Flexible spiral sensor and method for transformer lead bushing crack detection

Abstract

The application belongs to the field of transformer state monitoring and fault diagnosis, and particularly relates to a flexible spiral type sensor and method for detecting cracks in a transformer lead bushing, which mainly comprises distributed strain, temperature, pressure and piezoelectric sensors; all functional modules are integrated on the same flexible substrate, have certain bending deformation capacity, can be spirally wound on the surface of the transformer lead bushing, realize conformal winding of the flexible sensor and the tubular structure, distribute sensing and collecting multi-parameter information of the transformer bushing, combine the passive sensed transformer bushing state data and the active detected surface / interior defect data, realize online collection of the stress state information of the transformer bushing and comprehensive diagnosis of defects.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of transformer state monitoring and fault diagnosis, in particular to a flexible spiral sensor and method for detecting cracks in transformer lead bushings. BACKGROUND

[0002] The structure and dynamic performance of transformer lead bushings play a decisive role in the normal operation and intelligent operation and maintenance of transformers. How to monitor the state information of transformer lead bushings online is a current research hotspot. Especially under the influence of multiple factors such as high and low temperature, strong wind, and complex electromagnetic field, and especially the complex structure of the lead bushing itself poses new challenges to its online monitoring. It is urgent to develop a multifunctional sensor device for the complex curved surface structure of the transformer lead bushing to realize online health monitoring of the transformer lead bushing.

[0003] In recent years, with the rapid development of sensor and intelligent diagnosis technology, the development of remote monitoring technology and methods for power infrastructure such as transformers has been promoted, providing technical support for the remote online monitoring of the structural health state of transformer lead bushings. Patent CN202011080402.9 discloses a transformer bushing fault diagnosis method and system based on multiple source parameters, which performs discharge test on the high-voltage bushing based on the bushing frequency domain dielectric spectrum test method, and proposes a fault diagnosis method for the capacitor core of the high-voltage bushing under test based on the trained BP neural network. Invention patent 201610027929.2 discloses a power transformer fault diagnosis system based on Internet of Things and cloud computing, which includes a power transformer fault monitoring center and each power transformer monitor. Through the current sensor, temperature and humidity sensor, hydrogen concentration measurer, acetylene concentration measurer, and carbon monoxide concentration measurer, all power transformers in the area are monitored for faults, which can better and more comprehensively ensure the safe, reliable, high-quality, and economic operation of all power transformers in the area. Patent CN202020686233.2 discloses a power transformer fault diagnosis and monitoring device based on Internet of Things, which solves the problem that the existing power transformer fault diagnosis and monitoring device cannot be replaced when the non-dismountable handle is damaged and is not easy to move and transport. The current technology mainly solves the state health monitoring of transformer infrastructure based on multiple sensors and intelligent diagnosis methods, and gives an online monitoring and fault diagnosis method for transformer lead bushings. However, with the rapid development of power infrastructure and smart grid technology, it is urgent to develop a multifunctional sensor device for the complex curved surface structure of the transformer lead bushing to realize online health monitoring of the transformer lead bushing.

[0004] The application discloses a flexible spiral sensing and method for transformer lead bushing crack detection, mainly comprising distributed strain, temperature, pressure and piezoelectric sensors; all the functional modules are integrated on the same flexible substrate, have certain bending deformation ability, can be spirally wound on the surface of the transformer lead bushing, realize conformal winding of the flexible sensor and the tubular structure, distributed sensing and collection of multi-parameter information of the transformer bushing, combination of the passively sensed transformer bushing state data and the actively detected surface / interior defect data, realization of online collection of the stress state information of the transformer bushing and comprehensive diagnosis of defects. SUMMARY

[0005] In view of the above defects or improvement requirements of the prior art, the first purpose of the application is to provide a flexible spiral sensing for transformer lead bushing crack detection, mainly comprising distributed strain, temperature, pressure and piezoelectric sensors; the strain sensor adopts a piezoresistive sensor, constitutes multiple Wheatstone bridges, and is used for collecting surface deformation and structural vibration monitoring of the bushing; the temperature sensor adopts a thermal resistance film sensor, and is used for collecting environmental changes of the transformer bushing; the pressure sensor adopts a thin film capacitor sensor, and is used for collecting static pressure conditions of the bushing; the piezoelectric sensor generates a Lamb signal under the action of an external excitation signal, compares echo and guided wave signals, and is used for detecting columnar interior structural defect information of the bushing; all the functional modules are integrated on the same flexible substrate, have certain bending deformation ability, can be spirally wound on the surface of the transformer lead bushing, realize conformal winding of the flexible sensor and the tubular structure, distributed sensing and collection of multi-parameter information of the transformer bushing, and online collection of the stress state information of the transformer bushing and comprehensive diagnosis of faults.

[0006] In view of the above defects or improvement needs of the prior art, a second object of the present application is to provide a flexible spiral sensing structure optimization model and method for a transformer lead bushing curved surface, the conformal winding effect of the thin film sensing device and the bushing curved surface mainly depends on the material properties, thickness, width, and winding spacing of the thin film sensing, and the curvature of the bushing; a flexible sensing device conformal winding and structure optimization model is constructed based on the thickness, width, and winding spacing of the thin film sensing and the curvature of the bushing; the flexible spiral sensor is in a strip shape and can be spirally wound onto the surface of the transformer bushing, the thickness, width, and winding spacing of the thin film sensing are calculated and analyzed to adapt to the sensing device and the lead bushing, a sensing device structure optimization strategy based on the width and thickness of the device is proposed, the different properties of various materials of the sensing device are considered, the inadaptation problem between the sensing device and the lead bushing in terms of mechanics and geometric structure is solved, the winding conformal with the transformer lead bushing is realized, the performance response, anti-interference and stability of the 3D spiral sensing device are improved; a 3D spiral sensing device structure optimization strategy is proposed, a 3D spiral sensing implementation mechanism is given, a mechanical analysis model of the spiral sensing device and the transformer lead bushing is constructed, and the conformal sensing and online diagnosis of the transformer lead bushing crack defect are realized.

[0007] Further, a flexible spiral sensing and method for transformer lead bushing crack detection, the 3D spiral sensing device is designed and prepared from a planar 2D sensing device, different sensing unit integration of the 3D spiral sensing device is completed on a planar substrate, and it supports spiral mounting on the transformer lead bushing, ensuring distributed sensing of the 3D spiral sensing device on the lead bushing.

[0008] Further, a flexible spiral sensing and method for transformer lead bushing crack detection, the 3D spiral sensing device and the lead bushing between the spiral mounting process mainly includes: (1) flattening sub-process; (2) rotating sub-process; (3) stretching sub-process; (4) winding sub-process; the interface contact model of the 3D spiral sensing device and the lead bushing curved surface is mainly based on the four processes, the modeling of the 3D spiral sensing device and the lead bushing curved surface is completed, so as to analyze the influence of the material properties, thickness, width, and winding spacing of the thin film sensing, and the curvature of the bushing on the spiral winding performance effect, and solve the mechanical and geometric adaptation problem.

[0009] Further, the preparation and installation method of the flexible spiral sensing device for transformer lead bushing crack detection comprises the following process steps:

[0010] (1) different sensing unit design and processing preparation is completed, micro-electromechanical processing technology is used to complete the processing and preparation of flexible strain, temperature, pressure, and piezoelectric sensors;

[0011] (2) Construct a 3D spiral sensing and lead sleeve winding mechanical model, and calculate and analyze the interface mechanical behavior between the 3D spiral sensing and the lead sleeve for (1) flattening sub-process; (2) rotating sub-process; (3) stretching sub-process; (4) winding sub-process;

[0012] (3) Construct a flexible sensing device conformal winding and structure optimization model based on the thickness, width, and winding spacing of the film sensing and the curvature of the sleeve, and calculate and analyze the thickness, width, and winding spacing of the film sensing on the adaptability of the sensing device and the lead sleeve;

[0013] (4) Processing of long flexible circuit board to ensure matching between it and the flexible sensing unit, while meeting the winding adaptation process of the 3D spiral sensing;

[0014] (5) Mounting between the flexible sensing unit pin and the flexible circuit board soldering pin, according to the design scheme, to ensure directional mounting of the sensing unit and the soldering pin on the flexible circuit board, and complete the integration of the planar long flexible multifunctional sensing device;

[0015] (6) Packaging of the multifunctional sensing device, considering the influence of the lead sleeve on the electrical performance of the sensor, and the application of the sensing unit in the lead sleeve requires electrical isolation, a layer of organic polymer film is sprayed on the surface of the multifunctional sensing device to realize the packaging of the flexible multifunctional sensing device;

[0016] (7) Spiral winding of the 3D spiral sensing device on the lead sleeve, the flexible planar multifunctional sensing device is spirally wound on the curved surface of the transformer sleeve, which is used for distributed multifunctional data acquisition of the transformer sleeve.

[0017] Further, a transformer lead sleeve crack detection fusion sensing method based on flexible spiral sensing, focuses on comprehensive analysis of strain, temperature, pressure, and piezoelectric sensing data of the spiral winding on the lead sleeve, extraction of different sensing dynamic data features of the curved surface of the lead sleeve, construction of a sleeve defect diagnosis and recognition model based on deep learning, construction of a sleeve defect data set under typical defects for training and recognition of deep learning algorithm, fusion of multi-sensing state information of the lead sleeve under different defect morphologies, and online detection and classification recognition of transformer lead sleeve cracks based on flexible spiral sensing data.

[0018] Further, the flexible spiral sensing device for transformer lead sleeve crack detection, the fusion sensing algorithm for sleeve cracks mainly includes the following processes:

[0019] (1) Multi-sensing data acquisition and preprocessing: the flexible multifunctional sensor acquires strain, temperature, pressure, and guided wave / back wave information of the transformer lead sleeve online;

[0020] (2) The strain, temperature, pressure and guided wave / echo information of the transformer lead bushing are segmented and effectively represented by a dynamic time window;

[0021] (3) A training and test sample set for transformer lead bushing defect classification is constructed;

[0022] (4) Feature information of the transformer lead bushing state information is extracted respectively;

[0023] (5) The training process of the transformer lead bushing state information, the test sample set of the transformer lead bushing is trained, and the output performance of the transformer lead bushing under different defect states is analyzed;

[0024] (6) Test process: the test sample set is used to test and analyze the deep learning network of the transformer lead bushing defect constructed;

[0025] (7) The relationship between the algorithm calculation result and the actual state of the transformer lead bushing is compared, the recognition success rate is calculated, and the safety performance of the transformer lead bushing is quickly evaluated.

[0026] Further, the flexible spiral sensing device for detecting cracks in the transformer lead bushing proposes a comprehensive evaluation method for lead bushing cracks based on multiple sensing signals, which integrates strain, temperature, pressure, and piezoelectric sensing data spirally wound on the lead bushing, and is used to fuse the multi-sensing state information of the lead bushing under different defect morphologies. Based on the flexible spiral sensing data, online detection and classification recognition of transformer lead bushing cracks are realized, early defect detection of transformer lead bushing is realized, online diagnosis of bushing defects is realized, and post-maintenance is realized.

[0027] Further, the flexible spiral sensing device for detecting cracks in the transformer lead bushing, the flexible 3D spiral sensing device is conformally attached to the surface of the transformer lead bushing, and the electrical isolation between the flexible 3D spiral sensing device and the bushing is realized by using organic polymer packaging technology. Conformal spiral winding ensures the accuracy and reliability of the flexible 3D spiral sensing device in sensing and collecting bushing state information; the constructed interface mechanics model can be used to guide the structural optimization design of the flexible 3D spiral sensing device, and ensure the conformal attachment of the flexible 3D spiral sensing device and the bushing; a fusion sensing method based on flexible spiral sensing for detecting cracks in the transformer lead bushing is proposed, which is used to fuse the multi-sensing state information of the lead bushing under different defect morphologies, and realizes online detection and classification recognition of transformer lead bushing cracks.

[0028] Further, the flexible spiral sensing device for detecting cracks in the transformer lead bushing, the flexible 3D spiral sensing device can be applied to other curved surface structures besides being conformally attached to the surface of the transformer lead bushing, and the winding technology of the spiral sensing device can be widely applied to other curved surface structures.

[0029] Overall, compared with the prior art, the flexible 3D spiral sensing device and method for transformer lead bushing crack detection conceived by the present application have the advantages that the 3D spiral sensing device has advantages that cannot be matched by flexible planar electronic devices, can be conformally wound on the curved surface of the lead bushing, can distribute sensing information in different directions and different parts of the transformer lead bushing, and can realize omnidirectional sensing of the transformer lead bushing; the interface mechanics model can be used to guide the structural optimization design of the flexible 3D spiral sensing device, to ensure the conformal attachment of the flexible 3D spiral sensing device and the bushing; the fusion sensing method based on the flexible spiral sensing for transformer lead bushing crack detection is proposed, which is used to fuse the multi-sensing state information of the lead bushing under different defect morphologies, to realize online detection and classification identification of the transformer lead bushing crack.

[0030] The flexible 3D spiral sensing device for transformer lead bushing crack detection can be conformally attached on the surface of the transformer lead bushing, and can also be applied to the surfaces of other curved structures. The winding technology of the spiral sensing device can be widely applied to other curved structures, and can expand the application range of the flexible 3D spiral sensing in online monitoring of curved structures in different industries.

[0031] In summary, the flexible spiral sensing device for transformer lead bushing crack detection integrates distributed strain, temperature, pressure and piezoelectric sensing functions, is integrated on the same flexible substrate, has certain bending deformation capability, can be spirally wound on the surface of the transformer lead bushing, a flexible sensing device conformal winding and structural optimization model based on the thickness, width, winding spacing and curvature of the bushing is constructed, a sensing device structural optimization strategy based on the coordinated optimization of the device width and thickness is proposed, the material selection and structural optimization design of the flexible 3D spiral sensor are guided, the conformal attachment of the flexible 3D spiral sensing device and the transformer bushing is ensured, a fusion sensing method based on the flexible spiral sensing for transformer lead bushing crack detection is proposed, online detection and classification identification of the transformer lead bushing crack are realized, and the flexible film material can provide good tensile and compression performance, providing a wider application range for flexible 3D spiral electronic devices. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 For state monitoring and defect diagnosis of the transformer lead bushing based on the flexible 3D spiral sensing, (a) is an attachment and integration schematic diagram of the flexible 3D spiral multifunctional sensing device and the transformer lead bushing, (b) is a planar expansion schematic diagram of the flexible 3D spiral multifunctional sensing device and main sensing function units, and (c) is a multifunctional sensing principle for state monitoring of the transformer lead bushing.

[0033] Figure 2 For the 3D spiral electronic and the interface mechanical model of the column surface, (a) 3D spiral conformal mounting effect; (b) device specific schematic diagram; (c) multi-layer film device bending analysis modeling.

[0034] Figure 3 Temperature state information.

[0035] Figure 4 Strain and vibration state information.

[0036] Figure 5 Pressure state information.

[0037] Figure 6 Piezoelectric wave and echo information.

[0038] Figure 7 Transformer lead bushing multi-parameter fusion sensing method based on flexible 3D winding multifunctional sensing device.

[0039] Figure 8 Flexible 3D spiral multifunctional sensing device based transformer lead bushing state monitoring practical process, realize bushing stress, displacement, deformation monitoring, and fault defect diagnosis.

[0040] The meanings of the symbols in the figure are as follows:

[0041] 10-Transformer lead bushing.

[0042] 20-Flexible 3D spiral multifunctional sensing device.

[0043] 30-Different sensing units.

[0044] 21-Flexible substrate; 22-Capacitance sensing unit; 23-Thermal resistance sensing unit; 24-Strain sensing unit; 25-Metallic wire. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical scheme and advantages of the present application clearer and more apparent, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application. In addition, the technical features involved in each embodiment of the present application described below can be combined with each other as long as they do not conflict with each other.

[0046] Reference Figure 1a, the flexible 3D spiral multifunctional sensor 20 of the invention is wound on the complex structure surface of the transformer lead sleeve 10, and different sensing units 30 are integrated on the multifunctional sensor device 20, which is spirally wound on the surface of the lead sleeve for collecting strain, pressure and internal defect information of the lead sleeve surface, and realizing omnidirectional perception of the lead sleeve. Figure 1 b shows the main sensing function devices of the flexible 3D spiral multifunctional sensor 20 integrated with a thin film device, four flexible strain sensing units 24 form a Wheatstone strain gauge, which uses a metal thin film piezoresistive sensor to sense the structural deformation state information of the lead sleeve; the thermal resistance sensing unit 23 is used to sense the surface temperature information of the lead sleeve; the flexible capacitance sensing unit 22 adopts a three-layer thin film structure to collect the pressure information of the lead sleeve surface; the flexible piezoelectric unit is a three-layer thin film structure, including upper and lower electrodes 26, 28, and a middle layer of functional material 27 which is a piezoelectric material. Under the excitation of an external signal (five-peak sine wave), the piezoelectric unit generates Lamb signals that propagate on the surface and inside the sleeve structure. Other piezoelectric units are used to receive Lamb echo signals, compare the echo and Lamb guided wave signals, and detect internal and surface defect information of the lead sleeve; the strain sensing unit 24, the thermal resistance sensing unit 23, the capacitance sensing unit 22 and the piezoelectric sensing unit 27 are all integrated on a flexible substrate 21, completing the design of a planar flexible multifunctional sensor. All sensing unit signals are led out through metal wires 25 for easy monitoring of the transformer lead sleeve. Figure 1 (c) is the principle of multifunctional sensing for transformer lead sleeve state monitoring, (i) is a Wheatstone bridge for collecting the change value of the resistance sensor; (ii) is a variable pitch capacitance sensor for measuring the surface pressure of the sleeve; (iii) is a flexible piezoelectric active detection principle for detecting internal structural defects of the lead sleeve. Under the excitation of an external signal (five-peak sine wave), the piezoelectric unit generates Lamb signals that propagate on the surface and inside the sleeve structure. Other piezoelectric units are used to receive Lamb echo signals, compare the echo and Lamb guided wave signals, and detect internal and surface defect information of the lead sleeve.

[0047] The flexible planar multifunctional device can be spirally wound on the surface of the lead sleeve, and the design method from 2D to 3D needs to be studied to construct a 3D spiral sensing and lead sleeve winding mechanics model, as shown in Figure 2 a, and the local effect of the invented device is shown in Figure 2 b, the bending effect of the multi-layer thin film device is shown in Figure 2c. The flexible thin film sensor device is attached to the cylindrical surface in a 3D spiral winding manner, and the conformal winding effect of the thin film sensor device and the sleeve surface mainly depends on the material properties, thickness, width, and winding spacing of the thin film sensor and the curvature of the sleeve; a conformal winding and structure optimization model of the flexible sensor device based on the thickness, width, and winding spacing of the thin film sensor and the curvature of the sleeve is constructed, the adaptability of the thickness, width, and winding spacing of the thin film sensor to the sensor device and the lead sleeve is calculated and analyzed, a sensor device structure optimization strategy based on the collaborative optimization of the device width and thickness is proposed, the different properties of various materials of the sensor device are considered, the inadaptation problem between the mechanical and geometric structures of the sensor device and the lead sleeve is solved, the 3D spiral winding conformal with the transformer lead sleeve is realized, the performance response, anti-interference, and stability of the 3D spiral sensor device are improved; a 3D spiral sensor device structure optimization strategy is proposed, a 3D spiral sensor implementation mechanism is given, and a mechanical analysis model of the spiral sensor device and the transformer lead sleeve is constructed.

[0048] The flexible 3D winding multifunctional sensor device is spirally wound on the surface of the lead sleeve, and the preparation and installation method of the flexible spiral sensor device comprises the following process steps:

[0049] (1) Complete different sensor unit design and processing preparation, and complete flexible strain, temperature, pressure, and piezoelectric sensor processing and preparation by using microelectromechanical processing technology;

[0050] (2) Construct a 3D spiral sensor and lead sleeve winding mechanical model, and calculate and analyze the interface mechanical behavior between the 3D spiral sensor and the lead sleeve for (1) flattening sub-process; (2) rotating sub-process; (3) stretching sub-process; (4) winding sub-process;

[0051] (3) Construct a conformal winding and structure optimization model of the flexible sensor device based on the thickness, width, and winding spacing of the thin film sensor and the curvature of the sleeve, and calculate and analyze the adaptability of the thickness, width, and winding spacing of the thin film sensor to the sensor device and the lead sleeve;

[0052] (4) Processing of long strip flexible circuit board to ensure matching between the flexible sensor unit and the flexible circuit board, and to meet the winding adaptation process of the 3D spiral sensor;

[0053] (5) Mounting between the flexible sensor unit pin and the flexible circuit board soldering pin, according to the design scheme, ensure the directional mounting of the sensor unit and the soldering pin on the flexible circuit board, and complete the integration of the planar long strip flexible multifunctional sensor device;

[0054] (6) The packaging of the multifunctional sensing device considers the influence of the lead sleeve on the electrical performance of the sensor. The application of the sensing unit in the lead sleeve requires electrical isolation. A layer of organic polymer film is sprayed on the surface of the multifunctional sensing device to realize the packaging of the flexible multifunctional sensing device.

[0055] (7) The spiral winding of the 3D spiral sensing device on the lead sleeve. The flexible planar multifunctional sensing device is wound on the curved surface of the transformer sleeve in a spiral manner for distributed multifunctional data acquisition of the transformer sleeve.

[0056] The flexible 3D spiral multifunctional sensing device integrates temperature, strain (vibration), pressure, and active flaw detection functions. The data obtained by winding on the surface of the lead sleeve is shown in the figure. Figures 3-6 Figure 3 The flexible temperature sensor collects temperature state information on the surface of the lead sleeve and the environment, which can reflect the temperature change of the sleeve in real time. Figure 4 The flexible Wheatstone bridge-based strain and vibration state information of the lead sleeve is collected. It can be seen that the vibration signal has no obvious characteristics in the time domain, but has obvious characteristics in the frequency domain. Generally, the vibration signal is analyzed and processed through frequency domain transformation. Figure 5 The flexible capacitive sensor collects the pressure change information on the surface of the sleeve. Figure 6 The comparison information of the piezoelectric guided wave and the echo information. Under the condition of different defects of the sleeve, the echo signal is tested. The characteristic difference between the echo signal and the guided wave signal can be used to identify the internal defects of the lead sleeve.

[0057] The transformer lead sleeve multi-parameter fusion sensing method based on the flexible 3D winding multifunctional sensing device is described in Figure 7 , which mainly includes:

[0058] (1) The flexible 3D spiral multifunctional sensing device is wound on the surface of the transformer lead sleeve to collect strain, temperature, and pressure information on the surface of the transformer sleeve in real time, and actively detect internal defects of the lead sleeve through guided wave mode;

[0059] (2) Extract the strain, temperature, and pressure collected by the flexible 3D spiral multifunctional sensing device, and the characteristic information of the echo signal. Calculate the characteristic information of different parameters in the time-frequency domain, and calculate the damage factor corresponding to different defects.

[0060] (3) Train the test sample set constructed by CNN or fully connected network to obtain the main internal parameters of CNN or fully connected network and optimize the network performance.

[0061] ​(4) For the main defects of the surface deformation, external damage and internal defects of the transformer lead bushing, the mapping relationship between the defects and the characteristic information of strain, temperature, pressure and echo signals synchronously collected by the flexible 3D spiral multifunctional sensing device is constructed;

[0062] (5) The CNN or fully connected network is used to identify the characteristic information under the typical defect state, so as to realize the surface and internal defect identification of the transformer lead bushing;

[0063] (6) The flexible 3D spiral multifunctional sensing device realizes the online identification of the transformer bushing defects in combination with the deep learning algorithm (CNN, fully connected network), and evaluates the defect identification results, thereby improving the intelligent monitoring level of the transformer lead bushing.

[0064] In order to further promote the application of the flexible 3D spiral multifunctional sensing device in the monitoring of the transformer lead bushing, the practical and analysis process is shown in Figure 8 , which mainly includes the following:

[0065] (1) The flexible 3D spiral multifunctional sensing device collects the multi-parameter information of the lead bushing, and realizes the stress, displacement and deformation monitoring of the bushing;

[0066] (2) The dynamic time method is used to segment the collected multi-parameter information of the lead bushing, and the construction of the training and test sample set is completed;

[0067] (3) The principal component analysis method (PCA) is used to reduce the dimension of the multi-parameter multi-dimensional information of the lead bushing, thereby reducing the calculation amount;

[0068] (4) On the basis of dimension reduction, the characteristic information of the multi-parameter of the lead bushing under typical conditions is extracted;

[0069] (5) The least square method is used for cluster analysis of the characteristics;

[0070] (6) The parameter weight of the CNN or fully connected network is updated, and the algorithm iteration and update direction is clarified;

[0071] (7) The online analysis of the lead bushing fault is realized, the defect diagnosis and defect identification result analysis of the lead bushing is realized, and the recognition success rate of the lead bushing defect is calculated;

[0072] (8) The feature fusion analysis of the transformer lead bushing is realized through the CNN or fully connected network, and the typical fault of the transformer lead bushing is identified.

[0073] The application discloses a flexible 3D spiral sensing device and method for transformer lead bushing crack detection. The 3D spiral sensing device has the advantages of flexible planar electronic devices, can be conformally wound on the curved surface of the lead bushing, can distribute sensing information of the lead bushing in different directions and different parts, and can realize omnidirectional sensing of the lead bushing. An interface mechanics model can be used to guide the structural optimization design of the flexible 3D spiral sensing device, and ensure the conformal attachment of the flexible 3D spiral sensing device and the bushing. A fusion sensing method for transformer lead bushing crack detection based on the flexible spiral sensing is provided, which is used for fusing the multi-sensing state information of the lead bushing under different defect forms, and realizing the online detection and classification identification of the transformer lead bushing crack.

[0074] Those skilled in the art will easily understand that the above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement and improvement within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for preparing and installing a flexible spiral sensing device for detecting cracks in transformer lead bushings, characterized in that, The flexible helical sensing device includes distributed strain, temperature, pressure, and piezoelectric sensors. The strain sensors employ piezoresistive sensors, forming multiple Wheatstone bridges for collecting data on bushing surface deformation and structural vibration monitoring. The temperature sensors use thermal resistance thin-film sensors to collect data on changes in the bushing environment. The pressure sensors use thin-film capacitive sensors to collect data on the static pressure applied to the bushing. Under the action of an external excitation signal, the piezoelectric sensors generate Lamb signals, and by comparing the echo with the guided wave signal, they are used to detect defects in the cylindrical internal structure of the bushing. All functional modules are integrated on the same flexible substrate, which has a certain bending deformation capability and can be helically wound onto the surface of the transformer lead bushing, achieving conformal winding between the flexible sensing device and the tubular structure, and distributed sensing and collection of multi-parameter information of the transformer lead bushing. The conformal winding effect of a thin-film sensor and a bushing surface depends on the material properties, thickness, width, winding spacing, and curvature of the thin-film sensor. A conformal winding and structural optimization model for a flexible sensor based on the thickness, width, winding spacing, and bushing curvature of the thin-film sensor is constructed. The flexible helical sensor, being long and strip-shaped, can be helically wound onto the surface of a transformer lead bushing. The adaptability of the thickness, width, and winding spacing of the thin-film sensor to the sensor and lead bushing is calculated and analyzed. A structural optimization strategy for the sensor based on the coordinated optimization of the device width and thickness is proposed, considering the different properties of various materials used in the sensor, to solve the problem of mismatch between the mechanical and geometric structures of the sensor and the lead bushing. The spiral mounting process between the 3D spiral sensing device and the lead sleeve includes: (1) flattening sub-process; (2) rotation sub-process; (3) stretching sub-process; (4) winding sub-process; The interface contact model between the 3D spiral sensing device and the lead sleeve curved surface is based on these four processes to complete the modeling of the 3D spiral sensing device and the lead sleeve curved surface, so as to analyze the influence of the material properties, thickness, width, winding spacing and curvature of the thin film sensing on the spiral winding performance effect; The method further includes the following steps: (1) Complete the design and fabrication of different sensing units, and use microelectromechanical processing technology to complete the fabrication of flexible strain, temperature, pressure and piezoelectric sensors; (2) Construct a mechanical model of the winding between the 3D spiral sensing device and the lead sleeve, and calculate and analyze the interfacial mechanical behavior between the 3D spiral sensing device and the lead sleeve for the flattening sub-process, rotation sub-process, stretching sub-process and winding sub-process respectively. (3) Construct a conformal winding and structural optimization model for flexible sensing devices based on the thickness, width, winding spacing and sleeve curvature of thin film sensing, and calculate and analyze the adaptability of the thickness, width and winding spacing of thin film sensing to the sensing device and lead sleeve. (4) The processing of long strip flexible circuit boards ensures the matching between them and the flexible sensing unit, while also satisfying the winding adaptation process of the 3D spiral sensing device. (5) The mounting between the pins of the flexible sensing unit and the solder pads of the flexible circuit board ensures the directional mounting of the sensing unit and the solder pads on the flexible circuit board, thus completing the integration of the planar strip-shaped flexible multifunctional sensing device. (6) Packaging of multifunctional sensing device. Considering the influence of lead sleeve on the electrical performance of sensor, the application of sensing unit in lead sleeve needs to be electrically isolated. Spray an organic polymer film on the surface of multifunctional sensing device to realize the packaging of flexible multifunctional sensing device. (7) The spiral winding of the 3D spiral sensing device on the lead bushing, the flexible planar multi-functional sensing device is wound in a spiral manner on the curved surface of the transformer bushing, and is used for distributed multi-functional data acquisition of the transformer bushing.

2. A method for detecting and fusing sensing cracks in transformer lead bushings, the method being based on a flexible spiral sensing device for detecting cracks in transformer lead bushings. The flexible spiral sensing device is obtained by the preparation and installation method described in claim 1. It comprehensively analyzes strain, temperature, pressure, and piezoelectric sensing data spirally wound on the lead bushing, extracts different dynamic sensing data features from the curved surface of the lead bushing, constructs a deep learning-based bushing defect diagnosis and identification model, builds a typical defect dataset for training and identifying deep learning algorithms, and fuses multi-sensor state information of the lead bushing under different defect morphologies. Based on the flexible spiral sensing data, it achieves online detection and classification identification of transformer lead bushing cracks. Fusion sensing method for casing cracks The process includes the following steps: (1) Multi-sensor data acquisition and preprocessing: Flexible multi-functional sensing device acquires transformer lead bushing strain, temperature, pressure and guided / echo information online; (2) The strain, temperature, pressure and guided / echo information of the transformer lead bushing are segmented and effectively represented by a dynamic time window; (3) Construct a training and testing sample set for classifying defects in transformer lead bushings; (4) Extract the feature information of the transformer lead bushing status information respectively; (5) The training process of transformer lead bushing status information, training the test sample set of transformer lead bushing, and analyzing the output performance of transformer lead bushing under different defect states. (6) Testing process: The deep learning network for testing and analyzing transformer lead bushing defects was constructed using a test sample set; (7) Compare the relationship between the algorithm calculation results and the actual state of the transformer lead bushing, calculate the recognition success rate, and quickly evaluate the safety performance of the transformer lead bushing.

Images