Urban power grid optimization control system and method based on distribution network phase balance

Abstract

An urban area optimization control system and method based on phase balance of distribution network, the method includes: obtaining distribution network operation data in real time through intelligent sensors and measurement and control devices deployed at key nodes of distribution network; performing load forecasting based on the collected distribution network operation data and clustering loads through clustering algorithms, and evaluating load status based on load clustering results and load results; taking into account power grid topology, distribution network operation data, load forecasting results and load evaluation results, adopting multi-agent collaborative optimization to determine the optimal load dispatching instructions; converting the obtained optimal dispatching instructions into actual control actions based on phase balance to optimize the power grid. The present invention can reduce the three-phase imbalance of the distribution network from 23.6% to 8.4%, effectively improving the power supply quality and operation efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of power distribution network automation, and more specifically, relates to an urban power grid optimization control system and method based on distribution network interphase balance. Background Art

[0002] As my country's urbanization process continues to accelerate, the load of urban distribution networks has grown rapidly, and power grid planning and construction are facing new challenges. The "Distribution Network Construction and Reconstruction Action Plan (2015-2020)" clearly proposes to strengthen the construction of distribution network automation and improve power supply reliability and power quality.

[0003] As a key link between the transmission network and end users, urban power grids are characterized by dense loads, diverse types of electricity consumption, and high requirements for power supply reliability. However, affected by historical planning, equipment aging and other factors, urban power grids generally have problems such as unreasonable structure, low automation level, and uneven load distribution. In particular, some urban distribution networks generally have the problem of "heavy overload and light quality", and the imbalance of three-phase load is particularly prominent, which seriously affects the power supply quality and equipment health.

[0004] Taking a certain city center area as an example, the conventional distribution transformer capacity is between 315 and 630 kVA, while the actual maximum load can reach more than 1200 kVA, and the transformer overload rate is as high as 180%, which seriously threatens the equipment life and power supply safety. At the same time, due to the lack of effective monitoring means and control strategies, the problem of three-phase load imbalance is becoming increasingly prominent. Statistics show that more than 40% of the distribution transformers have a three-phase load imbalance rate of more than 20%, and the highest is 35%, far exceeding the national standard recommended value of 15%. Load imbalance leads to a series of problems such as neutral line overcurrent, increased voltage deviation, and reduced power quality. The voltage qualification rate is only 92%, the harmonic voltage distortion rate THD is as high as 7.8%, and the user complaint rate remains high. In addition, due to the lack of capacitor compensation devices, some old communities have a power factor lower than 0.8 for a long time, and the reactive power is as high as 62% of the active power. A large amount of reactive current exacerbates line losses.

[0005] Therefore, innovative intelligent technologies are urgently needed to improve the optimization and regulation capabilities of urban distribution networks. Summary of the invention

[0006] In order to solve the deficiencies in the prior art, the present invention provides an intelligent optimization control system based on phase balance of distribution network, which aims to solve the problem of three-phase load imbalance in the current urban distribution network and significantly improve the power supply quality and operation efficiency of the power grid. Compared with the traditional distribution network technology, the present invention realizes the dynamic balance and precise control of the load by innovatively combining multi-agent collaborative optimization and improving the Petri net control model. This technology can perceive the state of the power grid in real time, quickly adjust the load distribution, reduce the three-phase imbalance, optimize the power factor, reduce line losses, and improve the stability and safety of the power grid.

[0007] The present invention adopts the following technical solution.

[0008] A first aspect of the present invention provides an urban power grid optimization control method based on distribution network phase balance, comprising:

[0009] By deploying intelligent sensors and measurement and control devices at key nodes of the distribution network, the distribution network operation parameters can be obtained in real time;

[0010] Mining the collected distribution network operation parameters, conducting load forecasting and evaluating load status;

[0011] Taking into account the grid topology, distribution network operating parameters and load forecasting results, multi-agent collaborative optimization is used to determine the optimal load dispatch instructions;

[0012] The optimal dispatching instructions are converted into actual control actions to optimize the control of the power grid.

[0013] Preferably, obtaining the distribution network operating parameters includes:

[0014] Smart sensors continuously collect raw data and perform local preprocessing to filter out abnormal values;

[0015] The intelligent power distribution terminal performs frequency domain analysis on the data through discrete Fourier transform algorithm to form a standardized data frame;

[0016] The data frame is encapsulated according to the IEC 61850 standard and transmitted to the distribution automation master station via the MMS protocol;

[0017] The distribution automation master station parses the data frame and stores the data in the real-time database.

[0018] Preferably, load forecasting and assessment includes:

[0019] Clean and standardize the raw data to remove noise and outliers;

[0020] Through the improved K-means clustering algorithm, users are grouped according to the similarity of load curves, typical load patterns are identified, and load clustering is completed;

[0021] Combine historical data with real-time measurements to make load forecasts;

[0022] According to the load clustering and load forecasting results, the load status is evaluated in combination with the three-phase imbalance index and the potential overload risk index.

[0023] Preferably, a support vector regression model with adaptive weights is used for load forecasting, which can be expressed by the following formula:

[0024]

[0025] Among them, x i is the input feature vector, x t is the characteristic vector at time t, α i and is the Lagrange multiplier, b is the bias term, K(·,·) is the kernel function, and N is the total number of eigenvectors.

[0026] Preferably, the multi-agent collaborative optimization determines the optimal load dispatch instruction including:

[0027] The intelligent agent obtains the power grid topology model, real-time distribution network operation data and load forecast results of the area under its jurisdiction, and exchanges information with adjacent intelligent agents; each intelligent agent independently solves a multi-objective optimization problem. The objective function includes minimizing the three-phase load imbalance, minimizing the network loss and maximizing the voltage qualification rate, and the optimal solution is used as the optimal load scheduling instruction.

[0028] Preferably, the agent uses an adaptive differential evolution algorithm to solve the objective function, which can be expressed by the following formula:

[0029]

[0030] in, is the mutation vector, are three different individuals randomly selected from the tth generation, and F is the scaling factor.

[0031] Preferably, converting the optimal scheduling instruction into an actual control action includes:

[0032] After receiving the load transfer plan issued by the distribution automation master station, the intelligent distribution terminal first performs a safety check to evaluate the feasibility of the plan. After the evaluation is completed, the overall plan is decomposed into a switch operation sequence, and the switch operation sequence is modeled and optimized according to the spatial position and constraint relationship of the switch.

[0033] Preferably, based on the modeling and optimization of the switch operation sequence, the intelligent distribution terminal uses a heuristic search algorithm to search for the optimal switch operation path and generate control instructions. The control instructions are broadcast to each switch device through the GOOSE protocol of IEC 61850 to complete the real-time transfer of the load; at the same time, the intelligent distribution terminal monitors the load response in real time and feeds back the execution results to the distribution automation master station.

[0034] Preferably, a Petri net model is used to model and optimize the switch operation sequence, and its state equation can be expressed by the following formula:

[0035] M(t+1)=M(t)+A T ·U(t) (3)

[0036] Among them, M(t) is the identification vector at time t, A is the input matrix, and U(t) is the control vector.

[0037] The second aspect of the present invention provides an urban power grid optimization control system based on distribution network interphase balance, which runs the above-mentioned urban power grid optimization control method based on distribution network interphase balance, characterized in that:

[0038] Data acquisition module, used to obtain power grid operation parameters in real time;

[0039] Load analysis module, used to perform load forecasting and load evaluation based on power grid operation parameters;

[0040] Optimization decision module, used to determine the optimal load dispatch instructions;

[0041] The control execution module is used to convert the optimal load scheduling instructions into actual control actions.

[0042] Compared with the prior art, the beneficial effects of the present invention include at least: the interphase balance of the distribution network proposed by the present invention has achieved remarkable results in practical applications, innovatively combining multi-agent collaborative optimization with improved Petri net control to achieve dynamic balance and precise regulation of the distribution network load. The deployment of the system significantly improves the power supply quality indicators and provides a feasible solution for the construction of smart distribution networks. Future work will further optimize the load forecasting algorithm, introduce deep reinforcement learning technology to improve the system adaptability, and explore the feasibility of extending this technology to larger-scale distribution networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the layout of the optimization control system; DETAILED DESCRIPTION

[0044] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The described embodiments are only embodiments of a part of the present invention, not all embodiments. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0045] Phase balancing technology is an innovative distribution automation technology that aims to dynamically balance the load distribution of three-phase power grids in real time. Its basic principle is to collect key parameters such as current and voltage of each phase in real time, and use intelligent algorithms to analyze the imbalance of three-phase loads. According to the analysis results, the system automatically generates the optimal load adjustment strategy, and dynamically adjusts the load power of each phase through the coordinated control of distribution automation terminal equipment to balance the three-phase load. Taking a 10kV distribution network in a certain urban area as an example, by deploying a phase balancing control device, when it is detected that the load of phase A is 1.2MW, phase B is 0.9MW, and phase C is 0.6MW, the control strategy switches part of the load of phase C to phases A and B, so that the three-phase load ratio reaches the optimized level of 1:0.98:0.95, and the voltage deviation rate is reduced from 4.2% to less than 1.5%. This technology also integrates an adaptive power factor optimization model, which controls the overall power factor above 0.95 while ensuring phase balance, effectively reducing line losses. In addition, deep learning algorithms are introduced for load forecasting and control parameter self-tuning, which greatly improves the robustness and adaptability of the system. In short, phase balancing technology provides an efficient and reliable solution for distribution automation through real-time perception, intelligent decision-making, and precise control, which is of great significance to improving power supply quality, reducing line losses, and ensuring safe operation of equipment.

[0046] The first embodiment of the present invention provides an urban power grid optimization control method based on distribution network phase balance, comprising:

[0047] Step 1: By deploying intelligent sensors and measurement and control devices at key nodes of the distribution network, the grid operation parameters such as three-phase load data, voltage data and power factor of each distribution transformer are obtained in real time.

[0048] Various intelligent sensors continuously collect raw data and perform local preprocessing to filter out abnormal values. Then, the intelligent distribution terminal uses the discrete Fourier transform (DFT) algorithm to perform frequency domain analysis on the data, extracts fundamental wave, harmonics and other characteristic quantities, and forms a standardized data frame. The data frame is encapsulated according to the IEC 61850 standard and transmitted to the distribution automation master station through the MMS protocol. The distribution automation master station parses the data frame, stores the data in a real-time database, and triggers subsequent analysis and decision-making processes. The system introduces a data compression algorithm based on wavelet transform. For the voltage data U collected at the i-th measuring point at time t,i (t), after wavelet transformation, the wavelet coefficient W is obtained i (j,k), can be expressed by the following formula:

[0049]

[0050] Among them, j is the scale factor, k is the translation factor, and ψ is the wavelet basis function.

[0051] By threshold quantization and encoding of wavelet coefficients, a data compression ratio of up to 5:1 is achieved, which greatly reduces transmission and storage overhead while ensuring information integrity.

[0052] Step 2: Conduct in-depth mining of the grid operation data collected in step 1, evaluate the current load status, and predict future change trends.

[0053] The raw data is cleaned and standardized to remove noise and outliers and improve data quality. Then, clustering algorithms are introduced, including but not limited to the improved K-means clustering algorithm, to group users according to the similarity of load curves and identify typical load patterns. On this basis, the support vector regression (SVR) model with adaptive weights is used to combine historical data and real-time measurements to make rolling predictions of load changes in the future. For the load forecast value at time t, The SVR model can be expressed as follows:

[0054]

[0055] Among them, x i is the input feature vector, x t is the characteristic vector at time t, α i and is the Lagrange multiplier, b is the bias term, K(·,·) is the kernel function, and N is the total number of eigenvectors.

[0056] By introducing time decay factors and seasonal factors, the weights of each feature are dynamically adjusted to improve the accuracy and adaptability of the prediction. At the same time, the model also takes into account the influence of external factors such as temperature and humidity, enhancing the robustness of the prediction. Finally, based on load clustering and prediction results, the system evaluates indicators such as three-phase imbalance and potential overload risk, providing a basis for subsequent optimization decisions.

[0057] Step 3, taking into account multiple sources of information such as grid topology, grid operating parameters, and load forecasting, formulate the optimal load dispatching instructions, and adopt a multi-agent collaborative optimization framework. Each distribution transformer corresponds to an agent. Through distributed decision-making and negotiation mechanisms, the optimal load dispatching instructions are obtained to achieve optimized operation of the entire distribution network.

[0058] The agent first obtains the grid topology model, real-time distribution network operation data and load forecast results of the area under its jurisdiction, and exchanges information with adjacent agents. Then, each agent independently solves a multi-objective optimization problem. The objective function includes minimizing the three-phase load imbalance, minimizing network loss, maximizing the voltage qualification rate, etc. At the same time, the optimization problem is subject to operating constraints such as voltage amplitude and line capacity. The agent uses an adaptive differential evolution algorithm to solve the problem, continuously iterating and updating the load transfer plan, and using the optimal solution as the load dispatch instruction. Suppose the solution vector of the i-th agent in the t-th generation is Then its update formula can be expressed as follows:

[0059]

[0060] in, is the mutation vector, are three randomly selected distinct individuals in the tth generation, and F is the scaling factor. The algorithm balances global exploration and local development capabilities by adaptively adjusting F and crossover probability. Finally, each agent broadcasts its own optimization results to its neighbors and adjusts its decision based on feedback information. After multiple rounds of iterations, it converges to the global optimal solution and generates a detailed load transfer plan as a load scheduling instruction to guide subsequent control execution.

[0061] Step 4: Convert the dispatching instruction obtained in step 3 into actual control action.

[0062] After receiving the load transfer plan issued by the distribution network automation master station, the intelligent distribution terminal first performs a safety check to evaluate the feasibility of the plan. Then, the overall plan is decomposed into a series of switch operation sequences. According to the spatial position and constraint relationship of the switch, the improved Petri net model is used to model and optimize the switch operation sequence. The Petri net model can formally describe the static structure and dynamic behavior of the discrete event system, and its state equation can be expressed as:

[0063] M(t+1)=M(t)+A T ·U(t) (4)

[0064] Among them, M(t) is the identification vector at time t, A is the input matrix, and U(t) is the control vector. By introducing the prohibition arc and the suppression arc, the occurrence of unsafe conditions can be effectively avoided. On this basis, the intelligent distribution terminal adopts a heuristic search algorithm, such as A * The algorithm searches for the optimal switch operation path and generates detailed control instructions. Finally, the instructions are broadcast to each switch device through the GOOSE protocol of IEC 61850 to complete the real-time transfer of load. At the same time, the intelligent distribution terminal monitors the load response in real time and feeds back the execution results to the distribution automation master station to achieve closed-loop control.

[0065] like Figure 1 As shown, embodiment 2 of the present invention provides an urban power grid optimization control system based on distribution network phase balance, and runs an urban power grid optimization control method based on distribution network phase balance described in embodiment 1, including: a data acquisition module, a load analysis module, an optimization decision module and a control execution module.

[0066] First, the data acquisition module monitors the operating parameters of each distribution transformer, such as three-phase load data, voltage data and power factor, in real time; second, the load analysis module processes and analyzes the collected data, evaluates the degree of three-phase load imbalance, and predicts future load change trends; third, the optimization decision module formulates the optimal phase-to-phase load adjustment plan, including load transfer path and adjustment timing, based on the analysis results, combined with the distribution network topology and operation constraints; finally, the control execution module implements the optimization plan through remote control equipment, adjusts the load distribution of each phase, monitors the execution effect in real time, and makes dynamic adjustments when necessary.

[0067] The data acquisition module is the information source of the entire system. It obtains the grid operation parameters in real time through the intelligent sensors and measurement and control devices deployed at the key nodes of the distribution network. This module mainly includes intelligent sensors such as current, voltage, and power, as well as intelligent distribution terminals, distribution automation master stations and other equipment. Various intelligent sensors are connected to the intelligent distribution terminals through the RS-485 bus, with a sampling frequency of up to 10kHz and a measurement accuracy better than 0.5. The intelligent distribution terminal uploads the summary data to the distribution automation master station every 5 minutes, and achieves high-speed and reliable transmission through a dedicated optical fiber network.

[0068] The load analysis module is one of the core components of the system. Its main task is to conduct in-depth mining of the collected power grid operation data, evaluate the current load status, and predict future change trends. The module is based on a big data analysis platform, adopts a distributed computing architecture, and is equipped with high-performance servers and massive storage arrays to ensure the real-time and scalability of data processing.

[0069] The optimization decision module comprehensively considers multiple sources of information such as power grid topology, equipment parameters, load forecasting, etc. to formulate the optimal load scheduling strategy. This module adopts a multi-agent collaborative optimization framework, with each distribution transformer corresponding to an agent, and achieves the optimal operation of the entire system through distributed decision-making and negotiation mechanisms.

[0070] The control execution module is responsible for converting the dispatching instructions into actual control actions. This module consists of intelligent distribution terminals, switchgear, communication gateways, etc., and realizes real-time communication with the distribution automation master station through highly reliable industrial Ethernet.

[0071] In order to verify the practical application effect of the power grid optimization control system of the present invention, the third embodiment of the present invention selects a typical power supply unit of a 10kV distribution network in a certain urban area as an experimental scene, including an 800kVA distribution transformer and its three low-voltage feeders. The experiment uses a data acquisition device based on SIEMENS SICAM RTU, the sampling frequency is set to 2kHz, and the measurement accuracy is 0.2. The distribution terminal uses ABB REC670 intelligent terminal with a phase control accuracy of ±0.1%. The experiment is divided into three stages: the first stage (0-2h) collects the original operation data as a benchmark; the second stage (2-4h) enables the load analysis module and runs the improved K-means clustering algorithm (cluster number k=5, iteration number 200); the third stage (4-6h) activates the full system function and performs phase balance control. Evaluation indicators include: three-phase imbalance δ (%), voltage deviation rate ε (%), power factor cosφ, line loss rate η (%) and system response time τ (s). The three-phase imbalance is calculated based on the symmetrical component method, the sampling time window is 10 minutes, and the harmonic distortion rate of each phase current (THDi≤4.0%) is monitored at the same time to ensure data reliability. During the experiment, the ambient temperature was controlled at 25±2℃ and the relative humidity was 45±5%.

[0072] 4.2 Results Analysis

[0073] By comparing the operating data before and after the system is activated, the effectiveness of the proposed scheme is verified. Tables 1 and 2 show the performance index data and load distribution of the system at each stage.

[0074] Table 1 Comparison of performance indicators of each stage of the system

[0075]

[0076] Table 2 Three-phase load distribution before and after system optimization (kW)

[0077]

[0078] It can be seen from the experimental data that the system has achieved significant optimization effects in the third stage. The three-phase imbalance has dropped from the initial 23.6% to 8.4%, a decrease of 64.4%; the voltage deviation rate has improved from 4.8% to 1.6%, meeting the requirements of the national standard GB / T12325-2008; the power factor has increased to 0.95, meeting the assessment standards of the power department; the line loss rate has decreased by 2.5 percentage points, and the annual power saving can reach 42.3MWh. Especially in terms of system response time, thanks to the improved Petri net control model, the average execution time of inter-phase load adjustment is only 1.2s, which is 57.1% higher than the second stage. The data in Table 2 show that the system has achieved dynamic balance of three-phase loads, the load difference of each phase is controlled within ±3%, and the adjustment process is stable and controllable, without oscillation.

[0079] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for optimizing and controlling an urban power grid based on phase balance of a distribution network, characterized in that: include: By deploying intelligent sensors and measurement and control devices at key nodes of the distribution network, the distribution network operation data can be obtained in real time; Based on the collected distribution network operation data, load forecasting is carried out and load clustering is performed through clustering algorithm. Based on the load clustering results and load results, the load status is evaluated; Taking into account the grid topology, distribution network operation data, load forecasting results and load assessment results, multi-agent collaborative optimization is used to determine the optimal load dispatch instructions; The obtained optimal dispatching instructions are converted into actual control actions based on phase balance to optimize the control of the power grid.

2. The urban power grid optimization control method based on phase balance of distribution network according to claim 1, characterized in that: Obtaining distribution network operation data includes: Smart sensors continuously collect raw distribution network operation data and perform local preprocessing to filter out abnormal values; The intelligent power distribution terminal performs frequency domain analysis on the data through discrete Fourier transform algorithm to form a standardized data frame; Encapsulate the data frame and transmit it to the distribution automation master station through the MMS protocol; The distribution automation master station parses the data frame and stores the data in the real-time database.

3. The urban power grid optimization control method based on distribution network phase balance according to claim 1, characterized in that: Load forecasting and assessment include: Clean and standardize the raw data to remove noise and outliers; Through clustering algorithms, users are grouped according to the similarity of load curves, typical load patterns are identified, and load clustering is completed; Combine historical data with real-time measurements to make load forecasts; According to the load clustering and load forecasting results, the load status is evaluated in combination with the three-phase imbalance index and the potential overload risk index.

4. The urban power grid optimization control method based on distribution network phase balance according to claim 3 is characterized in that: The load forecasting using the support vector regression model with adaptive weights can be expressed as follows: Among them, x i is the input feature vector, x t is the characteristic vector at time t, α i and is the Lagrange multiplier, b is the bias term, K(·,·) is the kernel function, and N is the total number of eigenvectors.

5. The urban power grid optimization control method based on distribution network phase balance according to claim 1, characterized in that: The multi-agent collaborative optimization determines the optimal load dispatch instructions including: The intelligent agent obtains the power grid topology model, real-time distribution network operation data, load forecasting results and load assessment results of the area under its jurisdiction, and exchanges information with adjacent intelligent agents; each intelligent agent independently solves a multi-objective optimization problem. The objective function includes minimizing the three-phase load imbalance, minimizing the network loss and maximizing the voltage qualification rate, and the optimal solution is used as the optimal load scheduling instruction.

6. The urban power grid optimization control method based on distribution network phase balance according to claim 5, characterized in that: The agent uses an adaptive differential evolution algorithm to solve the objective function, which can be expressed as follows: in, is the mutation vector, are three different individuals randomly selected from the tth generation, and F is the scaling factor.

7. The urban power grid optimization control method based on distribution network phase balance according to claim 1, characterized in that: Converting the optimal scheduling instructions into actual control actions includes: After receiving the load transfer plan issued by the distribution automation master station, the intelligent distribution terminal first performs a safety check to evaluate the feasibility of the plan. After the evaluation is completed, the overall plan is decomposed into a switch operation sequence, and the switch operation sequence is modeled and optimized according to the spatial position and constraint relationship of the switch.

8. The urban power grid optimization control method based on distribution network phase balance according to claim 7, characterized in that: Based on the modeling and optimization of the switch operation sequence, the intelligent distribution terminal uses a heuristic search algorithm to search for the optimal switch operation path and generate control instructions. The control instructions are broadcast to each switch device through the GOOSE protocol of IEC 61850 to complete the real-time transfer of load. At the same time, the intelligent distribution terminal monitors the load response in real time and feeds back the execution results to the distribution automation master station.

9. The urban power grid optimization control method based on distribution network phase balance according to claim 7, characterized in that: The Petri net model is used to model and optimize the switch operation sequence, and its state equation can be expressed by the following formula: M(t+1)=M(t)+A T ·U(t) (3) Among them, M(t) is the identification vector at time t, A is the input matrix, and U(t) is the control vector.

10. An urban power grid optimization control system based on distribution network interphase balance, running an urban power grid optimization control method based on distribution network interphase balance according to any one of claims 1 to 9, characterized in that: Data acquisition module, used to obtain power grid operation parameters in real time; Load analysis module, used to perform load forecasting and load evaluation based on power grid operation parameters; Optimization decision module, used to determine the optimal load dispatch instructions; The control execution module is used to convert the optimal load scheduling instructions into actual control actions.

Images