A method and system for diagnosing mechanical status of a transformer

Abstract

The present invention relates to the technical field of transformers and discloses a method and system for diagnosing the mechanical state of a transformer. The method and system comprise: comprehensively analyzing the frequency domain amplitude vectors of a vibration signal, a current signal, and a voltage signal to obtain a first variation vector, a second variation vector, a third variation vector, and a fourth variation vector, which can accurately reflect key characteristic information of changes in the mechanical state of the transformer and effectively overcome the problem of low accuracy of existing vibration-based diagnostic methods. The method can quickly and accurately diagnose the mechanical state of the transformer and is suitable for rapid and accurate on-site analysis and judgment, greatly improving the efficiency and accuracy of the mechanical state diagnosis of the transformer and providing strong technical support for ensuring the safe and stable operation of the power system.

Description

Technical Field

[0001] The present invention relates to the technical field of transformers, and in particular to a transformer mechanical state diagnosis method and system. Background Art

[0002] In power systems, transformers are core equipment, and the stability of their mechanical condition plays a decisive role in the safe and reliable operation of the power grid. Accurate and timely diagnosis of transformer mechanical condition can detect potential faults in advance, avoiding equipment damage and power outages, which is of great significance for ensuring the stable operation of power systems.

[0003] Traditional transformer mechanical condition diagnosis methods, such as oil chromatography analysis and partial discharge detection, have a certain degree of accuracy in assessing transformer condition. However, these methods have significant limitations. Oil chromatography analysis requires regular sampling and testing of transformer oil, resulting in long testing cycles and an inability to reflect the transformer's operating status in real time. Partial discharge detection typically requires specialized equipment and complex procedures, resulting in high testing costs. Furthermore, some detection methods require the transformer to be powered off, which not only affects the grid's normal power supply but also increases the difficulty and cost of detection. Therefore, traditional methods struggle to meet the demand for rapid on-site diagnosis and are subject to numerous limitations in practical applications.

[0004] In recent years, vibration-based diagnosis of transformer mechanical conditions has gradually emerged. By collecting and analyzing vibration signals generated during transformer operation, this technology can effectively reflect changes in the transformer's internal mechanical and electrical conditions. Compared with traditional diagnostic methods, vibration-based diagnosis offers significant advantages such as non-invasiveness, real-time online operation, and low cost. It eliminates the need to shut down the transformer and does not disrupt normal operation. It can obtain real-time information on the transformer's operating status, providing a new approach for transformer condition monitoring and fault diagnosis.

[0005] Despite the numerous advantages of vibration-based transformer mechanical condition diagnosis, current applications still face challenges. Existing vibration-based diagnostic methods lack effective comprehensive analysis methods for processing vibration, current, and voltage signals, making it difficult to accurately extract key characteristic information reflecting changes in the transformer's mechanical condition. This results in inaccurate and unreliable diagnostic results, making them incapable of meeting the demand for rapid and accurate on-site diagnosis of transformer mechanical conditions. Summary of the Invention

[0006] Based on this, it is necessary to address the above problems and propose a transformer mechanical state diagnosis method and system, which can accurately reflect the key characteristic information of the transformer mechanical state changes, effectively overcoming the problem of low accuracy of the existing vibration-based diagnosis method. This method can quickly and accurately diagnose the mechanical state of the transformer, and is suitable for rapid and accurate on-site analysis and judgment, greatly improving the efficiency and accuracy of transformer mechanical state diagnosis, and providing strong technical support for ensuring the safe and stable operation of the power system.

[0007] To achieve the above objectives, the present invention provides, in a first aspect, a method for diagnosing a transformer mechanical state, the method comprising:

[0008] Determine the frequency domain vibration amplitude vector of the vibration signal, the frequency domain current amplitude vector of the current signal, and the frequency domain voltage amplitude vector of the voltage signal of the transformer to be tested in each time period;

[0009] Determine whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period, obtaining a first determination result, and determine whether the frequency-domain voltage amplitude vector of the i-th time period is the same as the frequency-domain voltage amplitude vector of the j-th time period, obtaining a second determination result;

[0010] If the first judgment result is the same, then determine the first change vector according to the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period, and determine the second change vector according to the frequency domain voltage amplitude vector of the i-th time period and the frequency domain voltage amplitude vector of the j-th time period; if the second judgment result is the same, then determine the third change vector according to the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period, and determine the fourth change vector according to the frequency domain current amplitude vector of the i-th time period and the frequency domain current amplitude vector of the j-th time period, where the initial value of i is 1 and the initial value of j is 1;

[0011] If j is less than the total number of time periods, set j=j+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period;

[0012] If j is equal to the total number of time periods, and i is less than the total number of time periods, then set i=i+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period;

[0013] If i is equal to the total number of time periods, then whether the mechanical state of the transformer to be tested has changed is diagnosed based on all first change vectors, all second change vectors, all third change vectors and all fourth change vectors.

[0014] Optionally, diagnosing whether the mechanical state of the transformer to be tested has changed based on all the first change vectors, all the second change vectors, all the third change vectors, and all the fourth change vectors includes:

[0015] Inputting all the second variation vectors and all the fourth variation vectors into a preset transfer parameter model to obtain a plurality of first predicted variation vectors and a plurality of second predicted variation vectors;

[0016] Based on all the first variation vectors and the corresponding multiple first predicted variation vectors, and all the third variation vectors and the corresponding multiple second predicted variation vectors, it is diagnosed whether the mechanical state of the transformer to be tested has changed.

[0017] Optionally, the preset transfer parameter model includes a first transfer parameter model between the vibration variation and the voltage variation and a second transfer parameter model between the vibration variation and the current variation, and inputting all the second variation vectors and all the fourth variation vectors into the preset transfer parameter model to obtain a plurality of first predicted variation vectors and a plurality of second predicted variation vectors includes:

[0018] inputting all second variation vectors into the first transfer parameter model in sequence to obtain a plurality of first predicted variation vectors;

[0019] All fourth variation vectors are sequentially input into the second transfer parameter model to obtain a plurality of second predicted variation vectors.

[0020] Optionally, diagnosing whether the mechanical state of the transformer to be tested has changed based on all the first change vectors and the corresponding multiple first predicted change vectors, and all the third change vectors and the corresponding multiple second predicted change vectors, includes:

[0021] diagnosing whether a mechanical state of the core of the transformer to be tested has changed according to a difference rate between each first variation vector and a corresponding first predicted variation vector;

[0022] Whether the mechanical state of the winding of the transformer to be tested has changed is diagnosed based on the difference rate between each third variation vector and the corresponding second predicted variation vector.

[0023] Optionally, diagnosing whether the mechanical state of the core of the transformer to be tested has changed according to the difference rate between each first variation vector and the corresponding first predicted variation vector includes:

[0024] Determining a sum of first difference rates between each element in each first variation vector and each element in the corresponding first predicted variation vector, and a first difference rate between each element in each first variation vector and a corresponding element in the corresponding first predicted variation vector;

[0025] Among the sum of all first difference rates and all first difference rates, if there is a sum of first difference rates whose sum of difference rates is greater than a first preset percentage, or if there is a first difference rate whose difference rate is greater than a second preset percentage, it is diagnosed that the mechanical state of the core of the transformer to be tested has changed; otherwise, it is diagnosed that the mechanical state of the core of the transformer to be tested has not changed.

[0026] Optionally, diagnosing whether the mechanical state of the winding of the transformer to be tested has changed according to the difference rate between each third variation vector and the corresponding second predicted variation vector includes:

[0027] Determining a sum of second difference rates between each element in each third variation vector and each element in the corresponding second predicted variation vector, and a second difference rate between each element in each third variation vector and the corresponding element in the corresponding second predicted variation vector;

[0028] Among all the second difference rate sums and all the second difference rates, if there is a first difference rate sum whose difference rate sum is greater than the third preset percentage, or a first difference rate whose difference rate is greater than the first preset percentage, it is diagnosed that the mechanical state of the winding of the transformer to be tested has changed; otherwise, it is diagnosed that the mechanical state of the winding of the transformer to be tested has not changed.

[0029] Optionally, determining the frequency-domain vibration amplitude vector of the vibration signal of the transformer to be tested in each time period includes:

[0030] Acquiring sensing signals of all vibration sensors, wherein all vibration sensors are evenly installed on the oil tank of the transformer to be tested;

[0031] The sensing signal of the nth vibration sensor is divided into time periods to obtain the vibration signals of each time period, where the initial value of n is 1;

[0032] The vibration signal of each time period is converted into time-frequency domain to obtain the frequency domain vibration amplitude vector of each time period;

[0033] The diagnosing whether the mechanical state of the transformer to be tested has changed based on all the first change vectors and the corresponding multiple first predicted change vectors, and all the third change vectors and the corresponding multiple second predicted change vectors, includes:

[0034] Diagnose whether a mechanical state of the transformer under test at an nth vibration sensor has changed based on all the first variation vectors and the corresponding plurality of first predicted variation vectors, and all the third variation vectors and the corresponding plurality of second predicted variation vectors;

[0035] Let n=n+1, and return to the step of dividing the sensing signal of the nth vibration sensor into time periods to obtain vibration signals of each time period, until n equals the total number of vibration sensors.

[0036] Optionally, determining the frequency-domain current amplitude vector of the current signal and the frequency-domain voltage amplitude vector of the voltage signal of the transformer to be tested in each time period includes:

[0037] Acquire a mutual inductance current signal of a current transformer and a mutual inductance voltage signal of a voltage transformer, wherein the current transformer and the voltage transformer are installed on a busbar on the high-voltage side of the transformer to be tested;

[0038] Dividing the mutual inductance current signal into time periods to obtain current signals in each time period, and dividing the mutual inductance voltage signal into time periods to obtain voltage signals in each time period;

[0039] The current signal of each time period is converted into the time-frequency domain to obtain the frequency-domain current amplitude vector of each time period, and the time-domain voltage signal of each time period is converted into the time-frequency domain to obtain the frequency-domain voltage amplitude vector of each time period.

[0040] Optionally, the method further includes:

[0041] When the transformer under test is put into operation for the first time or is operating normally, determining a normal frequency-domain vibration amplitude vector of a normal vibration signal, a normal frequency-domain current amplitude vector of a normal current signal, and a normal frequency-domain voltage amplitude vector of a normal voltage signal of the transformer under test in each time period;

[0042] Determining whether the normal frequency-domain current amplitude vector of the xth time period is the same as the normal frequency-domain current amplitude vector of the yth time period, to obtain a first normal determination result; and determining whether the normal frequency-domain voltage amplitude vector of the xth time period is the same as the normal frequency-domain voltage amplitude vector of the yth time period, to obtain a second normal determination result;

[0043] If the first normal judgment result is the same, then determine the first normal variation vector according to the normal frequency domain vibration amplitude vector of the x-th time period and the normal frequency domain vibration amplitude vector of the y-th time period, and determine the second normal variation vector according to the normal frequency domain voltage amplitude vector of the x-th time period and the normal frequency domain voltage amplitude vector of the y-th time period; if the second normal judgment result is the same, then determine the third normal variation vector according to the normal frequency domain vibration amplitude vector of the x-th time period and the normal frequency domain vibration amplitude vector of the y-th time period, and determine the fourth normal variation vector according to the normal frequency domain current amplitude vector of the x-th time period and the normal frequency domain current amplitude vector of the y-th time period, where the initial value of x is 1 and the initial value of y is 1;

[0044] If y is less than the total number of time periods, set y=y+1, and return to the step of determining whether the normal frequency domain current amplitude vector of the xth time period is the same as the normal frequency domain current amplitude vector of the yth time period;

[0045] If y is equal to the total number of time periods and x is less than the total number of time periods, set x=x+1, and return to the step of determining whether the normal frequency-domain current amplitude vector of the x-th time period is the same as the normal frequency-domain current amplitude vector of the y-th time period;

[0046] If x is equal to the total number of time periods, all first normal change vectors and all second normal change vectors are input into the first initial machine learning model for training to obtain the first transfer parameter model, and all third normal change vectors and all fourth change vectors are input into the second initial machine learning model for training to obtain the second transfer parameter model.

[0047] To achieve the above object, the present invention provides, in a second aspect, a transformer mechanical state diagnosis system, the system comprising a plurality of vibration sensors, a current transformer, a voltage transformer, and a processor;

[0048] A plurality of vibration sensors are evenly installed on the oil tank wall of the transformer to be tested, and the current transformer and the voltage transformer are installed on the busbar on the high-voltage side of the transformer to be tested;

[0049] The processor is configured to execute the method as described in any one of the first aspects.

[0050] To achieve the above-mentioned object, the present invention provides, in a third aspect, a transformer mechanical state diagnosis device, the device comprising:

[0051] A determination module, used to determine the frequency domain vibration amplitude vector of the vibration signal, the frequency domain current amplitude vector of the current signal, and the frequency domain voltage amplitude vector of the voltage signal of the transformer to be tested in each time period;

[0052] a judgment module, configured to judge whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period, to obtain a first judgment result, and to judge whether the frequency-domain voltage amplitude vector of the i-th time period is the same as the frequency-domain voltage amplitude vector of the j-th time period, to obtain a second judgment result;

[0053] an identical execution module, configured to determine, if the first judgment result is the same, a first variation vector based on the frequency-domain vibration amplitude vector of the i-th time period and the frequency-domain vibration amplitude vector of the j-th time period, and a second variation vector based on the frequency-domain voltage amplitude vector of the i-th time period and the frequency-domain voltage amplitude vector of the j-th time period; and, if the second judgment result is the same, determine a third variation vector based on the frequency-domain vibration amplitude vector of the i-th time period and the frequency-domain vibration amplitude vector of the j-th time period, and determine a fourth variation vector based on the frequency-domain current amplitude vector of the i-th time period and the frequency-domain current amplitude vector of the j-th time period, where an initial value of i is 1 and an initial value of j is 1;

[0054] a first return execution module, configured to, if j is less than the total number of time periods, set j=j+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period;

[0055] a second return execution module, configured to, if j is equal to the total number of time periods and i is less than the total number of time periods, set i=i+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period;

[0056] A diagnostic module is configured to diagnose whether the mechanical state of the transformer to be tested has changed based on all the first change vectors, all the second change vectors, all the third change vectors, and all the fourth change vectors if i is equal to the total number of time periods.

[0057] To achieve the above-mentioned objectives, the present invention provides, in a fourth aspect, a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the method as described in any one of the first aspects.

[0058] To achieve the above-mentioned objectives, the present invention provides a computer device in a fifth aspect, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the method as described in any one of the first aspects.

[0059] The embodiment of the present invention has the following beneficial effects: the above method determines the frequency domain vibration amplitude vector of the vibration signal of the transformer to be tested in each time period, the frequency domain current amplitude vector of the current signal and the frequency domain voltage amplitude vector of the voltage signal, and judges whether the frequency domain current amplitude vector of the i-th time period is the same as the frequency domain current amplitude vector of the j-th time period to obtain a first judgment result, and judges whether the frequency domain voltage amplitude vector of the i-th time period is the same as the frequency domain voltage amplitude vector of the j-th time period to obtain a second judgment result. If the first judgment result is the same, then according to the i-th time period, The first variation vector is determined based on the frequency domain vibration amplitude vector and the frequency domain vibration amplitude vector of the j-th time period, and the second variation vector is determined based on the frequency domain voltage amplitude vector of the i-th time period and the frequency domain voltage amplitude vector of the j-th time period. If the second judgment result is the same, the third variation vector is determined based on the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period, and the fourth variation vector is determined based on the frequency domain current amplitude vector of the i-th time period and the frequency domain current amplitude vector of the j-th time period. The initial value of i is 1, the initial value of j is 1, and if j is less than If the total number of time periods is less than the total number of time periods, then let j=j+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period. If j is equal to the total number of time periods and i is less than the total number of time periods, then let i=i+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period. If i is equal to the total number of time periods, then based on all the first variation vectors, all the second variation vectors, all the third variation vectors, and all the fourth variation vectors, diagnose whether the mechanical state of the transformer under test has changed; that is, by comprehensively analyzing the frequency-domain amplitude vectors of the vibration signal, the current signal, and the voltage signal, the first variation vector, the second variation vector, the third variation vector, and the fourth variation vector are obtained, which can accurately reflect the key characteristic information of the change in the mechanical state of the transformer, effectively overcome the problem of low accuracy of the existing vibration-based diagnosis method, and can quickly and accurately diagnose the mechanical state of the transformer, and is suitable for rapid and accurate on-site analysis and judgment, greatly improving the efficiency and accuracy of the transformer mechanical state diagnosis, and providing strong technical support for ensuring the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0061] in:

[0062] Figure 1 A schematic diagram of a transformer mechanical status diagnosis method according to an embodiment of the present application;

[0063] Figure 2 This is a schematic diagram of a transformer mechanical status diagnosis device according to an embodiment of the present application;

[0064] Figure 3 1 is a diagram of the internal structure of a computer device in some embodiments. DETAILED DESCRIPTION

[0065] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0066] In power systems, transformers are core equipment, and the stability of their mechanical condition plays a decisive role in the safe and reliable operation of the power grid. Accurate and timely diagnosis of transformer mechanical condition can detect potential faults in advance, avoiding equipment damage and power outages, which is of great significance for ensuring the stable operation of power systems.

[0067] Traditional transformer mechanical condition diagnosis methods, such as oil chromatography analysis and partial discharge detection, have a certain degree of accuracy in assessing transformer condition. However, these methods have significant limitations. Oil chromatography analysis requires regular sampling and testing of transformer oil, resulting in long testing cycles and an inability to reflect the transformer's operating status in real time. Partial discharge detection typically requires specialized equipment and complex procedures, resulting in high testing costs. Furthermore, some detection methods require the transformer to be powered off, which not only affects the grid's normal power supply but also increases the difficulty and cost of detection. Therefore, traditional methods struggle to meet the demand for rapid on-site diagnosis and are subject to numerous limitations in practical applications.

[0068] In recent years, vibration-based diagnosis of transformer mechanical conditions has gradually emerged. By collecting and analyzing vibration signals generated during transformer operation, this technology can effectively reflect changes in the transformer's internal mechanical and electrical conditions. Compared with traditional diagnostic methods, vibration-based diagnosis offers significant advantages such as non-invasiveness, real-time online operation, and low cost. It eliminates the need to shut down the transformer and does not disrupt normal operation. It can obtain real-time information on the transformer's operating status, providing a new approach for transformer condition monitoring and fault diagnosis.

[0069] Despite the numerous advantages of vibration-based transformer mechanical condition diagnosis, current applications still face challenges. Existing vibration-based diagnostic methods lack effective comprehensive analysis methods for processing vibration, current, and voltage signals, making it difficult to accurately extract key characteristic information reflecting changes in the transformer's mechanical condition. This results in inaccurate and unreliable diagnostic results, making them incapable of meeting the demand for rapid and accurate on-site diagnosis of transformer mechanical conditions.

[0070] In response to the above problems, the present application proposes a transformer mechanical state diagnosis method and system, which can accurately reflect the key characteristic information of the transformer mechanical state changes, effectively overcoming the problem of low accuracy of existing vibration-based diagnosis methods. This method can quickly and accurately diagnose the mechanical state of the transformer, and is suitable for rapid and accurate on-site analysis and judgment, greatly improving the efficiency and accuracy of transformer mechanical state diagnosis, and providing strong technical support for ensuring the safe and stable operation of the power system. The specific implementation principle will be described in detail in the following embodiments.

[0071] In a first aspect, the present application provides a method for diagnosing a transformer mechanical state.

[0072] See also Figure 1 , is a schematic diagram of a transformer mechanical state diagnosis method according to an embodiment of the present application, the method comprising:

[0073] Step 110: Determine the frequency-domain vibration amplitude vector of the vibration signal, the frequency-domain current amplitude vector of the current signal, and the frequency-domain voltage amplitude vector of the voltage signal of the transformer to be tested in each time period.

[0074] The transformer to be tested refers to a transformer that requires online diagnosis of its mechanical status.

[0075] It should be noted that since there is a correlation between the mechanical state changes of the transformer to be tested (such as winding deformation, loose core, etc.) and the vibration signal, current signal and voltage signal, this application diagnoses the mechanical state through the frequency domain vibration amplitude vector of the vibration signal, the frequency domain current amplitude vector of the current signal and the frequency domain voltage amplitude vector of the voltage signal.

[0076] It can be understood that for the mechanical state of the core of the transformer to be tested, its vibration signal is proportional to the square of the voltage signal, and for the mechanical state of the winding of the transformer to be tested, its vibration signal is proportional to the square of the current signal.

[0077] In some embodiments, a sensor can be installed on the oil tank wall of the transformer to be tested, and then the vibration signal of the transformer to be tested is collected by a vibration sensor, and finally the vibration signal is converted into a frequency domain vibration amplitude vector; similarly, the current signal and voltage signal of the transformer to be tested can be collected respectively by a current transformer and a voltage transformer, and converted respectively to obtain a frequency domain current amplitude vector and a frequency domain voltage amplitude vector.

[0078] Step 120: Determine whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period, and obtain a first judgment result; and determine whether the frequency-domain voltage amplitude vector of the i-th time period is the same as the frequency-domain voltage amplitude vector of the j-th time period, and obtain a second judgment result.

[0079] Step 130: If the first judgment result is the same, then determine the first change vector based on the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period, and determine the second change vector based on the frequency domain voltage amplitude vector of the i-th time period and the frequency domain voltage amplitude vector of the j-th time period; if the second judgment result is the same, then determine the third change vector based on the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period, and determine the fourth change vector based on the frequency domain current amplitude vector of the i-th time period and the frequency domain current amplitude vector of the j-th time period, where the initial value of i is 1 and the initial value of j is 1.

[0080] Regarding the method for determining the change vector, in some embodiments, the difference between the corresponding elements of the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period can be used as the first change vector, and the difference between the corresponding elements of the frequency domain voltage amplitude vector of the i-th time period and the frequency domain voltage amplitude vector of the j-th time period can be used as the second change vector; similarly, the difference between the corresponding elements of the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period can be used as the third change vector, and the difference between the corresponding elements of the frequency domain current amplitude vector of the i-th time period and the frequency domain current amplitude vector of the j-th time period can be used as the fourth change vector.

[0081] Step 140: If j is less than the total number of time periods, set j=j+1 and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period.

[0082] It should be noted that, for the first judgment result and the second judgment result, if there are different situations, it is not necessary to execute the same situation of step 130, and step 140 can be executed directly; for example, if the first judgment result is different, there is no need to execute the situation where the first judgment result is the same, and if the second judgment result is different, there is no need to execute the situation where the second judgment result is the same, and step 140 can be executed directly.

[0083] Step 150: If j is equal to the total number of time periods and i is less than the total number of time periods, set i=i+1 and return to the step of determining whether the frequency domain current amplitude vector of the i-th time period is the same as the frequency domain current amplitude vector of the j-th time period.

[0084] The total number of time periods is the total number of time periods in step 110 .

[0085] Step 160: If i is equal to the total number of time periods, diagnose whether the mechanical state of the transformer under test has changed based on all first change vectors, all second change vectors, all third change vectors, and all fourth change vectors.

[0086] In some embodiments, whether the mechanical state of the transformer under test has changed can be diagnosed based on the comparison results of all the first change vectors, all the second change vectors, all the third change vectors and all the fourth change vectors with the corresponding preset change vectors; in other embodiments, the corresponding preset change vector can be obtained based on all the second change vectors and all the fourth change vectors, and then the mechanical state of the transformer under test has changed can be diagnosed based on the comparison results between all the first change vectors and all the third change vectors and the corresponding preset change vector; wherein the corresponding preset change vector can be the change vector obtained before the mechanical state of the transformer changes.

[0087] It should be noted that this method of the present application can not only evaluate whether the mechanical state of the winding and the mechanical state or position of the core have changed, but also evaluate whether the mechanical state or position of other mechanical components in the transformer under test have changed; for example, the pads, clamps, fasteners, etc. of the transformer under test.

[0088] In an embodiment of the present application, by comprehensively analyzing the frequency domain amplitude vectors of the vibration signal, current signal and voltage signal to obtain the first change vector, the second change vector, the third change vector and the fourth change vector, the key characteristic information of the change in the mechanical state of the transformer can be accurately reflected, effectively overcoming the problem of low accuracy of the existing vibration-based diagnostic method. This method can quickly and accurately diagnose the mechanical state of the transformer, is suitable for rapid and accurate on-site analysis and judgment, greatly improves the efficiency and accuracy of the transformer mechanical state diagnosis, and provides strong technical support for ensuring the safe and stable operation of the power system.

[0089] In addition, the transformer mechanical state diagnosis method proposed in this application, in addition to being able to accurately reflect the key characteristic information of the transformer mechanical state changes, effectively overcome the problem of low accuracy of existing vibration-based diagnosis methods, can quickly and accurately diagnose the transformer mechanical state, is suitable for rapid and accurate on-site analysis and judgment, improves the efficiency and accuracy of transformer mechanical state diagnosis, and provides strong technical support for ensuring the safe and stable operation of the power system, it also has the following advantages: Accurate maintenance plan formulation: Since this method can quickly and accurately diagnose the mechanical state of the transformer, power operation and maintenance personnel can formulate more accurate maintenance plans based on the diagnosis results. For example, if the diagnosis finds that certain mechanical components of the transformer (such as pads, clamps) If there are early signs of wear but it has not yet reached the level of serious failure, targeted maintenance can be arranged in advance, such as replacing worn pads and tightening clips, to prevent the failure from further deteriorating, reduce unplanned power outages, and improve the reliability of grid operation; Reasonable allocation of maintenance resources: By accurately assessing the mechanical condition of the transformer, it can be determined which transformers need priority maintenance and which can have their maintenance cycle appropriately extended. This helps to reasonably allocate maintenance resources, avoid unnecessary maintenance on transformers in good condition, reduce maintenance costs, and ensure that transformers with potential problems are dealt with in a timely manner; Reduce the degree of equipment damage: Detecting changes in the mechanical condition of the transformer in advance allows measures to be taken in the early stages of the failure to avoid further damage. For example, when an abnormal winding mechanical state is detected, timely measures can prevent further deformation of the winding, reduce the difficulty and cost of maintenance, and extend the service life of the transformer. Reduce power outage losses: Rapid and accurate diagnosis helps to arrange transformer repair or replacement in a timely manner, reducing the duration and scope of power outages caused by transformer failures. This is crucial for industrial production, commercial activities, and residents' lives, and can avoid economic losses and social impacts caused by power outages. Data accumulation and analysis: During the implementation of this method, a large amount of transformer mechanical state data will be accumulated. This data can provide a basis for intelligent analysis of the power grid. Through in-depth mining of this data, the changing patterns and influencing factors of the transformer mechanical state can be understood, providing a more scientific basis for the design, manufacture, and operation of the transformer. Integration with other systems: This diagnostic method can be integrated with other monitoring and control systems of the power grid (such as SCADA systems and online monitoring systems) to achieve data sharing and interaction, which helps to build a more intelligent power grid operation and management system and improve the overall operational efficiency and reliability of the power grid. Promote technological innovation: This method comprehensively analyzes vibration signals, current signals, and voltage signals, providing new ideas and methods for transformer mechanical state diagnosis. This will encourage more researchers and companies to invest in research in this field and promote the continuous innovation and development of transformer diagnostic technology.Improve diagnostic standards: With the widespread application and practical verification of this method, a large amount of diagnostic cases and data can be accumulated, providing a basis for the formulation of more scientific and complete transformer mechanical condition diagnostic standards, which will help improve the accuracy and consistency of transformer mechanical condition diagnosis throughout the industry.

[0090] In a feasible implementation, step 160 in the above embodiment diagnoses whether the mechanical state of the transformer to be tested has changed based on all the first change vectors, all the second change vectors, all the third change vectors and all the fourth change vectors, including: inputting all the second change vectors and all the fourth change vectors into a preset transfer parameter model to obtain multiple first predicted change vectors and multiple second predicted change vectors; diagnosing whether the mechanical state of the transformer to be tested has changed based on all the first change vectors and the corresponding multiple first predicted change vectors, as well as all the third change vectors and the corresponding multiple second predicted change vectors.

[0091] Among them, the preset transfer parameter model here refers to a machine learning model that has been trained and can be directly used to predict the output of the first predicted change vector and the second predicted change vector based on the input second change vector and the fourth change vector.

[0092] It should be noted that the preset transfer parameter model of the present application can be used to reflect the transfer parameter relationship between the vibration change and the current change and voltage change.

[0093] It should be further explained that the number of the second change vector, the first predicted change vector and the first change vector is equal, and the number of the fourth change vector, the second predicted change vector and the third change vector is also equal.

[0094] Regarding the training method of the preset transfer parameter model, in some embodiments, a large number of first change vectors, second change vectors, third change vectors and fourth change vectors can be obtained, and then all the first change vectors, second change vectors, third change vectors and fourth change vectors are input into the initial machine learning model for training. After training to a certain extent, a trained preset transfer parameter model can be obtained; wherein, the first change vector and the third change vector can be used as the true value of the initial machine learning model, that is, by comparing the true value with the first predicted change vector and the second predicted change vector output during the training process one by one, it can be determined whether the initial machine learning model is well trained and has met the expected requirements.

[0095] In the embodiment of the present application, by introducing a preset transfer parameter model, the second and fourth change vectors are used to predict the first and second predicted change vectors, and then the mechanical state of the transformer is diagnosed, thereby improving the intelligence and accuracy of the diagnosis.

[0096] It can be understood that intelligent diagnosis is improved: by presetting the transfer parameter model, the second change vector and the fourth change vector are used as input, and the first predicted change vector and the second predicted change vector are automatically output. This method avoids the complex manual analysis and judgment in the traditional method, realizes the intelligence of the diagnosis process, and greatly improves the diagnosis efficiency; the accuracy is improved: the model can reflect the transfer parameter relationship between the vibration change and the current change and voltage change, and use this relationship to make predictions, so that the predicted change vector is closer to the actual situation. In the diagnosis process, the first change vector is compared with the first predicted change vector, and the third change vector is compared with the second predicted change vector, which can more accurately detect the changes in the mechanical state of the transformer, and effectively improve the accuracy of the diagnosis; deep mining of data relationships: the preset transfer parameters The training process of the model is actually an in-depth exploration of the relationship between a large number of first change vectors, second change vectors, third change vectors and fourth change vectors. Through this exploration, the model can learn the potential rules behind the data, so that in actual diagnosis, it can more accurately predict the corresponding predicted change vector based on the input change vector, providing strong support for accurate diagnosis of the mechanical state of the transformer; Adaptability and flexibility are enhanced: once the preset transfer parameter model training is completed, it can be directly applied to the mechanical state diagnosis of different transformers. Regardless of the model and specifications of the transformer, as long as the corresponding second change vector and fourth change vector can be obtained, the model can be used for prediction and diagnosis, which enhances the adaptability and flexibility of the diagnostic method and enables it to be widely used in the mechanical state diagnosis of various transformers.

[0097] In a feasible implementation, the preset transfer parameter model in the above embodiment includes a first transfer parameter model between the vibration variation and the voltage variation and a second transfer parameter model between the vibration variation and the current variation.

[0098] In the above embodiment, all the second change vectors and all the fourth change vectors are input into the preset transfer parameter model to obtain multiple first predicted change vectors and multiple second predicted change vectors, including: inputting all the second change vectors into the first transfer parameter model in sequence to obtain multiple first predicted change vectors; inputting all the fourth change vectors into the second transfer parameter model in sequence to obtain multiple second predicted change vectors.

[0099] Among them, the first transfer parameter model here refers to a machine learning model that has been trained and can be directly used to predict and output the first predicted change vector based on the input second change vector; the second transfer parameter model here refers to a machine learning model that has been trained and can be directly used to predict and output the second predicted change vector based on the input fourth change vector.

[0100] It should be noted that the first transfer parameter model of the present application can be used to reflect the transfer parameter relationship between the vibration change and the voltage change; the second transfer parameter model of the present application can be used to reflect the transfer parameter relationship between the vibration change and the current change.

[0101] The training method of the first transfer parameter model and the second transfer parameter model is similar to the training method of the preset transfer parameter model. You can refer to the relevant content of the above-mentioned preset transfer parameter model training method, which will not be repeated here.

[0102] In the embodiment of the present application, by adopting a dual-model architecture that reflects the relationship between vibration and voltage, and vibration and current transmission parameters respectively, the accuracy and intelligence level of transformer mechanical state diagnosis are further improved.

[0103] It is understandable that the precise diagnosis is refined: the first transfer parameter model focuses on the relationship between the vibration change and the voltage change, and the second transfer parameter model focuses on the relationship between the vibration change and the current change. This subdivision enables the model to more accurately capture the intrinsic connection between different physical quantities. During diagnosis, the predicted change vector output by each model is compared with the actual change vector, which can more keenly detect subtle changes in the mechanical state of the transformer, greatly improving the accuracy of diagnosis; the degree of intelligence is deepened: under the dual-model architecture, the second change vector is input into the first transfer parameter model to obtain the first predicted change vector, and the fourth change vector is input into the second transfer parameter model to obtain the second predicted change vector. The entire process is completed automatically without the need for complex manual analysis and judgment. This intelligent processing method not only improves the diagnostic efficiency, but also reduces the errors caused by human factors, making the diagnostic process More scientific and reliable; data relationship mining is more accurate: the two models are trained separately, and their corresponding data relationships are deeply mined. The first transfer parameter model mines the potential laws behind the vibration and voltage data, and the second transfer parameter model mines the potential laws behind the vibration and current data. This enables the model to more accurately predict the corresponding predicted change vector based on the input change vector in actual diagnosis, providing stronger support for accurate diagnosis of the mechanical state of the transformer; adaptability and flexibility expansion: the dual-model architecture enhances the adaptability and flexibility of the diagnostic method. Transformers of different models and specifications can be input into the corresponding models for prediction and diagnosis as long as the corresponding second change vector and fourth change vector can be obtained. This flexibility enables this method to be widely used in the mechanical state diagnosis of various transformers without being restricted by individual differences of transformers.

[0104] In a feasible implementation, the above embodiment diagnoses whether the mechanical state of the transformer to be tested has changed based on all the first change vectors and the corresponding multiple first predicted change vectors, as well as all the third change vectors and the corresponding multiple second predicted change vectors, including: diagnosing whether the mechanical state of the core of the transformer to be tested has changed based on the difference rate between each first change vector and the corresponding first predicted change vector; and diagnosing whether the mechanical state of the winding of the transformer to be tested has changed based on the difference rate between each third change vector and the corresponding second predicted change vector.

[0105] It should be noted that, since the vibration signal of the transformer core under test is proportional to the square of the voltage signal, the transfer parameter relationship between the vibration change and the voltage change can be used to diagnose the mechanical state of the transformer core under test; since the vibration signal of the transformer winding under test is proportional to the square of the current signal, the transfer parameter relationship between the vibration change and the current change can be used to diagnose the mechanical state of the transformer winding under test.

[0106] In the embodiment of the present application, by calculating the difference rates separately, it is possible to accurately and separately diagnose whether the core mechanical state and winding mechanical state of the transformer to be tested have changed, providing a strong basis for targeted maintenance of the transformer.

[0107] It can be understood that accurate diagnosis of the mechanical state of the core: since the mechanical state of the core of the transformer to be tested has a specific relationship with the vibration signal and the voltage signal, that is, the vibration signal is proportional to the square of the voltage signal, so by utilizing the transfer parameter relationship between the vibration change and the voltage change, by calculating the difference rate between each first change vector and the corresponding first predicted change vector, it is possible to accurately diagnose whether the mechanical state of the core of the transformer to be tested has changed. This diagnostic method based on a specific physical relationship makes the judgment of the mechanical state of the core more accurate and reliable, and helps to timely discover potential problems of the core; Accurate diagnosis of the mechanical state of the winding: For the mechanical state of the winding of the transformer to be tested, it has a similar relationship with the vibration signal and the current signal, and the vibration signal is proportional to the square of the current signal. Therefore, with the help of the transfer parameter relationship between the vibration change and the current change, By calculating the difference rate between each third variation vector and the corresponding second predicted variation vector, it is possible to accurately diagnose whether the mechanical state of the winding of the transformer under test has changed. This targeted diagnostic method can accurately capture subtle changes in the mechanical state of the winding and provide timely and effective information for the maintenance and upkeep of the winding; targeted maintenance support: By accurately diagnosing the mechanical state of the core and winding respectively, power operation and maintenance personnel can formulate more targeted maintenance plans based on the diagnostic results. For example, if the diagnosis finds that the mechanical state of the core is abnormal, further inspection and maintenance can be carried out on the core. If there is a problem with the mechanical state of the winding, corresponding measures can be taken for the winding in a timely manner. This targeted maintenance method avoids blind maintenance, improves maintenance efficiency, and reduces maintenance costs. It also helps to extend the service life of the transformer and ensure the safe and stable operation of the power system.

[0108] In a feasible implementation, the above embodiment diagnoses whether the core mechanical state of the transformer to be tested has changed based on the difference rate between each first change vector and the corresponding first predicted change vector, including: determining the sum of the first difference rates between each element in each first change vector and each element in the corresponding first predicted change vector, and the first difference rate between each element in each first change vector and the corresponding element in the corresponding first predicted change vector; among all the first difference rate sums and all the first difference rates, if there is a first difference rate sum whose sum of difference rates is greater than a first preset percentage, or there is a first difference rate whose difference rate is greater than a second preset percentage, then it is diagnosed that the core mechanical state of the transformer to be tested has changed; otherwise, it is diagnosed that the core mechanical state of the transformer to be tested has not changed.

[0109] The first preset percentage and the second preset percentage can be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics. Of course, they can also be set by the operator according to actual needs.

[0110] In some embodiments, the present application may set the first preset percentage to 10% and the second preset percentage to 5%.

[0111] For the first difference rate sum, in some embodiments, the formula can be used to determine the first difference rate. Determine the first difference rate sum and the first difference rate; wherein, is the sum of the first difference rates between each element in the d-th first variation vector and each element in the d-th first predicted variation vector, is the total number of elements in the variation vector, is the first difference rate between the mth element in the dth first variation vector and the mth element in the dth first predicted variation vector, is the mth element in the dth first variation vector, is the mth element in the dth first predicted change vector, is the reference value, which can be or .

[0112] In the embodiment of the present application, the mechanical state of the core is accurately diagnosed by setting a difference rate threshold, providing a reliable basis for targeted maintenance and ensuring stable operation of the transformer.

[0113] It can be understood that the core state can be accurately judged: by determining the sum of the first difference rates between each first change vector and the corresponding first predicted change vector and the first difference rate of each element, and setting the first preset percentage and the second preset percentage as the judgment threshold, it is possible to accurately judge whether the mechanical state of the core of the transformer under test has changed. This judgment method based on the specific numerical difference rate avoids fuzzy judgment and makes the diagnosis result more accurate and reliable; targeted maintenance support: once the mechanical state of the core is diagnosed to have changed, the power operation and maintenance personnel can quickly conduct further inspection and maintenance on the core. This targeted maintenance method avoids blindly maintaining the entire transformer, improves maintenance efficiency, and reduces maintenance costs. It also helps to promptly discover potential problems with the core, prevent the problem from further deteriorating, and extend the service life of the transformer; adapt to different situations: setting different preset percentage thresholds can adapt to the core mechanical state diagnosis needs of different transformers and different operating environments. Whether it is a slight state change or a more serious fault sign, it can be accurately judged through this method, providing a strong guarantee for the stable operation of the transformer.

[0114] In a feasible implementation, the above embodiment diagnoses whether the mechanical state of the winding of the transformer to be tested has changed based on the difference rate between each third change vector and the corresponding second predicted change vector, including: determining the sum of the second difference rates between each element in each third change vector and each element in the corresponding second predicted change vector, and the second difference rate between each element in each third change vector and the corresponding element in the corresponding second predicted change vector; among all the sums of the second difference rates and all the second difference rates, if there is a sum of the first difference rates whose sum of the difference rates is greater than the third preset percentage, or a first difference rate whose difference rate is greater than the first preset percentage, then it is diagnosed that the mechanical state of the winding of the transformer to be tested has changed; otherwise, it is diagnosed that the mechanical state of the winding of the transformer to be tested has not changed.

[0115] The third preset percentage may be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics. Of course, it may also be set by the operator based on actual needs.

[0116] In some embodiments, the present application may set the third preset percentage to 20%.

[0117] For the second sum of difference rates, the method for determining the second difference rate is similar to the method for determining the first sum of difference rates using the first difference rate. You can refer to the above-mentioned content on the method for determining the second sum of difference rates using the second difference rate, and will not repeat it here.

[0118] In the embodiment of the present application, the mechanical state of the winding is accurately diagnosed by setting a difference rate threshold, providing a reliable basis for targeted maintenance and ensuring stable operation of the transformer.

[0119] It can be understood that the winding state is accurately judged: by determining the sum of the second difference rates between each third variation vector and the corresponding second predicted variation vector and the second difference rate of each element, and setting the third preset percentage and the first preset percentage as the judgment threshold, it is possible to accurately judge whether the mechanical state of the winding of the transformer under test has changed. This judgment method based on the specific numerical difference rate avoids fuzzy judgment and makes the diagnosis result more accurate and reliable; targeted maintenance support: once the mechanical state of the winding is diagnosed to have changed, the power operation and maintenance personnel can quickly conduct further inspection and maintenance on the winding. This targeted maintenance method avoids blindly maintaining the entire transformer, improves maintenance efficiency, and reduces maintenance costs. It also helps to promptly discover potential problems with the winding, prevent the problem from further deteriorating, and extend the service life of the transformer; adapting to different situations: setting different preset percentage thresholds can adapt to the winding mechanical state diagnosis needs of different transformers and different operating environments. Whether it is a slight state change or a more serious fault sign, it can be accurately judged through this method, providing a strong guarantee for the stable operation of the transformer.

[0120] In a feasible implementation, step 110 in the above embodiment, determining the frequency domain vibration amplitude vector of the vibration signal of the transformer to be tested in each time period, includes: obtaining the sensing signals of all vibration sensors, wherein all vibration sensors are evenly installed on the oil tank of the transformer to be tested; dividing the sensing signal of the nth vibration sensor into time periods to obtain vibration signals of each time period, where the initial value of n is 1; performing time-frequency domain conversion on the vibration signal of each time period to obtain the frequency domain vibration amplitude vector of each time period.

[0121] The above embodiment diagnoses whether the mechanical state of the transformer to be tested has changed based on all the first change vectors and the corresponding multiple first predicted change vectors, as well as all the third change vectors and the corresponding multiple second predicted change vectors, including: diagnosing whether the mechanical state of the transformer to be tested has changed at the nth vibration sensor based on all the first change vectors and the corresponding multiple first predicted change vectors, as well as all the third change vectors and the corresponding multiple second predicted change vectors; setting n=n+1, returning to execute the step of dividing the sensing signal of the nth vibration sensor into time periods to obtain vibration signals of each time period, until n is equal to the total number of vibration sensors.

[0122] In some embodiments, the present application generally sets the number of vibration sensors to 6-9, that is, the total number of vibration sensors is 6-9.

[0123] In the embodiment of the present application, comprehensive and accurate status monitoring of the transformer is achieved through the layout of multiple vibration sensors and cyclic iterative diagnosis, providing all-round protection for the stable operation of the transformer.

[0124] It can be understood that comprehensive monitoring of the transformer status: by evenly installing multiple vibration sensors on the oil tank of the transformer to be tested, all key parts of the transformer can be fully covered. This layout method can ensure that the vibration signals of the transformer are collected from multiple angles and positions, thereby more comprehensively reflecting the mechanical state of the transformer. The cyclic iterative diagnosis method divides the sensing signal of each vibration sensor into time periods and converts it into time-frequency domain to obtain the frequency domain vibration amplitude vector, and further diagnoses the mechanical state of the transformer at different positions through difference rate analysis, thereby realizing comprehensive monitoring of the transformer status; Accurate diagnosis of different position states: The vibration signal collected by each vibration sensor represents the mechanical state of a specific position of the transformer. By analyzing and diagnosing the signal collected by each vibration sensor separately, it is possible to accurately determine whether the mechanical state of the transformer at different positions has changed. This ability to accurately diagnose the state of different positions helps to timely discover potential problems in various parts of the transformer and provide a strong basis for targeted maintenance; Improve diagnostic reliability: Due to the use of multiple vibration sensor layouts and cyclic iterative diagnosis methods, it is possible to diagnose the mechanical state of the transformer from multiple angles and positions. Even if the signal of a vibration sensor is interfered with or abnormal, it can be compensated and verified by the signals of other sensors to ensure the accuracy of the diagnostic results; Adapt to the needs of large transformers: For larger transformers, their structures are complex and mechanical state changes may involve multiple parts. The layout of multiple vibration sensors and cyclic iterative diagnosis methods can adapt to the needs of such complex structures and conduct comprehensive and accurate monitoring and diagnosis of transformers. This method helps to timely discover potential problems in various parts of large transformers and prevent damage to the entire transformer or power outages caused by local faults; Optimize maintenance strategies: Based on comprehensive and accurate diagnostic results, power operation and maintenance personnel can formulate more optimized maintenance strategies. For example, for parts diagnosed with potential problems, maintenance and inspection can be given priority, while for parts in good condition, the maintenance cycle can be appropriately extended. This optimized maintenance strategy helps to rationally allocate maintenance resources, improve maintenance efficiency, reduce maintenance costs, and ensure the stable operation of the transformer.

[0125] In a feasible implementation, step 110 in the above embodiment, determining the frequency domain current amplitude vector of the current signal and the frequency domain voltage amplitude vector of the voltage signal of the transformer to be tested in each time period, includes: obtaining the mutual inductance current signal of the current transformer and the mutual inductance voltage signal of the voltage transformer, wherein the current transformer and the voltage transformer have been installed on the busbar on the high voltage side of the transformer to be tested; dividing the mutual inductance current signal into time periods to obtain the current signal of each time period, and dividing the mutual inductance voltage signal into time periods to obtain the voltage signal of each time period; performing time-frequency domain conversion on the current signal of each time period to obtain the frequency domain current amplitude vector of each time period, and performing time-frequency domain conversion on the time domain voltage signal of each time period to obtain the frequency domain voltage amplitude vector of each time period.

[0126] In the embodiment of the present application, a current and voltage transformer is installed on the high-voltage side bus to obtain signals, thereby achieving comprehensive and accurate current and voltage monitoring, and further improving the accuracy and reliability of the transformer mechanical status diagnosis.

[0127] It can be understood that comprehensive current and voltage monitoring: installing current transformers and voltage transformers on the high-voltage side busbar of the transformer to be tested can accurately obtain the current and voltage signals of the transformer. This layout ensures the comprehensiveness and accuracy of the current and voltage signals, and provides a solid foundation for subsequent signal analysis and status diagnosis. By dividing the time period and converting the mutual inductance current signal and mutual inductance voltage signal into time-frequency domain, the frequency domain current amplitude vector and frequency domain voltage amplitude vector of each time period can be obtained, providing comprehensive current and voltage data support for the mechanical status diagnosis of the transformer; improving diagnostic accuracy: since the current and voltage signals are closely related to the mechanical status changes of the transformer, by accurately obtaining these signals and conducting in-depth analysis, the mechanical status changes of the transformer can be more accurately reflected. In the diagnostic process, the correlation between these signals and the vibration signal can be used to More accurately determine whether the transformer has mechanical abnormalities, thereby improving the accuracy of diagnosis; adapt to complex environments: installing current transformers and voltage transformers on the high-voltage side bus can adapt to complex and changeable power grid environments. Regardless of how the power grid load changes, the transformer's current and voltage signals can be accurately obtained, providing stable and reliable data support for the transformer's mechanical status diagnosis; optimize system architecture: installing current transformers and voltage transformers on the high-voltage side bus can optimize the architecture of the entire diagnostic system. This layout not only facilitates the installation and maintenance of sensors, but also reduces interference and loss during signal transmission, improves signal quality and stability, and is also conducive to integration with other monitoring and control systems (such as SCADA systems, online monitoring systems, etc.) to achieve data sharing and interaction, and build a more intelligent power grid operation and management system.

[0128] In a feasible implementation, the method in the above embodiment also includes: when the transformer to be tested is put into operation for the first time or operates normally, determining the normal frequency domain vibration amplitude vector of the normal vibration signal of the transformer to be tested, the normal frequency domain current amplitude vector of the normal current signal, and the normal frequency domain voltage amplitude vector of the normal voltage signal in each time period; judging whether the normal frequency domain current amplitude vector of the xth time period is the same as the normal frequency domain current amplitude vector of the yth time period, and obtaining a first normal judgment result, and judging whether the normal frequency domain voltage amplitude vector of the xth time period is the same as the normal frequency domain voltage amplitude vector of the yth time period, and obtaining a second normal judgment result; if the first normal judgment results are the same, determining a first normal variation vector according to the normal frequency domain vibration amplitude vector of the xth time period and the normal frequency domain vibration amplitude vector of the yth time period, and determining a second normal variation vector according to the normal frequency domain voltage amplitude vector of the xth time period and the normal frequency domain voltage amplitude vector of the yth time period; if the second normal judgment results are the same, The third normal variation vector is determined based on the normal frequency domain vibration amplitude vector of the xth time period and the normal frequency domain current amplitude vector of the yth time period, and the fourth normal variation vector is determined based on the normal frequency domain current amplitude vector of the xth time period and the normal frequency domain current amplitude vector of the yth time period, where the initial value of x is 1 and the initial value of y is 1; if y is less than the total number of time periods, let y=y+1, and return to the step of determining whether the normal frequency domain current amplitude vector of the xth time period is the same as the normal frequency domain current amplitude vector of the yth time period; if y is equal to the total number of time periods, and x is less than the total number of time periods, let x=x+1, and return to the step of determining whether the normal frequency domain current amplitude vector of the xth time period is the same as the normal frequency domain current amplitude vector of the yth time period; if x is equal to the total number of time periods, all the first normal variation vectors and all the second normal variation vectors are input into the first initial machine learning model for training to obtain a first transfer parameter model, and all the third normal variation vectors and all the fourth variation vectors are input into the second initial machine learning model for training to obtain a second transfer parameter model.

[0129] In the embodiment of the present application, by training the model based on normal data, the accuracy and adaptability of the diagnostic model are improved, providing a more accurate and reliable basis for the diagnosis of the mechanical state of the transformer.

[0130] It can be understood that to improve the accuracy of the diagnostic model: when the transformer to be tested is put into operation for the first time or operates normally, the normal vibration signals, current signals and voltage signals of each time period are collected and converted into corresponding frequency domain vectors. By comparing whether the frequency domain vectors of different time periods are the same, the normal change vector is determined. Subsequently, these normal change vectors are input into the initial machine learning model for training to obtain the first transfer parameter model and the second transfer parameter model. This model training method based on normal data can ensure that the model learns the data characteristics of the transformer under normal operating conditions, so that in the subsequent diagnosis process, abnormal conditions can be more accurately identified to improve the accuracy of diagnosis; Enhance model adaptability: Since the normal conditions of different transformers or the same transformer in different operating environments may be different, through model training based on actual operating data, the model can better adapt to these differences. This adaptability not only improves the application effect of the model on different transformers, but also enables the model to more accurately reflect the transformer under a specific operating environment. Mechanical state changes; Reduce human intervention: Traditional model training methods may require a lot of expert experience and manual adjustments, while automatic training methods based on normal data greatly reduce human intervention. By automatically collecting and analyzing normal data, the model can automatically learn the laws behind the data, thereby improving training efficiency and accuracy; Optimize the diagnosis process: Integrate the model training steps into the diagnosis process to make the entire diagnosis process more systematic and automated. Collect data and train the model when the transformer is first put into operation or operates normally, and then directly use the trained model for judgment in the subsequent diagnosis process, which simplifies the diagnosis process and improves the diagnosis efficiency; Improve system reliability: Through model training based on normal data, the transfer parameter model obtained can more accurately reflect the changing laws of the transformer's mechanical state. In actual applications, these models can more reliably identify potential faults of the transformer and provide strong guarantees for the stable operation of the power system. At the same time, since the model is trained based on actual data, its prediction results are more credible and practical.

[0131] In a second aspect, the present application provides a transformer mechanical state diagnosis system, which includes a plurality of vibration sensors, a current transformer, a voltage transformer, and a processor (not shown).

[0132] In a feasible implementation, multiple vibration sensors are evenly installed on the oil tank wall of the transformer to be tested, and the current transformer and the voltage transformer are installed on the busbar on the high-voltage side of the transformer to be tested; the processor is used to execute any method as in the first aspect.

[0133] In the embodiment of the present application, an integrated diagnostic system architecture is used to achieve efficient and accurate diagnosis of the mechanical state of the transformer, providing strong support for the stable operation of the power system.

[0134] It can be understood that the integrated system architecture: the transformer mechanical state diagnosis system proposed in this application integrates multiple vibration sensors, current transformers, voltage transformers and processors to form a complete diagnostic system. This integrated system architecture enables the signal acquisition, processing and diagnosis process to be completed on a unified platform, thereby improving the efficiency and accuracy of the diagnosis; comprehensive signal acquisition: multiple vibration sensors are evenly installed on the oil tank wall of the transformer to be tested, which can fully cover all key parts of the transformer, ensuring that the vibration signal of the transformer is collected from multiple angles and positions. At the same time, the current transformer and voltage transformer are installed on the busbar on the high-voltage side, which can accurately obtain the current and voltage signals of the transformer. This comprehensive signal acquisition method provides a rich data basis for subsequent diagnosis; efficient signal processing: the processor is the core component of the system and is responsible for executing any diagnostic method as in the first aspect. It can quickly and accurately process vibration signals, current signals and voltage signals, and extract key feature information reflecting the change in the mechanical state of the transformer. By introducing a preset transfer parameter model, the processor can use the second and fourth change vectors to predict the first and second predicted change vectors. This enables intelligent diagnosis of the transformer's mechanical state; accurate diagnostic results: Since the system integrates comprehensive signal acquisition and efficient signal processing functions, it can produce more accurate diagnostic results. By comparing the difference rate between the actual change vector and the predicted change vector, it can accurately determine whether the mechanical state of the core and the mechanical state of the winding of the transformer under test have changed. This accurate diagnostic result provides a reliable decision-making basis for power operation and maintenance personnel, which helps to take targeted maintenance measures in a timely manner to ensure the stable operation of the power system; strong adaptability: The diagnostic system has good adaptability and can be applied to transformers of different models and specifications. As long as the corresponding vibration signals, current signals and voltage signals can be obtained, diagnosis can be performed through the system. This flexibility enables this method to be widely used in the mechanical state diagnosis of various transformers without being restricted by individual differences of transformers; easy integration and expansion: The diagnostic system is easy to integrate with other monitoring and control systems (such as SCADA systems, online monitoring systems, etc.) to achieve data sharing and interaction. At the same time, the system architecture is also convenient for subsequent expansion and upgrading to adapt to changing diagnostic needs and technological development.

[0135] In a third aspect, the present application provides a transformer mechanical status diagnosis device.

[0136] See also Figure 2 , is a schematic diagram of a transformer mechanical status diagnosis device according to an embodiment of the present application, wherein the device 210 includes:

[0137] A determination module 211 is used to determine the frequency domain vibration amplitude vector of the vibration signal, the frequency domain current amplitude vector of the current signal, and the frequency domain voltage amplitude vector of the voltage signal of the transformer to be tested in each time period;

[0138] a judgment module 212 for judging whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period, thereby obtaining a first judgment result, and judging whether the frequency-domain voltage amplitude vector of the i-th time period is the same as the frequency-domain voltage amplitude vector of the j-th time period, thereby obtaining a second judgment result;

[0139] The same execution module 213 is configured to determine, if the first judgment result is the same, a first variation vector based on the frequency-domain vibration amplitude vector of the i-th time period and the frequency-domain vibration amplitude vector of the j-th time period, and a second variation vector based on the frequency-domain voltage amplitude vector of the i-th time period and the frequency-domain voltage amplitude vector of the j-th time period; and if the second judgment result is the same, determine a third variation vector based on the frequency-domain vibration amplitude vector of the i-th time period and the frequency-domain vibration amplitude vector of the j-th time period, and determine a fourth variation vector based on the frequency-domain current amplitude vector of the i-th time period and the frequency-domain current amplitude vector of the j-th time period, where the initial value of i is 1 and the initial value of j is 1;

[0140] A first return execution module 214 is configured to, if j is less than the total number of time periods, set j=j+1 and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period;

[0141] The second return execution module 215 is configured to, if j is equal to the total number of time periods and i is less than the total number of time periods, set i=i+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period;

[0142] The diagnosis module 216 is configured to diagnose whether the mechanical state of the transformer under test has changed based on all the first change vectors, all the second change vectors, all the third change vectors, and all the fourth change vectors if i is equal to the total number of time periods.

[0143] In the embodiment of the present application, the relevant contents of the above-mentioned determination module 211, judgment module 212, same execution module 213, first return execution module 214, second return execution module 215 and diagnosis module 216 can be referred to. Figure 1 The contents of the illustrated embodiments are not described in detail here.

[0144] It should be noted that the device 210 of the present application also includes some other modules. It can be understood that the method of the present application and the device 210 have a one-to-one correspondence. Therefore, the other modules of the device 210 of the present application are the contents corresponding to the method of the present application in the above-mentioned embodiment.

[0145] In a fourth aspect, the present application further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes a transformer mechanical state diagnosis method in the above method embodiment.

[0146] In a fifth aspect, the present application further provides a computer device including a memory and a processor, wherein the memory stores a computer program. When the computer program is executed by the processor, the processor executes a transformer mechanical state diagnosis method in the above method embodiment.

[0147] Figure 3 The internal structure diagram of the computer device in some embodiments is shown. The computer device can be a terminal, a server, or a gateway. Figure 3 As shown, the computer device includes a processor, a memory, and a network interface connected via a system bus.

[0148] The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program. When the computer program is executed by the processor, the processor can implement the various steps in the above method embodiment. The internal memory may also store a computer program. When the computer program is executed by the processor, the processor can implement the various steps in the above method embodiment. It will be understood by those skilled in the art that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0149] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0150] Among them, any reference to memory, storage, database or other media used in the various embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0151] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0152] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A method for diagnosing the mechanical state of a transformer, characterized in that: The method comprises: Determine the frequency domain vibration amplitude vector of the vibration signal, the frequency domain current amplitude vector of the current signal, and the frequency domain voltage amplitude vector of the voltage signal of the transformer to be tested in each time period; Determine whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period, obtaining a first determination result, and determine whether the frequency-domain voltage amplitude vector of the i-th time period is the same as the frequency-domain voltage amplitude vector of the j-th time period, obtaining a second determination result; If the first judgment result is the same, then determine the first change vector according to the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period, and determine the second change vector according to the frequency domain voltage amplitude vector of the i-th time period and the frequency domain voltage amplitude vector of the j-th time period; if the second judgment result is the same, then determine the third change vector according to the frequency domain vibration amplitude vector of the i-th time period and the frequency domain vibration amplitude vector of the j-th time period, and determine the fourth change vector according to the frequency domain current amplitude vector of the i-th time period and the frequency domain current amplitude vector of the j-th time period, where the initial value of i is 1 and the initial value of j is 1; If j is less than the total number of time periods, set j=j+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period; If j is equal to the total number of time periods, and i is less than the total number of time periods, then set i=i+1, and return to the step of determining whether the frequency-domain current amplitude vector of the i-th time period is the same as the frequency-domain current amplitude vector of the j-th time period; If i is equal to the total number of time periods, diagnosing whether the mechanical state of the transformer to be tested has changed based on all the first change vectors, all the second change vectors, all the third change vectors, and all the fourth change vectors; The step of diagnosing whether the mechanical state of the transformer to be tested has changed based on all the first variation vectors, all the second variation vectors, all the third variation vectors, and all the fourth variation vectors includes: Inputting all the second variation vectors and all the fourth variation vectors into a preset transfer parameter model to obtain a plurality of first predicted variation vectors and a plurality of second predicted variation vectors; Based on all the first variation vectors and the corresponding multiple first predicted variation vectors, and all the third variation vectors and the corresponding multiple second predicted variation vectors, it is diagnosed whether the mechanical state of the transformer to be tested has changed.

2. The transformer mechanical status diagnosis method according to claim 1, characterized in that: The preset transfer parameter model includes a first transfer parameter model between the vibration variation and the voltage variation and a second transfer parameter model between the vibration variation and the current variation. Inputting all second variation vectors and all fourth variation vectors into the preset transfer parameter model to obtain a plurality of first predicted variation vectors and a plurality of second predicted variation vectors includes: inputting all second variation vectors into the first transfer parameter model in sequence to obtain a plurality of first predicted variation vectors; All fourth variation vectors are sequentially input into the second transfer parameter model to obtain a plurality of second predicted variation vectors.

3. The transformer mechanical status diagnosis method according to claim 1, characterized in that: The diagnosing whether the mechanical state of the transformer to be tested has changed based on all the first change vectors and the corresponding multiple first predicted change vectors, and all the third change vectors and the corresponding multiple second predicted change vectors, includes: diagnosing whether a mechanical state of the core of the transformer to be tested has changed according to a difference rate between each first variation vector and a corresponding first predicted variation vector; Whether the mechanical state of the winding of the transformer to be tested has changed is diagnosed based on the difference rate between each third variation vector and the corresponding second predicted variation vector.

4. The transformer mechanical status diagnosis method according to claim 3, characterized in that: The diagnosing whether the mechanical state of the core of the transformer to be tested has changed according to the difference rate between each first variation vector and the corresponding first predicted variation vector includes: Determining a sum of first difference rates between each element in each first variation vector and each element in the corresponding first predicted variation vector, and a first difference rate between each element in each first variation vector and a corresponding element in the corresponding first predicted variation vector; Among the sum of all first difference rates and all first difference rates, if there is a sum of first difference rates whose sum of difference rates is greater than a first preset percentage, or if there is a first difference rate whose difference rate is greater than a second preset percentage, it is diagnosed that the mechanical state of the core of the transformer to be tested has changed; otherwise, it is diagnosed that the mechanical state of the core of the transformer to be tested has not changed.

5. The transformer mechanical status diagnosis method according to claim 3, characterized in that: The diagnosing whether the mechanical state of the winding of the transformer to be tested has changed according to the difference rate between each third variation vector and the corresponding second predicted variation vector includes: Determining a sum of second difference rates between each element in each third variation vector and each element in the corresponding second predicted variation vector, and a second difference rate between each element in each third variation vector and the corresponding element in the corresponding second predicted variation vector; Among all the second difference rate sums and all the second difference rates, if there is a first difference rate sum whose difference rate sum is greater than the third preset percentage, or a first difference rate whose difference rate is greater than the first preset percentage, it is diagnosed that the mechanical state of the winding of the transformer to be tested has changed; otherwise, it is diagnosed that the mechanical state of the winding of the transformer to be tested has not changed.

6. The transformer mechanical status diagnosis method according to claim 1, characterized in that: The step of determining the frequency domain vibration amplitude vector of the vibration signal of the transformer to be tested in each time period includes: Acquiring sensing signals of all vibration sensors, wherein all vibration sensors are evenly installed on the oil tank of the transformer to be tested; The sensing signal of the nth vibration sensor is divided into time periods to obtain the vibration signals of each time period, where the initial value of n is 1; The vibration signal of each time period is converted into the time-frequency domain to obtain the frequency-domain vibration amplitude vector of each time period; The diagnosing whether the mechanical state of the transformer to be tested has changed based on all the first change vectors and the corresponding multiple first predicted change vectors, and all the third change vectors and the corresponding multiple second predicted change vectors, includes: Diagnose whether a mechanical state of the transformer under test at an nth vibration sensor has changed based on all the first variation vectors and the corresponding plurality of first predicted variation vectors, and all the third variation vectors and the corresponding plurality of second predicted variation vectors; Let n=n+1, and return to the step of dividing the sensing signal of the nth vibration sensor into time periods to obtain vibration signals of each time period, until n equals the total number of vibration sensors.

7. The transformer mechanical status diagnosis method according to claim 1, characterized in that: Determining the frequency domain current amplitude vector of the current signal and the frequency domain voltage amplitude vector of the voltage signal of the transformer to be tested in each time period includes: Acquire a mutual inductance current signal of a current transformer and a mutual inductance voltage signal of a voltage transformer, wherein the current transformer and the voltage transformer are installed on a busbar on the high-voltage side of the transformer to be tested; Dividing the mutual inductance current signal into time periods to obtain current signals in each time period, and dividing the mutual inductance voltage signal into time periods to obtain voltage signals in each time period; The current signal of each time period is converted into the time-frequency domain to obtain the frequency-domain current amplitude vector of each time period, and the time-domain voltage signal of each time period is converted into the time-frequency domain to obtain the frequency-domain voltage amplitude vector of each time period.

8. The transformer mechanical status diagnosis method according to claim 2, characterized in that: The method further comprises: When the transformer under test is put into operation for the first time or is operating normally, determining a normal frequency-domain vibration amplitude vector of a normal vibration signal, a normal frequency-domain current amplitude vector of a normal current signal, and a normal frequency-domain voltage amplitude vector of a normal voltage signal of the transformer under test in each time period; Determining whether the normal frequency-domain current amplitude vector of the xth time period is the same as the normal frequency-domain current amplitude vector of the yth time period, to obtain a first normal determination result; and determining whether the normal frequency-domain voltage amplitude vector of the xth time period is the same as the normal frequency-domain voltage amplitude vector of the yth time period, to obtain a second normal determination result; If the first normal judgment result is the same, then determine the first normal variation vector according to the normal frequency domain vibration amplitude vector of the x-th time period and the normal frequency domain vibration amplitude vector of the y-th time period, and determine the second normal variation vector according to the normal frequency domain voltage amplitude vector of the x-th time period and the normal frequency domain voltage amplitude vector of the y-th time period; if the second normal judgment result is the same, then determine the third normal variation vector according to the normal frequency domain vibration amplitude vector of the x-th time period and the normal frequency domain vibration amplitude vector of the y-th time period, and determine the fourth normal variation vector according to the normal frequency domain current amplitude vector of the x-th time period and the normal frequency domain current amplitude vector of the y-th time period, where the initial value of x is 1 and the initial value of y is 1; If y is less than the total number of time periods, set y=y+1, and return to the step of determining whether the normal frequency domain current amplitude vector of the xth time period is the same as the normal frequency domain current amplitude vector of the yth time period; If y is equal to the total number of time periods and x is less than the total number of time periods, set x=x+1, and return to the step of determining whether the normal frequency-domain current amplitude vector of the x-th time period is the same as the normal frequency-domain current amplitude vector of the y-th time period; If x is equal to the total number of time periods, all first normal change vectors and all second normal change vectors are input into the first initial machine learning model for training to obtain the first transfer parameter model, and all third normal change vectors and all fourth change vectors are input into the second initial machine learning model for training to obtain the second transfer parameter model.

9. A transformer mechanical status diagnosis system, characterized in that: The system includes a plurality of vibration sensors, a current transformer, a voltage transformer, and a processor; A plurality of vibration sensors are evenly installed on the oil tank wall of the transformer to be tested, and the current transformer and the voltage transformer are installed on the busbar on the high-voltage side of the transformer to be tested; The processor is configured to execute the transformer mechanical state diagnosis method according to any one of claims 1 to 8.

Images