an insulating gas and a pressure sensor for detecting the pressure in the housing section.

[0015] The switchgear also includes an evaluation unit for acquiring measurement data from the pressure sensor. The evaluation unit is designed to determine a first value, which is a measure of the remaining service life, using measurement data from the pressure sensor that extends over a first period. It is further designed to determine a second value, which is a measure of the remaining service life, using measurement data from the pressure sensor that extends over a second period, wherein the second period is at least ten times longer than the first period.

[0016] Finally, the evaluation device is designed to determine a common assessment value for the remaining service life from the first and second values, whereby a comparison of the first value with a first time corridor is made, whereby the first time corridor is at most ten times as long as the first period.

[0017] The pressure sensor is a device that produces a gas pressure-dependent signal. This signal indicates the gas pressure inside the housing section or the difference between this pressure and the external pressure (relative pressure). It can be a conventional pressure sensor, such as a piezoelectric pressure sensor. However, the pressure sensor can also be a specialized sensor for the application, for example, a bellows filled with a reference gas, located inside the container, whose shape is determined by an equilibrium between the bellows' spring action, the pressure of the reference gas, and the pressure of the surrounding insulating gas. A change in shape can be detected outside the container via magnetic coupling. This visible change can be optically detected and used as an input for the algorithm.

[0018] Advantageously, the operating procedure and the switchgear enable the operator to obtain a concrete and reliable assessment of the plant's condition over a very long period. This assessment, through the combination of methods across different timeframes, provides both sufficient dynamic range for rapid processes and sufficient accuracy for slow pressure losses. Furthermore, the accuracy for the shorter initial period is advantageously evaluated by comparison with the first timeframe, allowing for a single, unified result—the key performance indicator—to be derived.

[0019] The assessment metric can be used to advantage for specific actions before the gas loss progresses to the point where the minimum filling pressure is reached. For example, in the case of a longer-term gas loss, this could involve prioritizing routine maintenance. In the case of a medium-term gas loss, a maintenance team can be dispatched directly. In the case of a short-term gas loss, the system can be shut down and the network reconfigured. The operator can then carry out activities in accordance with the desired risk management strategy. Thus, a "condition-based maintenance" concept is provided for the use case of "gas monitoring."

[0020] The invention was developed because it was recognized that no method with identical parameters exists that can predict both a pressure loss within minutes and a pressure loss over years. Therefore, a method adapted to the different time scales is used.

[0021] Advantageous embodiments of the operating method and the switchgear according to the invention are described in the dependent claims. The embodiment of the independent claims can be combined with the features of one of the dependent claims or, preferably, with those of several dependent claims. Accordingly, the following additional features can be provided:

[0022] In an advantageous embodiment of the invention, the first and second values ​​are determined as time points at which the remaining pressure in the housing section reaches a minimum filling pressure. Since different sets of measurement data are used to determine the first and second values, different time points are typically obtained, although the meaning of the result is the same. The accuracy of each time point is directly related to the time scale from which the measurements are taken, i.e., the respective time period.

[0023] In an advantageous embodiment of the invention, when determining the evaluation parameter, the first value is only considered if a comparison of the first value with the first time corridor shows that the remaining service life is no longer than the first time corridor. If, as described, the first value is the first point in time at which the pressure reaches the minimum filling pressure, then this value is only considered if it lies within the first time corridor, which extends from the point in time of the current calculation. If the first point in time is further in the future than the first time corridor allows, the first point in time is discarded as a result.

[0024] To determine the first, second, and potentially further values, a fitting process can be used on measurement data from the respective time period, in particular a fit with a linear function. A linear function in this context refers to a straight line (regression line) that does not necessarily pass through the origin. The first value is thus obtained by fitting the function to measurement data from the first time period.

[0025] Instead of a fit that results in a regression line, another function can be used, such as an exponential function, to obtain a more accurate result, especially over long periods. Alternatively, a physical model can be used and fitted to the measurement data.

[0026] The pressure sensor data is preferably acquired at intervals of at least 1 second and at most 30 seconds. Acquisition at such intervals allows the first period to last no more than 10 minutes while still enabling the use of multiple values ​​to compensate for fluctuations in the sensor signal. This allows the first period to be chosen in such a way that a prediction is possible even in the case of rapid pressure losses, i.e., large leaks.

[0027] The second period preferably has a length of between 10 minutes and 10 hours, particularly between 30 minutes and 2 hours. This period is above the thermal time constant of the switchgear. The switching station itself is air-conditioned, so no changes are expected during the day. The second period is therefore short enough to ignore longer-term fluctuations such as weekly variations or seasonal variations.

[0028] By selecting these time periods, it is advantageously exploited that, depending on the time scale, only certain aspects need to be considered with regard to both the causes and the physical parameters. For example, if a pressure loss is expected within one year, no seasonal temperature fluctuations or material aging need to be taken into account. Conversely, for a pressure loss over a long period, a significant leak is not realistic. A separate value is generated for each time scale. In an advantageous embodiment and further development of the invention, a third value is determined using measurement data from the pressure sensor that extends over a third period. This third value is a measure of the remaining service life, and the third period is at least ten times longer than the second. In other words, a third time scale is considered and evaluated, which is significantly longer than the second time scale.

[0029] If a third time period is used, in an advantageous embodiment of the invention, the second value is only considered if a comparison of the second value with a second time corridor shows that the remaining service life is no longer than the second time corridor. If, as described, the second value is a second point in time at which the pressure reaches the minimum filling pressure, then it is only considered if it lies within the second time corridor, which extends from the point in time of the current calculation. If the second point in time is further in the future than the second time corridor extends, the second point in time is also discarded as a result, since the prediction is then not sufficiently reliable.

[0030] The third period preferably has a length of between one day and one month, particularly between three and ten days. This allows for predictions over periods of several months, unlike the first and second periods.

[0031] The aforementioned time periods preferably overlap, i.e., the measured values ​​used partially cover the same period.

[0032] In a further development of the invention, values ​​for the residue of the respective fitting process are stored, and an evaluation signal for the reliability of the common evaluation parameter is determined from the trend of these values ​​for the fitting process. If the variance of the fitting process increases, the cause may be faulty temperature compensation. For this purpose, for example, a threshold value for the residue can be defined, and a warning signal can be triggered if this threshold is exceeded once or even just several times. This advantageously improves the long-term reliability of the process.

[0033] The invention will now be described and explained in more detail with reference to the exemplary embodiments shown in the figures. Figure 1 shows an exemplary schematic representation of a gas-insulated switchgear assembly.

[0034] Figure 2 shows an exemplary schematic representation of a gas holder of a gas-insulated switchgear,

[0035] Figures 3 to 5 show algorithms for determining values ​​that are each a measure of the remaining service life.

[0036] Figure 1 shows an exemplary schematic representation of a gas-insulated switchgear 1. The gas-insulated switchgear 1 has several display elements 4 and a control element 6, which can be used to perform switching operations and optionally display switching states, switching parameters, or other parameters or states. Inside the switchgear 1, not visible here, is a gas cylinder 8, which is shown by way of example in Figure 2. The gas-insulated switchgear 1 has a control unit, which is not shown in the figures.

[0037] Figure 2 shows an exemplary schematic representation of a gas holder 8 of the gas-insulated switchgear 1. The gas holder 8 has cable glands 7 on its underside (not shown) and its top side; in alternative embodiments, these glands can also be arranged on other surfaces of the gas holder, for example, on side surfaces. Furthermore, the gas space has at least one bushing 9, here a rotary bushing 9. The rotary bushing 9 is preferably designed to drive one or more switching elements inside the gas holder 8.

[0038] A pressure sensor 10 is arranged in or on the gas container 8. The pressure sensor 10 can be one of the known types of pressure sensors, for example, a piezoelectric pressure sensor. The pressure sensor 10 is preferably optimized for static pressure measurement. In certain embodiments, the pressure sensor 10 can also be a bellows filled with a reference gas, located in the container, and whose shape is determined by an equilibrium between the spring behavior of the bellows, the pressure of the reference gas, and the pressure of the surrounding insulating gas. A change in shape can be visible outside the container by magnetic coupling. This visible change can be optically detected and also used as an input for the algorithm.

[0039] The pressure sensor 10 is coupled to the controller, and the controller is configured to acquire and store measurement data from the pressure sensor 10. In other configurations, storage and processing can also take place outside the gas-insulated switchgear 1, for example, in a cloud service.

[0040] The following assumes that pressure sensor 10 provides a measurement every 15 seconds. The controller records and stores this measurement.

[0041] It then executes a first algorithm, which is schematically illustrated in Figure 3. In a first step 31, a new measurement is taken from the pressure sensor 10. Then, in a second step 32, the last 11 measurement data points from the pressure sensor 10 are taken and a fitting process is performed using a linear function, i.e., a regression line G1 is determined that best represents the measurement data used. Thus, measurement data from a first period Z1, which has a length of 150 s, is used. With this length, the first period Z1 is shorter than the thermal time constant of the gas-insulated switchgear 1 and also shorter than all other regular variations, such as daily or weekly load variations, daily temperature, or seasons. Therefore, all these influences do not need to be taken into account.

[0042] Using the regression line G1, in a third step 33 it is determined when the pressure reaches a specified minimum filling pressure p stored in the controller. M , for example, reaching 3 bar. This could also be the temperature-compensated minimum filling pressure, i.e., a filling pressure normalized to 20°C, which is then equivalent to the density. In a fourth step 34, it is checked whether the first time point ti determined in this way lies further in the future than a first time corridor ZK1, the length of which corresponds to ten times the first period Z1, i.e., 1500 s = 25 minutes. If so, the value ti determined in this way is discarded in a sixth step 36 and the process is continued with a second algorithm, since the accuracy of the measurement data is insufficient for such a long-range prediction.

[0043] If the first time point ti actually falls within the first time corridor ZK1, i.e., relatively close, then either a significant leak must be present, since the actual filling pressure is normally in the range of 5 bar to 10 bar and the minimum filling pressure PM is 3 bar. Alternatively, it is also possible that the pressure has been approaching the minimum filling pressure PM for some time. In any case, reaching the minimum filling pressure PM is foreseeable. In a fifth step 35, a warning signal is generated. This can be done in a known manner as an electronic signal to a control station or as a visual or audible warning signal. A shutdown can also be initiated as a direct reaction. For this purpose, it can be advantageous if such a prediction of reaching the minimum filling pressure is made several times. In other words, the event – ​​reaching the minimum filling pressure within the first time corridor – can be stored.

[0044] With the next measurement from pressure sensor 10, the algorithm is repeated, i.e., a new regression line is calculated using the new measurement and the previous measurements. If the first time point again falls within the first time corridor, this result is again registered and stored as an event.

[0045] It should be noted that the older measurement data used shifts with the new measurement; that is, the oldest measurement is discarded. The first time corridor also shifts by 15 seconds, as it always starts with the current time.

[0046] If such an event – ​​reaching the minimum filling pressure within the first time corridor – is registered several times with not too great an interval, for example four times in direct succession, then a warning signal or even a shutdown can be triggered.

[0047] To obtain a numerically coded overall result BG, the event can also be assigned a value of 4, for example. If, however, the first time point ti lies outside the first time corridor ZK1, then it is assigned a value of 0.

[0048] If, however, in the sixth step 36 the first time point ti is discarded because it is not in the first time corridor ZK1, but further in the future, then a second algorithm follows.

[0049] The second algorithm is similar to the first and is shown schematically in Figure 4. In the second algorithm, steps 41 through 46 are performed, which correspond to those of the first algorithm. The differences are described below.

[0050] The second algorithm uses a second time period Z2, which in this example is 3000 s. This time period Z2 is covered by 201 measurements. The second algorithm is not performed with every new measurement, even though this would be possible like the first algorithm, but only every 300 s, i.e., every 5 minutes. In the second step 42, a regression line G2 is then determined, in this case from the 201 measurements. Alternatively, instead of using the individual measurements, a reduced number of averaged measurements could be used, for example, if the evaluation is performed in the control system of the gas-insulated switchgear 1 and its processing power is limited.

[0051] In the third step (43), a second time point t2 is determined at which the minimum filling pressure PM is reached. In the fourth step (44), it is tested whether the second time point t2 is contained within the second time corridor ZK2, which is now 10 hours, i.e., slightly more than ten times the second period Z2. In other words, whether the second time point t2 lies more than 10 hours in the future. If this is the case, then again the accuracy of the prediction is insufficient to consider the second time point t2 realistic and reliable. Therefore, in the sixth step (46), a third algorithm is used.

[0052] If the second time point t2 actually lies within the second time corridor ZK2, then reaching the minimum filling pressure p is possible. MThis is foreseeable, though not in the next few minutes. For example, a warning signal will be generated in step 45, for instance. In this case, a notification can be triggered if such an event is registered several times in succession to avoid false alarms.

[0053] For the numerically coded overall result BG, a value of 3 can be assigned to the event. If, however, the second time point t2 lies outside the second time corridor ZK2, then a value of 0 is assigned to it. The largest value obtained (i.e., 0, 3, or 4) constitutes the overall result BG.

[0054] If the second time point t2 is also rejected because it is not in the second time corridor ZK2, but further in the future, then a third algorithm follows.

[0055] The third algorithm is similar to the first and second algorithms and is shown schematically in Figure 5. In the third algorithm, steps 51 through 56 are performed, which correspond to those of the first and second algorithms. The differences are described below.

[0056] In the third algorithm, a third period Z3 is used in step 52, which in this example is seven days (7 d). This choice of the third period Z3 compensates for weekly fluctuations in the load current, as measurement data is always used that includes all days of the week. At the same time, the third period Z3 is short enough to eliminate seasonal variations.

[0057] This period Z3 is covered by approximately 40,000 measurements. The third algorithm is not performed with every new measurement, but only once a day.

[0058] In the second step (52), a regression line G3 is determined, in this case from the approximately 40,000 measured values. Alternatively, instead of using the individual measured values, a reduced number of averaged measured values ​​could also be used.

[0059] In the third step (53), a third time point t3 is determined, at which the minimum filling pressure PM is reached. In the fourth step (54), it is tested whether the third time point t3 is contained within the third time corridor ZK3, which is now two months, i.e., slightly less than ten times the third period Z3; in other words, whether the third time point ta is more than two months in the future. If this is the case, a value of 1 is set as the numerically coded overall result BG in the sixth step (56).

[0060] If the third time point t3 actually lies within the third time corridor ZK3, then reaching the minimum filling pressure p is possible. MThe target date is foreseeable, but there is still a considerable amount of time before it is reached, during which maintenance measures can be implemented. For example, a warning signal will be generated in step 55. In turn, a notification can be triggered if such an event is registered several times in succession to avoid false alarms.

[0061] For the numerically coded overall result BG, a value of 2 can be assigned to the event. If, however, the third time point t3 lies outside the third time corridor ZK3, then a value of 1 is assigned to it. The largest value obtained still constitutes the overall result BG.

[0062] Even though the fifth steps already describe how measures will be taken, these measures can also be derived from the numerically encoded overall result BG, i.e., only after the completion of the entire algorithm consisting of the first, second, and third algorithms. To achieve a certain stability of the result, the number 4 must have appeared more than three times in an exemplary embodiment. After the number 3 appears in the second algorithm, an average is calculated for the last ten numerical values ​​from the results of the second and third algorithms, and the result is rounded up before being presented as the final output. The same procedure is performed when the number 2 appears as the overall result.

[0063] To ensure the long-term reliability of the procedure, each algorithm can be checked to determine whether, within the respective time period under consideration, there are systematic dependencies between the measurement data and the environmental or operating conditions of the gas-insulated switchgear 1, which might, for example, indicate a faulty temperature compensation. One indication of this is that the fitting process exhibits increased variance resulting from portions of the measurement data whose temporal profile correlates with the environmental and operating conditions. For this purpose, the residual of the fit is examined.

[0064] Reference symbol list

[0065] 1 gas-insulated switchgear

[0066] 4 display elements

[0067] 6 Control element

[0068] 7 Current feedthrough

[0069] 8 gas cylinders

[0070] 9 Rotary feedthrough

[0071] 10 pressure sensor

[0072] 31...36 process steps in the first algorithm

[0073] 41...46 Procedural steps in the second algorithm

[0074] 51...56 Procedural steps in the third algorithm

[0075] G1, G2, G3 regression line

[0076] Z1, Z2, Z3 time periods

[0077] ZK1...3 Time corridors tl, t2, tß Time points

[0078] PM minimum filling pressure

[0079] BG rating scale

Claims

Patent claims 1. Operating method for a gas-insulated switchgear (1) with a gas-tight enclosure section (8) filled with an insulating gas, wherein - Measurement data from a sensor (10) for detecting pressure or gas density in the housing section (8) are recorded, - using measurement data from the sensor (10) extending over a first period (Z1), a first value (h) is determined which is a measure of the remaining service life, - using measurement data from the sensor (10) extending over a second period (Z2), a second value (t2) is determined which is a measure of the remaining lifetime, wherein the second period (Z2) is at least ten times as long as the first period (Z1), - a common assessment value (BG) for the remaining service life is determined from the first and second values ​​(ti , t2), whereby a comparison of the first value (h) with a first time corridor (ZK1) is made, wherein the first time corridor (ZK1) is at most ten times as long as the first period (Z1).

2. Operating method according to claim 1, wherein the first value (h) is only taken into account when determining the evaluation parameter (BG) if the comparison of the first value (h) with the first time corridor (ZK1) shows that the remaining service life is not longer than the first time corridor (ZK1).

3. Operating method according to claim 1 or 2, wherein the first value (h) is a first time point (h) at which the remaining pressure in the housing section reaches a minimum filling pressure (p). M ) is reached and the second value (t2) is a second time point (t2) at which the remaining pressure in the housing section reaches the minimum filling pressure (p M ) reached.

4. Operating method according to one of the preceding claims, wherein a fitting process is carried out on measurement data of the respective period (Z1 , Z2) to determine the first and second value (ti, t2), in particular a fit with a linear function.

5. Operating method according to one of the preceding claims, wherein the measurement data of the sensor (10) are recorded at intervals of at least 1 s and at most 30 s.

6. Operating method according to one of the preceding claims, wherein the first and second periods (Z1 , Z2) overlap.

7. Operating method according to one of the preceding claims, wherein the first period (Z1) has a length of between 10 s and 600 s.

8. Operating method according to one of the preceding claims, wherein the second period (Z2) has a length of between 10 min and 10 h, in particular between 30 min and 2 h.

9. Operating method according to one of the preceding claims, wherein, using measurement data from the sensor (10) extending over a third period (Z3), a third value (t3) is determined which is a measure of the remaining service life, wherein the third period (Z3) is at least ten times as long as the second period (Z2).

10. Operating method according to one of the preceding claims, wherein the third period (Z3) has a length of between one day and one month, in particular between 3 days and 10 days.

11. Operating method according to one of the preceding claims, wherein values ​​for the residue of the fit process are stored and an evaluation signal for the reliability of the joint evaluation variable (JV) is determined from the course of these values.

12. Gas-insulated switchgear (1) comprising - a gas-tight housing section (8) filled with an insulating gas, - a pressure sensor (10) for detecting the pressure in the housing section (8), - an evaluation device for recording measurement data from the sensor (10), wherein the evaluation device is designed, - using measurement data from the sensor (10) extending over a first period (Z1), to determine a first value (ti) which is a measure of the remaining service life, - using measurement data from the sensor (10) extending over a second period (Z2), to determine a second value (t2) which is a measure of the remaining lifetime, wherein the second period (Z2) is at least ten times as long as the first period (Z1), - to determine a common assessment parameter (BG) for the remaining service life from the first and second values ​​(ti , t2), whereby a comparison of the first value (ti) with a first time corridor (ZK1) is carried out, wherein the first time corridor (ZK1) is at most ten times as long as the first period (Z1).

13. Gas-insulated switchgear (1) comprising a bellows as a pressure sensor (10).

Images