Intelligent terminal for commutator switches based on three-phase self-balancing and low-voltage fault location and isolation

Abstract

This invention discloses an intelligent terminal for phase-switching switches based on three-phase self-balancing and low-voltage fault location and isolation, belonging to the field of intelligent control technology for phase-switching switches. Through research on residual current change curves and characteristic components, distribution characteristics, and three-phase self-balancing techniques before and after a low-voltage distribution area fault, this invention proposes a residual current protection strategy using the transient component of the residual current as the logic action condition to achieve fault location and isolation in low-voltage distribution areas of the distribution network. It automatically acquires line impedance to deduce three-phase imbalance in the distribution area and combines it with phase-to-phase voltage difference to achieve three-phase self-balancing. Furthermore, through research on low-power wireless technology, it enables the uploading of measurement and fault information, providing data support for the analysis and processing of low-voltage faults and improving the automation level of the distribution network.

Description

Technical Field

[0001] This invention pertains to intelligent control terminals for phase switching, and particularly relates to an intelligent terminal for phase switching based on three-phase self-balancing and low-voltage fault location and isolation. Background Technology

[0002] With rapid social and economic development, my country's power distribution network has become increasingly large and complex. Due to historical reasons, my country has long maintained a management model that emphasizes "power generation and transmission" while neglecting "power distribution." Furthermore, the construction of the power distribution network has been influenced by national policies, overall planning, and environmental protection requirements, resulting in a low level of automation and insufficient power supply capacity to keep pace with social and economic development. At the same time, due to the numerous and widespread locations of power distribution networks, their complex line structures, and significant losses (accounting for over 50% of total grid losses), there is considerable potential for energy conservation.

[0003] In May 2017, the State Grid Corporation of China issued the "Notice on Carrying Out the Governance of Three-Phase Load Imbalance in Distribution Transmission Areas." The notice required that three-phase imbalance be addressed according to the principles of "source prevention, routine monitoring, scientific implementation, and dynamic governance." Three-phase imbalance is a crucial indicator of power quality, and imbalances are mostly caused by asymmetry in three-phase components, line parameters, or loads. In daily operation, unreasonable distribution of three-phase loads, constant changes in single-phase loads, and weakened monitoring of distribution transformer loads, as well as single-phase open-circuit faults and single-phase grounding faults in the distribution network, can all cause three-phase imbalance. In the future, with the rapid development of distributed residential photovoltaic systems, the imbalance of three-phase voltage and current at power supply points will be exacerbated, leading to increased losses in lines and distribution transformers. Furthermore, distribution transformers operating under three-phase load imbalance conditions will generate zero-sequence current, causing localized overheating of the transformer and accelerated aging of winding insulation due to overheating, resulting in reduced equipment lifespan. Moreover, it will also adversely affect motors at power supply points, seriously jeopardizing their normal operation. Currently, the rural low-voltage power grid has undergone significant changes in structure after renovation. Weak links in the grid structure have been largely resolved, the power supply capacity has been greatly enhanced, and voltage quality has significantly improved. The low-voltage line loss rate in most distribution substations has dropped below 11%. However, some distribution substations still suffer from persistently high line loss rates due to three-phase imbalance and other reasons, causing considerable difficulties for power supply companies in management and resulting in significant economic losses. In summary, unbalanced three-phase load operation will seriously endanger the safe and economical operation of electrical equipment. This is why the State Grid Corporation of China has continuously issued documents requiring the resolution of the three-phase load imbalance problem.

[0004] Manual adjustment is time-consuming and requires power outages, affecting the reliability of the power grid. Therefore, automatic adjustment devices are needed. Currently, the three commonly used devices for adjusting three-phase imbalance include phase-switching switch-type three-phase imbalance regulators, capacitor-type three-phase imbalance regulators, and power electronic three-phase imbalance regulators. These devices overcome the shortcomings of traditional manual rewiring to adjust three-phase imbalance, requiring no dedicated personnel for maintenance and management, automatically switching phases, and not interrupting power supply to users. However, each has its own drawbacks. Capacitor-type and power electronic-type three-phase unbalanced regulating devices only balance the three-phase load or current at the low-voltage outlet of the distribution transformer. However, since the load cannot be changed, they cannot fundamentally solve the problem of actual load balancing. Switch-type three-phase unbalanced regulating devices consist of an intelligent switching terminal and several switching units. The problems are: first, the switching time of the switching units is generally over 20ms, causing voltage flicker and affecting power quality; second, existing devices mainly rely on the terminal for switching decisions, and communication interruptions will affect normal operation; third, with load growth and transformer area upgrades, the number of switching units may increase, requiring extensive construction and commissioning work and incurring high maintenance costs for existing equipment. Therefore, it is necessary to develop fast-acting switching units to shorten switching time, improve power quality, and introduce self-sensing deep learning theory to achieve self-organizing networking and immediate use of the switching units, avoiding unnecessary commissioning and maintenance.

[0005] Low-voltage power grid outages have become a significant factor affecting power supply reliability. Due to the complexity of current distribution networks and the geographically dispersed and numerous electricity users, timely resolution and response to faults are often delayed. When a user or branch low-voltage fault occurs, the main circuit breaker will activate, leading to an expanded outage area. Currently, troubleshooting of traditional low-voltage distribution lines relies primarily on manual inspection and maintenance, which is inefficient, unsafe, and limited in implementation. Furthermore, considering the need for manual resetting after tripping, maintenance personnel must be physically present, increasing the workload. Therefore, there is a need to develop low-voltage fault location and isolation devices with anti-interference, anti-maloperation, and automatic reclosing functions. In addition, currently, fault location devices and three-phase imbalance control devices are generally installed and operated independently, resulting in redundancy in some measurement units. Therefore, integrating three-phase imbalance control and low-voltage side fault location and isolation functions into a single device can effectively reduce device size and application costs.

[0006] In some areas, the problem of three-phase imbalance in distribution transformers is quite prominent. Due to the lack of effective comprehensive consideration of three-phase load balance when low-voltage loads were connected in the early stage, and the large number of small power users in some areas, overload and low voltage caused by three-phase imbalance are prone to occur. Especially during peak summer seasons, there are many three-phase imbalances caused by uneven load distribution of single-phase users and overcapacity consumption by small power users. In particular, three-phase imbalances with a load imbalance greater than 25% and a neutral current greater than 40% of the transformer's rated current can easily lead to high line losses in the distribution area, seriously affecting the normal operation of the distribution transformer area. Effective measures need to be taken to address this issue.

[0007] II. Overview of Domestic and International Research Levels

[0008] 2.1 Current Status of Research on Three-Phase Imbalance Problems

[0009] Currently, there are three main solutions both domestically and internationally for problems such as excessive active power loss, excessive neutral current, and excessive reactive power caused by the imbalance of three-phase loads in low-voltage distribution networks: manual phase switching, reactive power compensation devices, and automatic switching devices.

[0010] To address the three-phase imbalance in low-voltage distribution networks caused by uneven load distribution, manual phase switching involves statistically analyzing the load power of each branch and the power of each phase on the busbar to calculate the optimal load distribution phase sequence at the current moment. Then, the phases of the load branches requiring phase sequence changes are manually switched, reducing the number of branches connected to heavily loaded phases and increasing the number of branches connected to lightly loaded phases, thus bringing the three-phase power of the busbar as close to three-phase balance as possible. This method has two main drawbacks: excessively long power outage times due to manual phase switching; and the inability to dynamically adjust the distribution busbar.

[0011] While reactive power compensation devices are one of the technical means to solve three-phase imbalance in power distribution systems, their drawbacks are relatively obvious. Firstly, reactive power compensation targets capacitive and inductive loads distributed throughout the power system, controlling them based on "power factor" or "reactive power," and cannot directly solve the three-phase imbalance caused by uneven load distribution. Secondly, the use of capacitors can easily lead to parallel resonance with inductive loads, amplifying harmonic currents. Furthermore, whether the device's compensation control strategy can adapt to the electricity needs of different types of users, and whether the cost of using the device is less than the losses caused by the imbalance, all require careful discussion.

[0012] Automatic switching devices are a new type of device that utilizes modern power electronics and communication technologies to switch the current power supply phase of an electrical load, thereby ensuring a reasonable and symmetrical distribution of single-phase electrical load among the three phases. However, the device's operation depends on control commands issued by the distribution station terminal; if the communication terminal is not reached, the device will be unable to operate.

[0013] The idea of ​​using power semiconductor converters to manage three-phase imbalance in power grids was proposed as early as the 1970s. In recent years, in-depth research has led to breakthroughs in both theory and specific devices. In March 2006, Tsinghua University and Shanghai Electric Power Company completed a 50MVA chain-type multilevel three-phase imbalance management device, which was successfully put into operation at the Huangdu West Substation in Shanghai. In 2016, Yangzhou Power Supply Company conducted research on three-phase load balancing technology for low-voltage distribution areas. By rationally dividing load areas, the load of each distribution transformer in each area is controlled in zones. Each distribution area is equipped with a load distribution controller and multiple commutation controllers. The commutation strategy adopts a centralized control method, issuing switching commands, and the commutation switches realize the switching of power phases. This equipment does not have functions such as fault location and isolation, time-sharing commutation control, autonomous control, and reclosing.

[0014] In my country, domestic experts and scholars have proposed various theoretical studies to address the three-phase load imbalance problem. Yan Hong of the China Jiliang University proposed a flexible switching system design based on a single-phase inverter power supply. This system not only achieves smooth transition during commutation but also solves the problem of short-term voltage interruptions caused by the use of switches and phase differences between two phases. However, the system is overly complex, employing an active power supply method from the inverter circuit, making the control strategy difficult and unable to guarantee the optimal load distribution scheme. Furthermore, it is expensive. Huazhong University of Science and Technology considered the three-phase load imbalance problem in the distribution network, providing a model of the three-phase load imbalance and using the Particle Swarm Optimization (PSO) algorithm for solution. However, the PSO algorithm suffers from premature convergence and other problems. The State Grid Jilin Electric Power Company published "A Brief Discussion on the Three-Phase Load Imbalance Problem in Low-Voltage Distribution Networks," which only analyzed the causes and conventional solutions for three-phase load imbalance in distribution networks, without providing case studies or proposing practical solutions. Zhejiang University of Science and Technology has established a microgrid load optimization allocation model and used the PSO algorithm to solve the model. However, the model construction is still within the scope of optimization scheduling and has not yet considered three-phase load imbalance. State Grid Shanxi Jincheng Power Supply Company has discussed the adjustment of three-phase load imbalance in the distribution network, using a power supply station as an example to discuss the causes of imbalance and adjustment schemes. However, its strategy is not yet intelligent, and the data is from two days ago and cannot be used for real-time reference.

[0015] The State Grid Corporation's "Notice on Carrying Out the Governance of Three-Phase Load Imbalance Problems in Distribution Transmission Areas" (Operation and Maintenance No. 3

[2017] 68) clearly outlines three governance methods: capacitor-type three-phase load automatic regulation devices, power electronic type three-phase load automatic regulation devices, and phase-switching switch type three-phase load automatic regulation devices. The first two types of devices are installed on the distribution area side and can only solve the three-phase imbalance problem in the distribution area, not the three-phase imbalance problem caused by residential users. Currently available phase-switching switch type three-phase load automatic regulation devices have an operating time of 20-50ms for each phase-switching switch. During the phase-switching process, users' refrigerators, air conditioners, computers, and other equipment may restart, affecting equipment lifespan and causing customer complaints. With the integration of distributed power into distribution transmission areas, such as rooftop photovoltaic systems with high penetration rates, the three-phase imbalance problem in distribution areas has been exacerbated. Currently, the main communication methods for phase-change switches that rely on communication include wireless GPRS, fiber optic, and low-voltage line-borne power communication. Wireless GPRS communication is costly and its signal is limited by the operator's base station signal. Although fiber optic communication is stable, it requires additional network cables to be laid in the distribution transformer area, which has a long construction period and high cost. Low-voltage line-borne power communication can communicate through existing power lines, but it is easily affected by factors such as harmonics in the distribution transformer area. Moreover, in Taizhou, most residents currently use power line carrier for remote meter reading. Phase-change switches using low-voltage carrier communication may interfere with residents' centralized meter reading, affecting the operation of the phase-change switches and the success rate of meter reading in the distribution transformer area, which will affect the performance evaluation of power personnel.

[0016] 2.2 Current status of research on fault location.

[0017] Regarding fault location, research on fault location in low-voltage distribution systems has long been in its initial stages due to limitations in technology and research focus. Domestic and international power industry research on fault location methods for low-voltage distribution lines has only implemented monitoring functions for the main outgoing lines of transformer substations in online monitoring systems, without monitoring of branch lines. Furthermore, it cannot provide more accurate fault location and handling functions. Therefore, at present, it is difficult to locate fault segments in low-voltage terminal faults, especially when the main outgoing line switch protection trips. Significant manpower is still required for troubleshooting. The development of smart grids places higher demands on the automation level of user safety electricity use. Automatic fault detection and location in user low-voltage distribution networks (400V low-voltage grids) is a key technology. This technology is used to automatically identify and locate fault types through distribution transformer terminals (with minimal manual inspection) after short-circuit and other accident protection actions.

[0018] When a power system fails, restoring normal system operation as quickly as possible is crucial for minimizing power outage time. Simultaneously, fault diagnosis is a prerequisite for fault recovery; therefore, the faulty area should be isolated quickly and accurately. Consequently, efficient fault diagnosis methods have always been a research hotspot in power distribution systems.

[0019] The main problems faced in fault location in low-voltage distribution networks are: 1) Uneven distribution of line parameters in low-voltage distribution networks. Most low-voltage distribution networks are three-phase unbalanced systems, with branch lines often contained on the main feeders, and intermediate loads mostly contained on the main feeders and side root branches. The faults occurring in these networks are mostly single-phase ground faults, with relatively high fault impedance and low fault current levels. 2) The impact of future distributed generation (DG) integration. The integration capacity of DG directly affects its contribution to fault current. DG integration can provide reverse fault current to faults upstream of its integration point, resulting in different fault current levels affecting the grid's contribution to fault location downstream of its integration point.

[0020] Currently, fault location methods proposed domestically and internationally for low-voltage distribution networks can be divided into two categories: methods based on distribution automation information and methods based on terminal measurement information. Some literature suggests that adding directional current detection devices can solve the problem, but this method increases the cost of retrofitting. With the development of distribution networks, feeder terminal units (FTUs) have been applied, capable of collecting voltage and current information at measurement points. This has led to widespread attention from researchers both domestically and internationally for fault location methods based on terminal measurement information in recent years. Proposed fault location methods can be categorized into five types based on their principles: matrix method, impedance method, path search method, high-frequency component method, and feature matching method. Most of these methods utilize single-point measurement information for fault determination and location. However, because multiple single-point measurement information does not contain the interrelationships between multiple measurement points, they lack good adaptability, mainly manifested in susceptibility to disturbances in non-fault states (fluctuations in distributed power sources, switching of distributed power sources or loads, etc.). Furthermore, the algorithm design is complex, and the fault location speed is slow; currently, these methods are primarily in the offline experimental research stage.

[0021] Furthermore, some circuit breakers with communication capabilities are currently available on the market. These can be connected to a host computer via their accompanying communication modules to achieve remote signaling, remote measurement, remote adjustment, and remote control functions. They can not only query the current status of the circuit breaker (closed, open, or tripped), but also retrieve the most recent fault record, including the faulty phase, fault type, current in each phase at the time of the fault, protection setting values, faulty phase current, and breaking time. However, circuit breakers with communication capabilities are expensive and are mostly used in medium- and high-voltage distribution networks. Circuit breakers in low-voltage distribution networks generally do not have waveform recording and communication functions. Therefore, currently, communication with circuit breakers cannot be used for fault location in low-voltage distribution networks. Using methods and devices from medium-voltage distribution networks would result in excessively high costs for fault location and isolation in low-voltage distribution networks, hindering their widespread application.

[0022] 2.3 Current Status of Deep Learning Technology Research

[0023] Deep learning is a new field of machine learning research. Its motivation lies in building and simulating neural networks that analyze and learn like the human brain. It mimics the mechanisms of the human brain to interpret data such as images, sounds, and text. Common deep learning models include convolutional neural networks, autoencoders, and deep belief networks.

[0024] (1) Convolutional Neural Network: A convolutional neural network is a type of feedforward neural network that is mainly used to recognize two-dimensional graphics that are invariant to translation, scaling and other forms of distortion. With its special structure of local weight sharing, convolutional neural networks have unique advantages in speech recognition and image processing.

[0025] (2) Autoencoder model: Sparse coding algorithm is an unsupervised learning method used to find a set of "overcomplete" basis vectors to more efficiently represent sample data. The purpose of sparse coding algorithm is to find a set of basis vectors such that we can represent the input vector X as a linear combination of these basis vectors.

[0026] (3) Deep Belief Networks: Deep Belief Networks are generative models that, by training the weights between neurons, allow the entire neural network to generate training data with the highest probability. Restricted Boltzmann Machines (RBMs) are the core component of DBN models. RBMs have two layers: a visible layer and a hidden layer. The classic DBN network structure is a deep neural network consisting of several RBM layers and one BP layer.

[0027] Applications of deep learning technology in the power industry:

[0028] (1) Power Grid Operation Monitoring: Real-time monitoring of power grid operation using deep learning technology, including fault diagnosis and analysis, and scheduling analysis of power equipment that cannot operate normally. Zhao Xuesong et al. proposed a power grid fault diagnosis method based on deep learning algorithms, which learns power grid fault information and obtains corresponding diagnostic models. Wu Shangyang proposed a wind energy forecasting algorithm based on deep learning, which creates a scheduling wind energy forecasting mechanism by introducing a restricted Boltzmann mechanism based on deep learning.

[0029] (2) User behavior analysis and prediction: Collect electricity consumption information of electricity users, analyze user electricity consumption behavior, predict user electricity demand, electricity purchase methods and other information, and improve the level of service to users. Lu Jun et al. implemented a feature-adaptive user electricity consumption behavior analysis method based on feature optimization strategy, and completed the optimized user electricity consumption behavior analysis.

[0030] (3) Power load analysis and forecasting: Power load forecasting is the basis for formulating power generation plans and power system development plans. Accurate load forecasting is of great significance for the economic, safe and reliable operation of the power system. Yang Jiaju applied the parallel decision tree based on the MapReduce model to the mining of electricity consumption habits in smart meter data to analyze and forecast power load.

[0031] Currently, the main solution to the three-phase imbalance problem in distribution transformer areas is to calculate the three-phase current imbalance at the terminal installed on the distribution transformer side, analyze the lightly loaded and heavily loaded phases, and then switch the commutation switch of the heavily loaded phase to the lightly loaded phase. Three-phase self-balancing technology based on deep learning can analyze historical data and combine it with real-time data for load forecasting to form the most economical commutation strategy.

[0032] Currently, fault location and isolation technology for distribution transformer areas relies on setting residual current thresholds for protection equipment. Exceeding these thresholds triggers equipment tripping. However, since load changes in distribution transformer areas are dynamic, the accuracy of these threshold settings affects the stability of the area. By employing deep learning technology and monitoring branch circuit data, the thresholds are continuously optimized based on residual current characteristic components, ultimately forming an optimized fault location and isolation strategy.

[0033] Currently, the mainstream algorithms used in power calculations are power flow algorithms and Newton-Rapson algorithms. These algorithms require given network topology, transformer turns ratios, and node injected power. However, the integration of new energy sources has altered the original network topology of distribution transformer areas. Furthermore, the network structure can change due to faults and maintenance, and user loads can fluctuate constantly. All these factors affect the accuracy of the algorithms. Using traditional algorithms may require periodic software upgrades. On-site upgrades increase the workload of maintenance personnel, and since equipment is typically installed at heights such as power poles, there is a risk of electric shock or falls during the upgrade process. Remote upgrades rely on signal stability; when equipment is installed in rural areas with weak signals, upgrades may be impossible or interrupted, leading to system downtime.

[0034] Deep learning has garnered significant attention from academia and industry in recent years, achieving remarkable results in image processing and classification, natural language processing, and biomedicine. However, its research and application in the power industry are still in their infancy. While related research has been limited in recent years, some successful examples exist. The research focus is primarily on fault diagnosis of power equipment, output prediction and defect detection of new energy generators, and power grid maintenance and upgrades. For the power sector, which possesses vast amounts of high-dimensional data, particularly in the research of three-phase self-balancing and low-voltage fault location and isolation technologies for distribution networks with massive user data, the introduction of deep learning theory is of considerable significance. Summary of the Invention

[0035] The technical problem to be solved by the present invention is to provide a smart terminal for a phase-switching switch based on three-phase self-balancing and low-voltage fault location and isolation, which overcomes the shortcomings of the prior art. It adopts self-learning and adaptive technology, load forecasting, communication-independent self-decision-making technology, and reclosing technology to achieve the effects of reducing losses and saving energy, improving emergency repair efficiency, and reducing economic losses. It can effectively reduce the operating and maintenance costs of equipment assets, improve the safe, stable and economical operation of the power grid and the level of safe production, and improve the reliability of power supply and the level of power quality.

[0036] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0037] The intelligent terminal for phase switching based on three-phase self-balancing and low-voltage fault location and isolation includes an intelligent distribution transformer terminal, a distribution side, a user side, and a distribution transformer. The intelligent distribution transformer terminal is connected to the distribution side and the user side respectively, and the distribution transformer is connected to the user side through the distribution side.

[0038] The intelligent distribution transformer terminal includes a data acquisition APP module and an automatic phase-switching load adjustment module connected to it.

[0039] The data acquisition APP module is used to collect the low-voltage total and branch line currents, voltages, active power, reactive power, and power factor on the distribution side.

[0040] An automatic phase-switching load regulation module is used to adjust the current, voltage, phase, power, and topology of each user on the user side; the distribution side includes a low-voltage main switch and multiple branch line switches;

[0041] The user side includes multiple commutation switches for the operating mechanism responsible for performing load commutation;

[0042] Specifically, it includes the following steps:

[0043] Step 1: Based on the residual current change curves, characteristic components, distribution characteristics, and three-phase self-balancing of the low-voltage distribution area before and after the fault, a residual current protection strategy with the transient component of the residual current as the logic action condition is proposed to realize the fault location and isolation of the low-voltage distribution area of ​​the distribution network.

[0044] Step 2: Automatically obtain the line impedance to deduce the three-phase imbalance in the transformer area and combine it with the phase-to-phase voltage difference to achieve three-phase self-balancing.

[0045] As a further preferred embodiment of the intelligent terminal for phase-changing switches based on three-phase self-balancing and low-voltage fault location and isolation of the present invention, the phase-changing switches communicate with each other via a CGMESH network. The CGMESH network is an IPv6-based mesh multi-hop network with CGR as the root node, which facilitates and flexibly extends and expands the wireless network coverage through a multi-level relay multi-hop method. In the CGMESH wireless mesh network, wireless nodes can conduct end-to-end direct IPv6 communication with the remote backend, and wireless nodes at each level within the wireless network can conduct parallel bidirectional communication with each other.

[0046] As a further preferred embodiment of the intelligent terminal for the phase-switching switch based on three-phase self-balancing and low-voltage fault location and isolation of the present invention, the phase-switching switch adopts a phase-switching switch with a bidirectional thyristor parallel contactor structure.

[0047] As a further preferred embodiment of the intelligent terminal for phase switching based on three-phase self-balancing and low-voltage fault location and isolation of the present invention, in step 1, the low-voltage distribution area fault includes grounding fault, open circuit fault, and mixed grounding and open circuit fault.

[0048] As a further preferred embodiment of the intelligent terminal for phase switching based on three-phase self-balancing and low-voltage fault location and isolation of the present invention, step 1 is specifically as follows:

[0049] Step 1.1: Investigate and statistically analyze different fault types on the distribution network side, classify the faults, and study the electrical quantity change characteristics under different fault types to provide a basis for accurate fault location.

[0050] Step 1.2: Investigate the three-phase imbalance problem and study the impact of single-phase access of distributed power sources on the three-phase voltage imbalance of distribution transformer areas; export historical data of distribution transformer areas for the past year from marketing, electricity consumption information collection and other systems; statistically analyze the changing trend of three-phase imbalance degree in transformer areas with large losses over the past year; and conduct multi-dimensional analysis of the three-phase imbalance problem in combination with factors such as transformer area load characteristics, climate, and season, to provide a basis for three-phase imbalance management strategies.

[0051] Meanwhile, the number of distribution substations containing distributed photovoltaic power was investigated. Through statistical analysis of the power load, number of electricity users, power consumption and power consumption trends of such substations, typical distribution substations were selected. Through multi-dimensional simulation analysis of distributed photovoltaic power generation, reactive power characteristics of distribution substations and power quality, the relationship between various factors was obtained.

[0052] Step 1.3: Statistically investigate the load characteristics of the transformer area, analyze the load consumption patterns based on historical data, identify the sensitive loads of the transformer area, and propose the selection and layout principles for the phase-switching switches; obtain relevant information such as historical load data of the transformer area through the system, analyze the electricity consumption, user characteristics, electricity consumption growth, etc., and determine the load characteristics of the transformer area.

[0053] Based on the classification of sensitive loads and the analysis of relevant data on the nature of the power load in the distribution area, it is possible to more accurately determine whether the distribution area contains sensitive loads, thereby selecting suitable distribution areas for addressing three-phase imbalance issues and proposing basic principles for the selection and layout of phase-switching switches.

[0054] As a further preferred embodiment of the intelligent terminal for phase-switching switches based on three-phase self-balancing and low-voltage fault location and isolation of the present invention, in step 1, fault location and isolation of low-voltage distribution areas in the distribution network are realized, as follows:

[0055] The anti-maloperation criterion strategy is optimized based on self-sensing learning technology. By analyzing the transient component of the residual current of the branch line, the pure power transfer quantity is used as the anti-maloperation criterion. The distribution network database is then input into the rough set computing system. Without providing any empirical knowledge, the rough set algorithm is used to directly analyze and reason about the data, discover the hidden power grid fault patterns, continuously optimize the criterion results, and improve the accuracy of the anti-maloperation criterion.

[0056] By using deep learning technology to analyze historical power grid data, fault diagnosis and analysis are performed on power equipment that cannot operate normally based on the operational data. This leads to the derivation of a power grid fault diagnosis method, forming a feature-adaptive distribution network fault analysis method, thereby accurately isolating faults.

[0057] As a further preferred embodiment of the intelligent terminal for phase switching based on three-phase self-balancing and low-voltage fault location and isolation of the present invention, step 2 is as follows:

[0058] Step 2.1: Establish the load characteristics and new energy access model of the transformer area, and use an improved quantum genetic algorithm with a double-chain structure to achieve three-phase self-balancing of the transformer area.

[0059] Establish a three-phase load imbalance optimization model and a new energy access model for the distribution area: represent the phase sequence of all users using a matrix, and consider the load type, commutation cost, and the impact of new energy access; adopt the daily load model of the distribution network, and establish a mathematical model of the three-phase imbalance degree with the goal of minimizing the economic cost under the condition of the lowest three-phase load imbalance.

[0060] The analysis of three-phase load imbalance in the distribution network involves analyzing user loads over a period of time to obtain the electrical load, average phase voltage, and power factor, and deriving a method for calculating the average phase current. By calculating the average three-phase load, a three-phase load function for the distribution network is established.

[0061] The improved quantum genetic algorithm adopts a double-chain structure to overcome the randomness of the encoding and the frequent decoding problem in the optimization process of the basic quantum genetic algorithm. It also adopts a dynamic adjustment strategy for the rotation angle and an adaptive adjustment algorithm search angle.

[0062] Step 2.2, Autonomous optimization algorithm for commutation criterion threshold based on historical switch operation data:

[0063] Based on historical voltage and current data collected by the phase-switching switch, a large amount of historical data is preprocessed to remove invalid and missing values; feature values ​​of the data are extracted by analyzing data fluctuation trends, electricity consumption dispersion coefficients, and variability.

[0064] A prediction commutation threshold function is constructed using bagging, and these functions are combined into a single prediction function. Given a weak learning algorithm and a training set, the learning algorithm is then used multiple times to obtain a sequence of prediction functions, which are then voted on. Finally, the accuracy of the result is improved, and the optimal threshold for the commutation criterion is obtained.

[0065] Step 2.3: Based on a large amount of historical data and the continuous updating of existing data, a low-voltage fault prediction algorithm is studied. The prediction results are continuously optimized and the prediction accuracy is improved by using the rough set method of fuzzy theory.

[0066] Fuzzy theory is used to identify faults using rough sets, abandoning the absolute membership relationship of either 0 or 1 in conventional fault location, forming a system where the membership value of the fault result can be selected within the interval [0,1]. The uncertainty of each fault selection result is represented by a membership function.

[0067] Based on the summary of fault types and the corresponding electrical quantity change characteristics, potential patterns are discovered through the analysis and reasoning of real-time data to achieve fault prediction. At the same time, the rough set method uses a large amount of historical data and existing data to continuously refresh and optimize the prediction results, thereby improving the prediction accuracy.

[0068] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0069] This invention, through research on residual current variation curves and characteristic components, distribution characteristics, and three-phase self-balancing techniques before and after low-voltage distribution area faults, proposes a residual current protection strategy that uses the transient component of residual current as the logic action condition to achieve fault location and isolation in low-voltage distribution areas of the distribution network; it automatically acquires line impedance to deduce three-phase imbalance in the distribution area and combines it with phase-to-phase voltage difference to achieve three-phase self-balancing; and through research on low-power wireless technology, it realizes the uploading of measurement information and fault information; providing data support for the analysis and processing of low-voltage faults and improving the automation level of the distribution network.

[0070] This invention transforms the original centralized control mode into a centralized + decentralized control mode. When communication is abnormal, the commutator can realize autonomous detection, autonomous acquisition, autonomous judgment, and autonomous action. By collecting the voltage and current information of the installation point, the optimal commutation strategy can be obtained. The coordination strategy is continuously optimized through the rough set method and the deep self-learning algorithm based on the historical data of operation to realize the setting of the difference between the set value and the time, thus solving the problem of malfunction caused by communication instability.

[0071] This invention proposes a wireless multi-hop self-organizing network based on CGMESH, enabling the phase-change switch to be installed and used immediately. The phase-change switch communicates automatically with the terminal through the CGMESH self-organizing network. When expanding the switch in the later stage, it can be installed and used immediately without debugging. The phase sequence is identified by matching the switch sampling with the terminal sampling, which solves the problem of difficult phase sequence identification of power cables in the field.

[0072] 4. This invention proposes a fault location algorithm for low-voltage distribution networks based on the transient characteristic components of residual current. Based on the voltage and current values ​​of each phase-switching switch and terminal acquisition installation point, the equivalent impedance of the entire distribution network can be derived. Based on the transient characteristic components of residual current, the distance between the fault point and the terminal and switch is calculated to locate the fault area, thus solving the problem of single-phase short-circuit fault location and isolation in low-voltage distribution networks. Attached Figure Description

[0073] Figure 1 This is a system architecture diagram of the intelligent terminal for phase-commutation switches based on three-phase self-balancing and low-voltage fault location and isolation of the present invention.

[0074] Figure 2 This is a functional block diagram of the intelligent terminal system for phase switching based on three-phase self-balancing and low-voltage fault location and isolation of the present invention;

[0075] Figure 3 This is the wiring diagram of the commutation switch type of the present invention;

[0076] Figure 4 Schematic wiring diagram of the capacitor-type three-phase load automatic regulating device of the present invention;

[0077] Figure 5 Wiring diagram of the principle of automatic three-phase load regulation by power electronic device of this invention;

[0078] Figure 6 This invention provides a schematic diagram of a distributed intelligent architecture for managing three-phase imbalance in a distribution transformer area in Jinhua.

[0079] Figure 7 A schematic diagram illustrating the installation and adjustment effect of the three-phase imbalance device in Ouhai District, Wenzhou, according to the present invention;

[0080] Figure 8 A schematic diagram illustrating the application of the intelligent three-phase imbalance adjustment device in Heze City, Shandong Province, according to this invention.

[0081] Figure 9 This invention relates to a commutator switch structure diagram of a commutator switch intelligent terminal based on three-phase self-balancing and low-voltage fault location and isolation. Detailed Implementation

[0082] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings:

[0083] A further detailed description of the invention, taken in conjunction with the accompanying drawings, will focus on at least one preferred embodiment. This description should be detailed enough to enable those skilled in the art to reproduce the invention or utility model without requiring further creative effort, such as exploration, research, and experimentation. The technical solution of the invention will now be described in further detail with reference to the accompanying drawings:

[0084] like Figure 1 and Figure 2 As shown, the intelligent phase-switching switch based on three-phase self-balancing and low-voltage fault location and isolation includes an intelligent distribution transformer terminal, a distribution side, a user side, and a distribution transformer. The intelligent distribution transformer terminal is connected to the distribution side and the user side respectively, and the distribution transformer is connected to the user side through the distribution side.

[0085] The intelligent distribution transformer terminal includes a data acquisition APP module and an automatic phase-switching load adjustment module connected to it.

[0086] The data acquisition APP module is used to collect the low-voltage total and branch line currents, voltages, active power, reactive power, and power factor on the distribution side.

[0087] An automatic phase-switching load regulation module is used to adjust the current, voltage, phase, power, and topology of each user on the user side; the distribution side includes a low-voltage main switch and multiple branch line switches;

[0088] The user side includes multiple commutation switches for the operating mechanism responsible for performing load commutation;

[0089] Specifically, it includes the following steps:

[0090] Step 1: Based on the residual current change curves, characteristic components, distribution characteristics, and three-phase self-balancing of the low-voltage distribution area before and after the fault, a residual current protection strategy with the transient component of the residual current as the logic action condition is proposed to realize the fault location and isolation of the low-voltage distribution area of ​​the distribution network.

[0091] Step 2: Automatically acquire line impedance to deduce three-phase imbalance in the distribution area and combine it with phase-to-phase voltage difference to achieve three-phase self-balancing. Through research on low-power wireless technology, realize the uploading of measurement information and fault information; provide data support for the analysis and handling of low-voltage faults and improve the automation level of the distribution network.

[0092] To achieve three-phase self-balancing and low-voltage fault location and isolation technology in distribution transformer areas, this invention uses a phase-switching switch as the main body and an improved quantum genetic algorithm with a double-chain structure to study the three-phase self-balancing technology of the distribution transformer area. The phase-switching switch adopts a two-phase thyristor structure to shorten the switching time. The rough set technology based on fuzzy theory is used to locate faults in branch lines. Communication between phase-switching switches and distribution transformer terminals is realized based on CGMESH network communication. At the same time, the device has a reclosing function.

[0093] Theoretical basis for the invention research content:

[0094] (1) Existing three-phase imbalance control technologies;

[0095] a) Three-phase load automatic regulating device with phase-switching switch type;

[0096] Technical principle:

[0097] The phase-switching switch type three-phase load automatic regulation device (low-voltage load automatic phase-switching device) consists of an intelligent phase-switching terminal (responsible for load monitoring and automatic phase-switching control) and several phase-switching switch units (operating mechanisms responsible for executing load phase-switching). The intelligent phase-switching terminal monitors the three-phase current of the low-voltage outgoing lines of the distribution transformer in real time. If the three-phase load imbalance on the low-voltage side of the distribution transformer exceeds the limit within a certain monitoring period, the intelligent phase-switching terminal reads the real-time data of current and phase sequence of the low-voltage outgoing lines of the distribution transformer and each load branch of all phase-switching switch units, performs optimization calculations, and issues the optimal phase-switching control command. Each phase-switching switch unit executes the phase-switching operation according to the prescribed phase flow, realizing the balanced distribution of three-phase load in the user's load area.

[0098] Applicable conditions:

[0099] The three-phase load imbalance in the distribution transformer area is caused by random variations in specific loads and is difficult to manage through routine operation and maintenance. The power factor on the low-voltage side of the distribution transformer is greater than 0.85. The low-voltage main line and major branch lines in the distribution transformer area are three-phase powered. There are no sensitive loads with high reliability requirements within the power supply range of the phase-switching switch. The wiring diagram of the phase-switching switch is as follows. Figure 3 As shown.

[0100] The intelligent commutation terminal communicates with all commutation switch units in real time via power line carrier. Based on the real-time data of current and phase sequence of the commutation switch units, the intelligent commutation terminal performs optimization calculations and issues the optimal commutation control command. The commutation switch units receive the command and execute the commutation operation to achieve balanced distribution of three-phase load in the user's negative distribution area.

[0101] b) Capacitor-type three-phase load automatic regulating device;

[0102] Technical principle:

[0103] The capacitor-type three-phase load automatic regulation device (phase-to-phase reactive power compensation device) connects a power capacitor between the phase lines to realize the transfer of active power and balance the active power between the phases. At the same time, the power capacitor connected between the phase line and the neutral line provides unequal reactive power compensation for each phase, balances the reactive power between the phases, reduces the three-phase imbalance, and improves the power factor.

[0104] Applicable conditions:

[0105] The three-phase load imbalance problem in the distribution transformer area is caused by random changes in specific loads and is difficult to manage through routine operation and maintenance. The distribution transformer area also suffers from both three-phase load imbalance and insufficient reactive power. The power supply radius of the distribution transformer area is relatively short. The wiring diagram of the capacitor-type three-phase load automatic regulation device is shown below. Figure 4 As shown.

[0106] Capacitor-type three-phase load automatic regulation devices, also known as phase-to-phase compensation type three-phase imbalance regulation devices, transfer active power by bridging power capacitors between phase lines, thereby balancing the active power between phases and reducing three-phase imbalance. However, this device only balances the three-phase load at the low-voltage outlet of the distribution transformer and cannot fundamentally solve the problem of balanced load distribution in practice.

[0107] c) Power electronic type three-phase load automatic regulation device;

[0108] Technical principle:

[0109] The power electronic three-phase load automatic regulation device (low-voltage static var compensator SVG, active power filter APF) is a comprehensive power quality management device that adopts high-power turn-off power electronic switching technology. It rapidly detects reactive, negative-sequence, and harmonic currents at the connection point, and generates trigger pulse signals based on the space vector pulse width modulation (SVPWM) control method to drive the thyristor output to produce a compensation current equal in magnitude but opposite in direction to the detected reactive, negative-sequence, and harmonic currents. This comprehensively solves problems such as reactive power, harmonics, voltage fluctuations, and three-phase load imbalance in the distribution area.

[0110] Applicable conditions:

[0111] The three-phase load imbalance problem in the distribution substation is caused by random variations in specific loads and is difficult to manage through routine operation and maintenance. Users have high requirements for power quality or simultaneously experience three-phase load imbalance, insufficient reactive power, and harmonic exceedance issues. The distribution substation has a short power supply radius. The wiring diagram for the power electronic three-phase load automatic regulation principle is shown below. Figure 5 As shown.

[0112] Load power imbalance causes three-phase current imbalance. By connecting a three-phase imbalance module regulator in parallel on the line, the electrical energy that needs to be transferred is transferred from the line to the regulator. Through DSP (Digital Signal Processing) to quickly calculate the distribution method, the transferred electrical energy is superimposed on different phases, forcing the three-phase current balance at the front end of the regulator. However, since the load cannot be changed, the current at the back end of the regulator is still unbalanced.

[0113] (2) Wireless Mesh network communication technology;

[0114] A wireless mesh network consists of mesh routers and mesh clients. The mesh routers form the backbone network and connect to the wired internet, providing multi-hop wireless internet connections for the mesh clients. It is a new type of wireless network technology that is completely different from traditional wireless networks.

[0115] In a wireless mesh network, any wireless device node can act as both an access point (AP) and a router. Each node in the network can send and receive signals, and each node can communicate directly with one or more peer nodes.

[0116] Wireless mesh networks have several unparalleled advantages:

[0117] 1) Easy Deployment and Installation: Installing a Mesh node is very simple; just take the device out of the box and plug it in. Due to this greatly simplified installation, users can easily add new nodes to expand the coverage and capacity of the wireless network. In a wireless Mesh network, not every Mesh node requires a wired cable connection, which is its biggest difference from a wired access point (AP).

[0118] 2) Stability: A common method for achieving network stability is to use multiple routers for data transmission. If one router fails, information is transmitted via backup paths by other routers. In a mesh network structure, each node has one or more paths for transmitting data. If the nearest node fails or is interfered with, data packets will be automatically routed to backup paths for continued transmission, and the operation of the entire network will not be affected.

[0119] 3) Flexible Structure: In multi-hop networks, devices can connect to the network simultaneously through different nodes, thus not leading to a decrease in system performance. Mesh networks also provide greater redundancy mechanisms and communication load balancing capabilities. In wireless mesh networks, each device has multiple transmission paths available, and the network can dynamically allocate communication routes based on the communication load of each node, effectively avoiding node congestion. In contrast, single-hop networks cannot dynamically handle communication interference and access point overload issues.

[0120] 4) High Bandwidth: The physical characteristics of wireless communication dictate that higher bandwidth is more easily achieved with shorter transmission distances. This is because as the wireless transmission distance increases, various interferences and other factors leading to data loss also increase. Therefore, choosing to transmit data via multiple short hops is an effective way to obtain higher network bandwidth, which is precisely the advantage of Mesh networks.

[0121] In a mesh network, a node can not only transmit and receive information, but also act as a router to forward information to its nearby nodes. As more nodes are interconnected and the number of possible paths increases, the total bandwidth also increases significantly.

[0122] (3) The algorithm used in the three-phase balancing technique;

[0123] Genetic algorithms are a class of randomized search methods that draw inspiration from the evolutionary principles of biology (survival of the fittest, the genetic mechanism of natural selection). Their main characteristics include direct operation on structural objects, without the constraints of differentiation or function continuity; inherent implicit parallelism and better global optimization capabilities; and the use of probabilistic optimization methods to automatically acquire and guide the search space, adaptively adjusting the search direction without requiring predetermined rules. These properties of genetic algorithms have led to their widespread application in combinatorial optimization, machine learning, signal processing, adaptive control, and artificial life.

[0124] Genetic algorithms, also known as genetic algorithms, are search heuristics used in computer science and artificial intelligence to solve optimization problems. They are a type of evolutionary algorithm. This heuristic is typically used to generate useful solutions for optimization and search problems. Evolutionary algorithms were initially developed by drawing inspiration from phenomena in evolutionary biology, including heredity, mutation, natural selection, and crossbreeding. However, genetic algorithms may converge to local optima rather than the global optimum if the fitness function is poorly chosen.

[0125] Genetic algorithms also have the following characteristics:

[0126] (a) Genetic algorithms start their search from a set of solutions to the problem, rather than from a single solution. This is a significant difference between genetic algorithms and traditional optimization algorithms. Traditional optimization algorithms iteratively find the optimal solution from a single initial value; they are prone to straying into local optima. Genetic algorithms, on the other hand, start their search from a set of solutions, providing broader coverage and facilitating global selection.

[0127] (b) Genetic algorithms process multiple individuals in the population at the same time, that is, they evaluate multiple solutions in the search space, which reduces the risk of getting trapped in local optima. At the same time, the algorithm itself is easy to parallelize.

[0128] (c) Genetic algorithms essentially do not require knowledge of the search space or other auxiliary information; they rely solely on the fitness function value to evaluate individuals and perform genetic operations based on this. The fitness function is not only unconstrained by continuous differentiability, but its domain can also be arbitrarily defined. This characteristic greatly expands the application scope of genetic algorithms.

[0129] (d) Genetic algorithms do not use deterministic rules, but rather probabilistic transition rules to guide their search direction.

[0130] (e) It possesses self-organizing, adaptive, and self-learning properties. When genetic algorithms utilize information gained from the evolutionary process to organize their search, individuals with higher fitness have a higher probability of survival and acquire gene structures that are more adapted to the environment.

[0131] (f) In addition, the algorithm itself can also adopt dynamic adaptive technology to automatically adjust the algorithm control parameters and coding precision during the evolution process.

[0132] (4) Fault location methods;

[0133] 1) Impedance Method: The impedance method measures the impedance between the fault point and the measuring end after a fault occurs, and then uses line parameters to establish a fault location equation to solve for the fault distance. This method is mostly based on the lumped parameter model of the line. Because of its simple principle and ease of implementation, it has always received widespread attention.

[0134] The bridge method is a widely used impedance-based fault location method. The bridge method includes two types: the resistance bridge method and the capacitance bridge method.

[0135] The resistance bridge method is only suitable for low-impedance measurements. The resistance at the fault point must not exceed 100 kΩ, and the maximum should not exceed 500 kΩ, with 2 kΩ being generally preferred. Its basic principle is as follows: short-circuit the non-faulty phase at the end of the cable under test with the faulty phase. Connect the two output arms of the bridge to the faulty and non-faulty phases respectively. Adjust the adjusting resistors on the bridge arms. Once the bridge is balanced, calculate the fault distance based on the bridge balance principle.

[0136] The capacitance bridge method can be used to measure cable breakage faults, and its principle is similar to that of the resistance bridge method.

[0137] The advantages of the bridge method are its convenience and simplicity. However, its disadvantages include unsuitability for measuring high-resistance and flashover faults. This is because, in cases of high fault resistance, the current in the bridge circuit is very small, and the sensitivity of a typical galvanometer is very low, making it difficult to detect whether the bridge is balanced. Another disadvantage of the bridge method is that it requires knowledge of the exact length of the cable and other original materials. When a cable line consists of two or more conductor materials or two or more cable sections with different cross-sectional areas, conversion is necessary. Furthermore, the bridge method cannot measure three-phase short-circuit faults.

[0138] This impedance method is based on a lumped parameter model of the line, and it becomes ineffective when the fault resistance is high. Therefore, some literature proposes an impedance method suitable for measuring high-resistance faults. This method is based on distributed parameters, establishes parametric equations, and calculates the fault distance. Its basic principle is: applying a sinusoidal high-voltage signal to the cable with the high-resistance fault causes flashover at the high-resistance fault point, at which point the high-resistance fault becomes an arc fault. Because arc faults are resistive, the current flowing through the fault point and the voltage across the fault point are in phase. By acquiring the current flowing through the line and the voltage across the line through a data acquisition system, the voltage and current at each point along the line can be calculated using distributed parameter theory, thus achieving fault location.

[0139] 2) Traveling wave: The traveling wave method is also a commonly used method in power cable fault location. The traveling wave method has the advantages of high speed and high accuracy, and specifically includes: type A, B, C and D traveling wave ranging methods.

[0140] (a) Type A Traveling Wave Distance Measurement: After a fault occurs at the fault point, the resulting traveling wave propagates back and forth between the fault point and the measuring end. The distance to the fault point can be determined by the time it takes for the traveling wave to travel to and from the fault point to the measuring end once and the traveling wave velocity. This method is not affected by transition resistance and the impedance of the load at the opposite end. The principle is simple, requires fewer devices, and can theoretically achieve high accuracy. For many years, due to a lack of in-depth understanding of the characteristics and propagation properties of the traveling wave generated at the fault point, coupled with the need for high-speed sampling systems and precise timing systems, it has not been widely used. In recent years, the more commonly used Type A traveling wave distance measurement methods are mainly the pulse current method and the pulse voltage method.

[0141] The pulse current method, developed in the early 1980s, is a testing method with advantages such as safety, reliability, and simple wiring. Its principle is to measure the pulse fault current signal generated when a cable fault breaks down using a linear current coupler. It achieves electrical coupling between the instrument and the high-voltage circuit, eliminating the need for capacitors and inductors connected in series between the capacitor and the cable, simplifying wiring, and making it easy to distinguish the pulse current signal coupled from the sensor. The pulse voltage method, also known as the impulse flashover method, works by first using a DC high voltage and a pulsed high voltage signal to break down the cable fault, and then measuring the distance by the time it takes for the discharge voltage pulse to travel back and forth between the measuring end and the fault point. Its advantages include not needing to break down high-resistance and flashover faults, directly utilizing the instantaneous pulse signal generated by the fault breakdown, resulting in faster testing speed and a simpler measurement process. Its disadvantages are: when using this method for distance measurement, the high-voltage capacitor is in a short-circuit state to the pulse signal, and a resistor or inductor needs to be connected in series to generate a voltage signal, which increases the complexity of wiring and reduces the voltage applied to the faulty cable when the capacitor discharges, making it difficult to break down the fault point; in addition, during the measurement process, the voltage waveform coupled by the voltage divider does not change significantly and is difficult to distinguish.

[0142] (b) Type B Traveling Wave Ranging: The basic principle of Type B traveling wave ranging is to use only the information from the first traveling wave front signal generated by the traveling wave at the fault point reaching both ends of the cable, and to perform fault location with the help of a communication channel. The advantage of this method is that it only uses the first wave front, so the reflected and transmitted waves at the fault point do not limit its application. However, this method still requires accurate time for the traveling wave to reach the measuring end.

[0143] (c) C-type traveling wave ranging: The basic principle of C-type traveling wave ranging is: when a cable fault occurs, the cable is disconnected from the power grid, a high-frequency pulse is emitted to the faulty cable through a pulse transmitting device, and then the round-trip time of the pulse signal between the device and the fault point is calculated, thereby calculating the fault distance.

[0144] Currently, the commonly used C-type traveling wave ranging methods mainly include the low-pressure pulse reflection method and the double pulse method.

[0145] The low-voltage pulse reflection method, also known as the radar method, is mainly used for locating low-resistance (or short-circuit) and open-circuit faults in cables. This method is relatively simple and intuitive, determining the distance by observing the time difference between the reflected pulse and the transmitted pulse at the fault point. Different fault types produce different reflected waves. If the transmitted pulse is positive and the echo pulse is also positive, it indicates an open-circuit fault or an open-ended termination, meaning the voltage cannot be fed to the other end. If the echo pulse is negative, it indicates a short-circuit to ground fault, meaning the insulation resistance at the cable fault point is lower than the cable's characteristic impedance, or even zero. The advantages of the low-voltage pulse reflection method are its simplicity, intuitiveness, lack of need for detailed original cable data, and ability to distinguish fault types based on the polarity of the reflected pulse. The disadvantage is that it cannot be used to measure high-resistance and flashover faults.

[0146] 3) Acoustic Measurement Method: The basic principle of the acoustic measurement method is to locate the fault point by using the sound signal generated when the power cable discharges. A sound sensor is placed above the power cable to detect the sound signal, and the location of the loudest sound is the location of the fault point. For faults where the cable sheath has been burned through, the discharge sound can often be heard directly from the ground. However, for cable faults where the sheath has not been burned through or for deeply buried cables, the discharge sound may be reduced. A highly sensitive sound-to-electric converter, such as a microphone or piezoelectric crystal, is needed to convert the weak seismic waves on the ground into an electrical signal. This signal is then amplified by relevant instruments, and the sound is reproduced through headphones or displayed on a screen to determine the location of the loudest discharge sound. This method is mainly used for locating high-resistivity faults in power cables.

[0147] 4) Audio current induction method:

[0148] If a cable experiences a low-resistance fault, such as a fault resistance less than 10Ω, it is difficult to detect the discharge sound at the fault point using acoustic detection methods, or there may be no discharge sound at all. Therefore, acoustic detection methods cannot be used for fault location in low-resistance faults. In this case, the audio current induction method can be used to determine the location of the fault point by detecting changes in the magnetic field on the ground. The basic principle is to use an audio signal generator of 1-15 kHz to pass an audio signal current through the cable under test. Due to electromagnetic coupling, an induced current is generated in the ground, thus forming a ground magnetic field. Then, a probe is used on the ground to receive the audio signal along the cable laying path, and the received audio signal is sent to a receiver for amplification and then sent to headphones. The location of the fault point can be detected by the change in sound in the headphones. The audio signal is strongest above the fault point, and as the probe continues to move forward, the audio signal gradually decreases. The point where the audio signal is strongest is the fault point.

[0149] The audio induction method is generally used to detect low-resistance faults with a fault resistance of less than 10Ω. Satisfactory results can be obtained when using the audio induction method to measure two-phase (three-phase) short circuits or two-phase (three-phase) short circuits with grounding faults. The absolute error of the measured fault location is generally 1-2m.

[0150] (5) Rough Set Theory: Rough set theory, as a data analysis and processing theory, was founded in 1982 by Polish scientist Z. Pawlak. As a scientific research area for intelligent computing, rough set theory has made great progress in both theoretical and practical applications. Rough set theory not only provides new scientific logic and research methods for information science and cognitive science, but also provides effective processing techniques for intelligent information processing.

[0151] As long as the database is input into the rough set operation system, without providing any prior knowledge, the rough set algorithm can automatically learn the knowledge, which is the root of its widespread application. In contrast, in set theory such as fuzzy sets and extension sets, we need to predefine the membership function.

[0152] Rough set theory, as an effective tool for handling various incomplete information such as imprecise, inconsistent, and incomplete information, benefits from its mature mathematical foundation and the fact that it does not require prior knowledge. Furthermore, it is easy to use. Since the purpose and starting point of rough set theory is to directly analyze and reason about data to discover implicit knowledge and reveal potential patterns, it is a natural data mining or knowledge discovery method. Compared with other methods for handling uncertainty, such as data mining based on probability theory, fuzzy theory, and evidence theory, its most significant difference is that it does not require any prior knowledge beyond the data set needed to address the problem, and it is highly complementary to other theories for handling uncertainty (especially fuzzy theory).

[0153] On the other hand, rough set theory can be combined with other intelligent algorithms, such as: combining with neural networks to perform data preprocessing and improve the convergence speed of neural networks; combining with support vector machines (SVM); combining with genetic algorithms; and especially combining with fuzzy theory, which has yielded many fruitful results. Although both rough set theory and fuzzy theory describe the uncertainty of sets, fuzzy theory focuses on describing the uncertainty of elements within a set, while rough set theory focuses on describing the uncertainty between sets. The two are not contradictory and are highly complementary.

[0154] The comprehensive management of three-phase imbalance in distribution transformer areas focuses on harmonic control and reactive power compensation regulation at the low-voltage side of the distribution transformer area, as well as the switching of three-phase loads at the end. Among these, harmonics, voltage fluctuations, flicker, and three-phase imbalance are closely related to the power load characteristics of users. These three indicators are difficult to monitor in real time and are generally measured periodically by the testing department.

[0155] 1) Three-phase voltage imbalance:

[0156] Implementation Standard: GB / T 15543-2008 "Power Quality - Permissible Unbalance of Three-Phase Voltage"

[0157] According to relevant implementation standards, the permissible values ​​for three-phase voltage unbalance, as well as their calculation, measurement, and determination methods, are specified. These standards apply to voltage unbalance at the point of common coupling caused by negative sequence components under normal power system operation. The national standard stipulates that the normal permissible voltage unbalance at the point of common coupling of a power system is 2%, and it should not exceed 4% for short periods. Generally, it should not exceed 1.3% for each user.

[0158] 2) Excessive Three-Phase Unbalance: The regulations stipulate that the load imbalance at the outlet of a distribution transformer should not be less than 10%, the overall distribution current imbalance should not exceed 15%, the neutral current should not exceed 25% of the rated current on the low-voltage side, and the current imbalance at the beginning of the low-voltage main line and major branches should be less than 20%. Based on current power grid upgrades, to achieve a line loss rate below 12%, the above indicators can only be tightened, not relaxed.

[0159] 3) Harmonic current limits:

[0160] Implementation standard: GB / T 14549-1993 "Power Quality - Harmonics in Public Power Grids";

[0161] Harmonic currents injected by harmonic users into the point of common coupling of the power system shall meet the requirements of the national standard GB / T 14549-1993 "Power Quality - Harmonics in Public Power Grids" and shall be calculated strictly in accordance with the methods specified in the national standard, based on the minimum short-circuit capacity of the point of common coupling, the capacity of the power supply equipment, and the capacity of the power consumption agreement.

[0162] 4) Distribution network planning and design technology:

[0163] Implementation Standard: DL / T 5729-2016 Technical Guidelines for Distribution Network Planning and Design

[0164] This standard applies to the planning and design of distribution networks at voltage levels of 110 (66) kV, 35 kV and below. It specifies requirements for power supply areas, planning basis, load forecasting and power balance, main technical principles, grid structure, equipment selection, basic requirements for intelligent systems, and user and power supply access requirements. It also proposes relevant requirements for distribution network planning calculation and technical-economic analysis. The goal is to build a distribution network with necessary capacity margin, appropriate load transfer capacity, certain self-healing and emergency handling capabilities, and reasonable distributed power supply acceptance capacity, thereby achieving safe, reliable, and economical power supply to users.

[0165] 5) Distribution network operation and maintenance procedures:

[0166] Implementation standard: Q / GDW 1519-2014 "Distribution Network Operation and Maintenance Regulations";

[0167] According to the "Distribution Network Operation and Maintenance Regulations" (Q / GDW 1519-2014), the load imbalance of distribution transformers should meet the following requirements: for transformers with Yyn0 connection, the load imbalance should not exceed 15%, and the neutral current should not exceed 25% of the transformer's rated current; for transformers with Dyn11 connection, the load imbalance should not exceed 25%, and the neutral current should not exceed 40% of the transformer's rated current.

[0168] (2) Current status of demonstration applications of existing equipment on the market:

[0169] Research and application of three-phase imbalance mitigation technology for distribution transformer substations in Jinhua distributed intelligent architecture;

[0170] In 2016, to address the three-phase imbalance problem in distribution networks in urban-rural fringe areas, a distributed intelligent architecture-based three-phase imbalance mitigation technology and solution for distribution transformer substations was developed, as shown in the figure. It consists of a back-end monitoring master station, an active power self-balancing load optimization control device (substation), load switching switches, and a communication system. A pilot application was conducted in the Liushi Xincun distribution transformer substation area of ​​Jinhua Power Company. The distributed intelligent architecture for three-phase imbalance mitigation in Jinhua distribution transformer substations is as follows: Figure 6 As shown.

[0171] The Liushi Xincun transformer substation has a capacity of 630KVA and serves approximately 100 households. Previously, the three-phase load imbalance generally exceeded 50%, leading to single-phase overload during peak electricity consumption periods. End-users complained of unstable voltage and malfunctioning appliances. In the pilot application, a transformer substation intelligent load balancing system, one active power self-balancing load optimization control device, and 20 load switching switches were installed.

[0172] Before application: At a certain time (22:30), the single-phase load current in this distribution area was 204A, 306A, and 588A, with a maximum single-phase load rate of 64.6% and a three-phase load imbalance of 65%. The transformer's remaining output was only 35.4% of its capacity. After application: The imbalance was controlled to 10%, the maximum single-phase load rate decreased to 44%, and the transformer's remaining capacity reached 56%, indirectly increasing the transformer's output by more than 20%. This significantly reduced the occurrence of single-phase overload faults, and line losses were reduced to approximately 3%. This solution uses a phase-switching switch-type three-phase load automatic adjustment device. Short-term power interruptions occur during the phase-switching process, requiring no highly sensitive loads with high reliability requirements within the power supply range. Furthermore, it may not function properly if the main station communication is interrupted.

[0173] Application of three-phase imbalance adjustment device in Ouhai District, Wenzhou.

[0174] The Panqiao No. 7 transformer substation of Louqiao Power Supply Station is a street-level residential transformer substation, mainly serving small workshops, shops, family workshops, and residents. The transformer capacity is 315kVA. A three-phase imbalance regulating device with a capacity of 100kvar is installed on the low-voltage side of the transformer. Before compensation, the maximum phase difference in three-phase current was over 200 amperes; after compensation, the maximum phase difference is generally kept within 10A. Because the installed three-phase imbalance device has a maximum regulating capacity of 150A, there are occasional times when the required adjustment current for all three phases exceeds 150A.

[0175] This scheme uses a capacitor-type three-phase load automatic regulation device, installed on the distribution transformer area side. The radius of the power supply area is relatively short, and it only achieves three-phase load balance at the low-voltage outlet of the distribution transformer, failing to fundamentally solve the problem of actual load balance distribution. For example... Figure 7 The image shows the installation and adjustment effect of the three-phase imbalance device in Ouhai District, Wenzhou.

[0176] Application of three-phase imbalance intelligent adjustment device in Heze City, Shandong Province:

[0177] In Juancheng County, Heze City, Shandong Province, 10 three-phase imbalance intelligent adjustment devices were installed in some administrative villages. This device combines passive reactive power compensation and active three-phase imbalance compensation. It uses the speed and practicality of active compensation to address imbalance, and the low cost of passive capacitors to improve the cost-effectiveness of the device, thus achieving the goal of comprehensive management of imbalance and reactive power.

[0178] This scheme employs a power electronic three-phase load automatic regulation device, installed on the distribution substation side. Given the short radius of the distribution substation, this device superimposes transferred electrical energy across different phases, forcing three-phase current balance at the regulator's front end. However, since the load cannot be changed, the current at the regulator's back end remains unbalanced. For example... Figure 8The diagram shows the application of a three-phase imbalance intelligent adjustment device in Heze City, Shandong Province.

[0179] The above-mentioned devices provide a practical basis for the research and application of the present invention. However, various devices still have various shortcomings in systematically solving problems. Therefore, it is urgent to develop a phase-switching switch type device and intelligent terminal that integrates three-phase imbalance management, self-learning and adaptive technology, load prediction, communication-independent self-decision technology, reclosing technology, and low-voltage side fault location and isolation.

[0180] (1) Self-decision commutation control strategy based on improved quantum genetic algorithm: The difficulty of this invention is that the equivalent impedance of the entire distribution area can be derived based on the voltage and current values ​​of the installation points of each commutation switch. The key point of this invention is to use the improved quantum genetic algorithm with a double-chain structure to predict the voltage change value before and after commutation adjustment and obtain the optimal self-decision commutation control strategy.

[0181] (2) The developed phase-changing switch should be stable and reliable with low power consumption: By adopting bidirectional thyristor technology and combining the advantages of semiconductor devices and contactor switches, a new type of phase-changing switch with a phase-changing time of less than 10ms and an adjustable phase-changing time is developed. Ensuring that short power outages do not affect users' power consumption is one of the difficulties and key points of this invention.

[0182] (3) Research on the optimization of anti-maloperation criteria based on deep learning: Based on the study of the transient component of the residual current of the branch line, a criterion for anti-maloperation is formed with pure power transfer quantity as the criterion. The criterion results are continuously optimized by the rough set method and the deep self-learning algorithm based on the historical data of operation, and the accuracy of the anti-maloperation criteria is improved. This is the key point of this invention.

[0183] Invention research content:

[0184] This study investigates and classifies low-voltage faults in Taizhou, and combines statistical data to study the changing characteristics of electrical quantities under different fault types. Based on historical data, it conducts multi-dimensional analysis of three-phase imbalance problems, studies the impact of new energy access on reactive power and voltage three-phase imbalance in distribution transformer areas, analyzes load consumption patterns based on historical data, statistically investigates the load characteristics of transformer areas and determines the sensitive loads in the transformer areas, and proposes principles for the selection and layout of phase-switching switches.

[0185] A model for the load characteristics and new energy access of the transformer substation was established. An improved quantum genetic algorithm with a double-chain structure was used to study the three-phase self-balancing technology of the transformer substation. An autonomous optimization algorithm for the commutation criterion threshold based on historical switch operation data was studied. Based on a large amount of historical data and the continuous updating of existing data, a low-voltage fault prediction algorithm was studied. The prediction results were continuously optimized and the prediction accuracy was improved by using the rough set method of fuzzy theory.

[0186] Based on the transient characteristic components of residual current, this study investigates low-voltage fault location algorithms and isolation criteria; for different types of transformer area faults, it studies and analyzes the residual current change curves, characteristic components, and distribution characteristics before and after the fault, and forms anti-interference and anti-maloperation criteria; considering the complex on-site conditions and the need to reduce the workload of operation, it studies the reclosing technology of the phase-switching switch in the low-voltage side fault location and isolation process.

[0187] Based on voltage and current data collected at different times from the low-voltage terminal phase-switching switch, this study investigates a method for automatically acquiring line impedance. Based on the calculated line impedance, using low-voltage power flow simulation technology, the mathematical relationship between the unbalance at the transformer substation head end and the line end impedance and phase-to-phase voltage difference is derived. A commutation criterion based on line impedance and phase-to-phase voltage difference is then investigated. Combined with load forecasting of the transformer substation, a self-decision control strategy for load switching of three-phase unbalanced switches based on self-organizing network communication is studied, and a prototype of a self-balancing phase-switching switch is developed.

[0188] Based on residual current monitoring, fault location, isolation, and low-power wireless transmission, this research designs the device architecture and its functional modules; develops the device hardware platform and functional modules; and prototypes and tests of new intelligent distribution transformer terminals. A pilot project is selected for demonstration application. Based on the research results of the aforementioned topics, the design, construction standards, and feasibility of the demonstration project are studied, and a feasibility study for the demonstration project is completed. Considering the specific conditions of the Taizhou low-voltage distribution network, a complete set of design for the demonstration project is carried out. Based on the demonstration project plan, construction and implementation of the demonstration project are conducted. Using the demonstration project as the evaluation object, the effectiveness of three-phase imbalance mitigation, fault location, and isolation reliability of the equipment are assessed.

[0189] Invention Implementation Scheme:

[0190] This invention is generally conducted in accordance with the approach of "basic research," "technical breakthroughs," "device development," and "demonstration application." "Basic analysis" establishes the theoretical and framework foundation for key technology research; "key technology research" serves as the backbone, combining practical power distribution network engineering examples to conduct key technology research; based on the key technology research, corresponding prototypes are developed and demonstration applications are implemented.

[0191] This invention, through research on residual current variation curves and characteristic components, distribution characteristics, and three-phase self-balancing techniques before and after low-voltage distribution area faults, proposes a residual current protection strategy that uses the transient component of residual current as the logic operation condition to achieve fault location and isolation in low-voltage distribution areas of the distribution network. It automatically acquires line impedance to deduce three-phase imbalance in the distribution area and achieves three-phase self-balancing by combining it with inter-phase voltage differences. Furthermore, through research on low-power wireless technology, it enables the uploading of measurement and fault information. This provides data support for the analysis and processing of low-voltage faults and improves the automation level of the distribution network.

[0192] (1) A survey and statistics were conducted on different fault types on the distribution network side in Taizhou area. The faults were classified and the electrical quantity change characteristics under different fault types were studied to provide a basis for accurate fault location. Low-voltage cable line faults are mainly grounding faults, open wire faults, and mixed grounding and open wire faults.

[0193] Grounding faults mainly include single-phase grounding and two-phase grounding; open-circuit faults mainly include one-phase open circuit, two-phase open circuit, and neutral line open circuit. The changes in electrical quantities during each type of fault are analyzed and summarized.

[0194] (2) Investigate the three-phase imbalance problem in Taizhou area and study the impact of single-phase access of distributed power sources on the three-phase voltage imbalance of distribution substations;

[0195] First, historical data of distribution transformer areas for the past year were exported from marketing and electricity consumption information collection systems. The changing trend of three-phase imbalance in transformer areas with significant losses over the past year was statistically analyzed. Furthermore, a multi-dimensional analysis of the three-phase imbalance problem was conducted, incorporating factors such as transformer load characteristics, climate, and season, to provide a basis for three-phase imbalance mitigation strategies.

[0196] Meanwhile, a survey was conducted on the number of distribution transformer substations in Taizhou that include distributed photovoltaic power. Through statistical analysis of the power load, number of users, power consumption, and power consumption trends of these substations, typical distribution transformer substations were selected. Through multi-dimensional simulation analysis of distributed photovoltaic power generation, reactive power characteristics of distribution transformer substations, and power quality, the relationships between various factors were obtained.

[0197] (3) Conduct statistical surveys on the load characteristics of the transformer area, analyze the load power consumption patterns in conjunction with historical data, determine the sensitive loads of the transformer area, and propose the principles for selecting and arranging phase-switching switches;

[0198] By acquiring historical data and other relevant information on the electricity load of the distribution area through the system, and analyzing factors such as electricity consumption, user type, and electricity consumption growth, the nature of the load in the distribution area is determined. Based on the load nature, it is determined whether the distribution area has sensitive loads. For loads / equipment in the power grid, if voltage fluctuations or sudden changes cause them to malfunction or experience a decline in function, these loads / equipment are called sensitive loads / equipment. Table 1 shows the definition of sensitive loads, as follows:

[0199] industry Sensitive load Electronics and electrical appliances RF generator, lithography machine, etching machine, diffusion furnace, PCB drilling rig, pick and place machine, reflow soldering machine Automobile manufacturing Spray painting, drying ovens, welding, conveyor chains / transmission systems, engine production lines Food and Medical Processing and packaging production lines, blow molding machines, pharmaceutical production lines, medical imaging equipment, refrigeration units, compressors, ventilation systems, temperature controllers, ultrapure water pumps, packaging equipment, and high-speed filling machines. Public facilities elevators, escalators Other automation Human-machine interface control panels, CNC machine tools, robots, servo controllers, data centers, telecommunications, printing machinery, dryers, raw material handling machines, extruders, PLCs, contactors, regulated power supplies, sensors, AC I / O cards, timers, counters, and safety components.

[0200] Based on the classification of sensitive loads in Table 1, and combined with the analysis of relevant data on the nature of the power load in the distribution area, it is possible to more accurately determine whether the distribution area contains sensitive loads, thereby selecting suitable distribution areas for addressing three-phase imbalance issues, and proposing basic principles for the selection and layout of phase-switching switches.

[0201] (1) Establish a model of load characteristics and new energy access in the transformer area, and study the three-phase self-balancing technology of the transformer area using an improved quantum genetic algorithm with a double-chain structure;

[0202] First, an optimization model for the three-phase load imbalance in the distribution area and a model for renewable energy access are established. The phase sequence of all users is represented by a matrix, taking into account load type, commutation costs, and the impact of renewable energy access. Using a conventional daily load model for the distribution network, and with the objective of minimizing economic costs while maintaining the lowest three-phase load imbalance, mathematical models for the three-phase imbalance degree are established.

[0203] The analysis of three-phase load imbalance in a distribution network involves analyzing user loads over a period of time to obtain the electrical load, average phase voltage, and power factor, leading to a method for calculating the average phase current. By calculating the average three-phase load, a three-phase load function for the distribution network is established.

[0204] An improved quantum genetic algorithm employs a double-chain structure to overcome the randomness of encoding and the frequent decoding problem in the optimization process of the basic quantum genetic algorithm. It also utilizes a dynamic rotation angle adjustment strategy and an adaptive algorithm to adjust the search angle. Ultimately, a phase sequence adjustment scheme is derived, resulting in a reduction of three-phase load imbalance in the distribution network, a decrease in line losses, a reduction in distribution transformer losses, a rational allocation of transformer output, and an improvement in the safe and efficient operation of electrical equipment.

[0205] (2) Research an autonomous optimization algorithm for the commutation criterion threshold based on historical switch operation data;

[0206] Based on historical data such as voltage and current collected from the phase-switching switch, the large amount of historical data is first preprocessed to remove invalid and missing values. Then, feature values ​​are extracted by analyzing data fluctuation trends, electricity consumption dispersion coefficients, and variability. While preserving the essential nature of the data as much as possible, the dimensionality of the extracted feature values ​​is reduced, and principal component analysis, factor analysis, and independent component analysis are comprehensively applied.

[0207] Next, the bagging method is used to construct a prediction commutation threshold function, and then these functions are combined into a prediction function in a certain way. First, a weak learning algorithm and a training set are given; then the learning algorithm is used multiple times to obtain a sequence of prediction functions, which are then voted on; finally, the accuracy of the result is improved, and the optimal threshold for the commutation criterion is obtained.

[0208] (3) Based on a large amount of historical data and the continuous updating of existing data, we study the low-voltage fault prediction algorithm and continuously optimize the prediction results through the rough set method of fuzzy theory to improve the prediction accuracy.

[0209] The output of common fault location methods can be represented as: 1 indicates a faulty line, and 0 indicates a fault-free line. Different lines may exhibit significant differences for a particular fault characteristic. Therefore, the fuzzy theory employed in this invention utilizes rough sets for fault discrimination, abandoning the absolute membership relationship of either 0 or 1 in conventional fault location. It establishes a system where membership values ​​for fault results can be selected within the interval [0,1]. The uncertainty of each fault selection result can be represented by a membership function. The probability of a fault occurring decreases as the membership value of the selected result approaches 0; conversely, the closer the membership value is to 1, the higher the probability of successfully locating the fault.

[0210] Based on the summary of fault types and corresponding electrical quantity change characteristics summarized in Project 1, potential patterns are discovered through analysis and reasoning of real-time data to achieve fault prediction. Simultaneously, the rough set method continuously optimizes the prediction results and improves prediction accuracy by using a large amount of historical and existing data for continuous updating.

[0211] (1) Study on low-voltage fault location algorithm and isolation criterion based on residual current transient characteristic components;

[0212] When a fault occurs in a low-voltage power grid, the switching switch can detect the fault overcurrent. Considering the characteristics of the low-voltage power grid itself, an ant colony algorithm is used for fault location analysis. The ant colony algorithm is a simulated evolutionary algorithm that mimics the foraging behavior of real ant colonies in nature. While the behavior of a single ant is simple, a group of simple individuals exhibits extremely complex behavior.

[0213] Since it's impossible to predict whether a fault is a single point or multiple points when it occurs, distributed computation is performed based on the ant colony algorithm. First, assuming a single point of failure, a global optimization calculation is performed to find the minimum value of the evaluation function and the fault point under the assumed conditions. Next, assuming two-point and three-point failures, the minimum value of the evaluation function and the fault point under the assumed conditions are calculated for each. Finally, the minimum values ​​of the evaluation function under the three assumed conditions are compared. The assumption with the minimum evaluation function value is valid, determining the global optimal solution. After identifying the faulty branch, the commutator switches activate to achieve isolation.

[0214] (2) For different types of faults in the transformer area, study and analyze the residual current change curves, characteristic components, distribution characteristics, etc. before and after the fault to form anti-interference and anti-maloperation criteria.

[0215] This study investigates the variation of residual current during faults and its value during fault-free periods to obtain the abrupt change in residual current during faults. The magnitude of this abrupt change is used as the logic operating condition for residual current protection. The transient characteristic components of the residual current are also studied to establish a criterion for preventing maloperation, based on pure power transfer.

[0216] Fault-free tripping of power equipment typically refers to equipment tripping without a short circuit or other fault occurring. Fault-free tripping is generally caused by sudden changes in current or power. Therefore, the magnitude of these sudden changes determines whether the device has entered the startup state. When the absolute value of the sudden change in current or power exceeds the current setting or power setting, it indicates that the device has entered the startup state.

[0217] After the device meets the start-up conditions for sudden change current or sudden change power, if it simultaneously meets the following conditions: the absolute value of the sudden change in power is greater than the active power setting before tripping, all three-phase currents are less than the operating current setting, and the time for meeting these conditions is greater than the fault-free tripping time setting, then the fault is judged as a fault-free trip. Ultimately, the pure power transfer quantity is used as the criterion for preventing maloperation.

[0218] The types of faults on site are very complex, and the time for different faults to transition from transient to steady state is also different. The residual current of a certain branch may suddenly increase and then return to normal. Therefore, the device is set to take action after detecting a fault to prevent the possibility of misjudgment. At the same time, the device delay action time and the residual current protection threshold can be manually set.

[0219] (3) Considering that when a fault occurs and the phase-switching switch trips, maintenance personnel need to manually reset it on-site, which increases their workload, the equipment is equipped with a reclosing function. When the line fault is instantaneous or temporary, the phase-switching switch detects the fault and isolates it. When the fault is detected and cleared, it automatically closes the switch, avoiding manual intervention. At the same time, the number of reclosing attempts and the time from tripping to reclosing can be manually set.

[0220] Self-balancing control based on self-sensing learning first utilizes electrical quantity data (three-phase current, three-phase voltage, etc.) from the power consumption information collection system at the transformer substation end and daily active power data from end users (smart meter users). It establishes an automatic topological association through smart meters, actively soliciting non-electrical quantity data from the meters to directly determine the optimal location and capacity for installing three-phase imbalance control measures (load transfer switch type) for each transformer substation, and provides a comparison of the effects before and after the switchover. Based on system data, it fully considers selecting more diverse and extensive electrical quantity characteristics to maximize the analysis of distribution network three-phase imbalance control strategies based on historical data mining. Considering the periodicity and regularity of the distribution network, it first analyzes historical data to find common features, then uses an improved quantum genetic algorithm to classify the data, and finally applies the analysis results to the distribution network three-phase dynamic self-balancing control strategy.

[0221] Data sources include: transformer ledger information, including transformer rated capacity, CT ratio, and transformer area wiring method; historical data sections from the transformer outlet side, extracting 24, 48, and 96 historical data points by date, including information such as three-phase current, three-phase voltage, active power, and reactive power, with as many sections as possible, and must include data on the dates when three-phase imbalance occurred; daily and monthly electricity consumption data for all low-voltage users in the transformer area, with as many sections as possible, preferably with dates consistent with the first-end section of the transformer area.

[0222] The commutator switches communicate automatically with the terminal through CGMESH's self-organizing network. When a terminal communication failure occurs, the CGMESH communication network selects a suitable switch as the terminal communication node according to the load rate of the commutator switches. The network is reorganized through self-sensing, which does not affect the mutual communication between the commutator switches and ensures the data acquisition and execution of the commutation function of the commutator switches.

[0223] Considering the integration of distributed residential photovoltaic (PV) systems into distribution transformer areas, simulation modeling of voltage and current data within the distribution transformer area is used to extract characteristic values ​​of data fluctuation trends. Simultaneously, taking into account the power generation characteristics of distributed residential PV, the power generation of distributed residential PV systems at different time periods is predicted, enabling advance prediction of active power dynamic self-balancing. In the event of communication anomalies, the commutator can autonomously detect, collect, judge, and operate. Based on the "four autonomous" control method of the commutator, the optimal commutation strategy can be derived by collecting voltage and current information at the installation point. Furthermore, the coordination strategy is continuously optimized using rough set methods and a deep self-learning algorithm based on historical operating data to achieve the setting of the difference between the setpoint and the time, thus resolving malfunctions caused by communication instability.

[0224] The anti-maloperation criterion strategy is optimized based on self-sensing learning technology. By analyzing the transient component of the residual current in branch lines, the pure power transfer quantity is used as the anti-maloperation criterion. The distribution network database is then input into the rough set computing system. Without providing any empirical knowledge, the rough set algorithm is used to directly analyze and reason about the data, thereby discovering the hidden power grid fault patterns, continuously optimizing the criterion results, and improving the accuracy of the anti-maloperation criterion.

[0225] By using deep learning technology to analyze historical power grid data, fault diagnosis and analysis are performed on power equipment that cannot operate normally based on the operational data. This leads to the derivation of a power grid fault diagnosis method, forming a feature-adaptive distribution network fault analysis method, thereby accurately isolating faults.

[0226] By using deep learning technology to monitor branch circuit data, the threshold is continuously optimized based on the residual current characteristic components. The residual current threshold of the protection equipment is set, and the equipment trips when the threshold is exceeded. Since the load change of the distribution transformer area is dynamic, the accuracy of the threshold setting is extremely important for the stability of the transformer area.

[0227] Based on the calculated line impedance value, the mathematical relationship between the unbalance at the beginning of the distribution area and the impedance at the end of the line and the phase-to-phase voltage difference is derived through low-voltage power flow simulation technology. According to the conditions for three-phase unbalance treatment, when the load unbalance of the distribution transformer is >15% and exceeds the limit for 1 hour in a day, treatment is required. At the same time, the phase-to-phase voltage difference is monitored in real time, and the commutation criterion is formed by using self-learning technology.

[0228] Traditional phase-commutation switch hardware uses relays and permanent magnet circuit breakers. These switches have short mechanical life and long commutation times. Therefore, this invention adopts a bidirectional thyristor parallel contactor structure, which eliminates impact during commutation and provides fast switching speed. The structural diagram is shown below. Figure 9 As shown.

[0229] Switching technology: Zero-crossing commutation technology is adopted, with a commutation time of no more than 10ms, which can avoid large inrush currents generated at the moment of load switching and avoid affecting the user's electrical equipment.

[0230] Energy-saving and reliable: It adopts a dual-phase thyristor parallel contactor structure, which avoids the heat generation problem caused by long-term operation of semiconductor devices in traditional phase-changing switches. It will greatly improve the stability and safety of distribution network operation. The device itself has low loss, close to zero loss, and has no electromagnetic pollution or noise pollution to the environment.

[0231] Since the peak power load periods in the distribution substation are 10:00-12:00, 16:00-18:00, and 20:00-22:00, during which the three-phase imbalance is greatest, the phase-switching switch designed in this invention can be set to manage the three-phase imbalance during these three time periods. Furthermore, the imbalance threshold can be manually set on the phase-switching switch interface according to different site conditions.

[0232] The switching switches communicate via a CGMESH network. The CGMESH network is an IPv6-based mesh multi-hop network. With the CGR as the root node, it facilitates the convenient and flexible extension and expansion of wireless network coverage through multi-level relay. In the CGMESH wireless mesh network, wireless nodes can conduct end-to-end direct IPv6 communication with the remote backend, and wireless nodes at each level within the wireless network can also conduct parallel bidirectional communication with each other.

[0233] (1) Based on residual current monitoring, fault location, isolation and low-power wireless transmission, study and design the architecture and functional modules of the prototype;

[0234] (2) Development of device hardware platform and functional modules. In order to facilitate the main station to query data, the terminal builds a communication backend from the intranet and promptly notifies the operation and maintenance personnel in case of failure.

[0235] (3) Conduct prototype manufacturing and testing.

[0236] (4) Based on the research results of the aforementioned topics, and considering the actual distribution area situation in Taizhou, study the design, construction standards and feasibility of the demonstration project, and complete the application and verification of the demonstration project.

[0237] This invention proposes a "four-autonomous" control method for commutator switches based on self-sensing deep learning: transforming the original centralized control mode into a centralized + decentralized control mode, enabling the commutator switch to autonomously detect, collect, judge, and act in the event of communication anomalies. By collecting voltage and current information at the installation point, the optimal commutation strategy can be derived, and the coordination strategy is continuously optimized through rough set methods and deep self-learning algorithms based on historical operating data to achieve the setting of the difference between the setpoint and the time, thus solving the problem of malfunctions caused by communication instability.

[0238] This invention proposes a CGMESH-based wireless multi-hop self-organizing network to enable plug-and-play functionality for phase-change switches: the phase-change switches automatically communicate with terminals via the CGMESH self-organizing network, allowing for plug-and-play operation without debugging during subsequent switch expansion. Phase sequence identification is achieved by matching switch sampling with terminal sampling, solving the problem of difficult phase sequence identification in field power cables.

[0239] This invention proposes a fault location algorithm for low-voltage distribution networks based on the transient characteristic components of residual current: based on the voltage and current values ​​of each phase-switching switch and terminal acquisition installation point, the equivalent impedance of the entire distribution network can be derived; based on the transient characteristic components of residual current, the distance between the fault point and the terminal and switch is calculated, and the fault area is located, thus solving the problem of single-phase short-circuit fault location and isolation in low-voltage distribution networks.

[0240] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention. The embodiments of the present invention have been described in detail above, but the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

[0241] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0242] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention. The embodiments of the present invention have been described in detail above, but the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A commutating switch intelligent terminal based on three-phase self-balancing, low-voltage fault location isolation, characterized in that: It includes an intelligent distribution transformer terminal, a distribution side, a user side, and a distribution transformer. The intelligent distribution transformer terminal is connected to the distribution side and the user side respectively, and the distribution transformer is connected to the user side through the distribution side. The intelligent distribution transformer terminal includes a data acquisition APP module and an automatic phase-switching load adjustment module connected to it. The data acquisition APP module is used to collect the low-voltage total and branch line currents, voltages, active power, reactive power, and power factor on the distribution side. An automatic phase-switching load regulation module is used to adjust the current, voltage, phase, power, and topology of each user on the user side; the distribution side includes a low-voltage main switch and multiple branch line switches; The user side includes multiple commutation switches for the operating mechanism responsible for performing load commutation; Specifically, it includes the following steps: Step 1: Based on the residual current change curves and characteristic components, distribution characteristics and three-phase self-balancing of the low-voltage distribution area before and after the fault, a residual current transient component is proposed as the logic action condition to realize the fault location and isolation of the low-voltage distribution area of ​​the distribution network. Step 1 is described in detail as follows: Step 1.1: Investigate and statistically analyze different fault types on the distribution network side, classify the faults, and study the electrical quantity change characteristics under different fault types to provide a basis for accurate fault location. Step 1.2: Investigate the three-phase imbalance problem and study the impact of single-phase access of distributed power sources on the three-phase voltage imbalance of distribution transformer areas; export historical data of distribution transformer areas for the past year from the marketing and electricity consumption information collection system; statistically analyze the changing trend of three-phase imbalance degree in transformer areas with large losses over the past year; and conduct multi-dimensional analysis of the three-phase imbalance problem in combination with the load characteristics, climate and seasonal factors of the transformer area to provide a basis for three-phase imbalance management strategies. Meanwhile, the number of distribution substations containing distributed photovoltaic power was investigated. Through statistical analysis of the power load, number of electricity users, power consumption and power consumption trends of such substations, typical distribution substations were selected. Through multi-dimensional simulation analysis of distributed photovoltaic power generation, reactive power characteristics of distribution substations and power quality, the relationship between various factors was obtained. Step 1.3: Statistically survey the load characteristics of the transformer area, analyze the load consumption patterns based on historical data, identify the sensitive loads of the transformer area, and propose the selection and layout principles for phase-switching switches; obtain relevant information on historical load data of the transformer area through the system, analyze the load, user characteristics, and power consumption growth, and determine the load characteristics of the transformer area. Based on the classification of sensitive loads and combined with the analysis of relevant data on the nature of the power load in the distribution area, it is possible to more accurately determine whether the distribution area contains sensitive loads, thereby selecting suitable distribution areas for addressing three-phase imbalance issues and proposing basic principles for the selection and layout of phase-switching switches. Step 2: Automatically obtain the line impedance to deduce the three-phase imbalance in the transformer area and combine it with the phase-to-phase voltage difference to achieve three-phase self-balancing.

2. The commutation switch intelligent terminal based on three-phase self-balancing, low-voltage fault location isolation according to claim 1, characterized in that: The phase-changing switches communicate with each other via a CGMESH network. The CGMESH network is an IPv6-based mesh multi-hop network with CGR as the root node. It extends and expands the wireless network coverage conveniently and flexibly through a multi-level relay multi-hop method. In the CGMESH wireless mesh network, wireless nodes can conduct end-to-end direct IPv6 communication with the remote backend, and wireless nodes at all levels within the wireless network can conduct parallel bidirectional communication with each other.

3. The commutation switch intelligent terminal based on three-phase self-balancing, low-voltage fault location isolation according to claim 2, characterized in that: The phase-commutation switch adopts a bidirectional thyristor parallel contactor structure.

4. The commutation switch intelligent terminal based on three-phase self-balancing, low-voltage fault location isolation according to claim 1, characterized in that: In step 1, low-voltage area faults include grounding faults, open-circuit faults, and mixed grounding and open-circuit faults.

5. The commutation switch intelligent terminal based on three-phase self-balancing, low-voltage fault location isolation according to claim 1, characterized in that: In step 1, fault location and isolation of low-voltage distribution areas in the power distribution network are achieved, as detailed below: The anti-maloperation criterion strategy is optimized based on self-sensing learning technology. By analyzing the transient component of the residual current of the branch line, the pure power transfer quantity is used as the anti-maloperation criterion. The distribution network database is then input into the rough set computing system. Without providing any empirical knowledge, the rough set algorithm is used to directly analyze and reason about the data, discover the hidden power grid fault patterns, continuously optimize the criterion results, and improve the accuracy of the anti-maloperation criterion. By using deep learning technology to analyze historical power grid data, fault diagnosis and analysis are performed on power equipment that cannot operate normally based on the operational data. This leads to the derivation of a power grid fault diagnosis method, forming a feature-adaptive distribution network fault analysis method, thereby accurately isolating faults.

6. The commutation switch intelligent terminal based on three-phase self-balancing, low-voltage fault location isolation according to claim 1, characterized in that: Step 2 is described in detail below: Step 2.1: Establish the load characteristics and new energy access model of the transformer area, and use an improved quantum genetic algorithm with a double-chain structure to achieve three-phase self-balancing of the transformer area. Establish a three-phase load imbalance optimization model and a new energy access model for the distribution area: represent the phase sequence of all users using a matrix, and consider the load type, commutation cost, and the impact of new energy access; adopt the daily load model of the distribution network, and establish a mathematical model of the three-phase imbalance degree with the goal of minimizing the economic cost under the condition of the lowest three-phase load imbalance. The analysis of three-phase load imbalance in the distribution network involves analyzing user loads over a period of time to obtain the electrical load, average phase voltage, and power factor, and deriving a method for calculating the average phase current. By calculating the average three-phase load, a three-phase load function for the distribution network is established. The improved quantum genetic algorithm adopts a double-chain structure to overcome the randomness of the encoding and the frequent decoding problem in the optimization process of the basic quantum genetic algorithm. It also adopts a dynamic adjustment strategy for the rotation angle and an adaptive adjustment algorithm search angle. Step 2.2, Autonomous optimization algorithm for commutation criterion threshold based on historical switch operation data: Based on historical voltage and current data collected by the phase-switching switch, a large amount of historical data is preprocessed to remove invalid and missing values; feature values ​​of the data are extracted by analyzing data fluctuation trends, electricity consumption dispersion coefficients, and variability. A prediction commutation threshold function is constructed using bagging, and these functions are combined into a prediction function. Given a weak learning algorithm and a training set, the learning algorithm is then used multiple times to obtain a sequence of prediction functions, which are then voted on. Finally, the accuracy of the result is improved, and the optimal threshold for the commutation criterion is obtained. Step 2.3: Based on a large amount of historical data and the continuous updating of existing data, a low-voltage fault prediction algorithm is studied. The prediction results are continuously optimized and the prediction accuracy is improved by using the rough set method of fuzzy theory. Fuzzy theory is used to identify faults using rough sets, abandoning the absolute membership relationship of either 0 or 1 in conventional fault location, forming a system where the membership value of the fault result can be selected within the interval [0,1]. The uncertainty of each fault selection result is represented by a membership function. Based on the summary of fault types and the corresponding electrical quantity change characteristics, potential patterns are discovered through the analysis and reasoning of real-time data to achieve fault prediction. At the same time, the rough set method uses a large amount of historical data and existing data to continuously refresh and optimize the prediction results, thereby improving the prediction accuracy.

Images