A high-voltage electromagnetic balancing power-saving device

Abstract

The present invention discloses a high-voltage electromagnetic balancing power-saving device, which relates to the field of high-voltage energy-saving technology. Compared with previous high-voltage line switchgear, the present invention solves the problem that existing high-voltage line switchgear or distribution cabinets have no power-saving device, have a single function, and lack electromagnetic balancing and on-load voltage stabilization functions, and cannot solve the problem of voltage fluctuation and power waste in the power grid line. The high-voltage power-saving device is added to the incoming distribution cabinet and switch cabinet of the high-voltage power supply enterprise, and the power distribution system is added with intelligent protection, voltage stabilization, power grid purification, remote monitoring, power-saving operation and other functions. The device can extend the power supply radius of the substation to suit long-distance power transmission in rural power grids, ensure the voltage qualification rate of the entire line, effectively reduce line loss and distribution transformer loss, and continue to operate without power outage when the voltage regulator is put into or out of operation. The voltage can also be adjusted automatically, manually or remotely according to the line load size, so that the entire power grid always operates in a safe, economical and efficient operating state.

Description

Technical Field

[0001] The present invention relates to the technical field of high-voltage energy saving, and in particular to a high-voltage electromagnetic balance power-saving device. Background Art

[0002] With the development of the electric power industry and the improvement of the national agricultural power grid transformation project, the previously low voltage at the end of the power grid has been generally improved. However, this has led to the widespread high voltage of the power grid. According to data recorded from 10 Class D voltage monitoring points in a certain area of ​​Zhejiang Province over a two-month period, the voltage exceeded 50.5% of the time, exceeded 1.4% of the time, and remained within acceptable limits 48.1% of the time. In particular, during periods of low load, voltage increases from 380V to 420V were common, resulting in increased losses in the power grid and low power quality. This has also increased power consumption for transformers and all other equipment used by businesses, leading to frequent motor burnout, shortened equipment lifespan, and increased maintenance and repair costs. Currently, switchgear and distribution cabinets for high-voltage lines lack energy-saving devices, are limited in functionality, and lack electromagnetic balancing and on-load voltage regulation, thus failing to address voltage fluctuations and energy waste in the power grid.

[0003] In order to solve the above problems, the present invention proposes a high-voltage electromagnetic balancing power-saving device. Summary of the Invention

[0004] The purpose of the present invention is to provide a high-voltage electromagnetic balancing power-saving device to solve the problems raised in the background technology:

[0005] Currently, the switch cabinets or distribution cabinets used for high-voltage lines on the market do not have energy-saving devices, their functions are single, and they do not have electromagnetic balance and load voltage stabilization functions, and cannot solve the problems of voltage fluctuations and energy waste in power grid lines.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A high-voltage electromagnetic balancing power-saving device, comprising: a three-dimensional wound iron core adjustable reactor, a safety protection automatic control module, a detection control module, a power quality management module, and an operation optimization management module;

[0008] The three-dimensional wound iron core adjustable reactor includes three identical single-frame rounded rectangular three-dimensional wound iron cores and three sets of reactor coils. The reactor coils include three sets of main winding reactor coils connected in series to the main circuit and three sets of autocoupling coil windings for parallel induction. The three-dimensional wound iron core adjustable reactor also adjusts the output voltage by tapping the autocoupling coil windings.

[0009] The safety protection automatic control module adopts microcomputer control to monitor the operation of the power supply line in real time. It also starts the corresponding protection program through the intelligent safety electricity fire alarm protector, and issues instructions to control the combined bypass knife switch to switch to the bypass power supply line, while sounding and lighting alarms;

[0010] The detection and control module is used to collect full power data, non-power signals and work log information, and summarize and integrate them;

[0011] The power quality management module is used to monitor the power supply quality of the power grid;

[0012] The operation optimization management module is used to provide continuous operation trend records, operation status records, startup voltage and current records, accident records, and insulation monitoring records for large electrical equipment, and optimize the system operation mode by summarizing the recorded information, and optimize the design capacity and operation arrangement of transformers, cables, switches, and motors;

[0013] The first ends of the three groups of main circuit reactors are respectively connected to the three-phase input ends of the high-voltage line, and the second ends of the three groups of main circuit reactors are respectively connected to the three-phase output ends of the high-voltage line; the first ends of the three groups of autocoupling coil windings are respectively connected to the second ends of the three groups of main circuit reactors; the three groups of autocoupling coil windings each have a neutral tap, and the neutral tap is connected to different gear positions of the voltage regulator to adjust the output voltage.

[0014] Preferably, the three-dimensional wound core is made of amorphous material, and its cross section is a polygonal structure.

[0015] Preferably, the detection control module is controlled by an intelligent control instrument, which includes a voltage signal feedback circuit, a current signal feedback circuit, an A / D module, a CPU, a display part, a communication interface, a remote control input and a remote control output; the voltage signal feedback circuit and the current signal feedback circuit are respectively connected to the A / D module, and the CPU is respectively connected to the A / D module, the display part, the communication interface, the remote control input and the remote control output.

[0016] Preferably, the CPU performs signal enhancement processing based on a convolutional neural network, and the input layer of the convolutional neural network extracts communication network signal features, specifically as follows:

[0017] The detected communication network signal X(k) is sampled by sampling to obtain the signal instantaneous amplitude sequence R(k):

[0018]

[0019] Where A(k) is the signal amplitude sequence obtained by sampling at the sampling frequency; k represents the length of the instantaneous amplitude sequence;

[0020] Get the maximum instantaneous amplitude F of the signal max The two characteristic parameters are the standard deviation σ of the absolute value of the instantaneous amplitude of the signal, and the calculation formula is as follows:

[0021]

[0022] in, represents the discrete welfare home transformation process; K represents the number of signal sampling points;

[0023] The convolution operation of the convolutional layer is as follows:

[0024]

[0025] Among them, y(i,j,l) is the output; x(i,j,l) is the input; d is the step size; n, m and s are the length ranges of different dimensions respectively; (i,j,l) corresponds to the multidimensional coordinate index; (h,w,r) is used to traverse the different dimensions of the filter; is the weight parameter of the lth filter; b l is the bias of the lth filter; ε is the expansion factor; f(·) is the activation function;

[0026] Use convolution kernel sizes of 1×1, 1×3, ... 1×2 N The +1 multi-scale dilated convolution operator connects feature signals of different scales and uses a zero-filled convolution operator to keep the output features and input features of the same size;

[0027] The convolutional neural network also incorporates an attention mechanism, as follows:

[0028] y out (g) = T Y [y(g)×im X (g)×im Y (g)×im Z (g)+y(g)]

[0029] Among them, y out (g) is the output of the attention layer; y(g) is the output of the convolution layer; T Y Indicates the pooling operation along the Y direction; im X (g), im Y (g) and im Z (g) Indicates the importance of different directions.

[0030] Preferably, the safety protection automatic control module also takes the power supply line operation status data as input and performs fault prediction and voltage regulation based on the convolutional neural network.

[0031] Preferably, the safety protection automatic control module is also provided with a lightning protection function.

[0032] Preferably, the power quality data monitored by the power quality management module include grid voltage phase loss, voltage imbalance, harmonic measurement and line crossing, harmonic distortion rate, voltage overshoot and undershoot, voltage fluctuation, and abnormal frequency fluctuation data.

[0033] Preferably, the operation optimization management module comprehensively collects continuous operation trend records, operation status records, startup voltage and current records, accident records, and insulation monitoring record data of large-scale electrical equipment, and optimizes the system operation mode based on an improved harmony optimization strategy. The improved harmony optimization strategy is specifically as follows:

[0034] The energy saving P={p1,p2,Λ,p Q} As the optimization target, Q is the total number of operation arrangement combinations of each device, and the relevant parameters are initialized;

[0035] p uv =p min +(p max -p min )×r1

[0036] Among them, p uv Arrange vector for initializing the operation of u-th device and v-th device; p max and p min are the upper and lower limits of the solution; r1 is a random number between (0,1);

[0037] Initialize the operation arrangement combination library GM, randomly generate q operation arrangement vectors in the search domain according to the above formula, and add them to the operation arrangement combination library GM shown by the matrix:

[0038]

[0039] Among them, FL[·] is the energy saving measurement function;

[0040] Update the operation schedule vector p before the e-th iteration uv (e) Update as follows:

[0041]

[0042] Among them, r2, r3, r4 and r5 are all random numbers between (0,1); p new,uv (e) is the updated operation arrangement vector after the e-th iteration; α is the iteration factor; E is the maximum number of iterations; BW is the combination limit of the operation arrangement of the e-th iteration; p1 is the storage probability of the operation arrangement combination library; p2(e) is the fine-tuning probability of the operation arrangement of the e-th iteration, p 2max and p2min Arrange the maximum and minimum values ​​of the fine-tuning probability for the run respectively;

[0043] The updated operation schedule vector p new,uv (e) The worst harmony vector p in the combined memory GM of the initial run arrangement worst The harmony metric function value is compared, if p new,uv The metric function value of (e) is less than p worst The metric function value is p new,uv (e) Replace p worst ; Otherwise, the run schedule vector in GM remains unchanged;

[0044] Check whether the algorithm termination condition is met; if so, output the global optimal solution and the algorithm ends; otherwise, return to continue updating the operation schedule vector.

[0045] Compared with the prior art, the present invention provides a high-voltage electromagnetic balancing power-saving device with the following beneficial effects:

[0046] The present invention adds a high-voltage power-saving device to the incoming distribution cabinet and switch cabinet of the high-voltage power supply enterprise, and adds functions such as intelligent protection, voltage stabilization, power grid purification, remote monitoring, and power-saving operation to the power distribution system; it can extend the power supply radius of the substation to be suitable for long-distance power transmission in rural power grids, ensure the voltage qualification rate of the entire line, effectively reduce line loss and distribution transformer loss, and continue to operate without power outage when the voltage regulator is put into or out of operation. It can also automatically, manually or remotely adjust the voltage according to the line load size, so that the entire power grid always operates in a safe, economical and efficient operation state. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a schematic block diagram of the device mentioned in Example 1 of the present invention;

[0048] Figure 2 Schematic diagram of the adjustable reactor core mentioned in Example 1 of the present invention;

[0049] Figure 3 This is a schematic diagram of the connection of the reactor coil mentioned in Example 1 of the present invention;

[0050] Figure 4 This is a schematic diagram of the detection control module mentioned in Example 1 of the present invention;

[0051] Figure 5 Schematic diagram of the convolution structure mentioned in Example 1 of the present invention;

[0052] Figure 6 Schematic diagram of the attention mechanism mentioned in Example 1 of the present invention.

[0053] The meaning of the marks in the figure:

[0054] 1. Three-dimensional wound iron core adjustable reactor; 2. Safety protection automatic control module; 3. Detection control module; 4. Power quality management module; 5. Operation optimization management module. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0056] The present invention adds high-voltage power-saving devices to the incoming distribution cabinets and switch cabinets of high-voltage power supply enterprises, and adds intelligent protection, voltage stabilization, grid purification, remote monitoring, and power-saving operation functions to the power distribution system. This system can extend the power supply radius of substations to suit long-distance power transmission in rural power grids, ensure the voltage compliance rate of the entire line, effectively reduce line losses and distribution transformer losses, and continue to operate without power outages when the voltage regulator is activated or deactivated. It can also automatically, manually, or remotely adjust the voltage according to the line load, ensuring that the entire power grid always operates in a safe, economical, and efficient state. Specifically, it includes the following contents.

[0057] Example 1:

[0058] See also Figure 1-6 The present invention provides a high-voltage electromagnetic balanced power-saving device, comprising: a three-dimensional wound iron core adjustable reactor 1, a safety protection automatic control module 2, a detection control module 3, a power quality management module 4, and an operation optimization management module 5;

[0059] The three-dimensional wound iron core adjustable reactor 1 includes three identical single-frame rounded rectangular three-dimensional wound iron cores and three groups of reactor coils. The actual picture can be referred to Figure 2 The three-dimensional wound core is made of amorphous alloy material with extremely low self-loss and a polygonal cross-section. The reactor coil includes three sets of main winding reactor coils connected in series to the main circuit and three sets of autocoupling coil windings for parallel induction. The output voltage is also adjusted by the tap of the autocoupling coil winding. Figure 3, A, B, and C are three-phase voltages respectively, and the first ends of the three sets of main circuit reactors are connected to them respectively. The second ends X, Y, and Z of the three sets of main circuit reactors are connected to the first ends a, b, and c of the three sets of autocoupling coil windings respectively. K1, K2, and K3 are the second ends of the three sets of autocoupling coil windings, that is, the neutral taps. The xl, yl, z1, x2, y2, z2, x3, y3, and z3 below are different voltage contacts of the voltage regulator. When the K1-xl / K2-yl / K3-z1 terminals are short-circuited, the output end line voltage drops by 3%; when the K1-x2 / K2-y2 / K3-z2 terminals are short-circuited, the output end line voltage drops by 5%; when the K1-x3 / K2-y3 / K3-z3 terminals are short-circuited, the output end line voltage drops by 7%. The 3D wound core changes the structure of the original silicon steel core, eliminating the need for Z-shaped coil winding. This reduces the number of contacts in the manufacturing process. The use of only a single neutral tap further reduces the number of contacts. Fewer contacts means fewer points of failure, significantly improving device stability and efficiency. Furthermore, compared to a laminated structure, the 3D wound core structure reduces the no-load current of a power-saving device of equivalent power by over 70%, significantly reducing its own losses and making them negligible compared to the energy-saving effect.

[0060] The safety protection automatic control module 2 utilizes a microcomputer control system, enabling real-time, accurate, and online monitoring of power supply line operation. If leakage, overload, short circuit, overvoltage, undervoltage, or phase loss occur, the intelligent safety power fire alarm protector immediately activates the corresponding protection program, issuing a command to bypass (without disconnecting) the power supply line to protect the safety and stability of the power system, and issues an audible and visual alarm. Furthermore, the safety protection automatic control module 2 is equipped with a lightning protection function to protect electrical equipment from damage caused by lightning.

[0061] The detection and control module collects and records all electrical data for each circuit, including U, I, P, COS∮, f, KWH, and KVARH. It also collects circuit breaker switching values, abnormal alarm signals, and non-electrical signals, as well as protection action information, event sequence records, and protection device operating status. It also provides real-time main wiring diagrams, trend curves, reports, and sequence of events (SOE) records.

[0062] Reference Figure 4 The detection control module 3 is controlled by an intelligent control instrument, which includes a voltage signal feedback circuit, a current signal feedback circuit, an A / D module, a CPU, a display part, a communication interface, a remote control input and a remote control output; the voltage signal feedback circuit and the current signal feedback circuit are respectively connected to the A / D module, and the CPU is respectively connected to the A / D module, the display part, the communication interface, the remote control input and the remote control output.

[0063] The CPU performs signal enhancement processing based on a convolutional neural network. The input layer of the convolutional neural network extracts the characteristics of the communication network signal, as follows:

[0064] The detected communication network signal X(k) is sampled by sampling to obtain the signal instantaneous amplitude sequence R(k):

[0065]

[0066] Where A(k) is the signal amplitude sequence obtained by sampling at the sampling frequency; k represents the length of the instantaneous amplitude sequence;

[0067] Get the maximum instantaneous amplitude F of the signal max The two characteristic parameters are the standard deviation σ of the absolute value of the instantaneous amplitude of the signal, and the calculation formula is as follows:

[0068]

[0069] Where ∮(·) represents the discrete welfare home transformation process; K represents the number of signal sampling points;

[0070] The convolution operation of the convolution layer can be referred to Figure 5 , as follows:

[0071]

[0072] Among them, y(i,j,l) is the output; x(i,j,l) is the input; d is the step size; n, m and s are the length ranges of different dimensions respectively; (i,j,l) corresponds to the multidimensional coordinate index; (h,w,r) is used to traverse the different dimensions of the filter; is the weight parameter of the lth filter; b l is the bias of the lth filter; ε is the expansion factor; f(·) is the activation function;

[0073] Use convolution kernel sizes of 1×1, 1×3, ... 1×2 N The +1 multi-scale dilated convolution operator connects feature signals of different scales and uses a zero-filled convolution operator to keep the output features and input features of the same size;

[0074] Convolutional neural networks also incorporate attention mechanisms, which can be found in Figure 6 , as follows:

[0075] y out (g) = T Y [y(g)×im X (g)×im Y (g)×im Z (g)+y(g)]

[0076] Among them, y out (g) is the output of the attention layer; y(g) is the output of the convolution layer; T Y Indicates the pooling operation along the Y direction; im X (g), im Y (g) and im Z (g) Indicates the importance of different directions.

[0077] The use of multi-scale convolutional layers enables the network to capture information at different scales, adapting to objects and features of varying sizes. Using convolution kernels with different dilation factors, it can cover a wide range of context, from local details to global structure, thereby better understanding the different layers of features in the input data. By simultaneously considering both past and future information, it further expands the receptive field and more comprehensively utilizes contextual information for analysis and prediction. Multi-scale non-causal dilated convolutions can effectively extract rich features without significantly increasing the computational load. Compared to traditional convolutions, they achieve more powerful representational capabilities with fewer parameters, improving computational efficiency. The convolutional layers of this embodiment demonstrate greater adaptability when processing data with complex structure and variation, such as natural images, audio signals, and video sequences. They can capture both regional details and global contextual information, better handling the diversity and uncertainty in the data. The convolutional layers of this embodiment can be easily combined with other deep learning architectures and techniques, offering high flexibility. They can be combined with attention mechanisms, residual connections, recurrent neural networks, and other techniques to build even more powerful models, depending on the needs of specific tasks.

[0078] Incorporating attention mechanisms into different dimensions can better focus on useful transmission signals, thereby better capturing important information in the transmitted data and achieving the goal of signal enhancement.

[0079] Power quality management monitors grid power quality, including voltage phase loss, voltage imbalance, harmonic measurement and overshoot, harmonic distortion, voltage overshoot, voltage sag, voltage fluctuation, and abnormal frequency fluctuation. Real-time monitoring of grid power quality can promptly detect problems such as voltage fluctuation, frequency deviation, and harmonic distortion. If not addressed promptly, these issues can lead to equipment failure, system crashes, and even widespread power outages. Real-time monitoring can issue alerts at the earliest stages of a problem, enabling operations and maintenance personnel to quickly take corrective measures and ensure stable power system operation. It can also quickly locate the fault point when a grid fault occurs and provide detailed fault information. This enables operations and maintenance personnel to respond quickly and implement effective fault isolation and repair measures, shortening outages and improving power supply reliability. It can also promptly detect abnormal conditions such as overvoltage, overcurrent, and harmonics, allowing appropriate protective measures to be taken, such as adjusting transformer taps and switching capacitor banks, to protect power equipment from damage.

[0080] Operation optimization management: Provide continuous operation trend records, operation status records, startup voltage and current records, accident records, and insulation monitoring records for large electrical equipment. Optimize system operation modes and optimize the design capacity and operation schedule of transformers, cables, switches, and motors.

[0081] The operation optimization management module 5 comprehensively collects the continuous operation trend records, operation status records, startup voltage and current records, accident records, and insulation monitoring record data of large electrical equipment, and optimizes the system operation mode based on the improved harmony optimization strategy. The improved harmony optimization strategy is as follows:

[0082] The energy saving P={p1,p2,Λ,p Q} As the optimization target, Q is the total number of operation arrangement combinations of each device, and the relevant parameters are initialized;

[0083] p uv =p min +(p max -p min )×r1

[0084] Among them, p uv Arrange vector for initializing the operation of u-th device and v-th device; p max and p min are the upper and lower limits of the solution; r1 is a random number between (0,1);

[0085] Initialize the operation arrangement combination library GM, randomly generate q operation arrangement vectors in the search domain according to the above formula, and add them to the operation arrangement combination library GM shown by the matrix:

[0086]

[0087] Among them, FL[·] is the energy saving measurement function;

[0088] Update the operation schedule vector p before the e-th iteration uv (e) Update as follows:

[0089]

[0090] Among them, r2, r3, r4 and r5 are all random numbers between (0,1); p new,uv (e) is the updated operation arrangement vector after the e-th iteration; α is the iteration factor; E is the maximum number of iterations; BW is the combination limit of the operation arrangement of the e-th iteration; p1 is the storage probability of the operation arrangement combination library; p2(e) is the fine-tuning probability of the operation arrangement of the e-th iteration, p 2max and p 2min Arrange the maximum and minimum values ​​of the fine-tuning probability for the run respectively;

[0091] The updated operation schedule vector p new,uv (e) The worst harmony vector p in the combined memory GM of the initial run arrangement worst The harmony metric function value is compared, if p new,uv The metric function value of (e) is less than p worst The metric function value is p new,uv (e) Replace p worst ; Otherwise, the run schedule vector in GM remains unchanged;

[0092] Check whether the algorithm termination condition has been met; if so, output the global optimal solution and the algorithm ends; otherwise, return to continue updating the operation schedule vector. By improving the harmony optimization strategy to optimize the system operation, the number of nodes can be greatly increased, thereby improving the node efficiency of the device.

[0093] The power-saving device of this embodiment solves the problem of long-distance power transmission by connecting a voltage regulator in series with the line, which can extend the power transmission distance exponentially and provide better voltage quality. The installation position and setting parameters of the voltage regulator can be determined according to the line parameters, and the voltage of the entire line can be simulated by the line automatic voltage regulator itself through measurement and calculation. After adjustment, it is ensured that the voltage of the entire line meets the national standard. By increasing the line operating voltage, the operating current can be reduced to reduce the power loss of the line when the same power is output. Controlling the line operating voltage near the rated value can reduce the distribution transformer loss. This device is also equipped with a specially designed combined bypass knife switch, which is ingenious in design and flexible and easy to operate. The voltage regulator can be put into operation without power outage during maintenance. Similarly, there is no need to shut down the power when the voltage regulator is put into operation, which improves the reliability of power supply.

[0094] The unique patented technology can automatically calculate and adjust the remote voltage according to the line load size, ensuring that the line can accurately regulate the voltage in any operating state, such as three-phase three-wire, two-wire and one-ground.

[0095] Example 2:

[0096] A high-voltage electromagnetic balancing power-saving device constructed in Example 1 was applied to an actual factory. The output voltage was adjusted to 10,000 volts, and the measured energy consumption data of factories with different voltages was monitored. For details, please refer to Table 1:

[0097] Table 1 Adjustment of the measured energy consumption data of factories with different voltages on the 10,000 V output side

[0098]

[0099] Then adjust the voltage to 6000V and monitor the actual energy consumption data under different working conditions. Please refer to Table 2 for details:

[0100] Table 2 Measured energy consumption data at different 6000V voltages

[0101]

[0102]

[0103] In summary, the power-saving device of Example 1 can achieve power savings ranging from 5% to 20%, depending on the enterprise's power load. When used by power supply departments, it can ensure a 100% voltage compliance rate and a plant power utilization rate of no less than 7.95% under various operating conditions, demonstrating its high application value. Furthermore, by extending the power line radius and reducing line losses, the entire power grid remains safe, economical, and efficient.

[0104] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A high-voltage electromagnetic balancing power-saving device, characterized in that: include: A three-dimensional wound iron core adjustable reactor (1), a safety protection automatic control module (2), a detection control module (3), a power quality management module (4), and an operation optimization management module (5); The three-dimensional wound iron core adjustable reactor (1) comprises three identical single-frame rounded rectangular three-dimensional wound iron cores and three groups of reactor coils, wherein the reactor coils comprise three groups of main winding reactor coils connected in series in a main circuit and three groups of autocoupling coil windings for parallel induction. The three-dimensional wound iron core adjustable reactor (1) further adjusts the level of the output voltage through the tap of the autocoupling coil winding. The safety protection automatic control module (2) is controlled by a microcomputer, monitors the operation of the power supply line in real time, starts the corresponding protection program through the intelligent safety electricity fire alarm protector, and issues a command to control the combined bypass knife switch to switch to the bypass power supply line, and simultaneously issues an audible and visual alarm; The detection control module (3) is used to collect full electrical quantity data, non-electrical quantity signals and work log information, and summarize and integrate them; The power quality management module (4) is used to monitor the power supply quality of the power grid; The operation optimization management module (5) is used to provide continuous operation trend records, operation status records, startup voltage and current records, accident records, and insulation monitoring records for large electrical equipment, and optimize the system operation mode by summarizing the recorded information, and optimize the design capacity and operation arrangement of transformers, cables, switches, and motors; The operation optimization management module (5) comprehensively collects continuous operation trend records, operation status records, startup voltage and current records, accident records, and insulation monitoring record data of large-scale electrical equipment, and optimizes the system operation mode based on the improved harmony optimization strategy. The improved harmony optimization strategy is specifically as follows: The energy saving P={p1,p2,Λ,p Q } As the optimization target, Q is the total number of operation arrangement combinations of each device, and the relevant parameters are initialized; p uv =p min +(p max -p min )×r1 Among them, p uv Arrange vector for initializing the operation of u-th device and v-th device; p max and p min are the upper and lower limits of the solution; r1 is a random number between (0,1); Initialize the operation arrangement combination library GM, randomly generate q operation arrangement vectors in the search domain according to the above formula, and add them to the operation arrangement combination library GM shown by the matrix: Among them, FL[·] is the energy saving measurement function; Update the operation schedule vector p before the e-th iteration uv (e) Update as follows: Among them, r2, r3, r4 and r5 are all random numbers between (0,1); p new,uv (e) is the updated operation arrangement vector after the e-th iteration; α is the iteration factor; E is the maximum number of iterations; BW is the combination limit of the operation arrangement of the e-th iteration; p1 is the storage probability of the operation arrangement combination library; p2(e) is the fine-tuning probability of the operation arrangement of the e-th iteration, p 2max and p 2min Arrange the maximum and minimum values ​​of the fine-tuning probability for the run respectively; The updated operation schedule vector p new,uv (e) The worst harmony vector p in the combined memory GM of the initial run arrangement worst The harmony metric function value is compared, if p new,uv The metric function value of (e) is less than p worst The metric function value is p new,uv (e) Replace p worst ; Otherwise, the run schedule vector in GM remains unchanged; Check whether the algorithm termination condition is met; if so, output the global optimal solution and the algorithm ends; otherwise, return to continue updating the operation schedule vector; The first ends of the three groups of main circuit reactors are respectively connected to the three-phase input ends of the high-voltage line, and the second ends of the three groups of main circuit reactors are respectively connected to the three-phase output ends of the high-voltage line; the first ends of the three groups of autocoupling coil windings are respectively connected to the second ends of the three groups of main circuit reactors; the three groups of autocoupling coil windings each have a neutral tap, and the neutral tap is connected to different gear positions of the voltage regulator to adjust the output voltage.

2. A high-voltage electromagnetic balancing power-saving device according to claim 1, characterized in that: The three-dimensional wound core is made of amorphous material, and its cross section is a polygonal structure.

3. A high-voltage electromagnetic balance power-saving device according to claim 1, characterized in that: The detection control module (3) is controlled by an intelligent control instrument, which includes a voltage signal feedback circuit, a current signal feedback circuit, an A / D module, a CPU, a display part, a communication interface, a remote control input, and a remote control output; the voltage signal feedback circuit and the current signal feedback circuit are respectively connected to the A / D module, and the CPU is respectively connected to the A / D module, the display part, the communication interface, the remote control input, and the remote control output.

4. A high-voltage electromagnetic balancing power-saving device according to claim 3, characterized in that: The CPU performs signal enhancement processing based on a convolutional neural network, and the input layer of the convolutional neural network extracts communication network signal features, specifically as follows: The detected communication network signal X(k) is sampled by sampling to obtain the signal instantaneous amplitude sequence R(k): Where A(k) is the signal amplitude sequence obtained by sampling at the sampling frequency; k represents the length of the instantaneous amplitude sequence; Get the maximum instantaneous amplitude F of the signal max The two characteristic parameters are the standard deviation σ of the absolute value of the instantaneous amplitude of the signal, and the calculation formula is as follows: Where ∮(·) represents the discrete Fourier transform process; K represents the number of signal sampling points; The convolution operation of the convolutional layer is as follows: Among them, y(i,j,l) is the output; x(i,j,l) is the input; d is the step size; n, m and s are the length ranges of different dimensions respectively; (i,j,l) corresponds to the multidimensional coordinate index; (h,w,r) is used to traverse the different dimensions of the filter; is the weight parameter of the lth filter; b l is the bias of the lth filter; ε is the expansion factor; f(·) is the activation function; Use convolution kernel sizes of 1×1, 1×3, ... 1×2 N The +1 multi-scale dilated convolution operator connects feature signals of different scales and uses a zero-filled convolution operator to keep the output features and input features of the same size; The convolutional neural network also incorporates an attention mechanism, as follows: y out (g)=T Y [y(g)×im X (g)×im Y (g)×im Z (g)+y(g)] Among them, y out (g) is the output of the attention layer; y(g) is the output of the convolution layer; T Y Indicates the pooling operation along the Y direction; im X (g), im Y (g) and im Z (g) Indicates the importance of different directions.

5. A high-voltage electromagnetic balancing power-saving device according to claim 4, characterized in that: The safety protection automatic control module (2) also takes the power supply line operation status data as input and performs fault prediction and voltage regulation based on the convolutional neural network.

6. A high-voltage electromagnetic balancing power-saving device according to claim 5, characterized in that: The safety protection automatic control module (2) is also provided with a lightning protection function.

7. A high-voltage electromagnetic balance power-saving device according to claim 1, characterized in that: The power quality data monitored by the power quality management module (4) include grid voltage phase loss, voltage imbalance, harmonic measurement and line crossing, harmonic distortion rate, voltage overshoot and undershoot, voltage fluctuation, and abnormal frequency fluctuation data.

Images