Dynamic calibration method and system for multi-physical field data fusion of oil-immersed transformer

Abstract

The invention discloses a dynamic calibration method and system for multi-physical field data fusion of an oil-immersed transformer, and relates to the technical field of power systems, and the method comprises the following steps: collecting multi-physical field data, including temperature data, vibration data, UHF signal data and gas data, of the oil-immersed transformer in real time through a plurality of sensors; dynamically calibrating the acquired data, selecting an optimal calibration model by adopting a self-adaptive optimization algorithm, and evaluating a calibration effect based on a target function; extracting dynamic characteristics of each physical field data, and generating comprehensive characteristic parameters through a weighted fusion method; and performing intelligent analysis on the multi-physics field data by utilizing a fusion diagnosis engine, calculating a comprehensive diagnosis score and judging a fault type. According to the dynamic calibration method and system for multi-physical field data fusion of the oil-immersed transformer, the accuracy of fault diagnosis is improved through fusion of temperature, vibration, gas, UHF and other sensor data.

Description

Technical Field

[0001] This invention relates to the field of power system technology, specifically to a dynamic calibration method and system for multi-physics data fusion of oil-immersed transformers. Background Technology

[0002] Oil-immersed transformers, as crucial equipment in power systems, are widely used in power transmission and distribution networks, especially in high-voltage and high-capacity power conversion scenarios. Their primary function is heat dissipation, protection, and insulation through oil-immersed insulation. Therefore, monitoring and maintaining their operating status is paramount. However, due to their complex operating environment, long-term operation, and susceptibility to external factors, oil-immersed transformers are prone to various faults, such as overheating, partial discharge, and mechanical damage. Failure to detect and address these issues promptly can lead to equipment failure or even major safety accidents.

[0003] Traditional methods for monitoring and diagnosing transformer faults mainly rely on single detection methods, such as temperature monitoring and gas monitoring. Although these methods can monitor the operating status of equipment to a certain extent, they lack sufficient comprehensiveness, cannot accurately and timely determine the health status of equipment, and are easily affected by external interference, leading to misdiagnosis or missed diagnosis. Therefore, how to achieve intelligent monitoring, health assessment, and fault early warning of oil-immersed transformers through comprehensive analysis of multiple physical field data has become an important issue facing the power industry.

[0004] Currently, with the rapid development of technologies such as smart grids, the Internet of Things, and artificial intelligence, the means of transformer health monitoring are gradually moving towards multi-dimensional data fusion and intelligent diagnosis. In order to improve the operational safety and maintenance efficiency of transformers, there is an urgent need for a transformer fault monitoring technology based on dynamic calibration and intelligent data fusion of multi-physical field collaborative perception. Summary of the Invention

[0005] The purpose of this invention is to provide a dynamic calibration method and system for multi-physics data fusion of oil-immersed transformers, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a dynamic calibration method for multi-physics data fusion of oil-immersed transformers, the method comprising the following steps:

[0007] Multi-physics data of the oil-immersed transformer are collected in real time by multiple sensors, including temperature data, vibration data, UHF signal data and gas data;

[0008] The collected data is dynamically calibrated, an adaptive optimization algorithm is used to select the optimal calibration model, and the calibration effect is evaluated based on the objective function.

[0009] Dynamic features of each physical field data are extracted, and comprehensive feature parameters are generated through a weighted fusion method.

[0010] The fusion diagnostic engine is used to intelligently analyze multiphysics data, calculate a comprehensive diagnostic score, and determine the fault type.

[0011] Early warning information is generated based on the health status assessment results, and operation and maintenance decision support is provided.

[0012] Further, the dynamic calibration steps include:

[0013] Linear regression, support vector regression, random forest regression, or neural networks are used as candidate calibration algorithms;

[0014] The calibration accuracy can be evaluated using mean square error, weighted mean square error, or adaptive error metrics.

[0015] Combine reinforcement learning, genetic algorithms, or particle swarm optimization to dynamically adjust the calibration model parameters.

[0016] Furthermore, the feature parameter extraction steps include:

[0017] The sliding window technique was used to calculate the trend, volatility, and rate of change of temperature data.

[0018] Frequency domain analysis and time domain statistics extraction were performed on the vibration data;

[0019] Wavelet transform was performed on the UHF signal to extract partial discharge features;

[0020] Calculate the concentration change trend and key gas component ratios from the gas data.

[0021] Furthermore, the fusion diagnostic engine uses a weighted fusion formula to calculate the comprehensive diagnostic score and dynamically adjusts it through an adaptive optimization algorithm.

[0022] Furthermore, the fault type determination includes:

[0023] When the temperature trend or gas concentration is abnormal, it is determined to be an overheating fault;

[0024] When the vibration spectrum is abnormal and accompanied by a sudden change in the UHF signal, it is determined to be a mechanical fault;

[0025] When the UHF signal is significantly abnormal and the gas composition changes, it is determined to be a partial discharge fault.

[0026] When multiple physical field data are abnormal at the same time, it is determined to be a multi-factor coupling fault.

[0027] Further steps in the health status assessment include:

[0028] Health levels are determined based on comprehensive diagnostic scores;

[0029] It combines historical data to predict potential failure risks and generate early warning signals.

[0030] A dynamic calibration system for multi-physics data fusion of oil-immersed transformers, used in the aforementioned dynamic calibration method for multi-physics data fusion of oil-immersed transformers, includes:

[0031] Multiphysics data acquisition module: used for real-time acquisition of temperature, vibration, UHF and gas data;

[0032] Dynamic calibration module: used to optimize sensor data calibration and improve data accuracy;

[0033] Feature parameter extraction module: used to extract dynamic features of each physical field data;

[0034] Fusion diagnostic engine: used to comprehensively analyze multiphysics data and calculate fault probability;

[0035] Health status assessment module: used to assess the operating status of transformers and generate early warning information;

[0036] Operation and maintenance management platform: used to visualize data, fault diagnosis results and support operation and maintenance decisions.

[0037] Furthermore, the dynamic calibration module includes: an adaptive optimization algorithm unit for dynamically selecting the optimal calibration model; a constraint unit for limiting the physical rationality and consistency of data calibration; and a penalty coefficient unit for correcting calibration results that exceed the constraint range.

[0038] Furthermore, the fusion diagnostic engine includes: a single-domain preliminary diagnostic unit for calculating fault scores for each physical field data; a multi-physical field data fusion unit for weighted fusion of data from various sensors; and a fault determination unit for determining the fault probability based on a logical function.

[0039] Furthermore, the operation and maintenance management platform includes: a data visualization interface for real-time display of multi-physics data and health status; an early warning push module for sending fault warning information to operation and maintenance personnel; and an intelligent decision support module for providing maintenance suggestions and optimizing operation and maintenance strategies.

[0040] This invention provides a dynamic calibration method and system for multi-physics data fusion of oil-immersed transformers, which has the following beneficial effects:

[0041] 1. This invention improves the accuracy of fault diagnosis by fusing data from multiple sensors such as temperature, vibration, gas, and UHF.

[0042] 2. This invention enables the system to adapt to changes in equipment during long-term operation through dynamic calibration and sliding window feature extraction. It maintains high accuracy by adjusting the parameters of feature extraction and model in real time through adaptive algorithms. The health status assessment and early warning mechanism can provide real-time and intelligent decision support for operation and maintenance personnel, ensuring the reliable operation of the equipment. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the process structure of a dynamic calibration method and system for multi-physics field data fusion of an oil-immersed transformer according to the present invention. Detailed Implementation

[0044] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0045] like Figure 1 As shown, a dynamic calibration method for oil-immersed transformers using multi-physics data fusion is presented. This method includes the following steps: real-time acquisition of multi-physics data from the oil-immersed transformer using multiple sensors, including temperature data, vibration data, UHF signal data, and gas data; dynamic calibration of the acquired data, employing an adaptive optimization algorithm to select the optimal calibration model, and evaluating the calibration effect based on an objective function; extraction of dynamic features from each physical field data, and generation of comprehensive feature parameters through a weighted fusion method; intelligent analysis of the multi-physics data using a fusion diagnostic engine, calculating a comprehensive diagnostic score and determining the fault type; and generation of early warning information based on the health status assessment results, providing support for operation and maintenance decisions.

[0046] Dynamic calibration includes:

[0047] Linear regression, support vector regression, random forest regression, or neural networks are used as candidate calibration algorithms; calibration accuracy is evaluated using mean square error, weighted mean square error, or adaptive error metrics; calibration model parameters are dynamically adjusted by combining reinforcement learning, genetic algorithms, or particle swarm optimization; temperature sensors installed on transformer windings and oil temperature are used to acquire transformer temperature information in real time; vibration sensors are installed on the transformer casing to collect vibration signals during transformer operation; UHF signal sensors are used to capture partial discharge signals that may occur inside the transformer; and gas sensors are used to detect the composition and concentration of dissolved gases in transformer oil, such as hydrogen, methane, and ethylene.

[0048] Feature parameter extraction includes: calculating the trend, fluctuation, and rate of change of temperature data using the sliding window technique; performing frequency domain analysis and time domain statistical extraction on vibration data; performing wavelet transform on UHF signals to extract partial discharge characteristics; calculating the concentration change trend and key gas component ratios for gas data; and employing linear regression, support vector regression, random forest regression, or neural networks as candidate calibration algorithms. In practical applications, a suitable algorithm can be selected for preliminary calibration based on the characteristics of the data and calibration requirements.

[0049] Calibration accuracy assessment: The calibration accuracy is assessed using mean square error, weighted mean square error, or adaptive error measures; for example, the mean square error of the data before and after calibration is calculated to evaluate the calibration effect.

[0050] Dynamic adjustment of calibration model parameters: Combining reinforcement learning, genetic algorithms or particle swarm optimization to dynamically adjust calibration model parameters. Taking reinforcement learning as an example, through continuous trial and error and feedback, the parameters of the calibration model are optimized to improve calibration accuracy.

[0051] The fusion diagnostic engine uses a weighted fusion formula to calculate the comprehensive diagnostic score and dynamically adjusts it through an adaptive optimization algorithm. The fusion diagnostic engine adopts a multi-level decision fusion architecture, which includes a data preprocessing layer, a single-domain diagnostic layer, a weight dynamic adjustment layer, and a comprehensive decision layer. Among them, the data preprocessing layer standardizes the feature parameters of each physical field and unifies the dimensions to the interval [-1,1]. The single-domain diagnostic layer uses a random forest classifier (temperature / gas data), a convolutional neural network (vibration data), and an LSTM neural network (UHF signal) to calculate the single-domain fault probability, with an output range of [0,1].

[0052] Fault type determination includes: when the temperature trend or gas concentration is abnormal, it is determined to be an overheating fault;

[0053] When the vibration spectrum is abnormal and accompanied by a sudden change in the UHF signal, it is determined to be a mechanical fault; when the UHF signal is significantly abnormal and the gas composition changes, it is determined to be a partial discharge fault; when multiple physical field data are abnormal simultaneously, it is determined to be a multi-factor coupling fault; when the temperature trend and gas concentration are abnormal, it is determined to be an overheating fault. For example, when the transformer winding temperature continues to rise and the concentration of gases such as carbon monoxide and carbon dioxide in the dissolved gases in the oil increases, it is determined to be an overheating fault. When the vibration spectrum is abnormal and accompanied by a sudden change in the UHF signal, it is determined to be a mechanical fault, such as when a specific frequency abnormal peak appears in the vibration signal and the UHF signal suddenly increases, it is determined to be a mechanical fault. When the UHF signal is significantly abnormal and the gas composition changes, it is determined to be a partial discharge fault. If the UHF signal intensity increases significantly and the concentration of gases such as hydrogen and acetylene in the dissolved gases in the oil increases, it is determined to be a partial discharge fault. When multiple physical field data are abnormal simultaneously, it is determined to be a multi-factor coupling fault. The abnormal value judgment mechanism has a fixed threshold: applicable to safety critical parameters, such as oil temperature ≥105℃ (emergency fault threshold). The dynamic threshold: calculated based on historical 7-day data.

[0054] A time window sliding mechanism (default 15 minutes) is adopted. When the overlap rate of different physical field abnormal signals is ≥30% within the time window, it is judged as a synchronization abnormality.

[0055] The health status assessment steps include: classifying health levels based on comprehensive diagnostic scores; predicting potential failure risks by combining historical data; and generating early warning signals.

[0056] A calibration method for a dynamic calibration system of multi-physics data fusion for oil-immersed transformers includes:

[0057] The system includes the following modules: Multi-physics data acquisition module (for real-time acquisition of temperature, vibration, UHF, and gas data); Dynamic calibration module (for optimizing sensor data calibration and improving data accuracy); Feature parameter extraction module (for extracting dynamic features from various physical field data); Fusion diagnostic engine (for comprehensive analysis of multi-physics data and calculation of fault probability); Health status assessment module (for assessing transformer operating status and generating early warning information); Operation and maintenance management platform (for visualizing data, fault diagnosis results, and providing operation and maintenance decision support); and Dynamic calibration module, which includes: Adaptive optimization algorithm unit (for dynamically selecting the optimal calibration model); Constraint condition unit (for limiting the physical rationality and consistency of data calibration); and Penalty coefficient unit (for correcting calibration results that exceed the constraint range).

[0058] The fusion diagnostic engine includes: a single-domain preliminary diagnostic unit for calculating fault scores for each physical field data; a multi-physical field data fusion unit for weighted fusion of data from various sensors; and a fault determination unit for determining the probability of a fault based on a logical function.

[0059] The operation and maintenance management platform includes: a data visualization interface that displays multi-physics data and health status in real time. It intuitively presents transformer operating data and health status through charts, curves, and other formats; an early warning push module that sends fault warning information to operation and maintenance personnel. When a transformer malfunctions, it promptly sends warning information to operation and maintenance personnel via SMS, email, etc.; and an intelligent decision support module that provides maintenance suggestions and optimized operation and maintenance strategies. Based on the transformer's health status and fault type, it provides corresponding maintenance suggestions and operation and maintenance strategies to improve operation and maintenance efficiency. The health level classification standards are as follows:

[0060] Level I (Normal) [0,20) Routine inspection;

[0061] Level II (Caution) [20,40) Enhanced monitoring;

[0062] Level III (early warning) [40,60) maintenance will be arranged;

[0063] Level IV (fault) [60,80] requires repair within a specified period;

[0064] Level V (Critical Fault) [80,100] Immediate shutdown.

[0065] By acquiring, dynamically calibrating, extracting features, and fusing analysis of multi-physics field data, intelligent diagnosis and health status assessment of oil-immersed transformer faults were achieved. Simultaneously, the operation and maintenance management platform provided maintenance personnel with intuitive data visualization and decision support, improving the operational reliability and maintenance efficiency of the transformers.

[0066] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A dynamic calibration method for multi-physics data fusion of an oil-immersed transformer, characterized in that: The method includes the following steps: Multi-physics data of the oil-immersed transformer are collected in real time by multiple sensors, including temperature data, vibration data, UHF signal data and gas data; The collected data is dynamically calibrated, an adaptive optimization algorithm is used to select the optimal calibration model, and the calibration effect is evaluated based on the objective function. Dynamic features of each physical field data are extracted, and comprehensive feature parameters are generated through a weighted fusion method. The fusion diagnostic engine is used to intelligently analyze multiphysics data, calculate a comprehensive diagnostic score, and determine the fault type. Early warning information is generated based on the health status assessment results, and operation and maintenance decision support is provided.

2. The dynamic calibration method for multi-physics data fusion of an oil-immersed transformer according to claim 1, characterized in that, Dynamic calibration includes: Linear regression, support vector regression, random forest regression, or neural networks are used as candidate calibration algorithms; The calibration accuracy can be evaluated using mean square error, weighted mean square error, or adaptive error metrics. Combine reinforcement learning, genetic algorithms, or particle swarm optimization to dynamically adjust the calibration model parameters.

3. The dynamic calibration method for multi-physics data fusion of an oil-immersed transformer according to claim 1, characterized in that, Feature parameter extraction includes: The sliding window technique was used to calculate the trend, volatility, and rate of change of temperature data. Frequency domain analysis and time domain statistics extraction were performed on the vibration data; Wavelet transform was performed on the UHF signal to extract partial discharge features; Calculate the concentration change trend and key gas component ratios from the gas data.

4. The dynamic calibration method for multi-physics data fusion of an oil-immersed transformer according to claim 1, characterized in that, The fusion diagnostic engine uses a weighted fusion formula to calculate the comprehensive diagnostic score and dynamically adjusts it through an adaptive optimization algorithm.

5. The dynamic calibration method for multi-physics data fusion of an oil-immersed transformer according to claim 1, characterized in that, Fault type determination includes: When the temperature trend or gas concentration is abnormal, it is determined to be an overheating fault; When the vibration spectrum is abnormal and accompanied by a sudden change in the UHF signal, it is determined to be a mechanical fault; When the UHF signal is significantly abnormal and the gas composition changes, it is determined to be a partial discharge fault. When multiple physical field data are abnormal at the same time, it is determined to be a multi-factor coupling fault.

6. The dynamic calibration method for multi-physics data fusion of an oil-immersed transformer according to claim 1, characterized in that, The steps for health status assessment include: Health levels are determined based on comprehensive diagnostic scores; It combines historical data to predict potential failure risks and generate early warning signals.

7. A dynamic calibration system for multi-physics data fusion of an oil-immersed transformer, used in the dynamic calibration method for multi-physics data fusion of an oil-immersed transformer as described in any one of claims 1-6, characterized in that, include: Multiphysics data acquisition module: used for real-time acquisition of temperature, vibration, UHF and gas data; Dynamic calibration module: used to optimize sensor data calibration and improve data accuracy; Feature parameter extraction module: used to extract dynamic features of each physical field data; Fusion diagnostic engine: used to comprehensively analyze multiphysics data and calculate fault probability; Health status assessment module: used to assess the operating status of transformers and generate early warning information; Operation and maintenance management platform: used to visualize data, fault diagnosis results and support operation and maintenance decisions.

8. The dynamic calibration system for multi-physics data fusion of an oil-immersed transformer according to claim 7, characterized in that, The dynamic calibration module includes: an adaptive optimization algorithm unit for dynamically selecting the optimal calibration model; a constraint unit for limiting the physical rationality and consistency of data calibration; and a penalty coefficient unit for correcting calibration results that exceed the constraint range.

9. The dynamic calibration system for multi-physics data fusion of an oil-immersed transformer according to claim 7, characterized in that, The fusion diagnostic engine includes: a single-domain preliminary diagnostic unit for calculating fault scores for each physical field data; a multi-physical field data fusion unit for weighted fusion of data from various sensors; and a fault determination unit for determining the probability of a fault based on a logical function.

10. A dynamic calibration system for multi-physics data fusion of an oil-immersed transformer according to claim 7, characterized in that, The operation and maintenance management platform includes: a data visualization interface for real-time display of multi-physics data and health status; an early warning push module for sending fault warning information to operation and maintenance personnel; and an intelligent decision support module for providing maintenance suggestions and optimizing operation and maintenance strategies.

Images