Low-noise low-loss reactor and processing method thereof

Abstract

The low-noise and low-loss electric reactor comprises an insulating base, a plurality of connecting supporting legs are arranged at the corners of the lower end of the insulating base, the lower ends of the connecting supporting legs are connected with a base plate through bolts to complete installation and fixation of a foundation, a plurality of limiting grooves are formed in the upper end of the insulating base at equal intervals, and the limiting grooves are connected with the base plate through bolts. A through hole matched with the copper bar is formed in the center of each limiting groove, a plurality of iron core cakes are arranged in each limiting groove, the bottom of the iron core cake on the lowermost side is clamped and matched with the corresponding limiting groove, and the adjacent iron core cakes are connected and fixed through a connecting frame. And the heat dissipation effect is further improved, so that the overall performance of the reactor is improved. The defect of one-sided monitoring of an existing method is overcome, and various hidden dangers such as overload, voltage fluctuation, condensation and dust blockage can be accurately captured; the problems of false alarm and missing alarm caused by a fixed threshold value are avoided, and the method can be flexibly adapted to various application scenes such as industrial power distribution, wind power and photovoltaic.

Description

Technical Field

[0001] This invention relates to the field of electrical component processing technology, specifically to low-noise, low-loss reactors and their processing methods. Background Technology

[0002] In terms of manufacturing process, traditional self-saturating reactors, after the core plates are wound, do not undergo vacuum pressure impregnation treatment, resulting in gaps between the plates. This makes the silicon steel sheets prone to electromagnetic vibration during reactor operation, with large amplitude vibrations, leading to a significant increase in equipment operating noise and affecting the surrounding environment and normal equipment operation. Existing patent CN120413256B discloses a reactor in which multiple damping pads are used to divide the core body into multiple segments, reducing noise. However, this connection method results in high processing costs, structural instability, and a prolonged feeding process. Furthermore, it is detrimental to the heat dissipation of the reactor body. Therefore, a low-noise, low-loss reactor and its processing method are now provided. Summary of the Invention

[0003] The purpose of this invention is to provide a low-noise, low-loss reactor and its processing method to solve the problem of unreasonable structure of existing reactors.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: The low-noise, low-loss reactor includes an insulating base. Multiple connecting legs are located at the lower corners of the insulating base. The lower ends of the connecting legs are connected to a base plate via bolts to complete the foundation installation and fixation. Multiple equidistant limiting grooves are provided on the upper end of the insulating base. A through hole matching a copper busbar is located at the center of each limiting groove. Multiple iron core discs are provided in each limiting groove. The bottom of the lowest iron core disc engages with the limiting groove. Adjacent iron core discs are connected and fixed together via a connecting frame.

[0005] A copper busbar is provided at the center of the iron core disc, and a positioning mechanism for locking the iron core disc is provided at the upper end of the insulating base. Further embodiment: The copper busbar is a tube structure with a central guide hole, the upper end of the copper busbar is provided with a first terminal, and the lower end of the copper busbar is provided with a second terminal.

[0006] Further embodiment: The positioning mechanism includes a positioning beam that presses down on the upper end of the iron core disc, with positioning vertical plates symmetrically arranged on both sides of the positioning beam, and a connecting seat at the lower end of the positioning vertical plate, the connecting seat being connected and fixed to the insulating base by bolts.

[0007] Further solution: The shape of the connecting frame matches the cross-section of the iron core disc. Multiple equally spaced isolation blocks are arranged at the center of the inner wall of the connecting frame. The isolation blocks are used to divide the upper and lower adjacent iron core discs, leaving gaps between them. Heat dissipation side holes are arranged on the surface of the connecting frame between adjacent isolation blocks. The setting of heat dissipation side holes allows the gaps between the iron core discs to communicate with the outside.

[0008] A further solution: The positioning beam has a central opening at its center that matches the copper busbar.

[0009] A method for manufacturing a low-noise, low-loss reactor includes the following steps: Step 1: First, an insulating base is manufactured by injection molding; Step 2: Use 0.27mm thick non-marked cold-rolled silicon steel sheets to wind an elliptical iron core disc 200, and then anneal and apply glue to the elliptical iron core disc; Step 3: Prepare the connecting frame. The connecting frame is made of nitrile rubber and is wrapped with rigid plastic on the outside for limiting the position. The support position is a buffer area. Step 4: Stack the prepared connecting frame and iron core disc alternately, and use the connecting frame to limit and fix the iron core disc to obtain a single-arm iron core; Step 5: Insert the bottom of the obtained single-arm iron core into the limiting groove at the upper end of the insulating base to complete the positioning; Step 6: Press down and fix the upper ends of multiple single-arm iron cores using the positioning crossbeam, and connect them to the insulating base with bolts using the connecting seat; Step 7: Pass the winding through the insulating base and the single-arm iron core, then position it to complete the fabrication.

[0010] A reactor monitoring method includes the following steps: Step 1: Multi-dimensional data acquisition, synchronously collecting electrical condition data of the reactor body, operating environment scenario data, structural vibration data, and temperature data of core components; Step 2: Data preprocessing, which involves filtering and noise reduction, format standardization, and outlier removal of the collected multi-dimensional raw data; Step 3: Multi-source data fusion analysis, using dynamic threshold calibration algorithm and multi-parameter linkage judgment algorithm to perform collaborative analysis on preprocessed data and identify the operating status category; Step 4: Intelligent diagnosis and prediction based on historical data. Fault characteristics are matched by fault mode recognition algorithm, and the remaining life of core components is output by multi-factor life prediction algorithm. Step 5: Early Warning and Linkage Control. Based on the status identification results and diagnostic prediction conclusions, early warning information is output in stages, and the actuators are linked to make adaptive adjustments. At the same time, the monitoring accuracy is optimized through data storage and model iteration mechanisms. Using "data acquisition-purification-fusion analysis-diagnosis prediction-closed-loop optimization" as the core link, multi-dimensional data acquisition is first used to supplement monitoring elements. Then, preprocessing ensures data quality. Next, algorithm fusion achieves accurate judgment of status and faults. Finally, early warning linkage and model iteration form a full lifecycle monitoring closed loop, ensuring that monitoring extends from "data perception" to "intelligent decision-making." This addresses the problems of existing monitoring methods such as single-dimensionality, poor scenario adaptability, and fragmented diagnosis, achieving comprehensive coverage and in-depth analysis of reactor operating status, significantly improving the completeness, accuracy, and continuity of monitoring, and reducing the risk of sudden faults.

[0011] A further proposed solution involves the following specific logic for multi-dimensional data acquisition in step 1: S11 Electrical operating condition data acquisition: Using non-contact or minimally invasive sensors, the main circuit current, control winding current, and voltage data at both ends are acquired. The sensor installation does not damage the original insulation and current-carrying structure of the reactor. S12 Operating environment scenario data acquisition: Environmentally adaptable sensors are deployed to acquire ambient temperature and relative humidity. In outdoor scenarios, altitude and air pressure data can be supplemented, and in industrial dust concentration scenarios, dust concentration data can be supplemented. S13 Structural vibration data acquisition: Low-power vibration sensors are deployed at locations where vibration transmission is concentrated to acquire vibration amplitude and frequency data. S14 Core component temperature data acquisition: Temperature sensors are embedded in core heat-generating components such as vertical copper busbars, single-arm iron cores, and control windings to acquire temperature data for each component. All analog signals acquired by the sensors are converted into digital signals by low-power chips and then stored in the processing module. Targeting the four key dimensions of reactors – electrical operation, environmental adaptation, structural status, and heat generation – sensors are selected as needed (non-contact / minimally invasive sensors ensure structural safety, while environmentally adaptable sensors are suitable for multiple scenarios). Through precise placement, signal digitization, and local storage, the security, relevance, and completeness of data acquisition are ensured, avoiding data loss or structural damage. This approach overcomes the limitations of existing technologies that only monitor temperature and vibration, comprehensively capturing various potential hazard signals such as electrical overload, voltage fluctuations, environmental condensation, and structural loosening. Furthermore, the sensor installation is compatible with the existing reactor structure, requiring no major modifications and lowering the application threshold.

[0012] According to the monitoring method of claim 1, the specific logic of the data preprocessing in step 2 and the dynamic threshold calibration algorithm in step 3 is as follows: S21 Filtering and Noise Reduction: Median filtering is used to remove pulse interference from vibration data, and moving average filtering is used to smooth data fluctuations for temperature, current, and voltage data; S22 Format Standardization: Heterogeneous data from different sensors are converted into a standardized data format, including data type, acquisition timestamp, sensor ID, and core fields of the measured value; S23 Outlier Removal: Outliers are identified using statistical criteria, and valid data is supplemented through interpolation to ensure data continuity; (4) Dynamic threshold calibration algorithm: S311 Scene classification storage: A three-dimensional classification system is constructed according to the application scene type, season and operating load level, and the historical normal operation data under the corresponding scene is stored; S312 Normal interval fitting: For the historical data under each scene, a polynomial fitting algorithm is used to fit the normal operation interval of each monitoring parameter. The fitting formula is: Where x is the runtime time series and y is the normal value of the parameter. , , , The fitting coefficients are solved using the least squares method; S313 Dynamic threshold generation: Based on the fitted normal operating range, a dynamic threshold range including safety redundancy is set; S314 Real-time threshold update: The latest historical data is periodically called to refit the normal range and dynamically adjust the threshold range. First, multi-source heterogeneous data is purified through "targeted filtering + standardization + outlier repair" to eliminate distortion caused by electromagnetic interference and format differences; then, through "3D scenario classification - normal range fitting - dynamic threshold generation - periodic updates", the judgment criteria are adapted to different operating conditions, replacing rigid fixed thresholds; This solves the problem of false alarms / missed alarms caused by poor data quality and thresholds that are not suitable for the scenario. The consistency and continuity of the preprocessed data are improved, and the dynamic threshold can accurately match multiple scenarios such as industrial power distribution and wind power, as well as different loads and seasonal operating conditions, significantly reducing the probability of invalid warnings.

[0013] The multi-dimensional intelligent monitoring method for self-saturating reactors according to claim 1 is characterized in that the specific logic of the multi-parameter linkage judgment algorithm in step 3 and the fault mode recognition algorithm in step 4 is as follows: (1) Multi-parameter linkage judgment algorithm: S321 constructs a hierarchical judgment rule base, which includes three types of judgment logic: normal working condition, early warning working condition, and fault working condition. All types of logic are based on the association combination conditions of multiple parameters such as "electrical working condition-environmental scene-vibration-temperature"; S322 real-time matching judgment: the preprocessed real-time data is matched with the rule base one by one, and the corresponding operating status category is output. The matching priority is fault working condition > early warning working condition > normal working condition. (2) Fault Mode Recognition Algorithm: S411 Constructing a Fault Feature Mapping Library: Establishing a mapping relationship according to the structure of "Scenario Type - Fault Type - Multi-parameter Feature Vector". The feature vector contains the abnormal features and associated co-change features of each monitoring parameter; S412 Feature Extraction: Extracting feature vectors such as parameter abnormality degree, change rate and multi-parameter correlation coefficient from the early warning or fault condition data; S413 Similarity Matching: Using the cosine similarity algorithm to calculate the similarity between the real-time feature vector and the feature vectors of each fault type in the mapping library. The similarity calculation formula is: Where A is the real-time feature vector, B is the fault feature vector in the database, and n is the feature dimension. , These are the i-th components of two vectors, respectively. S414 Fault Determination: Fault type, suspected fault, or novel fault is determined based on similarity results. Multi-parameter linkage judgment utilizes a "hierarchical rule base + priority matching" approach, leveraging the correlation of multi-dimensional data to locate the operating condition type, avoiding the limitations of single-parameter judgment. Fault pattern recognition achieves accurate matching of fault features with historical cases through "feature library construction - feature extraction - similarity calculation - hierarchical judgment," covering known faults, suspected faults, and novel faults. It addresses the problem of existing technologies being unable to accurately locate the root cause of faults. Multi-parameter linkage can distinguish between operating conditions with similar symptoms but different root causes, such as "overload heating" and "dust blockage in heat dissipation." Cosine similarity matching significantly improves fault diagnosis accuracy and can capture suspected and novel faults, providing clear guidance for operation and maintenance.

[0014] The multi-dimensional intelligent monitoring method for self-saturating reactors according to claim 1 is characterized in that the specific logic of the multi-factor lifetime prediction algorithm in step 4 and the early warning and linkage control and model iteration in step 5 is as follows: (1) Multi-factor life prediction algorithm: S421 Determine influencing factors: Select operating temperature (T), ambient humidity (H), voltage fluctuation amplitude (U), and vibration accumulation (V) as the influencing factors of the life of core components, among which vibration accumulation (t is the running time, and is the real-time vibration amplitude); S422 establishes the basic life model: based on the Arrhenius equation, the multi-factor life model is extended, and the formula is: Where L is the predicted remaining lifetime. The rated life under standard operating conditions. , , , These are standard operating parameters. , , , The influence coefficient is defined as follows: S423 Model Calibration: Historical operating data under the corresponding scenario is called, and the gradient descent algorithm is used to iteratively optimize the influence coefficient; S424 Remaining Life Calculation: Real-time monitoring data and historical cumulative data are substituted into the calibrated model to calculate the remaining life of the core components; (2) Early warning and linkage control: output early warning information according to the category of operating status, and activate actuators such as heat dissipation, dehumidification and load adjustment according to the type of operating condition. The corresponding protection mechanism is triggered when the fault condition occurs. (3) Data storage and model iteration: The processing module stores recent high-frequency data, while the cloud database stores historical summary data, feature mapping library, and model parameters. The parameters and rule library of various algorithms are updated regularly based on new data to continuously optimize the monitoring adaptability and accuracy. Lifespan prediction improves the accuracy of prediction by comprehensively considering the key influencing factors of core component aging through "multi-factor selection - Arrhenius equation extension - historical data calibration - real-time calculation". Early warning linkage achieves differentiated operation and maintenance for different risk levels through "tiered push + targeted execution". Model iteration allows the monitoring method to be continuously optimized as the operating data accumulates through "layered data storage - regular parameter updates". It solves the problems of existing lifespan prediction relying on a single factor, large error, and passive operation and maintenance. The error of multi-factor lifespan prediction is significantly reduced, tiered early warning clarifies the operation and maintenance priority, linkage control can quickly respond to hidden dangers, and model iteration continuously improves the adaptability and accuracy of the monitoring method, realizing the upgrade from "passive maintenance" to "preventive maintenance" and reducing operation and maintenance costs.

[0015] The present invention has the following beneficial effects: This invention optimizes the reactor structure, creating heat dissipation gaps between the single-arm iron cores in the reactor. This snap-fit ​​method simplifies installation. In addition, heat dissipation channels are set in the copper busbars, further improving the heat dissipation effect and thus enhancing the overall performance of the reactor.

[0016] This invention overcomes the shortcomings of existing methods in monitoring, and can accurately capture various hidden dangers such as overload, voltage fluctuation, condensation, and dust blockage. Relying on a three-dimensional scene classification system and a dynamic threshold calibration algorithm, the threshold can be dynamically adapted to the scene, season and load level, avoiding the false alarm and missed alarm problems caused by fixed thresholds. It can flexibly adapt to various application scenarios such as industrial power distribution, wind power and photovoltaic. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of one side of the structure of the present invention.

[0018] Figure 2 This is a schematic diagram of the lower structure of the present invention.

[0019] Figure 3 This is a schematic diagram of the connecting frame in this invention.

[0020] Figure 4 This is a schematic diagram of the copper busbar structure in this invention.

[0021] Figure 5 This is a logic block diagram of the monitoring method of the present invention.

[0022] In the diagram: Insulating base 100, connecting leg 101, limiting groove 102, through hole 103; Iron core disc 200, connecting frame 201, isolation block 202, heat dissipation side hole 203; Positioning vertical plate 300, positioning horizontal beam 301, connecting seat 302; Copper busbar 400, flow guide hole 401, first terminal 402, second terminal 403; Data cable 501, processing module 500. Detailed Implementation

[0023] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0024] Example 1

[0025] refer to Figures 1-4 As shown, the low-noise, low-loss reactor includes an insulating base 100. Multiple connecting legs 101 are located at the lower corners of the insulating base 100. The lower ends of the connecting legs 101 are connected to a base plate via bolts to complete the foundation installation and fixation. Multiple limiting grooves 102 (three in this case) are equidistantly located on the upper end of the insulating base 100. A through hole 103 matching a copper busbar 400 is located at the center of each limiting groove 102, allowing the copper busbar 400 to pass through the insulating base 100. Multiple core discs 200 are located in each limiting groove 102. The bottom of the lowest core disc 200 engages with the limiting groove 102 to fix the position of the core disc 200. Adjacent core discs 200 are connected and fixed by a connecting frame 201. The connecting frame 201 can connect multiple iron core discs 200. The connecting frame 201 itself can absorb the energy generated by the vibration of the iron core discs 200, thereby reducing noise pollution. In addition, the breathable structure retained on the connecting frame 201 allows heat to escape better, which helps to improve the heat dissipation effect of the product. The shape of the connecting frame 201 matches the cross-section of the iron core discs 200. Multiple equally spaced isolation blocks 202 are arranged at the center of the inner wall of the connecting frame 201. The isolation blocks 202 are used to divide the upper and lower adjacent iron core discs 200, leaving gaps between them. The surface of the connecting frame 201 between adjacent isolation blocks 202 is provided with heat dissipation side holes 203. The setting of heat dissipation side holes 203 allows the gaps between the iron core discs 200 to communicate with the outside, so as to allow heat to escape. A copper busbar 400 is fitted at the center of the iron core disc 200 through hole. The copper busbar 400 is a tubular structure with a central guide hole 401. The upper end of the copper busbar 400 has a first terminal 402, and the lower end has a second terminal 403 for wiring the reactor. The tubular structure of the copper busbar 400 allows internal heat to flow out with the airflow, improving heat dissipation. This means the copper busbar 400 not only functions as a conductor but also acts as a heat dissipation and heat conduction device. The upper end of the insulating base 100... A positioning mechanism is provided to lock the iron core disc 200. The positioning mechanism includes a positioning beam 301 that presses down on the upper end of the iron core disc 200. Positioning vertical plates 300 are symmetrically arranged on both sides of the positioning beam 301. A connecting seat 302 is provided at the lower end of the positioning vertical plate 300. The connecting seat 302 is connected and fixed to the insulating base 100 by bolts. The positioning beam 301 exerts downward pressure on the iron core disc 200, which, together with the limiting groove 102, limits the bottom of the iron core disc 200, thereby finally locking the position of the iron core disc 200. The positioning beam 301 has a central opening that matches the copper busbar 400. The central opening minimizes the contact area between the positioning beam 301 and the iron core disc 200, which helps with heat dissipation and saves materials. Specific processing steps: Step 1: First, the insulating base 100 is manufactured by injection molding; Step 2: Use 0.27mm thick non-marked cold-rolled silicon steel sheets to wind an elliptical iron core disc 200, and then anneal and apply glue to the elliptical iron core disc 200. Step 3: Prepare the connecting frame 201. The connecting frame 201 is made of nitrile rubber. The outside of the connecting frame 201 is wrapped with rigid plastic for limiting the position. The support position is a buffer area. Step 4: The prepared connecting frame 201 and iron core disc 200 are stacked alternately, and the connecting frame 201 is used to limit and fix the iron core disc 200 to obtain a single-arm iron core. Step 5: Insert the bottom of the obtained single-arm iron core into the limiting groove 102 at the upper end of the insulating base 100 to complete the positioning; Step 6: Press down and fix the upper ends of multiple single-arm iron cores by positioning beam 301, and connect them to insulating base 100 by connecting seat 302 with bolts; Step 7: Pass the winding through the insulating base 100 and the single-arm iron core, then position it to complete the fabrication; Example

[0026] refer to Figure 5As shown, in order to better monitor the working status of the reactor, the positioning beam 301 is equipped with a sensing module for the working information of the iron core disc 200 and the copper busbar 400. The sensor module is connected to the processing module 500 through the data line 501. Multi-dimensional data acquisition: Electrical condition of the main body: A snap-on Rogowski coil current sensor is installed on the outside of the vertical copper busbar of the main circuit, and a miniature voltage divider voltage sensor (insulation class H) is embedded at both ends of the control winding to collect the main circuit current, control winding current and voltage data; Environmental scenario: An IP65-rated temperature and humidity sensor is installed on the outside of the insulated base to collect ambient temperature and relative humidity; Vibration data: A snap-on low-power vibration sensor is installed on the top of the insulating pressure plate to collect vibration amplitude and frequency; Temperature data: Miniature PT100 temperature sensors are embedded in the vertical copper busbar near the winding, the middle of the single-arm iron core, and the output end of the control winding to collect the temperature of each part; All sensor signals are converted into digital signals by a low-power chip and then stored in the processing module.

[0027] Filtering: Vibration data are filtered using a median filter with a window size of 5 to remove pulse interference, while temperature, current, and voltage data are filtered using a moving average filter with 3 sampling periods to smooth fluctuations. Standardization processing: Convert all data to JSON format, including the fields "data type-collection timestamp-sensor ID-measurement value", with the timestamp accurate to milliseconds; Outlier removal: Outliers are identified using the 3σ criterion, and valid data are supplemented using linear interpolation.

[0028] Dynamic threshold calibration: ①Scenario Classification: Store historical normal data according to the category of "Industrial Power Distribution - Autumn - Medium Load"; ② Fitting within normal intervals: A third-order polynomial is used to fit the normal intervals of each parameter, and the fitting coefficients are solved using the least squares method. , , , ; ③ Threshold generation: Set the temperature parameter safety redundancy to 3℃, the current to 5% of the rated value, the voltage to 3% of the rated value, and the vibration amplitude to 0.1mm / s to form a dynamic threshold range; ④ Threshold update: Refit the data from the past 7 days every 24 hours and adjust the threshold accordingly; Multi-parameter linkage judgment: If the real-time data meets the conditions of "current > 20% of rated value + copper busbar temperature > dynamic threshold upper limit - 2℃ + ambient temperature < 35℃", it is determined to be a warning condition (high load active heating). If the following conditions are met: "voltage fluctuation > ±10% + sudden increase in vibration amplitude > 30% + winding temperature rise > 5℃ / h", it is determined to be a fault condition (voltage surge causes core loosening).

[0029] Fault mode identification: Extract the feature vector of "voltage fluctuation 12% + vibration amplitude sudden increase 35% + winding temperature rise 6℃ / h", calculate the cosine similarity with the feature vector of "voltage impact causes core loosening" in the fault feature mapping library, and the similarity is 88%, which is determined to be this fault type; Lifespan prediction: The connection frame is selected as the monitoring object, and the rated lifespan under standard operating conditions is [number] years. Standard parameter ℃ , , , Influence coefficient , , , ; Using nearly one year of historical data from this industrial power distribution scenario, the influence coefficient was calibrated using the gradient descent algorithm; real-time monitoring data (°C) was then input. , , , ), calculate the remaining lifespan Year.

[0030] Early warning output: Issue a Level 1 early warning and push fault information and handling suggestions such as "stop the machine to check the core fixing status" through background pop-ups, SMS and APP. Interlocking control: Triggers the control loop protection mechanism and issues a shutdown suggestion; Data storage and iteration: High-frequency data from the past 3 months is stored locally, while historical summary data and fault cases are stored in the cloud. At the end of each quarter, algorithm parameters and rule bases are updated based on newly added data.

[0031] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. 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.

Claims

1. A low-noise, low-loss reactor, characterized in that, The system includes an insulating base (100), with multiple connecting legs (101) at the lower corner of the insulating base (100). The lower ends of the connecting legs (101) are connected to the base plate by bolts to complete the installation and fixation of the base. Multiple limiting grooves (102) are provided at equal intervals on the upper end of the insulating base (100). A through hole (103) matching the copper busbar (400) is provided at the center of the limiting groove (102). Multiple iron core discs (200) are provided in each limiting groove (102). The bottom of the lowest iron core disc (200) is engaged with the limiting groove (102). Adjacent iron core discs (200) are connected and fixed by a connecting frame (201). A copper busbar (400) is provided at the center of the iron core disc (200), and a positioning mechanism for locking the iron core disc (200) is provided at the upper end of the insulating base (100).

2. The low-noise, low-loss reactor according to claim 1, characterized in that, The copper busbar (400) is a tube structure with a central guide hole (401). The upper end of the copper busbar (400) is provided with a first terminal (402), and the lower end of the copper busbar (400) is provided with a second terminal (403).

3. The low-noise, low-loss reactor according to claim 1, characterized in that, The positioning mechanism includes a positioning beam (301) that presses down on the upper end of the iron core disc (200). Positioning vertical plates (300) are symmetrically arranged on both sides of the positioning beam (301). A connecting seat (302) is provided at the lower end of the positioning vertical plate (300). The connecting seat (302) is connected and fixed to the insulating base (100) by bolts.

4. The low-noise, low-loss reactor according to claim 1, characterized in that, The shape of the connecting frame (201) matches the cross-section of the iron core disc (200). Multiple equally spaced isolation blocks (202) are arranged at the center of the inner wall of the connecting frame (201). The isolation blocks (202) are used to divide the upper and lower adjacent iron core discs (200) and retain a gap between them. The surface of the connecting frame (201) between adjacent isolation blocks (202) is provided with heat dissipation side holes (203). The setting of the heat dissipation side holes (203) allows the gap between the iron core discs (200) to communicate with the outside.

5. The low-noise, low-loss reactor according to claim 4, characterized in that, The positioning beam (301) has a central opening at its center that matches the copper busbar (400).

6. A method for manufacturing a low-noise, low-loss reactor according to any one of claims 1-5, characterized in that, The steps include: Step 1: First, the insulating base (100) is produced by injection molding. Step 2: Use 0.27mm thick non-marked cold-rolled silicon steel sheets to wind an elliptical iron core cake (200), and then anneal and apply glue to the elliptical iron core cake (200); Step 3: Prepare the connecting frame (201). The connecting frame (201) is made of nitrile rubber. The outside of the connecting frame (201) is wrapped with hard plastic for limiting the position. The support position is a buffer area. Step 4: The prepared connecting frame (201) and iron core disc (200) are stacked alternately, and the connecting frame (201) is used to limit and fix the iron core disc (200) to obtain a single-arm iron core; Step 5: Insert the bottom of the obtained single-arm iron core into the limiting groove (102) at the upper end of the insulating base (100) to complete the positioning; Step 6: Press down and fix the upper ends of multiple single-arm iron cores by positioning crossbeam (301), and connect them to the insulating base (100) by bolts through connecting seat (302); Step 7: Pass the winding through the insulating base (100) and the single-arm iron core, and then complete the positioning to finish the fabrication.

7. A method for monitoring a reactor, characterized in that, Includes the following steps: Step 1: Multi-dimensional data acquisition, synchronously collecting electrical condition data of the reactor body, operating environment scenario data, structural vibration data, and temperature data of core components; Step 2: Data preprocessing, which involves filtering and noise reduction, format standardization, and outlier removal of the collected multi-dimensional raw data; Step 3: Multi-source data fusion analysis, using dynamic threshold calibration algorithm and multi-parameter linkage judgment algorithm to perform collaborative analysis on preprocessed data and identify the operating status category; Step 4: Intelligent diagnosis and prediction based on historical data. Fault characteristics are matched by fault mode recognition algorithm, and the remaining life of core components is output by multi-factor life prediction algorithm. Step 5: Early warning and linkage control. Based on the status identification results and diagnostic prediction conclusions, early warning information is output in stages and the actuators are linked to make adaptive adjustments. At the same time, the monitoring accuracy is optimized through data storage and model iteration mechanisms.

8. The monitoring method according to claim 7, characterized in that, Step 2, the data preprocessing, includes: S21. Filtering and noise reduction: Median filtering algorithm is used to remove pulse interference from vibration data, and moving average filtering algorithm is used to smooth data fluctuations from temperature, current and voltage data. S22. Format standardization processing: Convert heterogeneous data from different sensors into a standardized data format, including data type, acquisition timestamp, sensor ID, and core fields of measurement value; S23. Outlier removal: Outliers are identified using statistical criteria, and valid data is supplemented by interpolation to ensure data continuity.

9. The monitoring method according to claim 7, characterized in that, Step 3: Dynamic threshold calibration algorithm: S311, Scene Classification Storage: Construct a three-dimensional classification system according to application scenario type, season and operating load level, and store historical normal operation data under the corresponding scenario; S312. Normal Interval Fitting: For historical data under each scenario, a polynomial fitting algorithm is used to fit the normal operating interval of each monitoring parameter. The fitting formula is: Where x is the runtime time series and y is the normal value of the parameter. , , , The fitting coefficients are obtained using the least squares method. S313, Dynamic Threshold Generation: Based on the fitted normal operating range, set a dynamic threshold range including safety redundancy; S314. Real-time threshold update: Periodically call the latest historical data to refit the normal range and dynamically adjust the threshold range.

10. The monitoring method according to claim 7, characterized in that, Fault Mode Recognition Algorithm in Step Four: S411. Construct a fault feature mapping library: Establish a mapping relationship according to the structure of "scenario type - fault type - multi-parameter feature vector". The feature vector contains the abnormal features and related and coordinated change features of each monitoring parameter. S412, Feature Extraction: Extract feature vectors such as the degree of parameter anomaly, rate of change, and multi-parameter correlation coefficients from early warning or fault condition data; S413. Similarity Matching: The cosine similarity algorithm is used to calculate the similarity between the real-time feature vector and the feature vectors of each fault type in the mapping database. The similarity calculation formula is as follows: Where A is the real-time feature vector, B is the fault feature vector in the database, and n is the feature dimension. , These are the i-th dimension components of the two vectors, respectively; S414. Fault Determination: Determine the fault type, suspected fault, or new type of fault based on the similarity results.

Images