Water cooling control method, terminal and storage medium for gas test transformer

Abstract

The present invention belongs to the technical field of power grid equipment control, and specifically relates to a water-cooling control method, terminal, and storage medium for a gas test transformer. The method comprises collecting multi-dimensional parameters of a gas test transformer and a supporting water cooling system of the transformer in real time through a sensor network. The multi-dimensional parameters include transformer temperature, load voltage and current, transformer winding temperature, cooling medium temperature, cooling medium flow rate, ambient temperature and humidity, and speed of a variable-frequency fan in the water cooling system. Combined with a heat conduction model established based on Fourier's law of heat conduction, the method can accurately predict the temperature distribution of the transformer, extract corresponding PID parameters from a pre-stored temperature distribution and PID parameter correspondence table, and optimize the PID parameters through a pre-stored LMS algorithm.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid equipment control, and in particular relates to a water cooling control method, a terminal and a storage medium for a gas test transformer. Background Art

[0002] In power systems, gas-fired test transformers are critical high-voltage testing equipment. Their operational stability and safety are directly related to the reliability of the entire power system. However, under existing technologies, the water-cooling control system for gas-fired test transformers still has many deficiencies, hindering further improvement in their performance.

[0003] The application of traditional PID control algorithms in gas test transformer water cooling systems primarily relies on manually setting fixed PID parameters. While this approach is simple and straightforward, its limitations are particularly evident when faced with complex and changing operating conditions. Once set, PID parameters are difficult to automatically adjust. When external conditions such as load current and ambient temperature change, traditional PID control systems often fail to respond promptly, resulting in poor cooling performance and potentially even safety hazards such as transformer overheating.

[0004] Furthermore, traditional water-cooling systems also have significant shortcomings in temperature monitoring. They often rely on a single temperature monitoring method, such as monitoring cooling water temperature or transformer casing temperature. This limited monitoring method cannot fully reflect the true temperature conditions within the transformer, particularly temperature fluctuations in critical areas such as the windings. Since the windings generate the most heat within the transformer, their temperature fluctuations have a crucial impact on the transformer's operating status. Therefore, the limitations of traditional water-cooling systems in temperature monitoring make it difficult to accurately grasp the transformer's actual operating status, which in turn affects the accuracy and effectiveness of control decisions. Summary of the Invention

[0005] In view of the defects of the conventional PID control algorithm in the water cooling system of a gas test transformer in the prior art, the application of the conventional PID control algorithm mainly relies on manually fixed setting of PID parameters. When external conditions such as load current and ambient temperature change, the algorithm often fails to respond in time, resulting in poor cooling effect. The algorithm also relies on a single temperature monitoring method, which cannot fully reflect the actual temperature conditions inside the transformer. The present invention provides a water cooling control method, terminal and storage medium for a gas test transformer to solve the above technical problems.

[0006] In a first aspect, the present invention provides a water cooling control method for a gas test transformer, comprising:

[0007] The sensor network collects multi-dimensional parameters of the gas test transformer and its supporting water cooling system in real time. These parameters include transformer temperature, load voltage and current, transformer winding temperature, cooling medium temperature, cooling medium flow rate, ambient temperature and humidity, and variable frequency fan speed in the water cooling system.

[0008] Use local PLC to pre-process the collected multi-dimensional parameters and compress and transmit them to the host computer; pre-processing includes outlier filtering and missing value filling;

[0009] The host computer performs feature selection on the preprocessed multi-dimensional parameters based on L1 regularization to eliminate redundant data;

[0010] The multi-dimensional data after redundant data removal is input into the heat conduction model established based on Fourier's heat conduction law to predict the temperature distribution of the gas test transformer;

[0011] According to the temperature distribution, the corresponding PID parameters are extracted from the pre-stored temperature distribution and PID parameter correspondence table, and the extracted PID parameters are optimized based on the pre-stored LMS algorithm to adjust the operating state of the transformer supporting water cooling system.

[0012] Further improvements to this technical solution include real-time acquisition of multi-dimensional parameters of the gas test transformer and its supporting water cooling system through a sensor network. The multi-dimensional parameters include transformer temperature, load voltage and current, transformer winding temperature, cooling medium temperature, cooling medium flow rate, ambient temperature and humidity, and variable frequency fan speed in the water cooling system. The method includes:

[0013] Use PT100 platinum resistance sensor to collect transformer surface temperature ;

[0014] Using fiber grating sensor to collect winding temperature ;

[0015] The load current is collected by Rogowski coil , use voltage transformer to collect load voltage ;

[0016] Use thermocouple sensor to collect cooling medium temperature ;

[0017] Use ultrasonic flowmeter to collect cooling medium flow rate ;

[0018] Use digital temperature and humidity sensor to collect ambient temperature and relative humidity ;

[0019] Use Hall effect sensors to collect the speed of variable frequency fans in water cooling systems .

[0020] A further improvement of this technical solution is that the host computer performs feature selection on the preprocessed multi-dimensional parameters based on L1 regularization to eliminate redundant data. The method includes:

[0021] The preprocessed multi-dimensional parameters are combined into feature vectors , each dimension corresponds to a collected parameter and is standardized;

[0022] Construct an objective function including an L1 regularization term, solve the eigenvector by minimizing the weighted sum of the prediction error and parameter sparsity, and iteratively update the weight corresponding to the eigenvector using the coordinate descent method;

[0023] The features are sorted according to their absolute values, and features with weights less than the preset weight threshold are eliminated, while the features with the strongest correlation with the temperature distribution are retained.

[0024] A further improvement of this technical solution is that the objective function including the L1 regularization term is:

[0025] ;

[0026] Where n is the number of samples; is the eigenvector of the i-th sample; is the target value of the i-th sample; Mean squared error loss function; is the feature weight vector with dimension m; is the regularization parameter; is the L1 norm penalty term.

[0027] A further improvement of this technical solution is to input the multi-dimensional data after removing redundant data into a heat conduction model established based on Fourier's heat conduction law to predict the temperature distribution of the gas test transformer. The method includes:

[0028] Based on the pre-stored physical structural parameters of the gas test transformer, the interior of the gas test transformer is divided into three-dimensional grid cells: the sulfur hexafluoride gas area, the transformer shell metal layer area, and the cooling pipe area, and an initial heat transfer coefficient is assigned to each cell.

[0029] Based on the differential form of Fourier's heat conduction law Establish a heat conduction model, iteratively calculate the heat flux density vector of each grid unit in the heat conduction model, and solve the temperature gradient equation corresponding to each unit by gradient descent method , and generate the corresponding real-time temperature distribution map; among them, is the direction and magnitude of heat flow through a unit area per unit time; is the thermal conductivity coefficient corresponding to the grid unit, the unit is W / (m·K), the larger the value, the stronger the thermal conductivity; is the temperature field in the grid cell; is the temperature gradient; is the heat source term;

[0030] Input the transformer winding temperature into the heat conduction model to initialize the temperature field of the heat conduction model;

[0031] Input the cooling medium flow rate into the heat conduction model and dynamically adjust the heat conduction coefficient corresponding to the cooling pipe area;

[0032] The load current and voltage are converted into the heating power of the winding and used as the heat source input of the temperature gradient equation to adjust the temperature field in the sulfur hexafluoride gas area;

[0033] Input the ambient temperature and humidity into the heat conduction model to adjust the heat dissipation boundary conditions of the metal layer area of ​​the transformer shell; input the variable frequency fan speed into the heat conduction model to adjust the heat dissipation boundary conditions of the metal layer area of ​​the transformer shell;

[0034] The heat conduction model outputs a 3D temperature contour map of the transformer windings, transformer housing, and cooling medium, and marks areas where the temperature exceeds a preset temperature threshold.

[0035] A further improvement of the technical solution is to extract corresponding PID parameters from a pre-stored temperature distribution and PID parameter correspondence table according to the temperature distribution, and optimize the extracted PID parameters based on a pre-stored LMS algorithm to adjust the operating state of the transformer supporting water cooling system. The method includes:

[0036] According to the high temperature area position and temperature gradient value marked in the three-dimensional temperature cloud map, the matching load differential coefficient, integral coefficient and threshold setting combination are retrieved from the pre-stored temperature distribution and PID parameter correspondence table;

[0037] The extracted PID parameters are adjusted based on the pre-stored LMS algorithm, and the weight coefficients are updated in real time by minimizing the mean square error between the current temperature distribution and the target temperature distribution;

[0038] According to the optimized PID parameters, the cooling medium flow rate of the transformer supporting water cooling system is adjusted synchronously and variable frequency fan speed in the water cooling system .

[0039] A further improvement of this technical solution is to adjust the extracted PID parameters based on a pre-stored LMS algorithm, and update the weight coefficient in real time by minimizing the mean square error between the current temperature distribution and the target temperature distribution. The method includes:

[0040] Based on the real-time 3D temperature cloud map output by the heat conduction model, the actual temperature value of each grid cell is extracted and compared point by point with the pre-stored target temperature distribution data;

[0041] The difference between the current temperature distribution and the ideal state is quantified by calculating the mean square error of all units. The calculation formula is: , where N is the total number of network units, is the predicted temperature value of the i-th unit, is the target temperature threshold corresponding to the i-th unit; and integrates the input signal vector u(n) under the current working condition, including load current, cooling medium flow rate and ambient temperature and humidity data;

[0042] Based on the pre-stored LMS algorithm, the weight vector w(n) of the PID parameters is iteratively updated according to the mean square error e(n) and the input signal vector u(n).

[0043] A further improvement of this technical solution is that the formula for iteratively updating the weight vector w(n) of the PID parameters is: ,in, It is the step size parameter in the LMS algorithm, which controls the amplitude of weight update.

[0044] In a second aspect, the present invention provides a terminal, comprising:

[0045] processor, memory, wherein

[0046] The memory is used to store computer programs,

[0047] The processor is used to call and run the computer program from the memory, so that the terminal executes the above-mentioned terminal method.

[0048] In a third aspect, the present invention provides a computer storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the methods described in the above aspects.

[0049] The beneficial effects of the present invention are:

[0050] This invention uses a sensor network to collect multi-dimensional parameters of the gas test transformer and water cooling system in real time. Specifically, multiple sensors are used to comprehensively monitor the transformer's surface temperature, winding temperature, and cooling medium temperature, enabling precise monitoring of the transformer's internal temperature distribution. This multi-dimensional, high-precision temperature monitoring method provides a reliable basis for precise control of the cooling system.

[0051] The present invention uses a host computer to perform feature selection on pre-processed multi-dimensional parameters based on L1 regularization, eliminating redundant data and improving the generalization ability and prediction accuracy of the model. Combined with a heat conduction model established based on Fourier's law of heat conduction, it can accurately predict the temperature distribution of the transformer, extract the corresponding PID parameters from a pre-stored temperature distribution and PID parameter correspondence table, and optimize the PID parameters using a pre-stored LMS algorithm. Compared to traditional PID control that relies on manually fixed parameter settings, the present invention can dynamically adjust the PID parameters according to real-time operating conditions, achieving more precise control and effectively responding to changes in external conditions such as load current and ambient temperature, thereby improving the cooling effect and reducing the risk of transformer overheating. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0053] Figure 1 A schematic flow chart of a method according to an embodiment of the present invention.

[0054] Figure 2 A schematic diagram of the structure of a terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, features, and advantages of the present invention more obvious and easy to understand, the technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings in the specific embodiments. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0057] PLC, Programmable Logic Controller, programmable logic controller.

[0058] L1 regularization, L1 Regularization, L1 regularization (also known as Lasso regularization) is a technique commonly used in machine learning and statistics to prevent model overfitting and achieve feature selection by adding an L1 norm penalty term to the loss function.

[0059] PID, Proportional-Integral-Derivative, is a control algorithm widely used in industrial control systems. It achieves precise control of the system by adjusting the three parameters of proportion, integration and differentiation.

[0060] RTD, Resistance Temperature Detector, is a sensor that uses the property of resistance changing with temperature to measure temperature. It is often used in situations where high-precision temperature measurement is required.

[0061] Figure 1 The present invention provides a schematic flow chart of a water cooling control method for a gas test transformer. The order of the steps in the flow chart may be changed, and some steps may be omitted, depending on different requirements.

[0062] like Figure 1 As shown, the method includes:

[0063] Step 110, collecting multi-dimensional parameters of the gas test transformer and the transformer's supporting water cooling system in real time through a sensor network, the multi-dimensional parameters including transformer temperature, load voltage and current, transformer winding temperature, cooling medium temperature, cooling medium flow rate, ambient temperature and humidity, and variable frequency fan speed in the water cooling system;

[0064] Step 120: pre-process the collected multi-dimensional parameters using the local PLC and compress and transmit them to the host computer; the pre-processing includes outlier filtering and missing value filling;

[0065] Step 130: The host computer performs feature selection on the preprocessed multi-dimensional parameters based on L1 regularization to eliminate redundant data;

[0066] Step 140 , inputting the multi-dimensional data after removing redundant data into a heat conduction model established based on Fourier's heat conduction law to predict the temperature distribution of the gas test transformer;

[0067] Step 150 : extracting corresponding PID parameters from a pre-stored temperature distribution and PID parameter correspondence table according to the temperature distribution, optimizing the extracted PID parameters based on a pre-stored LMS algorithm, and adjusting the operating state of the transformer supporting water cooling system.

[0068] To facilitate understanding of the present invention, the water cooling control method for the gas test transformer provided by the present invention is further described below based on the principle of the water cooling control method for the gas test transformer of the present invention and the process of controlling the water cooling system in the gas test transformer in the embodiment.

[0069] Specifically, a sensor network is used to collect multi-dimensional parameters of a gas test transformer and its supporting water cooling system in real time. The multi-dimensional parameters include transformer temperature, load voltage and current, transformer winding temperature, cooling medium temperature, cooling medium flow rate, ambient temperature and humidity, and variable frequency fan speed in the water cooling system. The method includes:

[0070] Use PT100 platinum resistance sensor to collect transformer surface temperature ;

[0071] Using fiber grating sensor to collect winding temperature ;

[0072] The load current is collected by Rogowski coil , use voltage transformer to collect load voltage ;

[0073] Use thermocouple sensor to collect cooling medium temperature ;

[0074] Use ultrasonic flowmeter to collect cooling medium flow rate ;

[0075] Use digital temperature and humidity sensor to collect ambient temperature and relative humidity ;

[0076] Use Hall effect sensors to collect the speed of variable frequency fans in water cooling systems .

[0077] PT100 platinum resistance sensors are evenly distributed on the surface of the transformer casing (top, side and bottom), with one PT100 platinum resistance sensor installed every 0.5 meters. The PT100 platinum resistance sensor outputs a resistance signal, which is converted into a temperature value (accuracy of ±0.5°C) by the RTD module of the local PLC. The sampling frequency is 10Hz.

[0078] Fiber Bragg grating sensors are embedded in the transformer winding coils, with temperature measurement points arranged at 20 cm intervals along the winding axis. The fiber optic sensor analyzes temperature changes through a wavelength demodulator (integrated in the local PLC), and the data is transmitted to the host computer in real time with a sampling frequency of 50 Hz.

[0079] The Rogowski coil is connected to the high-voltage wire on the primary side of the transformer; the output current signal is converted into an analog quantity by an integrator, and the AI ​​module of the local PLC collects and filters it (cut-off frequency 1kHz) with a sampling frequency of 1kHz.

[0080] The capacitive voltage transformer is connected in parallel to the high-voltage terminal and the ground terminal of the primary side of the transformer; the output signal is conditioned by the voltage divider circuit and input into the PLC with an accuracy of ±0.2%.

[0081] K-type thermocouple sensors are embedded at the inlet and outlet of the cooling pipe, with a group of sensors arranged every 1 meter; the thermocouple signal is processed by the cold-end compensation circuit and input into the PLC, with a temperature resolution of 0.1°C and a sampling frequency of 5Hz.

[0082] The ultrasonic flowmeter is installed in the straight section of the cooling pipe (meeting the straight pipe requirement of 10D in front and 5D in the back, where D is the pipe diameter). It measures flow velocity based on the time difference method and outputs a 4-20mA signal to the PLC with an accuracy of ±1% and a response time of ≤100ms.

[0083] A digital temperature and humidity sensor (SHT35) is placed within 1 meter of the transformer body (avoiding heat radiation areas). The I²C digital signal is directly input into the PLC. The temperature range is -40-125°C, the humidity range is 0-100% RH, and the sampling frequency is 1 Hz.

[0084] Hall effect sensor (integrated in the fan motor): A magnet is installed on the motor shaft, and the Hall effect sensor is fixed to the motor housing. It outputs a fixed pulse per revolution (e.g., 60 pulses per revolution). The PLC calculates the real-time speed through a high-speed counter module (resolution ±1 rpm), with a sampling frequency of 100 Hz.

[0085] All sensor data is pre-processed by PLC, including:

[0086] Outlier filtering: A sliding window method (window size 50 sampling points) is used to identify and remove outliers;

[0087] Missing value filling: linear interpolation of adjacent timestamps;

[0088] Standardization: normalize each parameter to the interval [0,1] to eliminate dimensional differences;

[0089] Data transmission: The PLC compresses and transmits the pre-processed data to the host computer via the Ethernet communication protocol (Modbus TCP). The transmission cycle is 200ms to ensure real-time data.

[0090] The present invention uses a multi-type, high-precision sensor network to collect multi-dimensional parameters of the gas test transformer and water cooling system in real time, ensuring the comprehensiveness and accuracy of the data. Each type of sensor uses specialized measurement methods for different monitoring objects, improving data reliability. At the same time, the local PLC preprocesses the collected data, including outlier filtering, missing value filling, and standardization, further improving data quality. The preprocessed data is compressed and transmitted to the host computer via the Ethernet communication protocol, ensuring real-time data while reducing transmission bandwidth requirements. This technical approach provides a reliable data foundation for subsequent precise control, fault prediction, and other aspects, helping to improve system operational stability and safety.

[0091] Afterwards, the host computer performs feature selection on the preprocessed multi-dimensional parameters based on L1 regularization to eliminate redundant data. The method includes:

[0092] The preprocessed multi-dimensional parameters are combined into feature vectors , each dimension corresponds to a collected parameter and is standardized;

[0093] Construct an objective function including an L1 regularization term, solve the eigenvector by minimizing the weighted sum of the prediction error and parameter sparsity, and iteratively update the weight corresponding to the eigenvector using the coordinate descent method;

[0094] The features are sorted according to their absolute values, and features with weights less than the preset weight threshold are eliminated, while the features with the strongest correlation with the temperature distribution are retained.

[0095] Furthermore, the objective function including the L1 regularization term is:

[0096] ;

[0097] Where n is the number of samples; is the eigenvector of the i-th sample; is the target value of the i-th sample; Mean squared error loss function; is the feature weight vector with dimension m; is the regularization parameter; is the L1 norm penalty term.

[0098] Initialize the weight vector is a zero vector; update the weights according to the feature dimensions: ,like ,but ,otherwise, ;in, is the correlation between the jth feature and the model residual; is the variance of the jth feature, which is used to standardize the weight update amplitude.

[0099] The present invention performs feature selection on the pre-processed multi-dimensional parameters based on L1 regularization by the host computer, effectively eliminating redundant data and improving data processing efficiency and model performance. Specifically, after the multi-dimensional parameters are composed into feature vectors and standardized, an objective function containing an L1 regularization term is constructed, and the feature weights are optimized by minimizing the weighted sum of the prediction error and the parameter sparsity. The weights are iteratively updated using the coordinate descent method to ensure the convergence and stability of the algorithm. Finally, the low-weight features are sorted and eliminated according to the absolute value of the feature weights, retaining the features with the strongest correlation with the temperature distribution, providing high-quality, low-redundancy input data for subsequent heat conduction model prediction and PID parameter optimization, and enhancing the overall reliability and control accuracy of the system.

[0100] In addition, the multi-dimensional data after redundant data removal is input into a heat conduction model established based on Fourier's heat conduction law to predict the temperature distribution of the gas test transformer. The method includes:

[0101] Based on the pre-stored physical structural parameters of the gas test transformer, the interior of the gas test transformer is divided into three-dimensional grid cells: the sulfur hexafluoride gas area, the transformer shell metal layer area, and the cooling pipe area, and an initial heat transfer coefficient is assigned to each cell.

[0102] Based on the differential form of Fourier's heat conduction law Establish a heat conduction model, iteratively calculate the heat flux density vector of each grid unit in the heat conduction model, and solve the temperature gradient equation corresponding to each unit by gradient descent method , and generate the corresponding real-time temperature distribution map; among them, is the direction and magnitude of heat flow through a unit area per unit time; is the thermal conductivity coefficient corresponding to the grid unit, the unit is W / (m·K), the larger the value, the stronger the thermal conductivity; is the temperature field in the grid cell; is the temperature gradient; is the heat source term;

[0103] Input the transformer winding temperature into the heat conduction model to initialize the temperature field of the heat conduction model;

[0104] The cooling medium flow rate Input the heat conduction model and dynamically adjust the heat conduction coefficient corresponding to the cooling channel area;

[0105] The load current and voltage are converted into the heating power of the winding and used as the heat source input of the temperature gradient equation to adjust the temperature field in the sulfur hexafluoride gas area;

[0106] Input the ambient temperature and humidity into the heat conduction model to adjust the heat dissipation boundary conditions of the metal layer area of ​​the transformer shell; input the variable frequency fan speed into the heat conduction model to adjust the heat dissipation boundary conditions of the metal layer area of ​​the transformer shell;

[0107] The heat conduction model outputs a 3D temperature contour map of the transformer windings, transformer housing, and cooling medium, and marks areas where the temperature exceeds a preset temperature threshold.

[0108] Heat source term Although the load voltage parameter U is not directly used to calculate the heating power of the winding, it still plays a key role in the following: Exceeding the load voltage limit (such as overvoltage or undervoltage) may trigger the protection mechanism, adjusting the cooling strategy in advance or cutting off the power supply to prevent equipment damage. For example, when the voltage is abnormal, even if the current is normal, the cooling intensity may be actively increased due to the increased risk of insulation breakdown. In addition, the voltage U can be combined with the load current I to calculate the actual power. , assists in evaluating the transformer operating efficiency. The system adjusts the cooling strategy by the power factor, for example, reducing cooling energy consumption when the power factor is high. The load voltage U is used as the characteristic vector Components of , participating in L1 regularized feature selection. Although they may eventually be removed (if the weight is low), their redundancy or relevance needs to be verified through data analysis.

[0109] Variable frequency fan speed Adjusting the forced convection heat transfer coefficient h on the shell surface affects the heat dissipation boundary conditions of the metal layer area of ​​the transformer shell. The calculation formula of the forced convection heat transfer coefficient h is: .

[0110] The ambient temperature and humidity modify the natural convection heat transfer boundary conditions on the shell surface. When the humidity is higher than the preset threshold, the heat dissipation efficiency is reduced, affecting the heat dissipation boundary conditions of the metal layer area of ​​the compressor shell.

[0111] Sulfur hexafluoride gas area: filled with insulating gas, with a default thermal conductivity coefficient (k1) of 0.015 W / (m·K); transformer housing metal layer area (aluminum housing): k2 = 237 W / (m·K); cooling pipe area (fluorinated liquid flowing in stainless steel pipes): k3 = 50 W / (m·K).

[0112] Initialize the temperature field ; Based on the current temperature gradient , update the temperature distribution ; Iterate until the convergence condition (temperature change is less than 0.1°C) or the maximum number of iterations (such as 1000).

[0113] Adjust the equivalent heat transfer coefficient of the cooling channel area according to the flow rate data .

[0114] By meticulously dividing the transformer interior into three-dimensional grid cells composed of different material regions and assigning each region appropriate initial thermal conductivity coefficients, the model accurately simulates the complex heat conduction processes within the transformer, including the sulfur hexafluoride gas region, the transformer housing metal layer region, and the cooling duct region, ensuring both accuracy and comprehensiveness. The heat conduction model is constructed using the differential form of Fourier's law of heat conduction. The heat flux density vector and temperature gradient equation for each grid cell are iteratively calculated. The model dynamically reflects the temperature changes of the transformer under different operating conditions, generating a real-time temperature distribution map. This not only facilitates the timely identification of potential high-temperature areas but also provides accurate data support for subsequent cooling control. The model output includes a three-dimensional temperature cloud map of the transformer windings, housing, and cooling medium, specifically marking areas where temperatures exceed a preset threshold. This allows operators to intuitively understand the transformer's thermal status and take timely measures to prevent overheating. Combined with cooling medium flow rate data collected by an ultrasonic flowmeter, the cooling path is optimized through fluid dynamics simulation and the coefficients in the heat conduction model are dynamically adjusted, further enhancing the model's prediction accuracy and practicality. This adaptive adjustment mechanism ensures that the cooling system can intelligently adjust to actual operating conditions, effectively improving the transformer's operating efficiency and safety.

[0115] In addition, according to the temperature distribution, corresponding PID parameters are extracted from a pre-stored temperature distribution and PID parameter correspondence table, and the extracted PID parameters are optimized based on a pre-stored LMS algorithm to adjust the operating state of the transformer supporting water cooling system. The method includes:

[0116] According to the high temperature area position and temperature gradient value marked in the three-dimensional temperature cloud map, the matching load differential coefficient, integral coefficient and threshold setting combination are retrieved from the pre-stored temperature distribution and PID parameter correspondence table;

[0117] The extracted PID parameters are adjusted based on the pre-stored LMS algorithm, and the weight coefficients are updated in real time by minimizing the mean square error between the current temperature distribution and the target temperature distribution;

[0118] According to the optimized PID parameters, the cooling medium flow rate of the transformer supporting water cooling system is adjusted synchronously and variable frequency fan speed in the water cooling system .

[0119] Based on historical experimental data, the mapping relationship between temperature distribution characteristics (such as high temperature area location, temperature gradient value) and PID parameters is established, as shown in Table 1.

[0120] Table 1: Mapping relationship between temperature distribution characteristics and PID parameters

[0121]

[0122] The retrieval logic from Table 1 is as follows: the host computer matches the corresponding PID parameter combination (Kd = 1.0, Ki = 1.5, threshold 95°C) based on the high-temperature area (such as the winding hot spot accounting for 25%) and temperature gradient (such as 85°C / m) marked in the real-time three-dimensional temperature cloud map.

[0123] Furthermore, the extracted PID parameters are adjusted based on a pre-stored LMS algorithm, and the weight coefficients are updated in real time by minimizing the mean square error between the current temperature distribution and the target temperature distribution. The method includes:

[0124] Based on the real-time 3D temperature cloud map output by the heat conduction model, the actual temperature value of each grid cell is extracted and compared point by point with the pre-stored target temperature distribution data;

[0125] The difference between the current temperature distribution and the ideal state is quantified by calculating the mean square error of all units. The calculation formula is: , where N is the total number of network units, is the predicted temperature value of the i-th unit, is the target temperature threshold corresponding to the i-th unit; and integrates the input signal vector u(n) under the current working condition, including load current, cooling medium flow rate and ambient temperature and humidity data;

[0126] Based on the pre-stored LMS algorithm, the weight vector w(n) of the PID parameters is iteratively updated according to the mean square error e(n) and the input signal vector u(n).

[0127] Specifically, the formula for iteratively updating the weight vector w(n) of the PID parameters is: ,in, It is the step size parameter in the LMS algorithm, which controls the amplitude of weight update.

[0128] By establishing a mapping table between temperature distribution characteristics and PID parameters, the system can quickly retrieve matching PID parameter combinations based on real-time three-dimensional temperature cloud maps, achieving precise response to cooling requirements under different operating conditions. This mapping relationship, based on historical experimental data, ensures scientific and rational PID parameter selection and avoids the blind and inefficient manual parameter adjustment. Secondly, the extracted PID parameters are optimized in real time using the LMS algorithm. By minimizing the mean square error between the current and target temperature distributions, the weight coefficients are dynamically adjusted, enabling the cooling system to adaptively adjust according to actual temperature conditions. This optimization mechanism not only improves cooling effectiveness but also effectively reduces energy waste and enhances system efficiency. Finally, the cooling medium flow rate and variable frequency fan speed are synchronously adjusted based on the optimized PID parameters, achieving precise control of the water cooling system's operating status. This closed-loop control strategy ensures that the transformer maintains a stable operating temperature under various operating conditions, effectively preventing safety hazards such as overheating, extending the transformer's service life, and improving the overall reliability of the power system. Furthermore, this technical approach offers high flexibility and scalability, adapting to gas test transformers of varying models and specifications, providing strong support for the intelligent management of power equipment.

[0129] Figure 2 This is a structural diagram of a terminal 200 provided in an embodiment of the present invention. The terminal 200 can be used to execute the water cooling control method for a gas test transformer provided in an embodiment of the present invention.

[0130] The terminal 200 may include a processor 210, a memory 220, and a communication module 230. These components communicate via one or more buses. Those skilled in the art will appreciate that the server structure shown in the figure does not limit the present invention. The server structure may be a bus structure or a star structure, and may include more or fewer components than shown, or may combine certain components or arrange the components differently.

[0131] The memory 220 can be used to store execution instructions of the processor 210. The memory 220 can be implemented by any type of volatile or non-volatile storage terminal, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk. When the execution instructions in the memory 220 are executed by the processor 210, the terminal 200 can perform some or all of the steps in the above-described method embodiments.

[0132] The processor 210 is the control center of the storage terminal. It uses various interfaces and lines to connect various parts of the entire electronic terminal. It executes various functions of the electronic terminal and / or processes data by running or executing software programs and / or modules stored in the memory 220, and calling data stored in the memory. The processor can be composed of an integrated circuit (IC), for example, it can be composed of a single packaged IC, or it can be composed of multiple packaged ICs with the same or different functions. For example, the processor 210 can only include a central processing unit (CPU). In the embodiment of the present invention, the CPU can be a single computing core or multiple computing cores.

[0133] The communication module 230 is used to establish a communication channel so that the storage terminal can communicate with other terminals, receive user data sent by other terminals, or send user data to other terminals.

[0134] The present invention also provides a computer storage medium, wherein the computer storage medium may store a program that, when executed, may include some or all of the steps of each embodiment provided herein. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0135] Those skilled in the art will clearly understand that the techniques in the embodiments of the present invention can be implemented using software plus a necessary general-purpose hardware platform. Based on this understanding, the technical solutions in the embodiments of the present invention, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, among other media capable of storing program code, and includes instructions for causing a computer terminal (which can be a personal computer, a server, or a second terminal, a network terminal, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention.

[0136] In this specification, the same or similar parts between the various embodiments can be referred to each other. In particular, for the terminal embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method embodiment.

[0137] Although the present invention has been described in detail with reference to the accompanying drawings and in conjunction with preferred embodiments, the present invention is not limited thereto. Without departing from the spirit and essence of the present invention, persons of ordinary skill in the art may make various equivalent modifications or substitutions to the embodiments of the present invention, and such modifications or substitutions shall be within the scope of the present invention. Any changes or substitutions that can be easily conceived by persons skilled in the art within the technical scope disclosed in the present invention shall be within the scope of protection of the present invention.

Claims

1. A water cooling control method for a gas test transformer, characterized in that: include: The sensor network collects multi-dimensional parameters of the gas test transformer and its supporting water cooling system in real time. These parameters include transformer temperature, load voltage and current, transformer winding temperature, cooling medium temperature, cooling medium flow rate, ambient temperature and humidity, and variable frequency fan speed in the water cooling system. Use local PLC to pre-process the collected multi-dimensional parameters and compress and transmit them to the host computer; pre-processing includes outlier filtering and missing value filling; The host computer performs feature selection on the preprocessed multi-dimensional parameters based on L1 regularization to eliminate redundant data; The multi-dimensional data after redundant data removal is input into the heat conduction model established based on Fourier's heat conduction law to predict the temperature distribution of the gas test transformer; According to the temperature distribution, the corresponding PID parameters are extracted from the pre-stored temperature distribution and PID parameter correspondence table, and the extracted PID parameters are optimized based on the pre-stored LMS algorithm to adjust the operating status of the transformer supporting water cooling system; The multi-dimensional data after redundant data removal is input into a heat conduction model established based on Fourier's heat conduction law to predict the temperature distribution of the gas test transformer. The method includes: Based on the pre-stored physical structural parameters of the gas test transformer, the interior of the gas test transformer is divided into three-dimensional grid cells: the sulfur hexafluoride gas area, the transformer shell metal layer area, and the cooling pipe area, and an initial heat transfer coefficient is assigned to each cell. Based on the differential form of Fourier's heat conduction law Establish a heat conduction model, iteratively calculate the heat flux density vector of each grid unit in the heat conduction model, and solve the temperature gradient equation corresponding to each unit by gradient descent method , and generate the corresponding real-time temperature distribution map; among them, is the direction and magnitude of heat flow through a unit area per unit time; is the thermal conductivity coefficient corresponding to the grid unit, the unit is W / (m·K), the larger the value, the stronger the thermal conductivity; is the temperature field in the grid cell; is the temperature gradient; is the heat source term; Input the transformer winding temperature into the heat conduction model to initialize the temperature field of the heat conduction model; Input the cooling medium flow rate into the heat conduction model and dynamically adjust the heat conduction coefficient corresponding to the cooling pipe area; The load current and voltage are converted into the heating power of the winding and used as the heat source input of the temperature gradient equation to adjust the temperature field in the sulfur hexafluoride gas area; Input the ambient temperature and humidity into the heat conduction model to adjust the heat dissipation boundary conditions of the metal layer area of ​​the transformer shell; input the variable frequency fan speed into the heat conduction model to adjust the heat dissipation boundary conditions of the metal layer area of ​​the transformer shell; The heat conduction model outputs a three-dimensional temperature cloud map of the transformer winding, transformer housing, and cooling medium, and marks areas where the temperature exceeds the preset temperature threshold; Secondly, according to the temperature distribution, corresponding PID parameters are extracted from a pre-stored temperature distribution and PID parameter correspondence table, and the extracted PID parameters are optimized based on a pre-stored LMS algorithm to adjust the operating state of the transformer supporting water cooling system. The method includes: According to the high temperature area position and temperature gradient value marked in the three-dimensional temperature cloud map, the matching load differential coefficient, integral coefficient and threshold setting combination are retrieved from the pre-stored temperature distribution and PID parameter correspondence table; The extracted PID parameters are adjusted based on the pre-stored LMS algorithm, and the weight coefficients are updated in real time by minimizing the mean square error between the current temperature distribution and the target temperature distribution; According to the optimized PID parameters, the cooling medium flow rate of the transformer supporting water cooling system is adjusted synchronously and variable frequency fan speed in the water cooling system .

2. The water cooling control method for a gas test transformer according to claim 1, characterized in that: The multi-dimensional parameters of the gas test transformer and the transformer's supporting water cooling system are collected in real time through a sensor network. The multi-dimensional parameters include transformer temperature, load voltage and current, transformer winding temperature, cooling medium temperature, cooling medium flow rate, ambient temperature and humidity, and variable frequency fan speed in the water cooling system. The method includes: Use PT100 platinum resistance sensor to collect transformer surface temperature ; Using fiber grating sensor to collect winding temperature ; The load current is collected by Rogowski coil , use voltage transformer to collect load voltage ; Use thermocouple sensor to collect cooling medium temperature ; Use ultrasonic flowmeter to collect cooling medium flow rate ; Use digital temperature and humidity sensor to collect ambient temperature and relative humidity ; Use Hall effect sensors to collect the speed of variable frequency fans in water cooling systems .

3. The water cooling control method for a gas test transformer according to claim 2, characterized in that: The host computer performs feature selection on the preprocessed multi-dimensional parameters based on L1 regularization to eliminate redundant data. The method includes: The preprocessed multi-dimensional parameters are combined into feature vectors , each dimension corresponds to a collected parameter and is standardized; Construct an objective function including an L1 regularization term, solve the eigenvector by minimizing the weighted sum of the prediction error and parameter sparsity, and iteratively update the weight corresponding to the eigenvector using the coordinate descent method; The features are sorted according to their absolute values, and features with weights less than the preset weight threshold are eliminated, while the features with the strongest correlation with the temperature distribution are retained.

4. The water cooling control method for a gas test transformer according to claim 3, characterized in that: The objective function including the L1 regularization term is: ； Where n is the number of samples; is the ith sample feature vector; is the target value of the i-th sample; Mean squared error loss function; is the feature weight vector with dimension m; is the regularization parameter; is the L1 norm penalty term.

5. The water cooling control method for a gas test transformer according to claim 3, characterized in that: The extracted PID parameters are adjusted based on the pre-stored LMS algorithm, and the weight coefficients are updated in real time by minimizing the mean square error between the current temperature distribution and the target temperature distribution. The method includes: Based on the real-time 3D temperature cloud map output by the heat conduction model, the actual temperature value of each grid cell is extracted and compared point by point with the pre-stored target temperature distribution data; The difference between the current temperature distribution and the ideal state is quantified by calculating the mean square error of all units. The calculation formula is: , where N is the total number of network units, is the predicted temperature value of the i-th unit, is the target temperature threshold corresponding to the i-th unit; and integrates the input signal vector u(n) under the current working condition, including load current, cooling medium flow rate and ambient temperature and humidity data; Based on the pre-stored LMS algorithm, the weight vector w(n) of the PID parameters is iteratively updated according to the mean square error e(n) and the input signal vector u(n).

6. The water cooling control method for a gas test transformer according to claim 5, characterized in that: The formula for iteratively updating the weight vector w(n) of PID parameters is: ,in, It is the step size parameter in the LMS algorithm, which controls the amplitude of weight update.

7. A terminal, characterized in that: include: processor; a memory for storing execution instructions of the processor; The processor is configured to execute the method according to any one of claims 1 to 6.

8. A computer-readable storage medium storing a computer program, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images