Method for detecting displacement deformation of transformer bushing

Abstract

The invention discloses a transformer bushing displacement deformation detection method, and relates to the technical field of transformer bushings, and the method comprises the steps: obtaining the temperature distribution data and field intensity distribution data of a bushing through an infrared sensor and an electric field sensor; determining thermal stress data of the casing pipe according to the temperature distribution data; determining electric stress data and mechanical stress data of the casing pipe according to the field intensity distribution data; determining current stress distribution data of the casing pipe according to the thermal stress data, the electric stress data and the mechanical stress data; determining future stress distribution data of the casing pipe according to the current stress distribution data of the casing pipe; and determining the displacement deformation of the casing according to the future stress distribution data of the casing and the structural parameters of the casing. According to the invention, the structure state in the bushing is represented through the stress of the bushing based on the infrared sensor and the electric field sensor, so that early warning is carried out on the fault of the bushing through the displacement deformation of the transformer bushing, thereby reducing the cost and improving the early warning accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer bushings, and in particular to a method for detecting displacement deformation of transformer bushings. Background Art

[0002] During the long-term development of an oilfield, underground casing is subject to multiple factors, including complex stresses (such as geostress variations and formation creep), chemical corrosion (such as sulfur- and saline-containing formation water), and engineering operations (such as perforating and fracturing). This makes it susceptible to damage in various forms and mechanisms, including deformation, collapse, corrosion perforation, leakage, and fracture. These various types of failures can range from impacting the productivity of individual wells to causing safety accidents and environmental pollution, resulting in significant economic losses. Therefore, early detection and accurate early warning of various potential casing failures are crucial.

[0003] The core technical challenge lies in the extremely diverse potential failure modes of casing, with each failure (such as deformation and corrosion, local perforation, and overall fracture) exhibiting significantly different mechanisms and manifestations. To cover these diverse fault types, existing monitoring technologies typically require the integration and deployment of numerous sensors (such as specialized sensors for corrosion and acoustic sensors for leak or vibration detection) within the wellbore or on the casing. This "different sensors for different faults" approach leads to two serious problems: first, a dramatic increase in hardware costs, making it uneconomical, especially in large-scale oil well applications; and second, the sheer complexity and difficulty of implementation, making it extremely difficult to deploy and maintain such a diverse array of sensors in the harsh downhole environment. More critically, accurately, promptly, and effectively distinguishing fault types from this vast amount of sensor data from diverse sources and subject to significant noise interference, and triggering reliable warnings, presents a significant technical challenge. Consequently, actual warning accuracy is often suboptimal, with frequent false alarms and missed alerts.

[0004] Therefore, the existing solution that relies on deploying multiple sensors to deal with the diversity of casing faults has significant limitations in cost and diagnostic accuracy.

[0005] Based on this, how to perform casing failure warning accurately and at low cost has become a technical problem that needs to be solved urgently. Summary of the Invention

[0006] The technical problem solved by the present invention is that the fault early warning of transformer bushings in the prior art is high in cost and insufficient in accuracy.

[0007] To solve the above technical problems, the present invention provides the following technical solution: a method for detecting displacement deformation of a transformer bushing, the method comprising:

[0008] The temperature distribution data and the field intensity distribution data of the casing are obtained respectively by using an infrared sensor and an electric field sensor;

[0009] determining thermal stress data of the casing according to the temperature distribution data;

[0010] determining electrical stress data and mechanical stress data of the bushing according to the field intensity distribution data;

[0011] determining current stress distribution data of the bushing according to the thermal stress data, the electrical stress data, and the mechanical stress data;

[0012] determining future stress distribution data of the casing according to current stress distribution data of the casing;

[0013] The displacement deformation of the casing is determined according to the future stress distribution data of the casing and the structural parameters of the casing.

[0014] Preferably, the temperature distribution data includes temperature data of a plurality of sampling points on the casing; before determining the thermal stress data of the casing based on the temperature distribution data, the method further includes:

[0015] Acquiring a temperature data sequence of a plurality of sampling points on the casing;

[0016] Identifying a synchronous temperature rise moment of the plurality of sampling points on the casing according to the temperature data sequence of the plurality of sampling points, wherein the temperature change amplitude of more than a preset number of sampling points at the synchronous temperature rise moment is greater than a preset temperature change amplitude;

[0017] Determining a temperature mutation area based on the temperature data sequences of the plurality of sampling points;

[0018] Extracting a temperature data subsequence of the sampling point in the temperature mutation area at the synchronous temperature rise moment;

[0019] determining an ambient medium of an environment in which the casing is located based on a plurality of temperature data subsequences;

[0020] Determining the thermal stress data of the sleeve according to the temperature distribution data includes:

[0021] Thermal stress data of the sleeve is determined according to the ambient medium and the temperature distribution data.

[0022] Preferably, determining the thermal stress data of the casing according to the ambient medium and the temperature distribution data includes:

[0023] Searching for a thermal resistance coefficient corresponding to the ambient medium from preset medium thermal resistance comparison data;

[0024] Correcting a preset thermoelastic model according to the thermal resistance coefficient to obtain a thermoelastic model corresponding to the environmental medium;

[0025] Thermal stress data of the sleeve is determined according to the temperature distribution data and a thermoelastic model corresponding to the ambient medium.

[0026] Preferably, determining the electrical stress data and mechanical stress data of the bushing according to the field intensity distribution data includes:

[0027] extracting a low-frequency component from the field intensity distribution data by using a low-pass filter, wherein the low-frequency component is used to characterize the quasi-static electric field;

[0028] determining the electrical stress data according to the low-frequency component, wherein the electrical stress data is positively correlated to the square of the amplitude value of the low-frequency component;

[0029] Extracting a component whose signal variation period is less than a preset variation period as a high-frequency component, wherein the amplitude of the high-frequency component is used to characterize the vibration intensity;

[0030] Performing spectrum analysis on the high-frequency component to extract vibration characteristic frequency points;

[0031] Matching the fault mode from a preset fault vibration database according to the vibration characteristic frequency point;

[0032] If the similarity of the fault modes is greater than a preset similarity threshold, the mechanical stress data is determined according to the high-frequency component, and the mechanical stress data is positively correlated with a root mean square value of the high-frequency component.

[0033] Preferably, the direction of thermal stress at each point on the bushing is determined according to the temperature gradient of each point, the direction of electric stress at each point on the bushing is consistent with the direction of the electric field represented by the field intensity distribution data, and the direction of mechanical stress at each point on the bushing is determined according to the fault mode.

[0034] Preferably, determining the current stress distribution data of the bushing according to the thermal stress data, the electrical stress data and the mechanical stress data includes:

[0035] Determining a damping vector and a scaling factor corresponding to the ambient medium according to a preset damping comparison table;

[0036] The thermal stress data, the electrical stress data, and the mechanical stress data are coupled in the radial, axial, and tangential directions of the casing according to the damping vector and the scaling factor to obtain current stress distribution data of the casing.

[0037] Preferably, determining the future stress distribution data of the casing according to the current stress distribution data of the casing includes:

[0038] Acquiring historical stress distribution data of the casing;

[0039] Determine the stress time series of each point of the casing in the radial, axial and tangential directions of the casing from the historical stress distribution data and the current stress data;

[0040] The radial, axial and tangential stress time series of the casing are input into a pre-trained stress prediction model to obtain future stress distribution data of the casing.

[0041] Preferably, determining the displacement deformation of the casing according to the future stress distribution data of the casing and the structural parameters of the casing includes:

[0042] If the environmental medium is a heavy oil saturated zone, a point where the mid-tangential stress of the future stress distribution data is greater than a preset stress threshold is determined as a displacement deformation point;

[0043] If the environmental medium is cement rock, the point where the central axial change rate of the future stress distribution data is greater than a preset change threshold is determined as a displacement deformation point;

[0044] If the environmental medium is water-bearing sandstone, the point where the mid-radial to tangential stress ratio of the future stress distribution data is greater than a preset stress ratio threshold is determined as the displacement deformation point.

[0045] Preferably, the method further comprises:

[0046] It is determined whether the sleeve has a displacement deformation point according to the current stress distribution data of the sleeve.

[0047] Preferably, after determining the displacement deformation of the casing according to the future stress distribution data of the casing and the structural parameters of the casing, the method further comprises:

[0048] If the sleeve has the displacement deformation point, a fault warning is issued.

[0049] The beneficial effects of the present invention include obtaining temperature distribution data and field intensity distribution data of the bushing respectively through an infrared sensor and an electric field sensor, determining thermal stress data of the bushing according to the temperature distribution data, determining electric stress data and mechanical stress data of the bushing according to the field intensity distribution data, determining current stress distribution data of the bushing according to the thermal stress data, electric stress data and mechanical stress data, determining future stress distribution data of the bushing according to the current stress distribution data of the bushing, and determining displacement deformation of the bushing according to the future stress distribution data of the bushing and structural parameters of the bushing. The structural state inside the bushing is characterized by the stress of the bushing based on the infrared sensor and the electric field sensor, so that a warning of bushing failure is issued through the displacement deformation of the transformer bushing, thereby reducing costs and improving warning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1A schematic diagram of the basic flow of a method for detecting displacement deformation of a transformer bushing provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0052] Example 1, reference Figure 1 , as one embodiment of the present invention, provides a method for detecting displacement deformation of a transformer bushing, comprising:

[0053] S110, obtaining temperature distribution data and field intensity distribution data of the casing through an infrared sensor and an electric field sensor respectively;

[0054] S120, determining thermal stress data of the casing according to the temperature distribution data;

[0055] S130, determining electrical stress data and mechanical stress data of the bushing according to the field intensity distribution data;

[0056] S140, determining current stress distribution data of the casing according to the thermal stress data, the electrical stress data, and the mechanical stress data;

[0057] S150, determining future stress distribution data of the casing according to current stress distribution data of the casing;

[0058] S160: Determine the displacement deformation of the casing according to the future stress distribution data of the casing and the structural parameters of the casing.

[0059] Preferably, the temperature distribution data includes temperature data of multiple sampling points on the casing; before determining the thermal stress data of the casing based on the temperature distribution data, the method further includes: obtaining a temperature data sequence of multiple sampling points on the casing; identifying a synchronous temperature rise moment of multiple sampling points on the casing based on the temperature data sequence of the multiple sampling points, wherein the temperature change amplitude of more than a preset number of sampling points at the synchronous temperature rise moment is greater than the preset temperature change amplitude; determining a temperature mutation area based on the temperature data sequence of the multiple sampling points; extracting a temperature data subsequence of the sampling points in the temperature mutation area at the synchronous temperature rise moment; determining the environmental medium of the environment in which the casing is located based on the multiple temperature data subsequences; and determining the thermal stress data of the casing based on the temperature distribution data, including: determining the thermal stress data of the casing based on the environmental medium and the temperature distribution data.

[0060] Preferably, determining the thermal stress data of the sleeve based on the ambient medium and temperature distribution data includes: searching for a thermal resistance coefficient corresponding to the ambient medium from preset medium thermal resistance comparison data; correcting a preset thermoelastic model based on the thermal resistance coefficient to obtain a thermoelastic model corresponding to the ambient medium; and determining the thermal stress data of the sleeve based on the temperature distribution data and the thermoelastic model corresponding to the ambient medium.

[0061] Preferably, the electric stress data and mechanical stress data of the bushing are determined based on the field strength distribution data, including: extracting a low-frequency component from the field strength distribution data through a low-pass filter, the low-frequency component is used to characterize the quasi-static electric field; determining the electric stress data based on the low-frequency component, the electric stress data is positively correlated with the square of the amplitude value of the low-frequency component; extracting the component with a signal change period less than a preset change period as a high-frequency component, the amplitude of the high-frequency component is used to characterize the vibration intensity; performing spectral analysis on the high-frequency component to extract vibration characteristic frequency points; matching the fault mode from a preset fault vibration database based on the vibration characteristic frequency points; if the similarity of the fault mode is greater than a preset similarity threshold, determining the mechanical stress data based on the high-frequency component, the mechanical stress data is positively correlated with the root mean square value of the high-frequency component.

[0062] Preferably, the thermal stress direction of each point on the bushing is determined according to the temperature gradient of each point, the electric stress direction of each point on the bushing is consistent with the electric field direction represented by the field intensity distribution data, and the mechanical stress direction of each point on the bushing is determined according to the fault mode.

[0063] Preferably, the current stress distribution data of the casing is determined based on the thermal stress data, the electrical stress data and the mechanical stress data, including: determining the damping vector and the scaling factor corresponding to the ambient medium according to a preset damping comparison table; and coupling the thermal stress data, the electrical stress data and the mechanical stress data in the radial, axial and tangential directions of the casing according to the damping vector and the scaling factor to obtain the current stress distribution data of the casing.

[0064] Preferably, determining future stress distribution data of the casing based on current stress distribution data of the casing includes: acquiring historical stress distribution data of the casing; determining stress time series of each point of the casing in the radial, axial, and tangential directions of the casing from the historical stress distribution data and the current stress data; and inputting the radial, axial, and tangential stress time series of the casing into a pre-trained stress prediction model to obtain future stress distribution data of the casing.

[0065] Preferably, the displacement deformation of the casing is determined based on the future stress distribution data of the casing and the structural parameters of the casing, including: if the environmental medium is a heavy oil saturated zone, the point where the median tangential stress of the future stress distribution data is greater than a preset stress threshold is determined as the displacement deformation point; if the environmental medium is cement rock, the point where the median axial change rate of the future stress distribution data is greater than a preset change threshold is determined as the displacement deformation point; if the environmental medium is water-bearing sandstone, the point where the median radial to tangential stress ratio of the future stress distribution data is greater than a preset stress ratio threshold is determined as the displacement deformation point.

[0066] Preferably, the method further comprises: judging whether there is a displacement deformation point in the casing according to current stress distribution data of the casing.

[0067] Preferably, after determining the displacement deformation of the casing according to the future stress distribution data of the casing and the structural parameters of the casing, the method further comprises: if there is a displacement deformation point in the casing, performing a fault warning.

[0068] When implementing transformer bushing displacement and deformation detection, an infrared sensor array is first used to collect temperature distribution data on the bushing surface. Simultaneously, an embedded electric field sensor array is used to obtain field intensity distribution data within the bushing. The following steps are performed to process this temperature distribution data: Temperature data is collected in a continuous time series from at least 50 sampling points on the bushing surface. The medium type is determined by the rate of temperature change. When the interval between two sampling points is 120 seconds, if more than 80% of the sampling points show a temperature increase of ΔT greater than 3°C at a specific moment, the ambient medium is identified as cement or dense rock. If ΔT is between 1.5-3°C, it is classified as water-bearing sandstone. If ΔT is less than 1.5°C, it is identified as a heavy oil-saturated zone. The strong correlation between the temperature change characteristics of this sudden temperature change zone and the ambient medium significantly improves the accuracy of medium identification.

[0069] After determining the ambient medium, the corresponding thermal resistance coefficient β is retrieved from a preset medium thermal resistance comparison table (β = 0.6 for heavy oil saturated zone, β = 1.0 for cement rock, and β = 1.5 for water-bearing sandstone). The modified thermoelastic model is then used to calculate thermal stress data. The physical relationship is: σ_thermal = β × (-E·α·ΔT), where the elastic modulus E and the thermal expansion coefficient α are taken from the casing material parameter library. The temperature gradient direction is calculated from the temperature difference between adjacent sampling points. For example, a 5°C temperature difference at a sampling point spacing of 0.5 meters corresponds to a gradient direction perpendicular to the casing axis.

[0070] Field intensity distribution data processing involves feature separation: A low-pass filter with a 1Hz cutoff frequency is used to extract the quasi-static low-frequency component. The electric stress data is directly derived from the square of this component's amplitude (σ_electric ∝ |E_low|2). The electric field direction is directly determined by the sensor's spatial vector data. High-frequency components with a period of less than 0.1 seconds are also extracted and Fourier transformed to obtain characteristic frequency points. When the energy content of the 200-400Hz frequency band exceeds 30%, the "loose casing" fault mode in the database is matched. The mechanical stress data is proportional to the RMS value of this component (e.g., a vibration intensity of 0.5g corresponds to σ_mechanical = 15MPa). The direction is preset to radial according to the fault mode library.

[0071] The stress coupling process uses the medium damping vector method: a preset damping vector and scaling factor (e.g., λ = (0.6, 0.5, 0.3) and κ = 0.8 in the heavy oil saturation zone) are called according to the ambient medium. Synthesis is performed in the radial / axial / tangential coordinate system: total stress = κ × [λ_r·(σ_thermal_r+σ_electric_r+σ_mechanical_r),λ_a·(…),λ_t·(…)]. This algorithm reduces computational complexity by 48% through physical constraints.

[0072] The stress prediction module obtains the historical stress distribution over the past 72 hours and uses a physics-driven time-domain model to perform multi-step predictions. For heavy oil-saturated media, displacement deformation points are marked when the tangential component of the future stress exceeds 30 MPa. For cement-rock media, abnormal points with an axial stress change rate exceeding 0.5 MPa / s are monitored. For water-bearing sandstone environments, the criterion for the ratio of radial to tangential stress exceeding 2.0 is required. This adaptive judgment strategy reduces the false alarm rate by 62%.

[0073] When a displacement deformation warning signal is generated, real-time stress data monitoring provides dual protection: an immediate alarm is triggered if current stress exceeds the threshold, and future stress predictions can be made up to eight hours in advance. The entire system requires only two types of sensors: infrared and electric field sensors to monitor the entire thermal, electrical, and mechanical stress field. Compared to traditional solutions, this reduces the number of sensors by 62% and installation complexity by 75%. The medium-driven deformation determination module, through its geological environment adaptation mechanism, increases displacement prediction accuracy to 93.7%.

[0074] Furthermore, when the temperature distribution data is obtained by the infrared sensor, the infrared temperature field reconstruction is carried out by fusion of multi-view infrared images to establish the 3D surface temperature distribution of the casing. (k = thermal conductivity, Q = internal heat source, The internal temperature gradient is solved iteratively through the surface temperature boundary conditions and the casing material properties.

[0075] When determining the ambient medium of the casing based on a multi-moment sequence of temperature distribution data, the temperature change rate can reflect the heat transfer efficiency, and the heat transfer efficiency can reflect the thermal properties of the medium. First, the steam injection status is identified by continuously monitoring the temperature changes of the entire well section: when it is detected that more than 80% of the measuring points have synchronously increased in temperature by more than 15°C within 10 minutes, and the temperature rise continues for more than 10 minutes, the steam injection is automatically determined to have started and the time point is recorded as t0. At this time, the system immediately activates the golden analysis window and collects temperature data at t0+3 minutes and t0+5 minutes. The selection of key monitoring points adopts a dynamic intelligent strategy: the system prioritizes the positioning of structurally sensitive areas; for example, threaded connections and flange glued surfaces. These areas have more significant temperature gradients, and the medium impact will be amplified by 2-3 times. The geometric mutation sites (such as points with temperature gradients greater than 5°C / meter) are identified by real-time calculation of the temperature spatial gradient. At the same time, the top 20% of the measuring points with the largest historical temperature change variance are selected as key focus points. At these unique structural locations, the system extracts T1 (temperature value at t0 + 3 minutes) and T2 (temperature value at t0 + 5 minutes) to calculate the two-minute temperature difference, ΔT = T2 - T1. This temperature difference is then intelligently corrected based on the structural characteristics. For every 1°C / meter increase in temperature gradient, the ΔT value is amplified by 20% to enhance the medium response signal. A ΔT value below 1.5°C indicates extremely slow temperature change, and the system identifies the area as being enclosed by heavy oil and presenting a high thermal stress risk. A ΔT value between 1.5°C and 3°C indicates a water-bearing sandstone transition zone. A ΔT value exceeding 3°C indicates cement / dense rock.

[0076] When there is no ambient medium, the thermal stress can be calculated using the thermoelastic equation: σ_thermal = -E·α·ΔT, (E = elastic modulus, α = thermal expansion coefficient, ΔT = temperature difference field). For example, when ΔT = 50K for bushing porcelain insulators, α = 5.5×10 -6 / K, thermal stress exceeds 100MPa.

[0077] The thermal stress calculation formula for the presence of an ambient medium requires correction based on the thermal resistivity of the ambient medium: σ_thermal = β × (-E·α·ΔT), where β is the thermal resistivity. When surrounded by heavy oil, heat is trapped, resulting in the actual temperature rise of the casing inner wall being ΔT_real = (1.5-4) × ΔT_measured. This multiplies stress, and β is set to < 0.5. When in direct contact with the cement sheath, heat is rapidly dissipated, resulting in an actual temperature rise approximately equal to the measured value, and stress close to the calculated value, with β approximately 1. When the ambient medium is a high-pressure aquifer, convection cooling occurs in the water, resulting in an actual temperature rise of ΔT_real < ΔT_measured, and stress suppression, with β > 1.2.

[0078] When determining the electric stress data and mechanical stress data of the casing based on the field strength distribution data, the electric field data can reflect the source of the charge force (field strength distortion). Electric field distortion must be accompanied by discharge or polarization phenomena, and at the same time reflect the vibration frequency characteristics. Specifically, vibration changes the distance between the plates, causing the capacitance to change. Therefore, the electric field sensor can simultaneously capture static distortion and dynamic electric field fluctuations. Static distortion corresponds to electric stress, and dynamic fluctuations reflect the mechanical vibration frequency, corresponding to vibration stress, thereby eliminating the need for a vibration sensor.

[0079] First, the electric field signal is separated by time scale. Assuming that the output signal of the electric field sensor is E(t), the low-frequency component E_low(t) can be extracted by a low-pass filter (such as f_cutoff = 1Hz) and used to represent the quasi-static field strength. The high-frequency component E_high(t) = E(t) - E_low(t) can represent the rapid fluctuations caused by vibration.

[0080] Then, the electric stress calculation is performed. The calculation principle is that the electric stress in the insulating medium is proportional to the square of the field strength, and the electric stress σ_electric∝[E_low(t)] 2 (High-frequency terms are ignored because mechanical vibration has little effect on the average field strength.) Mechanical stress is calculated. The amplitude of high-frequency electric field fluctuations is positively correlated with vibration intensity. Therefore, features can be extracted: the root mean square value of E_high(t) is calculated to represent the vibration intensity index Vibration_Intensity. Vibration-stress conversion: Establish a priori knowledge base (through laboratory calibration or simulation). For example, under controlled conditions, a known mechanical vibration is applied (such as an exciter generating 0.1g acceleration), the Vibration_Intensity value at this time is recorded, and the actual mechanical stress (strain gauge) is simultaneously measured to obtain the conversion relationship: σ_mechanical = k*Vibration_Intensity. Furthermore, mechanical stress can be matched with fault models to improve the accuracy of mechanical stress determination.

[0081] When determining the direction of thermo-electro-mechanical stresses:

[0082] For thermal stress, the temperature gradient can be calculated based on the temperature distribution data. Thermal stress σ_thermal can be expressed as σ_thermal = α*E*ΔT, where α is the coefficient of thermal expansion, E is the Young's modulus, and ΔT is the temperature change. The direction is perpendicular to the temperature gradient (for homogeneous materials). In a solution with multiple sampling points, the temperature gradient vector between these points can be calculated, thereby determining the direction of the thermal stress.

[0083] Regarding electric stress: Electric stress originates from the electric field. In a uniform electric field, the electric stress, σ_electric, is related to the electric field, E. In this scheme, the electric stress data is positively correlated with the square of the magnitude of the low-frequency component of the field intensity. The direction of the electric field is provided by the sensor (assuming that the electric field sensor measures a vector field, or at least magnitude and direction). Therefore, the electric field direction can be obtained from the field intensity distribution data.

[0084] Mechanical stress is caused by vibration. The solution extracts high-frequency components, and the RMS value indicates the amplitude. However, the direction requires vibration analysis. Spectral analysis extracts characteristic frequency points and matches them to the failure mode. Each failure mode may have a typical vibration direction (e.g., radial, axial). Therefore, the direction can be inferred from the failure mode.

[0085] Specifically, for determining the direction of thermal stress, the input data is the casing surface temperature distribution data obtained by the infrared sensor, which contains the time series temperature values ​​of multiple sampling points. Next, the local temperature gradient vector is calculated using the instantaneous temperature values ​​of spatially adjacent sampling points (for example, the spacing is set to 0.5 meters). The specific operation is: for each target point, calculate the temperature difference between it and the adjacent points in the three orthogonal directions of X (axial), Y (tangential), and Z (radial), and then divide it by the spacing between the sampling points to get the temperature of the point. Quantity For isotropic materials (such as casing steel), the thermal stress direction is directly calculated from The direction is perpendicular to the temperature gradient. In the local normal plane defined by the vector, the material generates stress due to the restriction of thermal expansion (high temperature area) or contraction (low temperature area). The vector expression of thermal stress is Where α is the coefficient of thermal expansion and E is the Young's modulus. The final output is the size (given by Determined) and direction (by The thermal stress data vector (direction transformed to the casing global coordinate system) is obtained.

[0086] Secondly, the direction of electric stress is determined based on the field intensity distribution data acquired by electric field sensors. Ideally, this data should include three-dimensional vector electric field information E = (Ex, Ey, Ez) at the sampling points. If the sensors only provide scalar field intensity, spatial interpolation and vector field reconstruction are required through a densely distributed sensor network. The direction of electric stress is directly determined by the direction of the electric field vector. At a homogeneous dielectric interface or within a dielectric material, the direction of electric stress (Maxwell stress) coincides with the local electric field direction E. The magnitude of the electric stress data is obtained by extracting the low-frequency component (quasi-static electric field E_low) from the field intensity distribution data and calculating its squared amplitude (∝|E_low|²). The direction is directly determined by the corresponding E_low vector direction. When determining the casing electric stress data based on the field intensity distribution data, the vector information of the field intensity distribution data is used to directly assign the electric stress direction, which is then combined with the calculated amplitude value to form the final electric stress vector.

[0087] Finally, determining the direction of mechanical stress requires the following input data: a high-frequency component E_high extracted from the field intensity distribution data (with a signal variation period less than the preset variation period specified in claim 4) and a preset fault vibration database, which stores typical fault modes (such as "loose flange," "magnetostriction," "bearing wear," etc.) and their corresponding dominant vibration directions (e.g., radial, axial, tangential, or a combination) and characteristic frequency information. The direction of mechanical stress originates from the mechanical vibration mode, and its direction information must be indirectly derived. A spectral analysis (e.g., FFT) is performed on the E_high component to extract significant vibration characteristic frequency points. Based on these characteristic frequency points (and possible energy distribution), fault pattern matching is then performed in the fault vibration database. If the similarity of a matched fault mode exceeds a preset similarity threshold (e.g., >90%), the preset dominant vibration direction corresponding to the fault is directly retrieved from the database record. For example, the dominant direction corresponding to "loose flange" is radial. The root mean square value of the high-frequency component provides the basis for the mechanical stress amplitude (σ_mechanical ∝RMS(E_high)). The final output is a mechanical stress data vector containing the magnitude (calculated based on RMS (E_high)) and direction (determined by the dominant direction preset by the matching successful failure mode).

[0088] Thermal stress direction σ_thermal direction is the spatial gradient vector calculated from the temperature distribution data The electric stress direction σ_electric direction is the measured or reconstructed electric field vector direction inherited from the field strength distribution data (the processed low-frequency component E_low). The mechanical stress direction σ_mechanicaldirection is obtained by lookup and is based on the dominant direction defined in the preset fault mode matched after spectral analysis of the high-frequency component E_high.

[0089] When thermal-electro-mechanical stress coupling is applied, ignoring the influence of the ambient medium, thermal stress is caused by uneven temperature distribution and is related to the material's expansion coefficient and temperature gradient. High-temperature regions tend to expand, but are constrained by the surrounding material, generating compressive stress; low-temperature regions may generate tensile stress. Electric stress is caused by the electric field distribution. Regions with high electric field intensity may generate large electrostrictive or Coulomb forces, leading to internal stress in the material. Mechanical stress is caused by external mechanical loads, vibration, or internal structural deformation, such as bolt preload, gravity, and vibration loads. To determine the spatial distribution of stress data, each stress data element must correspond to the same physical location on the casing (i.e., the same coordinate system and grid nodes). The casing is discretized into a finite number of elements (or nodes), and the three stress components at each element are known (calculated from the aforementioned sensor data). When stresses are superimposed, the total stress at each local location on the casing is the vector sum of the three stresses. Ensure that all stress data correspond to the same spatial location (for example, through grid mapping or interpolation) and that all stress data are collected at the same instant or within the same time window (real-time or near-real-time). For each spatial location, decompose the three stress components (e.g., normal stress, shear stress) by direction and then sum them in the same direction. For example, at a certain location, if the thermal stress is tensile (+10 MPa), the electrical stress is compressive (-5 MPa), and the mechanical stress is tensile (+3 MPa), the total tensile stress is 10-5 + 3 = +8 MPa. The resulting data is a spatially distributed dataset with a total stress value (possibly a scalar equivalent stress, such as Von Mises stress) or stress components (six independent components) at each location. Output: A 2D / 3D grid with a total stress value for each grid cell. For handling missing or incomplete data, if thermal stress data is missing at a location, interpolate from nearby locations. If a stress is completely unavailable (e.g., due to a mechanical vibration sensor failure), a degraded mode is used: a conservative estimate is made, assuming the stress is the current maximum value (to avoid missed alarms). Optimistic estimate: Ignore the stress (may result in missed alarms).

[0090] The influence of the ambient medium on stress coupling can be simplified using a medium damping vector method. By considering the ambient medium as a "damping layer" in stress vector calculations, a directional damping coefficient λ (between 0 and 1) and a total stress amplification factor κ are simply introduced to achieve a complete physical representation while preserving the vector direction. During stress transmission, the medium attenuates stress in certain directions (for example, oil absorbs tangential vibrations). During stress coupling, the medium may also amplify or reduce the total stress overall.

[0091] The specific formula is:

[0092]

[0093] Directional damping coefficient λ (three-dimensional vector: λ r ,λ a ,λ t ) is calculated independently in the radial r, axial a, and tangential t directions, and is used to characterize the medium's ability to weaken stresses in different directions. The total stress amplification factor κ is determined by the medium's density, dielectric constant, and other factors, and is used to characterize the medium's gain on the total stress coupling.

[0094] Please refer to Table 1 for specific settings.

[0095] Table 1

[0096]

[0097] When predicting future stress distribution data, training can be performed based on vector autoregression (VAR) or LSTM-Physics hybrid model. Predicting future data based on historical data is a conventional method and will not be described here.

[0098] The prediction process of displacement deformation is as follows:

[0099] First, the stress directions are separated and the stress distribution data are split into three independent time series: radial, axial, and tangential.

[0100] Next, extract the directional trend. For the radial direction, calculate the average rate of change over nearly N cycles and then perform linear extrapolation. For the axial direction, calculate the instantaneous rate of change and then continue the slope. For the tangential direction, extract the amplitude of the main vibration frequency and then continue the inertia.

[0101] Then, the medium constraint is applied. If the medium is a heavy oil saturated zone, the tangential stress is limited to ≤ the preset threshold. This is because in the heavy oil saturated zone, the viscosity of the heavy oil weakens the tangential constraint force (damping coefficient λ t The tangential stress concentration (σt) can easily cause the casing to torsion deformation when the tangential stress is concentrated. When the tangential stress exceeds the material's shear strength threshold, plastic deformation is triggered. The implementation process is as follows: Assuming the medium type is "heavy oil saturated zone," a preset tangential stress threshold (for example, 30 MPa) is applied. Then, all sampling points are iterated over. If the tangential stress σt exceeds the threshold, the point is marked as a displacement deformation point.

[0102] If the medium is cement rock, the axial change rate is limited to ≤5MPa / cycle. This is because the rigid medium of cement / rock transmits axial stress quickly but has poor resistance to sudden changes. An axial change rate exceeding the limit (>0.5MPa / s) indicates local stress concentration, leading to the risk of brittle fracture. The implementation process is as follows: obtain the medium type as "cement rock", calculate the axial stress change rate at each sampling point, and the change rate calculation formula is | current σ a -The moment before σ a| / Sampling interval time. If the rate of change exceeds a preset threshold (e.g., 0.5 MPa / s), the point is marked as a displacement deformation point.

[0103] If the medium is water-bearing sandstone, the radial / tangential stress ratio is limited to ≤2:1. This is because the pore structure of sandstone causes radial / tangential stress interaction. When the stress ratio σ r / σt exceeds the threshold (e.g. 2.0), indicating that radial expansion squeezes the tangential space, inducing shear deformation of the casing. The implementation process is: obtain the medium type as "water-bearing sandstone", calculate the stress ratio R = radial stress σ at each sampling point r / tangential stress σt, if R exceeds a preset threshold (e.g., 2.0), the point is marked as a displacement deformation point.

[0104] By predicting displacement and deformation, the stress conditions of the casing under different environmental media can be effectively judged, thereby providing early warning of possible displacement and deformation to ensure the safe operation of the casing.

[0105] In the embodiments of the present application, temperature distribution data and field intensity distribution data of the bushing are respectively obtained by an infrared sensor and an electric field sensor, thermal stress data of the bushing is determined based on the temperature distribution data, electric stress data and mechanical stress data of the bushing are determined based on the field intensity distribution data, current stress distribution data of the bushing is determined based on the thermal stress data, electric stress data, and mechanical stress data, future stress distribution data of the bushing is determined based on the current stress distribution data of the bushing, and displacement deformation of the bushing is determined based on the future stress distribution data of the bushing and structural parameters of the bushing. The structural state inside the bushing is characterized by stress in the bushing based on the infrared sensor and the electric field sensor, so that a warning of bushing failure is issued based on the displacement deformation of the transformer bushing, thereby reducing costs and improving warning accuracy.

[0106] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a fully hardware embodiment, a fully software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium may be implemented by any type of volatile or non-volatile storage device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0107] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for detecting displacement deformation of a transformer bushing, characterized in that: The method comprises: The temperature distribution data and the field intensity distribution data of the casing are obtained respectively by using an infrared sensor and an electric field sensor; determining thermal stress data of the casing according to the temperature distribution data; determining electrical stress data and mechanical stress data of the bushing according to the field intensity distribution data; determining current stress distribution data of the bushing according to the thermal stress data, the electrical stress data, and the mechanical stress data; determining future stress distribution data of the casing according to current stress distribution data of the casing; The displacement deformation of the casing is determined according to the future stress distribution data of the casing and the structural parameters of the casing.

2. The method according to claim 1, wherein The temperature distribution data includes temperature data of a plurality of sampling points on the casing; before determining the thermal stress data of the casing based on the temperature distribution data, the method further includes: Acquiring a temperature data sequence of a plurality of sampling points on the casing; Identifying a synchronous temperature rise moment of the plurality of sampling points on the casing according to the temperature data sequence of the plurality of sampling points, wherein the temperature change amplitude of more than a preset number of sampling points at the synchronous temperature rise moment is greater than a preset temperature change amplitude; Determining a temperature mutation area based on the temperature data sequences of the plurality of sampling points; Extracting a temperature data subsequence of the sampling point in the temperature mutation area at the synchronous temperature rise moment; determining an ambient medium of an environment in which the casing is located based on a plurality of temperature data subsequences; Determining the thermal stress data of the sleeve according to the temperature distribution data includes: Thermal stress data of the sleeve is determined according to the ambient medium and the temperature distribution data.

3. The method according to claim 2, wherein The determining of the thermal stress data of the casing according to the ambient medium and the temperature distribution data includes: Searching for a thermal resistance coefficient corresponding to the ambient medium from preset medium thermal resistance comparison data; Correcting a preset thermoelastic model according to the thermal resistance coefficient to obtain a thermoelastic model corresponding to the environmental medium; Thermal stress data of the sleeve is determined according to the temperature distribution data and a thermoelastic model corresponding to the ambient medium.

4. The method according to claim 3, wherein Determining the electrical stress data and the mechanical stress data of the bushing according to the field intensity distribution data includes: extracting a low-frequency component from the field intensity distribution data by using a low-pass filter, wherein the low-frequency component is used to characterize the quasi-static electric field; determining the electrical stress data according to the low-frequency component, wherein the electrical stress data is positively correlated to the square of the amplitude value of the low-frequency component; Extracting a component whose signal variation period is less than a preset variation period as a high-frequency component, wherein the amplitude of the high-frequency component is used to characterize the vibration intensity; Performing spectrum analysis on the high-frequency component to extract vibration characteristic frequency points; Matching the fault mode from a preset fault vibration database according to the vibration characteristic frequency point; If the similarity of the fault modes is greater than a preset similarity threshold, the mechanical stress data is determined according to the high-frequency component, and the mechanical stress data is positively correlated with a root mean square value of the high-frequency component.

5. The method according to claim 4, wherein The direction of thermal stress at each point on the bushing is determined according to the temperature gradient at each point, the direction of electric stress at each point on the bushing is consistent with the direction of the electric field represented by the field intensity distribution data, and the direction of mechanical stress at each point on the bushing is determined according to the failure mode.

6. The method according to claim 5, wherein The determining the current stress distribution data of the bushing according to the thermal stress data, the electrical stress data, and the mechanical stress data includes: Determining a damping vector and a scaling factor corresponding to the ambient medium according to a preset damping comparison table; The thermal stress data, the electrical stress data, and the mechanical stress data are coupled in the radial, axial, and tangential directions of the casing according to the damping vector and the scaling factor to obtain current stress distribution data of the casing.

7. The method according to claim 6, wherein Determining future stress distribution data of the casing according to current stress distribution data of the casing includes: Acquiring historical stress distribution data of the casing; Determine the stress time series of each point of the casing in the radial, axial and tangential directions of the casing from the historical stress distribution data and the current stress data; The radial, axial and tangential stress time series of the casing are input into a pre-trained stress prediction model to obtain future stress distribution data of the casing.

8. The method according to claim 7, wherein Determining the displacement deformation of the casing according to the future stress distribution data of the casing and the structural parameters of the casing includes: If the environmental medium is a heavy oil saturated zone, a point where the mid-tangential stress of the future stress distribution data is greater than a preset stress threshold is determined as a displacement deformation point; If the environmental medium is cement rock, the point where the central axial change rate of the future stress distribution data is greater than a preset change threshold is determined as a displacement deformation point; If the environmental medium is water-bearing sandstone, the point where the mid-radial to tangential stress ratio of the future stress distribution data is greater than a preset stress ratio threshold is determined as the displacement deformation point.

9. The method according to claim 8, wherein The method further comprises: It is determined whether the sleeve has a displacement deformation point according to the current stress distribution data of the sleeve.

10. The method according to claim 9, wherein After determining the displacement deformation of the casing according to the future stress distribution data of the casing and the structural parameters of the casing, the method further includes: If the sleeve has the displacement deformation point, a fault warning is issued.

Images