Review of Digital Twin Technology in Low-Voltage Distribution Area and the Implementation Path Based on the ‘6C’ Development Goals

Abstract

Low-voltage distribution area is the “last kilometer” connecting the distribution network and users, and the traditional distribution system is difficult to digitally manage in the low-voltage area, resulting in untimely and imprecise handling of voltage overruns, short-circuit outages, and other abnormal problems. With the deployment of smart meters, new sensors, smart gateways, and other devices in distribution areas, digital intelligent monitoring and management based on digital twins in LV distribution areas has gradually become the focus of distribution network research. In view of the profound changes that are taking place in the low-voltage distribution area, this paper first summarizes the characteristics and shortcomings of the existing digital twin research in the low-voltage distribution area, then puts forward the ‘6C’ development goals for the digital transformation of the low-voltage distribution area, introduces the practice work of Guangxi Power Grid Corporation around the ‘6C’ development goals in the low-voltage distribution area. Finally, the future research work of the ‘6C’ development goals for the digital transformation of the low-voltage distribution area is promising.

1. Introduction

With global warming, in order to reduce the application of fossil fuels to achieve carbon emission targets, on the one hand, a large amount of distributed new energy is connected to the power grid to solve the energy crisis, and the new energy power generation capacity in countries around the world has been growing rapidly in recent years [1]. On the other hand, as the power conversion efficiency of power electronic devices is usually greater than 90%, a large amount of electrical energy is converted by power electronic devices to drive consumer loads [2], and power electronic devices are widely distributed in equipment or systems such as LED lighting, electric vehicle charging piles, and HVDC transmission systems [3]. The use of massive distributed new energy sources and power electronic devices has a significant impact on the morphological structure and operational characteristics of distribution networks.

Massive distributed resources and power electronic device access leads to a new distribution system showing a high proportion of new energy and a high proportion of power electronic “double high” characteristics, system state complexity, mainly manifested as follows: (1) Distributed new energy output uncertainty, spatial and temporal distribution of power peaks and valleys is not uniform, the new distribution system of photovoltaic back-feeding, short-time heavy overload and other power balance problems are prominent [4]. (2) A large number of distributed new energy generator sets are gradually replacing the existing rotating generator sets in the distribution network, at the same time, with the power electronic loads, new energy storage equipment, distribution network flexible interconnection equipment and other power electronic devices applied in large quantities, the degree of power electronics in the distribution network has deepened significantly. The new distribution system trend direction compared to the traditional distribution network has changed significantly. By the power electronic device, multi-time scale operation characteristics of the distribution network voltage, power angle, frequency, and other types of instability problems are endless; the new distribution system’s safe and stable operation control is difficult [5]. The above factors lead to a series of major problems and challenges in promoting the green transformation of energy use in new distribution systems and the high-quality development of distribution networks [6].

Low-voltage distribution area (LVDA) is the “last kilometer” connecting the distribution network and users, and the safe and stable operation of LVDA is directly related to the quality of users’ electricity consumption. The typical LVDA structure is shown in Figure 1. As LVDA is at the end of the grid and has a large number and a wide distribution range, it is difficult to manage the grid. Existing studies have shown that improving the digitalization level of distribution networks has a positive effect on improving the efficiency of energy utilization and enhancing the system control capability. At present, although the 10 kV and voltage level MV distribution system has basically achieved digitization [6,7], the digitization capability of LVDA is still significantly deficient, which is mainly manifested as follows: (1) Small measurement coverage. Most of the current LVDAs are only installed with smart meters, but this is far from enough for LVDAs. Existing smart meters collect data every 15 min on average, while distributed power output is strictly affected by real-time weather, wind speed, temperature, and other factors, and the output fluctuation time is usually less than 5 min [8,9,10,11,12]. In particular, accidents such as user short-circuit faults and power outages occur more randomly, and the duration may even be less than 1 min, which may lead to the end of the accident before it can be sensed [13,14]. (2) Limited state sensing capability. Although LVDA can achieve accurate measurement of electrical and non-electrical quantities by installing sensors, the results of LVDA state estimation will be different due to the impact of the sensor sampling rate, sampling accuracy, transmission rate, etc. The algorithms related to how to give full play to the data advantages of sensors and accurately perceive the state of the LVDA have not yet been adequately studied. (3) Limited control measures. Existing power dispatch departments usually only control the power supply or disconnection status of the station area, and the management of LVDA is mainly achieved by controlling the switching of the 10 kV/0.4 kV transformer on the medium voltage side [5,15,16]. In the LV distribution area, the existing technology mainly realizes the protection of LV stations by setting the capacity of mechanical circuit breakers on a step-by-step basis. The overall informatization capacity of the desk area distribution room is insufficient, and the management of various types of switches, reactive power compensation devices, and other equipment in the distribution room mainly relies on manual in situ operation, which is inefficient [17]. (4) Lack of comprehensive scheduling of distributed resources. Although the station area has comprehensively analyzed the impact of the capacity of distributed power supply on the station area in terms of early planning, it has not taken into account the growth of the scale of user electricity consumption and the dynamic changes in the characteristics of user electricity consumption, and it is less likely to regulate the output of distributed power supply in the operation process, resulting in prominent problems of overvoltage and tidal current back-feeding in the station area during the subsequent operation process, and the utilization of distributed resources in the station area has a poorer effect, which has increased the cost of power grid operation [18].

The rapid development of technologies such as digital twin, big data, internet of things (IoT), and artificial intelligence is ushering in the Smart Grid 3.0 era [19]. The more convenient information interconnection and intercommunication between distribution network equipment make it possible to build a digital twin distribution network. The construction of digital twin LVDA will fundamentally solve the state sensing and control challenges faced by traditional LVDA. Existing research has started the discussion from several aspects of key devices, power business analysis algorithms, and software, systems, and platforms required for LVDA construction, and achieved certain research results [20]. However, since the research on LVDA digital transformation has just started, no research has been conducted specifically to summarize the current situation of LVDA digital twin development, and no planning has been made for the development objectives and development path of LVDA digital stations. In addition, most of the existing research focuses on the innovation of algorithmic software and fails to systematically and comprehensively discuss the disclosure of LVDA digital twin operation and management cases. Therefore, this paper will review the application of digital twin technology in LVDA data collection, status perception, flexible control, and other aspects, and summarize the existing problems in the application of digital twin technology in LVDA. In addition, in response to the problems identified in the existing literature review, this paper will propose the ‘6C’ development goals for the construction of LVDA digital twins and introduce the working practice of LVDA construction based on this idea. The main contributions of this article are as follows:

(1) The summary of the research that has been carried out on the construction of LVDA digital twins, including state estimation, electrical quantity measurement, distributed resource aggregation, electrical equipment regulation, and other aspects. Based on the technical focuses of the existing LVDA research in this paper, the existing research is divided into five types, namely state estimation techniques, sensing technologies, regulation and control technologies. Aggregation technologies for distributed energy resources, emerging devices, and terminals. At the same time, the characteristics and technical deficiencies of different types of research were analyzed, which can provide guidance for researchers to carry out subsequent related work.

(2) The ‘6C’ development goal of LVDA is proposed, that is, “Clear Measurement, Cognizable State, Controllable Devices, Configurable Parameters, Combinable Resources, Coordinated Autonomy”. By summarizing the shortcomings of the existing research work, this paper proposes the ‘6C’ development goals of the LVDA digital twin. The autonomous goal of LVDA is proposed for the first time, which is a significantly different idea from the existing standards and frameworks of the distribution network digital twin.

(3) It summarizes the practices work carried out by Guangxi Power Grid Company in Guangxi Province of China around the goal of ‘6C’ and elaborates on the systematic work in establishing the LVDA digital twin system with specific cases. A new management model for distribution networks has been proposed. That is, integration of data sources and equipment in the marketing, distribution, and dispatching business of LVDA. This approach enables data interconnection and interoperability among the Marketing Department, the power distribution department, and the dispatching department, significantly enhancing the operational efficiency of the power grid. Furthermore, a digital construction model of LVDA, namely the digital power supply station, has been proposed. Digital power supply stations have changed the traditional “cloud-edge” collaborative management mode of distribution networks, incorporated edge-side computing units, and improved the efficiency of LVDA status perception and regulation.

(4) The practical difficulties that still exist in achieving the ‘6C’ goals of the new distribution system are analyzed, and the future research direction of the new distribution system in the distribution area is pointed out.

The rest of this paper is as follows: Section 2 summarizes the current research status of LVDA digital twins, and analyzes the state estimation, state monitoring and sensing, system regulation methods, distributed resource aggregation, and new devices and terminals, respectively. Section 3 explains the specific meaning of LVDA’s ‘6C’ development goals and introduces the case of LVDA digital twin construction around the ‘6C’ goals. Section 4 summarizes some future research work in the area of LVDA.

2. A Study on the Current Landscape and Emerging Trends of Digital Twin Technology in Low-Voltage Distribution Networks

2.1. Context for the Development of Digital Twin Technology in Low-Voltage Distribution Systems

A Low Voltage Distribution Area (LVDA) typically refers to the low-voltage section downstream of a medium-voltage transformer, with a standard three-phase voltage of 400 V in China. Owing to the sheer number, broad geographic dispersion, and significant variability among LVDAs, their management presents substantial challenges. Consequently, prior studies on distribution network monitoring and regulation have predominantly targeted medium-voltage (≥10 kV) and higher-level systems, with limited attention devoted to low-voltage segments. As distributed generation technologies such as microgrids and nano-grids rapidly proliferate, their integration into LVDAs has indeed helped decrease reliance on conventional fossil fuels and improve power supply reliability for end-users, but it has also introduced severe impacts on the secure and stable operation of the grid. In addition, the mass deployment of novel power electronic loads—exemplified by EV charging piles—has further intensified the operational complexity within LVDAs. Since the grid lacks holistic monitoring and control of LVDAs, essential parameters—such as distributed generation output, fluctuation dynamics, and operation modes—remain inadequately sensed, potentially causing the control strategies at the medium-voltage level to clash with those governing low-voltage distributed generation, ultimately rendering regulation ineffective.

Existing research demonstrates that augmenting the number of measurement points and maximizing data acquisition within distribution systems significantly enhances the ability to perceive and regulate their operational states [21]. Conventional technologies leveraged the extensive deployment of smart meters in distribution networks to construct advanced metering infrastructure (AMI), which, by utilizing metered data to infer grid conditions, has effectively reduced energy losses [22] and enhanced the utilization efficiency of multiple energy sources [23]. Through the adoption of non-intrusive measurement (NIM) techniques, smart meters have simultaneously achieved grid state awareness and safeguarded user privacy [24]. In recent years, major advances in information and communications technology (ICT)—including integrated circuit chips, the Internet of Things (IoT), and Artificial Intelligence (AI)—have not only reshaped AMI systems [25,26], but have also been progressively adopted in low-voltage distribution areas, delivering promising outcomes in overcoming obstacles related to state monitoring and control [27].

ICT technologies have profoundly influenced the measurement frameworks, communication infrastructures, computing architectures, and operational paradigms of low-voltage distribution areas (LVDAs). On one front, novel low-power and non-intrusive sensors are being explored in LVDAs to tackle the challenges of extensive data acquisition and large-scale sensor deployment [28]. Technologies like LORA and power-line communications (PLC) have been adopted for extensive data gathering from widely distributed devices in LVDAs [29,30]. Considering the interaction between vast numbers of edge devices and cloud-based grid platforms, computing frameworks oriented around “Cloud-Edge-End-Core” collaboration have been proposed and implemented in certain distribution areas [31,32]. As comprehensive data from LVDAs becomes increasingly accessible, observability emerges as a central focus of research. Prior research has shown that smart meter data can be used to assess the observability of distribution networks [33], enabling the construction of digital models to infer their operating states [34]. Unlike in higher-voltage systems, a complete AMI framework has yet to be implemented in LVDAs. Although smart meters exist, they are merely components of the broader AMI and grid infrastructure, and LVDA users lack data ownership and management rights. Consequently, research has noted that amid fast-paced ICT developments, the relationship between data and services will become increasingly transparent, making data feature extraction and processing central to LVDA research [35]. End-users are now capable of equipping their premises or LVDAs with extensive sensor arrays and computing modules to monitor real-time operational status and formulate energy-optimized control strategies [36,37]. Hence, a paradigm shift in data measurement—from medium and high-voltage systems to LVDAs—is inevitable. The advent of digital twin-enabled LVDAs is bound to transform the structure and operation of future distribution networks profoundly. It is foreseeable that, in the future, personal energy regulation strategies may diverge from those of grid dispatchers—since the grid prioritizes system-wide supply–demand balance, whereas LVDA users are chiefly concerned with avoiding localized power interruptions. Accordingly, advancing research on digital twin technologies for LVDAs has become an urgent and strategic imperative.

This paper constructs the combination of ‘Low voltage distribution network’, ‘Distribution’, ‘Digital twin’, ‘Measurement’, ‘Control’, ‘Artificial intelligence’, “Smart electricity meter”, “Edge computing equipment”, and “Devices”, etc. Then this paper searches for a combination of different words on “Web of Science” and determines the 250 papers with the highest relevance. Through individual screening, papers that are not related to the field (such as applications targeting medium-voltage distribution networks, ICT technology and industry reviews, and technologies with communication and AI algorithms as the core of research) were removed. Drawing on current research progress in LVDA digital twin technologies, this paper classifies the field into five key dimensions—state estimation, operational status sensing, system regulation strategies, distributed resource aggregation, and emerging devices and terminals—which will be detailed in the subsequent sections.

2.2. State Estimation Techniques in LVDA-Oriented Digital Twin Systems

Reliable data forms the foundation of LVDA digital twins, and the precision of measurement data critically influences system state perception and operational control; therefore, accurate state estimation becomes especially vital. Prior research has investigated the sources of metering errors in AMI systems. By aggregating data from low-voltage meters and employing orthogonal matching pursuit combined with the orthogonal matching pursuit and recursive model, researchers were able to filter out anomalous readings and identify meters suffering from aging or measurement degradation [38]. In [39], clustering techniques were applied to monitor data from LVDA meters, followed by the development of a decision tree algorithm to detect meters affected by abnormal faults or aging. Furthermore, some studies contend that measurement errors in smart meters are inherently unavoidable; hence, continuous enhancement in both resolution and accuracy is required to enable effective monitoring of transient high-power loads [40]. In [41], a novel data processing and error estimation method tailored for low-voltage smart meters was introduced, significantly mitigating metering inaccuracies.

2.3. Sensing Technologies for LVDA Digital Twin Systems

Leveraging digital twin data to accurately perceive the operational state of LVDAs provides substantial support for grid companies’ business operations at the LVDA level—this has become a focal point of ongoing research. This study identifies and categorizes 14 typical scenarios for LVDA state perception, detailed as

Table 1. The 14 typical LVDA State Awareness Scenarios.

LVDA State	Core Perceived Objectives	Main Technical Approaches	Validation Approach	Representative Literature
Three-phase grid voltage state recognition	Accurate acquisition of voltage amplitude, phase, and voltage difference between nodes	Deep learning, interval arithmetic electrical-free model algorithms algorithms for locally observed data	Both simulation and field test	[42,43,44,45,46,47,48,49]
Tidal current calculation	Estimated regional current distribution for low-voltage distribution	Calculation method for PV-enriched stations AMI data-based tidal current calculation	Both simulation and field test	[50,51]
Line Energy Loss Calculation	Optimization of the distribution structure to reduce energy losses	Calculation of three-phase unbalance conditions, GIS data fusion, and integrated three-phase estimation	Field test	[52,53,54]
Topological identification	Identify frequently changing low-voltage network topologies	Bayesian estimation, voltage sensitivity factor, fusion of OSM maps and meter data, graph learning techniques	Both simulation and field test	[55,56,57,58,59,60,61,62]
Estimation of non-technical losses	Detecting power theft and reducing economic losses	Load time series prediction, deep learning, LV line temperature monitoring, edge device calibration optimization	Field test	[63,64,65,66,67,68]
Voltage overrun sensing and prediction	Predicting voltage anomalies to ensure power quality	Monte Carlo algorithm, near-real-time machine-learning framework, and linear correlation analysis of historical data	Simulation	[69,70,71,72]
Identification of three-phase voltage unbalance	Monitoring of the three-phase unbalanced state of the power grid	Power carrier communication devices, single-phase distributed power supply impact analysis, and load impact regulation	Field test	[73,74,75,76]
Fault Detection and Location	Quickly locate short-circuit faults to improve power supply reliability	Smart metering device recording, AI model-free algorithms, impedance state measurement, fault indicator fusion	Simulation	[77,78,79,80,81,82,83]
User-transformer relationship identification	Accurately correlate transformers and users to avoid billing errors	Gaussian clustering algorithms, data-driven, and principal component analysis	Field test	[84,85,86,87]
Load identification and behavioral forecasting	Identify multiple types of loads and predict short-term electricity behavior	Global computing frameworks, deep learning models, load classification management, and Monte Carlo statistical forecasting	Simulation	[88,89,90,91,92,93,94]
Phase recognition	Accurate identification of grid phases at low sampling rates	Generative Adversarial Network (GAN) augmented data, and nonlinear dimensionality reduction algorithms	Simulation	[95,96]
Power quality monitoring	Extension of power quality anomaly detection to low-voltage areas	Accurate timing strategy, hardware chip design, and reuse of existing meter platforms	Field test	[97,98,99,100]
Equipment condition monitoring	Predicting distributed power output and transformer health status	Meteorological data prediction, medium and low voltage electrical quantity fusion, causal information fusion	Simulation	[101,102]
Estimated cost-benefit of grid operation	Optimize operational strategies and analyze costs and benefits	Dynamic optimization of operating strategies, voltage-driven reinforcement cost estimation	Simulation	[103,104]

.

2.4. Regulation and Control Technologies for LVDA Digital Twin Systems

Regarding the regulation of LVDAs, existing studies primarily target voltage and power adjustments to maintain power balance and ensure voltage stability within acceptable operational limits.

Table 2. Typical LVDA Regulation and Control Technologies.

Control Methods	Principles	Applicability	limitations	Representative Literature
Localized control for voltage support	Reactive power coordination of residential PV sources	Medium	User privacy restrictions	[105]
User electrical equipment adjustment	Adjust the power of electrical appliances according to the grid voltage	Medium	Performance limitations of household appliances	[106]
Photovoltaic and charging pile collaborative technology	Energy storage stabilizes the fluctuations in photovoltaic power	High	The cost is slightly high	[107,108]
Distributed energy storage, OLTC, and SVR coordination	The wide-range operating characteristics of the transformer voltage	Medium	Coordinate among multiple parties	[109]
STATCOM operates flexibly based on the information from the electricity meter	Reactive power compensation	High	Certification of management methods by the power grid	[110]
Distributed power cluster regulation based on electricity meters	Reactive power compensation	High	User privacy restrictions	[111]
Phase reconfiguration voltage regulation method	Phase reconfiguration	High	The effect of overloading LVDA is limited	[112,113]
Distributed power supply control based on the probability of voltage imbalance	Phase reconfiguration	Medium	User privacy restrictions	[114]
Regulation of individual photovoltaic panels based on voltage state	MPPT technology for photovoltaic panels	Low	The diversity of equipment manufacturers and user privacy	[115]
LVDA power adjustment based on photovoltaic power regulation	Distributed secondary control	Low	The location and output of distributed power sources in LVDA are variable	[116,117]
Power flow adjustment of LVDA	Electric power flow theory	Medium	LVDA requires the extensive deployment of voltage and current sensors	[118]

shows the comparison of different voltage and power regulation methods of LVDA.

For voltage regulation, Ref. [105] introduced a novel joint coordination method for LVDA operation, leveraging both local measurements and global consensus to enable effective voltage adjustment. Ref. [106] suggested a voltage regulation method for LVDAs that involves curtailing end-user appliance power consumption. Ref. [107] developed a centralized coordination strategy for adjusting LVDA voltage by aggregating data from electric vehicle charging piles and photovoltaic systems. Ref. [108] examined the influence of PV–battery storage systems on LVDA voltage, identified conflicts between user-side energy storage operation and grid voltage control, and proposed a decentralized voltage optimization strategy leveraging user-level storage systems. Ref. [109] introduced a voltage optimization scheme for LVDAs that coordinates distributed storage systems, OLTCs, and step voltage regulators (SVRs). Ref. [110] presented a control framework wherein smart meters are integrated with STATCOM systems to enable LVDA voltage regulation. Ref. [111] developed a new voltage control architecture for LVDAs that adjusts DG output based on smart meter data, achieving both voltage optimization and effective mitigation of phase unbalance. Ref. [112] proposed a voltage unbalance estimation and regulation technique based on local observations from phase-reconfiguration devices (PRDs), eliminating reliance on smart meter infrastructure. Ref. [113] introduced a voltage balancing approach by reassigning phase connections for single-phase customers. Ref. [114] developed a probabilistic control framework for mitigating voltage unbalance in LVDAs, dynamically adjusting DG output using smart meter feedback and outperforming conventional phase-switching devices in both responsiveness and accuracy. Ref. [115] advocated the use of IoT technologies to integrate PV panels into the grid’s control scope, enabling voltage regulation by modulating PV generation capacity.

Regarding power regulation in LVDAs, Ref. [116] used the Maltese grid as a case study to examine how the magnitude of PV output affects the operational state of LVDAs. Ref. [117] proposed a secondary control scheme for distributed energy resources that enables LVDA power regulation under communication-free conditions. Ref. [118] introduced an LVDA power flow optimization strategy based on coordinated active and reactive power control, which significantly enhances both voltage magnitudes and voltage balance profiles.

2.5. Aggregation Technologies for Distributed Energy Resources in LVDA Digital Twin Systems

Distributed resource aggregation in LVDAs mainly encompasses two types: aggregation of distributed generators and aggregation of electrical loads. Ref. [119] reviewed the key features of current DG and load aggregation approaches and introduced a standardized framework and methodology for integrated aggregation. Ref. [120] conducted a time-segmented analysis of smart meter readings and identified 10 characteristic features that facilitate load aggregation. Ref. [121] explored how load clustering techniques are applied to transformer health diagnostics and electric vehicle charger safety control. Ref. [122] presented a load classification and clustering analysis approach based on deep reinforcement learning, designed to enhance non-intrusive load monitoring (NILM) capabilities. Ref. [123] introduced a user load curve association technique that automatically aggregates load curves into representative profiles, thereby eliminating the subjectivity introduced by manual classification.

2.6. Emerging Devices and Terminals for LVDA Digital Twin Applications

Against the backdrop of rapid advancement in LVDA digital twin technologies, a new wave of software and equipment is emerging, among which intelligent agents are taking shape as core digital infrastructures. Ref. [124] introduced a service agent tailored for EV charging stations, capable of precisely monitoring user charging time and load profiles, thereby enhancing user information management within the power grid. Ref. [125] developed an intelligent distribution panel that integrates diverse IoT devices and connects with multiple electrical components, facilitating LVDA control and signaling a future trend in the evolution of smart meters and gateways within LVDA systems.

3. Development Goals and Practices of the ‘6C’ in LVDA

Section 2 analyzes the research characteristics of LVDA digital twins. Existing research has carried out rich work in aspects such as LVDA data measurement, data processing, flexible control, specific business applications, and core devices. However, the current research is relatively scattered, lacking a development path plan for LVDA digital twin technology and evaluation methods for different development stages of digital twin LVDA, making it difficult to provide guidance for the further development direction of LVDA. Therefore, this section will propose the LVDA digital twin ‘6C’ development goals and describe in detail the connotations of the ‘6C’ goals.

3.1.‘6C’Development Goals for LVDA Digital Twin

The previous section analyzed and summarized the research status of Digital Twin in LVDA. Based on the classification of existing research, this article summarizes the development goals of the ‘6C’ development goals in LVDA, which are called Clear Measurement, Cognizable State, Controllable Devices, Configurable Parameters, and Combinable Resources, Coordinated Autonomy. The ‘6C’ technology development goals represent the deep application and extension of the “smart grid” technology concept in LVDA and represent the overall development trend and development connotation of LVDA at different stages. The detailed introduction of the ‘6C’ goals is as follows:

Clear Measurement. At this stage, a massive number of multi-type sensors are applied to the distribution area, and station area data is collected as comprehensively as possible. The operational data of key elements in the distribution area, such as distributed generation (DG), distribution lines, and user side loads, are fully accessible to the grid, including obtaining node voltage, branch current, equipment temperature, switch open/close status, and user power outage information.

Cognizable State. At this stage, the operational status of the distribution area is fully revealed. Based on comprehensive visibility of distribution area data, multi-dimensional state indicators that capture the core operational characteristics of the distribution area can be fully computed. These indicators are deeply integrated with the diversified operations of operation, distribution, and dispatch, reflecting multi-dimensional characteristics such as economy, stability, and safety. Key distribution area metrics include power quality metrics, power outage status indicators, and electricity theft risk indicators.

Controllable State. All equipment in the distribution area is subject to firm control by the higher-level power authority. The authority uses circuit breakers or control signals to control equipment start–stop and adjust operation modes, such as fault zone isolation and PV inverter start–stop control.

Configurable Parameters. All equipment with flexible regulation capability in the distribution area can accept parameter adjustments from the higher-level power authority. The authority uses control signals to modify operating parameters, such as PV inverter power output and adjustment of flexible load power consumption.

Combinable Resources. All DG and loads in the distribution area can achieve multi-timescale, wide-area dynamic aggregation, enabling source-load resources to be flexibly integrated according to higher-level dispatch requirements. Typical applications include centralized dispatch of distributed renewable energy for grid integration, flexible load aggregation management, and other scenarios.

Coordinated Autonomy. The distribution area can autonomously assess its operating status and independently adjust its operation to achieve energy balance and a self-consistent state. The autonomy of the distribution area encompasses all the first five functions of the ‘6C’ goals, representing an organic and flexible integration of these functions. It enables capabilities such as steady-state optimization, rapid fault clearance, and self-healing through automatic fault isolation and restoration.

The ‘6C’ development goals represent the LVDA digital twin development goal and basic development path. It is proposed by the authors based on the actual situations of LVDA. The “6C” development goals are similar to the evaluation of a car’s autonomous driving capability using the four levels of L1 to L4 in the field of autonomous driving [126]. At present, ideas similar to “6C” have already been put forward in China’s power grid. State Grid Corporation of China and China Southern Power Grid Company are committed to achieving the “4 Can” goals of distributed power sources, namely “can be seen”, “can be known”, “can be controlled”, and “can be dispatched”. Based on the review of existing research and in combination with the idea of the “4 Can” goals, this article summarizes the “6C” development goals of the LVDA digital twin.

The 6C goal is highly consistent with the development ideas and concepts of smart grids in many regions around the world. Existing research has discussed the development direction of low-voltage power grids under the current grid structure and information technology framework, including the typical application of IEC international standards in power grids in regions such as North America and Europe [27,127,128,129]. Existing research indicates that distribution, digitalization, and intelligence will be the mainstream development directions and characteristics of LVDA in the future. Communication connections and data interactions among LVDA will become increasingly frequent. Reference [130] proposes the main applications of data analysis in the future smart grid based on existing standards and puts forward the operation framework of the smart grid. In the reference [131], based on the development trend of the smart grid, operation and control strategies of the low-inertia system in the virtual power plant mode are summarized. References [132,133] point out the multi-layer management mechanism of the smart grid under the IEC 61850 framework, covering multiple aspects such as state perception and operation control. The above research is mainly based on the IEC standard architecture and has made summaries and improvements in the aspects of grid operation status perception, distributed resource aggregation, and efficient and flexible system control. These are consistent with the development ideas of the first five aspects of the 6C goals proposed in this paper. However, the architecture of the above research mainly relies on the “cloud-edge” interaction approach to manage LVDA. With the increase in distributed power sources, the traditional “cloud-end” interaction management model of the power grid is facing significant challenges, and power grid companies find it difficult to manage a large number of distributed power sources efficiently. The emergence of technologies such as deep learning and artificial intelligence will effectively make up for this deficiency [134]. By adding edge-side artificial intelligence devices to manage individual LVDA or individual microgrids, the energy balance of LVDA can be achieved, reducing the burden on power management departments. This is a feasible technical approach. Considering this development trend, the idea of “coordinated autonomy” of LVDA was proposed in the last aspect, which is also a key point that significantly distinguishes this paper from existing technologies. The current standards for digital power grids mainly focus on interactive connections in communication, power, and other aspects, emphasizing the interconnection and data transfer [135]. Our ‘6C’ development goals elaborate on the different development stages of LVDA’s digital transformation, with the ultimate goal of achieving the autonomy of the station area. The emphasis is on enhancing the autonomous perception and control capabilities of the LVDA regarding its operational status.

3.2. Typical Practice Work of Provincial-Level New Distribution System Station Areas Based on the ‘6C’ Technical Development Goals

Guangxi Power Grid Company is implementing the ‘6C’ technical development goals to address the construction needs of new distribution system station areas at the provincial level. Key innovative initiatives include:

(1) The integration of data sources and equipment in the marketing, distribution, and dispatching business of LVDA. The integration of data sources and equipment in the marketing, distribution, and dispatching business of LVDA refers to the installation of only one edge intelligent unit within a LVDA, achieving” one LVDA, one terminal “while meeting the business needs of production, marketing, and dispatch departments, and reporting data to marketing, distribution, dispatch, and other departments to ensure” one data source “. Business data can be exchanged on the terminal side as needed, supporting business scenarios such as power outage and restoration perception, power quality monitoring, lean line loss management, and power metering in the distribution area, enhancing low voltage distribution network status perception, substation transparency management, and flexible scheduling of distributed resources. The diagram of the integration of data sources and equipment in the marketing, distribution, and dispatching business of LVDA is shown in Figure 2. The edge intelligent unit is equipped with powerful communication capabilities and features multiple communication interfaces such as high-speed power line carrier communication (HPLC), RS485, LORA, NFC, 4G/5G, Ethernet, and Bluetooth. The edge–edge intelligent unit can achieve information interaction with different end-side devices through multiple communication interfaces. The communication objects include electrical equipment such as smart circuit breakers, distributed energy storage, low-voltage branch topology monitoring devices, etc. The communication objects also include non-electrical quantity sensors, such as a water immersion sensor and a water level sensor. The edge intelligent unit can interact with the cloud platform for information exchange via 4G/5G/Ethernet, enabling LVDA data upload and receiving control instructions from the upper layer. Typical cloud platforms include provincial-level power dispatching platforms, provincial-level whole-region Internet of Things, metering platforms, production operations support systems, and energy internet platforms. The edge intelligent unit enables the data flow and business flow between the cloud, edge, and end to flow with each other, achieving good information interaction among the cloud platforms, edge intelligent unit, and terminals. This mode enables the edge intelligent unit to more flexibly regulate the status of the end-side terminals, facilitating the flexible energy flow between the end-side and the edge-side devices (or the upper-level branch lines of the LVDA).

The ‘integration of data sources and equipment in the marketing, distribution, and dispatching business’ is a new model of power grid operation. It demonstrates the significant impact of LVDA’s digital twin technology on the way power grids operate. As can be seen from the figure, after the LVDA data is collected by the edge-side computing unit, it is uploaded to multiple platforms, such as the Internet of Things platform, metering platform, and dispatching platform. The data collection time and sampling rate of these several platforms are exactly the same. Before the implementation of LVDA digital twin technology, the data of these platforms were isolated from each other, and the time and rate of data collection were different. It was difficult to determine valid data through multiple platforms. The “integration of data sources and equipment in the marketing, distribution, and dispatching business” effectively avoids the problem of cross-departmental data interaction. This method will bridge the data silos between various departments of the power grid, effectively improving the comprehensive visibility of data, laying the foundation for the follow-up work of the ‘6C’.

(2) Cloud–Edge–End Device Collaboration. To achieve efficient collaboration among cloud–edge–end devices, the Electric HarmonyOS IoT Operating System (E-HarmonyOS) and flexible development platform are being promoted: (a) E-HarmonyOS is an IoT operating system independently developed by Southern Power Grid Corporation for the power industry. It supports interoperability across different brands and types of power equipment, enabling plug-and-play integration and massive data connectivity between cloud–edge–end devices. (b) Flexible Development Platform: Based on cloud orchestration technology, this platform is a modular tool for developing and deploying business applications (apps). Traditional distribution area apps are typically pre-installed by manufacturers in smart terminals/gateway devices, but face challenges such as poor cross-device compatibility, prolonged development and debugging cycles, and high operational maintenance costs. In the new provincial-level distribution system, the large-scale deployment of LVDA and the dynamic nature of distribution network operations make it difficult to meet the rapid development demands for cloud–edge–end coordination due to the low efficiency of traditional app development. The cloud–edge–end device collaboration serves as a core foundational element for implementing the ‘6C’ technical development roadmap in LVDA. It focuses on achieving comprehensive visibility and a deep understanding of LVDA status, while driving the realization of an efficient maintenance system with the goals of “unified business workflows” and “single-terminal operations”.

(3) Construction of a real-world test environment for “integrated source-end fusion”. To simulate the operational conditions of LVDAs planned for “integrated source-end fusion” technology upgrades, a practical test platform is being built in the real-world distribution network environment. This real-world test environment will provide experimental support for the ‘6C’ technical development roadmap at each implementation stage, enhance the quality of on-site installation and commissioning for diverse equipment, and achieve the goals of “real-time quality control + pre-commissioning + modular installation + plug-and-play maintenance”. It will reduce on-site installation costs and improve the efficiency of LVDA transformation.

(4) Large-scale demonstration construction based on digital power supply stations of LVDA. To implement large-scale demonstrations of digital twin LVDA, digital power supply stations are being constructed. As the most fundamental unit of the power grid, power supply stations handle critical tasks, such as fault repair and equipment inspection, and create rich application scenarios for the LVDA digital twin. However, the fragmented professional systems within these digital power supply stations lack a centralized management platform and intelligent support tools. The architecture of the substation digital power supply station is illustrated in Figure 3. The photovoltaic inverters, smart switches, and other facilities of different consumers can achieve information exchange with the edge intelligent units through their respective protocol adapters. Protocol converters typically use radio frequency (RF) communication and power line carrier communication (PLC) to achieve the conversion between different communication protocols. The edge intelligent unit acquires the global status information of the LVDA by interacting with the protocol adapters of different users, laying the foundation for further regulation of the LVDA.

In the digital power supply station of the LVDA, the cloud platform stores all the business data and management data of the LVDA, and the edge intelligent units collect various types of data from distributed power sources and smart electricity meters. By integrating analysis software within the cloud platform and the edge intelligent units, establishing the correlation between massive data and the operational status of LVDA, and through information exchange and logical optimization between the cloud platform and the edge measurement platform, status perception, regulation, and other tasks can be flexibly completed within LVDA. After the data is uploaded to the cloud platforms, all the LVDA business in the digital power supply station becomes clearer and more standardized. Most LVDA services can be presented in the form of applications (Apps), and the status of different business Apps can be queried on the cloud platforms and the edge intelligent units. In the completed digital power supply station of the LVDA, tasks such as three-phase voltage unbalance monitoring, overvoltage monitoring, and LVDA topology identification have been realized, which cannot be achieved in the traditional LVDA. The measurement data in the actual LVDA indicates that the user’s average power outage fault detection time is less than 1 min, while in the traditional LVDA, this time is over 10 min. In addition, the digital power supply stations in the transformer areas that have been demonstrated have basically achieved full visibility and are rapidly developing towards a method that is fully knowable, fully controllable, and adjustable as needed.

4. Discussion of Future Work

From the existing research, it can be seen that the LVDA digital twin has made great progress, and is particularly prominent in LVDA state sensing, system regulation, and distributed resource aggregation, but the existing research is still mainly based on AMI data to achieve LVDA state perception and regulation, ignoring the extensive linkage relationship of LOTs devices within LVDA. With the deepening integration of IoT technology and LVDA, new technologies and equipment for LVDA will continue to emerge, which will promote LVDA performance continuously. This section summarizes some future research work in response to this development trend, as follows:

(1) In terms of LVDA state estimation, the existing technology only conducts state estimation based on the metering data of a single smart electricity meter. With the advancement in LVDA digital twin technology, the integration of diverse sensors with new end-side devices has become inevitable, such as non-intrusive sensors and new reactive power compensation devices. By integrating data from multiple types of sensors, a vast number of end-side devices, and smart meters, the accuracy of LVDA status perception will be enhanced. This improvement will mainly be reflected in two aspects. On the one hand, the data from smart meters can be used to correct the measurement data of other smart distributed devices to determine the aging degree of key LVDA equipment. On the other hand, smart meters have a certain degree of aging, and the operation data of new equipment can compensate for the measurement accuracy of electricity meters. The above two aspects have not been fully studied and will become the key directions for future research.

(2) In terms of LVDA state sensing, although existing research has carried out relevant work around topology identification, load and phase identification, power quality monitoring and other scenarios, it is mainly based on the credibility of the source of the station metering data, and the problems of erroneous and missing station data are prominent, and the existing technologies are less likely to carry out research on LVDA state awareness to address this problem. In addition, existing LVDA state awareness techniques are mainly based on big data clustering and artificial intelligence, mostly using historical data as model training samples, which have the problems of a long model training cycle and difficulty in migrating a single LVDA model to other LVDAs. Therefore, universal and lightweight state-aware models are the focus of LVDA research in the future.

(3) In terms of LVDA distributed resource aggregation, the existing research focuses on load behaviour identification and operation curve analysis. In the future, the implementation of ‘virtual power plant’ technology in LVDA is an important development direction. However, the current research has not discussed the aggregation method of load behaviour in different time scales, and the research on aggregation considering the operation characteristics of different distributed power sources has not been discussed, which seriously affects the effect of state regulation of LVDA. In terms of time scale, future research will focus on establishing second-level, minute-level, and hour-level dynamic aggregated response curves for LVDA new energy units, respectively meeting the demands of transient control, energy trading, power grid regulation, and other aspects. In terms of spatial scale, the dynamic aggregation and trading of multiple LVDA will be an important research direction in the future, and a balance needs to be struck between power peak shaving and user rights.

(4) In terms of the LVDA regulation field, the existing research mainly focuses on the voltage and power regulation strategy of LVDA; however, the existing research ignores the impact of the regulation strategy on the security and stability of the grid operation, and it cannot provide a sure and effective basis for the regulation behavior. This leads to the fact that, in many cases, although relatively good digitalization has been achieved in LVDA, the phenomenon of regulatory failure still occurs from time to time. Although solving the problem of on-site regulation failure still relies on manual on-site operation, it is obviously inefficient for power workers to check each issue one by one. On the other hand, most of the existing regulation methods still rely on the feedback information from smart meters. The closed-loop feedback link is relatively long, and the degree of utilization of the information from multiple sensors is low. Therefore, the security evaluation method of the LVDA regulation strategy, LVDA predictive control based on multi-source data fusion, and other technologies will be one of the future directions of LVDA research.

5. Conclusions

The main conclusions of this paper are as follows:

(1) The access to a high proportion of new energy sources and a high proportion of power electronic devices make the LVDA power balance and safe and stable operation problems prominent. Realizing the digital transformation of the LVDA is the key to solving such problems. Digital twin technology will transform the operation mode of traditional LVDA from aspects such as measurement, perception, and control, facilitating the automation of the data collection and summary process, the intelligence of status perception, and the unmanned operation control.

(2) Although LVDA has basically realized the measurement of LVDA electrical quantity by installing smart meters, smart sensors, and other smart devices, the perception of the LVDA digital twin state is not completely clear. Accurately estimating the operational status of LVDA based on multi-type heterogeneous and multi-source data is the main difficulty and challenge currently faced. In addition, the aggregation methods of multiple types of micro-sources at multiple time scales and multiple spatial scales in LVDA have not been fully clarified, and the joint regulation technology of switchgear, power electronic equipment, and other devices is still in the initial exploration stage. In the future, maximizing the data efficiency of LVDA’s massive devices will be the focus of research.

(3) The ‘6C’ development goals define the basic path of LVDA digital transformation. The ‘6C’ development goals are dedicated to achieving the autonomous collaboration of the LVDA, which is a key factor that significantly distinguishes it from existing technologies. The connotation of the ‘6C’ development goals will be continuously expanded with the further deepening of LVDA digital transformation.

(4) The ‘integration of data sources and equipment in the marketing, distribution, and dispatching business’ and ‘digital power supply station has changed the traditional ‘cloud-end’ operation mode of LVDA. By proposing the edge intelligent unit, this model was transformed into a ‘cloud–edge–terminal’. The proposed edge intelligent unit will undertake most of the work of LVDA status perception and operation control in the future.

(5) A large number of ‘integration of data sources and equipment in the marketing, distribution, and dispatching business of LVDA’ and ‘digital power supply station’ project will continue to promote the practical work of LVDA around the ‘6C’ goals in the future.

(6) The digital transformation of LVDA will change the operation logic and operation mode of the traditional grid’s marketing, distribution, dispatching, and other business departments, and the grid will be transformed from a single power supplier to a power data service provider.