The Influence of Decentralized Charging Pile Connection on Voltage Deviation of Distribution Network

In this study, a mathematical model was developed to describe the charging process of a charging pile. The PWM charging process was determined, and a mathematical model for the DC-DC output of the charging pile was constructed. Prior to the photovoltaic connection, the voltage distribution of the distribution network was determined, and the line voltage drop was obtained. The impact of the access capacity of distributed charging piles on the voltage of the distribution network was analyzed. Furthermore, the line voltage drop between nodes was determined, and the nodes were balanced to calculate the voltage of each node at this time. This analysis enabled the investigation of the impact of distributed charging pile access on the voltage deviation of the distribution network. The results indicate that the calculation error of voltage deviation in the distribution network is only 0.002, with a maximum prediction accuracy of 99.0%.


INTRODUCTION
In order to cope with the development of new energy vehicles, the supporting infrastructure system of EV charging is also constantly improving [1] .Charging piles have become an important load in urban power supply, and charging stations composed of dozens of charging piles can be seen everywhere.When a large number of electric vehicles are charged at the same time, it will lead to a rapid change in power supply load.If a large number of charging piles are connected to the power grid at peak times, the peak-valley difference of power supply load in the region will be further aggravated on the basis of the original, and the distribution network [2,3] will be overloaded when the situation is bad.In addition, it is more and more difficult to forecast load and dispatch power in advance.In order to reduce the pollution to the ecological environment, countries are vigorously developing the photovoltaic industry and put into use.In recent years, China's solar energy industry has developed rapidly, and the annual installed capacity may continue to increase.However, with the gradual maturation of the industry, solar companies urgently need independent innovation.Therefore, relevant researchers have made some progress in the voltage study of charging piles connecting to distribution networks [4,5] .
In view of the above problems, this paper proposes a new method to analyze the influence of decentralized charging pile access on the voltage deviation of the distribution network.

Structure analysis of electric vehicle charging station
An electric car charging station comprises a power supply and distribution system, a vehicle charging system, a charging monitoring system, and supporting facilities [6] .When multiple charging piles of varying types are connected to the power grid for charging electric vehicle batteries simultaneously, the system consumes a considerable amount of reactive power, resulting in a decrease in terminal voltage and a decline in power quality.
As the core foundation to promote the further development of the electric vehicle field, the charging pile is an important guarantee for the realization of the electrification transformation of the global automotive industry.The charging methods of electric vehicle batteries can be divided into three types: conventional charging, rapid charging, and mechanical replacement of batteries according to different charging times; According to different charging locations, it can be divided into household charging facilities and centralized charging stations; According to different connection modes, it can be divided into non-contact charging and wired charging.

Construction of mathematical model for charging process of decentralized charging pile
After simplifying the PWM charging process, the charging process circuit of the decentralized charging pile is shown in Figure 1.

Simplified circuit of DC charging pile charging process
In the ideal state, the simplified dynamic mathematical model of the rectifier part of the DC charging pile is: where p is a differential operator; f R is the equivalent resistance of DC-DC conversion circuit.The resistance value can be approximately calculated by the law of conservation of energy as 2 ( ) ( ) where l R is the charging load; η is the efficiency of the charging equipment.In fact, η exists as a variable in the whole charging operation process, but because it changes little, it is often treated as a constant.
The simplified mathematical model of the DC-DC output end of the charging pile is where D is the duty cycle of the power change circuit.
In the ideal state, the mathematical model of the whole charging pile can be obtained by combining Formulas (1), (2) and (3).as load branches, the charging process of the automotive battery can be equivalent to that shown in Figure 2.Among them, since the equivalent resistance and inductance values in the whole charging model circuit are very small relative to the capacitance values, they can be ignored.Therefore, from the equivalent charging model, it can be seen that the charging process of the battery load is capacitive.

Equivalent simplified circuit of the charging process
Under the given charging voltage, after connecting any charging device, it is It can be seen from Formula (4) that f (5) shows that under the mathematical model of the whole charging process, the load is capacitive.From this model, it can also be concluded that the electric vehicle charging pile will consume the system reactive power in the process of charging the battery.

Voltage calculation model of distribution network with decentralized charging piles (1) Voltage distribution before connection
The relationship between m U and 1 m U − can be obtained from the power flow direction before the decentralized charging pile is connected to the radial distribution network.
In the formula, 1 m U − Δ and 1 m U δ − respectively represent the longitudinal and transverse components of voltage drop.The calculation formula for voltage drop is: We construct a calculation expression for voltage difference: where Li P represents active power, and Li Q represents the reactive power of the same line.The calculation formula for node voltage at this point can be expressed as: (2) Voltage distribution after connection After connection, the voltage difference between the two nodes is: The impact of decentralized charging station connection on voltage can be obtained through the above calculation.The specific impact analysis is as follows.
(1) If the capacity of the distributed charging pile is low, the line voltage increases when the decentralized charging pile is connected compared to when it is not connected.As the connection point gets closer to the end of the line, the voltage drops gradually.
(2) If the access capacity exceeds the load capacity of all locations beyond the access point, including the multi-load capacity of that access location, but is less than the total capacity of all node loads, the line voltage will initially drop, then rise, and finally drop again as the access location approaches the end of the line.
(3) When the access capacity exceeds the total load capacity of nodes, the decentralized charging pile distributes the required load to each node and returns any excess power to the grid.This creates power reverse transmission, causing the line voltage to rise and then fall as the access location moves towards the end of the line.
Due to the presence of multiple lines connecting the node m and the balance node, voltage drops occur.Consequently, the voltage level at node m can be expressed as: The voltage calculation formula for the point p is:

Impact analysis results
After the decentralized charging pile is connected to the distribution network, the voltage of the distribution network is affected as follows.
(1) The integration of distributed photovoltaic systems can impact the normal voltage distribution in traditional radial power grids.Normally, the flow direction is singular, resulting in a single voltage drop from the source node to the end of each branch.However, distributed photovoltaic systems add negative power loads to each node, reducing the total power transmitted to the feeder.Additionally, some generators may transmit reactive power back into the system, creating voltage at certain nodes in the distribution network.As the network experiences a rise in voltage, factors such as the capacity and location of the distributed solar system, as well as the load capacity of each node, can influence the increase at each node.
(2) There are various factors that can cause voltage fluctuation and flicker.In a traditional distribution network, the proximity of a node to the edge of the feeder tends to result in greater voltage fluctuation.The integration of distributed photovoltaic systems can worsen this issue due to the uneven output power.If the absorbed energy causes the voltage to fluctuate in the positive direction of the solar cell's power generation, it can help mitigate load changes.However, if the voltage fluctuates in the opposite direction, then the voltage fluctuation becomes more severe.External factors, such as sudden changes in light or unexpected starts/stops in the photovoltaic production process, can have a significant impact on the output power of the system, leading to voltage flicker.
For all areas of the power industry, it is difficult to achieve balanced power consumption, and the problem of peak valley difference is always inevitable.Generally, the time and place when users charge automotive batteries are relatively scattered, which can be predicted.However, for users who choose to use special vehicle charging stations, the centralized charging behaviour of large-scale NESP-2023 Journal of Physics: Conference Series 2592 (2023) 012095 IOP Publishing doi:10.1088/1742-6596/2592/1/0120955 vehicles at the same time and place has further increased the peak valley difference of power consumption.Electric vehicle charging behaviour can have a significant impact on the power grid, particularly during peak evening hours.When many users simultaneously charge their electric vehicles, it can cause a concentration of energy demand in a specific region, resulting in an increase in regional load.This increase in load can lead to a decrease in voltage and an increase in reactive power loss, which can ultimately affect the stability of the power grid.If the natural power factor of the load is too low, it can exacerbate the issue and lead to increased active and reactive power losses throughout the entire system.This can result in lower efficiency and higher costs for electricity providers, as well as potentially impacting the reliability of the power supply.In addition to these issues, connecting electric vehicle charging equipment to the power grid can also result in voltage drops at line nodes, particularly at terminal voltage.These voltage fluctuations can disrupt normal equipment operation and potentially damage sensitive electronic devices.To address these challenges, electricity providers are exploring various solutions, such as smart charging systems that can distribute the energy demand more evenly throughout the day, as well as implementing advanced monitoring and control systems to detect and mitigate voltage fluctuations.Additionally, improving the natural power factor of loads can also help to reduce reactive power losses and improve the stability of the power grid.
When it is reduced to an irreparable degree, it even causes voltage collapse and large-scale power failure.Since the voltage amplitude of the power grid is mainly affected by reactive power, when the reactive load continues to increase, new reactive power sources can be considered to offset part of the load consumption.According to China's power grid standards, the maximum voltage offset of 10KV voltage level is 7% for the power system in normal operation.If the system operates beyond the maximum range for a long time, reactive power compensation must be adopted for voltage adjustment.

Preparation Before Experiment
The IEEE-33 bus system is used in the example simulation, and the reactive power output of DG in the system will not exceed the upper limit.The capacitor bank will also meet the upper limit of switching times.The six nodes (10, 11, 13, 17, 24, 32) of the system are respectively connected with fans, two sets of capacitor banks, a photovoltaic power supply and two sets of static var compensators (SVCS).The basic parameters are as follows: the active capacity of DG is 1.5MW, the reactive capacity of the two sets of static var compensators is 1Mvar, each group of the two sets of capacitor banks has 20 capacitors, and the capacity of a single group is 25kvar.

Calculation error
To assess the effectiveness of this method, we compared it with the methods proposed in [4] and [5]  and evaluated the distribution network voltage deviation error.The results of these comparisons are presented in Table 1.Additionally, it is worth noting that the distribution network voltage deviation error is a crucial metric for evaluating the performance of any voltage control method.Therefore, the comparison of our method with existing approaches using this metric provides a comprehensive understanding of its effectiveness.
Table 1.Calculation error results of different methods

Node serial number
Calculation error [4] method [5]  The analysis presented in Table 1 indicates that for a single node, the calculation error of the method in [4] is 0.632, the calculation error of the method in [5] is 0.673, and the calculation error of the method proposed in this paper is only 0.002.For three nodes, the calculation error of the method in [4] is 0.713, the calculation error of the method in [5] is 0.958, and the calculation error of the method proposed in this paper is 0.012.Finally, for six nodes, the calculation error of the method in [4] is 0.956, the calculation error of the method in [5] is 0.891, and the calculation error of the method proposed in this paper is 0.018.These results demonstrate that the proposed method consistently achieves low calculation errors when estimating distribution network voltage deviation.Therefore, it can be concluded that this method effectively improves the accuracy of distribution network voltage deviation calculation.

Prediction accuracy
To validate the effectiveness of the proposed method, we compared it with the methods presented in [4]  and [5].Table 2 summarizes the results achieved by each method in terms of the accuracy of distribution network voltage deviation prediction.According to Table 2, in 100 experiments, the error prediction accuracy of the [4] method was 65.2%, the error prediction accuracy of the [5] method was 70.5%, and the error prediction accuracy of the method in this paper was 95.2%.In 300 experiments, the deviation prediction accuracy of the [4]  method is 72.9%, that of the [5] method is 72.1%, and that of the method in this paper is 99.0%.The prediction accuracy of the deviation of the proposed method is always high, which indicates that the method can effectively improve the prediction effect of the voltage deviation of the distribution network.

CONCLUSION
This paper presents a new method for analyzing the impact of decentralized charging piles on voltage deviation in distribution networks.The approach involves analyzing the charging pile principle, determining the PWM charging process, and developing a mathematical model for the DC-DC output of the charging pile.Additionally, the voltage distribution of the distribution network is determined before photovoltaic access, and the line voltage drop is obtained.The voltage deviation after connecting the decentralized charging pile to the distribution network is calculated, and the influence of the charging pile access capacity on the distribution network voltage is analyzed.The voltage drop of the line between the node and the balance node is determined, and the node voltage is calculated at this point, enabling the impact analysis of decentralized charging pile access on the voltage deviation of the distribution network.Experimental results demonstrate that the proposed method has a calculation error of only 0.002 and a maximum prediction accuracy of 99.0% for voltage deviation.These findings demonstrate the effectiveness of the proposed method in improving the overall performance of voltage deviation prediction in distribution networks.

)3
If all relevant components of the DC-DC conversion circuit on the right side of f C are regarded

Table 2 .
Prediction accuracy results of different methods