Research on Dispatching Coordination Control Method of Distribution Network with Photovoltaic Power Generation

The coordination control of photovoltaic power distribution networks faces challenges such as inaccurate load power prediction and long coordination times. To address these issues, this paper proposes a scheduling coordination control method for photovoltaic power distribution networks. The method involves analyzing the structure of the photovoltaic power generation system, obtaining the operation state of the network using disturbance observation methods, building a dispatching coordination control model, designing constraints for dispatching coordination control, building a power balance function, and solving the total demand power of the load. An algorithm flow is designed to optimize the coordination control of photovoltaic distribution networks. Experimental results demonstrate that the accuracy of photovoltaic unit power prediction using this method reaches 98%, with power grid dispatching and coordination control times always less than 3.8 seconds. This approach effectively improves the effectiveness of distribution network dispatching and coordination control.

power generation [3].However, air pollution, environmental degradation, and other survival problems have become a common global challenge faced by mankind, which cannot be solved by one person or one country.They explore environmentally friendly energy systems and achieve the goal of carbon peak and carbon neutrality, which have increasingly become the common consideration of all countries in the world [4].In order to solve this challenge and make up for the shortcomings of traditional power distribution systems, active distribution network technology, including distributed power, is proposed.
Chen et al. put forward based on the improved method of a second-order cone with photovoltaic power distribution network optimization scheduling method [5], the fuzzy sets method for photovoltaic power load data, according to set pair analysis to obtain the active load data type, modified iteration method via the boundaries of the second-order cone photovoltaic power load scheduling optimization, this method can effectively enhance photovoltaic power dispatching.However, the dispatching time of the photovoltaic distribution network is high.Zhang et al. put forward by photovoltaic power distribution network optimization scheduling method based on load demand response [6], according to the number of twin technology for pv power load data, using such as micro rate criterion for photovoltaic power transient stability, a cost-benefit analysis method is adopted to define the cost of photovoltaic power distribution network economy, according to the load demand response photovoltaic power network dispatching.This method can improve economic cost efficiency, but the accuracy of load power prediction is poor.
A new coordinated dispatch control method for power grids is proposed to improve the operational stability of distribution networks in response to the problems existing in existing control methods.

Analysis of photovoltaic power distribution network
Solar energy is a renewable, pollution-free energy.Energy is huge.In a sense, it is also inexhaustible.The development of solar energy can effectively alleviate the current situation of global energy shortage.Photovoltaic power generation can directly convert solar energy into electric energy, can take advantage of the advantages of building, fully absorb sunlight, and is not affected by the region, has a very wide range of application markets.At the same time, the national goal of carbon peak and carbon neutrality promotes the research and development of photovoltaic power generation technology [7].According to statistics from the Qian Zhan industry research institute, a photovoltaic industry association, by 2020, China's cumulative installed PV capacity had reached 253GW, with 48.2GW of new installed capacity.Among them, centralized installed capacity accounted for about 70 percent of the new installed capacity.
Photovoltaic systems rely on the photoelectric effect of semiconductors to generate electricity [8].This effect causes electrons to absorb light energy and escape from the surface of a metal.In a solar cell, a potential difference exists between the P and N poles.When sunlight shines on the solar panel, holes move from the P pole to the N pole while electrons move from the N pole to the P pole, thereby creating a current [9].The output of photovoltaic power stations is affected by many factors, among which the weather has the most significant impact on the output of photovoltaic power stations.Similarly, the impact of different weather on the output of photovoltaic power stations is also different.The influence degree of different weather on photovoltaic power station output is shown in Table 1.  1 demonstrates that the output of a photovoltaic power station is primarily influenced by temperature and solar radiation, which accounts for over 80% of the impact.Air pressure, humidity, and cloud cover have a moderate effect on the output, with an influence degree ranging from 30% to 50%.The impact of wind direction and wind speed on the output of a photovoltaic power station is minimal, accounting for only about 20% of the effect.

Dispatching and coordination control model of photovoltaic power distribution network
Photovoltaic materials can generate electricity with the help of semiconductor materials with photoelectric conversion properties.By doping a suitable amount of phosphorus and boron in the crystal silicon, make the charge in the material is in a state of imbalance is the semiconductor material, and then through the two p-type and n-type semiconductor materials joining together form the p-n junction, then have the special electrical properties of semiconductor materials, if there is light on the p-n junction, will be formed between the p-n junction emf [10].The equivalent circuit model of the photovoltaic power distribution network is shown in Figure 1.The formula for calculating the volt-ampere characteristics is: where ph I is the photoelectric effect current; s I is the reverse saturation current of the diode; In order to optimize the output power of a photovoltaic power distribution network, it is essential to operate the network close to its maximum power point during normal operation.This approach can significantly enhance the photoelectric conversion capacity of the system and improve its overall performance.To achieve this, it is necessary to implement maximum power point tracking (MPPT) control for the network [11].One of the commonly used methods for MPPT control is the disturbance observation method.This approach involves constantly altering the operating state of the photovoltaic power distribution network and comparing the changes in output power and voltage before and after the state change.Based on this information, the MPPT circuit can be adjusted to ensure the photovoltaic power distribution network, which operates close to its maximum power point [12].By operating in close proximity to its maximum power point, a photovoltaic power distribution network can achieve a higher level of energy efficiency and generate more power from the same amount of sunlight.This can be particularly beneficial for systems installed in areas with limited access to sunlight or during times of low solar irradiance.Furthermore, it is worth noting that the characteristic expression of a photovoltaic power distribution network operating close to its maximum power point may vary depending on the specific system and its operating conditions.Therefore, it is important to carefully monitor the network's performance and adjust the MPPT control accordingly to ensure optimal operation.0 Right side of maximum power point / 0 Maximum Power Point 0 Left side of maximum power point dp dv where dp represents the output power change of the photovoltaic power distribution network in adjacent sampling periods; dv represents the output voltage variation of the photovoltaic power distribution network in adjacent sampling periods.During the whole tracking process, only voltage and current information need to be sampled.When the output power of the photovoltaic power distribution network hardly changes after the disturbance is applied, that is, the difference between two adjacent samples and the calculated output power tends to 0, it means that the photovoltaic power distribution network has tracked to the maximum power point.

DISPATCHING COORDINATION CONTROL OF THE DISTRIBUTION NETWORK,
INCLUDING PHOTOVOLTAIC POWER GENERATION When the types of distributed energy incorporated are too complex, defects such as excessive voltage offset of distribution networks and weak grid structure will still be exposed [13].An active distribution network is a kind of distribution network system that integrates and coordinates all kinds of distributed energy and has high flexibility of load and energy storage units.The active distribution network has the natural advantage of easy reconfiguration of network topology.It can optimize and manage the energy flow direction of the whole power system, so as to make the whole power system operate more safely and reliably.Distributed energy can provide a stable and safe energy supply to the power system under its strict control.Generally, the network structure of an active distribution network is a ternary combination mode: "source side" refers to various DGS; "Middle side" mainly includes voltage converters, power towers, relay protection, and other power devices; "Load side" refers to the adjustable flexible load in the system.
The combined structure of an active distribution network enables effective control of electric energy, monitoring, and coordination of various sudden working conditions.This eliminates the uncertainty of distributed energy and ensures the reliable operation of the system while meeting safety constraints [14].As active distribution network technology matures, economic benefits have become a new focus, especially with the introduction of peak valley TOU prices.Economic dispatching strategy is now a milestone for the development and progress of the technology.To ensure bus voltage safety and stability, this paper evaluates bus voltage quality using a voltage safety evaluation index as the basis for specific action reward value [15].First, we obtain the bus voltage k V at the first m time under time t .From this, the average voltage at time t can be calculated: Therefore, the deviation of the voltage at time t is: Furthermore, the voltage safety index at time t is: We take the j moments before the action, and the reward value of action a is set as: where 1 λ is the reward coefficient.
The global reward is set as: where 2 λ is the global reward coefficient; I is the number of all units; J is the working mode of each unit; ij φ indicates the operation status of each unit, In the formula, parameters 1 ξ and 2 ξ are weight coefficients.
The output power of each unit: where i P represents the output power of unit i ; min i P represents the minimum output power of unit i ; max i P represents the maximum output power of the i th cell.
State of charge of energy storage:

Predicted output of the photovoltaic unit
To evaluate the effectiveness of distribution network scheduling and coordinated control, including photovoltaic power generation, Figure 3 shows the power prediction results of photovoltaic units after implementing such control using three different methods.Based on the analysis shown in Figure 3, it can be observed that our proposed method achieves higher power prediction accuracy compared to the [5] and [6] methods for nodes 3, 15, and 33.Specifically, for node 3, the power prediction accuracy of the [5] method is 45%, while the power prediction accuracy of the [6] method is 54%.In contrast, our proposed method achieves a significantly higher power prediction accuracy of 90%.Similarly, for node 15, the power prediction accuracy of the [5] method is 45%, and the power prediction accuracy of the [6] method is 68%.However, our proposed method achieves an even higher power prediction accuracy of 93%.Moreover, for node 33, the power prediction accuracy of the [5] method is 65%, whereas the power prediction accuracy of the [6] method is 80%.Nevertheless, our proposed method achieves the highest power prediction accuracy of 98%.It is worth highlighting that our proposed method consistently outperforms other methods, demonstrating its effectiveness in power prediction.

Efficiency of distribution network dispatching coordination and control
In order to verify the efficiency of dispatching and coordination control of the distribution network, including photovoltaic power generation, the methods of [5], [6] and the method of this paper are used for dispatching and coordination control of the distribution network, and the results are shown in Table 2.
Table 2 According to the analysis presented in Table 2, the coordinated control time of the [5] method was 22.5 seconds when the number of experiments was 100.The coordinated control time of the [6] method was 32.9 seconds, while the coordinated control time of the method proposed in this paper was only 0.3 seconds.Similarly, when the number of experiments increased to 300, the coordinated control time of the [5] method was 78.2 seconds, the coordinated control time of the [6] method was 98.3 seconds, and the coordinated control time of the method proposed in this paper was only 1.2 seconds.Finally, when the number of experiments further increased to 600, the coordinated control time of the [5] method was 236.0 seconds, the coordinated control time of the [6] method was 286.1 seconds, and the coordinated control time of the method proposed in this paper was only 3.8 seconds.These results indicate that the proposed method exhibits significantly higher coordinated control performance thanthe methods.

CONCLUSION
Build the photovoltaic power distribution network scheduling coordination control model, design the constraints of photovoltaic power distribution network scheduling coordination control, build the power balance function of the photovoltaic power distribution network, solve the total demand power of load, and realize the optimization of photovoltaic power distribution network scheduling coordination control.The experimental results show that: (1) This method's power prediction accuracy of photovoltaic units can reach up to 98%.This method can effectively improve the scheduling and coordination control effect of distribution networks containing photovoltaic power generation.
(2) The time of dispatching and coordination control of the distribution network, including photovoltaic power generation in this method, is always less than 3.8s, and this method always has high efficiency of dispatching and coordination control of the distribution network.

Figure 1 .
Figure 1.Equivalent circuit model of photovoltaic power distribution network of charge carried by an electron; U and I are the output voltage and output current of the photovoltaic system, respectively.α is the fitting coefficient of the diodeconstant; T is the absolute temperature when the system is working.

fT
represents the fuel cost of the i th unit.When it is renewable energy, 0 i f = ; ij P represents the power of the i th unit in j working mode; i C refers to the maintenance cost of the i th unit; ij σ means that the i th unit works in j mode, represents the startup cost of the i th unit.The weighted sum of global reward and local reward, the final form of reward function is:

Figure 3 .
Figure 3.Power prediction accuracy of the photovoltaic unit

Table 1
.Dispatching coordination and control efficiency of the distribution network