System and method for monitoring through-fault current
The method addresses the challenge of monitoring through-fault current in power transformers by calculating electrical stress and state changes, using a machine learning model to predict transformer readiness for future faults, ensuring timely maintenance and reducing damage.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- GENERAL ELECTRIC TECH GMBH
- Filing Date
- 2024-04-24
- Publication Date
- 2026-07-07
AI Technical Summary
Existing power transformer systems lack effective methods to accurately monitor through-fault current and assess the impact of through faults on transformer condition, leading to potential damage from electrical and mechanical stresses.
A method involving detection of through-faults, calculation of electrical stress and peak current, determination of state change rates, assignment of weights and criticalities, and training of a machine learning model to predict mechanical and thermal state changes, enabling assessment of the transformer's readiness for subsequent faults.
Enables accurate monitoring and prediction of transformer health, allowing for timely maintenance and preparation for future faults, thereby reducing damage and extending transformer lifespan.
Smart Images
Figure 2026522174000001_ABST
Abstract
Description
Technical Field
[0001] This application and the resulting patents generally relate to power transformer systems, and more particularly, to systems and methods for monitoring through fault current in a power transformer system.
Background Art
[0002] Generally, a through fault is a system fault in a power transformer system that is outside the protection zone associated with the power transformer. The protection zone for the differential relay of a power transformer can be defined by the location of the secondary circuit of the current transformer. A through fault can expose the power transformer to electrical and mechanical stresses, which can then reduce the insulation and mechanical strength of the power transformer. Damage to the power transformer can depend on the magnitude and duration of the through fault current.
[0003] Therefore, there is an increasing need for a method to accurately monitor the through fault current and accurately determine the impact of the through fault on the condition of the power transformer. Furthermore, there is an increasing need for a method to evaluate the state changes associated with the power transformer after a through fault and determine whether the power transformer is prepared to withstand subsequent through faults.
Prior Art Documents
Patent Documents
[0004]
Patent Document 1
Summary of the Invention
[0005] Therefore, this application and the resulting patent provide a method for monitoring through-fault current. The method includes the steps of: detecting a first occurrence of a first through-fault in a power transformer system; calculating a first electrical stress associated with the first through-fault, a first peak current associated with the first through-fault, and a first duration of the first through-fault; determining a first set of percentage state change associated with the first through-fault and a second set of percentage state change associated with the first through-fault; assigning a first set of weights and a first set of criticality to the first set of percentage state change; calculating a first mechanical state change based at least in part on the first set of percentage state change, the first set of weights, and the first set of criticality; assigning a second set of weights and a second set of criticality to the second set of percentage state change; and The method may include the steps of: calculating a first thermal state change based at least partially on a set of 2, a second set of weights, and a second set of criticality; calculating a first cumulative state change based at least partially on a first mechanical state change and a first thermal state change; and training a machine learning model using a first electrical stress, a first peak current, a first duration, a first set of state change rates, a second set of state change rates, a first mechanical state change, a first thermal state change, and a first cumulative state change.
[0006] This application and the resulting patent further provide a method for monitoring through-fault currents. The method may include the steps of: detecting the first occurrence of a first through-fault in a power transformer system; calculating a first electrical stress associated with the first through-fault, a first peak current associated with the first through-fault, and a first duration of the first through-fault; determining a first set of state change rates associated with the first through-fault and a second set of state change rates associated with the first through-fault; calculating a first mechanical state change based at least partially on the first set of state change rates; calculating a first thermal state change based at least partially on the second set of state change rates; calculating a first cumulative state change based at least partially on the first mechanical state change and the first thermal state change; and training a machine learning model using the first electrical stress, the first peak current, the first duration, the first set of state change rates, the second set of state change rates, the first mechanical state change, the first thermal state change, and the first cumulative state change.
[0007] This application and the resulting patent further provide a power transformer system. The power transformer system may include a power transformer and a controller, the controller of detecting a first occurrence of a first through fault in the power transformer system, calculating a first electrical stress associated with the first through fault, a first peak current associated with the first through fault, and a first duration of the first through fault, determining a first set of state change rates associated with the first through fault and a second set of state change rates associated with the first through fault, assigning a first set of weights and a first set of criticality to the first set of state change rates, and determining a first machine based at least in part on the first set of state change rates, the first set of weights and the first set of criticality. The system is configured to calculate mechanical state changes, assign a second set of weights and a second set of criticality to a second set of state change rates, calculate a first thermal state change based at least partially on the second set of state change rates, the second set of weights, and the second set of criticality, calculate a first cumulative state change based at least partially on the first mechanical state change and the first thermal state change, and train a machine learning model using a first electrical stress, a first peak current, a first duration, a first set of state change rates, a second set of state change rates, a first mechanical state change, a first thermal state change, and a first cumulative state change.
[0008] These features and improvements of this application and the resulting patent, as well as other features and improvements, will become apparent to those skilled in the art by examining the following detailed description in conjunction with some drawings and the attached claims. [Brief explanation of the drawing]
[0009] [Figure 1] This flowchart shows a method for monitoring through-fault current according to one or more exemplary embodiments of the present disclosure. [Figure 2]This flowchart shows a method for monitoring through-fault current according to one or more exemplary embodiments of the present disclosure. [Figure 3] This flowchart shows a method for monitoring through-fault current according to one or more exemplary embodiments of the present disclosure. [Figure 4] This flowchart shows a method for monitoring through-fault current according to one or more exemplary embodiments of the present disclosure. [Modes for carrying out the invention]
[0010] Referring here to the drawings, similar symbols throughout several figures refer to similar elements, and Figure 1 is a flowchart 100 of a method for monitoring through-fault current. Flowchart 100 may be applicable to through-faults in a power transformer system. In block 102, the occurrence of a through-fault in a power transformer system may be detected. The occurrence of a through-fault may be identified based on at least a set of circuit breaker statuses associated with the power transformer system. The set of circuit breaker statuses may be associated with the primary, secondary, and / or tertiary sides of the power transformer in the power transformer system. In some cases, the occurrence of a through-fault may be identified based on at least a determination that a set of circuit breakers has not tripped and the detection of a high symmetric current. The definition of a high current may vary depending on the category of the power transformer.
[0011] In block 104, the electrical stress associated with the through-fault, the peak current associated with the through-fault, and the duration associated with the through-fault can be calculated. The electrical stress may be a function of the through-fault current and the duration of the through-fault. The peak current may be calculated for each of the three phases of the current associated with the through-fault. The start and end dates and times of the through-fault may also be recorded. Then, the duration associated with the through-fault may be calculated based on the start and end dates and times. In some cases, the duration may be between 10 milliseconds and 1 second. The through-fault current associated with the start and end dates and times of the through-fault may also be recorded. The cumulative number of through-faults, the accumulated stress associated with the through-faults (as a function of the through-fault current and the duration of the through-faults), and the root mean square current associated with the through-faults may be determined.
[0012] In block 106A, a first set of state change rates related to through-hole failures may be determined. In block 106B, a second set of state change rates related to through-hole failures may be determined. The first set of state change rates may include state change rates related to dissolved gas analysis of the power transformer system, state change rates related to the rate of change of the power transformer system, state change rates related to the gas ratio related to the power transformer system, and / or other related state change rates. The second set of state change rates may include state change rates related to winding hotspot temperature or thermal output of the power transformer system, state change rates related to aging degradation factors related to the power transformer system, state change rates related to loss of life related to the power transformer system, and / or other related state change rates.
[0013] Determining a first set of state change rates and a second set of state change rates may involve identifying a set of mechanical measurements associated with the power transformer system and a set of thermal measurements associated with the power transformer system. Each mechanical measurement in the set of mechanical measurements and each thermal measurement in the set of thermal measurements may be associated with a corresponding condition. In one example, each condition may be one of a predetermined set of conditions. For example, the predetermined set of conditions may include conditions 1, 2, 3, and 4. Each mechanical measurement and each thermal measurement associated with a corresponding condition can then be used to determine the first set of state change rates and the second set of state change rates.
[0014] In block 108A, a first set of weights and a first set of criticalities may be assigned to a first set of state change rates. In block 108B, a second set of weights and a second set of criticalities may be assigned to a second set of state change rates. For example, the first set of weights and the second set of weights may include weights ranging from 0.1 to 1.0. In one example, each weight in the first set of weights and the second set of weights may range from 0.1 to 0.3 if the state change rate in the first set of state change rates or the second set of state change rates relates to condition 1, from 0.3 to 0.5 if the state change rate relates to condition 2, from 0.5 to 0.7 if the state change rate relates to condition 3, and from 0.7 to 1.0 if the state change rate relates to condition 4. In one example, each criticality in the first set of criticalities and the second set of criticalities may range from 0.1 to 1.0.
[0015] In block 110A, a first set of state change rates, a first set of weights, and a first set of criticalities may be used to calculate the mechanical state change associated with a through-fault. In some cases, the mechanical state change may be calculated by (1) multiplying each state change rate by an applicable weight and then by an applicable criticality, (2) summing the values obtained in step (1), and (3) dividing the sum obtained in step (2) by the number of state change rates. In block 110B, a second set of state change rates, a second set of weights, and a second set of criticalities may be used to calculate the thermal state change associated with a through-fault. In some cases, the thermal state change may be calculated by (1) multiplying each state change rate by an applicable weight and then by an applicable criticality, (2) summing the values obtained in step (1), and (3) dividing the sum obtained in step (2) by the number of state change rates.
[0016] In block 112, cumulative state changes can be calculated based on mechanical and thermal state changes. In some cases, cumulative state changes may indicate the impact of through-faults on the integrity of the power transformer. Cumulative state changes may be further calculated based on a third set of weights applied to mechanical state changes and a fourth set of weights applied to thermal state changes. The third and fourth sets of weights may relate to the power transformer's past failure history. In some cases, cumulative state changes may be calculated by (1) multiplying the mechanical state changes by the third set of weights, (2) multiplying the thermal state changes by the fourth set of weights, (3) summing the values obtained in steps (1) and (2), and (4) dividing the sum obtained in step (3) by 2.
[0017] In block 114, a machine learning model may be trained using electrical stress, peak current, duration, a first set of state change rates, a second set of state change rates, mechanical state changes, thermal state changes, and cumulative state changes. When the machine learning model is being trained, the power transformer system may be configured to operate in a healthy mode. During the training of the machine learning model, the first set of state change rates and the second set of state change rates may be correlated with the electrical stress, peak current, and duration of a through-fault in order to observe the healthy behavior of the power transformer.
[0018] Figure 2 is a flowchart 200 of a method for monitoring through-fault current. In block 202, the state of the power transformer system may be determined using a machine learning model. The state of the power transformer system may be determined in light of its occurrence after another through-fault in the power transformer system. When a subsequent through-fault is detected, the electrical stress, peak current, and duration associated with the subsequent through-fault may be calculated. A first set and a second set of state change rates associated with the subsequent through-fault may also be determined. In block 204A, the measured mechanical and thermal state changes associated with the subsequent through-fault may be calculated based on the first and second sets of state change rates. Furthermore, in block 204A, the measured cumulative state change may also be calculated using the measured mechanical and thermal state changes.
[0019] Next, in block 204B, a machine learning model may be used to estimate the predicted mechanical state changes, predicted thermal state changes, and predicted cumulative state changes associated with subsequent thermal failures. In block 206, a first difference between the predicted mechanical state changes and the measured mechanical state changes may be calculated. Furthermore, in block 206, a second difference between the predicted thermal state changes and the measured thermal state changes may be calculated. Furthermore, in block 206, a third difference between the predicted cumulative state changes and the measured cumulative state changes may be calculated. Next, in block 208, the first, second, and third differences may be used to determine the state of the power transformer system.
[0020] Next, in block 210, the status of the power transformer system may be used to determine whether the power transformer system is prepared for another through-fault. In some cases, the degradation caused by the through-fault may be calculated, and the risk of the power transformer system being susceptible to another through-fault may be assessed. An indication of whether the power transformer system is prepared may be output to the operator. If the power transformer system is not prepared for another through-fault, the indication may include an alarm, warning, and / or any other indication to the operator. In some cases, the alarm, warning, and / or any other indication may show the measured cumulative state change and / or predicted cumulative state change, along with a summary explaining why the power transformer system is not prepared for another through-fault. This summary may include a description of the power transformer subsystems that may have caused the increase in the measured cumulative state change and / or predicted cumulative state change. Such a notification system allows the operator to schedule early maintenance to address issues related to the affected subsystems so that the power transformer system can be configured to withstand another through-fault.
[0021] In block 212, a risk index of the power transformer can be calculated. In some cases, the risk index of the power transformer may indicate the state of the power transformer. For example, the risk index of the power transformer may be on a scale of 1 to 5, where 1 indicates that the power transformer is in good condition and 5 indicates that the power transformer contains defects that may cause a failure. In some cases, the risk index of the power transformer may affect a set of weights and criticalities used in the calculation of applicable mechanical state changes, thermal state changes, and cumulative state changes.
[0022] In some cases, when certain conditions are met, an alarm, a warning, and / or some other indication may be output to the operator. The above conditions may include a total accumulated damage exceeding a predetermined threshold and / or a through-fault counter exceeding a predetermined threshold. The total accumulated damage and the through-fault counter can be judged for each phase.
[0023] In some cases, the state of the power transformer system may indicate the total amount of accumulated damage to the power transformer system. The total accumulated damage may be the cumulative damage caused by previous through-faults. The total amount of accumulated damage may be judged differently based on the category of the transformer. For example, depending on the frequency of through-faults, only the thermal state change, rather than both the mechanical state change and the thermal state change, may be used to judge the cumulative state change. The total amount of accumulated damage may also be judged differently according to the total accumulated damage, or the total accumulated damage for each phase. For example, if the total accumulated damage, or the total accumulated damage for each phase, is below a predetermined threshold, only the thermal state change, rather than both the mechanical state change and the thermal state change, may be used to judge the cumulative state change.
[0024] In block 214, free-ranking can be updated based on the risk index of the power transformer. Free-ranking can be an organizational system for classifying various power transformers in the system based on their respective risk indices of the power transformers.
[0025] Figure 3 is a flowchart 300 of a method for monitoring through-fault current. In block 302, the power transformer is activated to operate in healthy mode. In block 304, the power transformer system evaluates whether an event has occurred. If no event is detected, the power transformer continues to operate in healthy mode. If an event is detected, the power transformer system determines whether the event is a through-fault. If the event is a through-fault, in block 306, the following data may be captured: dissolved gas analysis, rate of change, gas ratio, winding hotspot temperature or thermal output, aging factors, loss of life, electrical stress, and / or other applicable data. The captured data may then be used as input to a machine learning model 308. During a through-fault, the power transformer system may further evaluate the electrical stress during the through-fault, the duration of the through-fault, and the peak current associated with the through-fault. The electrical stress, duration, and peak current may also be used as input to a machine learning model 308.
[0026] Next, the machine learning model 308 can correlate the data inputs and output thermal state changes, mechanical state changes, and cumulative state changes. In block 310, the machine learning model 308 can be used for statistical analysis and to extract thresholds related to the power transformer. For example, the thresholds may relate to any applicable variables, such as electrical stress, peak current, thermal state changes, mechanical state changes, cumulative state changes, power transformer state, dissolved gas analysis, rate of change, gas ratio, winding hotspot temperature or thermal output, aging degradation factors, loss of life, or any other relevant variables.
[0027] Figure 4 is a flowchart 400 of a method for monitoring through-fault current. In block 402, the power transformer is activated to operate in monitoring mode. In block 404, the power transformer system evaluates whether an event has occurred. If no event is detected, the power transformer continues to operate in monitoring mode. If an event is detected, the power transformer system determines whether the event is a through-fault. If the event is a through-fault, in block 406, the following data may be captured: dissolved gas analysis, rate of change, gas ratio, winding hotspot temperature or thermal output, aging factors, loss of life, electrical stress, and / or other applicable data. The captured data may then be used as input to a machine learning model. Then, in block 408, the machine learning model may be used to estimate predicted thermal state changes, predicted mechanical state changes, and predicted cumulative state changes.
[0028] During a through-fault, the power transformer system can further evaluate the electrical stress during the through-fault, the duration of the through-fault, and the peak current associated with the through-fault. The captured data, along with the electrical stress, the duration of the through-fault, and the peak current associated with the through-fault, can be used to calculate the thermal state change, the mechanical state change, and the cumulative state change. In block 410, the calculated thermal state change, the calculated mechanical state change, and the calculated cumulative state change can be compared with the predicted thermal state change, the predicted mechanical state change, and the predicted cumulative state change. Furthermore, in block 410, the captured data, the electrical stress, the peak current associated with the through-fault, and the duration of the through-fault can be compared with predetermined thresholds, such as thresholds determined according to Figure 3. In block 412, the power transformer state and cumulative state change can be extracted based on the comparisons performed in block 410. In block 414, the power transformer system can determine, based on the power transformer state and cumulative state change, whether the power transformer is prepared for another through-fault. If the power transformer system is prepared for another through-fault, the power transformer may continue to operate in monitoring mode. If the power transformer system is not prepared for another through-fault, an alarm, warning, or indication may be provided to the operator in block 416.
[0029] It is clear that the above pertains only to specific embodiments of this application and the resulting patent. Those skilled in the art can make numerous changes and modifications herein without departing from the general spirit and scope of the invention as defined by the following claims and equivalents.
[0030] Further aspects of the present invention are provided by the subject matter of the following clauses.
[0031] 1. A method for monitoring through-fault current, comprising: detecting the first occurrence of a first through-fault in a power transformer system; calculating a first electrical stress associated with the first through-fault, a first peak current associated with the first through-fault, and a first duration of the first through-fault; determining a first set of state change rates associated with the first through-fault and a second set of state change rates associated with the first through-fault; assigning a first set of weights and a first set of criticality to the first set of state change rates; and determining a first mechanical state change based at least in part on the first set of state change rates, the first set of weights, and the first set of criticality. A method comprising: calculating a change; assigning a second set of weights and a second set of criticality to a second set of change rate; calculating a first thermal change based at least partially on the second set of change rate, the second set of weights, and the second set of criticality; calculating a first cumulative change based at least partially on a first mechanical change and a first thermal change; and training a machine learning model using a first electrical stress, a first peak current, a first duration, a first set of change rate, a second set of change rate, a first mechanical change, a first thermal change, and a first cumulative change.
[0032] 2. The method according to Clause 1, wherein the first set of phase change rates includes at least a first phase change rate related to dissolved gas analysis, a second phase change rate related to the rate of change, and a third phase change rate related to the gas ratio.
[0033] 3. The method according to Clause 1 or 2, wherein the second set of state change rates includes at least a fourth state change rate related to winding hot spot temperature or thermal output, a fifth state change rate related to aging degradation factors, and a sixth state change rate related to loss of life.
[0034] 4. The method described in any one of clauses 1 to 3, wherein the machine learning model is trained when the power transformer system is in a healthy mode.
[0035] 5. The method according to any one of the clauses 1 to 4, wherein determining a first set of state change rates related to a first through-fault and a second set of state change rates related to a first through-fault further comprises identifying a set of mechanical measurements related to a power transformer system and a set of thermal measurements related to a power transformer system, determining the respective conditions related to each mechanical measurement in the set of mechanical measurements and each thermal measurement in the set of thermal measurements, and determining the first set of state change rates and the second set of state change rates based at least in part on the respective conditions related to each mechanical measurement in the set of mechanical measurements and each thermal measurement in the set of thermal measurements.
[0036] 6. Detecting a second occurrence of a second through-fault in a power transformer system; calculating a second electrical stress associated with the second through-fault, a second peak current associated with the second through-fault, and a second duration of the second through-fault; determining a third set of state change rates associated with the second through-fault and a fourth set of state change rates associated with the second through-fault; and determining a second mechanical state change associated with the second through-fault and a second mechanical state change associated with the second through-fault, at least partially based on the third set of state change rates and the fourth set of state change rates. The method according to any one of Clauses 1 to 5, further comprising: calculating the thermal state change and the second cumulative state change due to the second through-pass failure; using a machine learning model to estimate the third mechanical state change associated with the second through-pass failure, the third thermal state change associated with the second through-pass failure, and the third cumulative state change due to the second through-pass failure; and determining the state of the power transformer system based at least in part on the second mechanical state change, the second thermal state change, the second cumulative state change, the third mechanical state change, the third thermal state change, and the third cumulative state change.
[0037] 7. The method according to any one of the clauses 1 to 6, wherein determining the state of a power transformer system based at least in part on a second mechanical state change, a second thermal state change, a second cumulative state change, a third mechanical state change, a third thermal state change, and a third cumulative state change further comprises calculating a first difference between the second mechanical state change and the third mechanical state change, calculating a second difference between the second thermal state change and the third thermal state change, calculating a third difference between the second cumulative state change and the third cumulative state change, and determining the state of the power transformer system based at least in part on the first difference, the second difference, and the third difference.
[0038] 8. The method of any one of the clauses 1 to 7, further comprising determining, at least in part, that the power transformer system is not prepared for a third occurrence of a third through-fault, and outputting a display to the operator.
[0039] 9. The method according to any one of the clauses 1 to 8, wherein the first occurrence of a first through-fault is identified at least in part on the status of multiple circuit breakers associated with a power transformer system.
[0040] 10. A method for monitoring through-fault current, comprising: detecting a first occurrence of a first through-fault in a power transformer system; calculating a first electrical stress associated with the first through-fault, a first peak current associated with the first through-fault, and a first duration of the first through-fault; determining a first set of state change rates associated with the first through-fault and a second set of state change rates associated with the first through-fault; calculating a first mechanical state change based at least in part on the first set of state change rates; calculating a first thermal state change based at least in part on the second set of state change rates; calculating a first cumulative state change based at least in part on the first mechanical state change and the first thermal state change; and training a machine learning model using the first electrical stress, the first peak current, the first duration, the first set of state change rates, the second set of state change rates, the first mechanical state change, the first thermal state change, and the first cumulative state change.
[0041] 11. The method according to any one of the clauses 1 to 10, further comprising: assigning a first set of weights and a first set of criticalities to a first set of rate of change of state; assigning a second set of weights and a second set of criticalities to a second set of rate of change of state; calculating a first mechanical change of state based at least in part on the first set of rate of change of state, the first set of weights and the first set of criticalities; and calculating a first thermal change of state based at least in part on the second set of rate of change of state, the second set of weights and the second set of criticalities.
[0042] 12. The method according to any one of the clauses 1 to 11, wherein the first set of phase change rates includes at least a first phase change rate related to dissolved gas analysis, a second phase change rate related to the rate of change, and a third phase change rate related to the gas ratio.
[0043] 13. The method according to any one of the clauses 1 to 12, wherein the second set of state change rates includes at least a fourth state change rate related to winding hot spot temperature or thermal output, a fifth state change rate related to aging degradation factors, and a sixth state change rate related to loss of life.
[0044] 14. The method according to any one of the clauses 1 to 13, wherein determining a first set of state change rates related to a first through-fault and a second set of state change rates related to a first through-fault further comprises identifying a set of mechanical measurements related to a power transformer system and a set of thermal measurements related to a power transformer system, determining the respective conditions related to each mechanical measurement in the set of mechanical measurements and each thermal measurement in the set of thermal measurements, and determining the first set of state change rates and the second set of state change rates based at least in part on the respective conditions related to each mechanical measurement in the set of mechanical measurements and each thermal measurement in the set of thermal measurements.
[0045] 15. Detecting a second occurrence of a second through-fault in a power transformer system; calculating a second electrical stress associated with the second through-fault, a second peak current associated with the second through-fault, and a second duration of the second through-fault; determining a third set of state change rates associated with the second through-fault and a fourth set of state change rates associated with the second through-fault; and determining a second mechanical state change associated with the second through-fault and a second mechanical state change associated with the second through-fault, at least partially based on the third set of state change rates and the fourth set of state change rates. The method according to any one of Clauses 1 to 14, further comprising: calculating the thermal state change and the second cumulative state change due to the second through-pass failure; using a machine learning model to estimate the third mechanical state change associated with the second through-pass failure, the third thermal state change associated with the second through-pass failure, and the third cumulative state change due to the second through-pass failure; and determining the state of the power transformer system based at least in part on the second mechanical state change, the second thermal state change, the second cumulative state change, the third mechanical state change, the third thermal state change, and the third cumulative state change.
[0046] 16. The method according to any one of the clauses 1 to 15, wherein determining the state of a power transformer system based at least in part on a second mechanical state change, a second thermal state change, a second cumulative state change, a third mechanical state change, a third thermal state change, and a third cumulative state change further comprises calculating a first difference between the second mechanical state change and the third mechanical state change, calculating a second difference between the second thermal state change and the third thermal state change, calculating a third difference between the second cumulative state change and the third cumulative state change, and determining the state of the power transformer system based at least in part on the first difference, the second difference, and the third difference.
[0047] 17. A power transformer system comprising a power transformer and a controller, wherein the controller detects the first occurrence of a first through-fault in the power transformer system, calculates a first electrical stress associated with the first through-fault, a first peak current associated with the first through-fault, and a first duration of the first through-fault, determines a first set of state change rates associated with the first through-fault and a second set of state change rates associated with the first through-fault, assigns a first set of weights and a first set of criticality to the first set of state change rates, and determines a first mechanical fault based at least in part on the first set of state change rates, the first set of weights, and the first set of criticality. A power transformer system configured to calculate a change of state, assign a second set of weights and a second set of criticality to a second set of rate of change of state, calculate a first thermal change of state based at least partially on the second set of rate of change of state, the second set of weights and the second set of criticality, calculate a first cumulative change of state based at least partially on a first mechanical change of state and a first thermal change of state, and train a machine learning model using a first electrical stress, a first peak current, a first duration, a first set of rate of change of state, a second set of rate of change of state, a first mechanical change of state, a first thermal change of state and a first cumulative change of state.
[0048] 18. The power transformer system according to Clause 17, wherein the first set of phase change rates includes at least a first phase change rate relating to dissolved gas analysis, a second phase change rate relating to the rate of change, and a third phase change rate relating to the gas ratio.
[0049] 19. A power transformer system according to Clause 17 or 18, wherein the second set of state change rates includes at least a fourth state change rate related to winding hot spot temperature or thermal output, a fifth state change rate related to aging degradation factors, and a sixth state change rate related to loss of life.
[0050] 20. A power transformer system according to any one of clauses 17 to 19, wherein the determination of a first set of state change rates related to a first through-fault and a second set of state change rates related to a first through-fault further comprises identifying a set of mechanical measurements related to the power transformer system and a set of thermal measurements related to the power transformer system, determining the respective conditions related to each mechanical measurement in the set of mechanical measurements and each thermal measurement in the set of thermal measurements, and determining the first set of state change rates and the second set of state change rates based at least in part on the respective conditions related to each mechanical measurement in the set of mechanical measurements and each thermal measurement in the set of thermal measurements. [Explanation of symbols]
[0051] 100 flowcharts 200 flowcharts 300 flowcharts 308 Machine Learning Models 400 flowcharts
Claims
1. A method for monitoring through-fault current, To detect the first occurrence of a first through-fault in a power transformer system (102), Calculating the first electrical stress associated with the first through-fault, the first peak current associated with the first through-fault, and the first duration of the first through-fault (104), Determining a first set of state change ratios related to the first through-fault and a second set of state change ratios related to the first through-fault (106A, 106B), Assigning a first set of weights and a first set of criticality to the first set of state change rates (108A), Calculating a first mechanical state change (110A) based at least partially on the first set of state change rates, the first set of weights, and the first set of criticality, Assigning a second set of weights and a second set of criticality to the second set of state change rates (108B), Calculating the first thermal state change based at least partially on the second set of state change rates, the second set of weights, and the second set of criticality (110B), Calculating a first cumulative state change based at least partially on the first mechanical state change and the first thermal state change (112), Training a machine learning model (308) using the first electrical stress, the first peak current, the first duration, the first set of the state change rates, the second set of the state change rates, the first mechanical state change, the first thermal state change, and the first cumulative state change (114) Methods that include...
2. The method according to claim 1, wherein the first set of state change ratios includes at least a first state change ratio related to dissolved gas analysis, a second state change ratio related to the rate of change, and a third state change ratio related to the gas ratio.
3. The method according to claim 2, wherein the second set of state change rates includes at least a fourth state change rate related to winding hot spot temperature or thermal output, a fifth state change rate related to aging degradation factors, and a sixth state change rate related to loss of life.
4. The method according to claim 1, wherein the machine learning model (308) is trained when the power transformer system is in a healthy mode.
5. Determining a first set of state change ratios related to the first through-fault and a second set of state change ratios related to the first through-fault is Identifying a set of mechanical measurements related to the power transformer system and a set of thermal measurements related to the power transformer system, To determine the respective conditions related to each mechanical measurement in the set of mechanical measurement values and each thermal measurement in the set of thermal measurement values, Determining the first set of state change rates and the second set of state change rates based at least partially on the respective conditions related to each of the mechanical measurement values in the set of mechanical measurement values and each of the thermal measurement values in the set of thermal measurement values. The method according to claim 1, further comprising:
6. To detect the second occurrence of a second through-fault in the power transformer system, Calculating the second electrical stress associated with the second through-fault, the second peak current associated with the second through-fault, and the second duration of the second through-fault, To determine a third set of state change rates related to the second through-fault and a fourth set of state change rates related to the second through-fault, Based at least partially on the third set of state change ratios and the fourth set of state change ratios, the second mechanical state change related to the second through-fault, the second thermal state change related to the second through-fault, and the second cumulative state change due to the second through-fault are calculated. Using the machine learning model (308), estimate the third mechanical state change associated with the second through-fault, the third thermal state change associated with the second through-fault, and the third cumulative state change due to the second through-fault. The state of the power transformer system is determined at least partially based on the second mechanical state change, the second thermal state change, the second cumulative state change, the third mechanical state change, the third thermal state change, and the third cumulative state change. The method according to claim 1, further comprising:
7. The state of the power transformer system is determined at least partially based on the second mechanical state change, the second thermal state change, the second cumulative state change, the third mechanical state change, the third thermal state change, and the third cumulative state change. The first difference between the second mechanical state change and the third mechanical state change is calculated, The second difference between the second thermal state change and the third thermal state change is calculated, Calculate the third difference between the second cumulative state change and the third cumulative state change, The state of the power transformer system is determined at least in part based on the first difference, the second difference, and the third difference. The method according to claim 6, further comprising:
8. Based at least partially on the state of the power transformer system, it is determined that the power transformer system is not prepared for the third occurrence of the third through-fault, Outputting a display to the operator The method according to claim 6, further comprising:
9. The method according to claim 1, wherein the first occurrence of the first through-fault is identified at least in part on a plurality of circuit breaker statuses associated with the power transformer system.
10. Power transformers and, Controller and A power transformer system comprising, the controller, To detect the first occurrence of a first through-fault in the power transformer system, To calculate the first electrical stress associated with the first through-fault, the first peak current associated with the first through-fault, and the first duration of the first through-fault, To determine a first set of state change rates related to the first through-fault and a second set of state change rates related to the first through-fault, Assigning a first set of weights and a first set of criticality to the first set of state change rates, The first mechanical state change is calculated based at least partially on the first set of state change rates, the first set of weights, and the first set of criticality. Assigning a second set of weights and a second set of criticality to the second set of state change rates, The first thermal state change is calculated based at least partially on the second set of state change rates, the second set of weights, and the second set of criticality. The first cumulative state change is calculated based at least partially on the first mechanical state change and the first thermal state change, Training a machine learning model (308) using the first electrical stress, the first peak current, the first duration, the first set of the state change rates, the second set of the state change rates, the first mechanical state change, the first thermal state change, and the first cumulative state change. A power transformer system configured to perform the following actions.
11. The power transformer system according to claim 10, wherein the first set of state change rates includes at least a first state change rate related to dissolved gas analysis, a second state change rate related to the rate of change, and a third state change rate related to the gas ratio.
12. The power transformer system according to claim 11, wherein the second set of state change rates includes at least a fourth state change rate related to winding hot spot temperature or thermal output, a fifth state change rate related to aging degradation factors, and a sixth state change rate related to loss of life.
13. The determination of the first set of state change rates related to the first through-fault and the second set of state change rates related to the first through-fault is Identifying a set of mechanical measurements related to the power transformer system and a set of thermal measurements related to the power transformer system, To determine the respective conditions related to each mechanical measurement in the set of mechanical measurement values and each thermal measurement in the set of thermal measurement values, Determining the first set of state change rates and the second set of state change rates based at least partially on the respective conditions related to each of the mechanical measurement values in the set of mechanical measurement values and each of the thermal measurement values in the set of thermal measurement values. The power transformer system according to claim 10, further comprising:
14. The power transformer system according to claim 10, wherein the machine learning model (308) is trained when the power transformer system is in a healthy mode.
15. The power transformer system according to claim 10, wherein the first occurrence of the first through-fault is identified at least in part on a plurality of circuit breaker statuses associated with the power transformer system.