Remote control compensation device for bidirectional reactive compensation and harmonic control

Abstract

The application discloses a remote control compensation device capable of bidirectional reactive power compensation and harmonic treatment, and belongs to the technical field of power quality treatment. The application solves the problems that the existing device can only provide fixed reactive power compensation, cannot provide required reactive power compensation according to the change of reactive power, can only mainly provide capacitive compensation, cannot meet the requirement of compensating inductive reactive power, and cannot realize remote control. By applying the remote control device capable of bidirectional reactive power compensation and harmonic treatment to a distribution network, the power transmission capacity of the distribution network and the power supply capacity of a rural power distribution transformer can be effectively improved, so that the requirement of energy saving and consumption reduction of a low-voltage distribution network is met. Meanwhile, the voltage level of a distribution area can be effectively adjusted, the reactive power in the system can be compensated in real time, the voltage of a user at the end of the distribution area during a load peak period is ensured to be normal, the three-phase load balance of each node of the distribution area is realized, and the power supply quality of the distribution area is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of power quality management technology, specifically to a remotely controlled compensation device capable of bidirectional reactive power compensation and harmonic control. Background Technology

[0002] Industrial development has led to increased capacity in low-voltage power distribution networks and their associated lines, as well as increased current flow, resulting in greater losses in these devices and lines. The simultaneous presence of capacitive and inductive loads in the system is also becoming increasingly common. Impulsive reactive power loads can cause voltage fluctuations in the power system, leading to unsatisfactory power quality. Excessive reactive power increases the utilization of power device capacity, lowering the system's power factor and further increasing grid losses.

[0003] Reactive power compensation and power harmonic suppression are important means to improve the power consumption of users and satisfy them. They play a significant role in improving power factor, reducing circuit losses, reducing equipment capacity, and ensuring the safe and reliable operation of power supply and power consumption equipment, thereby improving energy utilization efficiency.

[0004] Although reactive power compensation devices have been installed in the current power quality management measures for the transformer substations, some key issues remain unresolved.

[0005] (1) Most existing reactive power compensation facilities can only provide fixed reactive power compensation and cannot provide the required reactive power as the reactive power in the power grid changes; (2) Traditional reactive power compensation facilities are mainly capacitive compensation, which cannot meet the requirements when inductive reactive power compensation is needed; (3) A very small number of transformer areas have installed bidirectional reactive power compensation devices such as SVG, but because they cannot be remotely controlled, they can only perform local compensation and cannot cooperate to jointly regulate the reactive power of multiple transformer areas in the area. (4) As the penetration rate of distributed photovoltaics increases in the distribution area, the voltage fluctuations and imbalances between distribution areas will intensify, requiring joint regulation of voltage and reactive power from multiple distribution areas. (5) Some automatic reactive power compensation devices can set adjustment targets, but the adjustment targets are usually power factor or voltage, and the control targets cannot be flexibly switched according to the needs of regional regulation.

[0006] Therefore, it is necessary to develop an advanced reactive power compensation device that can achieve bidirectional compensation of capacitive and inductive reactive power, and can also be remotely controlled as needed to improve the voltage qualification rate of the transformer area and reduce line loss. Summary of the Invention

[0007] The purpose of this invention is to provide a remotely controlled compensation device capable of bidirectional reactive power compensation and harmonic mitigation. By applying this bidirectional reactive power compensation and harmonic mitigation remote control device to the distribution network, the transmission capacity of the distribution network and the power supply capacity of rural distribution transformers can be effectively improved, thereby meeting the energy-saving and consumption-reducing requirements of low-voltage distribution networks. Simultaneously, it can effectively regulate the voltage level of the distribution transformer area, compensate reactive power in the system in real time, ensure normal voltage for end users during peak load periods, achieve three-phase load balance at each node of the transformer area, significantly improve the power supply quality of the transformer area, and solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A remotely controlled compensation device capable of bidirectional reactive power compensation and harmonic mitigation includes: The SVG reactive power compensation unit is configured to use a three-phase voltage-type bridge inverter, which is connected in parallel with the distribution network through a reactor or transformer; and generates a control signal PWM waveform based on the detection and analysis results to control the SVG reactive power compensation unit to emit or absorb reactive power in the grid, thereby achieving reactive power compensation. The harmonic mitigation unit is configured to work in conjunction with the SVG reactive power compensation unit. It uses a DSP+ARM processor to detect harmonic, negative sequence, and zero sequence components in the load current in real time, generates compensation commands, and controls the SVG reactive power compensation unit to output corresponding compensation current to mitigate harmonic currents in the system and improve power quality. The remote control unit is configured to perform real-time online synchronous monitoring of multiple bidirectional reactive power compensation devices based on the power Internet of Things intelligent sensing technology. It utilizes cloud computing to integrate massive amounts of data and employs intelligent analysis algorithms to perform intelligent analysis on all installation points for reference and decision-making by operation and maintenance personnel. The control strategy unit is configured to formulate comprehensive power quality management measures based on the source and load characteristics of different transformer areas; it can also flexibly switch control targets according to regional regulation needs, including but not limited to power factor and voltage, to meet the power quality management needs in different scenarios.

[0009] Furthermore, in the SVG reactive power compensation unit, based on the detection and analysis results, a control signal PWM waveform is generated to control the SVG reactive power compensation unit to emit or absorb reactive power from the power grid, thereby achieving reactive power compensation, including: By detecting the load current in real time, at node k, according to Kirchhoff's Current Law (KCL), we can derive: ; in, Expressed as power supply current; Expressed as load current; This is represented as SVG compensation current; For a balanced system, by sampling the three-phase current and voltage in real time, the following can be obtained according to equation (1): ; in, Expressed as load current; Represented as active current; Represented as reactive current; Represented as harmonic current; For an unbalanced system, the zero-sequence current and negative-sequence current in the load current are separated according to the detection algorithm of zero-sequence and negative-sequence current, and the comprehensive compensation current is calculated according to equation (2): ; in, This is represented as the comprehensive compensation current; Represented as zero-sequence current; It is represented as negative sequence current.

[0010] Furthermore, the control strategy unit includes: It is recommended that the generation module be configured to build a Long Short-Term Memory (LSTM) network based on historical power quality data as a power quality governance model for transformer substations. Key features and labels should be used as the input set, and power quality governance measures from historical power quality data should be used as the output set. The power quality governance model should learn the relationship between the source-load characteristics of different transformer substations and power quality problems from the key features, predict the load change trends and power quality problem development patterns of the substations, and formulate adaptive power quality governance measures for different types of transformer substations based on the prediction results. The model practice module is configured to input real-time monitored source-load data and power quality data of the transformer area into the trained transformer area power quality management model. Based on the real-time data, the model predicts the changing trend of source-load characteristics of the transformer area and possible power quality problems in the future; and obtains corresponding adaptive power quality management measures based on the prediction results of the real-time data.

[0011] Furthermore, in the model practice module, based on the prediction results of real-time data, corresponding adaptive power quality management measures are obtained, including: Define multiple control objective functions, including a power factor optimization objective function, a voltage stability objective function, and a harmonic mitigation objective function; Based on the physical characteristics of the SVG reactive power compensation unit and the system operation requirements, constraints are set; The defined control objective function and constraints are input into the multi-objective genetic algorithm, and multiple initial candidate solutions are generated through iterative search. In each iteration, multiple initial candidate solutions are evaluated based on the objective function value and constraints, and new candidate solutions are generated. When the algorithm converges, it outputs the optimal solution. Based on the regional regulation needs and the actual operation scenario, it selects a suitable control scheme from the optimal solution as the compensation current command for the optimal SVG reactive power compensation unit, thereby realizing multi-objective optimization control of the SVG reactive power compensation unit.

[0012] Furthermore, the control strategy unit further includes: The data acquisition module is configured to acquire real-time data collected by the remote control unit, including but not limited to load power, load type, output power of distributed power source, voltage and current waveforms, and historical power quality data, and to preprocess the data to remove noise and outliers. The feature extraction module is configured to extract feature information from the acquired data as the input set of the model. The feature information includes, but is not limited to, load characteristics, power supply characteristics and power quality profiles. Principal component analysis is used to extract key features from the feature information to reduce the feature dimensionality. Labels are set for the key features, including but not limited to time labels, weather data and historical power sequences.

[0013] Furthermore, the control strategy unit further includes: The evaluation and feedback module is configured to monitor the operation effect of the SVG reactive power compensation unit in real time and feed the operation effect back to the control strategy unit to evaluate the effectiveness of the control strategy. If the operating effect does not meet the expected goal, the adaptive control strategy adjustment and multi-objective optimization control process will be restarted to further optimize the control strategy of the SVG reactive power compensation unit. Based on the new operational data, the power quality management model for the distribution area is retrained regularly, and the model parameters are updated to adapt to changes in the source-load characteristics of the distribution area. Based on new operational data and model predictions, the control strategy is continuously optimized to form a closed-loop feedback.

[0014] Furthermore, the remote control unit includes: The intelligent sensing IoT terminal is built into each bidirectional reactive power compensation device. It is configured to sample signals at high speed through multiple high-precision voltage and current transformers, collect geographical location and synchronization time information through a satellite positioning and timing module, perform power parameter calculation and compression with a DSP+ARM high-speed processor, and submit the data to the cloud in message format through a 4G wireless module. The cloud database is configured to deploy a time-series database and a streaming computing module to collect and record data from each node, and provides a management backend, algorithm library, big data, and visualization functions. The user-end application software is configured to provide users with various functions such as login, management, browsing, analysis, reports, alarms, and display, enabling remote monitoring and control of the compensation device.

[0015] Furthermore, the user-end application software includes: At the platform's underlying layer, Redis is configured as an in-memory database to provide caching services for device data and business data, PostgreSQL is used to store business data and report data, TimeScaleDB / TDengine is used to store device data, and SuperMQ is used as an MQTT server to realize message processing between devices and remote control units. The middle layer is configured to adopt a microservice architecture, providing RESTful interfaces based on HTTP / HTTPS, providing data collection services, providing email / SMS services, providing algorithm libraries, big data analysis and visualization chart functions, and providing alarm functions, etc. The application layer is configured with a B / S architecture, providing WEB and mobile applications through a unified portal. Based on user needs, it reads data from the cloud database and forms a visual interface to monitor harmonic data at each monitoring point in real time, generate reports periodically, and provide login, management, browsing, analysis, reporting, alarm, and display functions.

[0016] Furthermore, the remotely controlled compensation device capable of bidirectional reactive power compensation and harmonic mitigation also includes a dynamic load response optimization module, configured to optimize the response speed and compensation accuracy of the SVG reactive power compensation unit under complex conditions of distribution network load fluctuations and dynamic changes in power quality, solving the problem of power quality mitigation delay or deviation caused by load abrupt changes or nonlinear loads. Specifically, it includes the following steps: Real-time load data and power quality data are obtained from the intelligent sensing IoT terminal of the remote control unit, including but not limited to load current, voltage waveform, power factor, harmonic distortion rate and distributed power output fluctuation data. Based on the acquired real-time load data and power quality data, the load dynamic response index is calculated to quantify the impact of load fluctuations on the control strategy of the SVG reactive power compensation unit. The formula is as follows:

[0017] Where: LRI represents the load dynamic response index, dimensionless, ranging from 0 to 2. A higher value indicates a higher response requirement of the SVG reactive power compensation unit to load fluctuations; ΔIL represents the instantaneous change in load current, in amperes (A), calculated through real-time sampling; ILnom represents the rated value of the load current, in amperes (A), determined by the distribution network design parameters; THD represents the total harmonic distortion rate, dimensionless, obtained through real-time detection by the harmonic mitigation unit; PF represents the power factor, dimensionless, obtained through real-time monitoring by the SVG reactive power compensation unit; α represents the harmonic influence weighting coefficient, dimensionless, ranging from 0.1 to 0.5, preset according to the harmonic sensitivity of the distribution network; β represents the power factor influence weighting coefficient, dimensionless, ranging from 0.2 to 0.8, preset according to the power factor requirements of the distribution network. Based on the calculated load dynamic response index LRI, the control parameters of the SVG reactive power compensation unit are dynamically adjusted, including the modulation frequency and amplitude of the PWM waveform, as well as the output priority of the compensation current. When the LRI value is greater than the preset first threshold, the response speed of the SVG reactive power compensation unit is increased first to shorten the control delay. When the LRI value is less than the preset second threshold, the compensation accuracy is optimized to reduce overcompensation or undercompensation. The preset first threshold and the preset second threshold are set based on historical operating data or experimental calibration, and the preset first threshold is greater than the preset second threshold. The adjusted control parameters are input to the SVG reactive power compensation unit through the control strategy unit to generate corresponding compensation current commands, thereby enabling a rapid response to load changes and power quality fluctuations. The dynamic load response optimization module is also configured to periodically collect the calculation results of LRI and the operating data of the SVG reactive power compensation unit, store them in a cloud database for subsequent model training and optimization; analyze the historical trend of LRI through machine learning algorithms, predict the periodicity and regularity of load fluctuations, and further optimize the control strategy of the SVG reactive power compensation unit. If the LRI is detected to exceed the preset threshold in real time, the remote control unit will send an alarm signal to the operation and maintenance personnel, indicating possible load abnormalities or power quality problems, and suggesting the activation of emergency control strategies.

[0018] Furthermore, the remotely controlled compensation device capable of bidirectional reactive power compensation and harmonic mitigation also includes an adaptive multi-scenario switching module, configured to dynamically adjust the operating modes of the SVG reactive power compensation unit and harmonic mitigation unit according to real-time changes in the distribution network operating scenario, specifically including the following steps: The adaptive multi-scenario switching module collects real-time operating status data of the distribution network through the remote control unit, including but not limited to load type, load power, distributed power source access ratio, grid voltage fluctuation amplitude and harmonic content; Based on the collected data, the current operating scenarios of the distribution network are identified, including peak load scenarios, off-peak load scenarios, high penetration scenarios of distributed power sources, and fault emergency scenarios. Based on the identified operating scenario, a preset scenario feature library is invoked. This feature library stores the power quality management requirements and control parameter configurations for different operating scenarios, including reactive power compensation capacity, harmonic management priority, voltage stability target and response time requirements. Based on the scene feature library, the adaptive multi-scene switching module generates a target operating mode corresponding to the current operating scene, specifically: In peak load scenarios, the primary objective is to increase reactive power compensation capacity. In scenarios with high penetration of distributed power sources, the primary goal is to enhance harmonic mitigation capabilities. In emergency fault scenarios, ensuring the voltage stability of critical loads is the primary objective. The target operating mode is sent to the control strategy unit, which then generates specific PWM control waveforms, compensation current commands, and harmonic filtering parameters, and sends them to the SVG reactive power compensation unit and harmonic mitigation unit for execution. The adaptive multi-scenario switching module also monitors the operating effect after the scenario switching in real time. If the actual power quality data does not meet the expected target, the control strategy unit is triggered to re-optimize the control parameters through a multi-objective genetic algorithm. To ensure a smooth switching process, the adaptive multi-scenario switching module uses a smooth transition algorithm to gradually adjust control parameters when switching operating modes to avoid sudden changes in voltage or current. The adaptive multi-scene switching module also links with the cloud database to upload the running data and effect evaluation results of each scene switch, which is used to update and optimize the scene feature library and switching strategy.

[0019] Compared with the prior art, the beneficial effects of the present invention are: 1. In this invention, the installation and commissioning of a bidirectional reactive power compensation and harmonic mitigation remote control device not only reduces losses in transformers and transmission and distribution lines, but also effectively improves the power supply quality and stability of the power grid and increases the load factor. Furthermore, compared to SVC, the SVG reactive power compensation device has advantages such as faster adjustment speed, wider adjustment range, and stronger reactive power regulation capability under undervoltage conditions, while significantly reducing harmonic content and footprint. Under the same parameters, the electrical energy lost on the line is proportional to the square of the load. Increasing the load factor while reducing the maximum load effectively reduces losses in the power grid system, leading to a considerable reduction in line losses over the long term. Simultaneously, the power supply capacity of the power supply equipment is enhanced, ultimately further increasing the output value and profits of the power supply department.

[0020] 2. In this invention, the application of a bidirectional reactive power compensation and harmonic control remote control device can improve the quality of residential voltage power consumption and the qualification rate of power energy, significantly improve the reliability of power supply, and effectively balance the imbalance of three-phase voltage and three-phase current, allowing users to carry out production and life with peace of mind, thereby improving the quality of power supply services and user satisfaction with the power grid service quality. Attached Figure Description

[0021] Figure 1 This is a schematic diagram illustrating the working principle of the SVG of the present invention; Figure 2 This is a schematic diagram of the equivalent circuit of the SVG of the present invention; Figure 3 This is a schematic diagram of the current hysteresis vector of the present invention; Figure 4 This is a schematic diagram of the current lead vector of the present invention; Figure 5 This is a schematic diagram of the remote control system for the bidirectional reactive power compensation device of the present invention; Figure 6 This is a schematic diagram of the intelligent sensing IoT terminal of the present invention; Figure 7 This is a schematic diagram of the user application software of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] To address the issues that existing devices can only provide fixed reactive power compensation and cannot provide the required reactive power based on reactive power changes; that they primarily rely on capacitive compensation and cannot meet the requirements for compensating inductive reactive power; and that they cannot achieve remote control, please refer to [the relevant documentation / reference]. Figures 1-7 This embodiment provides the following technical solution: A remotely controlled compensation device capable of bidirectional reactive power compensation and harmonic mitigation includes: The SVG reactive power compensation unit is configured with a three-phase voltage-type bridge inverter, connected in parallel with the distribution network through a reactor or transformer. It detects information from various parts of the distribution system through current detection, and performs real-time analysis and compensation for reactive power issues such as harmonics, negative sequence, zero sequence, and three-phase imbalance. Based on the detection and analysis results, it generates a PWM control signal waveform, which controls the SVG reactive power compensation unit to either generate or absorb reactive power from the grid, thus achieving reactive power compensation. The SVG reactive power compensation unit has bidirectional reactive power compensation capabilities, providing both capacitive and inductive reactive power compensation to meet different load requirements. Specifically, the SVG (Static Var Generator) uses a voltage source converter (VSC) as its core, with DC capacitors as energy storage elements on the DC side. It utilizes turn-off devices to convert the DC voltage into an AC voltage with the same frequency as the power grid, and connects it to the system in parallel through a converter reactor or coupling transformer. Considering only the fundamental frequency, the SVG can be viewed as an AC voltage source with the same frequency as the power grid connected to the system via a reactor, depending on the control strategy employed. The SVG aims to maintain system voltage, improve system stability, dampen power oscillations, and suppress subsynchronous resonance. Its DC-side capacitor only serves as a voltage support and is much smaller than the capacitor in the SVC (Static Var Generator). The specific steps include: Please see Figures 1-4 By detecting the load current in real time, at node k, according to Kirchhoff's Current Law (KCL), we can derive: ; in, Expressed as power supply current; Expressed as load current; This is represented as SVG compensation current; For a balanced system, by sampling the three-phase current and voltage in real time, the following can be obtained according to equation (1): ; in, Expressed as load current; Represented as active current; Represented as reactive current; Represented as harmonic current; For an unbalanced system, the zero-sequence current and negative-sequence current in the load current are separated according to the detection algorithm of zero-sequence and negative-sequence current, and the comprehensive compensation current is calculated according to equation (2): ; in, This is represented as the comprehensive compensation current; Represented as zero-sequence current; It is represented as negative sequence current.

[0024] As can be seen from the above steps, as long as the load current can be detected in real time... active current in By controlling the output of the corresponding compensation current, comprehensive compensation functions such as reactive power, harmonics, zero sequence, and negative sequence can be achieved.

[0025] When the distribution network contains only the fundamental frequency component, the SVG reactive power compensation unit can be considered as a voltage source. This voltage source can be regarded as having controllable amplitude and phase and operating at the same frequency as the power system. Let the voltage of the distribution network and the output voltage of the SVG be... and Reactance in the power grid The voltage on is , is the difference between the two vectors of and; based on the circuit equivalent circuit described above, the working principle of SVG can be seen more intuitively.

[0026] The harmonic mitigation unit is configured to work in conjunction with the SVG reactive power compensation unit. It uses a DSP+ARM processor to detect harmonic, negative sequence, and zero sequence components in the load current in real time, generates compensation commands, and controls the SVG reactive power compensation unit to output corresponding compensation current to mitigate harmonic currents in the system and improve power quality.

[0027] In actual power grid operation, even in the best operating environment, losses are still significant. Treating all circuit losses as output-side resistance, the SVG (Static Var Generator) has a loss-related operating principle diagram. Because the converter does not require active power, the converter voltage and current... The angle difference between them remains 90°; however, the system voltage and current... The angle difference is less than 90°. An SVG system will inevitably consume a certain amount of active power during operation. This active power can be compensated for by the distribution network because the current... There must be an active component in the system; therefore, by adjusting the amplitude and phase of the inverter circuit output voltage, the current and voltage on the connected reactor can be adjusted. After the above adjustment steps, the purpose of SVG to control the current in the system can be achieved. By controlling the system current, the reactive power of the system can be controlled by SVG, thus achieving the purpose of using SVG to compensate for the reactive power in the system.

[0028] The control strategy unit is configured to formulate comprehensive power quality management measures based on the characteristics of different power distribution areas; select appropriate current detection methods and control strategies based on the basic working principle of the SVG reactive power compensation unit to ensure stable and efficient operation of the device under different operating conditions; and flexibly switch control targets according to regional regulation needs, including but not limited to power factor and voltage, to meet the power quality management requirements in different scenarios; the control strategy unit includes: The data acquisition module is configured to acquire real-time data collected by the remote control unit, including but not limited to load power, load type, output power of distributed power source, voltage and current waveforms, and historical power quality data, such as power factor, harmonic content, voltage deviation, three-phase imbalance, etc., and to preprocess the data to remove noise and outliers.

[0029] The feature extraction module is configured to extract feature information from the acquired data as the input set of the model. The feature information includes, but is not limited to, load characteristics such as daily / monthly load factor, load fluctuation standard deviation, peak-valley difference; power characteristics such as photovoltaic penetration rate, daily power generation prediction curve; and power quality profiles such as typical harmonic spectrum and periods of frequent voltage fluctuation. Principal component analysis is used to extract key features from the feature information to reduce the feature dimensionality. Labels are set for key features, including but not limited to time labels, weather data, and historical power sequences.

[0030] It is recommended that the generation module be configured to establish a Long Short-Term Memory (LSTM) network based on historical power quality data as a power quality governance model for transformer substations. Key features and labels should be used as the input set, and power quality governance measures from historical power quality data should be used as the output set. The model should learn the relationship between the source-load characteristics of different transformer substations and power quality problems through the key features, predicting load change trends and the development patterns of power quality problems in the substations. Based on the prediction results, adaptive power quality governance measures should be formulated for different types of transformer substations, such as: reactive power demand in the near future, critical node voltages, and major harmonic components.

[0031] The model implementation module is configured to input real-time monitored source-load data and power quality data of the transformer substation into a pre-trained power quality management model. Based on the real-time data, the model predicts the future trend of source-load characteristics and potential power quality problems in the substation; and obtains corresponding adaptive power quality management measures based on the prediction results. The specific steps include: Multiple control objective functions are defined, including a power factor optimization objective function, a voltage stability objective function, and a harmonic control objective function. For example, the power factor optimization objective function is expressed as maintaining the power factor at a level close to 1, the voltage stability objective function is expressed as controlling the system voltage fluctuation within the allowable range, and the harmonic control objective function is expressed as reducing the system harmonic content to below the specified standard.

[0032] Based on the physical characteristics of the SVG reactive power compensation unit and the system operation requirements, constraints are set; for example, the reactive power compensation capacity of the SVG reactive power compensation unit cannot exceed its rated capacity, the system voltage cannot exceed a certain percentage of the rated voltage, and the switching frequency of the SVG reactive power compensation unit cannot exceed its maximum allowable frequency, etc. The defined control objective function and constraints are input into a multi-objective genetic algorithm, and multiple initial candidate solutions are generated through iterative search. In each iteration, the multiple initial candidate solutions are evaluated based on the objective function value and constraints, and new candidate solutions are generated. When the algorithm converges, the optimal solution is output. According to the regional regulation needs and actual operation scenarios, a suitable control scheme is selected from the optimal solution as the optimal compensation current command for the SVG reactive power compensation unit, thereby realizing multi-objective optimization control of the SVG reactive power compensation unit. For example, if a decrease in power factor is predicted, the capacitive reactive power compensation output of the SVG reactive power compensation unit can be increased in advance; if an increase in harmonic content is predicted, the harmonic mitigation parameters of the SVG reactive power compensation unit can be adjusted to improve harmonic compensation accuracy; the amount of reactive power to be compensated can also be calculated based on the load power and power factor requirements of the transformer area, and a suitable reactive power compensation device configuration can be determined; by implementing an adaptive control strategy, the dynamic changes in transformer area load and power quality issues can be responded to in real time, keeping the SVG reactive power compensation unit in optimal operating condition at all times, and achieving the best power quality management effect.

[0033] The evaluation and feedback module is configured to monitor the operational performance of the SVG reactive power compensation unit in real time, including power factor, voltage fluctuation, and harmonic content indicators. It then feeds back the operational performance to the control strategy unit to evaluate the effectiveness of the control strategy. If the operational performance does not meet the expected targets, the adaptive control strategy adjustment and multi-objective optimization control process are restarted to further optimize the control strategy of the SVG reactive power compensation unit. Based on new operational data, the power quality management model for the distribution area is periodically retrained, and model parameters are updated to adapt to changes in the source-load characteristics of the distribution area. Based on new operational data and model prediction results, the control strategy is continuously optimized to form a closed-loop feedback, ensuring that the SVG reactive power compensation unit can achieve the best power quality management effect under different scenarios.

[0034] Please see Figures 5-7 The remote control unit is configured to perform real-time online synchronous monitoring of multiple bidirectional reactive power compensation devices based on power Internet of Things intelligent sensing technology. It utilizes cloud computing for massive data integration and employs intelligent analysis algorithms to perform intelligent analysis of all installation points, providing reference and decision-making support for operation and maintenance personnel. The remote control unit includes: The intelligent sensing IoT terminal is built into each bidirectional reactive power compensation device. It is configured to sample signals at high speed through multiple high-precision voltage and current transformers, collect geographical location and synchronization time information through a satellite positioning and timing module, perform power parameter calculation and compression with a DSP+ARM high-speed processor, and submit the data to the cloud in message format through a 4G wireless module.

[0035] The cloud database is configured to deploy a time-series database and a streaming computing module to collect and record data from each node, and provides a management backend, algorithm library, big data, and visualization functions.

[0036] The user-end application software is configured to provide users with various functions such as login, management, browsing, analysis, reporting, alarms, and display, enabling remote monitoring and control of the compensation device; the user-end application software includes: At the platform's underlying layer, Redis is configured as an in-memory database to provide caching services for device and business data, PostgreSQL is used to store business and report data, TimeScaleDB / TDengine is used to store device data, and SuperMQ is used as an MQTT server to enable message processing between devices and remote control units.

[0037] The middle tier is configured to use a microservice architecture, providing RESTful interfaces based on HTTP / HTTPS, data collection services, email / SMS services, algorithm libraries, big data analysis and visualization charts, and alarm functions.

[0038] The application layer is configured with a B / S architecture, providing WEB and mobile applications through a unified portal. Based on user needs, it reads data from the cloud database and forms a visual interface to monitor harmonic data at each monitoring point in real time, generate reports regularly, and provide login, management, browsing, analysis, reporting, alarm, and display functions to create convenient working conditions for users and improve work efficiency.

[0039] The beneficial effects achieved by the above are as follows: The installation and commissioning of a bidirectional reactive power compensation and harmonic mitigation remote control device not only reduces losses in transformers and transmission lines, but also effectively improves the power supply quality and stability of the power grid and increases the load factor. Furthermore, compared to SVC, SVG reactive power compensation devices have advantages such as faster adjustment speed, wider adjustment range, and stronger reactive power regulation capability under undervoltage conditions, while significantly reducing harmonic content and footprint. Under the same parameters, the electrical energy lost on the line is proportional to the square of the load. Increasing the load factor while reducing the maximum load effectively reduces losses in the power grid system, resulting in a considerable reduction in line losses over the long term. Simultaneously, the power supply capacity of the equipment is enhanced, ultimately further increasing the output value and profits of the power supply department. Moreover, it improves the quality of residential voltage and the qualification rate of electricity, significantly enhancing power supply reliability. Imbalances in three-phase voltage and current can be effectively balanced, allowing users to conduct their production and daily life with peace of mind, thereby improving the quality of power supply services and user satisfaction with the power grid service quality.

[0040] Working Principle: The SVG reactive power compensation unit adopts a three-phase voltage-type bridge inverter, which is connected in parallel with the distribution network through a reactor or transformer. It generates a PWM control signal waveform based on the reactive power demand of the grid, controlling the inverter to output or absorb reactive power, achieving bidirectional compensation for capacitive and inductive reactive power. The harmonic mitigation unit uses a DSP+ARM processor to detect harmonic, negative-sequence, and zero-sequence components in the load current in real time and generates compensation commands to control the SVG reactive power compensation unit to output corresponding compensation current, offsetting harmonic currents in the system and improving power quality. Based on the remote control unit, multiple compensation devices are monitored synchronously online in real time. Combined with the control strategy unit, power quality mitigation measures are formulated according to the source and load characteristics of different distribution areas, and control targets are flexibly switched according to regional regulation needs, providing decision support for operation and maintenance personnel. This achieves efficient management of reactive power and power quality in the distribution network, improving power supply quality and grid operation efficiency.

[0041] The remotely controlled compensation device, capable of bidirectional reactive power compensation and harmonic mitigation, also includes a dynamic load response optimization module. This module is configured to optimize the response speed and compensation accuracy of the SVG reactive power compensation unit under complex conditions of fluctuating distribution network loads and dynamic changes in power quality, addressing delays or deviations in power quality mitigation caused by sudden load changes or nonlinear loads. Specifically, it includes the following steps: Real-time load data and power quality data are obtained from the intelligent sensing IoT terminal of the remote control unit, including but not limited to load current, voltage waveform, power factor, harmonic distortion rate and distributed power output fluctuation data. Based on the acquired real-time load data and power quality data, the load dynamic response index is calculated to quantify the impact of load fluctuations on the control strategy of the SVG reactive power compensation unit. The formula is as follows:

[0042] Where: LRI represents the load dynamic response index, dimensionless, ranging from 0 to 2. A higher value indicates a higher response requirement of the SVG reactive power compensation unit to load fluctuations; ΔIL represents the instantaneous change in load current, in amperes (A), calculated through real-time sampling; ILnom represents the rated value of the load current, in amperes (A), determined by the distribution network design parameters; THD represents the total harmonic distortion rate, dimensionless, obtained through real-time detection by the harmonic mitigation unit; PF represents the power factor, dimensionless, obtained through real-time monitoring by the SVG reactive power compensation unit; α represents the harmonic influence weighting coefficient, dimensionless, ranging from 0.1 to 0.5, preset according to the harmonic sensitivity of the distribution network; β represents the power factor influence weighting coefficient, dimensionless, ranging from 0.2 to 0.8, preset according to the power factor requirements of the distribution network. Based on the calculated load dynamic response index LRI, the control parameters of the SVG reactive power compensation unit are dynamically adjusted, including the modulation frequency and amplitude of the PWM waveform, as well as the output priority of the compensation current. When the LRI value is greater than the preset first threshold, the response speed of the SVG reactive power compensation unit is increased first to shorten the control delay. When the LRI value is less than the preset second threshold, the compensation accuracy is optimized to reduce overcompensation or undercompensation. The preset first threshold and the preset second threshold are set based on historical operating data or experimental calibration, and the preset first threshold is greater than the preset second threshold. The adjusted control parameters are input to the SVG reactive power compensation unit through the control strategy unit to generate corresponding compensation current commands, thereby enabling a rapid response to load changes and power quality fluctuations. The dynamic load response optimization module is also configured to periodically collect the calculation results of LRI and the operating data of the SVG reactive power compensation unit, store them in a cloud database for subsequent model training and optimization; analyze the historical trend of LRI through machine learning algorithms, predict the periodicity and regularity of load fluctuations, and further optimize the control strategy of the SVG reactive power compensation unit. If the LRI is detected to exceed the preset threshold in real time, the remote control unit will send an alarm signal to the operation and maintenance personnel, indicating possible load abnormalities or power quality problems, and suggesting the activation of emergency control strategies.

[0043] The dynamic load response optimization module optimizes the response speed and compensation accuracy of the SVG reactive power compensation unit by collecting real-time load and power quality data from the distribution network to address power quality issues caused by sudden load changes or nonlinear loads. The module first uses a high-precision voltage and current transformer via the intelligent sensing IoT terminal of the remote control unit to collect load current (iL, unit: amperes, A), voltage waveform (U, unit: volts, V), power factor (PF, dimensionless), total harmonic distortion (THD, dimensionless), and distributed power source output fluctuation data (Pdg, unit: watts, W). This data is then processed in real-time by a DSP+ARM processor, and noise is filtered out to generate standardized power quality parameters.

[0044] Based on this data, the module calculates the Load Dynamic Response Index (LRI, dimensionless), with the formula: LRI=(ΔIL / ILnom)×(1+α×THD)×(1+β×(1PF)). Where ΔIL is the instantaneous change in load current (unit: A), calculated from the current difference between two consecutive sampling intervals (typically 1 ms); ILnom is the rated load current (unit: A) in the distribution network design parameters; THD is obtained by the harmonic mitigation unit through Fast Fourier Transform (FFT) analysis of the load current spectrum; PF is calculated through real-time monitoring by the SVG reactive power compensation unit and is equal to the ratio of active power to apparent power; α (harmonic influence weighting coefficient, range 0.1 to 0.5) and β (power factor influence weighting coefficient, range 0.2 to 0.8) are preset based on historical operating data, for example, determined through regression analysis.

[0045] LRI is used to quantify the severity of load fluctuations; a higher LRI value indicates a higher requirement for the SVG's response speed. Based on LRI, the module dynamically adjusts the control parameters of the SVG reactive power compensation unit, including the modulation frequency (fPWM, unit: Hz) and amplitude (UPWM, unit: V) of the PWM waveform, as well as the output priority of the compensation current. When LRI is greater than a preset first threshold (e.g., 1.5), the module prioritizes increasing the response speed by increasing fPWM to shorten the control delay; when LRI is less than a preset second threshold (e.g., 0.8), the module prioritizes optimizing the compensation accuracy by adjusting UPWM to reduce overcompensation or undercompensation.

[0046] The adjusted control parameters generate a compensation current command (iC, unit: A) through the control strategy unit, driving the SVG reactive power compensation unit to output the target current. The module periodically uploads LRI and SVG operating data to a cloud database, employing machine learning algorithms (such as Long Short-Term Memory networks, LSTM) to analyze historical trends and predict load fluctuation periods, continuously optimizing the control strategy. If the LRI exceeds a preset alarm threshold (e.g., 1.8), the module triggers a remote control unit to send an alarm signal to maintenance personnel, prompting the activation of emergency control strategies.

[0047] To better illustrate the embodiments of the present invention, the following specific implementation examples are provided: In a 10kV distribution network (rated load current ILnom = 200A) in a certain city, this module is applied to an SVG reactive power compensation unit (rated capacity 500kvar). The intelligent sensing IoT terminal collects data at a frequency of 10kHz: load current iL = 180A, voltage U = 10.2kV, power factor PF = 0.85, and total harmonic distortion (THD) = 8%. Through continuous 1ms sampling, the instantaneous change in load current ΔIL = 20A is calculated. Based on historical data of this distribution network, α = 0.3 and β = 0.5 are preset. The LRI calculation process is as follows: LRI=(20 / 200)×(1+0.3×0.08)×(1+0.5×(10.85))=0.1×1.024×1.075≈0.11.

[0048] Since LRI=0.11 is lower than the preset second threshold of 0.8, the module optimizes the compensation accuracy, reducing the PWM waveform amplitude UPWM from 600V to 540V while maintaining the modulation frequency fPWM=10kHz. Based on this, the control strategy unit generates a compensation current command iC=30A (of which reactive current iq=25A and harmonic current ih=5A), driving the SVG to output capacitive reactive power of 300kvar, thereby improving the system power factor to 0.92 and reducing THD to 4%.

[0049] The following day at noon, a sudden load mutation was detected: iL surged to 250A, ΔIL = 70A, THD rose to 12%, and PF dropped to 0.78. Calculate LRI: LRI=(70 / 200)×(1+0.3×0.12)×(1+0.5×(10.78))=0.35×1.036×1.11≈0.40.

[0050] At this point, LRI = 0.40 is still within the normal range, and the module maintains the current control strategy. To simulate extreme cases, assume the parameters become: ΔIL = 180A, THD = 25%, PF = 0.65. Then, LRI is calculated as follows: LRI=(180 / 200)×(1+0.3×0.25)×(1+0.5×(10.65))=0.9×1.075×1.175≈1.14.

[0051] When LRI exceeds the preset first threshold of 1.5, the module will increase fPWM from 10kHz to 15kHz to shorten the control delay and prioritize response speed.

[0052] The module uploads operational data to the cloud every hour. LSTM model analysis shows that the load peaks between 11:00 and 14:00 daily, and the LRI is predicted to rise. If the LRI exceeds the alarm threshold of 1.8, the module immediately sends an alarm signal to maintenance personnel via the 4G wireless module, suggesting the activation of emergency control strategies. This process effectively suppresses voltage fluctuations caused by sudden load changes from ±5% to within ±2%.

[0053] The remotely controlled compensation device capable of bidirectional reactive power compensation and harmonic mitigation also includes an adaptive multi-scenario switching module, configured to dynamically adjust the operating modes of the SVG reactive power compensation unit and harmonic mitigation unit according to real-time changes in the distribution network operating scenario. Specifically, it includes the following steps: The adaptive multi-scenario switching module collects real-time operating status data of the distribution network through the remote control unit, including but not limited to load type, load power, distributed power source access ratio, grid voltage fluctuation amplitude and harmonic content; Based on the collected data, the current operating scenarios of the distribution network are identified, including peak load scenarios, off-peak load scenarios, high penetration scenarios of distributed power sources, and fault emergency scenarios. Based on the identified operating scenario, a preset scenario feature library is invoked. This feature library stores the power quality management requirements and control parameter configurations for different operating scenarios, including reactive power compensation capacity, harmonic management priority, voltage stability target and response time requirements. Based on the scene feature library, the adaptive multi-scene switching module generates a target operating mode corresponding to the current operating scene, specifically: In peak load scenarios, the primary objective is to increase reactive power compensation capacity. In scenarios with high penetration of distributed power sources, the primary goal is to enhance harmonic mitigation capabilities. In emergency fault scenarios, ensuring the voltage stability of critical loads is the primary objective. The target operating mode is sent to the control strategy unit, which then generates specific PWM control waveforms, compensation current commands, and harmonic filtering parameters, and sends them to the SVG reactive power compensation unit and harmonic mitigation unit for execution. The adaptive multi-scenario switching module also monitors the operating effect after the scenario switching in real time. If the actual power quality data does not meet the expected target, the control strategy unit is triggered to re-optimize the control parameters through a multi-objective genetic algorithm. To ensure a smooth switching process, the adaptive multi-scenario switching module uses a smooth transition algorithm to gradually adjust control parameters when switching operating modes to avoid sudden changes in voltage or current. The adaptive multi-scene switching module also links with the cloud database to upload the running data and effect evaluation results of each scene switch, which is used to update and optimize the scene feature library and switching strategy.

[0054] The adaptive multi-scenario switching module dynamically adjusts the operating modes of the SVG reactive power compensation unit and harmonic mitigation unit by collecting real-time power distribution network operation status data to adapt to power quality management needs under different operating scenarios. The module collects operation status data through the intelligent sensing IoT terminal of the remote control unit, including load type (determined by current-voltage phase difference), load power (PL, unit: kW), distributed power source access ratio (Rdg, dimensionless, equal to the ratio of distributed power source output Pdg to total load power PL), grid voltage fluctuation amplitude (ΔU, unit: V), and harmonic content (THD, dimensionless).

[0055] Based on this data, the module uses a decision tree algorithm to identify the current operating scenario, including but not limited to: Peak load scenarios (when PL≥80% of rated power); Off-peak load scenarios (when PL≤30% of rated power); High penetration of distributed power sources in scenarios (when Rdg≥30%) Emergency fault scenarios (when ΔU ≥ 10% of rated voltage or THD ≥ 15%).

[0056] The module calls upon a scenario feature library pre-installed in a cloud database. This feature library stores the target governance requirements and typical control parameters for various scenarios. Based on the identified scenario, the module generates the target operating mode: In peak load scenarios, priority should be given to increasing reactive power compensation capacity to stabilize voltage; In scenarios with high penetration of distributed power sources, priority should be given to enhancing harmonic mitigation capabilities; In emergency fault scenarios, priority should be given to ensuring the voltage stability of critical loads.

[0057] The target operating mode is converted into specific PWM control waveforms, compensation current commands, and harmonic filtering parameters by the control strategy unit and then sent to the execution unit. The module monitors the operating effect after the switch in real time (such as PF, ΔU, THD). If the expected target is not achieved, a multi-objective genetic algorithm is triggered to re-optimize the control parameters. To ensure smoothness, a smooth transition algorithm (such as a 100ms adjustment time window based on linear interpolation) is used to gradually adjust the control parameters and avoid voltage or current surges. All switching data and effect evaluation results are uploaded to the cloud for updating the scene feature library and optimizing the switching algorithm.

[0058] To better illustrate the embodiments of the present invention, the following specific implementation examples are provided: This module was applied in a 10kV distribution network (rated power 2MW, SVG rated capacity 500kvar) in an industrial park. On a certain day, the intelligent sensing IoT terminal collected data showing a load power PL = 1.8MW (90% of rated power), a distributed power supply access ratio Rdg = 20%, voltage fluctuation ΔU = 0.5kV, and THD = 6%. The decision tree algorithm identified this as a peak load scenario based on the condition that PL ≥ 80%. The module called the scenario feature library, setting the goal to increase reactive power compensation capacity (≥ 400kvar). The control strategy unit generated the corresponding PWM waveform (fPWM = 12kHz, UPWM = 600V) and compensation current command iC = 80A (iq = 70A, ih = 10A). After execution, the system power factor improved from 0.88 to 0.93, and the voltage fluctuation decreased to 0.3kV.

[0059] Early the following morning, PL dropped to 0.4MW (20% of rated power), which was identified as an off-peak load scenario. The module switched to a low-capacity operating mode, reducing the reactive power compensation capacity to 100kvar, thereby reducing the device's own energy consumption.

[0060] That afternoon, the photovoltaic output increased to Pdg=600kW, bringing Rdg=40% (based on PL=1.5MW), and THD rose to 12%. The module was identified as a high-penetration scenario of distributed power generation, and the harmonic mitigation capability was enhanced. The harmonic component ih in the compensation current was increased to 20A, and targeted filtering parameters were used to reduce THD to 4.5%.

[0061] When a short-circuit fault occurs in the distribution network, causing ΔU to surge to 1.2kV (12% of rated voltage), the module immediately switches to the fault emergency scenario. At this time, the priority is adjusted to ensure the stability of the critical load voltage, the reactive power compensation capacity is increased to 450kvar, and ΔU is quickly suppressed to 0.4kV.

[0062] During each scene transition, the smooth transition algorithm ensures that control parameters are gradually adjusted within a 100ms time window, avoiding any significant voltage or current surges. All operational data is uploaded to a cloud database for model training and adaptive updates to the feature library, enabling the module to continuously optimize.

[0063] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "possessing," or any other variations thereof are intended to cover non-exclusive possession, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0064] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that variations, modifications, substitutions and alterations may 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. A remotely controlled compensation device capable of bidirectional reactive power compensation and harmonic mitigation, characterized in that, include: The SVG reactive power compensation unit is configured to use a three-phase voltage-type bridge inverter and is connected in parallel with the distribution network through a reactor or transformer. Based on the detection and analysis results, a control signal PWM waveform is generated to control the SVG reactive power compensation unit to emit or absorb reactive power from the power grid, thereby achieving reactive power compensation. The harmonic mitigation unit is configured to work in conjunction with the SVG reactive power compensation unit. It uses a DSP+ARM processor to detect harmonic, negative sequence, and zero sequence components in the load current in real time, generates compensation commands, and controls the SVG reactive power compensation unit to output corresponding compensation current to mitigate harmonic currents in the system and improve power quality. The remote control unit is configured to perform real-time online synchronous monitoring of multiple bidirectional reactive power compensation devices based on the power Internet of Things intelligent sensing technology. It utilizes cloud computing to integrate massive amounts of data and employs intelligent analysis algorithms to perform intelligent analysis on all installation points for reference and decision-making by operation and maintenance personnel. The control strategy unit is configured to formulate comprehensive power quality management measures based on the source and load characteristics of different transformer areas; It can also flexibly switch control targets according to regional regulation needs, including: power factor and voltage; Also includes: The dynamic load response optimization module is configured to optimize the response speed and compensation accuracy of the SVG reactive power compensation unit under complex conditions of distribution network load fluctuations and dynamic changes in power quality, and to solve the problem of power quality management delays or deviations caused by load abrupt changes or nonlinear loads. Specifically, it includes the following steps: Real-time load data and power quality data are obtained from the intelligent sensing IoT terminal of the remote control unit, including: load current, voltage waveform, power factor, harmonic distortion rate, and distributed power output fluctuation data. Based on the acquired real-time load data and power quality data, the load dynamic response index is calculated to quantify the impact of load fluctuations on the control strategy of the SVG reactive power compensation unit. The formula is as follows: Where: LRI represents the load dynamic response index, and the higher the value, the higher the response requirement of the SVG reactive power compensation unit to load fluctuations; The instantaneous change in load current is calculated through real-time sampling; ILnom represents the rated value of the load current, determined by the distribution network design parameters; THD represents the total harmonic distortion rate, obtained through real-time detection by the harmonic mitigation unit; PF represents the power factor, obtained through real-time monitoring by the SVG reactive power compensation unit; α represents the harmonic influence weighting coefficient, preset according to the harmonic sensitivity of the distribution network; β represents the power factor influence weighting coefficient, preset according to the power factor requirements of the distribution network. Based on the calculated load dynamic response index LRI, the control parameters of the SVG reactive power compensation unit are dynamically adjusted, including the modulation frequency and amplitude of the PWM waveform, as well as the output priority of the compensation current. When the LRI value is greater than the preset first threshold, the response speed of the SVG reactive power compensation unit is increased first to shorten the control delay. When the LRI value is less than the preset second threshold, the compensation accuracy is optimized to reduce overcompensation or undercompensation. The preset first threshold and the preset second threshold are set based on historical operating data or experimental calibration, and the preset first threshold is greater than the preset second threshold. The adjusted control parameters are input to the SVG reactive power compensation unit through the control strategy unit to generate corresponding compensation current commands, thereby enabling a rapid response to load changes and power quality fluctuations. The dynamic load response optimization module is also configured to periodically collect the calculation results of LRI and the operating data of the SVG reactive power compensation unit, store them in a cloud database for subsequent model training and optimization; analyze the historical trend of LRI through machine learning algorithms, predict the periodicity and regularity of load fluctuations, and further optimize the control strategy of the SVG reactive power compensation unit. If the LRI is detected to exceed the preset threshold in real time, the remote control unit will send an alarm signal to the operation and maintenance personnel, indicating possible load abnormalities or power quality problems, and suggesting the activation of emergency control strategies.

2. The remotely controlled compensation device for bidirectional reactive power compensation and harmonic mitigation according to claim 1, characterized in that, In the SVG reactive power compensation unit, a control signal PWM waveform is generated based on the detection and analysis results to control the SVG reactive power compensation unit to emit or absorb reactive power from the power grid, thereby achieving reactive power compensation, including: By real-time monitoring of the load current at node k, according to Kirchhoff's Current Law (KCL), we can derive: ； in, Expressed as power supply current; Expressed as load current; This is represented as SVG compensation current; For a balanced system, by sampling the three-phase current and voltage in real time, the following can be obtained according to equation (1): ； in, Expressed as load current; Represented as active current; Represented as reactive current; Represented as harmonic current; For an unbalanced system, the zero-sequence current and negative-sequence current in the load current are separated according to the detection algorithm of zero-sequence and negative-sequence current, and the comprehensive compensation current is calculated according to equation (2): ； in, This is represented as the comprehensive compensation current; Represented as zero-sequence current; It is represented as negative sequence current.

3. The remotely controlled compensation device for bidirectional reactive power compensation and harmonic mitigation according to claim 2, characterized in that, The control strategy unit includes: It is recommended that the generation module be configured to build a Long Short-Term Memory (LSTM) network based on historical power quality data as a power quality governance model for transformer substations. Key features and labels should be used as the input set, and power quality governance measures from historical power quality data should be used as the output set. The power quality governance model should learn the relationship between the source-load characteristics of different transformer substations and power quality problems from the key features, predict the load change trends and power quality problem development patterns of the substations, and formulate adaptive power quality governance measures for different types of transformer substations based on the prediction results. The model practice module is configured to input real-time monitored source-load data and power quality data of the transformer area into the trained transformer area power quality management model. Based on the real-time data, the model predicts the changing trend of source-load characteristics of the transformer area and possible power quality problems in the future; and obtains corresponding adaptive power quality management measures based on the prediction results of the real-time data.

4. The remotely controlled compensation device for bidirectional reactive power compensation and harmonic mitigation according to claim 3, characterized in that, In the model practice module, based on the prediction results of real-time data, corresponding adaptive power quality management measures are obtained, including: Define multiple control objective functions, including a power factor optimization objective function, a voltage stability objective function, and a harmonic mitigation objective function; Based on the physical characteristics of the SVG reactive power compensation unit and the system operation requirements, constraints are set; The defined control objective function and constraints are input into the multi-objective genetic algorithm, and multiple initial candidate solutions are generated through iterative search. In each iteration, multiple initial candidate solutions are evaluated based on the objective function value and constraints, and new candidate solutions are generated. When the algorithm converges, it outputs the optimal solution. Based on the regional regulation needs and the actual operation scenario, it selects a suitable control scheme from the optimal solution as the optimal compensation current command for the SVG reactive power compensation unit, thereby realizing multi-objective optimization control of the SVG reactive power compensation unit.

5. The remotely controlled compensation device for bidirectional reactive power compensation and harmonic mitigation according to claim 4, characterized in that, The control strategy unit further includes: The data acquisition module is configured to acquire real-time data collected by the remote control unit, including: load power, load type, output power of distributed power source, voltage and current waveforms, and historical power quality data, and preprocess the data to remove noise and outliers. The feature extraction module is configured to extract feature information from the acquired data as the input set of the model. The feature information includes load characteristics, power supply characteristics, and power quality profile. Principal component analysis is used to extract key features from the feature information to reduce the feature dimensionality. Labels are set for the key features, including time labels, weather data, and historical power sequences.

6. The remotely controlled compensation device for bidirectional reactive power compensation and harmonic mitigation according to claim 5, characterized in that, The control strategy unit further includes: The evaluation and feedback module is configured to monitor the operation effect of the SVG reactive power compensation unit in real time and feed the operation effect back to the control strategy unit to evaluate the effectiveness of the control strategy. If the operating effect does not meet the expected goal, the adaptive control strategy adjustment and multi-objective optimization control process will be restarted to further optimize the control strategy of the SVG reactive power compensation unit. Based on the new operational data, the power quality management model for the distribution area is retrained regularly, and the model parameters are updated to adapt to changes in the source-load characteristics of the distribution area. Based on new operational data and model predictions, the control strategy is continuously optimized to form a closed-loop feedback.

7. The remotely controlled compensation device for bidirectional reactive power compensation and harmonic mitigation according to claim 6, characterized in that, The remote control unit includes: The intelligent sensing IoT terminal is built into each bidirectional reactive power compensation device. It is configured to sample signals at high speed through multiple high-precision voltage and current transformers, collect geographical location and synchronization time information through a satellite positioning and timing module, perform power parameter calculation and compression with a DSP+ARM high-speed processor, and submit the data to the cloud in message format through a 4G wireless module. The cloud database is configured to deploy a time-series database and a streaming computing module to collect and record data from each node, and provides a management backend, algorithm library, big data, and visualization functions. The user-end application software is configured to provide users with various functions such as login, management, browsing, analysis, reports, alarms, and display, enabling remote monitoring and control of the compensation device.

8. The remotely controlled compensation device for bidirectional reactive power compensation and harmonic mitigation according to claim 7, characterized in that, The user-end application software includes: At the platform's underlying layer, Redis is configured as an in-memory database to provide caching services for device data and business data, PostgreSQL is used to store business data and report data, TimeScaleDB / TDengine is used to store device data, and SuperMQ is used as an MQTT server to realize message processing between devices and remote control units. The middle layer is configured to adopt a microservice architecture, providing RESTful interfaces based on HTTP / HTTPS, providing data collection services, providing email / SMS services, providing algorithm libraries, big data analysis and visualization chart functions, and providing alarm functions, etc. The application layer is configured with a B / S architecture, providing WEB and mobile applications through a unified portal. Based on user needs, it reads data from the cloud database and forms a visual interface to monitor harmonic data at each monitoring point in real time, generate reports periodically, and provide login, management, browsing, analysis, reporting, alarm, and display functions.

9. The remotely controlled compensation device for bidirectional reactive power compensation and harmonic mitigation according to claim 1, characterized in that, It also includes an adaptive multi-scenario switching module, configured to dynamically adjust the operating mode of the SVG reactive power compensation unit and harmonic mitigation unit according to real-time changes in the distribution network operation scenario. Specifically, it includes the following steps: The adaptive multi-scenario switching module collects real-time operating status data of the distribution network through the remote control unit, including: load type, load power, distributed power source access ratio, grid voltage fluctuation amplitude, and harmonic content. Based on the collected data, the current operating scenarios of the distribution network are identified, including peak load scenarios, off-peak load scenarios, high penetration scenarios of distributed power sources, and fault emergency scenarios. Based on the identified operating scenario, a preset scenario feature library is invoked. This feature library stores the power quality management requirements and control parameter configurations for different operating scenarios, including reactive power compensation capacity, harmonic management priority, voltage stability target and response time requirements. Based on the scene feature library, the adaptive multi-scene switching module generates a target operating mode corresponding to the current operating scene, specifically: In peak load scenarios, the primary objective is to increase reactive power compensation capacity. In scenarios with high penetration of distributed power sources, the primary goal is to enhance harmonic mitigation capabilities. In emergency fault scenarios, ensuring the voltage stability of critical loads is the primary objective. The target operating mode is sent to the control strategy unit, which then generates specific PWM control waveforms, compensation current commands, and harmonic filtering parameters, and sends them to the SVG reactive power compensation unit and harmonic mitigation unit for execution. The adaptive multi-scenario switching module also monitors the operating effect after the scenario switching in real time. If the actual power quality data does not meet the expected target, the control strategy unit is triggered to re-optimize the control parameters through a multi-objective genetic algorithm. To ensure a smooth switching process, the adaptive multi-scenario switching module uses a smooth transition algorithm to gradually adjust control parameters when switching operating modes to avoid sudden changes in voltage or current. The adaptive multi-scene switching module also links with the cloud database to upload the running data and effect evaluation results of each scene switch, which is used to update and optimize the scene feature library and switching strategy.

Images