Fast load flow calculation algorithm and module based on distributed processing

Abstract

The invention discloses a rapid load flow calculation module based on distributed processing, and the module is characterized in that the module comprises the steps: building a node distribution model, which is used for building a distribution model of each node in a power grid in detail; constructing an action space model, wherein the action space model is used for describing adjustment actions and operation measures in different power grid states; constructing a reward function model for optimizing system performance by quantitatively evaluating each index of power grid operation; and the model verification is used for verifying the accuracy and the practicability of the model. According to the method, all the nodes of the power grid are evaluated in detail, and the calculation strategy is dynamically adjusted according to real-time data, so that global optimization and balance of the power grid are realized; a distributed computing technology is utilized, a power grid is divided into a plurality of sub-regions, and load flow calculation of each sub-region is processed in parallel on a plurality of computing nodes, so that the computing efficiency is greatly improved, and the requirement of real-time computing can be met; and the precision of a load flow calculation result and the stability of power grid operation are obviously improved.

Description

Technical Field

[0001] The invention relates to a fast power flow calculation algorithm and module based on distributed processing. Technical Background

[0002] In the current low-voltage distribution network, with the large-scale access of distributed photovoltaic systems, traditional power flow calculation methods mainly rely on linear programming (LP) and Newton-Raphson methods. These methods are based on precise mathematical models and use detailed grid parameters to perform power flow analysis to determine the voltage, current distribution and power flow direction of each node. Traditional power flow calculation methods usually assume that the system is in a steady state, and the model is relatively simple, which is suitable for distribution networks without distributed power sources.

[0003] Linear programming methods are mainly used to deal with the supply and demand balance problem in power systems, and to meet the power demand at different times and locations by optimizing the power generation plan and power flow. The Newton-Raphson method is used to solve the nonlinear equations of the power system, calculate the voltage of each node and the line flow, and is widely used in power flow calculation and stability analysis of the power system.

[0004] The specific processes of these methods usually include the following steps:

[0005] 1. Model establishment: Construct a mathematical model of the low-voltage distribution network, including node voltage equations and line flow equations.

[0006] 2. Parameter input: Input the basic data of the system, such as the load demand of the node, the output of the power generation equipment, the impedance parameters of the line, etc.

[0007] 3. Power flow calculation: Use linear programming or Newton-Raphson method to solve the equations of the power system to obtain the voltage value of each node and the power flow direction of the line.

[0008] 4. Result analysis: Analyze the calculation results and evaluate the operating status of the system, such as voltage deviation, line loss, and power flow.

[0009] Although the traditional power flow calculation method can meet the basic needs of low-voltage distribution networks to a certain extent, it has the following main disadvantages when facing large-scale access to distributed photovoltaic systems:

[0010] 1. Insufficient response to dynamic changes: Traditional methods assume that the system is in a steady state and cannot respond in real time to the volatility of distributed photovoltaic power generation and the random changes in load demand. Photovoltaic power generation has obvious intermittent and random characteristics, and its output is greatly affected by weather conditions. Traditional power flow calculation methods are difficult to reflect the dynamic changes of the power grid in real time.

[0011] 2. High computational complexity: After the distributed photovoltaic system is connected, the number of nodes and power flow paths in the distribution network will increase significantly, resulting in a significant increase in the complexity and calculation time of power flow calculation. Especially in large-scale distribution networks, the computational efficiency of traditional methods is low and it is difficult to meet the requirements of real-time calculation.

[0012] 3. Insufficient stability and reliability: When high-penetration photovoltaic systems are connected, traditional power flow calculation methods are difficult to ensure the stability of calculation results and the reliable operation of distribution networks. When the photovoltaic power generation output changes dramatically, the system may experience voltage fluctuations, power reverse flow and other problems, and traditional methods cannot effectively cope with these challenges.

[0013] 4. Large data requirements: Traditional power flow calculation methods require a large amount of accurate input data, such as the load of each node, power output and line parameters, etc. The collection and processing of these data are difficult, and in actual applications, there are problems of incomplete and inaccurate data, which affects the reliability of the calculation results.

[0014] 5. Poor adaptability: Traditional methods are usually based on preset operating conditions and models and lack the ability to adapt to the actual operating environment. In the operation of the distribution network, power demand and supply conditions are constantly changing, and traditional methods are difficult to flexibly adjust the calculation strategy to adapt to these changes.

[0015] Therefore, the existing technology shows obvious shortcomings when dealing with the access of large-scale distributed photovoltaic systems. There is an urgent need for a new power flow calculation method that can respond to system changes in real time and improve computing efficiency and stability to meet the needs of modern low-voltage distribution networks.

[0016] With the transformation of the global energy structure and the rapid development of renewable energy, the proportion of distributed photovoltaic systems connected to low-voltage distribution networks continues to increase. Distributed photovoltaic systems have become an important direction for future energy development due to their advantages such as green environmental protection and distributed local consumption. However, this has also brought unprecedented challenges to traditional power systems.

[0017] The traditional low-voltage distribution network design is mainly optimized for centralized power generation and one-way power transmission. The power grid flow basically flows from the substation to each user end. However, the access of distributed photovoltaic systems has changed this model. Photovoltaic power generation is intermittent and random, and its output power is significantly affected by weather and day and night changes, which increases the uncertainty of power grid operation. Traditional power flow calculation methods are unable to cope with such changes.

[0018] First, the volatility and intermittency of the photovoltaic system pose challenges to the voltage and frequency stability of the power grid. When the photovoltaic power generation is greater than the local load demand, the excess power needs to be transmitted back to the main grid, resulting in reverse power flow, which may cause voltage deviation and line overload. The traditional power flow calculation method assumes that the power grid is in a steady state, which makes it difficult to reflect this dynamic change in real time, resulting in a large deviation between the calculation results and the actual situation.

[0019] Secondly, the large-scale access of distributed photovoltaic systems increases the complexity and computational load of the power grid. The traditional Newton-Raphson method and linear programming method are computationally inefficient when dealing with large-scale and complex power grids, especially in the case of high penetration of photovoltaic access, where the computational time and resource requirements increase significantly. This makes it difficult for traditional methods to meet the needs of real-time computing and rapid response.

[0020] In addition, existing power flow calculation methods often rely on accurate input data, such as the load of each node, power output and line parameters, etc. The collection and processing of these data are difficult, and there are problems of incomplete and inaccurate data in practical applications, which further limits the application effect of traditional methods.

[0021] Therefore, there is an urgent need for a new power flow calculation method that can analyze the distribution network operation status after the distributed photovoltaic system is connected in real time and accurately in a dynamic and complex power grid environment, so as to improve the operation efficiency and stability of the power grid. Summary of the invention

[0022] The purpose of the present invention is to provide a fast power flow calculation algorithm and module based on distributed processing. The method aims to solve the above problems by introducing improved algorithms and real-time data processing technology, and provide technical support for the safe and reliable operation of modern low-voltage distribution networks.

[0023] One of the technical solutions to achieve the above object is: a fast power flow calculation module based on distributed processing, characterized in that it includes:

[0024] Construct a node distribution model to establish a detailed distribution model of each node in the power grid;

[0025] Construct an action space model, which is used to describe the regulation actions and operation measures under different power grid states;

[0026] Construct a reward function model to optimize system performance by quantitatively evaluating various indicators of power grid operation;

[0027] Model verification is used to verify the accuracy and practicality of the above model.

[0028] In the above-mentioned fast power flow calculation module based on distributed processing, the step of constructing a node distribution model includes:

[0029] The power grid system structure records the overall topology of the power grid, including the connection relationship between each node and its position in the system;

[0030] Actual transmission line diagram, used to draw each actual transmission line in the power grid, and to identify the length and impedance parameters of the line in detail;

[0031] Grid topology, which is used to further refine the topology and covers the lines connected to each node;

[0032] The distribution of each node is used to analyze and record the voltage, current and load conditions of each node, providing a basis for subsequent modeling.

[0033] In the above-mentioned fast power flow calculation module based on distributed processing, the operation of constructing the action space model includes:

[0034] Power generation adjustment: by adjusting the output power of the photovoltaic power generation system to cope with the real-time demand changes of the power grid;

[0035] Transmission line configuration: Dynamically adjust the configuration of transmission lines to optimize the power flow path;

[0036] Demand response measures: Through demand response, users' electricity demand is dynamically adjusted to achieve load balance.

[0037] In the above-mentioned fast power flow calculation module based on distributed processing, the construction of the reward function model includes:

[0038] Reward function objectives, which are used to clearly optimize key objectives such as voltage stability and line loss reduction;

[0039] Reward and punishment mechanism, which is used to give rewards or punishments according to the effects of different operations, so as to guide the system to develop in the optimal direction;

[0040] Quantified reward value, used to quantify the reward value through a specific formula to facilitate calculation and adjustment in practical applications;

[0041] The reward function is adjusted in real time to adjust the weights and parameters of the reward function according to the actual operation of the power grid to achieve the optimal control effect.

[0042] In the above-mentioned fast power flow calculation module based on distributed processing, the model verification evaluates the performance of the model in different scenarios through load risk scenario simulation and solution implementation cost analysis, and realizes fast power flow calculation through joint operation.

[0043] In the above-mentioned fast power flow calculation module based on distributed processing, in the construction of the reward function model, the reward function is used to measure the effect of the power flow calculation method and optimize the operating state of the distribution network, including minimizing the voltage deviation.

[0044] The reward function includes a voltage deviation term, which encourages the algorithm to reduce the deviation of node voltage and keep the voltage stable. The formula is expressed as:

[0045]

[0046] Where V i is the voltage of the ith node; V ref is the reference voltage value; R V It is a quantitative indicator of voltage deviation.

[0047] In the above-mentioned fast power flow calculation module based on distributed processing, in the construction of the reward function model, the reward function is used to measure the effect of the power flow calculation method and optimize the operating state of the distribution network, including reducing line losses, optimizing power flow, reducing line losses, and improving the operating efficiency of the power grid. The formula is expressed as:

[0048]

[0049] In the formula, is the power loss of the jth line, R l It is a quantitative indicator of line loss.

[0050] In the above-mentioned fast power flow calculation module based on distributed processing, in the construction of the reward function model, the reward function is used to measure the effect of the power flow calculation method and optimize the operating state of the distribution network, including optimizing the power flow direction.

[0051]

[0052] Where P k is the power flow of the kth line; P max is the maximum power capacity of the line, R p Optimize quantitative indicators for line power flow.

[0053] The second technical solution to achieve the above purpose is: a fast power flow calculation algorithm based on distributed processing, comprising the following steps:

[0054] Step 1: Initial state setting

[0055] According to historical data and real-time monitoring data, set the initial state of the power grid, including initial voltage, current and power distribution;

[0056] Step 2: Algorithm iterative calculation

[0057] The improved power flow calculation algorithm is used to iteratively calculate the power grid state and update the voltage value of each node and the power flow direction of the line. The specific steps include:

[0058] Step 2.1, establish node voltage equation and line power flow equation;

[0059] Step 2.2, initialize algorithm parameters and set convergence conditions;

[0060] Step 2.3, iteratively solve the voltage equation and the power flow equation, and update the voltage and power values;

[0061] Step 3: Distributed Computing

[0062] The power grid is divided into several sub-areas. Distributed computing technology is used to process the power flow calculation of each sub-area in parallel on multiple computing nodes to speed up the calculation. The specific steps include:

[0063] Step 3.1, divide the grid area: divide the grid into multiple sub-areas according to the grid topology and computing requirements.

[0064] Step 3.2, parallel processing: Solve the power flow equations of each sub-area in parallel on multiple computing nodes.

[0065] Step 3.3, Summarize the results: Summarize the calculation results of each sub-area and update the overall power grid status.

[0066] Step 4: Result analysis and adjustment

[0067] In the above-mentioned fast power flow calculation algorithm based on distributed processing, in step 4, according to the calculation results, the operation status of the power grid is analyzed, including voltage deviation and line loss, and adjustment operations are performed in real time to optimize the operation of the power grid, which specifically includes the following steps:

[0068] Step 4.1, compare the calculation results with the actual monitoring data to evaluate the calculation accuracy;

[0069] Step 4.2, adjusting the voltage over-limit nodes and power overload lines, such as adjusting the photovoltaic power generation output and load demand.

[0070] The beneficial effects of the present invention are:

[0071] The present invention provides technical support for the safe and reliable operation of modern low-voltage distribution networks by introducing improved algorithms and real-time data processing technology. Compared with traditional power flow calculation methods, the present invention has made improvements in the following aspects:

[0072] 1. Global optimization and real-time response:

[0073] The present invention achieves global optimization and balance of the power grid by performing detailed evaluation on each node of the power grid and dynamically adjusting the calculation strategy according to real-time data.

[0074] 2. Improved computing efficiency:

[0075] By using distributed computing technology, the power grid is divided into several sub-areas, and the power flow calculations of each sub-area are processed in parallel on multiple computing nodes, which greatly improves the computing efficiency and can meet the needs of real-time computing.

[0076] 3. Result accuracy and stability:

[0077] Through the improved Newton-Raphson method and the precise design of the reward function, the accuracy of the power flow calculation results and the stability of the power grid operation are significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 It is the IEEE 33 node system topology

[0079] Figure 2 It is the 24-hour node voltage diagram

[0080] Figure 3 It is the 24-hour active load diagram

[0081] Figure 4 It is the 24-hour reactive load diagram

[0082] Figure 5 24-hour fan output diagram

[0083] Figure 6 This is the 24-hour CB output diagram

[0084] Figure 7 24-hour SVC output diagram

[0085] Figure 8 24-hour OLTC ratio diagram

[0086] Fig. 9 24-hour ESS battery charge and discharge power diagram

[0087] Fig.10 24-hour ESS battery SOC diagram

[0088] Fig.11 It is a block diagram of a fast power flow calculation module based on distributed processing of the present invention; DETAILED DESCRIPTION

[0089] The present invention will be further described below in conjunction with the accompanying drawings.

[0090] See also Fig.11The present invention provides a fast power flow calculation module based on distributed processing, which is specifically described as follows.

[0091] 1. Construct node distribution model: This module aims to establish the distribution model of each node in the power grid in detail. Specifically, it includes:

[0092] Grid system structure: This part records the overall topology of the grid, including the connection relationship between each node and its position in the system.

[0093] Actual transmission line diagram: Draw the actual transmission lines in the power grid, and identify detailed parameters such as line length and impedance.

[0094] Grid topology: The topology is further refined to cover the lines connecting each node to it.

[0095] Distribution of each node: Analyze and record the voltage, current and load conditions of each node to provide a basis for subsequent modeling.

[0096] 2. Construct an action space model: The action space model is used to describe the regulation actions and operation measures under different power grid states. Specific operations include:

[0097] Power generation adjustment: By adjusting the output power of the photovoltaic power generation system to respond to real-time demand changes in the power grid.

[0098] Transmission Line Configuration: Dynamically adjust the configuration of transmission lines to optimize the power flow path.

[0099] Demand response measures: Through demand response, users' electricity demand is dynamically adjusted to achieve load balance.

[0100] 3. Constructing a reward function model: In the present invention, the design of a reward function model is crucial, and the purpose is to optimize system performance by quantitatively evaluating various indicators of power grid operation. It includes the following:

[0101] Reward function objective: Clearly optimize key objectives such as voltage stability and reduce line losses.

[0102] Reward and punishment mechanism: rewards or punishments are given according to the effects of different operations, thus guiding the system to develop in the optimal direction.

[0103] Quantify the reward value: Quantify the reward value through a specific formula to facilitate calculation and adjustment in practical applications.

[0104] Real-time adjustment of reward function: According to the actual operation of the power grid, the weights and parameters of the reward function are adjusted in real time to achieve the optimal control effect.

[0105] 4. Model verification: This part focuses on verifying the accuracy and practicality of the above model. Through load risk scenario simulation and solution implementation cost analysis, the performance of the model in different scenarios is evaluated.

[0106] Through the joint operation of these modules, the proposed fast power flow calculation algorithm is realized. This algorithm uses dynamic adjustment and distributed processing technology to significantly improve the calculation efficiency and stability of the power grid when high penetration photovoltaic systems are connected, providing technical support for the safe and reliable operation of modern low-voltage distribution networks.

[0107] Model building

[0108] a. Definition of state space

[0109] The state space is composed of key parameters such as node voltage, line flow and load conditions of the distribution network, taking into account the volatility of photovoltaic power generation and the dynamic impact of load changes. The definition of the state space requires detailed recording and updating of the grid status of each node at different time points, such as Figure 1 As shown, including:

[0110] 1) Voltage level of each node

[0111] Includes voltage values ​​at photovoltaic access points and load points to reflect voltage fluctuations.

[0112] 2) Power flow of each line

[0113] Includes active power and reactive power to reflect the direction and size of the flow.

[0114] 3) Power generation output of photovoltaic system

[0115] PV power generation is predicted based on weather forecasts and historical data and updated in real time.

[0116] 4) Changes in user load demand

[0117] Dynamically adjust load demand based on power load forecasting model and real-time monitoring data.

[0118] b. Design of action space

[0119] The action space includes various possible grid regulation operations to cope with dynamically changing grid conditions. The main regulation operations include:

[0120] 1) Adjust photovoltaic power generation

[0121] Increase or decrease the power generation of the photovoltaic system to adapt to the real-time demand of the power grid. This can be achieved by adjusting the output power of the photovoltaic inverter.

[0122] 2) Start and stop the backup power supply

[0123] According to the load of the power grid, the backup power source is started or stopped to ensure the stability of power supply. For example, the backup diesel generator is started when the photovoltaic power generation is insufficient.

[0124] 3) Adjust load demand

[0125] Through demand response measures, users' electricity demand can be adjusted to balance the load on the power grid. For example, users can be encouraged to increase electricity consumption during the peak period of photovoltaic power generation and reduce electricity consumption during the off-peak period.

[0126] c. Design of reward function

[0127] The reward function is used to measure the effect of the power flow calculation method. Its design goal is to optimize the operation status of the distribution network, including:

[0128] 1) Minimize voltage deviation

[0129] The reward function includes a voltage deviation term to encourage the algorithm to reduce the deviation of node voltage and maintain voltage stability. The formula is:

[0130]

[0131] Where V i is the voltage of the ith node; V ref is the reference voltage value.

[0132] 2) Line loss reduction

[0133] By optimizing the power flow, reducing line losses, and improving the operating efficiency of the power grid, the formula is expressed as:

[0134]

[0135] In the formula, is the power loss of the jth line.

[0136] 3) Power flow optimization

[0137]

[0138] Where P k is the power flow of the kth line; P max is the maximum power capacity of the line.

[0139] d. Application of power flow calculation algorithm

[0140] An improved power flow calculation algorithm is used, namely the fast power flow calculation algorithm based on distributed processing. This algorithm combines the advantages of the Newton-Raphson method and distributed computing, improves the calculation efficiency through parallel computing, and adapts to the real-time power flow calculation needs of high-penetration photovoltaic systems. The specific steps are as follows:

[0141] 1) Initial state setting

[0142] According to historical data and real-time monitoring data, the initial state of the power grid is set, including initial voltage, current and power distribution.

[0143] 2) Algorithm iterative calculation

[0144] The improved power flow calculation algorithm is used to iteratively calculate the power grid status and update the voltage value of each node and the power flow direction of the line. The specific steps include:

[0145] 1. Establish node voltage equation and line power flow equation.

[0146] 2. Initialize algorithm parameters and set convergence conditions.

[0147] 3. Iteratively solve the voltage equation and power flow equation to update the voltage and power values.

[0148] 3) Distributed computing

[0149] The power grid is divided into several sub-areas, and distributed computing technology is used to process the power flow calculation of each sub-area in parallel on multiple computing nodes to speed up the calculation. The specific steps include:

[0150] 1. Divide the grid into regions: Divide the grid into multiple sub-regions based on the grid topology and computing requirements.

[0151] 2. Parallel processing: Solve the flow equations of each sub-area in parallel on multiple computing nodes.

[0152] 3. Summarize results: Summarize the calculation results of each sub-area and update the overall power grid status.

[0153] 4) Result analysis and adjustment

[0154] According to the calculation results, the operation status of the power grid is analyzed, such as voltage deviation, line loss, etc., and adjustments are made when necessary to optimize the operation of the power grid. The specific steps include:

[0155] 1. Compare the calculation results with the actual monitoring data and evaluate the calculation accuracy.

[0156] 2. Adjust voltage-exceeding nodes and power-overloaded lines, such as adjusting photovoltaic power generation output and load demand.

[0157] (3) Model verification

[0158] a. Simulation verification

[0159] The algorithm proposed in this paper is simulated and verified by the distribution system of IEEE33 nodes. The discount rate in this planning network is 10%, the engineering economic service life of the line is 20 years, the fixed operation rate of the project is 2.5%, the line investment is 350,000 yuan / km, the network loss price is 0.25 yuan / (kW·h), the basic load shedding compensation price is 6 yuan / (kW·h), the transmission line trip probability is 0.2 times / 100km, the average fault repair time is 5h, and w1, w2 and w3 are all 1. Add photovoltaic and energy storage elements to the basic network.

[0160] Use the YALM IP interface in MATLAB to call the CPLEX solver for optimization. Output the optimization results and check and analyze the results. Figure 2 .

[0161] b. Analysis of power flow voltage and power results

[0162] Figure 3 and Figure 4 The distribution of active load and reactive load of the distribution system within 24 hours is shown respectively. It can be seen that the fluctuation trends of active and reactive loads are similar during the day, both reaching peaks from the 5th to the 10th hour, and are at their lowest points from the 0th to the 5th hour and from the 15th to the 24th hour. In addition, there are significant differences in the load distribution of different nodes, with higher loads on low-numbered nodes and lower loads on high-numbered nodes. These characteristics reveal the temporal and spatial distribution of system loads, which are of great guiding significance for the optimization of power dispatching and reactive compensation equipment configuration. Through reasonable load management and reactive compensation, the pressure of peak load on the system can be effectively reduced, ensuring voltage stability and the economy of system operation.

[0163] c. Flow output results of each component of the system

[0164] Figure 5 , 6 and Figure 7The output of wind turbines, reactive power compensation capacitors (CBs) and static VAR compensators (SVCs) within 24 hours is shown, providing a detailed analysis basis for the operation of the distribution system. The output of wind turbines shows obvious fluctuation characteristics within a day. The output is relatively high from 0 to 5 hours and from 20 to 24 hours, and gradually decreases and reaches a trough between 5 and 20 hours, reflecting the availability of wind energy resources in different periods. The output change of CB shows that its demand for reactive power compensation is high in a specific time period, mainly concentrated in several periods for adjustment, but the overall operation frequency is relatively small, indicating that the system reactive power demand is low during most of the time. The frequent output change of SVC shows its important role in maintaining system voltage stability and reactive power balance. By continuously adjusting reactive power to adapt to real-time changes in load and voltage, the output change trends of different SVCs are not exactly the same, reflecting the differences in the demand for reactive power compensation at different locations in the system. These analysis results show that the reasonable configuration and scheduling of these devices can effectively improve the operating efficiency and reliability of the distribution system, and provide an important basis for optimizing the system's scheduling strategy and equipment configuration.

[0165] Figure 8 , Fig. 9 and Fig.10 The changes in the transformation ratio of the on-load tap-changing transformer (OLTC), the charging and discharging power of the energy storage system (ESS), and the battery state (Soc) within 24 hours are shown. OLTC effectively maintains the stability of the system voltage by adjusting the transformation ratio during the morning and evening peaks. The transformation ratio jumps to a high level around the 5th hour and decreases around the 20th hour. ESS frequently performs charging and discharging operations throughout the day, discharging to meet demand during high-load periods and charging to store energy during low-load periods. The battery Soc shows a trend of gradually increasing during the day and gradually decreasing at night, which is consistent with the charging and discharging strategy, reflecting the important role of the energy storage system in balancing the supply and demand of electricity. Overall, the coordinated regulation of OLTC and ESS effectively improves the operating efficiency and reliability of the distribution system, and provides key data support for optimizing scheduling strategies and equipment configuration.

[0166] The present invention achieves the following technical effects:

[0167] (1) Enhanced adaptability and responsiveness: The present invention learns the grid status and operation feedback in real time and dynamically adjusts the power flow calculation strategy to cope with sudden grid events and fluctuations, thereby enhancing the grid's adaptability to uncertainty and changes.

[0168] (2) Optimized resource utilization: By accurately calculating and predicting the power flow at each node of the power grid, the present invention can effectively adjust photovoltaic power generation and load demand, reduce voltage deviation and line loss, and improve the operating efficiency and energy utilization of the distribution network.

[0169] (3) Improved grid stability: By combining real-time monitoring data with improved computing algorithms, the present invention can identify and respond to potential grid risks in advance, reduce the probability of system failures and accidents, and improve the overall stability and safety of the grid.

[0170] (4) Efficient computing performance: Distributed processing technology is used to decompose the power flow calculation task into multiple subtasks for parallel processing, which greatly improves the computing efficiency. It can complete the power flow calculation of large-scale power grids in a short time and meet real-time computing needs.

[0171] (5) Promote efficient integration of renewable energy: By optimizing the flow calculation strategy and adjusting the grid load, the present invention improves the absorption capacity of renewable energy such as photovoltaics, supports the development of green energy, and reduces carbon emissions.

[0172] The invention proposes a fast power flow calculation algorithm based on distributed processing, which aims to optimize the power flow calculation after the distributed photovoltaic system is connected to the low-voltage distribution network. By introducing the improved Newton-Raphson method and distributed computing technology, the invention can respond to the dynamic changes of the power grid in real time, improve the calculation efficiency and result accuracy, and thus effectively solve the problems of insufficient dynamic change response, high calculation complexity, insufficient stability and reliability existing in the prior art.

[0173] In summary, the method of the present invention not only has significant advantages in calculation efficiency and result accuracy, but also improves the adaptability and stability of the power grid when high-penetration photovoltaic systems are connected through dynamic adjustment and distributed processing technology, providing strong technical support for the safe and reliable operation of modern low-voltage distribution networks.

[0174] The present invention is described in detail above in conjunction with the embodiments of the accompanying drawings. A person skilled in the art can make various variations of the present invention according to the above description. Therefore, certain details in the embodiments should not constitute a limitation of the present invention, and the scope of protection of the present invention shall be defined by the scope of the attached claims.

Claims

1. A fast power flow calculation module based on distributed processing, characterized in that: It includes: Construct a node distribution model to establish a detailed distribution model of each node in the power grid; Construct an action space model, which is used to describe the regulation actions and operation measures under different power grid states; Construct a reward function model to optimize system performance by quantitatively evaluating various indicators of power grid operation; Model verification is used to verify the accuracy and practicality of the above model.

2. The fast power flow calculation module based on distributed processing according to claim 1, characterized in that: The step of constructing a node distribution model comprises: The power grid system structure records the overall topology of the power grid, including the connection relationship between each node and its position in the system; Actual transmission line diagram, used to draw each actual transmission line in the power grid, and to identify the length and impedance parameters of the line in detail; Grid topology, which is used to further refine the topology and covers the lines connected to each node; The distribution of each node is used to analyze and record the voltage, current and load conditions of each node, providing a basis for subsequent modeling.

3. The fast power flow calculation module based on distributed processing according to claim 1, characterized in that: The operation of constructing the action space model includes: Power generation adjustment: by adjusting the output power of the photovoltaic power generation system to cope with the real-time demand changes of the power grid; Transmission line configuration: Dynamically adjust the configuration of transmission lines to optimize the power flow path; Demand response measures: Through demand response, users' electricity demand is dynamically adjusted to achieve load balance.

4. The fast power flow calculation module based on distributed processing according to claim 1, characterized in that: The reward function model is constructed, and its contents include: Reward function objectives, which are used to clearly optimize key objectives such as voltage stability and line loss reduction; Reward and punishment mechanism, which is used to give rewards or punishments according to the effects of different operations, so as to guide the system to develop in the optimal direction; Quantified reward value, used to quantify the reward value through a specific formula to facilitate calculation and adjustment in practical applications; The reward function is adjusted in real time to adjust the weights and parameters of the reward function according to the actual operation of the power grid to achieve the optimal control effect.

5. The fast power flow calculation module based on distributed processing according to claim 1, characterized in that: The model verification evaluates the performance of the model in different scenarios through load risk scenario simulation and solution implementation cost analysis, and realizes fast power flow calculation through joint operation.

6. The fast power flow calculation module based on distributed processing according to claim 4, characterized in that: In the reward function model, the reward function is used to measure the effect of the power flow calculation method and optimize the operation state of the distribution network, including minimizing the voltage deviation. The reward function includes a voltage deviation term, which encourages the algorithm to reduce the deviation of node voltage and keep the voltage stable. The formula is expressed as: Where V i is the voltage of the ith node; V ref is the reference voltage value; R V It is a quantitative indicator of voltage deviation.

7. The fast power flow calculation module based on distributed processing according to claim 4, characterized in that: In the reward function model, the reward function is used to measure the effect of the power flow calculation method and optimize the operation state of the distribution network, including reducing line losses, optimizing power flow, reducing line losses, and improving the operation efficiency of the power grid. The formula is expressed as: In the formula, is the power loss of the jth line, R l It is a quantitative indicator of line loss.

8. The fast power flow calculation module based on distributed processing according to claim 4, characterized in that: In the reward function model, the reward function is used to measure the effect of the power flow calculation method and optimize the operation state of the distribution network, including optimizing the power flow direction. Where P k is the power flow of the kth line; P max is the maximum power capacity of the line, R p Optimize quantitative indicators for line power flow.

9. A fast power flow calculation algorithm based on distributed processing of the module as claimed in claim 1, comprising the following steps: Step 1: Initial state setting According to historical data and real-time monitoring data, set the initial state of the power grid, including initial voltage, current and power distribution; Step 2: Algorithm iterative calculation The improved power flow calculation algorithm is used to iteratively calculate the power grid state and update the voltage value of each node and the power flow direction of the line. The specific steps include: Step 2.1, establish node voltage equation and line power flow equation; Step 2.2, initialize algorithm parameters and set convergence conditions; Step 2.3, iteratively solve the voltage equation and the power flow equation, and update the voltage and power values; Step 3: Distributed Computing The power grid is divided into several sub-areas. Distributed computing technology is used to process the power flow calculation of each sub-area in parallel on multiple computing nodes to speed up the calculation. The specific steps include: Step 3.1, divide the grid area: divide the grid into multiple sub-areas according to the grid topology and computing requirements. Step 3.2, parallel processing: Solve the power flow equations of each sub-area in parallel on multiple computing nodes. Step 3.3, Summarize the results: Summarize the calculation results of each sub-area and update the overall power grid status. Step 4: Result analysis and adjustment.

10. The fast power flow calculation algorithm based on distributed processing according to claim 9 is characterized in that: In step 4, according to the calculation results, the operation status of the power grid is analyzed, including voltage deviation and line loss, and real-time adjustment operations are performed to optimize the operation of the power grid, which specifically includes the following steps: Step 4.1, compare the calculation results with the actual monitoring data to evaluate the calculation accuracy; Step 4.2, adjusting the voltage over-limit nodes and power overload lines, such as adjusting the photovoltaic power generation output and load demand.

Images