Heat dissipation system automatic control method based on digital twinborn model of high-voltage electrical cabinet

Abstract

The invention relates to the technical field of electrical cabinet intelligent control, in particular to an automatic control method of a heat dissipation system based on a high-voltage electrical cabinet digital twinborn model, the system comprises a high-voltage cabinet, an electrical apparatus element and a fan assembly, and the fan assembly can realize air draft and air supply operation and is provided with an air guide assembly to change an airflow gathering area. The control method comprises the following steps: constructing a digital twinborn model and a temperature prediction model to realize temperature prediction in the cabinet; further constructing an oxidation-temperature rise prediction model, generating an in-cabinet temperature-oxidation hot spot distribution diagram in combination with the temperature prediction model, and performing potential overheating area analysis; according to the analysis result, the rotating speed of the fan assembly and the deflection angle of the air guide assembly are dynamically adjusted, and airflow distribution in the overheating area is optimized; in addition, the system can automatically adjust the operation state of the fan according to the blockage condition of the filter screen to achieve cleaning, and efficient operation of the heat dissipation system is ensured.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for electrical cabinets, and in particular to an automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet. Background Technology

[0002] In existing technologies, to prevent excessively high temperatures inside electrical cabinets, cooling fans are typically installed on the top or sides of the cabinet to dissipate heat. However, traditional methods for controlling the heat dissipation of electrical cabinets have some problems that urgently need to be addressed.

[0003] Traditional heat dissipation control systems often employ timed start / stop or temperature threshold control, meaning the fan starts and stops according to a set time or temperature threshold. This method does not fully consider the dynamic changes in temperature inside the cabinet and the impact of load, easily leading to large temperature fluctuations, which in turn affects the working performance and reliability of electrical components.

[0004] Chinese Patent Publication No. CN202411585720.9 discloses an intelligent electrical cabinet with stable heat dissipation function and its heat dissipation control method. First, load power data, temperature data, and heat dissipation power are acquired. Then, based on the changing characteristics of historical load power data, future power change trend parameters are predicted. Next, based on the fluctuation characteristics of historical temperature data, heat dissipation effect parameters are obtained. Finally, based on the current power change trend parameters and heat dissipation effect parameters, combined with the current heat dissipation power, a predicted heat dissipation power is obtained. Finally, based on the predicted heat dissipation power, the heat dissipation of the intelligent electrical cabinet is controlled. By predicting the changing trend of the load power that the electrical cabinet may bear and evaluating the current heat dissipation effect of the electrical cabinet, the predicted heat dissipation power is obtained, and the heat dissipation of the intelligent electrical cabinet is controlled, allowing for timely adjustment of the heat dissipation power, improving heat dissipation stability, and ensuring the stable and safe operation of the electrical cabinet.

[0005] While the aforementioned patent documents improve heat dissipation stability to some extent by predicting load power changes and adjusting the current electrical cabinet's cooling system accordingly, the complex and variable internal environment of electrical cabinets in actual engineering applications makes cavitation of electrical contacts due to long-term operation a key factor affecting heat dissipation. Specifically, under the long-term influence of multiple factors such as arc erosion and environmental corrosion, an oxide layer gradually forms on the surface of the electrical contacts inside the cabinet. This oxide layer not only increases the contact resistance between contacts, leading to localized heat accumulation, but also causes localized temperature rises inside the cabinet. Circuit breaker contacts, as critical components in electrical cabinets, are particularly prone to oxidation due to frequent switching operations. However, existing technologies rarely include real-time monitoring of the oxidation level of vibrating electrical contacts and precise control of the cooling system based on this monitoring, which limits further improvements in the heat dissipation performance of electrical cabinets. Summary of the Invention

[0006] The purpose of this invention is to provide an automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet, so as to solve the technical problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: An automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet, the automatic control method comprising: S1. Constructing a digital twin model: Analyze the structural information, electrical component parameters, thermodynamic parameters and gas flow information of the high-voltage switchgear, construct an electromagnetic-thermal coupling model, and use digital twin technology to construct a digital twin model; S2. Construct a temperature prediction model: Real-time acquisition and preprocessing of the internal operating data of the high-voltage switchgear, generation of internal temperature distribution map of the high-voltage switchgear using finite element simulation technology, and construction of temperature prediction model using ELM algorithm to realize internal temperature prediction of the high-voltage switchgear and obtain relevant prediction information. S3. Construct an oxidation-temperature rise prediction model: Monitor the contact temperature of circuit breakers / busbars in real time, combine contact resistance and current data to obtain contact oxidation dataset, construct an oxidation-temperature rise prediction model and train the model. S4. Construct a hotspot distribution map: Couple the temperature prediction model with the oxidation-temperature rise prediction model to construct a hotspot distribution map of the internal temperature of the high-voltage cabinet, conduct potential overheating area analysis, and adjust the heat dissipation system and provide early warning based on the analysis results.

[0008] Preferably, the construction of the digital twin model includes the following steps: S101. Obtain the overall structural parameters of the high-voltage switchgear by importing the 3D model and analyzing the structural drawings; S102. Extract the information of electrical components inside the high-voltage cabinet, and refer to the high-voltage cabinet equipment specification to obtain the operating parameters and thermodynamic parameters of each electrical component, and calculate the heat generation data of each component inside the high-voltage cabinet accordingly. S103. Extract the heat dissipation structure features from the high-voltage cabinet structure, analyze the gas flow direction inside the high-voltage cabinet based on this, construct a gas flow model, and simulate the airflow distribution under two conditions: natural convection and forced ventilation. S104. Based on Maxwell's equations and the heat conduction equation, establish an electromagnetic field-temperature field coupling model to simulate the Joule heating effect and local temperature rise generated when current passes through conductive components. S105. Using the above structural information, physical parameters and physical field model, construct a virtual twin of the high-voltage switchgear in the digital twin platform to achieve real-time data synchronization and state mapping with the actual high-voltage switchgear.

[0009] Preferably, the construction of the temperature prediction model includes the following steps: S201. Install multiple sets of sensors inside the high-voltage cabinet to achieve real-time acquisition of multi-dimensional parameters of key parts; S202. Clean the raw data, eliminate noise and outliers, and improve the quality of model input; S203. Based on the finite element simulation platform and combined with the electromagnetic-thermal coupling model, input the current operating condition parameters to simulate the internal temperature field of the high-voltage cabinet and obtain a high-precision temperature distribution map. S204. Train the model using the ELM algorithm on historical monitoring data and simulation data; S205. Use cross-validation to evaluate the model's prediction accuracy and combine it with a feedback mechanism to optimize model parameters, thereby improving prediction stability and generalization ability.

[0010] Preferably, the input layer of the ELM algorithm consists of current, ambient temperature, and historical temperature sequences; the hidden layer uses randomly generated weights and biases, and the activation function is Sigmoid; the output layer is the predicted temperature.

[0011] Preferably, the construction of the oxidation-temperature rise prediction model includes the following steps: S301. Install a monitoring device at the connection point of electrical components to continuously monitor the temperature change of the contact point; S302. Obtain the contact resistance value and the current value flowing through the contacts of the circuit breaker in the closed state, as input parameters for the oxidation model; S303. Based on multi-dimensional data such as contact temperature, contact resistance, current, and ambient humidity, combined with laboratory accelerated oxidation test data, a contact oxidation database is constructed to indicate the degree of oxidation. S304. Use the neural network LSTM to train the oxidation-temperature rise prediction model; S305. Use cross-validation to evaluate model performance. First, pre-train on a simulation dataset, then fine-tune on a real-world dataset to optimize model structure parameters.

[0012] Preferably, the construction of the hotspot distribution map includes the following steps: S401. Couple the temperature prediction model with the oxidation prediction model to generate a temperature-oxidation hotspot distribution map inside the high-voltage cabinet. S402. Define temperature threshold and oxidation degree threshold, mark the area that exceeds the threshold at the same time, and define the area that exceeds the threshold at the same time as the overheated area; S403. By confirming the three-dimensional coordinates of the hot zone inside the high-voltage cabinet, dynamically adjust the speed of the fan assembly and the deflection angle of the air guide assembly to optimize the airflow distribution in the overheated area.

[0013] A heat dissipation system based on a digital twin model of a high-voltage electrical cabinet includes a high-voltage cabinet, which contains electrical components. The high-voltage cabinet is equipped with a fan assembly, which includes a first fan and a second fan symmetrically arranged at the top of the high-voltage cabinet, and a third fan and a fourth fan symmetrically arranged at the bottom of the high-voltage cabinet.

[0014] Preferably, each of the fan components is equipped with a controller, which is used to control the rotation direction of the fan component to realize the ventilation and air supply operations.

[0015] Preferably, each of the fan components is provided with an air guide component at the end facing the inside of the high-voltage cabinet. The air guide component is used to guide the air inlet / outlet direction of the fan component to change the accumulation area of ​​airflow inside the high-voltage cabinet.

[0016] Preferably, the air guiding assembly includes an array of air guiding plates, which are connected to a drive frame mounted on the fan assembly. The air guiding plates can be oscillated by the up-and-down movement of the drive frame. The drive frame is connected to a drive component located inside the high-voltage cabinet.

[0017] The technical effects and advantages of this invention are as follows: This invention achieves effective heat dissipation by coordinating the exhaust and supply of air by fans at different positions within the fan assembly and guiding airflow through the air guide assembly, thereby changing the concentration area. It can also dynamically adjust the fan speed and air guide angle based on a temperature prediction model to adapt to actual heat generation. Specifically, when the filter is clogged, the fan rotation direction is changed, using reverse airflow to clean the filter, while simultaneously increasing the relevant fan speed to prevent poor ventilation and the formation of localized eddies. Furthermore, by constructing an oxidation-temperature rise prediction model and generating a hotspot distribution map based on the temperature prediction model, the three-dimensional coordinates of overheated areas can be determined according to the degree of oxidation of electrical contacts, enabling targeted heat dissipation. Based on the location of the overheated area within the cabinet, the fan assembly speed and air guide angle are dynamically adjusted to optimize airflow distribution, and the air guide angle adjustment is based on the three-dimensional coordinate distribution of the overheated area, further improving heat dissipation efficiency. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the main process of the present invention; Figure 2 This is a schematic diagram illustrating the construction of the digital twin model of the present invention; Figure 3 This is a schematic diagram illustrating the construction of the temperature prediction model of the present invention; Figure 4 This is a schematic diagram illustrating the construction of the oxidation-temperature rise prediction model of this invention; Figure 5 This is a schematic diagram illustrating the construction of the hotspot distribution map of the present invention; Figure 6This is a schematic diagram of the first working state of the wind turbine assembly of the present invention; Figure 7 This is a schematic diagram of the second working state of the wind turbine assembly of the present invention; Figure 8 This is a schematic diagram of the third working state of the wind turbine assembly of the present invention; Figure 9 This is a schematic diagram of the fourth working state of the wind turbine assembly of the present invention; Figure 10 This is a schematic diagram of the fifth working state of the wind turbine assembly of the present invention; Figure 11 This is a schematic diagram of the sixth working state of the wind turbine assembly of the present invention.

[0019] In the picture: 100. High-voltage switchgear; 200. Fan assembly; 201. First fan; 202. Second fan; 203. Third fan; 204. Fourth fan; 300. Air guide assembly. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] Example 1 Reference Figures 6 to 11 As shown, the present invention provides a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet, including a high-voltage cabinet 100, electrical components installed inside the high-voltage cabinet 100, and a fan assembly 200 on the high-voltage cabinet 100. The fan assembly 200 includes a first fan and a second fan 202 symmetrically arranged on the top of the high-voltage cabinet 100, and a third fan 203 and a fourth fan 204 symmetrically arranged on the bottom of the high-voltage cabinet 100.

[0022] The fan has the same structural composition, mainly including a housing that is fixedly connected to the high-voltage cabinet 100. The housing includes a rectangular housing and a circular housing. A fan is installed inside the housing, and a filter screen is installed at the end of the housing away from the high-voltage cabinet 100.

[0023] In this embodiment, the housing is preferably a rectangular housing, and an array of fans are arranged inside it.

[0024] Each fan assembly 200 is equipped with a controller, which is used to control the rotation direction of the fan assembly 200 to realize the ventilation and air supply operations.

[0025] Reference Figures 6 to 11 As shown, each end of the fan assembly 200 facing the interior of the high-voltage cabinet 100 is provided with an air guide assembly 300. The air guide assembly 300 is used to guide the inlet / outlet air direction of the fan assembly 200 to change the accumulation area of ​​airflow inside the high-voltage cabinet 100.

[0026] The air guide assembly 300 includes an array of air guide plates, which are connected to a drive frame mounted on the fan assembly 200. The air guide plates are swayed by the up-and-down movement of the drive frame. The drive frame is connected to a drive component mounted inside the high-voltage cabinet 100.

[0027] The driving components include an electric push rod and a reciprocating mechanism consisting of a gear and rack. The swaying motion of the air guide plate is achieved by the up-and-down movement of the drive frame. This is existing technology and will not be described in detail here.

[0028] During operation, the control terminal located inside the high-voltage cabinet 100 controls the fan assembly 200 to start, so as to perform heat dissipation and cooling operations on the inside of the high-voltage cabinet 100.

[0029] During heat dissipation and cooling operations, the third fan 203 and the fourth fan 204, located at the bottom of the high-voltage cabinet 100, are controlled by their respective internal controllers to operate in a supply mode, drawing air from outside the high-voltage cabinet 100 and drawing it into the cabinet. Conversely, the first fan 201 and the second fan 202, located at the top of the high-voltage cabinet 100, are controlled by their respective internal controllers to operate in a suction mode, drawing air from inside the high-voltage cabinet 100 and expelling it from the cabinet. This achieves heat dissipation for the high-voltage cabinet 100. See [link to details] for further information. Figure 6 As shown.

[0030] Example 2 Although the above embodiments can achieve heat dissipation of the high-voltage switchgear 100, in actual applications, the operating state of the fan in the high-voltage switchgear 100 is generally in a fixed set state, and it is impossible to dynamically adjust the actual heat generation of the high-voltage switchgear 100 to adapt to the actual heat generation of the high-voltage switchgear 100. In view of this, the present invention provides an automatic control method for the heat dissipation system, as follows: Reference Figures 1 to 11 As shown, this invention provides an automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet. The automatic control method includes: S1. Constructing a digital twin model: Analyze the structural information, electrical component parameters, thermodynamic parameters and gas flow information of the high-voltage switchgear 100, construct an electromagnetic-thermal coupling model, and use digital twin technology to construct a digital twin model; S101. Obtain the overall structural parameters of the high-voltage switchgear 100 by importing the 3D model and analyzing the structural drawings. Structural parameters include the cabinet dimensions of the high-voltage switchgear 100, the location of the ventilation openings, the layout of the internal partitions, the arrangement of the busbars, and the installation location of the circuit breakers.

[0031] S102. Extract the information of electrical components inside the high-voltage cabinet 100, and refer to the equipment specification of the high-voltage cabinet 100 to obtain the operating parameters and thermodynamic parameters of each electrical component. Based on this, calculate the heat generation data of each component inside the high-voltage cabinet 100. Operating parameters include rated current, voltage level, operating frequency, load fluctuation range, and heating power; thermodynamic parameters include thermal conductivity, heat capacity, and surface emissivity.

[0032] S103. Extract the heat dissipation structure features from the structure of the high-voltage cabinet 100, analyze the gas flow direction inside the high-voltage cabinet 100 based on this, construct a gas flow model, and simulate the airflow distribution under the two conditions of natural convection and forced ventilation. S104. Based on Maxwell's equations and the heat conduction equation, establish an electromagnetic field-temperature field coupling model to simulate the Joule heating effect and local temperature rise generated when current passes through conductive components. S105. Using the above structural information, physical parameters and physical field model, construct a virtual twin of the high-voltage switchgear 100 in the digital twin platform to achieve real-time data synchronization and state mapping with the actual high-voltage switchgear 100.

[0033] S2. Construct a temperature prediction model: Real-time acquisition and preprocessing of operating data inside the high-voltage switchgear 100, generation of temperature distribution map inside the high-voltage switchgear 100 using finite element simulation technology, and construction of temperature prediction model using ELM algorithm to achieve temperature prediction inside the high-voltage switchgear 100 and obtain relevant prediction information. S201. Install multiple sets of sensors inside the high-voltage cabinet 100 to realize real-time acquisition of multi-dimensional parameters of key parts; Temperature, humidity, current, voltage and air pressure sensors are installed inside the high-voltage cabinet 100 to collect real-time operating data of key components.

[0034] S202. Clean the raw data, eliminate noise and outliers, and improve the quality of model input; Median filtering is used to eliminate spike noise caused by electromagnetic interference and moving average filtering is used to smooth instantaneous noise caused by load fluctuations. Based on the Z-score method, a threshold (e.g., |Z|≥3) is set to mark and isolate abnormal data points.

[0035] The processed data undergoes data synchronization and normalization. Specifically, the clocks of each sensor are synchronized via PTP (Precise Time Protocol) to ensure that temperature, current, and air pressure data are fused under the same time base, thereby achieving timestamp alignment. At the same time, non-temperature data (such as current and air pressure) are normalized using Min-Max, mapping them to the [0,1] interval for subsequent ELM model input.

[0036] It is important to note that for short-term data loss (such as single-cycle data loss), cubic spline interpolation is used. For example, if a temperature sensor loses two sampling points due to a communication failure, the data can be recovered by interpolating the preceding and following data. For long-term data loss (such as sensor failure), the system switches to backup sensor data. For example, when the main temperature sensor fails, thermocouple data from symmetrical locations is used.

[0037] S203. Based on the finite element simulation platform and combined with the electromagnetic-thermal coupling model, input the current operating condition parameters to simulate the temperature field inside the high-voltage cabinet 100 and obtain a high-precision temperature distribution map. Based on Maxwell's equations, the electromagnetic field distribution within a high-voltage switchgear 100 was simulated using ANSYS Maxwell electromagnetic field simulation software. The software obtains the current density distribution in the conductor by solving Maxwell's equations, and then uses the formula... Precise calculation of Joule heat in conductors.

[0038] Using the Joule heat obtained from electromagnetic field simulation as the heat source input, and combining the heat conduction equation and fluid dynamics equation, the conduction and convection processes of heat in the high-voltage cabinet 100 can be fully considered, and an electromagnetic-thermal coupling model can be constructed to provide a reliable basis for temperature prediction and heat dissipation optimization of the high-voltage cabinet 100.

[0039] Heat conduction equation: S204. Train the model using the ELM algorithm on historical monitoring data and simulation data; Based on the electromagnetic-thermal coupling model constructed in the digital twin model, real-time operating parameters are input, and a high-precision temperature distribution map of the high-voltage cabinet 100 is generated through finite element simulation.

[0040] Operating parameters include current load current and ambient temperature.

[0041] A temperature prediction model was established by training historical and simulation data using the Elastic Dynamic Model (ELM) algorithm. The input layer of the ELM consisted of current, ambient temperature, and historical temperature sequences; the hidden layer used randomly generated weights and biases, and the activation function was sigmoid; the output layer was the predicted temperature.

[0042] The output weights are solved using the least squares method with historical datasets to achieve fast training.

[0043] S205. Use cross-validation to evaluate the model's prediction accuracy, and combine feedback mechanisms to adjust the number of hidden layer nodes and activation function type of ELM to improve prediction stability and generalization ability.

[0044] In this embodiment, K-fold cross-validation is used to divide the historical data. In this example, the value of K is 5, that is, 4 folds of data are used for each training, and the remaining 1 fold is used for validation. Finally, the average mean squared error of 5 validations is taken as the model performance index.

[0045] After the model is deployed, dynamic optimization is achieved based on the sliding window update strategy: every hour, the weights of the hidden layers of the neural network are fine-tuned through the backpropagation algorithm using the real-time running data of the most recent hour to adapt to environmental disturbances (such as the impact of seasonal temperature and humidity changes on the heat dissipation of the high-voltage switchgear 100).

[0046] Under typical operating conditions such as rated load and 20% overload, the measured temperature of key components such as busbars is collected by an infrared thermal imager and compared with the model prediction value.

[0047] If the maximum deviation between the predicted temperature and the measured value does not exceed the set value, the engineering accuracy requirements are met. At this time, the fan assembly 200 distributed on the high-voltage cabinet 100 maintains its current working state, that is, the fan speed and direction remain unchanged, and the angle of the air guide plate remains unchanged.

[0048] If the maximum deviation between the predicted temperature and the measured temperature exceeds the set value, the engineering accuracy requirements are not met. In this case, the operating status of the fan assembly 200 distributed on the high-voltage cabinet 100 is adaptively adjusted. Specifically, if the predicted temperature value is less than the measured value, the fan speed should be reduced while the other operating statuses remain unchanged; if the predicted temperature value is greater than the measured value, the fan speed should be increased while the other operating statuses remain unchanged.

[0049] The steps for adjusting the fan speed are as follows: First, based on the temperature prediction model, the temperature rise rate of the high-voltage switchgear 100 within a set time period is determined. : in C represents the specific heat capacity.

[0050] Then, by reversing the heat dissipation equation, the required increase in heat transfer coefficient is calculated. : Finally, based on the fan performance curve (airflow-speed-static pressure relationship), a mathematical model of h (convective heat transfer coefficient) and N (fan speed) is established to determine whether the fan speed needs to be increased or decreased: Where k is the characteristic constant of the wind turbine, which needs to be calibrated experimentally.

[0051] Meanwhile, the digital twin platform of this application includes a three-dimensional model of the fan assembly 200, including the geometric parameters of key components such as the fan, filter, and motor. A current sensor is installed at the fan motor, and air pressure sensors are arranged at the fan inlet / outlet. These sensors are used to collect the motion parameters of the fan assembly 200 and input the real-time collected fan parameters into the digital twin model to drive the operation of the physical model of the fan.

[0052] The actual operating parameters of the actual wind turbine component 200 are compared with the expected parameters simulated in the digital twin model to determine the operating status of the wind turbine.

[0053] If the air pressure difference threshold continues to exceed the set value and the flow rate at the fan inlet decreases, a blockage warning is triggered. When the difference between the measured current value and the predicted value exceeds the set value and the fan speed decreases, it is determined that the filter is blocked. When the filter is blocked, the controller controls the fan assembly 200 to change the rotation direction in order to clean the filter.

[0054] Specifically, when the filters of the third fan 203 and the fourth fan 204 become clogged, the fans of the first fan 201 and the second fan 202 located at the top of the high-voltage cabinet 100 are controlled to rotate in opposite directions, changing their operation from exhaust to supply air. This means that air outside the high-voltage cabinet 100 is drawn into the cabinet. Conversely, the third fan 203 and the fourth fan 204 located at the bottom of the high-voltage cabinet 100 are controlled by their respective internal controllers to be in exhaust mode, drawing air from inside the high-voltage cabinet 100 and expelling it. The airflow impacts the filters on the third fan 203 and the fourth fan 204 in a reverse manner, thus cleaning the filters. (See details below.) Figure 7 As shown.

[0055] It should be noted that when the first fan 201 and the second fan 202 are in the air supply state, and the third fan 203 and the fourth fan 204 are in the air exhaust state, the speed of the first fan 201 and the second fan 202 should be increased. Preferably, the speed of the first fan 201 and the second fan 202 should be increased by 20% on the original basis. This increases the upper air pressure of the high-voltage cabinet 100 to overcome the blockage resistance and avoid insufficient positive pressure inside the high-voltage cabinet 100, which would prevent the third fan 203 and the fourth fan 204 from effectively exhausting air and forming local vortices.

[0056] Example 3 Although the above embodiments can automatically adjust the motion parameters of the fan assembly 200 according to the temperature prediction model, in actual applications, oxidation of electrical contacts can also cause local temperature rise inside the high-voltage cabinet 100. However, due to the complex and variable internal environment of the high-voltage cabinet 100, the existing technology cannot determine the three-dimensional coordinates of the overheated area based on the degree of oxidation of the electrical contacts, and therefore cannot carry out targeted heat dissipation treatment. In view of this, this application makes technical improvements based on Embodiment 2, and the improved technical solution is as follows: Reference Figures 1 to 11 As shown, this invention provides an automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet. The automatic control method includes: S1. Constructing a digital twin model: Analyze the structural information, electrical component parameters, thermodynamic parameters and gas flow information of the high-voltage switchgear 100, construct an electromagnetic-thermal coupling model, and use digital twin technology to construct a digital twin model; S2. Construct a temperature prediction model: Real-time acquisition and preprocessing of operating data inside the high-voltage switchgear 100, generation of temperature distribution map inside the high-voltage switchgear 100 using finite element simulation technology, and construction of temperature prediction model using ELM algorithm to achieve temperature prediction inside the high-voltage switchgear 100 and obtain relevant prediction information. S3. Construct an oxidation-temperature rise prediction model: Monitor the contact temperature of circuit breakers / busbars in real time, combine contact resistance and current data to obtain contact oxidation dataset, construct an oxidation-temperature rise prediction model and train the model. S301. Install a monitoring device at the connection point of electrical components to continuously monitor the temperature change of the contact point; For the connection points of electrical components, i.e., near the contacts of circuit breakers / busbars, install non-contact temperature sensors suitable for the high-voltage switchgear environment, such as infrared temperature sensors. Infrared temperature sensors are based on the principle that the infrared radiation energy emitted by an object is related to its temperature, and can accurately measure the temperature change of the surface of the object being measured without contact.

[0057] S302. Obtain the contact resistance value and the current value flowing through the contacts of the circuit breaker in the closed state, as input parameters for the oxidation model; This non-contact contact resistance measurement method, based on the principle of electromagnetic induction, involves installing a specially designed induction coil around the circuit breaker / busbar. When current flows through the contacts, it generates an alternating magnetic field in the surrounding space. The induction coil senses the change in the magnetic field and converts it into an electrical signal. By analyzing and processing this electrical signal, combined with known circuit parameters and electromagnetic theory models, the contact resistance value of the contacts can be calculated.

[0058] The current value flowing through the contacts can be obtained using the existing current transformer in the high-voltage switchgear 100 or by installing a dedicated non-contact current sensor. Non-contact current sensors, based on electromagnetic induction or the Hall effect, can accurately measure the current magnitude without direct contact with the current circuit being measured.

[0059] The output signals of the infrared temperature sensor, contact resistance measuring device, and current sensor are connected to the data acquisition system to achieve synchronous acquisition of temperature, contact resistance, and current data.

[0060] It should be noted that the calculation of contact resistance is mainly based on Joule's law ( ), contact resistance With temperature rise They show a positive correlation. Contact temperature is monitored using an infrared thermal imager, combined with thermal resistance parameters from a digital twin model. Inverting contact resistance: S303. Based on multi-dimensional data such as contact temperature, contact resistance, current, and ambient humidity, combined with laboratory accelerated oxidation test data, a contact oxidation database is constructed to indicate the degree of oxidation. In addition to real-time acquisition of contact temperature, contact resistance, and current data, ambient humidity data is also required. This can be achieved by installing a humidity sensor inside the high-voltage switchgear 100 to monitor ambient humidity in real time. Integrating multi-dimensional data such as temperature, contact resistance, current, and ambient humidity forms a complete dataset that describes the state of the contacts under different operating conditions.

[0061] To determine the relationship between the degree of contact oxidation and multidimensional data, accelerated oxidation tests were conducted in the laboratory. The actual operating environment of the high-voltage switchgear 100 was simulated in the laboratory, with different conditions for temperature, humidity, current, and contact resistance, to perform accelerated oxidation tests on the contact samples. During the tests, the degree of oxidation of the contact samples was measured periodically using non-contact detection methods.

[0062] Data obtained from accelerated oxidation experiments in the laboratory is correlated with real-time acquired multidimensional data to construct a contact oxidation database. Within this database, each data record is labeled with its corresponding contact oxidation degree, forming a complete labeled dataset. This database will provide rich sample data for training the oxidation-temperature rise prediction model, helping the model learn the relationship between contact oxidation degree and factors such as temperature, contact resistance, current, and humidity.

[0063] S304. Use the neural network LSTM to train the oxidation-temperature rise prediction model; A neural network LSTM is used as the architecture for the oxidation-temperature rise prediction model. In the high-voltage switchgear 100, parameters such as contact temperature, contact resistance, current, and humidity are time-varying sequential data, and the oxidation process of the contacts is also a long-term, cumulative process. The LSTM network can remember historical information and make predictions based on the current input and historical states.

[0064] The data in the contact oxidation database was preprocessed, and the preprocessed dataset was divided into training, validation, and test sets. The training data was input into an LSTM model, and the model parameters were continuously adjusted using the backpropagation algorithm to minimize the error between the model's predicted output and the actual labeled oxidation degree and temperature rise value. During training, appropriate optimization algorithms were employed to accelerate model convergence.

[0065] The model's input consists of real-time collected data on current, contact resistance, temperature, and humidity, which are fed into the LSTM model in time series format. The model's output is the degree of contact oxidation and the corresponding predicted temperature rise. The degree of oxidation can be represented by a continuous numerical value or discretized; the predicted temperature rise is the increase in contact temperature relative to the ambient temperature under the current operating conditions.

[0066] S305. The model performance is evaluated by cross-validation. It is first pre-trained on the simulation dataset and then fine-tuned on the actual test dataset to optimize the model structure parameters and improve the accuracy and stability of oxidation prediction.

[0067] Cross-validation is used to evaluate the model's performance. The dataset is divided into k subsets. Each time, k-1 subsets are selected as the training set, and the remaining subset is used as the validation set. This process is repeated k times. The model's generalization ability and stability are evaluated by calculating the average of the k validation results, such as mean absolute error (MAE) and mean squared error (MSE). Cross-validation makes full use of the dataset and reduces evaluation bias caused by improper data partitioning.

[0068] Given the limited amount of measured data, the model was first pre-trained using a simulation dataset. Based on the actual operating principle and contact oxidation mechanism of the high-voltage switchgear 100, a simulation model was established, generating a large amount of simulation data. This simulation data included information such as contact temperature, contact resistance, current, humidity, and oxidation degree under different operating conditions. The simulation data was then input into the LSTM model for pre-training, allowing the model to initially learn the relationship between contact oxidation and various parameters, thus improving the model's initial performance.

[0069] After pre-training on the simulation dataset, the model was fine-tuned using the real-world dataset. The real-world dataset was divided into training and validation sets using cross-validation. The pre-trained model parameters were used as initial parameters for further training and optimization on the real-world dataset. Through fine-tuning, the model can better adapt to the operating environment and data characteristics of actual equipment, improving the accuracy and stability of oxidation prediction.

[0070] During model training and evaluation, the model's structural parameters are optimized based on the validation results. For example, hyperparameters such as the number of layers in the LSTM network, the number of neurons per layer, and the learning rate are adjusted to find the optimal model structure. Hyperparameter optimization algorithms such as grid search and random search can be used to automatically find the optimal parameter combinations and improve model performance.

[0071] S4. Construct a hotspot distribution map: Couple the temperature prediction model with the oxidation-temperature rise prediction model to construct a temperature hotspot distribution map inside the high-voltage cabinet 100, conduct potential overheating area analysis, and adjust the heat dissipation system and provide early warning based on the analysis results.

[0072] S401. Couple the temperature prediction model with the oxidation prediction model to generate a temperature-oxidation hotspot distribution map inside the high-voltage cabinet 100. Align the outputs (predicted temperature value and predicted oxidation degree value) of the temperature prediction model and the oxidation prediction model by timestamp and spatial coordinates to ensure that the prediction results at the same time and physical location are correlated. For example, the predicted temperature value of contact A is bound to the predicted oxidation degree value of contact A.

[0073] Based on model accuracy and business priority, weights are allocated, with the weight of the temperature prediction model being preferred over that of the oxidation prediction model. The weighting is 70%, and the oxidation prediction model ( The weighting is 30%.

[0074] Calculate a risk score for each area (or contact) within the high-voltage switchgear: in, To predict temperature, =65℃; For the degree of oxidation (e.g., contact resistance increment), O threshold = 20%; WT = 0.7, WO = 0.3.

[0075] The risk score is mapped to the 100-dimensional digital twin model of the high-voltage switchgear, and the overheated area is marked by color gradient.

[0076] S402. Define temperature threshold and oxidation degree threshold, mark the area that exceeds the threshold at the same time, and define the area that exceeds the threshold at the same time as the overheated area; First, a temperature threshold is set based on industry standards and experimental data, and an oxidation threshold is determined through accelerated aging tests of contacts. The preferred temperature threshold is 65°C, as exceeding this temperature may cause insulation aging. The oxidation threshold is determined through accelerated aging tests of contacts, and the preferred oxidation threshold is when the contact resistance increases by 20%.

[0077] Next, the internal space of the high-voltage switchgear 100 is divided into spatial grid regions (the grid size is preferably 10cm×10cm×10cm), with each grid center point corresponding to a predicted position; at the same time, the actual three-dimensional coordinates of the circuit breaker and busbar contacts are marked in the digital twin model.

[0078] Then, a risk assessment is performed on the grid or contact point coordinates. If the risk score of a certain grid or contact point exceeds the first set value, it is marked as a level 1 overheating zone; if the risk score does not exceed the first set value but exceeds the second set value, it is marked as a level 2 overheating zone.

[0079] It is important to note that for the primary overheating zone, the system sends its 3D coordinates to the backend control center via network connection for early warning, allowing staff to address contact oxidation areas based on these coordinates. Simultaneously, the system expands a grid outwards from the primary overheating zone to create a continuous overheating area, preventing the omission of edge risks. The number of grids expanding outwards is between 1 and 5.

[0080] S403. By confirming the three-dimensional coordinates of the overheated area inside the high-voltage cabinet 100, the speed of the fan assembly 200 and the deflection angle of the air guide assembly 300 are dynamically adjusted to optimize the airflow distribution in the overheated area.

[0081] In this application, the internal area of ​​the high-voltage switchgear 100 is divided into S1-S4 areas, which correspond to the upper left, upper right, lower left and lower right positions of the high-voltage switchgear 100, respectively.

[0082] As shown in the figure, when the primary superheated zones are concentrated in regions S2 and S4, but the number of primary superheated zones in region S4 is greater than that in region S2, adaptive adjustments are made by controlling the fan assembly 200 on the high-voltage switchgear 100. The specific adjustment strategy is as follows: By adjusting the position of the drive frame, the guide vane on the third fan 203 is tilted towards the top of the high-voltage cabinet 100, and the guide vane on the second fan 202 is tilted towards the bottom of the high-voltage cabinet 100. This causes the airflow delivered by the third fan 203 to be guided by the guide vanes towards the second fan 202. The second fan 202, guided by the guide vanes, draws out the airflow from the direction of the third fan 203, so that the airflow is concentrated more in the S2 and S4 regions (see details). Figure 8 (As shown).

[0083] When adjusting the air guide plate, the temperature rise rate of regions S2 and S4 within a set time is determined by the temperature prediction model; the required increase in heat transfer coefficient is calculated by solving the heat dissipation equation; and a mathematical model of convective heat transfer coefficient and fan speed is established based on the fan performance curve to determine the required increase in speed value for the second fan 202 and the third fan 203, and adjustments are made accordingly.

[0084] If the number of primary superheated zones in region S4 is less than that in region S2, the supply / exhaust direction of the high-voltage cabinet 100 needs to be changed. That is, the first fan 201 and the second fan 202 are switched to supply mode, and the third fan 203 and the fourth fan 204 are switched to exhaust mode. After the supply / exhaust mode is switched, the speed of the first fan 201 and the second fan 202 is increased to increase the upper air pressure of the high-voltage cabinet 100, so as to avoid insufficient positive pressure inside the high-voltage cabinet 100, which would prevent the third fan 203 and the fourth fan 204 from effectively exhausting air and forming a local vortex dead zone.

[0085] Simultaneously, by adjusting the position of the drive frame, the guide vane on the first fan 201 is tilted towards the bottom of the high-voltage cabinet 100, and the guide vane on the fourth fan 204 is tilted towards the top of the high-voltage cabinet 100. This causes the airflow delivered by the first fan 201 to be guided by the guide vanes towards the fourth fan 204. The fourth fan 204, guided by the guide vanes, draws out the airflow from the direction of the third fan 203, thus concentrating the airflow more in the S2 and S4 regions (see details). Figure 10 (As shown).

[0086] When adjusting the air guide plate, the temperature rise rate of regions S2 and S4 within a set time is determined by the temperature prediction model; the required increase in heat transfer coefficient is calculated by solving the heat dissipation equation; and a mathematical model of convective heat transfer coefficient and fan speed is established based on the fan performance curve to determine the required increase in speed value for the first fan 201 and the fourth fan 204, and adjustments are made accordingly.

[0087] For cases where the primary superheated zones are concentrated in regions S1 and S3, but the number of primary superheated zones in region S3 is greater or less than that in region S1, adjust the fan speed and the deflection direction of the guide vanes as described above. (Specific details are not provided in the original text.) Figure 9and Figure 11 As shown.

[0088] It should be noted that the adjustment of the air guide plate angle of each fan in this application is not a fixed angle adjustment, but based on the three-dimensional coordinate distribution of the overheated zone. By adjusting the angle of the air guide plate, the airflow can cover the three-dimensional coordinate distribution of the overheated zone, thereby improving the heat dissipation efficiency.

[0089] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0090] Although embodiments of the invention have been shown and described, those skilled in the art will recognize that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet, characterized in that, The automatic control method includes: S1. Constructing a digital twin model: Analyze the structural information, electrical component parameters, thermodynamic parameters and gas flow information of the high-voltage switchgear (100), construct an electromagnetic-thermal coupling model, and use digital twin technology to construct a digital twin model; S2. Constructing a temperature prediction model: Real-time acquisition and preprocessing of the internal operating data of the high-voltage cabinet (100), using finite element simulation technology to generate a temperature distribution map inside the high-voltage cabinet (100), and using the ELM algorithm to construct a temperature prediction model to realize the prediction of the internal temperature of the high-voltage cabinet (100) and obtain relevant prediction information. S3. Construct an oxidation-temperature rise prediction model: Monitor the contact temperature of circuit breakers / busbars in real time, combine contact resistance and current data to obtain contact oxidation dataset, construct an oxidation-temperature rise prediction model and train the model. S4. Construct a hotspot distribution map: Couple the temperature prediction model with the oxidation-temperature rise prediction model to construct a hotspot distribution map of the internal temperature of the high-voltage cabinet (100), conduct potential overheating area analysis, and adjust the heat dissipation system and provide early warning based on the analysis results.

2. The automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet according to claim 1, characterized in that, The construction of the digital twin model includes the following steps: S101. Obtain the overall structural parameters of the high-voltage switchgear (100) by importing the three-dimensional model and analyzing the structural drawings; S102. Extract the information of electrical components inside the high-voltage cabinet (100), and refer to the equipment specification of the high-voltage cabinet (100) to obtain the operating parameters and thermodynamic parameters of each electrical component, and calculate the heat generation data of each component inside the high-voltage cabinet (100). S103. Extract the heat dissipation structure features from the high-voltage cabinet (100) structure, analyze the gas flow direction inside the high-voltage cabinet (100) accordingly, construct a gas flow model, and simulate the airflow distribution under the two conditions of natural convection and forced ventilation. S104. Based on Maxwell's equations and the heat conduction equation, establish an electromagnetic field-temperature field coupling model to simulate the Joule heating effect and local temperature rise generated when current passes through conductive components. S105. Using the above structural information, physical parameters and physical field model, construct a virtual twin of the high-voltage switchgear (100) in the digital twin platform to achieve real-time data synchronization and state mapping with the actual high-voltage switchgear (100).

3. The automatic control method for the heat dissipation system based on the digital twin model of a high-voltage electrical cabinet according to claim 2, characterized in that, The construction of the temperature prediction model includes the following steps: S201. Install multiple sets of sensors inside the high-voltage cabinet (100) to realize real-time acquisition of multi-dimensional parameters of key parts; S202. Clean the raw data, eliminate noise and outliers, and improve the quality of model input; S203. Based on the finite element simulation platform and combined with the electromagnetic-thermal coupling model, input the current operating condition parameters to simulate the internal temperature field of the high-voltage cabinet (100) and obtain a high-precision temperature distribution map. S204. Train the model using the ELM algorithm on historical monitoring data and simulation data; S205. Use cross-validation to evaluate the model's prediction accuracy and combine it with a feedback mechanism to optimize model parameters, thereby improving prediction stability and generalization ability.

4. The automatic control method for the heat dissipation system based on the digital twin model of a high-voltage electrical cabinet according to claim 3, characterized in that, The input layer of the ELM algorithm consists of current, ambient temperature, and historical temperature sequences; the hidden layer uses randomly generated weights and biases, and the activation function is Sigmoid; the output layer is the predicted temperature.

5. The automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet according to claim 3, characterized in that, The construction of the oxidation-temperature rise prediction model includes the following steps: S301. Install a monitoring device at the connection point of electrical components to continuously monitor the temperature change of the contact point; S302. Obtain the contact resistance value and the current value flowing through the contacts of the circuit breaker in the closed state, as input parameters for the oxidation model; S303. Based on multi-dimensional data such as contact temperature, contact resistance, current, and ambient humidity, combined with laboratory accelerated oxidation test data, a contact oxidation database is constructed to indicate the degree of oxidation. S304. Use the neural network LSTM to train the oxidation-temperature rise prediction model; S305. Use cross-validation to evaluate model performance. First, pre-train on a simulation dataset, then fine-tune on a real-world dataset to optimize model structure parameters.

6. The automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet according to claim 5, characterized in that, The construction of the hotspot distribution map includes the following steps: S401. Couple the temperature prediction model with the oxidation prediction model to generate a temperature-oxidation hot spot distribution map inside the high-voltage cabinet (100). S402. Define temperature threshold and oxidation degree threshold, mark the area that exceeds the threshold at the same time, and define the area that exceeds the threshold at the same time as the overheated area; S403. By confirming the three-dimensional coordinates of the hot zone inside the high-voltage cabinet (100), the rotation speed of the fan assembly (200) and the deflection angle of the air guide assembly (300) are dynamically adjusted to optimize the airflow distribution in the overheated area.

7. A heat dissipation system based on a digital twin model of a high-voltage electrical cabinet, characterized in that, The high-voltage switchgear (100) according to any one of claims 1-6 is provided with electrical components installed inside the high-voltage switchgear (100) and a fan assembly (200) is provided on the high-voltage switchgear (100). The fan assembly (200) includes a first fan (201) and a second fan (202) symmetrically arranged on the top of the high-voltage switchgear (100) and a third fan (203) and a fourth fan (204) symmetrically arranged on the bottom of the high-voltage switchgear (100).

8. The automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet according to claim 7, characterized in that, Each of the fan components (200) is equipped with a controller, which is used to control the rotation direction of the fan component (200) to realize the ventilation and air supply operations.

9. The automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet according to claim 7, characterized in that, Each of the fan assembly (200) is provided with an air guide assembly (300) at one end facing the inside of the high-voltage cabinet (100). The air guide assembly (300) is used to guide the air inlet / outlet direction of the fan assembly (200) to change the accumulation area of ​​airflow inside the high-voltage cabinet (100).

10. The automatic control method for a heat dissipation system based on a digital twin model of a high-voltage electrical cabinet according to claim 9, characterized in that, The air guide assembly (300) includes an array of air guide plates, which are connected to a drive frame mounted on the fan assembly (200). The air guide plates are swayed by the up-and-down movement of the drive frame. The drive frame is connected to a drive component mounted inside the high-voltage cabinet (100).

Images