Apparatus for deriving a failure probability value for a transformer component and system having such apparatus

Abstract

The invention relates to a device (2) for determining a failure probability value for a transformer component (16) of an electrical transformer (12). The device (2) has a signal input interface (4), a processor unit (6) and a storage unit (8). The processor unit (6) is couplable to a measurement system (10) via the signal input interface (4), which detects a plurality of properties of the power transformer (12) by means of sensors (14). The processor unit (6) can determine the failure probability value for the transformer component (16) on the basis of the actual state of the power transformer by means of a Bayesian network, which at least partially depicts the transformer (12) and the measurement system (10). The invention also comprises a system (28), which comprises the device (2).

Description

[0001] Transformers are known in principle from the prior art. Here, a transformer capable of converting an input voltage into an output voltage is discussed. Typically, the input voltage is an input AC voltage, and the output voltage is an output AC voltage. In a very simple design, the transformer has a primary coil and a secondary coil coupled to each other via a common core. The coils and / or the core may be at least partially arranged in a tank containing coolant to cool the coils. The primary coil, secondary coil, core, coolant, and tank may each be a transformer component of the transformer. The transformer may also have additional transformer components.

[0002] In practice, transformers are frequently monitored using measurement systems. These systems typically employ multiple sensors configured to detect the transformer's physical and / or chemical properties. Measurement systems are often used when the transformer is configured as a so-called power transformer. This is true, for example, when the electrical power transferred from the primary coil to the secondary coil exceeds 10 kilowatts. During active operation, the two coils of a power transformer generate significant heat, which is at least partially absorbed by the transformer's coolant. The coolant can be circulated in the power transformer's coolant circuit, where a so-called cooler is integrated to transfer the heat transferred to the coolant to the surrounding environment and / or other media. The measurement system can therefore, for example, have sensors configured to detect the temperature of the power transformer's coolant. A fan can be arranged on the cooler to achieve the desired cooling capacity. The current flowing through the fan motor can be detected, for example, using another sensor in the measurement system. Oil, also known as insulating oil, is frequently used as the coolant. The insulating properties of the oil prevent electrical short circuits within or between the coils. The measurement system can have an additional sensor configured to detect the water content in the oil, either directly or indirectly. Thus, the sensor can, for example, be in direct contact with the oil to directly detect the water content. However, it is also possible that the sensor is arranged in the tank of the power transformer, where it is configured to detect the concentration of gases with hydrocarbon chains. A concentration of hydrocarbon-chain gases exceeding a predetermined threshold detected by the sensor can indicate a specific minimum water content in the oil. Another sensor in the measurement system can, for example, be configured to detect the mechanical winding tension between the windings of one coil in the coil. This sensor can, for example, be configured to detect mechanical pressure or force, wherein the sensor is arranged on the primary or secondary coil of the power transformer to detect the mechanical tension of the respective coil. Another sensor in the measurement system can be arranged and / or configured to detect the current flowing through the primary or secondary coil. Furthermore, the measurement system can have a voltage sensor configured to detect the voltage present in the primary or secondary coil. Additional sensors in the measurement system can be configured, for example, to detect the condition of transformer leads, wherein the sensors are configured, for example, to determine the capacitance and / or loss factor of the transformer leads. Additional sensors in the measurement system can be configured, for example, to detect vibrational acoustic signals. This sensor can, for example, be configured to detect acoustic signals generated when the tap changer of the power transformer is operated. The measurement system can be configured to generate state signals representing multiple different state parameters.The measurement system can be configured to determine at least one state parameter based on the sensor signal of exactly one of the sensors and / or one state parameter based on the sensor signals of multiple sensors. Each sensor is arranged on the power transformer to detect the physical and / or chemical characteristics of the power transformer. Therefore, the physical and / or chemical characteristics can correspondingly be the physical and / or chemical characteristics of components of the power transformer. Components of the power transformer are also referred to as transformer components. The coolant of the transformer can, for example, be a component of the power transformer. However, the voltage or current on the first or second coil of the power transformer can also be a physical characteristic of the power transformer. Each state parameter determined by the measurement system represents the state condition of the characteristic of the power transformer detected by the at least one configured sensor, as a normal condition N or as an abnormal condition. If the state parameter is determined, for example, based on the sensor signal of a sensor used to detect the temperature of the coolant, the normal condition of the measurement system can be determined for the transformer component "coolant," more precisely, especially when the temperature of the coolant detected by the sensor is less than a reference value. Otherwise, the abnormal condition can be determined by the state measurement system for the state condition. Therefore, a functional relationship exists between each state parameter and at least one corresponding transformer component (whose characteristics are detected directly or indirectly by means of at least one associated sensor). Thus, it can also be said that each state parameter is coupled to at least one transformer component via a functional relationship.

[0003] The state parameters obtained by the measurement system can represent the state of the power transformer, more specifically, as either a normal state or an abnormal state. If the state parameters are related to a single transformer component, then the state condition directly indicates whether the corresponding transformer component is in a normal or abnormal state. Conversely, if the state parameters are obtained based on multiple sensor signals from multiple sensors, there is often no direct and explicit relationship between the state condition represented by the corresponding state parameter and a single transformer component. More precisely, a functional relationship may exist between the state condition and multiple transformer components. Furthermore, a transformer component may influence multiple sensor signals. A transformer component in an abnormal state may thus lead the measurement system to determine multiple state parameters such that each state parameter represents a state condition as an abnormal state.

[0004] When multiple state parameters represent abnormal conditions as state states, it is not always possible to determine, in a meaningless and explicit way, whether a transformer component and which transformer components are in an abnormal condition, or whether only one transformer component is in an abnormal condition, which is the reason why multiple state parameters simultaneously represent abnormal conditions as state states.

[0005] Therefore, the object of the present invention is to provide an apparatus and a system that allows for a particularly simple assessment of which transformer components of a power transformer are faulty and which transformer components are faulty. According to a first aspect of the invention, this object is achieved by an apparatus having the features of the invention. Thus, an apparatus is provided for determining the failure probability values ​​of transformer components. The apparatus has a signal input interface, a processor unit, and a storage unit. The signal input interface is configured to be directly or indirectly coupled to a measurement system for the power transformer, wherein the measurement system has a plurality of sensors, each configured to detect physical and / or chemical characteristics of the power transformer, wherein the power transformer has a plurality of transformer components, and each sensor is coupled to at least one of the transformer components of the power transformer via a corresponding direct or indirect connection, such that the characteristics of the power transformer detected by the corresponding sensor are affected by at least one of the transformer components. Furthermore, the signal input interface is configured to receive status signals from the measurement system, wherein the status signals represent multiple different status parameters, each status parameter being assigned to at least one of the sensors such that each status parameter is in a functional relationship with at least one of the transformer components, wherein each status parameter represents the state of the power transformer as detected by the at least one assigned sensor as either a normal state or an abnormal state. Additionally, data sets are stored in a storage unit, wherein the data sets have a first probability value for each status parameter, such that the corresponding status parameter represents an error state; a second probability value for each transformer component, such that a fault occurs in the corresponding transformer component; and a third probability value for each functional relationship from one of the transformer components to one of the status parameters, such that a fault in the corresponding transformer component causes an abnormal state for the corresponding status parameter. The processor unit is configured to, for each transformer component, determine the fault probability value of the corresponding transformer component in a fault condition based on a Bayesian network and the state conditions represented by state parameters, wherein the Bayesian network represents a directed graph from the transformer component to the state parameters based on first, second, and third probability values.

[0006] The description of the power transformer and measurement system is preferably referenced in a similar manner to that mentioned at the beginning. The measurement system and power transformer are preferably not part of the device. More precisely, it is preferably specified that the signal input interface is configured to receive status signals from the measurement system. These status signals represent multiple different status parameters, each representing a state of the power transformer's characteristic detected by at least one associated sensor of the measurement system as either a normal or abnormal state. Preferably, the status is either normal or abnormal. The receiveable status signals thus provide multiple different status parameters in the device, each representing an abnormal or normal state of a corresponding characteristic of the power transformer. However, only from these status parameters can one empirically determine which transformer components might actually be faulty. To obtain repeatable and quantifiable statements regarding possible faults in the transformer components, the device includes a processor unit and a storage unit.

[0007] The storage unit stores a data set, which includes a first probability value for each state parameter. This first probability value preferably gives the probability that the state represented by the state parameter (i.e., normal or abnormal) is an erroneous state, independent of the actual state represented by the state parameter. If the state parameter represents, for example, a normal state, then there is a certain probability, that is, a first probability value corresponding to the state represented by the state parameter that is actually not a normal state but must be an abnormal state. If the first probability value is, for example, 0.05, then there is a 5% probability that the state represented by the corresponding state parameter is erroneous.

[0008] The data set stored in the storage unit also includes a second probability value for each transformer component of the power transformer. This second probability value indicates the probability that a fault may have occurred in the corresponding transformer component. For example, statistical surveys and / or practical experience with transformers can determine that there are transformer components that fail more frequently than other transformer components of the power transformer. Therefore, the statistical probability of faults in transformer components can be deduced from statistical surveys and / or experience. Based on these statistical probabilities of transformer components, the second probability value can be predetermined. For example, a second probability value representing the probability of a fault in the primary winding of the power transformer can be predetermined for the primary winding. Correspondingly, a second probability value can be predetermined for a predetermined number of transformer components of the power transformer.

[0009] The data set stored in the storage unit also includes a corresponding third probability value for each functional relationship between a transformer component and a state parameter. The corresponding third probability value thus indicates the probability that a fault in the corresponding transformer component causes an anomaly in the state parameter within the corresponding functional relationship. If the state parameter is provided to a sensor capable of detecting the physical and / or chemical characteristics of exactly one transformer component, the third probability can represent whether a fault in that transformer component leads to an anomaly in the state parameter caused by the sensor. However, it is also possible that the transformer component has direct or indirect relationships with multiple sensors, such that a fault in the corresponding transformer component can equally affect multiple state parameters. Whether this situation occurs statistically for multiple sensors and corresponding state parameters is represented by the corresponding third probability value.

[0010] As can be at least indirectly understood from the foregoing explanation, sensors and therefore state parameters do not always allow direct inference of the state of exactly one transformer component that may be faulty. More precisely, multiple transformer components that may be faulty are typically considered simultaneously when one or more state parameters represent an abnormal situation as a state condition. Therefore, there exists a functionally oriented relationship from transformer component to sensor and thus also an oriented relationship from transformer component to state parameter; however, there is no explicit relationship from state parameter to transformer component. In order to still be able to derive conclusions about which transformer components and whether one of the transformer components is faulty, the device's processor unit is configured to, for each transformer component, calculate, based on a Bayesian network and the state condition represented by the state parameter, the corresponding fault probability value for whether the transformer component is in a fault condition. Each fault probability value preferably gives the probability (fault probability) of each transformer component being in a fault condition. A fault condition can be understood as an abnormal situation of a transformer component. If, for example, a fault probability value of 0.001 is calculated for a transformer component, then there is a 0.1% probability that each transformer component is faulty. This is relatively small and often indicates that the corresponding transformer component is not faulty. However, if the fault probability value of another transformer component is calculated to be 0.7 using a processor unit, this indicates that each transformer component has a 70% probability of being faulty. The fault probability value is calculated based on a Bayesian network and state parameters, which are provided to the device by a measurement system via state signals. Bayesian networks are known in principle from the prior art and therefore do not require further and detailed explanation here. In the Bayesian network, the transformer components form a first set of nodes and the state parameters form a second set of nodes, wherein there exists a specifically oriented functional relationship between each transformer component and at least one of the state parameters.

[0011] The processor unit can be configured to implement the fault probability value by evaluating the Bayesian network and taking into account the actual state conditions, more specifically, for example, by means of one of the following algorithms: a sampling algorithm or a variable elimination algorithm. Both algorithms are known in principle from the prior art. Other algorithms may also be used. The processor unit can be constructed to implement each of the two algorithms or other algorithms. However, it is sufficient that the processor unit is configured to implement one of the two algorithms to obtain the fault probability value.

[0012] The Bayesian network used to calculate fault probability values ​​is represented by a directed graph from transformer components to state parameters. The data set has a first probability value for each transformer component, a second probability value for each state parameter, and a third probability value for each directional relationship from transformer components to state parameters. Since the Bayesian network represents the directed graph from transformer components to state parameters, this corresponds to a mapping of the true statistical relationship between the power transformer and the measurement system. However, it is possible, based on the Bayesian network, to determine which fault probability values ​​and their respective fault probabilities exist for a given transformer component in a faulty or abnormal condition, while considering the actual state parameters.

[0013] Therefore, the advantage provided by the device is that, despite the directional functional relationship between transformer components and state parameters, it is still possible to infer a quantitative failure probability value for the transformer components, which allows it to indicate, for example, which transformer component has the highest probability of failure.

[0014] An advantageous design of the device is characterized in that the processor unit is configured to generate multiple samples based on a Bayesian network, the Bayesian network representing a directed graph from transformer components to state parameters based on first, second, and third probability values, wherein each sample represents a corresponding sample condition for each state parameter and for each transformer component, either a normal condition or an abnormal condition. The processor unit is configured to determine a reference group of samples from the multiple samples, such that in each sample of the reference group, the sample condition to which the state parameter belongs is consistent with the state condition of the state signal's state parameter. The processor unit is further configured to, for each transformer component, calculate, from the samples of the reference group the corresponding fault probability value for the transformer component in an abnormal condition based on the sample condition for that transformer component.

[0015] The aforementioned design of the device is based on the fundamental concept of first generating a large number of samples, and then determining which samples from the previously generated samples to be assigned to the reference group based on the actual state conditions of the state parameters. Therefore, the samples from the reference group each have sample conditions corresponding to the actual state conditions for the state parameters, which are represented by the state parameters of the state signals received by the device from the measurement system. In other words, the processor unit determines, with the aid of the actual state conditions, which samples of the corresponding state parameters have sample conditions corresponding to the actual state conditions. These determined samples form the sample set of the reference group. However, the sample conditions of the transformer components are not considered when determining the sample set of the reference group. More precisely, these sample conditions are used in a separate step to determine the failure probability values ​​of the transformer components.

[0016] As previously explained, the condition of transformer components cannot usually be directly inferred from their state status. However, each sample from the reference group also has a corresponding sample status for each transformer component. Due to the multiple samples in the reference group, multiple sample statuses are derived for each transformer component, and the probability of the corresponding transformer component being in an abnormal condition can be deduced from the distribution of these sample statuses. Therefore, the processor unit can calculate the individual fault probability values ​​for each transformer component based on the samples from the reference group; these fault probability values ​​represent the probability that each transformer component is in an abnormal condition.

[0017] The processor unit is configured to generate the plurality of samples, wherein each sample represents a corresponding sample condition as either a normal or abnormal condition for each state parameter and for each transformer component. The data set stored by the storage unit includes a first probability value for each state parameter. Multiple samples are generated, for example, multiple sample conditions are generated for each state parameter. However, this generation of sample conditions is not arbitrary, but is performed with regard to a first fault probability value, thereby generating sample conditions for the state parameter with probabilities corresponding to the first probability value as abnormal conditions. For example, if the first probability value is 0.2, then multiple sample conditions for the corresponding state parameter are generated such that 20% of the sample conditions represent abnormal conditions and the remaining 80% represent normal conditions. Multiple sample conditions are generated for the remaining state parameters in a similar manner. Furthermore, the corresponding situation also applies to the sample conditions for transformer components. Because for each transformer component, a second probability value is represented from the data set. Each sample represents a corresponding sample condition as either a normal or abnormal condition for each transformer component. However, the generation of sample cases for transformer components is not arbitrary, but rather takes into account a second fault probability value, thereby generating sample cases of transformer components with probabilities corresponding to the second probability value as abnormal conditions. For example, if the second probability value is 0.15, multiple samples are generated by the processor unit such that 15% of the sample cases for the corresponding transformer component represent abnormal conditions and the remaining 85% represent normal conditions. It should also be noted that multiple samples are generated randomly while taking into account the fault probability value. The processor unit is configured accordingly. The larger the sample size, the more accurately the fault probability value can be calculated for each transformer component, which indicates the probability that the corresponding transformer component is in an abnormal condition.

[0018] Another advantageous design feature of the device is that it has an output interface, wherein the device is configured to generate an output signal that represents at least the maximum failure probability value and the corresponding transformer component among the failure probability values, and the device is configured to transmit the output signal via the signal output interface.

[0019] As described in the above design scheme, a fault probability value is calculated for each transformer component. If there are multiple transformer components, a corresponding number of fault probability values ​​are also calculated. These fault probability values ​​can be different. However, the most relevant is the maximum fault probability value, as it indicates which transformer component is most likely to be in an abnormal condition. The processor unit is therefore preferably configured to determine the maximum fault probability value and its associated transformer component from the multiple fault probability values. Furthermore, the device, especially the associated processor unit, is configured to generate an output signal representing the maximum fault probability value and indicating the associated transformer component. Thus, the output signal could, for example, represent a transformer component with the digit "01" and an associated fault probability value of "0.7". However, it is also possible that the output signal represents the corresponding transformer component in plaintext. Thus, the output signal could, for example, represent the transformer component "coolant" and an associated fault probability value of "0.7". The output signal can be transmitted via the device's signal output interface. The device can be configured accordingly. The output signal can, for example, be transmitted to a display unit to visually display the fault probability value represented by the output signal in association with the transformer component.

[0020] Another advantageous design feature of this device is that it is configured to generate an output signal such that the output signal represents at least two or three maximum failure probability values ​​and the corresponding transformer component. In practice, the transformer component with the maximum failure probability value is usually of the greatest interest. However, the corresponding failure probability value also represents a probability, so the corresponding transformer component does not necessarily have to be faulty. Therefore, transformer components with large failure probability values ​​already determined in practice are also of interest. Thus, the processor unit of the device can be configured to determine two or three maximum failure probability values. Furthermore, the device, and especially the corresponding processor unit, can be configured to generate an output signal such that the output signal represents the previously determined two or three maximum failure probability values ​​and at least indicates the corresponding transformer component. The foregoing description of the maximum failure probability values ​​and the output signal similarly applies to two or three maximum failure probability values ​​and the corresponding transformer component. However, using the output signal representing two or three maximum failure probability values ​​and the corresponding transformer component, those skilled in the art obtain information that which transformer components are also likely to be faulty, especially when the transformer component determined to have the maximum failure probability value is actually not faulty. Therefore, those skilled in the art can continue their inspection, and, for example, also inspect the transformer components for which the second highest failure probability value has been determined.

[0021] Another advantageous design feature of this device is that the state parameters represented by the state signals are ternary state parameters, each representing exactly one state condition from the following groups of possible state situations: normal, abnormal, or in-installation. This excludes the possibility that the state condition can represent any situation other than normal, abnormal, or in-installation. Alternatively, it can be specified that the group of possible situations is limited to two conditions, namely, normal and abnormal. In this case, the state condition can represent either normal or abnormal, and not any other condition.

[0022] Another advantageous design of the device is characterized in that the processor unit is configured to generate a Bayesian network based on the data set. The Bayesian network is preferably a directed acyclic graph, wherein each node of the graph is assigned a probability value, and the correlation with the node is also represented by a probability value. In the present case, the Bayesian network represents a first group of nodes consisting of state parameters, each assigned a first probability value. A second group of nodes in the Bayesian network is represented by transformer components of a power transformer, wherein each of these nodes is assigned a second probability value. The correlation between the state parameters and the transformer components is represented by corresponding third probability values. The first, second, and third probability values ​​exist in the data set, which is stored in the device's storage unit. Therefore, particularly based on the foregoing description, the processor unit can generate a Bayesian network based on probability values ​​from the data set.

[0023] Another advantageous design of the device is characterized in that the Bayesian network is constructed as a bipartite graph, representing a first group of nodes, a second group of nodes, and the functional relationship from the first group of nodes to the second group of nodes, wherein the first group of nodes is determined by a first probability value, the second group of nodes is determined by a second probability value, and the functional relationship is determined by a third probability value. Therefore, according to this design variant, the Bayesian network is used as a special Bayesian network in the form of a bidirectional graph. This ensures that there is only a functional relationship from transformer components to state parameters, but no inverse correlation. In other words, the bipartite graph is constructed such that the state parameters depend only on the characteristics of the transformer components, but the characteristics of the transformer components do not depend on the state parameters. Preferably, relationships other than those described above are excluded. Thus, all relationships between transformer components and state parameters can be represented as a two-dimensional table, the entries of which are preferably determined by expertise and can be easily configured.

[0024] Another advantageous design feature of this device is that the processor unit is configured to generate samples using a probability-weighted algorithm. Here, the processor unit preferably considers the first, second, and third probability values ​​of the data set.

[0025] Another advantageous design of the device features that the processor unit is configured to generate the output signal such that the output signal represents the corresponding fault probability value and transformer component as a first graph, in which the fault probability value is graphically represented and the transformer component is indicated and / or labeled. Such a graph could, for example, represent the fault probability value as a bar graph, wherein each bar indicates and / or labels the corresponding transformer component. The corresponding transformer component can be indicated, for example, by numbers or as text, such as "coolant". The probability value can be represented graphically in the image quantitatively and / or qualitatively. The qualitative representation of the probability value can be represented, for example, by hue and / or by the length of the bars. The quantitative representation of the probability value can, for example, be a digital graphical display of the probability value. The device can, for example, use the output signal to display the corresponding image by means of the device's display unit. However, it can also be specified that the output signal is transmitted via an output interface. For this purpose, the processor unit and, in particular, the corresponding output interface can be constructed and / or configured.

[0026] According to a second aspect of the invention, the task described at the outset is solved by a system having the features of the invention. Thus, a system having an apparatus and a display unit is provided. Preferably, the apparatus is constructed according to the first aspect of the invention and / or one of the advantageous designs thereto. The display unit has a screen. The signal interface of the apparatus is coupled to the display unit via a direct or indirect first signal connection to transmit an output signal to the display unit. The display unit is also configured to display a second diagram on the screen based on the output signal, indicating the transformer components and their associated fault probability values. The aforementioned second diagram does not necessarily imply that the first diagram already exists. Furthermore, advantageous designs, preferred features, effects, and / or advantages are referenced in a manner similar to that of the system according to the second aspect of the invention, as illustrated in conjunction with the apparatus according to the first aspect and / or one of the advantageous designs thereto.

[0027] The system includes a device and a display unit with a screen. A second diagram can be displayed on the screen, graphically representing at least one fault probability value represented by an output signal and the corresponding transformer component. This representation may include a qualitative and / or quantitative representation of the at least one fault probability value and a quantitative and / or qualitative representation of the at least one corresponding transformer component. If the output signal represents, for example, a maximum fault probability value and the corresponding transformer component, the second diagram can graphically represent the maximum fault probability value and the corresponding transformer component accordingly. The display unit can be physically separated from the device, such that the first signal connection also extends between the device and the display unit. The first signal connection can be a wired signal connection or a radio signal connection. However, in principle, the display unit can also be integrated into the device, so that the system is composed of a device with an integrated display unit. The system may also have additional components.

[0028] Another advantageous design feature of the system is that the display unit is configured such that the second graph displays the transformer components represented by the output signals in list form, and qualitatively displays their respective fault probability values ​​through different colors and / or stripes. The fault probability values ​​can also be supplementarily displayed quantitatively by the display unit. Furthermore, it is preferably specified that the display unit is configured to display the fault probability values ​​sequentially in list form according to their magnitude. Therefore, the observer obtains an overview in the shortest possible time of which transformer components are likely to be in abnormal condition, and which of these possible transformer components have the highest probability of exhibiting the corresponding abnormal condition.

[0029] Another advantageous design feature of this system is that it has a measurement system for the power transformer, wherein the measurement system is coupled to a signal input interface of the device to transmit the status signal to the signal input interface, wherein the measurement system has multiple sensors, each configured to detect the physical and / or chemical characteristics of the power transformer. The power transformer has multiple transformer components. Each sensor is coupled to at least one transformer component of the power transformer via a corresponding direct or indirect connection, such that the characteristics of the power transformer detected by the corresponding sensor are affected by at least one transformer component. The measurement system, sensors, and power transformer are described in a similar manner to the initial description. Therefore, the measurement system can have multiple sensors, each configured to detect the physical and / or chemical characteristics of the power transformer. Here, it is not always possible to directly detect the desired characteristics of the transformer components. Typically, the characteristics of the power transformer are detected by means of sensors, where the characteristics depend only indirectly on the condition of the transformer components. The measurement system can be arranged adjacent to the power transformer, with sensors of the measurement system directly mounted to the power transformer to detect corresponding physical and / or chemical properties of the power transformer. Each sensor can be coupled to an evaluation unit of the measurement system, thereby configuring the measurement system to generate the status signal by means of the evaluation unit. The evaluation unit of the measurement system can also be configured to transmit the status signal to a signal input interface of the device. Preferably, the measurement system and the device are constructed separately. However, it is also possible that the measurement system and the device can be constructed at least partially or entirely integrally. For example, the device can form part of the measurement system, or vice versa.

[0030] Another advantageous design of the system features that the measurement system is configured to sensitively detect at least two of the following characteristics of the power transformer: hydrocarbon chain gas in the liquid insulating medium, water in the liquid insulating medium, the filling level of the liquid insulating medium, voltage, current, winding tension, fan speed, fan motor power, fan volume, tap changer motor torque, tap changer switching sequence, temperature (e.g., hot spot temperature or top oil temperature), vibration intensity and functional effectiveness of the air dehumidifier, loss factor of one or more transformer leads, capacity of one or more transformer leads, and vibration intensity of the tap changer. The power transformer preferably includes a tank filled with coolant. The coolant is also referred to as an insulating fluid or insulating medium because it is preferably configured to be electrically insulating. The coolant can preferably be oil, especially mineral oil. Coolants are often esters composed of rapeseed oil. Coolants that can also be called insulating fluids are therefore, for example, oil, preferably based on esters, and particularly preferably liquid insulating media, such as rapeseed oil. The coolant preferably has two main functions. The coolant should cool the primary and secondary coils and should be electrically insulating. The transformer core and the primary and secondary coils are preferably also located in the aforementioned tank. Therefore, most of the transformer's maximum heat is released in the tank. The coolant will dissipate the heat induced on the coils. The transformer can therefore have a wiring system and a cooler constructed and arranged such that the coolant is guided from the tank through the wiring system to the cooler and back to the tank, allowing heat to be dissipated through the cooler. The power transformer is preferably constructed to connect different power grids with different voltages. For example, the power transformer can electrically couple a first power grid with a lower voltage to a second power grid with a higher voltage, or vice versa, to transfer electrical energy from one power grid to another. The power grids are preferably three-phase. Therefore, in each power grid, a corresponding current and a corresponding voltage appear on each of the three phases. The corresponding current and / or voltage can be sensed. For this purpose, the measuring system can have sensors constructed and / or arranged accordingly. Theoretically, all currents and voltages in both power grids can be detected by sensors. In practice, in most cases, only a portion of the currents and voltages in both power grids are detected. The measurement system can therefore be constructed to detect a portion of the voltage and current in two power grids, and the power transformer can be coupled to the measurement system.

[0031] Another advantageous design feature of this system is that the display unit is integrated into the measurement system. The measurement system is typically arranged adjacent to the power transformer. The device can be arranged physically separate from both the measurement system and the power transformer. In particular, it can be specified that the device is formed by a computer network, especially a cloud network, including data storage. Therefore, it is possible that the device is equipped with a processor unit with exceptionally high computing power. When the display device is part of the measurement system and therefore directly near the power transformer, displaying a second image on the display device is particularly advantageous, as this information is frequently required during the maintenance and / or repair of the power transformer.

[0032] Another advantageous design feature of the system is that it includes a remote monitoring system, wherein the signal output interface of the device is coupled to the remote monitoring system via a direct or indirect second signal connection to transmit the output signal to the remote monitoring system. The remote monitoring system is geographically separate from the device, measurement system, and power transformers. The remote monitoring system may, for example, be formed by a control center that monitors multiple power transformers.

[0033] Another advantageous design feature of the system is that the remote monitoring system is based on a cloud architecture. Therefore, the remote monitoring system can be formed by a cloud network with multiple computer units.

[0034] Another advantageous design feature of the system is that the system's display device is integrated into the remote monitoring system. Therefore, the display device can be arranged separately from the equipment, measurement system, and power transformer.

[0035] Another advantageous design feature of this system is that it includes a power transformer. Therefore, the system can include a power transformer, a measurement system, and devices. Furthermore, it can be specified that the system additionally includes the remote monitoring system.

[0036] Further features, advantages, and applications of the invention will become apparent from the following description and accompanying drawings of the embodiments. Here, all described and / or illustrated features, in themselves and in any combination, also form the subject matter of the invention, independent of their composition in the various claims or their references. Furthermore, the same reference numerals in the drawings denote the same or similar objects.

[0037] Figure 1 Advantageous design schemes for the device and system are illustrated with schematic diagrams.

[0038] Figure 2 A schematic diagram illustrates a favorable design scheme for a Bayesian network.

[0039] Figure 3 This is an exemplary tabular representation of the sample.

[0040] Figure 4 An exemplary tabular representation of a sample of the reference group is shown.

[0041] Figure 5 An exemplary representation of an image that can be reproduced on a display unit is shown.

[0042] exist Figure 1 An advantageous design of device 2 is schematically illustrated. Device 2 has a processor unit 6, a signal input interface 4, and a storage unit 8. The processor unit 6 is connected to the storage unit 8, enabling the processor unit 6 to access data stored in the storage unit 8. Thus, for example, the processor unit 6 can access a data set stored in the storage unit 8. Furthermore, the processor unit 6 is connected to the signal input interface 4, so that signals received by the signal input interface 4 are provided directly or indirectly to the processor unit 6. The signal input interface 4 can, for example, perform data preprocessing to provide the data represented by the received signals to the processor unit 6. It has proven advantageous that device 2 has a signal output interface 36. Signals generated by device 2, and especially by the processor unit 6, such as the output signal Q, can be transmitted via the signal output interface 36. The processor unit 6 can be directly or indirectly coupled to the signal output interface 36 to directly or indirectly control the signal output interface 36, enabling the transmission of the output signal Q via the signal output interface 36.

[0043] In addition, Figure 1 An advantageous design of system 28 is shown, which includes device 2 and display unit 30. System 28 is schematically shown by dashed lines to show that display unit 30 is not necessarily part of device 2. Device 2 may have a dedicated display unit 34. Device 2 and display unit 30 may be coupled by a signal connection extending between device 2 and display unit 30, also referred to as a first signal connection. The first signal connection may extend from signal output interface 36 of device 2 to display unit 30. Display unit 30 has a screen.

[0044] If the advantageous design, preferred features, technical effects and / or advantages of device 2 are described below, then the corresponding advantageous design, preferred features, technical effects and / or advantages can be applied to system 28 in a similar manner. Therefore, repetition is omitted.

[0045] exist Figure 1 In addition to device 2, a measuring system 10 and a power transformer 12 are also schematically shown. The power transformer 12 has a terminal 44 on the primary side and another terminal 46 on the secondary side. The power transformer 12 is used to convert voltage from the primary side to the secondary side. Different power grids can be coupled to the primary and secondary sides, allowing electrical energy to be transferred from the primary side to the secondary side, or vice versa.

[0046] The power transformer 12 includes multiple transformer components 16. Therefore, the power transformer 12 can serve as transformer components 16, for example having terminals 44, 46, a primary coil, a secondary coil, a core, a tank for coolant, coolant, a cooling system, a cooler, a fan for cooling the cooler, a tap changer, a motor for driving the tap changer, leads, valves, solid insulation for winding insulation, and / or other technical components.

[0047] The measurement system 10 includes a plurality of sensors 14. Each of the sensors 14 is arranged on the power transformer 12 such that each sensor 14 can detect the physical and / or chemical properties of the power transformer 12. Thus, the sensors 14 may, for example, be arranged to detect the temperature of the coolant in the power transformer 12. Additional sensors 14 may, for example, be configured and / or correspondingly arranged to detect the mechanical tension between the windings of the primary and secondary coils. Tension is also referred to as winding tension. Therefore, a corresponding sensor 14 may be configured to detect winding tension. Additional sensors 14 may be arranged, for example, to detect the voltage between terminals 44 on the primary side. Additional sensors 14 of the measurement system 10 may, for example, be configured to detect the condition of the transformer leads, wherein the sensors are, for example, configured to determine the capacitance and / or loss factor of the transformer leads. Additional sensors 14 of the measurement system 10 may, for example, be configured to detect vibrational acoustic signals. This sensor 14 may, for example, be a sensor for detecting acoustic signals generated when the tap changer of the power transformer 12 is operated. Figure 1 The sensor shown is merely an exemplary sensor 14. In practice, many other sensors 14 and / or other sensors 14 may be provided for the measurement system 10. Sensor 14 may be configured to detect force, mechanical stress, current, voltage, temperature, pressure, mass flow rate, and / or sound.

[0048] Each transformer component in transformer assembly 16 may have different characteristics, particularly different physical and / or chemical properties. In practice, it has been found that not all characteristics of transformer assembly 16 can be detected by sensor 14. For example, the temperature in the primary or secondary coil can typically only be detected indirectly by the temperature of the coolant. Therefore, in principle, there is a functional relationship between the temperature of the primary or secondary coil and the temperature of the coolant detected by sensor 14. However, it is not possible to directly infer from the detected coolant temperature whether one of the two coils has a temperature and what temperature it has. Especially when the coolant temperature is too high, it is impossible to definitively determine which of the two coils is, for example, faulty, thus in any case, the faulty coil dissipates too much heat into the coolant, causing the aforementioned excessively high coolant temperature.

[0049] The above example makes it clear that the physical and / or chemical characteristics of the power transformer 12 detected by sensor 14 can indeed indicate that one of the transformer components 16 is in an abnormal, i.e., an anomalous condition. This abnormal or anomalous condition is also referred to as an anomalous condition of the corresponding transformer component 16. Therefore, there is a need to provide an apparatus 2 configured to, in particular, to obtain a specific estimate regarding whether one or more transformer components 16 are faulty and which transformer components are faulty, based on multiple physical and / or chemical characteristics detected by sensor 14 when significant characteristics of the power transformer 12 are sensed.

[0050] The problems described above are utilized as follows: Figure 1 The solution is exemplarily and schematically illustrated by device 2. Device 2 has the aforementioned signal input interface 4, which is configured for direct or indirect coupling with measurement system 10. Measurement system 10 has a plurality of sensors 14, which are respectively configured to detect the physical and / or chemical characteristics of power transformer 12. Power transformer 12 has a plurality of transformer components 16. Each sensor 14 of measurement system 10 is coupled to at least one transformer component of transformer component 16 of power transformer 12 via its respective direct or indirect connection. Preferably, a functional connection is involved here. Through the direct or indirect connection between sensor 14 and transformer component 16, the technical effect that the characteristics of power transformer 12 detected by the corresponding sensor 14 are affected by at least one transformer component of transformer component 16 can be achieved.

[0051] The measurement system 10 may have an input interface 38, an output interface 40, and an evaluation unit 42. Sensor 14 is preferably coupled to the input interface 38, so that the characteristics detected by sensor 14 are preferably transmitted to the evaluation unit 42 via sensor signals. The evaluation unit 42 may be configured to evaluate the characteristics detected by sensor 14 and generate a state signal S representing multiple different state parameters. Each state parameter may, for example, indicate whether one of the sensors 14 has detected a chemical and / or physical characteristic of the power transformer 12 that is greater than or less than a corresponding limit value. The state signal S represents different state parameters, wherein each state parameter is assigned to at least one of the sensors 14, such that each state parameter is coupled to at least one transformer component of the transformer component 16 via a functional relationship. Here, it is not a mechanical coupling, but a functional coupling in the sense of a functional relationship. Each state parameter represents the state of the characteristic of the power transformer 12 detected by the at least one assigned sensor 14 as either a normal state N or an abnormal state A. In other words, the state can be either normal (N) or abnormal (A), where the corresponding state is represented by a state parameter. For example, if the characteristic of a transformer is the temperature of the coolant, then when the temperature is below a predetermined maximum temperature, the state of the coolant temperature can be in the normal state, and when the temperature reaches the maximum temperature or higher, it can be in the abnormal state.

[0052] The status signal S represents multiple different status parameters. Furthermore, the signal input interface 4 of device 2 is configured to receive the status signal S from measurement system 10. After measurement system 10 detects the characteristics of power transformer 12 using sensor 14, measurement system 10 can therefore transmit the status signal S from its respective output interface 40 to signal input interface 4 of device 2.

[0053] The processor unit 6 of device 2 is preferably coupled to the signal input interface 4, such that a state parameter represented by a state signal S is provided to the processor unit 6. The processor unit 6 of device 2 is configured to calculate, for each transformer component 16, a fault probability value based on a Bayesian network 18 and the state condition represented by the state parameter of the state signal, wherein the corresponding fault probability value indicates the probability that the corresponding transformer component 16 is in a fault condition. The fault condition can be an abnormal condition of each transformer component 16.

[0054] exist Figure 2An example of such a Bayesian network 18 is shown, which is designed as a directed graph 20 from transformer components 16 (referred to in this case as first transformer component T1 and second transformer component T2) to state parameters S1, S2. The directed graph 20 is based on a first probability value, a second probability value, and a third probability value. Graph 20 is a directed, acyclic, and bipartite graph, such that only the relationships between transformer components T1, T2 and state parameters S1, S2 exist. State parameters S1, S2 therefore depend on the states of transformer components T1, T2. However, the states of transformer components T1, T2 do not depend on the states of state parameters S1 and S2. Furthermore, there is no relationship between state parameters S1 and S2. It is also preferably specified that graph 20 does not depict the relationships between transformer components T1, T2. Transformer components T1, T2 form the first set of nodes in graph 20. State states S1, S2 form the second set of nodes in graph 20. The probability values ​​mentioned above will now be discussed.

[0055] Assign a first probability value (first W value) to each state parameter S1 and S2, so that the corresponding state parameters S1 and S2 represent the error condition. Figure 2 The first table 48 shows that a first probability value of 0.2 is set for the first state parameter S1, and a first probability value of 0.3 is set for the second state parameter S2. Therefore, the first probability value of 0.2 set for the first state parameter S1 indicates that the state represented by S1 is incorrect. If the first state parameter S1 represents, for example, the normal state N, then there is a 20% probability that the normal state N is incorrect and the actual situation must be the abnormal state A. The corresponding situation applies to the second state parameter S2 and its corresponding first probability value of 0.3.

[0056] For each transformer component T1 and T2, a second probability value (2W value) is set such that a fault occurs in the corresponding transformer component T1 or T2. If the first transformer component T1 is, for example, composed of a primary coil and the second transformer component T2 is composed of a secondary coil, then statistical investigation can determine that there is a 20% probability, for example, that the primary coil has a fault and the secondary coil has a 30% probability of fault. For each functional relationship from one of the transformer components T1 and T2 to one of the state parameters S1 and S2, represented by arrows in Figure 20, a third probability value (3W value) is set such that a fault in the corresponding transformer component T1 or T2 causes an abnormal condition A in the corresponding state parameter S1 or S2. Therefore, for example, when transformer component T1 is in abnormal condition A, the third probability value for the relationship between the first transformer component T1 and the first state parameter S1 is 0.5. Therefore, when transformer component T1 exhibits abnormal condition A, there is a 50% probability that the state condition represented by state parameter S1 is abnormal condition A. Therefore, a third probability value of 0.5 is set for the relationship between transformer component T1 and state parameter S1. As can be seen from Figure 20, however, there is also a case where the relationship between transformer component T1 and the second state parameter S2 exists. As can be seen from the third table 52, when the first transformer component T1 is in abnormal condition A, based on the corresponding third probability value, there is a 90% probability that the state condition represented by the second state parameter S2 is abnormal condition A. Device 2 has a storage unit 8, wherein a data set including a first probability value, a second probability value, and a third probability value is stored in the storage unit 8. The processor unit 6 of device 2 is configured to read the data set and / or the probability values ​​in the data set from the storage unit 8.

[0057] Based on the previously described relationships between the first, second, and third probability values ​​and Figure 20, processor unit 6 can use the corresponding data to generate a Bayesian network 18. However, it is also possible that the Bayesian network 18 already exists, more precisely, on the directed Figure 20 and the aforementioned probability values. The Bayesian network 18 can therefore represent a specific mapping between transformer components 16 or T1, T2 and state parameters S1, S2. The state signal S of the measurement system 10 received via the signal input interface 4 of device 2 represents multiple different state parameters, more precisely, the state parameters that actually exist in this case. Based on the Bayesian network 18 and the state conditions represented by the state parameters of the measured signals, processor unit 6 can therefore calculate the corresponding fault probability value for each transformer component 16 or T1, T2, i.e., the probability that the corresponding transformer component 16 or T1, T2 is in a fault condition. Each fault probability value thus indicates the probability that the corresponding transformer component 16 or T1, T2 is actually in that fault condition. Processor unit 6 can be configured to obtain the fault probability value by evaluating the Bayesian network 18 and taking into account the actual state conditions, more specifically, by means of, for example, one of the following algorithms: a sampling algorithm or a variable elimination algorithm. Processor unit 6 can be constructed and / or configured to implement the corresponding algorithm.

[0058] In particular, processor unit 6 is configured to generate multiple samples 22 based on Bayesian network 18 and first, second, and third probability values, wherein each sample 22 represents a corresponding sample case 24 for each state parameter and for each transformer component T1, T2, and T6 as either normal case N or abnormal case A. Figure 3 The table shown shows a total of twelve samples 22, numbered from 1 to 12, where each sample 22 represents sample condition 24 as normal condition (N) or abnormal condition (A) for each transformer component T1, T2 and each state parameter S1, S2.

[0059] Therefore, sample number 1 represents sample cases N, N, N, N, more precisely, in the following order: first transformer component T1, second transformer component T2, first state parameter S1, and second state parameter S2. Correspondingly, second sample number 2 represents sample cases A, N, N, A. Other sample cases can be obtained from... Figure 3 As shown in the table, processor unit 6 generates sample cases 24 such that the probability values ​​reflected in tables 48, 50, and 52 are satisfied. To achieve this, in practice, a large number of samples with corresponding sample cases are often generated. Therefore, for example, processor unit 6 is configured to generate at least 10,000 samples 22. Figure 3The multiple samples 22 shown are provided exemplarily only and are not required to meet all the conditions in Tables 48, 50, and 52. However, if we consider sample cases 24 in the sample 22 for the first transformer component T1, then from the sample cases 24 for the first ten samples, two sample cases 24 represent abnormal case A and the remaining sample cases 24 represent normal case N. Thus, the probability value of abnormal case A for the first transformer component T1 is 20%. Figure 2 The first table 48 shows the corresponding probability value for anomaly A of the first transformer component T1, namely, probability value 0.2.

[0060] The status signal S is transmitted from the measurement system 10 to the signal input interface 4 of the device 2. The transmitted status parameters represent the actual state conditions, either normal condition N or abnormal condition A. If the first status parameter S1 represents normal condition N and the second status parameter S2 represents abnormal condition A, then... Figure 3 Two samples 22 were precisely identified in the table, which represent the first state parameter S1 as normal condition N and the second state parameter S2 as abnormal condition A. Here, these are samples numbered 2 and 8. Figure 4 Only the two samples numbered 2 and 8 are shown. Processor unit 6 is preferably configured to implement the corresponding method steps. Therefore, it is specified that processor unit 6 is configured to process data from multiple samples 22 (…). Figure 3 The processor unit 6 is configured to obtain a reference group 26 for samples 22 (sample numbers 2 and 8 in this example) such that in each sample (number 2 and number 8) of the reference group 26, the sample condition 24 of the state parameters S1 and S2 is consistent with the state condition of the state parameter of the state signal S. Furthermore, the processor unit 6 is configured to, for each transformer component T1 and T2, obtain the corresponding fault probability value of the transformer component T1 and T2 in abnormal condition A from samples 22 (numbers 2 and 8) of the reference group 26 based on the sample condition 24 for the transformer component T1 and T2. Figure 3The table illustrates the samples (sample numbers 2 and 8) of reference group 26, which have state parameters S1 and S2 corresponding to the actual state parameters S1 and S2 obtained by measurement system 10. However, for each sample in reference group 22, sample case 24 is also described for each of the two transformer components T1 and T2. The failure probability value for each transformer component can be obtained from the average of the described sample cases for the corresponding transformer components T1 and T2. For the first transformer component T1, sample number 2 describes abnormal condition A and sample number 8 describes normal condition N. Therefore, the average is 50%, and the failure probability value for the first transformer component T1 is 0.5. In the second transformer component T2, sample case 24 is given as normal condition N in the two samples with numbers 2 and 8, and the average is 100% normal condition N. Therefore, the failure probability value for the second transformer component T2 is 0. In summary, it can be determined that when the measurement system 10 determines normal condition N for the first state parameter and abnormal condition A for the second state parameter, then the device 2 calculates a fault probability value of 0.5 for the first transformer component T1 and a fault probability value of 0 for the second transformer component T2. This means that the second transformer component T2 is in normal condition N and the first transformer component T1 is in fault condition or abnormal condition A with a 50% probability. Fault condition or abnormal condition A can be synonymous.

[0061] As by Figure 1 As schematically shown, device 2 can be coupled to display device 30 using a separate signal line. Alternatively or additionally, device 2 can have an integrated display unit 34. Display unit 34 can be coupled to processor unit 6. Processor unit 6 can be configured and / or constructed to control display unit 34, so that display unit 34 reproduces an image that graphically represents the obtained fault probability value and indicates and / or identifies the corresponding transformer component 16. Thus, a person observing display unit 34 of device 2 can directly identify which transformer component 16 has an abnormal condition A or a fault condition with what probability. Such an understanding can be derived empirically from the state conditions generated by measurement system 10, not indirectly and directly.

[0062] Possibly, device 2 has an output signal interface 36, wherein device 2 is configured to generate an output signal Q, which at least represents the maximum fault probability value among the calculated fault probability values ​​and the corresponding transformer component 16. Furthermore, device 2 is preferably configured and / or arranged to control the signal output interface 36 in such a way that the output signal Q is transmitted, in particular, sent, more precisely preferably transmitted to the display device 30 via the signal output interface 36.

[0063] As previously explained, device 2 can be part of system 28, which, in addition to device 2, also includes a display unit 30 with a screen. Device 2 can be constructed and / or configured to transmit an output signal Q to display unit 30 via signal output interface 36, whereby display unit 30 displays an image that graphically represents the fault probability value and indicates and / or identifies the corresponding transformer component 16. Display unit 30 can be constructed accordingly for this purpose. A signal connection, particularly a wired and / or wireless signal connection, can exist between signal output interface 36 and display unit 30.

[0064] In practice, the focus is often on the transformer component 16 for which the maximum failure probability value has been calculated, because in this case, it can be assumed that the transformer component 16 is very likely in abnormal condition A. However, the failure probability value represents the probability that each transformer component 16 is in abnormal condition A. Therefore, there is a residual probability that the corresponding transformer component 16 is not in abnormal condition A but in normal condition N. Therefore, the second or third largest failure probability value is also usually of interest. Therefore, it is preferred that the device 2 is configured to generate the output signal Q such that the output signal Q represents at least two or three maximum failure probability values ​​and the corresponding transformer component 16. The display unit 30 of the system 28 is preferably configured to display a second graph indicating the transformer component 16 and its corresponding failure probability value on the screen of the evaluation unit 30 based on the output signal Q. Therefore, the display unit 30 can be configured such that the image displayed on the screen shows the transformer component 16 represented by the output signal Q in list form and qualitatively displays the corresponding failure probability value by different colors and / or bars. Figure 5 The example shown is a screen image reflected on a screen, which includes the previously mentioned image, by representing transformer components 16 in list form and qualitatively representing their respective fault probability values ​​through (colored) bars and the length of the bars. The longer the bar, the greater the fault probability obtained for the corresponding transformer component 16. Transformer components 16 are listed according to the magnitude of their fault probability values, with the transformer component 16 with the highest fault probability value arranged at the top. From this list-based representation of the transformer components 16 and their respective fault probability values, a person can at a glance identify which transformer component 16 is most likely to be in abnormal condition A, and simultaneously identify which other transformer components 16 are also in abnormal condition A with a certain probability by comparing the color and / or length of the bars. This facilitates maintenance and / or troubleshooting in the power transformer 12.

[0065] exist Figure 1The system 28 shown schematically includes a device 2 and a display unit 30. It may be specified that the system 28 additionally includes a measurement system 10 having an associated sensor 14. Here, the display unit 30 may be connected to the measurement system 10 or may be integrated into the measurement system 10. Furthermore, it is possible that the system 28 additionally includes a power transformer 12.

[0066] The measurement system 10 may include a first sensor 14 configured to detect the temperature of the coolant in the reservoir of the power transformer 12. A further sensor 14 may be configured to detect mechanical stress or force acting on the primary coil. A third sensor 14 of the measurement system 10 may be configured to detect the voltage between terminals 44 on the primary side of the power transformer 12. A fourth sensor 14 of the measurement system may be configured to detect the loss factor and / or capacitance in one or more leads of the power transformer 12. A fifth sensor 14 of the measurement system 10 may be configured to detect acoustic signals generated when the tap changer of the power transformer 12 is operated. Furthermore, other sensors or additional sensors 14 are also possible for the measurement system 10.

[0067] Furthermore, it is preferably specified that the device 2 is configured to periodically re-detect the fault probability value of the transformer component 16, more specifically, preferably always after receiving a state signal S with new multiple different state parameters from the measurement system 10. The measurement system 10 can also be configured to periodically evaluate the characteristics of the power transformer 12 detected by its associated sensor 14. After each periodic detection of the characteristics by the associated sensor 14, the measurement system 10 can transmit the state signal S to the signal input interface 4 of the device 2. Then, the fault probability of the transformer component 16 is calculated by means of the processor unit 6. Furthermore, the output signal Q can be updated by means of the processor unit 6 based on the last calculated fault probability, so that a current image of which transformer components 16 have the highest probability of failure is always output on the display device.

[0068] It should also be noted that "having" does not exclude other elements or steps, and "a" or "an" does not exclude multiple. Furthermore, it should be noted that features described with reference to one embodiment of the above exemplary embodiments may also be used in combination with other features of the other embodiments described above. Reference numerals in the claims should not be considered limiting.

[0069] List of reference numerals

[0070] A. Abnormal situation

[0071] Q output signal

[0072] S state signal

[0073] N Normal situation

[0074] S1 State Parameter

[0075] S2 State Parameters

[0076] T1 Transformer Components

[0077] T2 Transformer Components

[0078] 2. Device

[0079] 4. Signal input interface of the device

[0080] 6 processor units

[0081] 8 storage units

[0082] 10 Measurement System

[0083] 12 Power Transformers

[0084] 14 Sensors

[0085] 16 Transformer Components

[0086] 18 Bayesian Networks

[0087] 20 charts

[0088] 22 samples

[0089] 24 Sample Details

[0090] 26 Reference Group

[0091] 28 System

[0092] 30 display units

[0093] 34 Display unit of the device

[0094] 36. Signal output interface of the device

[0095] 38 Input Interfaces

[0096] 40 Output Interfaces

[0097] 42 Evaluation Units

[0098] 44. Terminals on the primary side

[0099] 46 Secondary side terminals

[0100] 48 Table 1 (1W values)

[0101] 50. Second Table (2W Values)

[0102] 52. Table 3 (3W values)

Claims

1. An apparatus (2) for determining the failure probability value for a transformer component (16), the apparatus comprising: Signal input interface (4). Processor unit (6), and Storage unit (8), in, The signal input interface (4) is configured to be directly or indirectly coupled to a measurement system (10) for a power transformer (12), wherein the measurement system (10) has a plurality of sensors (14) configured to detect physical and / or chemical properties of the power transformer (12), wherein the power transformer (12) has a plurality of transformer components (16), and each sensor (14) is coupled to at least one of the transformer components (16) of the power transformer (12) by a corresponding direct or indirect connection, such that the properties of the power transformer (12) detected by the corresponding sensor (14) are affected by at least one of the transformer components (16); The signal input interface (4) is configured to receive the status signal S of the measurement system (10), wherein the status signal S represents a plurality of different status parameters S1, S2, wherein each status parameter S1, S2 is assigned to at least one of the sensors (14), such that each status parameter S1, S2 is in a functional relationship with at least one of the transformer components (16), wherein each status parameter S1, S2 represents the state of the power transformer (12) as normal condition N or as abnormal condition A, as detected by the assigned at least one sensor (14); The data set is stored in storage unit (8). Each data set has a first probability value for each state parameter S1, S2, such that the corresponding state parameters S1, S2 represent an error condition. Each data set also has a second probability value for each transformer component (16), such that a fault occurs in the corresponding transformer component (16). Furthermore, each data set has a third probability value for each functional relationship from one transformer component (16) to one of the state parameters S1, S2, such that a fault in the corresponding transformer component (16) causes an abnormal condition A in the corresponding state parameters S1, S2. The processor unit (6) is configured to obtain the fault probability value for each transformer component (16) based on a Bayesian network (18) and the state conditions represented by state parameters S1 and S2, such that the corresponding transformer component (16) is in a fault condition. The Bayesian network represents a directed graph (20) from the transformer component (16) to the state parameters S1 and S2 based on the first, second, and third probability values.

2. The device (2) according to claim 1, characterized in that, The processor unit (6) is configured to generate multiple samples (22) based on a Bayesian network (18), the Bayesian network representing a directed graph (20) from transformer components (16) to state parameters S1, S2 based on the first, second, and third probability values, wherein each sample (22) represents a corresponding sample case (24) belonging to either normal case N or abnormal case A for each state parameter S1, S2 and for each transformer component (16); wherein the processor unit (6) is configured to generate multiple samples (22) from the multiple samples (24) 2) Determine a reference group (26) for the sample (22) such that the sample condition (24) for the state parameters S1, S2 in each sample (22) of the reference group (26) is consistent with the state condition of the state parameters S1, S2 of the state signal S; and the processor unit (6) is configured to obtain the fault probability value for each transformer component (16) from the sample (22) of the reference group (26) based on the sample condition (24) for the transformer component (16), such that the corresponding transformer component (16) is in abnormal condition A.

3. The device (2) according to claim 1, characterized in that, The device (2) has an output interface (40), wherein the device (2) is configured to generate an output signal Q, the output signal representing at least the maximum fault probability value and the corresponding transformer component (16) in the fault probability values, and the device (2) is configured to transmit the output signal Q via a signal output interface (36).

4. The device (2) according to claim 3, characterized in that, The device (2) is configured to generate an output signal Q such that the output signal Q represents at least two or three maximum failure probability values ​​and the corresponding transformer component (16).

5. The apparatus (2) according to any one of claims 1 to 4, characterized in that, The state parameters S1 and S2 represented by the state signal S are ternary state parameters, which respectively represent exactly one of the following groups of possible state conditions: normal condition N, abnormal condition A, or non-installation condition.

6. The apparatus (2) according to any one of claims 1 to 4, characterized in that, The processor unit (6) is configured to generate a Bayesian network (18) based on a data set.

7. The apparatus (2) according to any one of claims 1 to 4, characterized in that, The Bayesian network (18) is constructed as a bipartite graph (20), which represents a first group of nodes, a second group of nodes, and a functional relationship from the first group of nodes to the second group of nodes, wherein the first group of nodes is determined by a first probability value, the second group of nodes is determined by a second probability value, and the functional relationship is determined by a third probability value.

8. The apparatus (2) according to any one of claims 1 to 4, characterized in that, The processor unit (6) is configured to generate samples (22) using a probability-weighted algorithm.

9. The apparatus (2) according to claim 4, characterized in that, The processor unit (6) is configured to generate an output signal Q such that the output signal Q represents the fault probability value and the transformer component (16) in a first figure, in which the fault probability value is graphically represented and the transformer component (16) is indicated and / or labeled.

10. A system (28) comprising: The device (2) according to any one of claims 4 to 8, and A display unit (30) with a screen. in, The signal output interface (36) of the device (2) is coupled to the display unit (30) via a direct or indirect first signal connection so as to transmit the output signal Q to the display unit (30), and The display unit (30) is configured to display a second graph on the screen based on the output signal Q, indicating the transformer component (16) and the associated fault probability value.

11. The system (28) according to claim 10, characterized in that, The display unit (30) is configured such that the second figure displays the transformer components (16) represented by the output signal Q in a list form and qualitatively displays the corresponding fault probability values ​​by different colors and / or stripes.

12. The system (28) according to claim 10 or 11, characterized in that, The system (28) has a measurement system (10) for a power transformer (12), wherein the measurement system (10) is coupled to a signal input interface (4) of the device (2) to transmit a status signal S to the signal input interface (4), wherein the measurement system (10) has a plurality of sensors (14) configured to detect physical and / or chemical properties of the power transformer (12), wherein the power transformer (12) has a plurality of transformer components (16), and each sensor (14) is coupled to at least one of the transformer components (16) of the power transformer (12) by a corresponding direct or indirect connection, such that the properties of the power transformer (12) detected by the corresponding sensor (14) are affected by at least one of the transformer components (16).

13. The system (28) according to claim 12, characterized in that, The measurement system (10) is configured to sense at least two of the following characteristics of the power transformer (12): hydrocarbon chain gas in the liquid insulating medium, water in the liquid insulating medium, the filling level of the liquid insulating medium, voltage, current, winding tension, fan speed, fan motor power, fan volume, tap changer motor torque, tap changer switching sequence, hot spot temperature, top oil temperature, loss factor of one or more leads of the power transformer (12), capacity of one or more leads of the power transformer (12), acoustic signal of the tap changer, vibration sound intensity of the air dehumidifier, and functional effectiveness.

14. The system (28) according to claim 10 or 11, characterized in that, The display unit (30) forms part of the measurement system (10).

15. The system (28) according to claim 10 or 11, characterized in that, The system (28) has a remote monitoring system, wherein the signal output interface (36) of the device (2) is coupled to the remote monitoring system via a direct or indirect second signal connection so as to transmit the output signal Q to the remote monitoring system.

16. The system (28) according to claim 15, characterized in that, The remote monitoring system is based on a cloud-ground architecture.

17. The system (28) according to claim 15, characterized in that, The display unit (30) forms part of the remote monitoring system.

18. The system (28) according to claim 10 or 11, characterized in that, The system (28) includes a power transformer (12).

Images