Parameter optimization method and device for coupled transformer

Abstract

The present invention relates to the field of transformer parameter control, and in particular to a parameter optimization method and device for a coupled transformer. The method comprises the following steps: obtaining multi-dimensional real-time working monitoring parameters and real-time working environment state parameters of a coupled transformer; performing nonlinear dynamic analysis and global state perception modeling on the multi-dimensional real-time working monitoring parameters, thereby constructing a transformer global state perception model; performing multi-parameter combination definition on the multi-dimensional real-time working monitoring parameters, and performing dynamic operation simulation based on the transformer global state perception model, collecting operation simulation response data of multiple schemes; performing multi-objective dynamic balance analysis on the operation simulation response data of multiple schemes, thereby obtaining an optimal transformer working parameter combination; performing environmental dynamic parameter change analysis on the real-time working environment state parameters, and constructing a real-time environmental condition distribution field. The present invention realizes efficient and adaptively adjusted transformer parameter optimization.

Description

Technical Field

[0001] The present invention relates to the field of transformer parameter control, and in particular to a method and device for optimizing parameters of a coupled transformer. Background Art

[0002] With the continuous expansion of the scale of power systems and the application of intelligent technologies, the stability and reliability of transformer performance, as a key device in power transmission and distribution, directly affect the safe operation of the power system. Traditional transformer operation management methods mainly rely on regular inspections, manual inspections and simple fault diagnosis. However, with the increasing load demand and the complex and changeable operating environment, traditional methods have many shortcomings in dealing with transformer performance optimization and fault prediction under complex working conditions, such as delayed diagnosis, low efficiency, and poor fault prediction accuracy.

[0003] Especially under long-term, heavy-load, and high-frequency operating conditions, the various operating parameters of the transformer will be affected by internal and external factors such as load fluctuations, temperature changes, current and voltage instability, and equipment aging. The interaction of these factors will not only lead to reduced transformer efficiency, but also cause failures, and even equipment damage in extreme cases. In order to avoid this situation, the working status of the transformer needs to be monitored and analyzed comprehensively, in real time, and accurately throughout its life cycle. In modern power systems, with the improvement of automation and intelligence levels, traditional single diagnosis and parameter adjustment methods can no longer meet rapidly changing operating requirements. Therefore, there is an urgent need for a parameter optimization method based on a coupled transformer that can realize the perception, analysis, optimization, and prediction of multi-dimensional real-time working monitoring parameters. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a method and device for optimizing parameters of a coupled transformer to solve at least one of the above technical problems.

[0005] To achieve the above object, the present invention provides a method for optimizing parameters of a coupled transformer, comprising the following steps:

[0006] Step S1: obtaining multi-dimensional real-time working monitoring parameters and real-time working environment state parameters of the coupled transformer; performing nonlinear dynamic analysis and global state perception modeling on the multi-dimensional real-time working monitoring parameters, thereby constructing a transformer global state perception model;

[0007] Step S2: defining a multi-parameter combination of the multi-dimensional real-time working monitoring parameters, performing a dynamic operation simulation based on a transformer global state perception model, and collecting operation simulation response data of multiple schemes;

[0008] Step S3: performing multi-objective dynamic balance analysis on the operation simulation response data of multiple schemes, and performing local parameter optimal balance optimization, so as to obtain the optimal transformer operating parameter combination;

[0009] Step S4: performing environmental dynamic parameter change analysis on the real-time working environment state parameters, and performing environmental condition distribution fitting to construct a real-time environmental condition distribution field;

[0010] Step S5: Based on the optimal transformer working parameter combination and the real-time environmental condition distribution field, a long-term working state simulation is performed, and then a fault trend prediction is performed to obtain transformer fault prediction trend data;

[0011] Step S6: Perform local dynamic parameter tuning based on the fault prediction trend data of the transformer, and perform global synchronous optimization to build a transformer global parameter optimization engine.

[0012] The present invention obtains the transformer's multi-dimensional working monitoring parameters (such as current, voltage, temperature, vibration, etc.) and environmental state parameters (such as temperature, humidity, air pressure, etc.) in real time to grasp the transformer's operating state and external environmental impact in real time. The working state of the transformer is usually nonlinear. Through the nonlinear dynamic analysis of the monitoring parameters, the complex correlation relationship is revealed, and potential fault signs are identified, so as to provide early warning. By constructing a global state perception model, the comprehensive operating state of the transformer can be comprehensively and accurately evaluated, its working performance and potential risks can be identified, and the basis for subsequent decision-making can be optimized. By reasonably combining multiple monitoring parameters, the operating conditions of the transformer can be described more accurately, avoiding the limitations brought by a single parameter. With the help of the global state perception model, dynamic operation simulation is performed to predict the performance of the transformer under different operating conditions, including possible faults or abnormal conditions. Find the best balance point between multiple optimization objectives (such as energy efficiency, stability, life, etc.), so that the transformer can achieve the best performance in different usage scenarios. By adjusting local parameters (such as cooling system efficiency, voltage adjustment, etc.) in a targeted manner, it is ensured that the transformer performs the most efficient performance under specific working conditions. Changes in environmental conditions (such as temperature and humidity) have a significant impact on transformer performance. Timely capture of these changes can help optimize the working strategy of the transformer. By fitting the distribution field of environmental changes, the operating performance of the transformer under different environmental conditions can be accurately predicted, providing a basis for subsequent parameter optimization. Through long-term simulation, the performance of the transformer under different conditions can be fully understood, and the fault mode that occurs can be identified in advance. By performing trend analysis on the operating data, the faults that occur can be predicted in advance, providing a time window for maintenance and repair, and reducing the risk of sudden failures. Based on the fault prediction trend data, preventive maintenance can be implemented, maintenance plans can be optimized, and the overall reliability and operating life of the transformer can be improved. According to the fault prediction results, the local parameters of the transformer (such as load, current, etc.) can be adjusted in a targeted manner to optimize the operating status. On the basis of local tuning, global synchronous optimization is performed to ensure that all parameters are optimized at the overall system level, thereby improving the overall efficiency and stability of the transformer. By building a global parameter optimization engine, automatic real-time adjustment and optimization can be achieved, which not only improves the operating efficiency of the transformer, but also realizes adaptive adjustment and improves the intelligence level of the system.

[0013] Preferably, step S1 comprises the following steps:

[0014] Step S11: obtaining multi-dimensional real-time working monitoring parameters and real-time working environment status parameters of the coupled transformer;

[0015] Step S12: performing multi-dimensional comprehensive state mining on the multi-dimensional real-time working monitoring parameters to obtain multi-dimensional state characteristics of the transformer;

[0016] Step S13: performing nonlinear dynamic analysis according to the multi-dimensional state characteristics of the transformer to extract the nonlinear dynamic characteristics of the transformer;

[0017] Step S14: performing state characteristic evolution on the nonlinear dynamic characteristics of the transformer to generate a nonlinear dynamic state evolution law;

[0018] Step S15: Perform global state perception modeling based on the nonlinear dynamic state evolution law, so as to construct a transformer global state perception model.

[0019] The present invention can reveal deeper transformer status characteristics by conducting a comprehensive analysis of multi-dimensional working monitoring parameters. It not only focuses on parameter changes in a single dimension, but also comprehensively considers the interaction of multiple factors. By mining state characteristics, the dynamic behavior pattern and potential operating rules of the transformer are extracted, laying the foundation for subsequent fault prediction and parameter optimization. The working state of the transformer usually has nonlinear characteristics. By performing nonlinear dynamic analysis, complex dynamic behaviors and potential nonlinear correlations can be identified, which helps to more accurately describe the performance of the transformer and transcend the limitations of linear models. Nonlinear dynamic characteristics reveal the complex behavior patterns and potential fault trends of the system, providing a more accurate basis for subsequent optimization and fault prediction. The evolution analysis of state characteristics reveals the transformer The dynamic evolution laws under different operating conditions help predict the future state change trend of the transformer. The evolution of nonlinear dynamic characteristics helps to predict the long-term operating trend of the transformer, such as life decline, performance degradation, etc., and identify fault modes in advance. Based on the evolution laws of nonlinear dynamic states, a global state perception model is constructed to grasp the overall operation of the transformer from a macro level. Not only the influence of a single parameter or local factor is considered, but all factors and their relationships are comprehensively considered. The global state perception model can accurately predict the current state and future development trend of the transformer, provide data support for decision-making, and reduce operation and maintenance costs. After constructing the global state perception model, the dynamic optimization and regulation of the transformer can be more intelligent, and the operation strategy can be adjusted through the real-time feedback provided by the model to achieve optimal work efficiency.

[0020] Preferably, the specific steps of step S12 are:

[0021] The multi-dimensional real-time working monitoring parameters include mechanical vibration frequency, current and voltage parameters, magnetic field strength value and temperature parameters;

[0022] Identify the vibration frequency fluctuation of the mechanical vibration frequency and extract the vibration frequency fluctuation characteristics of the transformer;

[0023] Calculate the real-time transformer load based on current and voltage parameters;

[0024] Performing load time sequence variation analysis on the real-time load of the transformer to obtain load time sequence variation characteristics;

[0025] Fitting the magnetic field intensity distribution to the magnetic field intensity value to construct a magnetic field intensity distribution field;

[0026] The magnetic saturation degree is estimated based on the magnetic field intensity distribution field to obtain the real-time magnetic saturation degree value of the transformer;

[0027] Conduct temperature rise trend analysis on temperature parameters and generate transformer temperature rise trend data;

[0028] Perform discrete fitting of temperature changes on transformer temperature rise trend data to construct transformer temperature rise trend curve;

[0029] Multi-dimensional comprehensive state mining is performed on the transformer's vibration frequency fluctuation characteristics, load timing change characteristics, transformer real-time magnetic saturation value and transformer temperature rise trend curve to obtain the transformer's multi-dimensional state characteristics.

[0030] The present invention can help accurately reflect the mechanical state of the transformer by extracting the vibration frequency fluctuation characteristics, identify the factors affecting the operation stability, and thus improve the operation reliability of the equipment. By continuously monitoring the vibration frequency fluctuation, the accidental failure caused by equipment aging or improper use is reduced, and the accuracy of maintenance decisions is improved. The real-time load of the transformer is calculated by current and voltage parameters, and the working load of the transformer is accurately grasped to avoid energy waste caused by overload or low load. Real-time monitoring of the load level provides a basis for adjusting the operation mode of the transformer and improving the load efficiency, thereby optimizing the working efficiency of the transformer. Through the analysis of load time series changes, the regularity and periodicity of load changes are identified, which helps to predict the impact of load fluctuations on transformer performance. Understanding the characteristics of load time series changes provides data support for transformer operation scheduling, ensuring that the optimal working state is maintained under different load conditions. By fitting the magnetic field intensity distribution, the distribution of the transformer magnetic field is accurately understood, especially the difference in magnetic field intensity in each area, which provides support for subsequent state analysis and optimization. The magnetic field intensity distribution field helps identify the area of ​​uneven magnetic field in the transformer, which is a signal of loss, thermal effect or local overload, and helps to diagnose problems early. The degree of magnetic saturation directly affects the working efficiency and stability of the transformer. Real-time estimation of the magnetic saturation value can timely identify the impact of magnetic saturation on transformer performance. The magnetic saturation of the transformer causes power loss, excessive temperature rise or unstable operation. Through real-time monitoring of the magnetic saturation degree, measures can be taken in advance to avoid equipment damage or failure. Temperature rise trend analysis helps to monitor the temperature changes of the transformer in real time to ensure that it operates within a safe temperature range and avoid failures or shortened life due to excessive temperature rise. Through long-term monitoring of the temperature rise trend, potential overheating problems can be discovered in advance, and adjustments or maintenance can be made in time to prevent equipment damage caused by overheating. Through discrete fitting, the temperature rise trend data is converted into a temperature rise trend curve, which intuitively displays the laws and trends of temperature changes. The temperature rise trend curve helps to predict the future temperature changes of the transformer, timely discover the pattern of abnormal temperature changes, and provide support for fault prediction and optimized scheduling. Through accurate temperature rise trend curves, transformer failure or shortened life due to overheating can be avoided, and the long-term reliability of the equipment can be improved. Comprehensive multi-dimensional features such as vibration frequency, load timing, magnetic saturation degree, and temperature rise trend help to evaluate the operating status of the transformer from a global perspective and avoid the limitations of single-dimensional analysis. Multi-dimensional comprehensive analysis can more accurately identify the operating status of the transformer, especially in complex working environments, and provide more comprehensive fault warnings and performance evaluations. Comprehensive analysis based on multi-dimensional state characteristics provides a basis for intelligent optimization and scheduling of transformers, improving the working efficiency and intelligence level of transformers.

[0031] Preferably, step S2 specifically comprises the following steps:

[0032] Step S21: performing coupling correlation analysis between parameters on the multi-dimensional real-time working monitoring parameters to obtain coupling correlation characteristics between multiple parameters;

[0033] Step S22: performing normalization range analysis of each parameter based on the multi-dimensional real-time working monitoring parameters to generate a normalization range for each parameter;

[0034] Step S23: defining a combination of multiple parameters based on the normalization range of each parameter and the coupling correlation characteristics between multiple parameters, thereby obtaining multiple parameter combination schemes;

[0035] Step S24: using multiple parameter combination schemes to perform dynamic operation simulation on the coupled transformer, and collecting operation simulation response data of the multiple schemes.

[0036] The present invention predicts the overall performance of the transformer more accurately by discovering the coupling relationship between parameters. The synergy between parameters such as temperature, load, and vibration can reveal whether the transformer is operating in the best working state, which is helpful to optimize operating conditions and control strategies. The coupling relationship between parameters can reveal potential fault signals. For example, abnormal fluctuations in certain parameters are a precursor to abnormalities in other parameters. Through coupling correlation analysis, abnormalities in the system can be identified in advance, and preventive measures can be taken to avoid faults. Different monitoring parameters (such as current, voltage, vibration frequency, temperature, etc.) have different dimensions and value ranges. Normalization range analysis is performed to convert all parameters into a unified standard range, which is convenient for comprehensive comparison and analysis. This provides a unified benchmark for subsequent multi-parameter optimization and fault diagnosis. The normalization range helps to clarify the normal working range of each parameter. Once the parameter exceeds the normalization range, the abnormal situation can be quickly identified, thereby providing an accurate basis for fault diagnosis and performance optimization. Through the combination of normalization range and coupling association characteristics, multiple reasonable parameter combination schemes are defined. These schemes are optimized according to different operating scenarios and goals and can adapt to different operating requirements. The generation of multiple schemes provides a variety of choices for subsequent dynamic simulations. Through different parameter combination schemes, the optimal adjustment scheme is provided for different working conditions (such as load changes, environmental changes, etc.), thereby improving the working efficiency and stability of the transformer under various working conditions. Through dynamic operation simulation, the working performance of the transformer under different parameter combinations is reproduced in a virtual environment, and real operation response data is obtained. This simulation can avoid unnecessary burden on the equipment in actual operation. At the same time, a large number of experiments are carried out without affecting the operation of the equipment. By simulating multiple parameter combination schemes, the advantages and disadvantages of each scheme can be evaluated, thereby providing a basis for subsequent optimization, and evaluating which parameter combination schemes can make the operation of the transformer more efficient and which schemes can better reduce the risk of failure.

[0037] Preferably, step S3 specifically comprises the following steps:

[0038] Step S31: performing scheme performance calculation on the operation simulation response data of multiple schemes one by one to generate a comprehensive performance evaluation value of each scheme;

[0039] Step S32: performing multi-objective dynamic balance analysis on the operation simulation response data of the multiple schemes to generate dynamic balance characteristics of each scheme;

[0040] Step S33: extracting a better parameter combination scheme based on the dynamic balance characteristics of each scheme and the comprehensive performance evaluation value of each scheme;

[0041] Step S34: performing local parameter optimal balance optimization on the preferred parameter combination scheme, so as to obtain an optimal transformer operating parameter combination.

[0042] The present invention can quantify the performance of each parameter combination in different operating scenarios by performing performance calculations on the operation simulation response data of each solution. This provides a clear basis for comparing the advantages and disadvantages of different solutions. The comprehensive performance evaluation value takes into account multiple factors (such as temperature, load, efficiency, etc.), helping to comprehensively evaluate the performance of each solution in actual operation, including its stability, energy efficiency, failure rate and other multi-dimensional indicators. The performance optimization of the transformer is not just the adjustment of a single parameter, but the balance of multiple goals (such as power efficiency, thermal stability, load capacity, etc.). Through multi-objective dynamic balance analysis, multiple key parameters are comprehensively considered to ensure that the final solution performs well in all aspects. In the dynamic operation of the transformer, stability and efficiency are often mutually influential goals. Multi-objective dynamic balance analysis can help find the best balance between stability and efficiency, ensuring that the transformer can maintain long-term stability while operating efficiently. By combining the dynamic balance characteristics and comprehensive performance evaluation values ​​of each solution, those parameter combinations that perform better in various dimensions can be effectively screened out. This is a key step in selecting the most suitable solution from many solutions, ensuring that the multiple needs of the solution are balanced, not only considering performance improvement, but also taking into account efficiency. The optimal solution extracted is often the best trade-off between performance and risk. Through scientific feature extraction methods, the interference of subjective judgment can be reduced and the transparency and rationality of the decision-making process can be improved, which helps to find the most suitable parameter combination for the transformer among multiple competing solutions. After selecting the optimal solution, the optimal balance optimization of local parameters helps to further adjust some details to ensure that each subsystem or local parameter can operate in the optimal state, which can not only improve the overall performance of the transformer, but also improve the efficiency of each subsystem. Local optimization helps to adjust those parameters that have a greater impact on the performance of the transformer, such as current, voltage, temperature, etc., so that it can reach the best state under specific working conditions. Through this refinement step, the operating efficiency and stability of the transformer are greatly improved. By optimizing the local parameters, the optimal working parameter combination finally obtained is not only the global optimal, but also the most effective under specific working conditions, so that the transformer can remain stable and efficient in a changing operating environment.

[0043] Preferably, the specific steps of step S31 are:

[0044] The operation efficiency of each scheme is calculated for the operation simulation response data of multiple schemes one by one to obtain the operation efficiency of each scheme;

[0045] Conduct quantitative analysis of the operation cost of the operation simulation response data of multiple schemes to generate the quantitative cost value of each scheme;

[0046] Perform reliability calculation on the operation simulation response data of multiple schemes to obtain the operation reliability evaluation value of each scheme;

[0047] Multi-objective performance calculations are performed on the operating efficiency of each solution, the quantified cost value of each solution, and the operational reliability evaluation value of each solution to generate a comprehensive performance evaluation value for each solution.

[0048] The present invention calculates the operating efficiency of each solution and quantifies the energy use effect of the transformer under different working conditions. This helps to determine which solutions can most effectively convert input electrical energy into output power and reduce energy waste. The calculation of operating efficiency can help identify inefficient solutions or parameter combinations, optimize them in time, ensure that the transformer reduces energy consumption as much as possible during operation, and improve the energy utilization rate of the system. The operating cost quantitative analysis can quantify factors such as the maintenance cost, energy consumption, and failure cost of the transformer under different solutions, so as to comprehensively evaluate the economic feasibility of each solution. By evaluating the cost quantification value of each solution, the solution with excessively high cost is found and optimized to reduce unnecessary energy consumption, reduce the occurrence rate of failures or reduce the probability of equipment damage, thereby effectively reducing the overall operating cost. By calculating the reliability of each solution, the failure probability, failure mode and its impact on the overall system of each solution can be clearly identified, which helps to identify possible failure problems in advance and take measures to reduce risks Risk, reliability assessment can help identify the weak links of the system, especially the potential high-failure risk points, take preventive measures in time, and improve the overall stability and reliability of the transformer. By performing multi-objective performance calculations on operating efficiency, cost quantification and reliability, the comprehensive performance of each solution is comprehensively evaluated to ensure that all key factors (such as efficiency, cost, and reliability) are reasonably considered. There are certain conflicts between different optimization goals. For example, improving efficiency will lead to increased costs or reduced reliability. Through multi-objective calculations, reasonable trade-offs can be made for each goal to find the best comprehensive solution. The comprehensive performance evaluation value provides a clear basis for the final decision. Decision makers choose the best performance solution based on these evaluation values, which helps to make scientific decisions in operation, maintenance and management. By comprehensively considering multiple goals, it can ensure that the selected parameter combination has excellent performance at different levels, so that the overall performance of the transformer system can reach the best state.

[0049] Preferably, the specific steps of step S4 are:

[0050] Step S41: performing an environmental dynamic parameter change analysis on the real-time working environment state parameter to generate an environmental dynamic parameter change feature;

[0051] Step S42: identifying environmental mutation points based on the environmental dynamic parameter change characteristics and extracting environmental state mutation points;

[0052] Step S43: performing periodic fluctuation deep mining on the characteristics of environmental dynamic parameter changes to generate a periodic fluctuation representation of the environmental state;

[0053] Step S44: Perform environmental condition distribution fitting according to the environmental condition mutation point and the environmental condition periodic fluctuation characterization to construct a real-time environmental condition distribution field.

[0054] The present invention helps to deeply understand the changing laws of environmental factors (such as temperature, humidity, atmospheric pressure, electromagnetic interference, etc.) through the analysis of the change of environmental dynamic parameters. Through this analysis, it is possible to accurately capture the slight changes in the environment, thereby providing a scientific basis for subsequent optimization. By analyzing the changes in environmental state parameters, a mathematical model of environmental behavior is established to provide a basis for further simulation and prediction. This makes the operation of the transformer no longer a reaction to existing data, but a prediction and optimization of future changes through the model. Environmental mutation point identification effectively captures drastic or sudden changes in environmental states (such as sudden temperature rise, sudden drop in humidity, etc.). These mutation points are usually precursors to equipment failure or performance degradation, and can detect abnormalities in advance to prevent equipment from being damaged due to rapid environmental changes. By identifying the mutation points of environmental states, a more accurate early warning mechanism is designed for the transformer system. The control system of the transformer can respond when the environment changes drastically, such as adjusting the working state of the cooling system, reducing the load, etc., to avoid failures caused by sudden environmental changes. Environmental parameters (such as temperature, humidity, etc.) often have obvious periodic fluctuation characteristics. In-depth exploration of these fluctuation characteristics can reveal the long-term laws of environmental changes, such as day and night temperature differences, seasonal changes, etc. This helps to more accurately predict the operating status of the transformer in different cycles. By deeply exploring periodic fluctuations, the changes in environmental factors in different time periods can be predicted, helping to formulate more reasonable operation scheduling strategies. Taking into account the periodic fluctuations of certain environmental parameters, the operating parameters of the transformer can be adjusted in advance, such as adjusting the working intensity or load of the cooling system. The environmental condition distribution field can combine the characteristics of environmental mutation points and periodic fluctuations to construct a comprehensive environmental change model. The model can reflect the distribution changes of environmental parameters in space and time in real time, providing an accurate representation of the environmental state. By constructing the environmental condition distribution field, the environment in which the transformer is located can be monitored and predicted in real time. When certain environmental parameters tend to be extreme, the system responds in advance and adjusts the working state of the transformer to ensure that the equipment can operate smoothly even under unstable environmental conditions.

[0055] Preferably, the specific steps of step S5 are:

[0056] Step S51: performing real-time operating parameter control on the coupled transformer based on the optimal transformer operating parameter combination, and continuously collecting operating monitoring parameters;

[0057] Step S52: performing a long-term working state simulation on the working monitoring parameters according to the real-time environmental condition distribution field to obtain transformer long-term operating state simulation data;

[0058] Step S53: performing operating state parameter trend evolution on the transformer long-term operating state simulation data to obtain long-term operating state trend characteristics in advance;

[0059] Step S54: performing fault trend prediction on the long-term operation status trend characteristics to obtain fault prediction trend data of the transformer.

[0060] The present invention can ensure that the transformer can operate with the parameters that are most suitable for the current state at every moment by using the optimal transformer working parameter combination. This real-time parameter control helps the transformer maintain optimal performance under different loads and working environments. Continuously collect working monitoring parameters (such as voltage, current, temperature, load, etc.) to provide data support for subsequent analysis. These data help to understand the working state of the transformer in real time, judge whether the system is operating normally, and provide a basis for future optimization. Based on the real-time environmental condition distribution field, the long-term operating state of the transformer is simulated to predict the performance of the transformer under different environmental conditions. This not only helps to understand the current working state, but also provides data support for future operating changes. Real-time environmental data is combined with working monitoring parameters for long-term simulation to improve the accuracy of prediction. The operation of the transformer is no longer just adjusted according to the current state, but is optimized based on the prediction of future trends. Through the trend evolution analysis of long-term simulation data, the changing trend characteristics of the transformer operating state can be extracted. These trend characteristics can reveal the performance of the transformer over a long period of time, such as load changes, temperature fluctuations, and reduced efficiency, thereby providing guidance for subsequent maintenance and optimization. The evolution analysis of the trend characteristics of the long-term operating state helps to warn potential abnormalities in advance. If the system detects that a parameter is rising or falling in trend, it indicates that the equipment is aging or the performance of some components is declining, and the possibility of failure can be predicted in advance. By analyzing the trend characteristics of long-term operating status, potential failure trends can be identified and early warnings can be provided. Continuous temperature increases and gradual load increases indicate that some components or systems are close to failure. Failure trend prediction can help operators identify potential equipment failures in advance, so as to carry out preventive maintenance in time, avoid major failures and downtime, and reduce equipment maintenance and downtime costs.

[0061] Preferably, the specific steps of step S6 are:

[0062] Step S61: locating the fault location based on the fault prediction trend data of the transformer, and marking the transformer fault location point;

[0063] Step S62: performing local dynamic parameter tuning on the transformer fault location point, thereby generating dynamic tuning parameters of the fault point;

[0064] Step S63: performing global synchronous optimization of the transformer based on the dynamic tuning parameters of the fault point, thereby obtaining global synchronous optimization parameters;

[0065] Step S64: Perform iterative transfer learning on the global synchronous optimization parameters to build a transformer global parameter optimization engine.

[0066] The present invention accurately identifies the potential fault points of the transformer by analyzing the fault prediction trend data, which helps to locate the problem area in advance through the precursor signal of the fault, avoids blind inspection of the entire system, saves time and resources, and after marking the fault location point, the maintenance personnel can check the key parts of the transformer in a targeted manner, greatly improving the efficiency of fault diagnosis and processing. The rapid location of the fault also helps to reduce downtime and ensure that the transformer resumes normal operation as soon as possible. The local dynamic parameter tuning of the located fault point can adjust the operating parameters of the part according to the actual fault performance to ensure that the part resumes normal operation. The local tuning can quickly respond to changes in local faults and avoid unnecessary energy consumption caused by global adjustment. By adjusting the dynamic parameters of the fault point, the operating state of the transformer is immediately improved to avoid more serious faults. This enables the transformer to achieve rapid repair and recovery when facing local problems. On the basis of local dynamic tuning, global synchronous optimization can ensure that the entire transformer system can still be maintained after the local fault is repaired. Maintain optimal performance. Global synchronous optimization helps to coordinate the operating status of various components to avoid the optimization measures of a certain component affecting the operation of other components. Through global synchronous optimization, the imbalance effect caused by local optimization can be effectively eliminated to ensure that the various parts of the transformer can still be coordinated in the adjusted operation to avoid system instability. Through iterative transfer learning, the transformer system can continuously learn from historical fault data and operation data and adjust the optimization strategy. Each optimization is based on previous experience, making the system gradually more intelligent and adaptive. With the continuous progress of iterative transfer learning, the parameter optimization engine of the transformer will gradually achieve performance improvement. Each learning is based on new data and fault modes. The optimization engine will be able to continuously adapt to the changes of the transformer in different operating environments, thereby further improving its operating efficiency and reducing the failure rate. Iterative transfer learning can not only optimize the existing operating parameters, but also help identify new potential fault modes and trends, so that the transformer can be prevented and adjusted at an early stage to reduce the probability of failure.

[0067] In this specification, a coupled transformer parameter optimization device is provided, which is used to execute the coupled transformer parameter optimization method as described above, including:

[0068] A state perception module is used to obtain multi-dimensional real-time working monitoring parameters and real-time working environment state parameters of the coupled transformer; perform nonlinear dynamic analysis and global state perception modeling on the multi-dimensional real-time working monitoring parameters, thereby constructing a transformer global state perception model;

[0069] A dynamic operation simulation module is used to define a multi-parameter combination of the multi-dimensional real-time working monitoring parameters, perform dynamic operation simulation based on the transformer global state perception model, and collect operation simulation response data of multiple schemes;

[0070] The local parameter optimization module is used to perform multi-objective dynamic balance analysis on the operation simulation response data of multiple schemes and perform local parameter optimal balance optimization to obtain the optimal transformer operating parameter combination;

[0071] An environmental condition distribution module is used to analyze the environmental dynamic parameter changes of the real-time working environment state parameters, perform environmental condition distribution fitting, and construct a real-time environmental condition distribution field;

[0072] The fault trend prediction module is used to simulate the long-term working state based on the optimal transformer working parameter combination and the real-time environmental condition distribution field, and then perform fault trend prediction to obtain the transformer fault prediction trend data;

[0073] The global parameter optimization module is used to perform local dynamic parameter tuning based on the fault prediction trend data of the transformer, and perform global synchronous optimization to build a transformer global parameter optimization engine.

[0074] The present invention can fully perceive the operating state of the transformer by acquiring real-time working monitoring parameters and environmental state parameters, which provides reliable data support for further fault detection, performance optimization and safety warning. Nonlinear dynamic analysis helps to identify the complex interactive relationship between parameters, so that the global state perception model is not limited to static information, but can also reflect dynamic changes in real time and accurately predict the long-term operating trend of the transformer. By combining multi-dimensional parameters, the performance of the transformer under different operating conditions can be simulated, and then the best parameter combination scheme can be found. The module can simulate the dynamic response of the transformer under different load and environmental conditions, evaluate the effects of various operating schemes, discover potential problems in advance and pre-optimize. Through multi-objective dynamic balance analysis, multiple performance indicators (such as efficiency, load, temperature, etc.) are optimized at the same time to ensure that the various parameters of the transformer are kept in the best balance during operation. Through the optimization and adjustment of local parameters, the bottleneck problem under certain working conditions can be effectively solved, thereby improving the overall operating efficiency and stability of the transformer. The operation of the transformer is not only affected by its own parameters, but also by the strong influence of the external environment (such as temperature, humidity, etc.). By analyzing the dynamic changes of the environment, we can better understand the impact of the environment on the operation of the transformer and make environmental adaptability adjustments. By fitting the distribution field of environmental conditions, we can predict the impact of environmental changes on the transformer, provide a basis for subsequent fault prediction and optimization solutions, and enhance the transformer's adaptability to environmental changes. Based on the optimal working parameters and environmental changes, long-term state simulation is performed to identify potential fault trends in advance, which provides a basis for preventive maintenance of the transformer and helps to avoid sudden equipment failures. By predicting fault trends, maintenance personnel can take timely measures to overhaul or adjust the equipment to avoid long-term downtime and ensure the stable operation of the power system. By combining fault prediction trend data to optimize local and global parameters simultaneously, it is ensured that the transformer always maintains the best performance throughout its life cycle, reduces energy loss and faults, and makes dynamic adjustments based on fault prediction data to avoid faults to the greatest extent possible. At the same time, it optimizes the working efficiency and service life of the transformer. The global optimization engine automatically adjusts transformer parameters to reduce manual intervention, improves operation and maintenance efficiency, and ensures that the transformer can continue to operate in the best state. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 A schematic flow chart of the steps of a method for optimizing parameters of a coupled transformer according to the present invention;

[0076] Figure 2 Detailed implementation flow chart of step S1;

[0077] Figure 3 Detailed implementation flow chart of step S2;

[0078] Figure 4Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0079] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0080] The present application example provides a method and device for optimizing the parameters of a coupled transformer. The execution subjects of the method and device for optimizing the parameters of the coupled transformer include but are not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0081] See also Figures 1 to 4 The present invention provides a method for optimizing the parameters of a coupled transformer, and the method for optimizing the parameters of a coupled transformer comprises the following steps:

[0082] Step S1: obtaining multi-dimensional real-time working monitoring parameters and real-time working environment state parameters of the coupled transformer; performing nonlinear dynamic analysis and global state perception modeling on the multi-dimensional real-time working monitoring parameters, thereby constructing a transformer global state perception model;

[0083] Step S2: defining a multi-parameter combination of the multi-dimensional real-time working monitoring parameters, performing a dynamic operation simulation based on a transformer global state perception model, and collecting operation simulation response data of multiple schemes;

[0084] Step S3: performing multi-objective dynamic balance analysis on the operation simulation response data of multiple schemes, and performing local parameter optimal balance optimization, so as to obtain the optimal transformer operating parameter combination;

[0085] Step S4: performing environmental dynamic parameter change analysis on the real-time working environment state parameters, and performing environmental condition distribution fitting to construct a real-time environmental condition distribution field;

[0086] Step S5: Based on the optimal transformer working parameter combination and the real-time environmental condition distribution field, a long-term working state simulation is performed, and then a fault trend prediction is performed to obtain transformer fault prediction trend data;

[0087] Step S6: Perform local dynamic parameter tuning based on the fault prediction trend data of the transformer, and perform global synchronous optimization to build a transformer global parameter optimization engine.

[0088] The present invention obtains the transformer's multi-dimensional working monitoring parameters (such as current, voltage, temperature, vibration, etc.) and environmental state parameters (such as temperature, humidity, air pressure, etc.) in real time to grasp the transformer's operating state and external environmental impact in real time. The working state of the transformer is usually nonlinear. Through the nonlinear dynamic analysis of the monitoring parameters, the complex correlation relationship is revealed, and potential fault signs are identified, so as to provide early warning. By constructing a global state perception model, the comprehensive operating state of the transformer can be comprehensively and accurately evaluated, its working performance and potential risks can be identified, and the basis for subsequent decision-making can be optimized. By reasonably combining multiple monitoring parameters, the operating conditions of the transformer can be described more accurately, avoiding the limitations brought by a single parameter. With the help of the global state perception model, dynamic operation simulation is performed to predict the performance of the transformer under different operating conditions, including potential faults or abnormal conditions. Find the best balance point between multiple optimization objectives (such as energy efficiency, stability, life, etc.), so that the transformer can achieve the best performance in different usage scenarios. By adjusting local parameters (such as cooling system efficiency, voltage adjustment, etc.) in a targeted manner, it is ensured that the transformer performs the most efficient performance under specific working conditions. Changes in environmental conditions (such as temperature and humidity) have a significant impact on transformer performance. Timely capture of these changes can help optimize the working strategy of the transformer. By fitting the distribution field of environmental changes, the operating performance of the transformer under different environmental conditions can be accurately predicted, providing a basis for subsequent parameter optimization. Through long-term simulation, the performance of the transformer under different conditions can be fully understood, and the fault mode that occurs can be identified in advance. By performing trend analysis on the operating data, the faults that occur can be predicted in advance, providing a time window for maintenance and repair, and reducing the risk of sudden failures. Based on the fault prediction trend data, preventive maintenance can be implemented, maintenance plans can be optimized, and the overall reliability and operating life of the transformer can be improved. According to the fault prediction results, the local parameters of the transformer (such as load, current, etc.) can be adjusted in a targeted manner to optimize the operating status. On the basis of local tuning, global synchronous optimization is performed to ensure that all parameters are optimized at the overall system level, thereby improving the overall efficiency and stability of the transformer. By building a global parameter optimization engine, automatic real-time adjustment and optimization can be achieved, which not only improves the operating efficiency of the transformer, but also realizes adaptive adjustment and improves the intelligence level of the system.

[0089] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a method for optimizing the parameters of a coupled transformer according to the present invention. In this example, the steps of the method for optimizing the parameters of the coupled transformer include:

[0090] Step S1: obtaining multi-dimensional real-time working monitoring parameters and real-time working environment state parameters of the coupled transformer; performing nonlinear dynamic analysis and global state perception modeling on the multi-dimensional real-time working monitoring parameters, thereby constructing a transformer global state perception model;

[0091] In this embodiment, in order to ensure comprehensive monitoring of the operating status of the transformer, a variety of sensors are configured, such as temperature sensors, vibration sensors, current transformers and oil pressure sensors. These sensors should be distributed in key parts of the transformer, such as windings, cooling systems and casings. The data acquisition frequency is set, usually 1 to 10 times per second, to obtain real-time monitoring data. In certain circumstances (such as load changes or sudden failures), the sampling frequency can be increased to capture instantaneous data and collect state parameters related to the operating environment of the transformer, such as ambient temperature, humidity, air pressure, etc. These environmental factors have a significant impact on the performance of the transformer, so real-time monitoring must be carried out. Environmental sensors are used to obtain these data and ensure their compatibility with the transformer monitoring system. Before performing nonlinear dynamic analysis, the acquired multi-dimensional data is preprocessed, including operations such as denoising, normalization and interpolation. A low-pass filter can be used to remove high-frequency noise to ensure the smoothness and accuracy of the data. Through data cleaning, missing values ​​and outliers are removed to improve the reliability of subsequent analysis. Nonlinear dynamic models, such as state space models or nonlinear regression models, are used to analyze multi-dimensional monitoring parameters. These models can handle complex nonlinear relationships and are suitable for the dynamic operation of transformers. In order to, combine time series analysis, model the changes of parameters over time, identify potential dynamic features and patterns, use recurrent neural networks (RNN) or long short-term memory networks (LSTM) to capture the dynamic characteristics of time series data, build a global state perception model based on the previous nonlinear dynamic analysis results, the model should integrate all monitoring parameters and environmental conditions to fully reflect the operating status of the transformer, select suitable modeling tools and frameworks, such as TensorFlow or PyTorch, design deep learning models to improve the accuracy and adaptability of the model, use the acquired multi-dimensional monitoring data to train the global state perception model, divide the data into training set, validation set and test set to ensure the generalization ability of the model on different data sets, optimize the model performance through cross-validation and hyperparameter tuning to ensure that it can accurately predict the performance of the transformer under different working conditions, evaluate the performance of the global state perception model, use indicators such as mean square error (MSE), determination coefficient (R²), etc. for evaluation, and further adjust the model structure or parameters according to the evaluation results to improve the prediction accuracy, record the structure, parameters and evaluation results of the model, generate a state perception model report, and provide a basis for subsequent monitoring and control.

[0092] Step S2: defining a multi-parameter combination of the multi-dimensional real-time working monitoring parameters, performing a dynamic operation simulation based on a transformer global state perception model, and collecting operation simulation response data of multiple schemes;

[0093] In this embodiment, key parameters are selected from the acquired multi-dimensional real-time working monitoring parameters for combination definition. Common parameters include temperature (such as winding temperature, oil temperature), current (such as load current, no-load current), pressure (such as oil pressure), vibration (such as shaft vibration, housing vibration), etc. These parameters are classified, for example, temperature-related parameters are classified into one category, and current-related parameters are classified into another category, so as to facilitate subsequent combination. The combination method in combinatorial mathematics is used to define different parameter combinations and different working states (such as normal, overload, short circuit, etc.), and a corresponding parameter combination is selected for each state. Each combination should contain at least three parameters. Ensure that the operating status of the transformer can be fully reflected. Use matrix form to record these combinations for subsequent simulation and analysis. Build a dynamic operation simulation model based on the transformer global state perception model. Use MATLAB Simulink or other simulation tools to create the model. Ensure that the model can handle multi-parameter inputs. Use the previously defined multi-parameter combination as input and set the corresponding initial conditions and boundary conditions. If the selected parameter combination includes temperature, load current and oil pressure, enter the initial values ​​of these parameters in the model. Design multiple simulation schemes to simulate different working conditions. Scheme 1 simulates normal working conditions, Scheme 2 simulates overload conditions, and Scheme 3 simulates short circuit conditions. For each scheme, set different running times (such as 1 hour, 3 hours, etc.) and record key events that occur during the simulation (such as temperature rise, vibration enhancement, etc.). Start the dynamic operation simulation, simulate according to the set scheme, monitor the output data in real time, and use a data acquisition system (such as LabVIEW) to record the simulation results, including output parameters such as temperature, current, vibration, etc. During the simulation process, ensure that the monitoring data is updated in real time to capture instantaneous changes. If the temperature exceeds the set threshold at a certain time point, the system should immediately record the event and issue an alarm. Conduct a preliminary analysis of the response data, identify the key performance indicators (such as temperature changes, vibration amplitude, etc.) under each scheme, use statistical analysis methods (such as mean, standard deviation) to conduct descriptive analysis of the data, and draw a change curve chart of the response data for each scheme to intuitively display the parameter changes under different states. Record the operation simulation response data of all schemes, including input parameters, output results, and timestamps of key events, generate data logs to provide a basis for subsequent analysis and decision-making, and prepare an operation simulation report to describe in detail the input and output of each scheme and its impact on the operating status of the transformer, providing data support for subsequent optimization and control.

[0094] Step S3: performing multi-objective dynamic balance analysis on the operation simulation response data of multiple schemes, and performing local parameter optimal balance optimization, so as to obtain the optimal transformer operating parameter combination;

[0095] In this embodiment, data integrity and consistency are ensured, input parameters (such as load current, temperature, vibration, etc.) and output results (such as efficiency, failure rate, energy consumption, etc.) of each solution are recorded, response data are preprocessed, outliers and noise are removed, the accuracy of data analysis is ensured, statistical methods (such as Z-score standardization) are used to identify and process outliers, key performance indicators (KPIs) such as operating efficiency, temperature fluctuation, vibration amplitude, etc. are extracted from the sorted response data, these indicators are used as reference points for multi-objective optimization, and weights of performance indicators are formulated. Based on the working requirements and actual application scenarios of the transformer, efficiency and stability are given higher weights. weight, while the failure rate is given a lower weight. A multi-objective optimization algorithm (such as non-dominated sorting genetic algorithm NSGA-II or particle swarm optimization PSO) is used for dynamic balance analysis. These algorithms can handle the trade-offs between multiple objectives and find the best solution. The objective function is defined, which usually includes minimizing energy consumption, maximizing efficiency, minimizing failure rate, etc. These objective functions are integrated together using mathematical expressions. The extracted feature data is input into the multi-objective optimization algorithm for dynamic balance analysis. The system will evaluate the advantages and disadvantages of different solutions according to the set objective function, record the changes in performance indicators in each round of iteration, and generate a Pareto chart for visualization. The trade-off relationship between different solutions, through the Pareto frontier, identify the best and suboptimal solutions, select local optimization algorithms (such as gradient descent or Newton's method), refine and optimize the preferred solutions identified in the multi-objective dynamic balance analysis, these algorithms effectively adjust parameters to further improve performance, determine the range and steps of optimization parameters, such as setting reasonable adjustment ranges for key parameters such as current, temperature, and load, apply local optimization algorithms to the selected preferred solutions, gradually adjust parameters, observe changes in performance indicators, record the results after each adjustment, and compare them with previous results, set stop conditions, such as when the performance indicator improvement is less than the set When the optimization threshold is reached, the optimization process ends. At this time, the final parameter combination is recorded, the optimal transformer working parameter combination is verified, and retested using simulation tools (such as MATLAB or Simulink) to ensure its effectiveness under actual operating conditions. The performance indicators before and after optimization are compared, the effect of optimization is evaluated, and it is checked whether energy consumption is reduced, efficiency is improved, and failure rate is reduced. An optimization result report is generated, recording the input, output, and performance indicators of each solution in detail, describing the problems encountered in the optimization process and their solutions, and including the optimal parameter combination and its corresponding performance indicators in the report as a reference for subsequent transformer operation and maintenance.

[0096] Step S4: performing environmental dynamic parameter change analysis on the real-time working environment state parameters, and performing environmental condition distribution fitting to construct a real-time environmental condition distribution field;

[0097] In this embodiment, real-time working environment status parameter data, such as temperature, humidity, air pressure, noise, etc., are collected. These data should be collected through environmental sensors, and the data sampling frequency is set (such as once per second) to ensure that sufficient time series data is obtained. Data cleaning is performed to remove missing values, outliers and redundant data. Interpolation methods (such as linear interpolation and spline interpolation) are used to process missing data to maintain data continuity and integrity. Time series analysis methods are used to analyze the dynamic changes of the collected environmental parameters. An autoregressive integrated moving average (ARIMA) model or a long short-term memory network (LSTM) is used to capture the time dependency and trend of the data. An appropriate time window is set to analyze the environmental parameters in different time periods (such as day and night). The changing characteristics of the number of parameters are analyzed to identify periodic fluctuations and mutation points. The temperature changes differently during the day and at night. The changes in humidity in different seasons should also be considered. The analysis results are recorded, including the trend, periodic changes and abnormal conditions of each environmental parameter. An environmental dynamic change analysis report is generated to provide a basis for subsequent distribution fitting. Visual tools (such as Matplotlib or Seaborn) are used to draw environmental parameter change curves to intuitively display the changes of different parameters over time, which is convenient for identifying potential problems. According to the characteristics of environmental parameters, appropriate distribution models are selected for fitting. Methods such as Gaussian mixture model (GMM), multivariate normal distribution or Kriging interpolation are used to describe the spatial distribution of environmental states and determine the parameters of the model. For example, Such as the number of components and covariance structure in the Gaussian mixture model to ensure the accuracy and applicability of the model. The collected environmental parameter data are input into the selected distribution model for fitting. The scikit-learn library in Python or the MASS package in R language is used for model training and fitting. The model parameters are estimated through maximum likelihood estimation (MLE) or Bayesian inference method to obtain the best fitting effect. The model parameters are gradually adjusted, and the fitting effect is evaluated through cross-validation. The accuracy of the fitting results is evaluated using indicators such as mean square error (MSE) and determination coefficient (R²). The difference between the fitted data and the actual data is compared to identify the deficiencies of the model. The parameter changes during the fitting process are recorded to generate the environment. The condition distribution fitting report describes the model selection, parameter settings and fitting effects, providing a basis for subsequent applications. Based on the fitting results, the environmental condition distribution field is constructed, and the spatial distribution of environmental parameters is visualized using geographic information system (GIS) tools or visualization software (such as Matplotlib, Plotly). In the distribution field, the environmental parameters of different areas are represented by colors, contour lines, etc., to intuitively display the changes in environmental conditions. A real-time monitoring mechanism is established, and the environmental condition distribution field is updated regularly. Whenever new data is collected, it is automatically updated dynamically to ensure that the distribution field reflects the current environmental status, record the changes in key parameters during the update process, and generate a real-time environmental condition distribution field report to provide support for subsequent monitoring and decision-making.

[0098] Step S5: Based on the optimal transformer working parameter combination and the real-time environmental condition distribution field, a long-term working state simulation is performed, and then a fault trend prediction is performed to obtain transformer fault prediction trend data;

[0099] In this embodiment, the previously constructed global state perception model is used, combined with the optimal transformer operating parameter combination (such as temperature, current, load, etc.), to build a long-term working state simulation model, select MATLAB Simulink, Ansys or other professional simulation tools for modeling, input the data of the real-time environmental condition distribution field as external conditions to ensure that the simulation reflects the real working environment and that the model can handle multi-dimensional inputs, including the time-varying nature of parameters, determine the simulation time range, usually set to 72 hours to one week, to capture changes in long-term operating conditions, the simulation can be set to sample once an hour, record key parameters such as temperature, load current, vibration, etc., set different load conditions during the simulation process, and conditions (such as full load, half load and no load), and simulate changes in environmental conditions (such as temperature fluctuations and humidity changes) to ensure that various potential working conditions are covered, start long-term working state simulation, monitor various output parameters in real time, use data acquisition systems (such as LabVIEW) or built-in recording functions to record the state data at each time point during the simulation process. The recorded parameters include but are not limited to temperature, load current, oil pressure, vibration, etc., to ensure that the data fully reflects the long-term operating status of the transformer, generate data logs for subsequent analysis, select a suitable fault prediction model based on the long-term simulation results, use machine learning methods (such as random forests, support vector machines) or deep learning methods (such as LSTM) to predict fault trends, and collect Historical fault data to ensure that the model can learn the characteristics and laws of fault occurrence. Combine historical data with simulation results to form a complete data set. Extract key features from simulation data, such as temperature change rate, vibration amplitude, load fluctuation, etc., and use these features as input to the model. Set target variables, such as the time window and type of fault occurrence, to train the fault prediction model. Ensure the accuracy and consistency of data labels. Use the extracted features and historical fault data to train the fault prediction model. Use cross-validation to evaluate the performance of the model to ensure that the model has good generalization ability. Adjust model hyperparameters (such as learning rate, number of trees, etc.) to improve prediction accuracy. Monitor the model during training. Loss value and accuracy, carry out necessary iterative optimization, input the state data obtained by long-term simulation into the trained fault prediction model, generate fault prediction trend data, record the fault probability and fault type at each time point, set the prediction time, for example, predict the probability of fault occurrence in the next 72 hours, generate the corresponding trend curve, analyze the prediction results, identify high-risk time periods and fault types, use visualization tools (such as Matplotlib, Seaborn) to draw the fault prediction trend graph to intuitively show the change of fault probability over time, generate a fault prediction trend report, record the prediction results, model parameters and their impact on the safe operation of the transformer in detail, and provide decision support for subsequent maintenance.

[0100] Step S6: Perform local dynamic parameter tuning based on the fault prediction trend data of the transformer, and perform global synchronous optimization to build a transformer global parameter optimization engine.

[0101] In this embodiment, key parameters such as temperature, load current, oil pressure, vibration, etc. are extracted from the fault prediction trend data, the correlation between the change trend of these parameters and the occurrence of faults is analyzed, and several key parameters with the greatest impact on the faults are identified. Statistical analysis methods (such as correlation analysis and principal component analysis) are used to find the parameters with the highest correlation with the fault risk. If the positive correlation between temperature and the occurrence of faults reaches 0.8 or above, the temperature will be regarded as a key tuning parameter. For the identified key parameters, the target of local dynamic tuning is set. If it is found that high temperature is the main factor causing the fault, the goal is to reduce the temperature to a safe range. The scope and amplitude of the tuning are determined to ensure the safety and effectiveness of the tuning process, and the adjustment of the oil pressure is set. The range is 50-70psi, and the temperature is controlled between 70-85℃. Fuzzy control or PID control algorithm is used to perform real-time dynamic tuning of key parameters. The controller is set to monitor parameter changes in real time and adjust according to the set goals. If the temperature exceeds 85℃, the control system will automatically increase the coolant flow rate. If the temperature is lower than 70℃, the coolant flow rate will be reduced. A data recording system is used to record each parameter change during the adjustment process in real time. On the basis of local tuning, a global synchronous optimization model is constructed. Multi-objective optimization algorithms (such as genetic algorithms and particle swarm optimization algorithms) are used to comprehensively consider the mutual influence of all key parameters and determine the objective function of global optimization, such as minimizing energy consumption, maximizing equipment efficiency, and reducing failure rate. Set the weights of different objectives to the corresponding priorities to ensure the effectiveness of global optimization. Select the parameters to be optimized, including temperature, load current, oil pressure, vibration, etc. Set the optimization range of each parameter. The temperature is set between 70-85℃ and the load current is 50-100A. Use historical data and real-time monitoring data to estimate the impact of each parameter under different conditions to ensure that the parameter adjustment during the optimization process is scientific and reasonable. Start the global optimization algorithm, input the parameter values ​​obtained after local tuning, perform global optimization calculations, record the objective function values ​​and the changes of each parameter in each iteration, and find the optimal solution by evaluating the performance of different parameter combinations in each round of iteration. Visualize the results of the optimization process (such as Pareto The optimization engine is designed to visualize the trade-offs between different schemes, integrate the results of local tuning and global optimization into an optimization engine, use a real-time monitoring system (such as LabVIEW or MATLAB) to build a parameter optimization engine, ensure that it can process input data in real time and automatically adjust parameters, design the user interface of the optimization engine so that operators can easily monitor the current parameter status, adjust the optimization strategy, view real-time feedback, start the optimization engine, monitor the working status of the transformer in real time, and adjust parameters according to real-time data, record key data during the operation of the optimization engine to evaluate its performance, regularly evaluate the effect of the optimization engine, check indicators such as failure rate, energy consumption and equipment efficiency, and ensure its effectiveness and stability in long-term operation.

[0102] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0103] Step S11: obtaining multi-dimensional real-time working monitoring parameters and real-time working environment status parameters of the coupled transformer;

[0104] Step S12: performing multi-dimensional comprehensive state mining on the multi-dimensional real-time working monitoring parameters to obtain multi-dimensional state characteristics of the transformer;

[0105] Step S13: performing nonlinear dynamic analysis according to the multi-dimensional state characteristics of the transformer to extract the nonlinear dynamic characteristics of the transformer;

[0106] Step S14: performing state characteristic evolution on the nonlinear dynamic characteristics of the transformer to generate a nonlinear dynamic state evolution law;

[0107] Step S15: Perform global state perception modeling based on the nonlinear dynamic state evolution law, so as to construct a transformer global state perception model.

[0108] In this embodiment, sensors, including temperature sensors, pressure sensors, current sensors, and vibration sensors, are installed at key parts of the transformer to ensure that the operating status of the transformer can be fully monitored. High-precision sensors (such as PT100 temperature sensors and Hall effect current sensors) are used to improve the accuracy of the data. A data acquisition system (such as a PLC or embedded system) is configured, and the sampling rate is set to 10 times per second to obtain real-time monitoring data. Ensure that the data acquisition system is compatible with the sensor and has good data storage capabilities. During the operation of the transformer, multi-dimensional working monitoring parameters, including current, voltage, temperature, humidity, vibration, etc., are collected in real time. The data is stored in CSV or database format using a data recorder for subsequent analysis. At the same time, relevant environmental state parameters, such as ambient temperature, humidity, and air pressure, are obtained to analyze the impact of the environment on the working status of the transformer. Regularly perform quality checks on the collected data to ensure the integrity and accuracy of the data. Set reasonable thresholds to automatically identify and filter abnormal values, such as alarming when abnormal temperature fluctuations exceed the set range. Clean and preprocess the original monitoring data, including removing duplicate values, filling missing values, and standardizing. The Pandas library in Python is used for data processing to ensure data consistency and comparability. The Z-score standardization method is applied to transform the data into a standard normal distribution, so that the features are comparable and the analysis results are not affected by different dimensions. Dimensionality reduction techniques such as principal component analysis (PCA) or independent component analysis (ICA) are used to comprehensively analyze the multi-dimensional monitoring parameters and extract the main state characteristics. The number of principal components is set to 95% of the variance explanation rate to determine the important features. The extracted features are classified by combining machine learning methods (such as cluster analysis) to identify different working states. The K-means clustering algorithm is used for state classification, and the number of clusters is set to 3 (normal, warning, fault). The mined multi-dimensional state features are stored in a structured format to ensure that they can be used for subsequent analysis and modeling. Use visualization tools (such as Matplotlib or Seaborn) to draw state feature distribution diagrams to facilitate intuitive understanding of the changing trends of features. Select appropriate nonlinear dynamic models, such as nonlinear autoregressive models (NAR) or long short-term memory networks (LSTM) to adapt to the nonlinear characteristics of transformer states. Set the model input to the extracted state features and the output to the dynamic response of the transformer. Prepare a training data set to match the historical state features with the corresponding dynamic response data to facilitate model learning. Use the back propagation algorithm to train the nonlinear dynamic model, set the batch size to 32, the learning rate to 0.001, and the training rounds to 200. Evaluate the model performance through cross-validation to ensure that it can accurately capture the dynamic characteristics of the transformer. Use indicators such as mean square error (MSE) and R² to evaluate the model fitting effect to ensure that the extraction of nonlinear dynamic features is highly accurate.Perform time series analysis on the extracted nonlinear dynamic features and establish an evolution model. Use a nonlinear regression model or a state space model (such as Kalman filtering) to capture the evolution of features. Set a time window (such as 24 hours) and perform sliding window analysis on the dynamic features to ensure that the changes of features over time can be captured. Use the model to analyze the changes in key parameters during the evolution process and identify the laws of feature evolution. Determine the evolution path of features under different states by calculating the rate of change and trend line. Record the time series data of the evolution law to form a visualization of the feature evolution to help understand the changes of dynamic features over time. Compare the extracted evolution law with the actual monitoring data to verify the validity of the law. Use hypothesis testing methods (such as t-tests) to evaluate the significance of the evolution law. Apply the evolution law to the transformer state prediction model to improve the prediction ability of future states. Select a suitable modeling method (such as a neural network or Bayesian network) to build a global state perception model. The input of the model is the extracted dynamic features and evolution law, and the output is the global state evaluation of the transformer. Set the learning objectives of the model, including state prediction, fault diagnosis, and health assessment, to ensure that the model has comprehensive monitoring capabilities. Use historical data to train the global state perception model, set the batch size to 64, the learning rate to 0.001, and the training rounds to 150. Evaluate the performance of the model through cross-validation to ensure its generalization ability on unseen data. Use indicators such as confusion matrix and ROC curve to evaluate the classification performance of the model to ensure that the model can accurately identify different states. Apply the trained global state perception model to real-time monitoring to evaluate the state of the transformer in real time. Use the information output by the model for decision support, such as maintenance recommendations and fault warnings. Review the model performance regularly, and update and optimize the model in combination with new monitoring data to ensure that the model always maintains efficient state perception capabilities.

[0109] In this embodiment, the specific steps of step S12 are:

[0110] The multi-dimensional real-time working monitoring parameters include mechanical vibration frequency, current and voltage parameters, magnetic field strength value and temperature parameters;

[0111] Identify the vibration frequency fluctuation of the mechanical vibration frequency and extract the vibration frequency fluctuation characteristics of the transformer;

[0112] Calculate the real-time transformer load based on current and voltage parameters;

[0113] Performing load time sequence variation analysis on the real-time load of the transformer to obtain load time sequence variation characteristics;

[0114] Fitting the magnetic field intensity distribution to the magnetic field intensity value to construct a magnetic field intensity distribution field;

[0115] The magnetic saturation degree is estimated based on the magnetic field intensity distribution field to obtain the real-time magnetic saturation degree value of the transformer;

[0116] Conduct temperature rise trend analysis on temperature parameters and generate transformer temperature rise trend data;

[0117] Perform discrete fitting of temperature changes on transformer temperature rise trend data to construct transformer temperature rise trend curve;

[0118] Multi-dimensional comprehensive state mining is performed on the transformer's vibration frequency fluctuation characteristics, load timing change characteristics, transformer real-time magnetic saturation value and transformer temperature rise trend curve to obtain the transformer's multi-dimensional state characteristics.

[0119] In this embodiment, an acceleration sensor is installed at a key position of the transformer, and the sampling frequency is set to 1kHz to obtain high-resolution vibration data. The sensor should have good frequency response characteristics to cover the frequency range (0-500Hz) that occurs during the operation of the transformer. Under normal operating conditions of the transformer, the vibration signal is continuously collected and the data is recorded for at least 5 minutes to provide sufficient information for subsequent analysis. The collected vibration signal is analyzed in the frequency domain using the fast Fourier transform (FFT) to extract the vibration frequency component. The window length is set to 1024 points and the overlap rate is 50% to ensure good frequency resolution. The spectrum diagram is analyzed to identify the main frequency The vibration frequency fluctuation range and center frequency are recorded, and the statistical characteristics such as the mean, standard deviation and kurtosis of the vibration frequency are calculated to quantify the fluctuation characteristics of the vibration frequency, visualize the characteristic change trend, and help understand the change of the vibration state. The extracted characteristic data are recorded in the database for subsequent comprehensive analysis and comparison. The current transformer (CT) and voltage transformer (VT) are installed to monitor the input current and voltage of the transformer respectively. The sampling frequency is set to 1kHz to ensure the accuracy of real-time data. The current and voltage data are recorded to ensure that they are collected under different load changes so that the load condition of the transformer can be fully reflected. The formula P=U×I× , calculate the real-time load of the three-phase transformer, where P is power, U is voltage, and I is current. Calculate the active power, reactive power, and apparent power of the transformer, and record the calculation results for subsequent analysis. Perform time series processing on the real-time load data, use the sliding window method (for example, a 10-minute window) to calculate the load mean, peak value, and valley value within the window, use the time series analysis method to calculate the load change rate and fluctuation amplitude to capture the dynamic characteristics of load changes, identify the peak and valley periods of the load, record the time series characteristics of load changes, use MATLAB or Python for data visualization, display the load change curve, generate a load change feature report, and record the periodicity and regularity of load fluctuations for subsequent analysis. Hall sensors are installed at different positions of the transformer to measure the magnetic field strength. The sampling frequency is set to 1 Hz to obtain magnetic field data at different positions. The magnetic field strength data is recorded for at least 10 minutes to ensure that the operation cycle of the transformer is covered. The collected magnetic field strength data is fitted using interpolation methods (such as spline interpolation) to construct a magnetic field strength distribution field. According to the spatial coordinates of the measurement points, a three-dimensional magnetic field distribution model is generated. Visualization tools (such as Matplotlib or ParaView) are used to display the magnetic field strength distribution, analyze the concentration and dispersion characteristics of the magnetic field, set the magnetic saturation threshold, which is usually determined by the design parameters of the transformer, calculate the ratio of the actual magnetic field strength to the saturation threshold, record the real-time magnetic saturation degree of the transformer, and evaluate its impact on the transformer performance. The influence of magnetic saturation degree can be visualized over time for monitoring and early warning. Temperature sensors can be installed at different positions of the transformer to monitor the temperature changes of the transformer in real time. The sampling frequency can be set to 1 Hz to ensure the real-time nature of the temperature data. The temperature data can be recorded for 10 minutes to ensure that the temperature changes under different loads and working conditions are covered. The temperature data can be analyzed using linear regression or polynomial regression methods to generate a temperature rise trend curve. The time window can be set (such as 1 hour) to calculate the temperature change rate within each window. The temperature rise trend data can be visualized to help identify the regularity and abnormality of temperature rise. The temperature rise trend data can be discretized to generate time series data points. The piecewise linear fitting or spline fitting method can be used to ensure the continuity and peace of the fitting results. Slipperiness, calculate the standard deviation and mean of temperature change to evaluate the fluctuation during the temperature rise process, use the fitting results to generate a temperature rise trend curve, record the temperature change characteristics under different load and environmental conditions, visualize the curve, and show the evolution of temperature rise over time. The fitting results are stored in the database for subsequent analysis and comparison. The extracted vibration frequency fluctuation characteristics, load timing change characteristics, magnetic saturation degree values, and temperature rise trend curves are integrated to form a multi-dimensional feature vector. Feature selection methods (such as principal component analysis and correlation analysis) are used to screen the features and identify the features that have a significant impact on the transformer state. Machine learning models (such as random forests and support vector machines) are used to train the integrated features to identify the operating status of the transformer.Set training parameters such as batch size (64), learning rate (0.01), and training rounds (100), and evaluate the performance of the model through cross-validation to ensure that the model can accurately classify the status of the transformer (normal, warning, fault).

[0120] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0121] Step S21: performing coupling correlation analysis between parameters on the multi-dimensional real-time working monitoring parameters to obtain coupling correlation characteristics between multiple parameters;

[0122] Step S22: performing normalization range analysis of each parameter based on the multi-dimensional real-time working monitoring parameters to generate a normalization range for each parameter;

[0123] Step S23: defining a combination of multiple parameters based on the normalization range of each parameter and the coupling correlation characteristics between multiple parameters, thereby obtaining multiple parameter combination schemes;

[0124] Step S24: using multiple parameter combination schemes to perform dynamic operation simulation on the coupled transformer, and collecting operation simulation response data of the multiple schemes.

[0125] In this embodiment, real-time working monitoring parameter data of the transformer are collected, including temperature, current, voltage, vibration frequency, magnetic field strength, etc. Ensure the integrity and continuity of the data, and record at least one cycle (such as 24 hours) of data. The data should be stored in a structured format for subsequent analysis. It is recommended to use CSV or database format. Statistical analysis methods such as Pearson correlation coefficient and Spearman rank correlation coefficient are used to calculate the correlation between the parameters. Set a correlation threshold (such as 0.7) to screen out parameter pairs with high correlation. Use Pandas and NumPy libraries in Python for data processing, calculate the correlation coefficient matrix, and visualize it as a heat map to intuitively display the coupling relationship between parameters. According to the results of the correlation analysis, significant coupling correlation features are extracted. If the correlation between current and temperature is 0.85, record this feature for subsequent analysis and modeling. Generate a coupling correlation feature report to record the interaction relationship between parameters and their impact on transformer performance. Preprocess the collected monitoring parameter data, including removing outliers and filling missing values. Use the Z-score method or quartile method for outlier detection. Standardize the data of each parameter to ensure that the data is in the same dimension for easy comparison. Use statistical methods (such as the mean ± 2 times the standard deviation method) to calculate the normalization range of each parameter. For each parameter, calculate its mean and standard deviation, and determine the normal range. For the temperature parameter, if the mean is 75°C and the standard deviation is 3°C, the normalization range is (69°C, 81°C). Record the normalization range of each parameter in the database to form a standard for subsequent reference. Use visualization tools (such as Matplotlib) to draw a normalization range diagram of the parameters to intuitively display the normal working range of each parameter. According to the coupling correlation characteristics, identify the key parameter combinations that affect the performance of the transformer. Current, temperature, and vibration frequency constitute a combination because they affect each other and have a significant impact on the operating state of the transformer. Set the number and type of combination schemes to ensure that different working conditions and scenarios are covered, such as normal operation, overload, and fault conditions. Verify each combination scheme to ensure that it can effectively reflect the state of the transformer in actual work. Sensitivity analysis methods can be used to evaluate the impact of different parameter combinations on transformer performance. Record the characteristics of each combination and its applicable scenarios to ensure the effectiveness of subsequent simulation and analysis. Use simulation software (such as MATLAB / Simulink, ANSYS, etc.) to build a dynamic operation model of the transformer. Integrate multiple previously defined parameter combinations into the model to ensure that the actual operating environment can be simulated. Set simulation parameters, including simulation time (such as 10 minutes), sampling frequency (such as 1Hz), and load conditions (such as different load levels). Start the dynamic operation simulation, monitor the operating status of the transformer in real time, and record response data, including parameters such as temperature, load, current, and vibration. Ensure that multiple simulations are performed under different combinations to obtain sufficient data.Use a data logging system (such as a real-time database or data acquisition system) to store the simulated response data to ensure data integrity and accuracy. Analyze the collected response data to evaluate the impact of different parameter combinations on transformer performance. Use statistical analysis methods (such as variance analysis) to compare the effects of different combinations. Record the results of each simulation, including the trend of each parameter and the performance indicators of the transformer, for subsequent analysis and optimization.

[0126] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0127] Step S31: performing scheme performance calculation on the operation simulation response data of multiple schemes one by one to generate a comprehensive performance evaluation value of each scheme;

[0128] Step S32: performing multi-objective dynamic balance analysis on the operation simulation response data of the multiple schemes to generate dynamic balance characteristics of each scheme;

[0129] Step S33: extracting a better parameter combination scheme based on the dynamic balance characteristics of each scheme and the comprehensive performance evaluation value of each scheme;

[0130] Step S34: performing local parameter optimal balance optimization on the preferred parameter combination scheme, so as to obtain an optimal transformer operating parameter combination.

[0131] In this embodiment, the response data of each scheme is extracted from the previous simulation, which includes key parameters such as temperature, load, current, vibration frequency, etc. Ensure the integrity of the data for subsequent analysis. Organize the response data of each scheme into a structured format, and it is recommended to use Excel or a database for further processing. Determine the calculation formula for the comprehensive performance evaluation value, including the weighted average of multiple indicators. Consider the stability of temperature, the adaptability of load, and the stability of vibration, and set the weights (such as 40% for temperature, 30% for load, and 30% for vibration). Set the calculation method of the performance index, for example: temperature fluctuation amplitude = maximum temperature - minimum temperature load adaptability = ratio of actual load to rated load vibration stability = standard deviation of vibration frequency, and for each scheme, use the preset index calculation formula to calculate the comprehensive performance evaluation value. For scheme A, calculate its temperature fluctuation amplitude, load adaptability, and vibration stability, and calculate the comprehensive performance evaluation value according to the set weights. Record the comprehensive performance evaluation value of each scheme, and generate a corresponding performance evaluation report for subsequent analysis. Determine the analysis indicators of dynamic balance, such as load balance, temperature balance, and vibration balance. Each indicator is defined as: load balance: standard deviation of each phase load, temperature balance: standard deviation of temperature at different locations, vibration balance: uniformity measurement of vibration at each part, and the Pareto front analysis method is used to identify the trade-off between multiple objectives. In each scheme, the values ​​of different dynamic balance indicators are calculated, and a multi-dimensional balance feature diagram is generated. The SciPy library in Python is used for optimization calculation and balance feature diagram is drawn to identify the dynamic balance characteristics of each scheme. Dynamic balance characteristics are extracted from each scheme and relevant data are recorded. A dynamic balance feature report is generated, listing the balance indicators and corresponding values ​​of each scheme. All schemes are compared based on the comprehensive performance evaluation value and dynamic balance characteristics. A threshold is set, for example, the scheme with a comprehensive performance evaluation value greater than a certain value and a dynamic balance characteristic within a certain range is regarded as a better scheme. The decision matrix method is used to score and sort different schemes, combined with the weight of each indicator, so as to better identify the better scheme. The better schemes that meet the conditions are screened out and recorded in the database to form a list of better parameter combination schemes. Analyze each optimal solution to ensure its applicability and stability under different conditions. Determine the optimization goal, such as minimizing temperature fluctuations and vibration amplitudes or maximizing load adaptability while ensuring safe operation. Set the parameters of the optimization algorithm, such as genetic algorithm or particle swarm optimization algorithm, and optimize the parameters according to the goals and constraints. Apply the optimization algorithm to adjust local parameters for the selected optimal solution. Evaluate the impact of different parameter combinations on the objective function by simulating them. Record the performance evaluation value and dynamic balance characteristics after each parameter adjustment to ensure that the optimal solution can be found. Evaluate the optimal parameter combination solution after optimization to confirm its feasibility in actual operation.Perform multiple simulations to ensure the stability and consistency of the optimization results. Record the optimal parameter combination in the database and generate the final parameter optimization report for subsequent application and reference.

[0132] In this embodiment, the specific steps of step S31 are:

[0133] The operation efficiency of each scheme is calculated for the operation simulation response data of multiple schemes one by one to obtain the operation efficiency of each scheme;

[0134] Conduct quantitative analysis of the operation cost of the operation simulation response data of multiple schemes to generate the quantitative cost value of each scheme;

[0135] Perform reliability calculation on the operation simulation response data of multiple schemes to obtain the operation reliability evaluation value of each scheme;

[0136] Multi-objective performance calculations are performed on the operating efficiency of each solution, the quantified cost value of each solution, and the operational reliability evaluation value of each solution to generate a comprehensive performance evaluation value for each solution.

[0137] In this embodiment, the operating efficiency is generally defined as the ratio of output power to input power, Efficiency = Output Power / Input Power × 100%. The output power can be calculated by measuring the effective power provided by the transformer, and the input power is calculated by multiplying the current and voltage. For each solution, the input power and output power are calculated, and then the efficiency is calculated using the above formula. If the input power of a solution is 100kW and the output power is 90kW, the efficiency is 90%.

[0138] Record the operating efficiency of each solution and generate an efficiency evaluation report for subsequent comparison and analysis. Determine the components of operating costs, which usually include electricity costs, maintenance costs, personnel costs, etc. Set the electricity price based on electricity costs (e.g., 0.1 yuan per kilowatt-hour).

[0139] Collect the operating time and average load data of each solution for cost calculation. The operating cost is calculated as follows: Cost = electricity price × input current × operating time. If the input power of a solution is 100kW, the operating time is 10 hours, and the electricity price is 0.1 yuan / kWh, then its operating cost is 100kW×10h×0.1 yuan / kWh=100 yuan. Repeat the above cost calculation for each solution and record the cost quantification value of all solutions. Generate a cost analysis report to clarify the economic differences of each solution. Operational reliability is usually defined as the probability that a device will not fail within a specific time, which is calculated by failure rate and mean time between failures (MTBF).

[0140] Collect failure data (such as the number of failures, cumulative operating time, etc.) for each solution for reliability analysis. Use the following formula to calculate reliability: Use the following formula to calculate reliability: Reliability = MTBF / MTBF + MTTR. Among them, MTTR is the mean repair time. If the MTBF of a solution is 200 hours and the MTTR is 20 hours, perform the above calculations for each solution, record the operational reliability evaluation value of each solution, and generate a reliability evaluation report to assist in subsequent decision-making. Determine the indicators for calculating the comprehensive performance evaluation value, including operating efficiency, operating cost, and operating reliability. Define the weight of each indicator, for example, efficiency accounts for 40%, cost accounts for 30%, and reliability accounts for 30%. CompositePerformance = Wefficiency⋅Efficiency−Wcost⋅Cost+Wreliability⋅Reliability. For each solution, insert the respective efficiency, cost, and reliability data to calculate the comprehensive performance evaluation value. If the efficiency of a solution is 90%, the cost is 100 yuan, and the reliability is 90.9%, then the comprehensive performance evaluation value is: CompositePerformance=0.4⋅90−0.3⋅100+0.3⋅90.9. Record the comprehensive performance evaluation value of each solution and generate a comprehensive performance analysis report to facilitate comparison of the advantages and disadvantages of the solutions.

[0141] In this embodiment, the specific steps of step S4 are:

[0142] Step S41: performing an environmental dynamic parameter change analysis on the real-time working environment state parameter to generate an environmental dynamic parameter change feature;

[0143] Step S42: identifying environmental mutation points based on the environmental dynamic parameter change characteristics and extracting environmental state mutation points;

[0144] Step S43: performing periodic fluctuation deep mining on the characteristics of environmental dynamic parameter changes to generate a periodic fluctuation representation of the environmental state;

[0145] Step S44: Perform environmental condition distribution fitting according to the environmental condition mutation point and the environmental condition periodic fluctuation characterization to construct a real-time environmental condition distribution field.

[0146] In this embodiment, environmental state parameters are monitored in real time by sensors (such as temperature, humidity, air pressure, noise, etc.), and the data sampling frequency is set (such as once per second) to ensure the timeliness and accuracy of the data. Record at least 24 hours of environmental data to comprehensively analyze the dynamic changes of the environment. Use time series analysis methods to calculate the mean, standard deviation, maximum and minimum values ​​of each parameter in the time window to identify the trend of dynamic changes. For temperature data, calculate its change characteristics in different time periods (such as day and night), and use the moving average method to smooth the data to eliminate the impact of short-term fluctuations. Use visualization tools such as Matplotlib or Seaborn to draw the change curve of environmental parameters and intuitively display the dynamic change characteristics of the environment. Through visualization, it is easy to identify the fluctuation trend and abnormal situation of environmental parameters. Record the generated change characteristic data to form a report on the change characteristics of environmental dynamic parameters, which provides a basis for subsequent analysis. Select a suitable mutation point detection algorithm, such as CUSUM (cumulative sum control chart) or Pettitt test. These methods can effectively identify mutation points in time series data. Determine algorithm parameters, such as the control limit of CUSUM and the significance level of Pettitt test (such as 0.05). Input the real-time environmental dynamic parameter data into the mutation point detection algorithm, run the detection program, and identify the mutation points in the time series. If the temperature suddenly rises to a critical value at a certain time point, the detection algorithm will mark the point as a mutation point. Record all identified mutation points and their corresponding time and parameter values, and generate a mutation point identification report. Analyze the causes of the mutation points, which are related to external environmental factors (such as weather changes, equipment operating status, etc.), to provide a basis for subsequent analysis. Use frequency domain analysis methods such as Fourier transform or wavelet transform to analyze the periodic fluctuation characteristics of environmental parameters. Determine the frequency range of the analysis, such as 0.1Hz to 0.5Hz, to capture the periodic fluctuations. Use FFT to convert the time series data into the frequency domain and identify the main frequency components and corresponding amplitudes. If the temperature data has obvious periodic fluctuations within 24 hours, FFT will show the corresponding frequency peak. Record the periodic fluctuation characteristics of each parameter, including frequency, amplitude and phase information, and generate a periodic fluctuation characterization report. Use Matplotlib to draw a spectrum diagram to show the periodic fluctuation characteristics of environmental parameters. Visualization makes it easier to identify and understand the laws of periodic changes. Select an appropriate distribution modeling method, such as Gaussian process regression or kriging interpolation, to build a spatial distribution model of environmental conditions. Determine model parameters, such as hyperparameters for Gaussian processes or distance weights in kriging interpolation. Use mutation points and periodic fluctuation characteristics as input data and apply the selected model to fit the distribution of environmental conditions. Generate predicted values ​​of environmental conditions and compare them with actual measured values ​​to evaluate the accuracy and reliability of the model. Use visualization tools to draw a distribution map of environmental conditions to show the values ​​of environmental parameters in different regions. Generate a spatial distribution map of temperature or humidity to intuitively show changes in the environment.Record the fitting results and visualization diagrams to form an environmental status distribution field report to provide a basis for subsequent monitoring and decision-making.

[0147] In this embodiment, the specific steps of step S5 are:

[0148] Step S51: performing real-time operating parameter control on the coupled transformer based on the optimal transformer operating parameter combination, and continuously collecting operating monitoring parameters;

[0149] Step S52: performing a long-term working state simulation on the working monitoring parameters according to the real-time environmental condition distribution field to obtain transformer long-term operating state simulation data;

[0150] Step S53: performing operating state parameter trend evolution on the transformer long-term operating state simulation data to obtain long-term operating state trend characteristics in advance;

[0151] Step S54: performing fault trend prediction on the long-term operation status trend characteristics to obtain fault prediction trend data of the transformer.

[0152] In this embodiment, a real-time monitoring and control system is constructed, and a PLC (programmable logic controller) or SCADA (supervisory control and data acquisition) system is used to realize real-time control of the transformer operating parameters. The system needs to have interfaces with sensors and actuators to ensure timely transmission of data acquisition and control instructions, input the optimal transformer operating parameter combination (such as current, load, temperature, etc.) into the control system, set a reasonable control range and threshold to ensure that the transformer operates in a safe state, configure high-precision sensors (such as temperature sensors, current transformers, pressure sensors, etc.) on the transformer and its surrounding environment to monitor the operating parameters in real time, set the data sampling frequency to 1Hz to ensure the real-time nature of the data, and periodically collect the working monitoring data through the control system. Measure parameters, record operating status (such as load current, temperature, vibration, etc.), and store the data in the database for subsequent analysis. Analyze the collected data in real time, compare the optimal parameter combination, and determine whether the current operating status of the transformer is within the safe range. If the threshold is exceeded, the system will immediately issue an alarm to the operator and automatically adjust the parameters to restore to a safe state. Generate a working status report, record real-time working monitoring parameter changes, and form a data log to provide a basis for subsequent analysis. Use the previously established environmental condition distribution model to extract environmental parameters (such as temperature, humidity, pressure, etc.) and set the simulation time range (for example, 72 hours) to reflect the actual environmental changes in operation. Use the environmental condition distribution field data as input and combine it with the real-time monitoring parameters to generate a data log. Data, long-term operation status simulation, select appropriate simulation tools (such as MATLABSimulink, ANSYS, etc.) to build the dynamic model of the transformer, input real-time monitoring parameters and environmental status data for simulation calculation, set simulation conditions, including load changes, environmental fluctuations, etc., to ensure that the simulation results can reflect the actual operating conditions. During the simulation operation, long-term operation status simulation data is generated in real time, including current, temperature, pressure and other parameters, recorded in the database, and a simulation data report is generated, which records the simulation conditions, parameter changes and their impact on the transformer operation status in detail, providing a basis for subsequent analysis, and extracting key operating parameters (such as current, temperature, load, etc.) from the long-term simulation data. Use time series analysis methods ( Perform trend analysis on the data using machine learning algorithms (such as ARIMA models), calculate the mean, standard deviation, maximum and minimum values ​​of each parameter to identify the changing trend and amplitude of the parameters, identify the evolution law of the operating status parameters, and observe the changing trend of the parameters over time by drawing time series charts. If the temperature gradually rises within a certain period of time, record the trend as a potential precursor to failure, generate a trend evolution feature report, describe the changing trend and characteristics of each parameter in detail, and provide a basis for subsequent fault prediction. Select a suitable fault prediction model, such as a machine learning algorithm (such as random forest, support vector machine) or a statistical method (such as regression analysis), establish a fault prediction model related to the operating status parameters, and collect historical fault data to provide samples for model training.Use the extracted operating status parameter trend features as input and historical fault data as output to train the model, set training parameters (such as learning rate, number of iterations, etc.) to ensure that the model can accurately fit the data, use cross-validation to evaluate the model's prediction performance, ensure the model's accuracy and robustness, input the long-term operating status trend features into the trained fault prediction model, generate fault prediction trend data, record the prediction results, including the probability and time of fault occurrence, generate a fault trend prediction report, describe the prediction results and fault types in detail, and provide decision support for equipment maintenance and management.

[0153] In this embodiment, the specific steps of step S6 are:

[0154] Step S61: locating the fault location based on the fault prediction trend data of the transformer, and marking the transformer fault location point;

[0155] Step S62: performing local dynamic parameter tuning on the transformer fault location point, thereby generating dynamic tuning parameters of the fault point;

[0156] Step S63: performing global synchronous optimization of the transformer based on the dynamic tuning parameters of the fault point, thereby obtaining global synchronous optimization parameters;

[0157] Step S64: Perform iterative transfer learning on the global synchronous optimization parameters to build a transformer global parameter optimization engine.

[0158] In this embodiment, the fault prediction trend data of the transformer is collected, including information such as fault type, fault probability, and fault occurrence time. In addition, historical fault records are integrated for comparative analysis. Real-time monitoring data (such as temperature, vibration, oil pressure, etc.) are used as auxiliary information to help confirm the fault location. A model-based method, such as finite element analysis (FEA) or thermal analysis, is used to determine the fault location in combination with fault prediction data. Vibration analysis is used to identify that certain components of the transformer (such as windings and insulating materials) have faults. Cluster analysis (such as K-means clustering) is used to classify the fault data to find out the areas where faults frequently occur. Different parts of the transformer are used as cluster centers to analyze the corresponding fault characteristics, mark the determined fault locations, and record their locations, fault characteristics, and related parameters (such as temperature and pressure). A fault location report is generated, describing the fault information of each location point in detail for subsequent tuning and optimization. According to the characteristics of the fault location point, tuning targets are set, such as reducing temperature, reducing vibration amplitude, improving insulation performance, etc. The targets should be clear and quantifiable. Real-time control and optimization algorithms are used ( The system uses a control system equipped with sensors to monitor parameter changes in real time and determine tuning strategies, such as step-by-step adjustment and feedback control, to ensure the stability and effectiveness of the tuning process. Dynamic parameter tuning is performed on each fault location point. If the temperature of a winding is too high, the temperature can be lowered by adjusting the coolant flow rate. The parameter changes before and after tuning are recorded, and a dynamic tuning parameter report is generated. The parameter changes and optimization effects during the adjustment process are recorded to provide a basis for subsequent global optimization. A global optimization model is constructed based on local tuning parameters. All operating parameters of the transformer (such as current, temperature, load, etc.) and the coupling relationship between different components are comprehensively considered. Appropriate optimization algorithms (such as genetic algorithms and particle swarm optimization) are selected for global optimization to ensure that the global optimal solution can be found. Global synchronous optimization is implemented in the control system. Local tuning parameters are input and global parameters are adjusted. Optimization goals are set, such as minimizing energy consumption, maximizing efficiency, and reducing failure rate. Parameter changes during the optimization process are recorded, and the optimization effect is monitored. Simulation tools (such as MATLAB) are used to calculate the optimal solution. Simulink) to ensure the effectiveness of the optimization results, evaluate the global synchronous optimization parameters, generate a global optimization report, record in detail the optimization effect of each parameter and its impact on the transformer performance, ensure the scientific nature of the optimization decision, select a suitable transfer learning algorithm, such as Fine-tuning or DomainAdaptation in transfer learning, to apply the existing optimization experience to the new situation, collect historical optimization data and operating parameters to facilitate the training of the transfer learning model, use the global synchronous optimization parameters as training data, combine historical data for model training, set hyperparameters such as learning rate and number of iterations to ensure the accuracy and convergence of the model, evaluate the performance of the model in each iteration, and make adjustments to the parts with poor performance.In order to improve the predictive ability of the model, the trained transfer learning model is integrated into the control system of the transformer to form a global parameter optimization engine, which can analyze the operating data in real time, provide optimization suggestions and control strategies, record the operating effect of the optimization engine, generate engine performance reports, and evaluate its effectiveness and adaptability under different operating conditions.

[0159] In this embodiment, a coupled transformer parameter optimization device is provided, which is used to execute the coupled transformer parameter optimization method as described above, including:

[0160] A state perception module is used to obtain multi-dimensional real-time working monitoring parameters and real-time working environment state parameters of the coupled transformer; perform nonlinear dynamic analysis and global state perception modeling on the multi-dimensional real-time working monitoring parameters, thereby constructing a transformer global state perception model;

[0161] A dynamic operation simulation module is used to define a multi-parameter combination of the multi-dimensional real-time working monitoring parameters, perform dynamic operation simulation based on the transformer global state perception model, and collect operation simulation response data of multiple schemes;

[0162] The local parameter optimization module is used to perform multi-objective dynamic balance analysis on the operation simulation response data of multiple schemes and perform local parameter optimal balance optimization to obtain the optimal transformer operating parameter combination;

[0163] An environmental condition distribution module is used to analyze the environmental dynamic parameter changes of the real-time working environment state parameters, perform environmental condition distribution fitting, and construct a real-time environmental condition distribution field;

[0164] The fault trend prediction module is used to simulate the long-term working state based on the optimal transformer working parameter combination and the real-time environmental condition distribution field, and then perform fault trend prediction to obtain the transformer fault prediction trend data;

[0165] The global parameter optimization module is used to perform local dynamic parameter tuning based on the fault prediction trend data of the transformer, and perform global synchronous optimization to build a transformer global parameter optimization engine.

[0166] The present invention can fully perceive the operating state of the transformer by acquiring real-time working monitoring parameters and environmental state parameters, which provides reliable data support for further fault detection, performance optimization and safety warning. Nonlinear dynamic analysis helps to identify the complex interactive relationship between parameters, so that the global state perception model is not limited to static information, but can also reflect dynamic changes in real time and accurately predict the long-term operating trend of the transformer. By combining multi-dimensional parameters, the performance of the transformer under different operating conditions can be simulated, and then the best parameter combination scheme can be found. The module can simulate the dynamic response of the transformer under different load and environmental conditions, evaluate the effects of various operating schemes, discover potential problems in advance and pre-optimize. Through multi-objective dynamic balance analysis, multiple performance indicators (such as efficiency, load, temperature, etc.) are optimized at the same time to ensure that the various parameters of the transformer are kept in the best balance during operation. Through the optimization and adjustment of local parameters, the bottleneck problem under certain working conditions can be effectively solved, thereby improving the overall operating efficiency and stability of the transformer. The operation of the transformer is not only affected by its own parameters, but also by the strong influence of the external environment (such as temperature, humidity, etc.). By analyzing the dynamic changes of the environment, we can better understand the impact of the environment on the operation of the transformer and make environmental adaptability adjustments. By fitting the distribution field of environmental conditions, we can predict the impact of environmental changes on the transformer, provide a basis for subsequent fault prediction and optimization solutions, and enhance the transformer's adaptability to environmental changes. Based on the optimal working parameters and environmental changes, long-term state simulation is performed to identify potential fault trends in advance, which provides a basis for preventive maintenance of the transformer and helps to avoid sudden equipment failures. By predicting fault trends, maintenance personnel can take timely measures to overhaul or adjust the equipment to avoid long-term downtime and ensure the stable operation of the power system. By combining fault prediction trend data to optimize local and global parameters simultaneously, it is ensured that the transformer always maintains the best performance throughout its life cycle, reduces energy loss and faults, and makes dynamic adjustments based on fault prediction data to avoid faults to the greatest extent possible. At the same time, it optimizes the working efficiency and service life of the transformer. The global optimization engine automatically adjusts transformer parameters to reduce manual intervention, improves operation and maintenance efficiency, and ensures that the transformer can continue to operate in the best state.

[0167] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0168] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for optimizing parameters of a coupled transformer, characterized in that: The following steps are involved: Step S1: obtaining multi-dimensional real-time working monitoring parameters and real-time working environment state parameters of the coupled transformer; performing nonlinear dynamic analysis and global state perception modeling on the multi-dimensional real-time working monitoring parameters, thereby constructing a transformer global state perception model; Step S2: defining a multi-parameter combination of the multi-dimensional real-time working monitoring parameters, performing a dynamic operation simulation based on a transformer global state perception model, and collecting operation simulation response data of multiple schemes; Step S3: performing multi-objective dynamic balance analysis on the operation simulation response data of multiple schemes, and performing local parameter optimal balance optimization, so as to obtain the optimal transformer operating parameter combination; Step S4: performing environmental dynamic parameter change analysis on the real-time working environment state parameters, and performing environmental condition distribution fitting to construct a real-time environmental condition distribution field; Step S5: performing a long-term working state simulation based on the optimal transformer working parameter combination and the real-time environmental condition distribution field, and then performing a fault trend prediction to obtain transformer fault prediction trend data; Step S6: Perform local dynamic parameter tuning based on the fault prediction trend data of the transformer, and perform global synchronous optimization to build a transformer global parameter optimization engine; Among them, the specific steps of step S2 are: Step S21: performing coupling correlation analysis between parameters on the multi-dimensional real-time working monitoring parameters to obtain coupling correlation characteristics between multiple parameters; Step S22: performing a normalization range analysis of each parameter based on the multi-dimensional real-time working monitoring parameters to generate a normalization range for each parameter; Step S23: defining a combination of multiple parameters based on the normalization range of each parameter and the coupling correlation characteristics between multiple parameters, thereby obtaining multiple parameter combination schemes; Step S24: using multiple parameter combination schemes to perform dynamic operation simulation on the coupled transformer, and collecting operation simulation response data of the multiple schemes; The specific steps of step S4 are: Step S41: performing an environmental dynamic parameter change analysis on the real-time working environment state parameter to generate an environmental dynamic parameter change feature; Step S42: Identify the environmental mutation points based on the environmental dynamic parameter change characteristics and extract the environmental state mutation points; Step S43: performing periodic fluctuation deep mining on the characteristics of environmental dynamic parameter changes to generate a periodic fluctuation representation of the environmental state; Step S44: Perform environmental condition distribution fitting according to the environmental condition mutation point and the environmental condition periodic fluctuation characterization to construct a real-time environmental condition distribution field.

2. The method for optimizing the parameters of a coupled transformer according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: obtaining multi-dimensional real-time working monitoring parameters and real-time working environment status parameters of the coupled transformer; Step S12: performing multi-dimensional comprehensive state mining on the multi-dimensional real-time working monitoring parameters to obtain multi-dimensional state characteristics of the transformer; Step S13: performing nonlinear dynamic analysis according to the multi-dimensional state characteristics of the transformer to extract the nonlinear dynamic characteristics of the transformer; Step S14: performing state characteristic evolution on the nonlinear dynamic characteristics of the transformer to generate a nonlinear dynamic state evolution law; Step S15: Perform global state perception modeling based on the nonlinear dynamic state evolution law, thereby constructing a transformer global state perception model.

3. The method for optimizing the parameters of a coupled transformer according to claim 2, characterized in that: The specific steps of step S12 are: The multi-dimensional real-time working monitoring parameters include mechanical vibration frequency, current and voltage parameters, magnetic field strength value and temperature parameters; Identify the vibration frequency fluctuation of the mechanical vibration frequency and extract the vibration frequency fluctuation characteristics of the transformer; Calculate the real-time transformer load based on current and voltage parameters; Performing load time sequence variation analysis on the real-time load of the transformer to obtain load time sequence variation characteristics; Fitting the magnetic field intensity distribution to the magnetic field intensity value to construct a magnetic field intensity distribution field; The magnetic saturation degree is estimated based on the magnetic field intensity distribution field to obtain the real-time magnetic saturation degree value of the transformer; Conduct temperature rise trend analysis on temperature parameters and generate transformer temperature rise trend data; Perform discrete fitting of temperature changes on transformer temperature rise trend data to construct transformer temperature rise trend curve; Multi-dimensional comprehensive state mining is performed on the transformer's vibration frequency fluctuation characteristics, load timing change characteristics, transformer real-time magnetic saturation value and transformer temperature rise trend curve to obtain the transformer's multi-dimensional state characteristics.

4. The method for optimizing the parameters of a coupled transformer according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: performing scheme performance calculation on the operation simulation response data of multiple schemes one by one to generate a comprehensive performance evaluation value of each scheme; Step S32: performing multi-objective dynamic balance analysis on the operation simulation response data of the multiple schemes to generate dynamic balance characteristics of each scheme; Step S33: extracting a better parameter combination scheme based on the dynamic balance characteristics of each scheme and the comprehensive performance evaluation value of each scheme; Step S34: performing local parameter optimal balance optimization on the preferred parameter combination scheme, so as to obtain an optimal transformer operating parameter combination.

5. The method for optimizing the parameters of a coupled transformer according to claim 4, characterized in that: The specific steps of step S31 are: The operation efficiency of each scheme is calculated for the operation simulation response data of multiple schemes one by one to obtain the operation efficiency of each scheme; Conduct quantitative analysis of the operation cost of the operation simulation response data of multiple schemes to generate the quantitative cost value of each scheme; Perform reliability calculation on the operation simulation response data of multiple schemes to obtain the operation reliability evaluation value of each scheme; Multi-objective performance calculations are performed on the operating efficiency of each solution, the quantified cost value of each solution, and the operational reliability evaluation value of each solution to generate a comprehensive performance evaluation value for each solution.

6. The method for optimizing the parameters of a coupled transformer according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: performing real-time operating parameter control on the coupled transformer based on the optimal transformer operating parameter combination, and continuously collecting operating monitoring parameters; Step S52: performing a long-term working state simulation on the working monitoring parameters according to the real-time environmental condition distribution field to obtain transformer long-term operating state simulation data; Step S53: performing operating state parameter trend evolution on the transformer long-term operating state simulation data to obtain long-term operating state trend characteristics in advance; Step S54: performing fault trend prediction on the long-term operation status trend characteristics to obtain fault prediction trend data of the transformer.

7. The method for optimizing the parameters of a coupled transformer according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: locating the fault location based on the fault prediction trend data of the transformer, and marking the transformer fault location point; Step S62: performing local dynamic parameter tuning on the transformer fault location point, thereby generating dynamic tuning parameters of the fault point; Step S63: performing global synchronous optimization of the transformer based on the dynamic tuning parameters of the fault point, thereby obtaining global synchronous optimization parameters; Step S64: Perform iterative transfer learning on the global synchronous optimization parameters to build a transformer global parameter optimization engine.

8. A parameter optimization device for a coupled transformer, characterized in that: The method for optimizing the parameters of the coupled transformer according to claim 1 comprises: A state perception module is used to obtain multi-dimensional real-time working monitoring parameters and real-time working environment state parameters of the coupled transformer; perform nonlinear dynamic analysis and global state perception modeling on the multi-dimensional real-time working monitoring parameters, thereby constructing a transformer global state perception model; A dynamic operation simulation module is used to define a multi-parameter combination of the multi-dimensional real-time working monitoring parameters, perform dynamic operation simulation based on the transformer global state perception model, and collect operation simulation response data of multiple schemes; The local parameter optimization module is used to perform multi-objective dynamic balance analysis on the operation simulation response data of multiple schemes and perform local parameter optimal balance optimization to obtain the optimal transformer operating parameter combination; An environmental condition distribution module is used to analyze the environmental dynamic parameter changes of the real-time working environment state parameters, perform environmental condition distribution fitting, and construct a real-time environmental condition distribution field; The fault trend prediction module is used to simulate the long-term working state based on the optimal transformer working parameter combination and the real-time environmental condition distribution field, and then perform fault trend prediction to obtain the transformer fault prediction trend data; The global parameter optimization module is used to perform local dynamic parameter tuning based on the fault prediction trend data of the transformer, and perform global synchronous optimization to build a transformer global parameter optimization engine.

Images