Electrical equipment condition analysis

Abstract

A method for analyzing a state of an electric device 2 of a power supply system 1, comprising: The electrical device 2 comprises a housing 3 containing an insulating liquid 4, the method comprising the steps of: - collecting measurements indicative of gas dissolved in said insulating liquid 4; - determining equipment parameters; - adapting said measurements to a certain analytical criterion by means of said instrumental parameters; - analyzing the state of said electrical appliance 2 on the basis of said adapted measurements by means of at least one machine learning method; - determining at least one state of said electrical appliance 2; - outputting at least one state; Includes.

Description

[Technical field]

[0001] The present invention relates to a method for analysing the state of electrical equipment of a power supply system, to a device for analysing the state of electrical equipment of a power supply system, and to a power supply system comprising such a device. [Background technology]

[0002] In electric power technology, a power supply system refers to a network for transmitting and distributing electric power. The power supply system includes power transmission lines, such as overhead lines and underground cables, associated facilities, such as generating stations and substations, and the electrical equipment of the associated facilities.

[0003] Monitoring and analysis of electrical equipment in a power supply system, such as power transformers or tap changers, is an important means of ensuring the smooth running of the power grid.

[0004] A reliable method for inferring the condition of a power transformer or tap changer is the analysis of the insulating liquid. Due to the ageing of the insulating liquid, and more particularly due to abnormalities caused by heating or leakage during the operation, decomposition gases are generated from the insulating liquid, which are dissolved in the insulating liquid. From the amount and type of the dissolved gas, the rate of increase of the gas and the ratio of several types of gas to each other, the abnormal operation of the electrical equipment can be inferred and also the type of the abnormality. In this field, this analysis is known as the so-called gas-in-oil analysis (DGA).

[0005] In addition to the aging of the insulating liquid and the abnormal operation of the electrical equipment, parameters specific to the electrical equipment also play an important role in the gas generation and composition and interpretation of the DGA data.

[0006] Therefore, control transformers used, for example, in substations, such as transformers in HVDC transmission systems, have to meet different requirements. The same applies to on-load tap changers, where the turns ratio of each transformer is designed to match the requirements of the corresponding range of use.

[0007] Therefore, similar results from gas-in-oil analyses based on different equipment parameters may infer a different condition of the electrical equipment, which in turn may lead to misinterpretation of the gas analyses and adversely affect the condition analysis of the electrical equipment. Summary of the Invention [Problem to be solved by the invention]

[0008] Therefore, the object of the present invention is to provide an improved concept for analyzing the state of electrical equipment, which takes into account parameters specific to the electrical equipment during the state analysis and thus allows a more accurate analysis of the state of electrical equipment of a power supply system. [Means for solving the problem]

[0009] This problem is solved by the respective objects and methods defined in the independent claims. Further embodiments are defined in the dependent claims.

[0010] The improved concept is based on the idea of ​​making the data received for analyzing the condition into one constant analytical standard, facilitating comparability between individual electrical equipment and a representative group of electrical equipment and achieving a more accurate interpretation of the DGA data based on said comparability.

[0011] According to an aspect of the improved concept, there is provided a method for analyzing a state of an electrical device of a power supply system, the electrical device comprising a housing having an insulating liquid, the method comprising the steps of: - collecting measurements indicative of gas dissolved in said insulating liquid (4); - determining equipment parameters; - conforming said measurements to a certain analytical criterion by said instrumental parameters; - analyzing the state of said electrical equipment (2) by means of at least one machine learning method on the basis of said adapted measurements; - determining at least one state of said electrical device (2); - outputting at least one state.

[0012] Preferably, a DGA data set from a DGA analysis performed on the electrical equipment is used as a measurement indicative of the gas dissolved in the insulating liquid.

[0013] According to a preferred embodiment, a sample of the insulating fluid is taken and analyzed in a laboratory to determine the measurements, where an actual data set may be collected as the measurements and one or more historical data sets may be collected as the measurements.

[0014] According to another embodiment, the measurements are taken continuously, in particular at regular time intervals, for example, so that at least one sensor is provided for continuously measuring the dissolved gas in the insulating liquid.

[0015] The insulating liquid is preferably configured as an insulating oil.

[0016] According to a preferred embodiment, the equipment parameters include configuration-specific and design-specific parameters of the electrical equipment, the functional principle of the electrical equipment, the volume and properties of the insulating liquid, the properties of the housing, the processing capacity, the operating mode of the electrical equipment or the physical ambient conditions in which the electrical equipment is present.

[0017] According to another embodiment, the device parameters are specified only once.

[0018] According to another embodiment, the equipment parameters include in particular data sets determined during operation of the electrical equipment and / or during maintenance of the electrical equipment.

[0019] According to a preferred embodiment, the collected measurements and / or the device parameters are checked for plausibility. This can be done by checking predefined plausibility rules, for example based on predefined limits and / or value ranges. Known methods for such plausibility checks can be used as well, such as the local outlier factor method.

[0020] According to another embodiment, a probability of the existence of at least one state of the electrical device is determined and the state is output together with its probability of existence.

[0021] According to one embodiment, a first state and a second state with a probability of each state being present are output, for example, the first state includes a state of the electrical device without the identified anomaly, and the second state includes a state of the electrical device without the identified anomaly.

[0022] According to another embodiment, three or more states are output with a probability of each state being present, for example at least two of the states include at least two different identified abnormal states and at least one of the states includes a state without the identified abnormality.

[0023] According to another embodiment, the method comprises the steps of: a further step of identifying an indicator of uncertainty regarding the probability that the state exists; - Another step that outputs this indicator; Includes.

[0024] According to one embodiment, the indices are presented as numerical values ​​and / or as percentage values.

[0025] According to another embodiment, the uncertainty indicator is determined depending on the availability of the collected measurements and / or the determined instrument parameters.

[0026] When the index is shown as a numerical value, according to a preferred embodiment, the lower the availability of the collected measurements and / or the identified equipment parameters, the higher the numerical value becomes.

[0027] According to another embodiment, missing measurements and / or missing instrument parameters are identified by at least one statistical analysis method.

[0028] According to a preferred embodiment, the missing measurements are imputed when the condition analysis is performed.

[0029] According to another embodiment, the measurements and the missing measurements are matched to one fixed analysis criterion by the instrument parameters and the missing instrument parameters.

[0030] According to a further embodiment, the electrical device is configured as a tap changer, and the device parameters include data that are in particular specific to tap changers.

[0031] According to one embodiment, the tap changer specific data comprises, for example, the number or frequency of changes, the duration of use, the volume of insulating liquid, the typical load factor, the number of phases, data on the basic structure and design, the circuit topology or the functioning principle of the tap changer, whereby, for example, it can be identified whether the tap changer is based on vacuum or oil changing technology. Furthermore, the tap changer specific data also comprises data collected during operation of the tap changer or during maintenance work.

[0032] According to further embodiments, in addition to the tap changer-specific data, characteristic variables specific to the insulating liquid of the tap changer and / or the transformer and / or data relating to the operation of the tap changer and / or the transformer and / or data relating to the transformer are determined, where data relating to the operation is understood in particular as the operating mode of the tap changer, for example the use in a hermetically sealed transformer, in a feed-in control, in a high-voltage direct current transmission system or in a steelworks, or also as the ambient conditions of the transformer and / or the tap changer.

[0033] Here, the characteristic variables specific to the insulating liquid of the tap changer can be understood as, for example, the distinction between mineral oils, synthetic oils or natural oils, or they can also be understood as properties of the insulating liquid, such as, for example, the duration of use, the proportion of inhibitors, passivators and / or other additives.

[0034] According to another embodiment, the measured values ​​are further adapted to a certain analysis criterion by characteristic variables specific to the insulating liquid of the tap changer and / or the transformer and / or data relating to the operation of the tap changer and / or the transformer and / or data specific to the transformer.

[0035] According to another embodiment, the statistical analysis method for identifying missing measurements and / or missing instrument parameters is based on imputation methods, in particular single imputation and / or multiple imputation and / or nearest neighbor imputation methods.

[0036] According to another embodiment, the machine learning method for analysing the condition is based on regression methods and / or neural networks and / or support vector machines and / or linear discriminant methods and / or Gaussian process regression.

[0037] According to another embodiment, the method comprises the steps of: - other steps of outputting status messages and / or warning messages and / or recommended operations and / or performing actions related to the operation of the electrical equipment depending on the indicators related to the identified status and / or uncertainty.

[0038] According to a second aspect of the improved concept, there is further provided an apparatus for analyzing the state of an electrical device of a power supply system, the electrical device comprising a housing with an insulating liquid.

[0039] The features of the device correspond to the steps of the method according to the first aspect of the improved concept. Therefore, for the device according to the second aspect of the improved concept, reference is likewise made to the beneficial descriptions, preferred features, technical effects and / or advantages already explained for the corresponding embodiment of the method according to the first aspect, without repeating the explanation.

[0040] The apparatus includes an interface for collecting measurements and / or instrument parameters indicative of gas dissolved in the insulating liquid.

[0041] The interface may be configured to automatically and continuously collect measurements and / or device parameters, for example by means of sensors disposed on the electrical equipment.

[0042] The interface may also be configured as an input device to manually collect measurements and / or equipment parameters.

[0043] The interface may also be configured to collect historical measurements and / or device parameters from the upper network.

[0044] The apparatus further includes an analytical device configured to carry out a method configured according to the first aspect of the improved concept.

[0045] According to one embodiment, the apparatus further comprises an output device configured to output a status of the electrical device.

[0046] According to one embodiment, the output device is further configured to output an indication of the uncertainty on the probability that the state exists.

[0047] The output is preferably visually implemented as a graphical display.

[0048] According to a third aspect of the improved concept, there is further proposed a power supply system including at least one electric appliance and at least one device for analyzing a state of the at least one electric appliance configured according to the second aspect of the improved concept.

[0049] Other embodiments and implementations of the system are directly described in the different embodiments of the method and apparatus.

[0050] The present invention will be described in detail below based on representative embodiments with reference to the drawings. Components that are identical or functionally identical or have the same effect may be labeled with the same reference numerals. Components that are identical or have the same function will be described specifically only with respect to the drawing in which they first appear. The description will not be repeated in the subsequent drawings. [Brief description of the drawings]

[0051] [Figure 1] 1 shows a schematic representation of a preferred embodiment of a power supply system including an electric appliance and a device for analysing the state of the electric appliance; [Diagram 2] A preferred embodiment of the method according to the improved concept is shown. [Figure 3a] 1 illustrates an example of an output of the status of an electrical device. [Figure 3b] 1 illustrates an example of an output of the status of an electrical device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0052] FIG. 1 shows a schematic representation of a preferred embodiment of a power supply system including an electrical function 2 and a device 5 according to the improved concept.

[0053] The system 1 comprises a tap changer 2, which in this case constitutes electrical equipment and is provided for switching between winding taps of a transformer (not shown). In addition to the tap changer 2 and the transformer, the system 1 comprises a number of transmission lines for transmitting and distributing the power, associated installations such as generating stations, substations and electrical equipment of the associated installations. However, for the sake of better clarity, only the components important for the improved idea of ​​the power supply system 1 are shown in FIG. 1.

[0054] The tap changer 2 has a housing 3, which is filled with an insulating liquid 4. Furthermore, the system 1 comprises a device 5, which is provided for analyzing the state of the tap changer 2. The device 5 has an interface 6, which transmits measured values ​​related to the tap changer 2 to the device 5. To collect the measured values, the tap changer 2 has at least one suitable sensor 7, which transmits the collected measured values ​​to the device 5 via the interface 6. The measured values ​​include in particular dissolved gases in the insulating liquid 4. Thus, at least one sensor 7 is configured as a DGA sensor. Furthermore, further measured values ​​related to the tap changer 2 can be collected as real-time data of the tap changer 2 and taken into account during the state analysis, such as, for example, the system voltage, the load current, the operating time, the rotational torque, the course of the rotational torque, the tap position, the switching time, the number of positions or winding taps reached, the respective duration of stay in one position, the temperature of the insulating liquid 4, the ambient temperature or other data. Depending on the type of measured value, the collection can be carried out once or at regular intervals and can be automatically transmitted to the device 5 via the corresponding sensor 7 by the interface 6 or by manual input. Optionally, tap changer-specific parameters, such as, for example, the number or frequency of changes, the duration of use, the volume of insulating liquid, the load factor, the number of phases, data on the basic construction and design, the circuit topology or the functioning principle of the tap changer, are also taken into account during the condition analysis of the tap changer 2, or additional data collected during operation or maintenance of the tap changer. Furthermore, specific characteristic curves of the insulating liquid 4 of the tap changer 2, data on the operation of the tap changer 2, i.e. data that depends on the field of use of the tap changer, can be taken into account during the condition analysis, or transformer-specific parameters, such as, for example, data on the construction or design of the transformer, can be taken into account during the condition analysis. The parameters are recorded in a data memory 8 of the device 5. Furthermore, additional data may be recorded in the data memory 8 which is necessary for plausibility checks of the incoming measured values ​​and of the tap-changer specific parameters carried out during the condition analysis.To this end, for example, suitable validity rules, limit values, value ranges or program code may be recorded in the data memory 8 for automatically carrying out the validity checks.

[0055] Furthermore, the device 5 comprises an analysis device 9 configured to perform a condition analysis on the tap changer 2 using one or more machine learning methods and to output the condition of the tap changer 2 by means of an output device 10. Furthermore, the analysis device 9 is configured to perform an automatic plausibility check of the collected measured values ​​during the condition analysis with tap changer-specific parameters and to bring the collected measured values ​​into one constant analysis criterion with tap changer-specific parameters and / or transformer-specific parameters and / or insulating liquid-specific characteristic variables and / or operational data. For this purpose, the analysis device 9 is configured to identify missing measured values ​​and / or missing tap changer-specific or transformer-specific parameters using suitable statistical analysis methods. The missing measured values ​​are then completed when the condition analysis is performed. The missing tap changer- or transformer-specific parameters are used to bring the complete measured values ​​into one constant analysis criterion, for example a standard application of the tap changer in the network.

[0056] The functions of the analysis device 9 described may be performed by hardware, firmware, software, other machine-readable instruction code, or any combination thereof.

[0057] For this purpose, the device 5 for analyzing the condition can be connected to a higher-level network or the Internet 11 via an interface 6. The condition can then be output locally at the device 5 via the output device 10 as described above, or via the higher-level network 11 as a communication medium to various other terminal devices connected to this network 11. Another advantage of the connection to the higher-level network 11 is the access to further data from the electrical appliance 12 which is also connected to the higher-level network or to historical DGA data of the electrical appliance 2 which were collected and then transmitted to the network 11 via the interface 6. The historical measured values ​​and / or data of the electrical appliance 12 can be taken into account in the condition analysis or can be used, for example, to adapt the measured values ​​to a certain analysis criterion or to identify and complete missing measured values ​​and parameters. It is also possible by this connection to completely transfer the execution of the condition analysis, i.e. the functions of the evaluation device 9, to a server, cloud or another computing device connected to the higher-level network or the Internet 11.

[0058] A preferred embodiment of a method according to the improved concept is shown in flow chart form in FIG.

[0059] According to this preferred embodiment, the method is used to analyze the state of an electrical device of a power supply system, for example as shown in Figure 1. The system 1 thus comprises at least one electrical device configured as a tap changer 2 and a housing 3 with an insulating liquid 4.

[0060] According to step a of the method, firstly, measured values ​​indicative of dissolved gas in the insulating liquid are collected. Typically, the measured values ​​are data analyzed during a gas-in-oil analysis. In step b, the equipment parameters are identified. In this case, according to this embodiment, the equipment parameters are configured as tap changer-specific parameters, since the electrical equipment is a tap changer. In step c, further data relating to the tap changer are identified, i.e. further data relating to characteristic variables specific to the insulating liquid, such as, for example, the chemical composition and properties of the insulating liquid, and data relating to the area of ​​use of the tap changer, and possibly also data relating to the transformer in which the tap changer is installed. In step d, the collected data are checked for plausibility, for example by suitable plausibility rules, recorded limit values ​​or value ranges, or also by a program code that performs an automated plausibility check. In step e, missing measured values ​​are identified by at least one suitable imputation method.

[0061] The order in which steps b, c, d and e are performed is not important to the performance of the method, for example steps d and e may be performed after step a, step b or step c, or may be performed in parallel to steps b and c as shown in FIG.

[0062] In a next step f, data specific to the missing tap changer are identified by at least one suitable imputation method. Then, in step g, the measurements collected in step a and possibly supplemented in step e are adapted to a certain analysis criterion by the tap changer-specific parameters identified in step b and possibly supplemented in step f, and by the further determined data in step c, in particular the insulating liquid data and the usage data of the tap changer. For example, the measurements can be adapted to a given standard application, in order to be able to compare a particular tap changer with other tap changers. This allows an improved interpretation of the data and therefore also an improved analysis of the tap changer's condition.

[0063] Thereafter, in a next step h, a condition analysis is performed by at least one machine learning method, for example a support vector machine, based on the measurements collected, adapted in step e and possibly supplemented in step f, and on the remaining data collected in step c. Steps f, g and h are repeated in this order many times, for example several thousand times.

[0064] Then, depending on the state analysis in step i, possible states of the tap changer are identified with their respective probabilities of existence and an indication of the uncertainty for the probability of the existence of each state. Finally, in step j of the method, these states are output together with the indications of the probabilities and the uncertainty. Optionally, together with the output and depending on the state analysis, a corresponding warning message and a recommended action may be output to the tap changer operator or the tap changer may be automatically operated, e.g. stopped.

[0065] Possible representations of the output performed in step j are shown in figures 3a and 3b. One graph each is shown as two bar graphs. The states of the tap changer are shown along the x-axis. Each bar represents a possible state. In this example, a total of five states are identified, namely "abnormal A", "abnormal B", "abnormal C", "normal operation A" and "normal operation B". "Abnormal A" indicates, for example, the occurrence of one or more arcs, "abnormal B" indicates the occurrence of one or more partial discharges in the on-load tap changer and "abnormal C" indicates, for example, an unacceptable heating state. In the case of "normal state A", the on-load tap changer is considered to be in a state without an abnormality. "Normal state B" indicates, in this example, a state representing a stray gas phenomenon (unwanted gas generation due to the interaction of insulating liquid with mostly metallic components of the on-load tap changer at normal operating temperatures). However, since the stray gas phenomenon does not cause significant damage to the electrical equipment to a certain extent, this state is not considered to be abnormal.

[0066] The probability of each state being present is shown along the y-axis, so the height of the bar indicates the probability of a state being present.

[0067] FIG. 3a shows the result when fewer known measurement data and / or equipment parameters are taken into account during the condition analysis than in FIG. 3b. It can therefore be seen from FIG. 3a that the condition "abnormal A" exists with a probability of at least 20 percent. The lighter colored bar areas form an uncertainty indicator for the probability of the existence of the respective condition. Taking into account the uncertainty indicator, it can be seen that the probability of the existence of the condition "abnormal A" lies between 20 and 90 percent. The value of the uncertainty indicator is therefore approximately 70 percent. In contrast, it can be seen from FIG. 3b that the condition "abnormal A" exists with a probability of at least 25 percent. Here, the proportion of the uncertainty indicator is significantly smaller, namely 10 percent. The probability corresponding to the condition "abnormal A" therefore lies between 23 parameters and 22 percent.

[0068] Thus, the remaining bars in Figures 3a and 3b, namely states "Abnormal B", "Abnormal C", "Normal Operation A" and "Normal Operation B", can be interpreted in a similar manner.

[0069] The comparison of the two graphs makes it clear that the uncertainty indicator depends on the availability of measurement data and equipment parameters taken into account: in particular, this means that the more data is available and taken into account during the state analysis, the smaller the uncertainty in outputting the probability of the existence of a state. [Explanation of symbols]

[0070] 1. Power Supply System 2 Electrical equipment, tap changers 3. Housing 4. Insulating fluid 5 Equipment 6. Interface 7 Sensors 8 Data Memory 9 Analyzer 10 Output Devices 11 Network 12 Other Electrical Equipment

Claims

1. A method for analyzing a state of an electrical device (2) of a power supply system (1), comprising: The electrical device (2) comprises a housing (3) containing an insulating liquid (4), and is - collecting measurements indicative of gas dissolved in said insulating liquid (4); - identifying instrument parameters; - adapting said measurements to a certain analytical criterion by means of said instrumental parameters; - analyzing the state of said electrical appliance (2) on the basis of said adapted measurements by means of at least one machine learning method; - determining at least one state of said electrical appliance (2); - outputting at least one state; The method includes:

2. The method of claim 1, wherein the collected measurements and / or the instrument parameters are checked for plausibility.

3. - the probability of the existence of at least one state of said electrical device (2) is determined, The method according to claim 1 or 2, wherein the states are output together with the probability of being.

4. below, a further step of identifying a measure of uncertainty regarding the probability of the existence of said state; another step of outputting this indicator; The method according to any one of claims 1 to 3, comprising:

5. The method according to claim 4, wherein the uncertainty measure is determined depending on the availability of the collected measurements and / or instrument parameters.

6. The method according to any one of claims 1 to 5, wherein the missing measurements and / or missing instrument parameters are identified by at least one statistical analysis method.

7. The method according to any one of claims 1 to 6, wherein the electrical equipment is configured as a tap changer (2) and the equipment parameters include in particular tap changer-specific data.

8. The method according to claim 7, wherein in addition to the tap changer-specific data, insulating liquid-specific characteristic variables and / or operational data and / or transformer-specific data are identified.

9. The method according to any one of claims 1 to 8, wherein the step of adapting the measured values ​​to a certain analytical criterion is further carried out by characteristic variables specific to the insulating liquid and / or by data relating to the operation and / or by data specific to the transformer.

10. The method according to any one of claims 1 to 9, wherein the statistical analysis method for identifying missing measurements and / or missing instrument parameters is based on imputation methods.

11. The method according to any one of claims 1 to 10, wherein the machine learning methods for analysing the condition are based on regression methods and / or neural networks and / or support vector machines and / or linear discriminants and / or Gaussian process regression.

12. below, The method according to any one of claims 1 to 11, comprising a further step of outputting a recommended operation and / or performing an action related to the operation of the electrical appliance (2) depending on the indicators related to the identified conditions and / or uncertainties.

13. an interface (6) for collecting measurements and / or instrumental parameters representative of the gas dissolved in the insulating liquid (4); - an analysis device (9) configured to carry out a method configured according to any one of claims 1 to 12, wherein the electrical equipment (2) of a power supply system (1) has a housing containing the insulating liquid (4).

14. An apparatus (5) according to claim 13, comprising an output device (10) arranged to output a state of said electrical appliance (2).

15. The device (5) according to claim 13 or 14, wherein the output device (10) is further adapted to output an indication of the uncertainty on the probability that the state exists.

16. at least one electrical device (2), - at least one device (5) for analyzing the state of at least one electrical device (2) constructed according to any one of claims 13 to 15, A power supply system (1).

Images