Security constraint-based power grid load prediction method and regulation and control system

Abstract

The invention relates to the technical field of power grid load prediction, in particular to a power grid load prediction method based on security constraints and a regulation and control system. The method comprises the following steps: acquiring advanced sensor deployment data; performing multi-source data stream acquisition on the advanced sensor deployment data to obtain a multi-source power edge data set; performing power data association on the multi-source power edge data set to generate multi-source power association data; performing data fusion on the multi-source power edge data set through the multi-source power associated data to generate a power fusion data set; performing power grid feature extraction on the power fusion data set to obtain photovoltaic power generation feature data and user power consumption feature data; and performing time sequence curve superposition on the photovoltaic power generation characteristic data and the user power consumption characteristic data to generate a power grid power time sequence superposition curve. Through multi-source data fusion, dynamic load prediction, security attack protection and a user feedback mechanism, the timeliness, accuracy and security of power grid prediction scheduling are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid load forecasting, and in particular to a power grid load forecasting method and a control system based on safety constraints. Background Art

[0002] Early load forecasting mainly relied on historical data and simple statistical models, such as time series analysis and regression models. These methods have certain effects in predicting short-term loads, but due to the complex and changeable power grid environment, it is difficult to cope with medium- and long-term fluctuations and emergencies. With the improvement of computing power, artificial intelligence technology has gradually been introduced into load forecasting, especially neural networks, support vector machines and other algorithms are used to deal with complex nonlinear relationships, which significantly improves the prediction accuracy. However, the problem of power grid security has become more and more prominent, especially in high load, extreme weather or equipment failure, traditional methods are difficult to take into account both security constraints and load forecasting accuracy. In recent years, load forecasting methods based on security constraints have gradually developed. Such methods not only focus on the accuracy of load forecasting, but also integrate power grid security constraints (such as voltage stability and line capacity) into the prediction model. Through the combination of dynamic optimization and deep learning technology, they can effectively cope with the uncertainty and risks in power grid operation. However, the current traditional prediction methods are usually based on static data, which is difficult to cope with the rapidly changing load conditions in power grid operation. At the same time, there is a lack of effective protection mechanism in the face of security attacks, which leads to low timeliness, accuracy and security of power grid prediction and scheduling. Summary of the invention

[0003] Based on this, it is necessary to provide a power grid load forecasting method and control system based on safety constraints to solve at least one of the above technical problems.

[0004] To achieve the above object, a power grid load forecasting method based on safety constraints is provided, the method comprising the following steps:

[0005] Step S1: Acquire advanced sensor deployment data; collect multi-source data streams for the advanced sensor deployment data to obtain a multi-source power edge data set; perform power data association on the multi-source power edge data set to generate multi-source power association data; perform data fusion on the multi-source power edge data set through the multi-source power association data to generate a power fusion data set;

[0006] Step S2: Extract grid characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data and user power consumption characteristic data; overlap the time series curves of the photovoltaic power generation characteristic data and the user power consumption characteristic data to generate a grid power time series overlap curve; perform grid dynamic balance calculation on the grid power time series overlap curve to obtain a grid dynamic balance coefficient; perform grid load joint prediction on the photovoltaic power generation characteristic data and the user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint prediction data;

[0007] Step S3: Use the power grid load joint prediction data to dispatch the energy storage device charging and discharging resources on the power fusion data set to generate energy storage device dispatching data; perform security constraint attack simulation on the energy storage device dispatching data to generate a power grid dispatching attack simulation data set; use the power grid dispatching attack simulation data set to perform dispatching security protection on the energy storage device dispatching data to generate security constraint power grid dispatching data;

[0008] Step S4: Monitor the real-time dispatching status of the power grid based on the safety-constrained power grid dispatching data to generate a real-time operation feedback data set of the power grid; collect user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjust the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set, and generate a power grid load regulation report to execute the power grid load regulation monitoring operation.

[0009] The present invention collects multi-source power data from advanced sensors and performs data fusion to generate a power fusion data set that integrates key information of photovoltaic power generation and user power consumption. This fusion improves the integrity and accuracy of the data and helps power grid managers to fully understand the dynamic changes in power production and consumption. By calculating the time sequence overlap and dynamic balance of photovoltaic power generation and user power consumption characteristic data, the generated power grid dynamic balance coefficient helps to jointly predict the power grid load. This dynamic balance mechanism can not only predict the changing trend of the power grid load, but also help the power grid to better allocate resources during peak power consumption periods, reduce energy consumption and overload risks. By optimizing the dispatch of the joint prediction data of the power grid load, the charging and discharging operations of the energy storage equipment are reasonably allocated to ensure that the power system can effectively respond to demand fluctuations. In addition, by simulating attack scenarios, the dispatch data is protected, which helps to deal with malicious attacks and system vulnerabilities and ensure the safe operation of the power grid. The user demand feedback mechanism enables the power grid system to adjust according to the actual power consumption and flexibly respond to changes in power demand. At the same time, the generated power grid load control report provides detailed load control data for power grid managers to help optimize the power grid dispatch strategy. Real-time monitoring of safety-constrained grid dispatch data ensures that the operation status of the grid is always within the controllable range, and the real-time feedback mechanism enables the system to quickly respond to potential faults or abnormal conditions, reducing the possibility of system shutdown or power supply interruption. Therefore, the present invention improves the timeliness, accuracy and security of grid prediction and dispatch through multi-source data fusion, dynamic load forecasting, security attack protection and user feedback mechanism.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: Acquire advanced sensor deployment data of the photovoltaic power generation system, user power equipment and power grid system, wherein the advanced sensors include photovoltaic module monitoring sensors, user power usage sensors and power grid status sensors;

[0012] Step S12: performing multi-source data stream acquisition on the advanced sensor deployment data to obtain an original multi-source power data set; performing data preprocessing on the original multi-source power data set to generate a standard multi-source power data set, wherein the data preprocessing includes data cleaning, data denoising, data normalization and data standardization;

[0013] Step S13: performing edge computing on the standard multi-source power data set to generate a multi-source power edge data set;

[0014] Step S14: performing power data association on the multi-source power edge data sets to generate multi-source power association data; performing data fusion on the multi-source power edge data sets through the multi-source power association data to generate a power fusion data set.

[0015] The present invention can realize comprehensive multi-source power data collection and fusion by acquiring advanced sensor data of photovoltaic power generation systems, user power equipment and power grid systems. The correlation and fusion of data are used to help improve the overall efficiency of power management and achieve more accurate load forecasting and power distribution. Through pre-processing steps such as data cleaning, denoising, normalization and standardization, a standardized multi-source power data set is generated to ensure data consistency and accuracy. At the same time, edge computing can perform real-time processing close to the data source, reducing the burden on the central server and improving the speed and efficiency of overall data processing. Power data association and data fusion enable the system to analyze and process power data from multiple dimensions, improving the intelligence level of the power system. Through the fusion of multi-source power edge data, the dispatching and distribution of power can be optimized at multiple levels, further improving the stability and reliability of the system. Advanced sensor deployment provides real-time and accurate data, and combined with data fusion technology, it can better cope with complex changes in power demand, timely adjust the load balance of photovoltaic power generation and power grid, and ensure the stability of power supply and the matching of user needs.

[0016] Preferably, step S2 comprises the following steps:

[0017] Step S21: extract photovoltaic power generation characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data; extract user power consumption characteristics from the power fusion data set to obtain user power consumption characteristic data;

[0018] Step S22: performing a power generation time series analysis on the photovoltaic power generation characteristic data to generate photovoltaic power generation time series data; performing a user power consumption time series analysis on the user power consumption characteristic data to generate user power consumption time series data; performing curve conversion on the photovoltaic power generation time series data and the user power consumption time series data to generate a photovoltaic power generation time series curve and a user power consumption time series curve;

[0019] Step S23: Overlap the photovoltaic power generation time sequence curve and the user power consumption time sequence curve to generate a grid power time sequence overlap curve; segment the grid power time sequence overlap curve to generate grid power non-overlapping line segments and grid power overlapping line segments;

[0020] Step S24: using a power grid dynamic balance calculation formula to perform power grid dynamic balance calculation on power grid non-coincidence line segments and power grid coincidence line segments to obtain a power grid dynamic balance coefficient;

[0021] Step S25: Perform grid load joint prediction on the photovoltaic power generation characteristic data and the user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint prediction data.

[0022] The present invention can deeply analyze and understand the dynamic change characteristics of power generation and power consumption by extracting the characteristic data of photovoltaic power generation and user power consumption from the power fusion data set, and provide accurate basic data support for subsequent time series analysis and prediction. The characteristic data of photovoltaic power generation and user power consumption are analyzed in time series, and a time series curve is generated, which can vividly display the changing trend of power generation and power consumption, so that the power management system can more intuitively grasp the time series relationship between photovoltaic power generation and user power consumption. By overlapping the photovoltaic power generation time series curve and the user power consumption time series curve, the overlapping and non-overlapping parts of power generation and power consumption can be effectively identified. The non-overlapping line segments of power grid power reflect the imbalance between power supply and demand, and the overlapping line segments show the balance part of supply and demand. This segmentation provides a clear data basis for subsequent power grid load balance calculation and prediction. The dynamic balance calculation formula of the power grid is used to perform dynamic balance calculation on the overlapping and non-overlapping line segments, and the generated power grid dynamic balance coefficient can reflect the balance status of the power grid in different time periods. Through these coefficients, the system can adjust the matching of power generation and power consumption in real time, optimize the load management of the power grid, and ensure the stability and safety of the power system. Based on the dynamic balance coefficient of the power grid, the combined load forecast of photovoltaic power generation characteristics and user power consumption characteristics data can provide more accurate reference data for power dispatching. By accurately predicting future loads, the power grid can optimize the dispatching strategy in advance, reduce the risk of insufficient or excessive power supply, and improve the efficiency of the overall power system.

[0023] Preferably, the power grid dynamic balance calculation formula in step S24 is as follows:

[0024]

[0025] Where B is the dynamic balance coefficient of the power grid, P net (t) is the net power at time t, P loss (t) is the power loss at time t, P gen (t) represents the total power generation at time t, t0 represents the lower limit of the measurement time, t1 represents the upper limit of the measurement time, τ represents the time constant, k i Expressed as the amplitude of the i-th frequency component, ω i Expressed as the angular frequency of the i-th frequency component, φ i is represented as the phase angle of the i-th frequency component, and n is represented as the number of frequency components.

[0026] The present invention analyzes and integrates a dynamic balance calculation formula for the power grid, which is designed to evaluate the dynamic balance state of the power grid in a specific time period, taking into account the net power, loss, power generation and the impact of various periodic fluctuations on the balance of the power grid. The performance of the power grid in the time dimension is comprehensively analyzed by integration and addition, where the integral term in the formula The difference P between the actual load of the grid and the power generation is taken into account net (t), which reflects the supply and demand status of the power grid at a certain moment. Considering the power P that cannot be used in the power grid due to the loss of transmission lines and equipment loss (t), to ensure more accurate calculation. Total power generation P gen (t), as a benchmark for the power supply capacity of the power grid. By using the relative proportion method, the efficiency of the power grid can be observed more clearly. It reflects the influence of time on the balance of the power grid and reflects the lag of the system response. The time constant τ determines the intensity of this influence. The frequency component term in the formula Considering the amplitude parameter k i , which represents the relative influence of different frequency components on the dynamic balance of the power grid. Different frequency components correspond to different load modes. The periodic function sin(ω i t+φ i ), which reflects the periodic characteristics of various load changes in the power grid. By combining components of different frequencies together, the fluctuations of the power grid can be fully reflected. When using the conventional power grid dynamic balance calculation formula in the field, the power grid dynamic balance coefficient can be obtained. By applying the power grid dynamic balance calculation formula provided by the present invention, the power grid dynamic balance coefficient can be calculated more accurately. The formula can accurately evaluate the dynamic balance state of the power grid in different time periods, thereby providing a scientific basis for power grid dispatching and management. By introducing the power loss factor, a comprehensive analysis of the power grid efficiency is ensured, making the power grid operation more efficient. The introduction of frequency components enhances the adaptability of the power grid to various load modes and can better cope with emergencies in actual operation. Considering the influence of the time constant, the formula is more in line with the dynamic characteristics of the actual power grid, helping to optimize the dispatching decision. The structure of the formula allows the parameters to be adjusted according to the actual situation, so as to adapt to the needs of different types of power grids. In summary, the relationship between the formula and its parameters not only enhances the accuracy of the dynamic balance analysis of the power grid, but also provides an effective tool for optimizing power grid management and dispatching.

[0027] Preferably, step S25 includes the following steps:

[0028] Step S251: Perform power generation environment analysis on photovoltaic power generation characteristic data according to the dynamic balance coefficient of the power grid to generate photovoltaic power generation environment data; perform power consumption pattern analysis on user power consumption characteristic data according to the dynamic balance coefficient of the power grid to generate user power consumption pattern data;

[0029] Step S252: divide the photovoltaic power generation environment data and the user power consumption mode data into data sets to generate a model training set and a model test set; perform model training on the model training set using a long short-term memory network algorithm to generate a power grid load joint prediction pre-model;

[0030] Step S253: Use the model test set to iterate the model optimization of the power grid load joint prediction pre-model, so as to generate a power grid load joint prediction model; import the photovoltaic power generation characteristic data and the user's electricity consumption characteristic data into the power grid load joint prediction model to perform power grid joint load prediction and generate power grid load joint prediction data.

[0031] The present invention generates more targeted power generation environment data and user power consumption pattern data by performing detailed environmental analysis and pattern analysis on photovoltaic power generation characteristics and user power consumption characteristics according to the dynamic balance coefficient of the power grid, which helps to understand how the external environment affects photovoltaic power generation, as well as the user's power consumption behavior and habits, and provides more accurate and relevant basic data for load forecasting. By dividing the power generation environment data and power consumption pattern data into a model training set and a test set, and using a long short-term memory network (LSTM) for training, the dynamic change trend of time series data in the power system can be effectively captured. LSTM is good at processing long-term dependent time series data, and can significantly improve the prediction accuracy of the model in complex power scenarios. The model test set is used to optimize and iterate the preliminarily generated power grid load joint prediction pre-model, and the prediction ability of the model is continuously adjusted and improved. This optimization process can improve the generalization ability of the model, making it not only suitable for training data, but also maintaining good prediction performance on new data sets, and improving the overall stability and accuracy of the model. The finally generated power grid load joint prediction model can comprehensively import photovoltaic power generation feature data and user power consumption feature data to perform more accurate power grid load prediction. Such forecasts help the power system to dispatch in advance and balance power generation and consumption, especially to cope with the uncertainty of photovoltaic power generation and the volatility of user power demand. Through the joint forecast data of power grid load, the power dispatching system can adjust the power generation and power supply plan more flexibly and intelligently, reducing the risk of power overload or insufficient power supply. This optimization can not only ensure the stable operation of the power grid, but also improve the utilization rate of power resources and reduce waste.

[0032] Preferably, step S3 comprises the following steps:

[0033] Step S31: Performing grid load balance analysis on the power fusion data set through grid load joint prediction data to generate a grid load balance matrix; performing input scheduling optimization on the grid load balance matrix to generate grid load scheduling optimization data;

[0034] Step S32: matching the grid load dispatch optimization data with peak period characteristics to generate peak grid demand data; matching and adjusting the grid load dispatch optimization data according to the peak grid demand data to generate grid load matching adjustment data, wherein the matching and adjustment includes smoothing photovoltaic power generation fluctuations and adjusting user power consumption behavior;

[0035] Step S33: Performing energy storage device charging and discharging resource scheduling based on the grid load matching adjustment data to generate energy storage device scheduling data; performing grid-friendliness analysis on the energy storage device scheduling data to generate grid energy storage scheduling friendliness data;

[0036] Step S34: Perform security constraint attack simulation on energy storage device dispatch data using power grid energy storage dispatch friendliness data to generate a power grid dispatch attack simulation data set; perform dispatch security protection on energy storage device dispatch data using the power grid dispatch attack simulation data set to generate security constraint power grid dispatch data.

[0037] The present invention can more comprehensively understand the balance state between power supply and demand of the power grid by performing load balance analysis on the power fusion data set and generating a power grid load balance matrix. According to this balance state, the system can perform more accurate input scheduling optimization, thereby ensuring the efficiency and stability of power grid load scheduling. Through the analysis of peak power grid demand, the power system can effectively cope with the imbalance of supply and demand during peak power consumption. The smoothing of photovoltaic power generation fluctuations and the adjustment of user power consumption behavior make the power supply system more stable during peak periods, reduce the risks caused by power shortages or overloads, and improve the reliability of the power supply system. The charging and discharging scheduling of the energy storage device is optimized and scheduled through the power grid load matching adjustment data. The energy storage device can release electric energy when the power grid load is high and store electric energy when the load is low, thereby alleviating the imbalance of power supply and demand. The intelligent scheduling of the energy storage device not only improves the utilization efficiency of electric energy, but also provides strong support for the smooth operation of the power grid. Through the power grid friendliness analysis, the system can optimize the scheduling of the energy storage device to make it more consistent with the supply and demand changes of the power grid, avoid overloading or ineffective operation of the energy storage device, and this friendliness analysis improves the coordination between the energy storage system and the power grid, and enhances the adaptability of the power grid. By simulating potential security constraint attacks and protecting the energy storage equipment dispatch data, we can effectively identify and resist potential network attacks and ensure the safe operation of energy storage equipment in the power grid. By generating security constraint grid dispatch data, the power grid system can have higher security during the dispatch process, resist malicious attacks from the outside, and ensure the reliability and robustness of power dispatch.

[0038] Preferably, step S34 includes the following steps:

[0039] Step S341: Perform intrusion detection scenario simulation on energy storage device dispatch data through power grid energy storage dispatch friendliness data to generate a power grid attack scenario set, wherein the power grid attack scenario set includes a denial of service attack scenario and a false data injection attack scenario;

[0040] Step S342: Based on the denial of service attack scenario, a dispatch load response attack simulation is performed on the energy storage device dispatch data to generate dispatch response attack simulation data; based on the false data injection attack scenario, a misleading voltage and frequency regulation attack simulation is performed on the energy storage device dispatch data to generate false regulation attack simulation data;

[0041] Step S343: integrating the dispatch response attack simulation data and the false control attack simulation data to generate a power grid dispatch attack simulation data set; performing a dispatch risk assessment on the energy storage device dispatch data according to the power grid dispatch attack simulation data set to generate dispatch risk assessment data;

[0042] Step S344: Use the power grid dispatch attack simulation data set to perform dispatch security protection on the energy storage device dispatch data to generate security-constrained power grid dispatch data.

[0043] The present invention can comprehensively identify and analyze security threats by simulating intrusion detection scenarios on energy storage device scheduling data and generating a power grid attack scenario set. This simulation not only helps identify potential risks such as denial of service attacks and false data injection attacks, but also provides data support for subsequent protective measures. Based on different attack scenarios (such as denial of service attacks and false data injection attacks), the simulation of dispatching load response attacks and misleading voltage and frequency regulation attacks can be concretized. This targeted simulation provides a realistic attack scenario for the energy storage device scheduling system, which helps the system understand how to deal with various security challenges and thus formulate effective protection strategies. By integrating the dispatching response attack simulation data and the false regulation attack simulation data to generate a power grid dispatching attack simulation data set, the impact of various attacks on the dispatching system can be comprehensively evaluated. This integrated approach improves the efficiency of risk identification and evaluation, allowing the system to more accurately locate potential safety hazards. Dispatching risk assessment based on the power grid dispatching attack simulation data set can provide a scientific basis for the dispatching of energy storage equipment. The evaluation results will help managers understand the potential risks in the face of different types of attacks, provide support for dispatching decisions, and ensure the safety and effectiveness of the dispatching process. By using the power grid dispatch attack simulation data set to perform security protection on the energy storage device dispatch data, the generated security-constrained power grid dispatch data can effectively respond to identified security threats. This not only improves the security of power dispatch, but also ensures the stable operation of the energy storage system and reduces the risk of power outages caused by network attacks. Step S34 improves the power system's protection and response capabilities in the face of network attacks through intrusion detection, targeted attack simulation, risk assessment and security protection. These measures can not only ensure the safety of energy storage equipment, but also enhance the reliability and resilience of the entire power dispatch system, providing important support for the stable operation of the power grid.

[0044] Preferably, step S344 includes the following steps:

[0045] Step S3441: abnormal monitoring of the energy storage device dispatching data to generate abnormal dispatching data; scenario identification and matching of the abnormal dispatching data and the power grid dispatching attack simulation data set, when a denial of service attack scenario is identified, scheduling packet loss rate analysis is performed on the energy storage device dispatching data to generate scheduling packet loss rate data; first security constraints are performed on the energy storage device dispatching data according to the scheduling packet loss rate data to generate first security constraint data;

[0046] Step S3442: when a false data injection attack scenario is identified, the energy storage device scheduling data is subjected to scheduling output over-analysis to generate scheduling over-output data; a second security constraint is applied to the energy storage device scheduling data according to the scheduling over-output data to generate second security constraint data;

[0047] Step S3443: construct a safety constraint engine according to the first safety constraint data and the second safety constraint data to generate a safety constraint engine; use the safety constraint engine to perform backup energy storage device switching on the energy storage device scheduling data to generate backup energy storage device scheduling switching data;

[0048] Step S3444: synchronously adjust the energy storage device scheduling data with the backup energy storage device scheduling switching data to generate power adjustment data; perform scheduling safety verification on the power adjustment data to generate safety-constrained power grid scheduling data.

[0049] The present invention can timely discover potential scheduling anomalies and generate scheduling anomaly data by performing abnormal monitoring on the scheduling data of energy storage equipment. This real-time monitoring mechanism helps to quickly identify and respond to abnormal scheduling behaviors and enhance the responsiveness of the system. By matching the scheduling anomaly data with the power grid scheduling attack simulation data set for scene recognition, the system can effectively identify different attack scenarios such as denial of service attacks and false data injection attacks. This combination makes risk assessment more accurate and provides a scientific basis for subsequent safety measures. When a denial of service attack is identified, the first safety constraint is performed through scheduling packet loss rate analysis, and when a false data injection attack is identified, the scheduling over-output analysis is performed and the second safety constraint is implemented. This targeted safety constraint mechanism can effectively respond to different types of attacks and improve the safety of energy storage equipment scheduling. Integrating the first and second safety constraint data to construct a safety constraint engine can provide comprehensive safety protection for energy storage equipment scheduling. This engine can automatically adjust the scheduling strategy according to different safety constraints in real time to ensure the safe operation of the system. By switching the backup energy storage device through the safety constraint engine, the backup energy storage device can be quickly mobilized when an anomaly or attack is detected. This intelligent switching mechanism ensures that when a problem occurs in the main device, the system can still maintain a stable power supply and reduce the risk of power outages. By synchronously adjusting the dispatching power of the backup energy storage equipment dispatching switching data, the continuity and stability of the power supply can be ensured during the switching process. At the same time, the dispatching safety verification link ensures the rationality and security of the power adjustment data, providing a solid guarantee for the stable operation of the power grid. Step S344 comprehensively improves the power system's ability to respond to security threats through abnormal monitoring, scene recognition, targeted security constraints, security constraint engine construction, intelligent switching and power adjustment. These measures enhance the safety and reliability of energy storage equipment, ensure the stable operation of the power grid, and reduce the risk of power outages caused by external attacks.

[0050] Preferably, step S4 comprises the following steps:

[0051] Step S41: Performing real-time Internet of Things configuration monitoring on the safety-constrained power grid dispatching data to generate real-time dispatching monitoring data; performing real-time power grid dispatching status monitoring on the real-time dispatching monitoring data to generate a real-time power grid operation feedback data set;

[0052] Step S42: collecting user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjusting the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set to generate power grid load dynamic regulation data;

[0053] Step S43: Visualize the power grid load dynamic regulation data to generate a power grid load regulation report to perform power grid load regulation monitoring operations.

[0054] The present invention performs real-time configuration monitoring of the Internet of Things on the dispatching data of the safety-constrained power grid, and the generated real-time dispatching monitoring data enables the power grid manager to understand the operation status of the power grid at any time. This real-time monitoring improves the response speed of power grid management and timely discovers and solves potential problems. By collecting user demand feedback on the real-time operation feedback data set of the power grid, the user's electricity demand and preference can be better understood. Timely user demand feedback helps the power grid system adjust the power supply strategy according to actual needs, improve the flexibility of power supply services and user satisfaction. Based on the user demand feedback data, the dynamic load control strategy adjustment of the real-time operation feedback data of the power grid can realize the intelligent regulation of the power grid load. This dynamic adjustment mechanism optimizes the utilization efficiency of power resources and avoids resource waste and insufficient power supply. The dynamic control data of the power grid load is visualized, and the generated power grid load control report provides an intuitive view to help managers better understand the operation of the power grid. Data visualization not only makes complex information easier to understand, but also supports the data-driven decision-making process and improves the scientific nature of decision-making. The generated power grid load control report provides a basis for the power grid load control monitoring operation, ensuring that effective monitoring and control can be carried out in actual operation. This monitoring mechanism improves the adaptability of the power grid to load changes and ensures the stability and safety of power supply. Step S4 effectively improves the flexibility and efficiency of power grid management through real-time monitoring, user demand feedback, dynamic load control, data visualization, and monitoring and regulation. These measures not only optimize the utilization of power resources, but also enhance the power grid system's ability to respond to user needs, ensuring the stability and security of power supply.

[0055] In this specification, a power grid load control system based on security constraints is provided, which is used to execute the above-mentioned power grid load prediction method based on security constraints. The power grid load control system based on security constraints includes:

[0056] A multi-source data fusion module is used to obtain advanced sensor deployment data; collect multi-source data streams for advanced sensor deployment data to obtain a multi-source power edge data set; perform power data association on the multi-source power edge data set to generate multi-source power association data; perform data fusion on the multi-source power edge data set through the multi-source power association data to generate a power fusion data set;

[0057] The joint load forecasting module is used to extract grid characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data and user power consumption characteristic data; overlap the time series curves of photovoltaic power generation characteristic data and user power consumption characteristic data to generate a grid power time series overlap curve; perform grid dynamic balance calculation on the grid power time series overlap curve to obtain a grid dynamic balance coefficient; perform grid load joint forecasting on photovoltaic power generation characteristic data and user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint forecasting data;

[0058] The security constraint adjustment module is used to dispatch the charging and discharging resources of energy storage equipment to the power fusion data set through the joint prediction data of the power grid load, and generate the energy storage equipment dispatching data; perform security constraint attack simulation on the energy storage equipment dispatching data, and generate the power grid dispatching attack simulation data set; use the power grid dispatching attack simulation data set to perform dispatching security protection on the energy storage equipment dispatching data, and generate the security constraint power grid dispatching data;

[0059] The real-time feedback monitoring module is used to monitor the real-time dispatching status of the power grid with safety constraints on the dispatching data of the power grid, and generate a real-time operation feedback data set of the power grid; collect user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjust the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set, and generate a power grid load regulation report to perform power grid load regulation monitoring operations.

[0060] The beneficial effect of the present invention is that more comprehensive and accurate power data is obtained through advanced sensor deployment and multi-source data stream collection. Through data association and fusion, a power fusion data set is generated, which improves the consistency and relevance of the data and ensures that subsequent predictions and scheduling are based on more accurate inputs. Feature extraction of the power fusion data set can obtain feature data of photovoltaic power generation and user power consumption respectively, and effectively reflect the dynamic changes of the load of the power grid through timing curve overlap and dynamic balance calculation, further improving the accuracy of the joint prediction of the power grid load. The charging and discharging scheduling of energy storage equipment is optimized by using the joint prediction data of the power grid load to achieve more efficient resource management. At the same time, through attack simulation and scheduling security protection, the security of power grid scheduling and the ability to resist network attacks are enhanced to ensure the stability of power grid operation. Through real-time scheduling status monitoring and user demand feedback mechanism, the power grid load regulation is more flexible and can be dynamically adjusted, which improves the adaptability and responsiveness of power grid operation, ensures that the power supply meets real-time needs, and optimizes the power grid regulation strategy through feedback mechanism. Therefore, the present invention improves the timeliness, accuracy and security of power grid prediction and scheduling through multi-source data fusion, dynamic load prediction, security attack protection and user feedback mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1A schematic diagram of the steps of a power grid load forecasting method based on security constraints;

[0062] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0063] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0064] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0065] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0066] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0067] It should be understood that although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit, and the term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0068] To achieve this, please refer to Figures 1 to 3 , a power grid load forecasting method based on security constraints, the method comprising the following steps:

[0069] Step S1: Acquire advanced sensor deployment data; collect multi-source data streams for the advanced sensor deployment data to obtain a multi-source power edge data set; perform power data association on the multi-source power edge data set to generate multi-source power association data; perform data fusion on the multi-source power edge data set through the multi-source power association data to generate a power fusion data set;

[0070] Step S2: Extract grid characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data and user power consumption characteristic data; overlap the time series curves of the photovoltaic power generation characteristic data and the user power consumption characteristic data to generate a grid power time series overlap curve; perform grid dynamic balance calculation on the grid power time series overlap curve to obtain a grid dynamic balance coefficient; perform grid load joint prediction on the photovoltaic power generation characteristic data and the user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint prediction data;

[0071] Step S3: Use the power grid load joint prediction data to dispatch the energy storage device charging and discharging resources on the power fusion data set to generate energy storage device dispatching data; perform security constraint attack simulation on the energy storage device dispatching data to generate a power grid dispatching attack simulation data set; use the power grid dispatching attack simulation data set to perform dispatching security protection on the energy storage device dispatching data to generate security constraint power grid dispatching data;

[0072] Step S4: Monitor the real-time dispatching status of the power grid based on the safety-constrained power grid dispatching data to generate a real-time operation feedback data set of the power grid; collect user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjust the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set, and generate a power grid load regulation report to execute the power grid load regulation monitoring operation.

[0073] The present invention collects multi-source power data from advanced sensors and performs data fusion to generate a power fusion data set that integrates key information of photovoltaic power generation and user power consumption. This fusion improves the integrity and accuracy of the data and helps power grid managers to fully understand the dynamic changes in power production and consumption. By calculating the time sequence overlap and dynamic balance of photovoltaic power generation and user power consumption characteristic data, the generated power grid dynamic balance coefficient helps to jointly predict the power grid load. This dynamic balance mechanism can not only predict the changing trend of the power grid load, but also help the power grid to better allocate resources during peak power consumption periods, reduce energy consumption and overload risks. By optimizing the dispatch of the joint prediction data of the power grid load, the charging and discharging operations of the energy storage equipment are reasonably allocated to ensure that the power system can effectively respond to demand fluctuations. In addition, by simulating attack scenarios, the dispatch data is protected, which helps to deal with malicious attacks and system vulnerabilities and ensure the safe operation of the power grid. The user demand feedback mechanism enables the power grid system to adjust according to the actual power consumption and flexibly respond to changes in power demand. At the same time, the generated power grid load control report provides detailed load control data for power grid managers to help optimize the power grid dispatch strategy. Real-time monitoring of safety-constrained grid dispatch data ensures that the operation status of the grid is always within the controllable range, and the real-time feedback mechanism enables the system to quickly respond to potential faults or abnormal conditions, reducing the possibility of system shutdown or power supply interruption. Therefore, the present invention improves the timeliness, accuracy and security of grid prediction and dispatch through multi-source data fusion, dynamic load forecasting, security attack protection and user feedback mechanism.

[0074] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a power grid load forecasting method based on safety constraints of the present invention. In this example, the power grid load forecasting method based on safety constraints includes the following steps:

[0075] Step S1: Acquire advanced sensor deployment data; collect multi-source data streams for the advanced sensor deployment data to obtain a multi-source power edge data set; perform power data association on the multi-source power edge data set to generate multi-source power association data; perform data fusion on the multi-source power edge data set through the multi-source power association data to generate a power fusion data set;

[0076] In the embodiment of the present invention, photovoltaic power generation monitoring sensors are deployed to collect information such as power generation, temperature, and irradiance of photovoltaic modules. User power usage sensors are deployed to collect power usage of user devices, including voltage, current, power, and other data. Grid status sensors are deployed to monitor the operation status of the grid in real time, such as grid voltage, current, and frequency. Sensor deployment data is collected to obtain real-time power data from multiple sources to form a complete advanced sensor deployment data set. Using the Internet of Things communication protocol and edge computing nodes, real-time data streams are collected from photovoltaic power generation, user power consumption, and grid status sensors to obtain the original multi-source power data set. Data cleaning is performed on the original multi-source power data set to remove outliers and missing values. Denoising is performed on the data to ensure the accuracy and reliability of the power data. The data collected by different sensors are normalized and standardized to generate a multi-source power edge data set in a unified format. Combine photovoltaic power generation data, user power consumption data, and grid status data to perform cross-source correlation analysis to extract the relationship between the three, such as the correlation between photovoltaic power generation and user power load changes. Using the data association model, we conduct association analysis on multi-source power edge data sets to generate a multi-source power association data set containing multi-dimensional association features. Using the data fusion algorithm, we fuse photovoltaic power generation data, user power consumption data, and grid status data to comprehensively analyze the relationship and impact between different data sources. The fused data includes the real-time linkage between the photovoltaic power generation system and the user's power equipment, the dynamic balance state of the power grid, etc. The fusion results form a power fusion data set, which provides data support for subsequent dynamic grid dispatch and load forecasting.

[0077] Step S2: Extract grid characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data and user power consumption characteristic data; overlap the time series curves of the photovoltaic power generation characteristic data and the user power consumption characteristic data to generate a grid power time series overlap curve; perform grid dynamic balance calculation on the grid power time series overlap curve to obtain a grid dynamic balance coefficient; perform grid load joint prediction on the photovoltaic power generation characteristic data and the user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint prediction data;

[0078] In the embodiment of the present invention, by extracting features from the photovoltaic power generation related data in the power fusion data set, the extracted features include power generation, photovoltaic system efficiency, solar irradiance, temperature, etc., to generate photovoltaic power generation feature data. By extracting features from the user power consumption related data in the power fusion data set, the extracted features include the user's power consumption mode, power consumption, voltage, current, etc., to generate user power consumption feature data. By performing time series analysis on the photovoltaic power generation feature data, the photovoltaic power generation time series data that changes with time is obtained, and the data is converted into a photovoltaic power generation time series curve. By performing time series analysis on the user power consumption feature data, the user power consumption time series data that changes with time is obtained, and the data is converted into a user power consumption time series curve. The photovoltaic power generation time series curve and the user power consumption time series curve are overlapped and analyzed to generate a power grid power time series overlap curve, reflecting the dynamic relationship between photovoltaic power generation and user power consumption. By analyzing the power grid power time series overlap curve, the balance point and unbalance point between power generation and power consumption in the power grid are found. By using the power grid dynamic balance calculation formula, the supply and demand of photovoltaic power generation and user power consumption in different time periods are calculated to obtain the power grid dynamic balance coefficient. The dynamic balance coefficient reflects the supply and demand balance between photovoltaic power generation and user electricity consumption in the power grid in each time period. According to the dynamic balance coefficient of the power grid, the photovoltaic power generation characteristic data and the user electricity consumption characteristic data are analyzed to identify the laws of photovoltaic power generation fluctuations and user electricity consumption patterns. Based on these characteristics, a joint prediction model for power grid load is established. The photovoltaic power generation characteristic data and the user electricity consumption characteristic data are imported into the joint prediction model for power grid load to perform joint power grid load prediction. Through model prediction, the joint prediction data for power grid load is generated for subsequent power grid load scheduling and optimization.

[0079] Step S3: Use the power grid load joint prediction data to dispatch the energy storage device charging and discharging resources on the power fusion data set to generate energy storage device dispatching data; perform security constraint attack simulation on the energy storage device dispatching data to generate a power grid dispatching attack simulation data set; use the power grid dispatching attack simulation data set to perform dispatching security protection on the energy storage device dispatching data to generate security constraint power grid dispatching data;

[0080] In an embodiment of the present invention, the load demand of the power grid in the future time period is analyzed according to the joint prediction data of the power grid load, including the load fluctuation during the peak and valley periods of power supply. The load of the power grid is balanced by using energy storage equipment, and the energy storage is released during the peak period of power and charged during the valley period of power. According to the load fluctuation demand, the charging and discharging time and capacity of the energy storage equipment are dynamically adjusted to generate energy storage equipment scheduling data. The scheduling data is used to optimize the charging and discharging operation of the energy storage equipment, thereby improving the operating efficiency and stability of the power grid. According to the energy storage equipment scheduling data, an attack scenario simulation is introduced to simulate potential network attacks, such as denial of service attack (DoS) and false data injection attack (FDIA). By simulating different attack scenarios, a power grid scheduling attack simulation data set is generated, which includes the risk of attack on the energy storage equipment scheduling process, wherein the denial of service attack scenario is to simulate a large number of invalid requests so that the energy storage equipment cannot respond normally, and generate scheduling response attack simulation data. The false data injection attack scenario is to simulate the injection of misleading voltage or frequency data into the system to interfere with the normal operation of the energy storage equipment and generate false regulation attack simulation data. Different types of attack simulation data are integrated to form a grid dispatch attack simulation data set, which fully demonstrates the security threats faced by energy storage equipment during dispatch. According to the grid dispatch attack simulation data set, a detailed risk assessment is conducted on the energy storage equipment dispatch data to identify the security risks in the dispatch process. Through risk assessment, dispatch risk assessment data is generated to help identify and predict the impact of attacks on the dispatch operation of energy storage equipment. According to the dispatch risk assessment results, a corresponding security constraint mechanism is established, including protection measures for different attack scenarios. For example, for denial of service attacks, the first security constraint is performed by analyzing the dispatch packet loss rate; for false data injection attacks, the second security constraint is performed by monitoring the dispatch output excess. The established security constraint mechanism is used to dynamically adjust the energy storage equipment dispatch data to ensure that it can switch to the backup energy storage equipment or adjust the dispatch strategy in time when an attack occurs, generate security-constrained grid dispatch data, and ensure the security and stability of dispatch operations under network attack conditions.

[0081] Step S4: Monitor the real-time dispatching status of the power grid based on the safety-constrained power grid dispatching data to generate a real-time operation feedback data set of the power grid; collect user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjust the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set, and generate a power grid load regulation report to execute the power grid load regulation monitoring operation.

[0082] In the embodiment of the present invention, the real-time operation status of the power grid is monitored based on the safety constraint power grid dispatching data, and multi-dimensional data such as power transmission, load distribution, and energy storage equipment operation status are collected to generate a real-time operation feedback data set of the power grid. Real-time monitoring of key parameters such as voltage, current, and frequency of the power grid ensures the stable operation of the power grid. Anomaly detection is performed on the real-time monitoring data, and alarm information is generated in time, such as load overload, abnormal voltage fluctuation, abnormal charging and discharging of energy storage equipment, etc. The monitored status data is fed back to the system background to form a real-time operation feedback data set of the power grid for subsequent analysis and processing. The power consumption behavior data of the user end, including information such as power consumption, power consumption time period preference, and load demand, is collected to form a user demand feedback data set. The user's demand characteristics are classified, such as household electricity, industrial electricity, commercial electricity, etc., so as to more accurately adjust the load control strategy. The user demand feedback data set is combined with the real-time operation feedback data set of the power grid to dynamically adjust the load control strategy of the power grid. Based on the real-time power demand of the user, load fluctuations are predicted, and the power grid dispatching adjustment is performed in advance to balance supply and demand. According to the real-time grid operation feedback data set and user demand feedback data set, analyze the matching of current load distribution and power supply and demand, and dynamically adjust the load control strategy. Optimize the balance of grid load by scheduling the charging and discharging behavior of energy storage equipment, adjusting the power distribution path, and reasonably allocating power resources. After the load control strategy is adjusted, generate a grid load regulation report, including grid load distribution, energy storage equipment scheduling status, supply and demand balance, etc. The report should also include predictions and suggestions for future loads to guide subsequent scheduling monitoring and operations. According to the generated grid load regulation report, the system implements real-time load regulation operations, dynamically schedules the charging and discharging of energy storage equipment, and adjusts the power transmission path to balance load fluctuations. Continuously monitor the operation of the grid to ensure the effectiveness of the regulation strategy; when new load fluctuations or changes in user demand are detected, timely feedback is provided and a new round of scheduling adjustments is made.

[0083] Preferably, step S1 comprises the following steps:

[0084] Step S11: Acquire advanced sensor deployment data of the photovoltaic power generation system, user power equipment and power grid system, wherein the advanced sensors include photovoltaic module monitoring sensors, user power usage sensors and power grid status sensors;

[0085] Step S12: performing multi-source data stream acquisition on the advanced sensor deployment data to obtain an original multi-source power data set; performing data preprocessing on the original multi-source power data set to generate a standard multi-source power data set, wherein the data preprocessing includes data cleaning, data denoising, data normalization and data standardization;

[0086] Step S13: performing edge computing on the standard multi-source power data set to generate a multi-source power edge data set;

[0087] Step S14: performing power data association on the multi-source power edge data sets to generate multi-source power association data; performing data fusion on the multi-source power edge data sets through the multi-source power association data to generate a power fusion data set.

[0088] In an embodiment of the present invention, photovoltaic module monitoring sensors are deployed to collect photovoltaic power generation data, including power generation, temperature and light intensity. User power usage sensors are deployed to monitor the power consumption of user power devices and record power consumption patterns and peak values. Grid status sensors are deployed to obtain real-time data such as grid load, frequency and voltage. Multi-source data stream acquisition is performed to integrate data from different sensors into an original multi-source power data set. The original data set is preprocessed, including: data cleaning: removing missing values ​​and outliers. Data denoising: using filters or signal processing techniques to remove noise. Data normalization: adjusting the data to a uniform range for comparison. Data standardization: making the data normally distributed and enhancing the stability of the model. Data processing tasks are performed on edge devices to reduce transmission delays. Data aggregation and preliminary analysis are performed to generate a multi-source power edge data set, including real-time power monitoring results and statistical data. The data in the multi-source power edge data set is associated using an association algorithm (such as time series analysis and correlation analysis) to generate multi-source power associated data. By fusing edge data sets with multi-source power-related data and using fusion techniques (such as weighted averaging, Kalman filtering, etc.), a comprehensive power fusion data set is generated to provide more accurate power status and predictions.

[0089] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0090] Step S21: extract photovoltaic power generation characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data; extract user power consumption characteristics from the power fusion data set to obtain user power consumption characteristic data;

[0091] Step S22: performing a power generation time series analysis on the photovoltaic power generation characteristic data to generate photovoltaic power generation time series data; performing a user power consumption time series analysis on the user power consumption characteristic data to generate user power consumption time series data; performing curve conversion on the photovoltaic power generation time series data and the user power consumption time series data to generate a photovoltaic power generation time series curve and a user power consumption time series curve;

[0092] Step S23: Overlap the photovoltaic power generation time sequence curve and the user power consumption time sequence curve to generate a grid power time sequence overlap curve; segment the grid power time sequence overlap curve to generate grid power non-overlapping line segments and grid power overlapping line segments;

[0093] Step S24: using a power grid dynamic balance calculation formula to perform power grid dynamic balance calculation on power grid non-coincidence line segments and power grid coincidence line segments to obtain a power grid dynamic balance coefficient;

[0094] Step S25: Perform grid load joint prediction on the photovoltaic power generation characteristic data and the user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint prediction data.

[0095] In the embodiment of the present invention, by extracting relevant features such as power generation, temperature, light intensity, etc. from the power fusion data set, photovoltaic power generation characteristic data is generated, including indicators such as power generation efficiency and fluctuation. User power consumption characteristics such as total power consumption, peak and valley power consumption, etc. are extracted from the power fusion data set to generate user power consumption characteristic data, including information such as power consumption mode and equipment distribution. Time series analysis is performed on photovoltaic power generation characteristic data, and photovoltaic power generation time series data is generated using a time series model (such as ARIMA or LSTM). Time series analysis is performed on user power consumption characteristic data to generate user power consumption time series data. Photovoltaic power generation time series data and user power consumption time series data are converted into time series curves to generate photovoltaic power generation time series curves and user power consumption time series curves for easy visualization analysis. The photovoltaic power generation time series curve is overlapped with the user power consumption time series curve to generate a power grid power time series overlap curve. The power grid power time series overlap curve is analyzed to identify the overlap and non-overlap parts of the power line segment, and the power grid power non-overlap line segment and the power grid power overlap line segment are generated to facilitate subsequent calculations. The dynamic balance calculation formula of the power grid (such as the formula based on the principle of power conservation) is used to calculate the non-coincidence and coincidence segments of the power grid to obtain the dynamic balance coefficient of the power grid, which reflects the matching degree between the power grid load and the power generation capacity. According to the dynamic balance coefficient of the power grid, the photovoltaic power generation characteristic data and the user power consumption characteristic data are jointly predicted. The machine learning model (such as regression analysis or neural network) is used to predict the power grid load, generate the joint prediction data of the power grid load, and provide a reference for the future power grid load.

[0096] Preferably, the power grid dynamic balance calculation formula in step S24 is as follows:

[0097]

[0098] Where B is the dynamic balance coefficient of the power grid, P net (t) is the net power at time t, P loss (t) is the power loss at time t, P gen (t) represents the total power generation at time t, t0 represents the lower limit of the measurement time, t1 represents the upper limit of the measurement time, τ represents the time constant, k i Expressed as the amplitude of the i-th frequency component, ωi Expressed as the angular frequency of the i-th frequency component, φ i is represented as the phase angle of the i-th frequency component, and n is represented as the number of frequency components.

[0099] The present invention analyzes and integrates a dynamic balance calculation formula for the power grid, which is designed to evaluate the dynamic balance state of the power grid in a specific time period, taking into account the net power, loss, power generation and the impact of various periodic fluctuations on the balance of the power grid. The performance of the power grid in the time dimension is comprehensively analyzed by integration and addition, where the integral term in the formula The difference P between the actual load of the grid and the power generation is taken into account net (t), which reflects the supply and demand status of the power grid at a certain moment. Considering the power P that cannot be used in the power grid due to the loss of transmission lines and equipment loss (t), to ensure more accurate calculation. Total power generation P gen (t), as a benchmark for the power supply capacity of the power grid. By using the relative proportion method, the efficiency of the power grid can be observed more clearly. It reflects the influence of time on the balance of the power grid and reflects the lag of the system response. The time constant τ determines the intensity of this influence. The frequency component term in the formula Considering the amplitude parameter k i , which represents the relative influence of different frequency components on the dynamic balance of the power grid. Different frequency components correspond to different load modes. The periodic function sin(ω i t+φ i ), which reflects the periodic characteristics of various load changes in the power grid. By combining components of different frequencies together, the fluctuations of the power grid can be fully reflected. When using the conventional power grid dynamic balance calculation formula in the field, the power grid dynamic balance coefficient can be obtained. By applying the power grid dynamic balance calculation formula provided by the present invention, the power grid dynamic balance coefficient can be calculated more accurately. The formula can accurately evaluate the dynamic balance state of the power grid in different time periods, thereby providing a scientific basis for power grid dispatching and management. By introducing the power loss factor, a comprehensive analysis of the power grid efficiency is ensured, making the power grid operation more efficient. The introduction of frequency components enhances the adaptability of the power grid to various load modes and can better cope with emergencies in actual operation. Considering the influence of the time constant, the formula is more in line with the dynamic characteristics of the actual power grid, helping to optimize the dispatching decision. The structure of the formula allows the parameters to be adjusted according to the actual situation, so as to adapt to the needs of different types of power grids. In summary, the relationship between the formula and its parameters not only enhances the accuracy of the dynamic balance analysis of the power grid, but also provides an effective tool for optimizing power grid management and dispatching.

[0100] Preferably, step S25 comprises the following steps:

[0101] Step S251: Perform power generation environment analysis on photovoltaic power generation characteristic data according to the dynamic balance coefficient of the power grid to generate photovoltaic power generation environment data; perform power consumption pattern analysis on user power consumption characteristic data according to the dynamic balance coefficient of the power grid to generate user power consumption pattern data;

[0102] Step S252: divide the photovoltaic power generation environment data and the user power consumption mode data into data sets to generate a model training set and a model test set; perform model training on the model training set using a long short-term memory network algorithm to generate a power grid load joint prediction pre-model;

[0103] Step S253: Use the model test set to iterate the model optimization of the power grid load joint prediction pre-model, so as to generate a power grid load joint prediction model; import the photovoltaic power generation characteristic data and the user's electricity consumption characteristic data into the power grid load joint prediction model to perform power grid joint load prediction and generate power grid load joint prediction data.

[0104] In an embodiment of the present invention, by analyzing the environmental factors affecting photovoltaic power generation (such as climate change, light intensity, temperature, etc.) according to the dynamic balance coefficient of the power grid, photovoltaic power generation environmental data is generated, including power generation efficiency and potential power generation under different environmental conditions. According to the dynamic balance coefficient of the power grid, the user's power consumption pattern (such as peak period, trough period and seasonal changes) is analyzed to generate user power consumption pattern data to describe power consumption habits and demand changes. Photovoltaic power generation environmental data and user power consumption pattern data are divided to form a model training set (for example, 70%) and a model test set (for example, 30%). The model training set is trained using a long short-term memory network (LSTM) algorithm to construct a power grid load joint prediction pre-model. Adjust network parameters (such as learning rate, batch size, etc.) to optimize model performance and ensure that it can effectively capture long-term dependencies in time series. Use the model test set to evaluate the power grid load joint prediction pre-model and calculate the prediction error (such as mean square error, mean absolute error, etc.). Perform model optimization iterations based on the evaluation results, and adjust the network structure or parameters to improve the prediction accuracy of the model. Combine the optimized grid load joint prediction model with photovoltaic power generation characteristic data and user power consumption characteristic data. Input data for grid joint load prediction, generate grid load joint prediction data, and provide decision-making basis for grid management.

[0105] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0106] Step S31: Performing grid load balance analysis on the power fusion data set through grid load joint prediction data to generate a grid load balance matrix; performing input scheduling optimization on the grid load balance matrix to generate grid load scheduling optimization data;

[0107] Step S32: matching the grid load dispatch optimization data with peak period characteristics to generate peak grid demand data; matching and adjusting the grid load dispatch optimization data according to the peak grid demand data to generate grid load matching adjustment data, wherein the matching and adjustment includes smoothing photovoltaic power generation fluctuations and adjusting user power consumption behavior;

[0108] Step S33: Performing energy storage device charging and discharging resource scheduling based on the grid load matching adjustment data to generate energy storage device scheduling data; performing grid-friendliness analysis on the energy storage device scheduling data to generate grid energy storage scheduling friendliness data;

[0109] Step S34: Perform security constraint attack simulation on energy storage device dispatch data using power grid energy storage dispatch friendliness data to generate a power grid dispatch attack simulation data set; perform dispatch security protection on energy storage device dispatch data using the power grid dispatch attack simulation data set to generate security constraint power grid dispatch data.

[0110] In the embodiment of the present invention, by using the power grid load joint prediction data, the power fusion data set is analyzed to generate a power grid load balance matrix to reflect the load situation and supply and demand relationship of each period of the power grid. Based on the power grid load balance matrix, input scheduling optimization is performed, and an optimization algorithm (such as linear programming or genetic algorithm) is applied to generate power grid load scheduling optimization data to achieve load balancing and effective resource utilization. The power grid load scheduling optimization data is matched with the peak period characteristics, and the historical data is analyzed to identify the peak power demand characteristics and generate peak power grid demand data. According to the peak power grid demand data, the power grid load scheduling optimization data is matched and adjusted. Including smoothing the fluctuation of photovoltaic power generation (adjusting the output through energy storage equipment) and adjusting the user's power consumption behavior (such as demand response measures), generating power grid load matching adjustment data. Based on the power grid load matching adjustment data, the energy storage equipment is scheduled for charging and discharging resources, and the energy storage equipment scheduling data is generated to ensure that the power grid can be supported during the peak period. The energy storage equipment scheduling data is analyzed for power grid friendliness, the support degree of the energy storage system to the power grid is evaluated, and the power grid energy storage scheduling friendliness data is generated to optimize the energy storage resource configuration. Through the grid energy storage dispatch friendliness data, security constraint attack simulation is performed on the energy storage equipment dispatch data, potential dispatch safety risks are identified, and a grid dispatch attack simulation data set is generated. Using the grid dispatch attack simulation data set, dispatch safety protection is performed on the energy storage equipment dispatch data, and security mechanisms (such as redundant dispatch or real-time monitoring) are used to generate security constraint grid dispatch data to improve the safety and stability of the grid system.

[0111] Preferably, step S34 includes the following steps:

[0112] Step S341: Perform intrusion detection scenario simulation on energy storage device dispatch data through power grid energy storage dispatch friendliness data to generate a power grid attack scenario set, wherein the power grid attack scenario set includes a denial of service attack scenario and a false data injection attack scenario;

[0113] Step S342: Based on the denial of service attack scenario, a dispatch load response attack simulation is performed on the energy storage device dispatch data to generate dispatch response attack simulation data; based on the false data injection attack scenario, a misleading voltage and frequency regulation attack simulation is performed on the energy storage device dispatch data to generate false regulation attack simulation data;

[0114] Step S343: integrating the dispatch response attack simulation data and the false control attack simulation data to generate a power grid dispatch attack simulation data set; performing a dispatch risk assessment on the energy storage device dispatch data according to the power grid dispatch attack simulation data set to generate dispatch risk assessment data;

[0115] Step S344: Use the power grid dispatch attack simulation data set to perform dispatch security protection on the energy storage device dispatch data to generate security-constrained power grid dispatch data.

[0116] In an embodiment of the present invention, the intrusion detection scenario is simulated on the energy storage device scheduling data through the power grid energy storage scheduling friendliness data to generate a power grid attack scenario set, which specifically includes: Denial of service attack scenario: Simulate the attacker to exhaust system resources through overload requests, thereby causing service unavailability. False data injection attack scenario: Simulate the attacker to inject false data into the system to interfere with the normal operation of the system. Based on the denial of service attack scenario, the energy storage device scheduling data is simulated for scheduling load response attack, and scheduling response attack simulation data is generated to evaluate the performance of the system under high load. Based on the false data injection attack scenario, the energy storage device scheduling data is simulated for misleading voltage and frequency regulation attack, and false regulation attack simulation data is generated to analyze the system's response to erroneous information. The scheduling response attack simulation data and the false regulation attack simulation data are integrated to generate a power grid scheduling attack simulation data set. According to the power grid scheduling attack simulation data set, the energy storage device scheduling data is evaluated for scheduling risk, and scheduling risk assessment data is generated to identify the vulnerability and potential risks of the system under different attack scenarios. Using the power grid scheduling attack simulation data set, the energy storage device scheduling data is dispatched for scheduling security protection. Apply security strategies (e.g., redundancy, real-time monitoring, and attack detection algorithms) to enhance the robustness of the system and generate security-constrained grid dispatch data to ensure stable operation in the face of potential attacks.

[0117] Preferably, step S344 includes the following steps:

[0118] Step S3441: abnormal monitoring of the energy storage device dispatching data to generate abnormal dispatching data; scenario identification and matching of the abnormal dispatching data and the power grid dispatching attack simulation data set, when a denial of service attack scenario is identified, scheduling packet loss rate analysis is performed on the energy storage device dispatching data to generate scheduling packet loss rate data; first security constraints are performed on the energy storage device dispatching data according to the scheduling packet loss rate data to generate first security constraint data;

[0119] Step S3442: when a false data injection attack scenario is identified, the energy storage device scheduling data is subjected to scheduling output over-analysis to generate scheduling over-output data; a second security constraint is applied to the energy storage device scheduling data according to the scheduling over-output data to generate second security constraint data;

[0120] Step S3443: construct a safety constraint engine according to the first safety constraint data and the second safety constraint data to generate a safety constraint engine; use the safety constraint engine to perform backup energy storage device switching on the energy storage device scheduling data to generate backup energy storage device scheduling switching data;

[0121] Step S3444: synchronously adjust the energy storage device scheduling data with the backup energy storage device scheduling switching data to generate power adjustment data; perform scheduling safety verification on the power adjustment data to generate safety-constrained power grid scheduling data.

[0122] In an embodiment of the present invention, by performing real-time abnormal monitoring on the energy storage device scheduling data, potential scheduling abnormalities are identified and scheduling abnormality data is generated. The scheduling abnormality data is matched with the power grid scheduling attack simulation data set for scenario identification. When a denial of service attack scenario is identified: a scheduling packet loss rate analysis is performed, the packet loss caused by the network load is evaluated, and the scheduling packet loss rate data is generated. A first security constraint is performed on the energy storage device scheduling data according to the scheduling packet loss rate data, and a first security constraint data is generated to ensure the reliability of the scheduling. When a false data injection attack scenario is identified: a scheduling output over-analysis is performed, the output of the energy storage device under the input of erroneous data is evaluated, and scheduling over-output data is generated. A second security constraint is performed on the energy storage device scheduling data according to the scheduling over-output data, and a second security constraint data is generated to prevent abnormal output caused by false data. A security constraint engine is constructed based on the first security constraint data and the second security constraint data. The energy storage device scheduling data is switched to a backup energy storage device using the security constraint engine to ensure the continuous operation of the system under abnormal conditions, and generate scheduling switching data for backup energy storage devices. Through the dispatch switching data of the backup energy storage device, the dispatch power of the energy storage device dispatch data is adjusted synchronously to generate power adjustment data to ensure the stability of the power grid load. The power adjustment data is dispatched for safety verification to confirm whether the adjusted dispatch plan meets the safety constraints and ensure that the system can operate safely under the new dispatch conditions. After verification, the safety constraint grid dispatch data is generated to provide a safe and reliable dispatch plan for grid management.

[0123] Preferably, step S4 comprises the following steps:

[0124] Step S41: Performing real-time Internet of Things configuration monitoring on the safety-constrained power grid dispatching data to generate real-time dispatching monitoring data; performing real-time power grid dispatching status monitoring on the real-time dispatching monitoring data to generate a real-time power grid operation feedback data set;

[0125] Step S42: collecting user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjusting the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set to generate power grid load dynamic regulation data;

[0126] Step S43: Visualize the power grid load dynamic regulation data to generate a power grid load regulation report to perform power grid load regulation monitoring operations.

[0127] In an embodiment of the present invention, real-time data of each link of the power grid is collected by performing real-time configuration monitoring of the Internet of Things on the dispatching data of the safety constraint power grid, and real-time dispatching monitoring data is generated. Based on the real-time dispatching monitoring data, the operation status of the power grid is monitored in real time, and a real-time operation feedback data set of the power grid is generated to evaluate the current operation status and potential problems of the power grid. User demand feedback is collected on the real-time operation feedback data set of the power grid to obtain user demand information on power supply and load changes, and obtain a user demand feedback data set. Using the user demand feedback data set, the real-time operation feedback data set of the power grid is dynamically adjusted to ensure that the power grid can be flexibly adjusted according to user needs, and the dynamic control data of the power grid load is generated. The dynamic control data of the power grid load is visualized, and the control results and related data are presented in the form of charts, dashboards, etc., which are easy to analyze and understand. Based on the visualized data, a power grid load control report is generated, and the control measures, effects and user feedback are recorded in detail to perform the power grid load control monitoring operation.

[0128] In this specification, a power grid load control system based on security constraints is provided, which is used to execute the above-mentioned power grid load prediction method based on security constraints. The power grid load control system based on security constraints includes:

[0129] A multi-source data fusion module is used to obtain advanced sensor deployment data; collect multi-source data streams for advanced sensor deployment data to obtain a multi-source power edge data set; perform power data association on the multi-source power edge data set to generate multi-source power association data; perform data fusion on the multi-source power edge data set through the multi-source power association data to generate a power fusion data set;

[0130] The joint load forecasting module is used to extract grid characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data and user power consumption characteristic data; overlap the time series curves of photovoltaic power generation characteristic data and user power consumption characteristic data to generate a grid power time series overlap curve; perform grid dynamic balance calculation on the grid power time series overlap curve to obtain a grid dynamic balance coefficient; perform grid load joint forecasting on photovoltaic power generation characteristic data and user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint forecasting data;

[0131] The security constraint adjustment module is used to dispatch the charging and discharging resources of energy storage equipment to the power fusion data set through the joint prediction data of the power grid load, and generate the energy storage equipment dispatching data; perform security constraint attack simulation on the energy storage equipment dispatching data, and generate the power grid dispatching attack simulation data set; use the power grid dispatching attack simulation data set to perform dispatching security protection on the energy storage equipment dispatching data, and generate the security constraint power grid dispatching data;

[0132] The real-time feedback monitoring module is used to monitor the real-time dispatching status of the power grid with safety constraints on the dispatching data of the power grid, and generate a real-time operation feedback data set of the power grid; collect user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjust the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set, and generate a power grid load regulation report to perform power grid load regulation monitoring operations.

[0133] The beneficial effect of the present invention is that more comprehensive and accurate power data is obtained through advanced sensor deployment and multi-source data stream collection. Through data association and fusion, a power fusion data set is generated, which improves the consistency and relevance of the data and ensures that subsequent predictions and scheduling are based on more accurate inputs. Feature extraction of the power fusion data set can obtain feature data of photovoltaic power generation and user power consumption respectively, and effectively reflect the dynamic changes of the load of the power grid through timing curve overlap and dynamic balance calculation, further improving the accuracy of the joint prediction of the power grid load. The charging and discharging scheduling of energy storage equipment is optimized by using the joint prediction data of the power grid load to achieve more efficient resource management. At the same time, through attack simulation and scheduling security protection, the security of power grid scheduling and the ability to resist network attacks are enhanced to ensure the stability of power grid operation. Through real-time scheduling status monitoring and user demand feedback mechanism, the power grid load regulation is more flexible and can be dynamically adjusted, which improves the adaptability and responsiveness of power grid operation, ensures that the power supply meets real-time needs, and optimizes the power grid regulation strategy through feedback mechanism. Therefore, the present invention improves the timeliness, accuracy and security of power grid prediction and scheduling through multi-source data fusion, dynamic load prediction, security attack protection and user feedback mechanism.

[0134] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0135] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A power grid load forecasting method based on security constraints, characterized in that: The following steps are involved: Step S1: Obtaining high-level sensor deployment data; Collect multi-source data streams of advanced sensor deployment data to obtain multi-source power edge data sets; Perform power data association on multi-source power edge data sets to generate multi-source power association data; Perform data fusion on multi-source power edge data sets through multi-source power correlation data to generate a power fusion data set; Step S2: Extract grid characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data and user power consumption characteristic data; overlap the time series curves of the photovoltaic power generation characteristic data and the user power consumption characteristic data to generate a grid power time series overlap curve; perform grid dynamic balance calculation on the grid power time series overlap curve to obtain a grid dynamic balance coefficient; perform grid load joint prediction on the photovoltaic power generation characteristic data and the user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint prediction data; Step S3: dispatching energy storage device charging and discharging resources for the power fusion data set through the joint prediction data of the power grid load to generate energy storage device dispatching data; Perform security constraint attack simulation on energy storage equipment dispatch data to generate a power grid dispatch attack simulation data set; use the power grid dispatch attack simulation data set to perform dispatch security protection on energy storage equipment dispatch data to generate security constraint power grid dispatch data; Step S4: monitoring the real-time dispatching status of the power grid on the safety-constrained power grid dispatching data to generate a real-time operation feedback data set of the power grid; collecting user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; The user demand feedback data set is used to adjust the dynamic load control strategy of the power grid real-time operation feedback data set, and a power grid load regulation report is generated to perform power grid load regulation monitoring operations.

2. The power grid load forecasting method based on security constraints according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Acquire advanced sensor deployment data of the photovoltaic power generation system, user power equipment and power grid system, wherein the advanced sensors include photovoltaic module monitoring sensors, user power usage sensors and power grid status sensors; Step S12: performing multi-source data stream acquisition on the advanced sensor deployment data to obtain an original multi-source power data set; performing data preprocessing on the original multi-source power data set to generate a standard multi-source power data set, wherein the data preprocessing includes data cleaning, data denoising, data normalization and data standardization; Step S13: performing edge computing on the standard multi-source power data set to generate a multi-source power edge data set; Step S14: performing power data association on the multi-source power edge data sets to generate multi-source power association data; performing data fusion on the multi-source power edge data sets through the multi-source power association data to generate a power fusion data set.

3. The power grid load forecasting method based on safety constraints according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: extract photovoltaic power generation characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data; extract user power consumption characteristics from the power fusion data set to obtain user power consumption characteristic data; Step S22: performing a power generation time series analysis on the photovoltaic power generation characteristic data to generate photovoltaic power generation time series data; performing a user power consumption time series analysis on the user power consumption characteristic data to generate user power consumption time series data; performing curve conversion on the photovoltaic power generation time series data and the user power consumption time series data to generate a photovoltaic power generation time series curve and a user power consumption time series curve; Step S23: Overlap the photovoltaic power generation time sequence curve and the user power consumption time sequence curve to generate a grid power time sequence overlap curve; segment the grid power time sequence overlap curve to generate grid power non-overlapping line segments and grid power overlapping line segments; Step S24: using a power grid dynamic balance calculation formula to perform power grid dynamic balance calculation on power grid non-coincidence line segments and power grid coincidence line segments, to obtain a power grid dynamic balance coefficient; Step S25: Perform grid load joint prediction on the photovoltaic power generation characteristic data and the user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint prediction data.

4. The method for predicting power grid load based on safety constraints according to claim 3, characterized in that: The calculation formula for the dynamic balance of the power grid in step S24 is as follows: In the formula, B represents the dynamic balance coefficient of the power grid, P net (t) is the net power at time t, P loss (t) is the power loss at time t, P gen (t) represents the total power generation at time t, t0 represents the lower limit of the measurement time, t1 represents the upper limit of the measurement time, τ represents the time constant, k i Expressed as the amplitude of the i-th frequency component, ω i Expressed as the angular frequency of the i-th frequency component, φ i is represented as the phase angle of the i-th frequency component, and n is represented as the number of frequency components.

5. The method for predicting power grid load based on safety constraints according to claim 3, characterized in that: Step S25 includes the following steps: Step S251: Perform power generation environment analysis on photovoltaic power generation characteristic data according to the dynamic balance coefficient of the power grid to generate photovoltaic power generation environment data; perform power consumption pattern analysis on user power consumption characteristic data according to the dynamic balance coefficient of the power grid to generate user power consumption pattern data; Step S252: divide the photovoltaic power generation environment data and the user power consumption mode data into data sets to generate a model training set and a model test set; perform model training on the model training set using a long short-term memory network algorithm to generate a power grid load joint prediction pre-model; Step S253: Use the model test set to iterate the model optimization of the power grid load joint prediction pre-model, so as to generate a power grid load joint prediction model; import the photovoltaic power generation characteristic data and the user's electricity consumption characteristic data into the power grid load joint prediction model to perform power grid joint load prediction and generate power grid load joint prediction data.

6. The method for predicting power grid load based on safety constraints according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: Performing grid load balance analysis on the power fusion data set through grid load joint prediction data to generate a grid load balance matrix; performing input scheduling optimization on the grid load balance matrix to generate grid load scheduling optimization data; Step S32: matching the grid load dispatch optimization data with peak period characteristics to generate peak grid demand data; matching and adjusting the grid load dispatch optimization data according to the peak grid demand data to generate grid load matching adjustment data, wherein the matching and adjustment includes smoothing photovoltaic power generation fluctuations and adjusting user power consumption behavior; Step S33: Performing energy storage device charging and discharging resource scheduling based on the grid load matching adjustment data to generate energy storage device scheduling data; performing grid-friendliness analysis on the energy storage device scheduling data to generate grid energy storage scheduling friendliness data; Step S34: Perform security constraint attack simulation on energy storage device dispatch data using power grid energy storage dispatch friendliness data to generate a power grid dispatch attack simulation data set; perform dispatch security protection on energy storage device dispatch data using the power grid dispatch attack simulation data set to generate security constraint power grid dispatch data.

7. The method for predicting power grid load based on safety constraints according to claim 6, characterized in that: Step S34 includes the following steps: Step S341: Perform intrusion detection scenario simulation on energy storage device dispatch data through power grid energy storage dispatch friendliness data to generate a power grid attack scenario set, wherein the power grid attack scenario set includes a denial of service attack scenario and a false data injection attack scenario; Step S342: Based on the denial of service attack scenario, a dispatch load response attack simulation is performed on the energy storage device dispatch data to generate dispatch response attack simulation data; based on the false data injection attack scenario, a misleading voltage and frequency regulation attack simulation is performed on the energy storage device dispatch data to generate false regulation attack simulation data; Step S343: integrating the dispatch response attack simulation data and the false control attack simulation data to generate a power grid dispatch attack simulation data set; performing a dispatch risk assessment on the energy storage device dispatch data according to the power grid dispatch attack simulation data set to generate dispatch risk assessment data; Step S344: Use the power grid dispatch attack simulation data set to perform dispatch security protection on the energy storage device dispatch data to generate security-constrained power grid dispatch data.

8. The method for predicting power grid load based on safety constraints according to claim 7, characterized in that: Step S344 includes the following steps: Step S3441: abnormal monitoring of the energy storage device dispatching data to generate abnormal dispatching data; scenario identification and matching of the abnormal dispatching data and the power grid dispatching attack simulation data set, when a denial of service attack scenario is identified, scheduling packet loss rate analysis is performed on the energy storage device dispatching data to generate scheduling packet loss rate data; first security constraints are performed on the energy storage device dispatching data according to the scheduling packet loss rate data to generate first security constraint data; Step S3442: when a false data injection attack scenario is identified, the energy storage device scheduling data is subjected to scheduling output over-analysis to generate scheduling over-output data; a second security constraint is applied to the energy storage device scheduling data according to the scheduling over-output data to generate second security constraint data; Step S3443: construct a safety constraint engine according to the first safety constraint data and the second safety constraint data to generate a safety constraint engine; use the safety constraint engine to perform backup energy storage device switching on the energy storage device scheduling data to generate backup energy storage device scheduling switching data; Step S3444: synchronously adjust the energy storage device scheduling data with the backup energy storage device scheduling switching data to generate power adjustment data; perform scheduling safety verification on the power adjustment data to generate safety-constrained power grid scheduling data.

9. The method for predicting power grid load based on safety constraints according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Performing real-time Internet of Things configuration monitoring on the safety-constrained power grid dispatching data to generate real-time dispatching monitoring data; performing real-time power grid dispatching status monitoring on the real-time dispatching monitoring data to generate a real-time power grid operation feedback data set; Step S42: collecting user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjusting the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set to generate power grid load dynamic regulation data; Step S43: Visualize the power grid load dynamic regulation data to generate a power grid load regulation report to perform power grid load regulation monitoring operations.

10. A power grid load control system based on safety constraints, characterized in that: Used to execute the power grid load forecasting method based on safety constraints as claimed in claim 1, the power grid load regulation system based on safety constraints comprises: A multi-source data fusion module is used to obtain advanced sensor deployment data; collect multi-source data streams for advanced sensor deployment data to obtain a multi-source power edge data set; perform power data association on the multi-source power edge data set to generate multi-source power association data; perform data fusion on the multi-source power edge data set through the multi-source power association data to generate a power fusion data set; The joint load forecasting module is used to extract grid characteristics from the power fusion data set to obtain photovoltaic power generation characteristic data and user power consumption characteristic data; overlap the time series curves of photovoltaic power generation characteristic data and user power consumption characteristic data to generate a grid power time series overlap curve; perform grid dynamic balance calculation on the grid power time series overlap curve to obtain a grid dynamic balance coefficient; perform grid load joint forecasting on photovoltaic power generation characteristic data and user power consumption characteristic data according to the grid dynamic balance coefficient to generate grid load joint forecasting data; The security constraint adjustment module is used to dispatch the charging and discharging resources of energy storage equipment to the power fusion data set through the joint prediction data of the power grid load, and generate the energy storage equipment dispatching data; perform security constraint attack simulation on the energy storage equipment dispatching data, and generate the power grid dispatching attack simulation data set; use the power grid dispatching attack simulation data set to perform dispatching security protection on the energy storage equipment dispatching data, and generate the security constraint power grid dispatching data; The real-time feedback monitoring module is used to monitor the real-time dispatching status of the power grid with safety constraints on the dispatching data of the power grid, and generate a real-time operation feedback data set of the power grid; collect user demand feedback on the real-time operation feedback data set of the power grid to obtain a user demand feedback data set; adjust the dynamic load control strategy of the real-time operation feedback data set of the power grid through the user demand feedback data set, and generate a power grid load regulation report to perform power grid load regulation monitoring operations.

Images