A Power System State Analysis Method Based on Large-Scale Distributed Energy Access

Abstract

This invention discloses a power system state analysis method based on large-scale distributed energy access, relating to the field of distribution network technology. The method includes the following steps: collecting basic and historical equipment operation data of the distribution network to construct a virtual power plant model; collecting real-time equipment operation data to correct the virtual power plant model; constructing a photovoltaic output prediction model using real-time photovoltaic power generation data, outputting the prediction trajectory and calculating the photovoltaic output volatility; constructing a fuel output prediction model based on real-time load demand and fuel reserve data, obtaining the fuel output trajectory and calculating its stability, and combining the photovoltaic output volatility to derive the generation-side reliability index; calculating energy storage stability and controllable load stability based on energy storage charging and discharging power and controllable load adjustment, respectively, to obtain the grid-side stability index; and integrating the generation-side reliability index and the grid-side stability index to calculate the distribution network state assessment index. This invention achieves the identification of the distribution network state through a multi-dimensional assessment mechanism.

Description

Technical Field

[0001] This invention relates to the field of power distribution network engineering, specifically to a power system state analysis method based on large-scale distributed energy access. Background Technology

[0002] With the advancement of "dual carbon" goals and the transformation of the global energy structure, distributed energy, with its decentralized and clean characteristics, is rapidly being integrated into the power distribution network. By analyzing the status of the power distribution network in real time, problems such as equipment failures and line overloads can be detected in a timely manner, and maintenance and switching measures can be taken in advance to avoid power outages and ensure a continuous and stable power supply to users.

[0003] However, traditional methods mainly assess the state of the distribution network based on load factor zoning, but load factor zoning relies on human experience and cannot effectively quantify the coupling relationship between energy output fluctuations and grid stability, making it difficult to accurately identify and assess the state of the distribution network. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a power system state analysis method based on large-scale distributed energy access, which solves the problem that traditional methods cannot effectively quantify the coupling relationship between energy output fluctuations and grid stability, and that it is difficult to accurately identify and assess the state of the distribution network.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a power system state analysis method based on large-scale distributed energy access, comprising the following steps:

[0006] Step S1: Collect basic data of the power distribution network and historical equipment operation data, and construct a virtual power plant model based on the basic data of the power distribution network and historical equipment operation data;

[0007] Step S2: Collect real-time equipment operation data, and correct the virtual power plant model based on the real-time equipment operation data to obtain a corrected virtual power plant model;

[0008] Step S3: Collect real-time photovoltaic power generation data, construct a photovoltaic output prediction model based on the modified virtual power plant model, input the real-time photovoltaic power generation data into the photovoltaic output prediction model, output the photovoltaic power generation prediction trajectory, and calculate the photovoltaic output volatility based on the photovoltaic power generation prediction trajectory.

[0009] Step S4: Collect real-time load demand and fuel reserve data, construct a fuel output prediction model based on the modified virtual power plant model, input the real-time load demand and fuel reserve data into the fuel output prediction model, output the fuel output trajectory, calculate the fuel output stability based on the fuel output trajectory, and calculate the power generation side reliability index based on the photovoltaic output volatility and fuel output stability.

[0010] Step S5: Calculate the energy storage stability based on the energy storage charging and discharging power in the modified virtual power plant model, calculate the controllable load stability based on the controllable load adjustment amount in the modified virtual power plant model, and calculate the grid-side stability index by combining the energy storage stability and the controllable load stability.

[0011] Step S6: Calculate the distribution network status assessment index by combining the generation side reliability index and the grid side stability index.

[0012] Preferably, the construction of the virtual power plant model includes:

[0013] Building a virtual power plant model:

[0014] Using a 5km radius as a standard, and leveraging GIS spatial analysis capabilities, distributed energy resources within the same geographical area are grouped into a single aggregation unit. Dispatch homogeneity aggregation: Distributed energy resources belonging to the same operator or supporting unified command issuance are aggregated into a single unit regardless of their geographical location. Response consistency aggregation: Aggregation is based on the response characteristics of the equipment. For energy storage units, energy storage devices with charge / discharge rate deviations within 10% are aggregated together. For power generation equipment, photovoltaic power plants and micro gas turbines with active power regulation times of similar magnitude are aggregated to ensure that equipment within the aggregation unit can operate collaboratively when participating in grid regulation, thereby improving regulation efficiency.

[0015] Equipment list compilation: Real-time operating status: current output power of photovoltaic, SOC and charge / discharge status of energy storage, current fuel consumption rate of micro gas turbine;

[0016] Communication status: timestamp of the most recent data update, average communication latency;

[0017] Initialize model parameters:

[0018] For each aggregation unit, pre-calculate: Photovoltaics: installed capacity share, historical output volatility baseline; Micro gas turbines: rated output share, fuel reserve safety threshold; Energy storage: available adjustable capacity.

[0019] Preferably, the step of collecting real-time equipment operation data and correcting the virtual power plant model based on the real-time equipment operation data to obtain a corrected virtual power plant model includes:

[0020] Data access: Collect real-time measurement data and device operation logs from edge devices;

[0021] Dynamic deviation detection: A 1-minute sliding window is used to calculate the deviation between the model's predicted value and the measured value. If the photovoltaic output prediction error is greater than 5%, a correction is triggered. The source of the deviation is located through residual analysis.

[0022] Adaptive correction strategy execution:

[0023] Control strategy refitting: For equipment with excessive deviation, the power response curve is refitted based on the latest 1-hour data;

[0024] Influence path weight adjustment: Dynamically update the influence coefficients of virtual power plant behavior on state variables using recursive least squares method;

[0025] Temporary decoupling: Devices with communication interruptions exceeding 5 minutes are marked as "offline" and temporarily removed from the model;

[0026] Model validation after correction: The prediction error after correction should be less than 3%, and the output should be a corrected virtual power plant model with the latest parameters;

[0027] The latest parameters include energy storage charging and discharging power: the charging and discharging power of the energy storage system at each moment within N time steps; and controllable load adjustment amount: the adjustment amount of the controllable load at each moment.

[0028] Preferably, the step of calculating the photovoltaic power output fluctuation rate based on the predicted photovoltaic power output trajectory includes:

[0029] Output predictive force trajectory generation: Input the current real-time raw data at k time steps The model directly outputs the predicted sequence for the next n steps at the original scale: ;

[0030] The predicted power trajectory is obtained directly. ;

[0031] Predicted trajectory at the original scale Calculate the relative rate of change between adjacent time steps:

[0032]

[0033] in, Represents the relative rate of change between the i-th time step and the (i-1)-th time step. It is the predicted value at the (i-1)th time step, used as the benchmark value for calculating the relative rate of change. It is the predicted value at the i-th time step, where i is the time step index, representing a discrete time point;

[0034] Calculate the average of the relative rates of change based on the relative rates of change at adjacent time steps:

[0035]

[0036] in, The average value represents the relative rate of change, where n is the number of time steps in the prediction, and represents the number of predicted values ​​contained in the predicted trajectory vector. The relative rate of change between adjacent time steps, where i is the time step index, representing a discrete time point;

[0037] Volatility is defined as the standard deviation of a relative rate of change.

[0038] Calculation of photovoltaic power output volatility:

[0039]

[0040] in, This represents the photovoltaic power output volatility, where n is the prediction time step. The relative rate of change between adjacent time steps The average value of the relative rate of change is represented by the formula, which is directly based on the relative fluctuations of the original power sequence and reflects the force stability. i is the time step index, representing a discrete time point.

[0041] Preferably, the calculation of fuel output stability based on the fuel output trajectory includes:

[0042] Fuel output trajectory generation:

[0043] Calculate using the state equation and output equation That is, the future Predict fuel output trajectory step by step.

[0044] This model generates a smooth fuel output trajectory by using the rolling optimization mechanism of MPC, under the premise of satisfying fuel reserve and output constraints, and quantifies its stability by standard deviation, avoiding normalization preprocessing and directly modeling and calculating based on the original physical quantities.

[0045] Fuel output stability calculation:

[0046] Based on predicted trajectory The stability index is defined as the standard deviation of output fluctuation (reflecting the smoothness of the trajectory):

[0047] =

[0048] in, The smaller the value, the more stable the output and the smaller the fluctuation. Is Always Predicted fuel output at any given moment; Is Always Predicted fuel output at any given time. It is the total number of steps in the prediction time domain, i.e., predicting the future. Fuel output at any given moment.

[0049] Preferably, the calculation of the power generation side reliability index based on photovoltaic power output volatility and fuel output stability includes:

[0050] Based on the photovoltaic power output volatility and fuel output stability, the power generation side reliability index is calculated:

[0051]

[0052] in, It is the reliability index of the power generation side. and These are the weights for photovoltaic power output volatility and fuel power output stability, respectively. =1, It is the fluctuation rate of photovoltaic power output. It refers to fuel output stability.

[0053] Preferably, the step of calculating the energy storage stability based on the energy storage charging and discharging power in the modified virtual power plant model includes:

[0054] Based on a modified virtual power plant model, the impact of energy storage and controllable load regulation capabilities on grid stability is quantitatively evaluated, and regulation capability indicators are defined.

[0055] Energy storage charging and discharging power: denoted as { } represents the charging and discharging power of the energy storage system at each moment within N time steps;

[0056] Calculating the mean absolute deviation measures stability

[0057] Energy storage stability:

[0058] First, calculate the average energy storage:

[0059]

[0060] in, This represents the average value of the energy storage charging and discharging power, where N is the total number of time steps. Let t be the energy storage charging and discharging power at time t. This formula obtains the average value of the energy storage charging and discharging power by summing the energy storage power over N time steps and taking the average value, which reflects the overall level of energy storage power. t is the identifier of the discrete time step.

[0061] Calculate the mean absolute deviation of energy storage as a measure of energy storage stability:

[0062]

[0063] in, The stability of energy storage charging and discharging power is measured by the degree of deviation of the energy storage power from its mean. The smaller the deviation, the more stable the energy storage charging and discharging. This represents the average value of the energy storage charging and discharging power. Let N be the energy storage charging and discharging power at time t, N be the total number of time steps, and t be the identifier of the discrete time step.

[0064] Preferably, the step of calculating the controllable load stability based on the controllable load adjustment amount in the modified virtual power plant model includes:

[0065] Controllable load adjustment amount: denoted as { } represents the adjustment amount of the controllable load at each moment;

[0066] Find the mean value of the controllable load adjustment:

[0067]

[0068] in, This represents the average value of the controllable load adjustment. The total number of time steps. For a moment The controllable load adjustment amount, where t is the identifier of the discrete time step;

[0069] This formula is derived from... Controllable load adjustment at each time step The summation and averaging yields the mean of the controllable load adjustment, reflecting the overall level of controllable load adjustment.

[0070] Calculate the mean absolute deviation of the controllable load adjustment as a measure of controllable load stability:

[0071]

[0072] in, The stability of controllable load regulation is measured by the degree of deviation of the controllable load regulation from its mean. The smaller the deviation, the more stable the controllable load regulation. This represents the average value of the controllable load adjustment. The controllable load adjustment at each moment is first determined by the average value. Characterize the overall level of controllable load adjustment, and then utilize To quantify its stability, the smaller the value, the more stable the controllable load adjustment.

[0073] Preferably, the calculation of the grid-side stability index by combining the energy storage stability and the controllable load stability includes:

[0074] A nonlinear function is introduced to reflect the characteristic that "the larger the adjustment amount, the more significant the impact on stability":

[0075] Energy storage weight: ,when Increase Approaching 1, highlighting the dominant influence of energy storage on stability;

[0076] Controllable load weighting: ;

[0077] Considering energy storage and controllable load regulation Introducing secondary interaction items: ;

[0078] Grid-side stability index:

[0079]

[0080] in, and These are the average energy storage fluctuation intensity and the average controllable load fluctuation intensity, respectively. These are the weights of the average energy storage fluctuation intensity and the average controllable load fluctuation intensity, respectively. Weights for interaction items; It is an introduced secondary interaction term; when the fluctuation of a single adjustment term is strong, it is through... Dynamic weighting amplifies its influence; when both fluctuate simultaneously... Further increase the index value to reflect the risk superposition of "1+1>2".

[0081] Preferably, the calculation of the distribution network condition assessment index by combining the generation-side reliability index and the grid-side stability index includes:

[0082] The reliability index of the power generation side was determined using an expert evaluation method. Weight of grid-side stability index ;

[0083] Combining the generation-side reliability index and the grid-side stability index, the distribution network condition assessment index is calculated:

[0084] The distribution network condition assessment index is calculated using a weighted summation method, and the formula is as follows:

[0085]

[0086] in, It is a power distribution network condition assessment index. and These are the generation-side reliability index and the grid-side stability index. These are the weights of the generation-side reliability index and the grid-side stability index, respectively. This formula integrates the evaluation results from the generation side and the grid side, quantitatively reflecting the overall state of the distribution network and providing a basis for subsequent comprehensive analysis.

[0087] Beneficial effects

[0088] This invention provides a power system state analysis method based on large-scale distributed energy resource integration. It has the following beneficial effects:

[0089] (1) This power system state analysis method based on large-scale distributed energy access can accurately assess the impact of the generation side on the power supply reliability of the distribution network by calculating the reliability index of the generation side and comprehensively considering factors such as photovoltaic output volatility and fuel output stability. This helps to predict possible power outages or instability in advance, take corresponding measures to optimize and improve, thereby improving the ability of the distribution network to continuously and stably supply power to users, reducing the time and frequency of power outages, and improving the power supply reliability index.

[0090] (2) The power system status analysis method based on large-scale distributed energy access takes into account multiple factors such as generation side reliability and grid side stability through the comprehensive status assessment index. It can more comprehensively and accurately reflect the actual operating status of the distribution network, avoid the one-sidedness and inaccuracy that may be caused by single indicator assessment, and provide a more reliable basis for the operation and management of the distribution network.

[0091] (3) This power system state analysis method based on large-scale distributed energy access fully considers external and internal factors such as weather disturbances and the health status of distribution equipment when calculating the distribution network impact coefficient, making the risk assessment of the distribution network more accurate and comprehensive. Weather disturbance data can reflect the impact of the natural environment on the operation of distribution network equipment, such as equipment failures that may be caused by lightning strikes, strong winds, and rainstorms; distribution equipment health index data can detect equipment aging, wear and tear, and potential fault hazards in advance. By incorporating these factors into the assessment system, various risks faced by the distribution network can be identified in a timely manner, and targeted prevention and response measures can be taken according to the degree of risk to reduce the probability and impact of risks. Attached Figure Description

[0092] Figure 1 This is a flowchart of a power system state analysis method based on large-scale distributed energy access proposed in this invention.

[0093] Figure 2 This is a flowchart illustrating the modified virtual power plant model obtained from a power system state analysis method based on large-scale distributed energy access proposed in this invention.

[0094] Figure 3The flowchart of the power system state analysis method based on large-scale distributed energy access proposed in this invention is shown below to obtain the distribution network state assessment index. Detailed Implementation

[0095] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0096] Please see Figure 1 This invention provides a technical solution: a power system state analysis method based on large-scale distributed energy resource integration. For details, please refer to... Figure 1 The method includes the following steps:

[0097] Step S1: Collect basic data of the power distribution network and historical equipment operation data, and construct a virtual power distribution plant model based on the basic data of the power distribution network and historical equipment operation data.

[0098] Distribution network basic data collection:

[0099] Network topology data: Detailed network topology information is obtained using a Geographic Information System (GIS) and a distribution network automation system. The number and geographic coordinates of each node are recorded, accurate to the meter. For lines, the starting and ending node numbers, length, and conductor type are collected to determine line impedance, while transformer parameters such as turns ratio, rated capacity, and short-circuit impedance are also obtained. Switch status data is collected in real-time through communication with intelligent switching devices to ensure data timeliness and accuracy.

[0100] Distributed energy access information:

[0101] Photovoltaics: Supplement the inverter conversion efficiency curve (as it varies with temperature and light intensity) and maximum power point tracking (MPPT) parameters for initial parameter setting of the photovoltaic output prediction model;

[0102] Micro gas turbines: The new load-efficiency characteristic curves (fuel consumption rate under different outputs, unit: tons / kW·h) and minimum technical output (percentage of rated power) directly support the constraints of the fuel output prediction model;

[0103] By reviewing engineering construction documents, equipment ledgers, and conducting on-site surveys, the connection locations and voltage levels for distributed energy resources are determined, down to the specific pole or distribution room. For photovoltaic power generation devices, their installed capacity, photovoltaic module type, and inverter model are recorded; for micro gas turbines, their rated power and fuel type are recorded; for energy storage units, their capacity, charge / discharge efficiency, and battery type are recorded; and for controllable loads, detailed information such as their industry, load characteristics, and rated power is recorded.

[0104] Historical equipment operation data collection:

[0105] Data from power generation equipment: Operating data is collected through the monitoring module built into the photovoltaic inverter and the control system of the micro gas turbine. For photovoltaic modules, data such as output power, open-circuit voltage, short-circuit current, and module temperature are acquired, with a sampling frequency of once per minute. For the micro gas turbine, key operating parameters such as active power, reactive power, fuel consumption rate, turbine speed, and exhaust temperature are collected, with a sampling frequency of once every 5 minutes to ensure timely capture of changes in equipment operating status.

[0106] Energy storage device data: Utilizing the battery management system (BMS) of the energy storage system, the state of charge (SOC) of the energy storage unit is collected in real time with an accuracy of ±2%. Simultaneously, data such as charging and discharging power, battery voltage, current, and temperature are collected at a sampling frequency of once every 2 minutes to accurately assess the performance and remaining capacity of the energy storage system.

[0107] Load-related equipment data: For controllable loads, real-time power, power factor, voltage and current harmonics, etc. are collected through smart meters or load management terminals, with a sampling frequency of once every 15 minutes.

[0108] Data preprocessing:

[0109] Outlier cleaning: The 3σ principle is used to detect and clean outliers in the collected data. For a given data sequence, its mean and standard deviation are calculated. If a data point deviates from the mean by more than three times the standard deviation, it is considered an outlier.

[0110] Unified data format: All collected data is converted according to the standard format of the Common Information Model (CIM) to facilitate subsequent data storage, transmission and processing.

[0111] Building a virtual power plant model:

[0112] Using a 5km radius as a standard, and leveraging GIS spatial analysis, distributed energy resources within the same geographical area are grouped into a single aggregation unit. Dispatch homogeneity aggregation: Distributed energy resources belonging to the same operator or supporting unified command issuance are aggregated into a single unit regardless of their geographical location. Response consistency aggregation: Aggregation is based on the response characteristics of the equipment; for energy storage units, those with charge / discharge rate deviations within 10% are aggregated together. For power generation equipment, photovoltaic power plants and micro gas turbines with active power regulation times of similar magnitude are aggregated to ensure coordinated operation of equipment within the aggregation unit when participating in grid regulation, thereby improving regulation efficiency.

[0113] Equipment list compilation: Real-time operating status: Current output power of photovoltaic, SOC and charge / discharge status of energy storage (charging / discharging / standby), current fuel consumption rate of micro gas turbine;

[0114] Communication status: timestamp of the most recent data update, average communication delay (used as a basis for determining temporary decoupling).

[0115] Initialize model parameters (pre-configure the basic variables required for subsequent steps):

[0116] For each aggregation unit, pre-calculate: Photovoltaics: installed capacity ratio, historical output volatility baseline (used as a benchmark for photovoltaic output volatility comparison); Micro gas turbine: rated output ratio, fuel reserve safety threshold (S_min = fuel required for 3 hours minimum technical output); Energy storage: available adjustable capacity = (SOC upper limit - SOC current value) × rated capacity × charge and discharge efficiency, as the basic parameters for energy storage stability calculation in step S4.

[0117] Step S2: Collect real-time equipment operation data, and correct the virtual power plant model based on the real-time equipment operation data to obtain the corrected virtual power plant model.

[0118] Data access: Collect real-time measurement data (voltage / current accuracy 0.1%) from edge devices (PMU, smart meters) and device operation logs (inverter fault codes, energy storage BMS alarms).

[0119] Dynamic deviation detection: A 1-minute sliding window is used to calculate the deviation between the model's predicted value and the measured value (e.g., a photovoltaic power output prediction error > 5% triggers correction). Residual analysis is used to locate the source of the deviation (equipment-level fault / aggregation unit-level scheduling delay).

[0120] Adaptive correction strategy execution:

[0121] Control strategy refitting: For equipment with excessive deviation, the power response curve is refitted based on the latest 1-hour data (such as updating the load-efficiency curve of a micro gas turbine).

[0122] Influence path weight adjustment: The influence coefficients of virtual power plant behavior on state variables are dynamically updated using the recursive least squares (RLS) method (e.g., the voltage boost effect of energy storage discharge is adjusted from 0.5% / MW to 0.6% / MW).

[0123] Temporary decoupling: Devices with communication interruptions exceeding 5 minutes are marked as "offline" and temporarily removed from the model.

[0124] Model validation after correction: The prediction error after correction should be <3% (photovoltaic) / 2% (micro gas turbine), and the output should include the corrected virtual power plant model with the latest parameters.

[0125] The latest parameters include energy storage charging and discharging power: the charging and discharging power of the energy storage system at each moment within N time steps (unit: kW); controllable load adjustment: the adjustment of controllable load at each moment (such as reducing or increasing power consumption, unit: kW).

[0126] Step S3: Collect real-time photovoltaic power generation data, construct a photovoltaic output prediction model based on the modified virtual power plant model, input the real-time photovoltaic power generation data into the photovoltaic output prediction model, output the photovoltaic power generation prediction trajectory, and calculate the photovoltaic output fluctuation rate based on the photovoltaic power generation prediction trajectory.

[0127] Collect real-time photovoltaic power generation data :

[0128] The photovoltaic power generation is collected in real time at various moments through intelligent monitoring devices deployed in the photovoltaic power plant (such as power acquisition devices and inverter data interfaces). And store them in chronological order to form a data sequence. Data can be obtained directly through a photovoltaic monitoring system or data acquisition platform.

[0129] Based on the modified virtual power plant model, a photovoltaic power output prediction model is constructed:

[0130] Define input and output characteristics:

[0131] Input sequence: Historical photovoltaic power output sequence, taking the first few digits. Each time step is denoted as: ,in for The original photovoltaic power output at any given time (unit: kW), with the time step as the input dimension. .

[0132] Output sequence: Future The predicted output sequence at each time step is denoted as: ;

[0133] Objective: Through historical sequences Directly predict the original scale .

[0134] Training set: The original output sequence is divided into training samples ( , ):

[0135] in: , The sample set is ,in This represents the total number of time steps.

[0136] The photovoltaic power output prediction model based on the LSTM model has the following inner structure:

[0137] The core gating mechanism of LSTM remains unchanged, directly processing the input at the original data scale:

[0138] Forgotten Gate: ;

[0139] Input Gate: , ;

[0140] Cell status update: ;

[0141] Output gate: , ;

[0142] in, This represents the activation function, usually the sigmoid function. , , , These represent the weight matrices for the forget gate, input gate, cell state update, and output gate, respectively, and are used to learn the relationship between the input data and each gate. This represents the input vector at time step t, which is the original power sequence data. This represents the hidden state at time step t-1, containing information from previous time steps, and is used to pass past information to the current time step. , , , These are the bias terms for the forget gate, input gate, cell state update, and output gate, respectively, used to adjust the output of the activation function and increase the flexibility of the model. This represents element-level multiplication, which involves multiplying elements at corresponding positions and is used to control the flow of information. This represents the candidate cell state at time step t. It is activated by the tanh function, generating a vector in the interval (-1,1) to update the cell state. Represents the cell state at time step t, containing long-term memory information, and is updated through the control of the forget gate and the input gate. : Indicates the activation value of the output gate at time step t. : Represents the hidden state at time step t.

[0143] Multi-layer LSTM and fully connected layer: The last layer is directly mapped to the output of the original power scale through the fully connected layer, and the output dimension is consistent with the original data (unit: kW).

[0144] Model training objective: The loss function directly affects the prediction error of the original data, and the mean squared error (MSE) is:

[0145]

[0146] in, Mean square error, For model parameters, This represents the loss function, used to measure the model's predictive performance. These are model parameters; It is the total number of data points. Iterate through each data point ( ); Is the model in Predicted value at time, yes The true value of a moment; It represents the square of the L2 norm of the difference between the predicted value and the true value (i.e., the square of the difference).

[0147] The prediction error of the original data is minimized by optimizing using gradient descent.

[0148] Output predictive force trajectory generation: Input the current real-time raw data at k time steps The model directly outputs the predicted sequence for the next n steps at the original scale: ;

[0149] The predicted power trajectory is obtained directly. .

[0150] Predicted trajectory at the original scale Calculate the relative rate of change between adjacent time steps:

[0151]

[0152] in, Represents the relative rate of change between the i-th time step and the (i-1)-th time step. It is the predicted value at the (i-1)th time step, used as the benchmark value for calculating the relative rate of change. It is the predicted value at the i-th time step, where i is the time step index, representing a discrete time point;

[0153] Calculate the average of the relative rates of change based on the relative rates of change at adjacent time steps:

[0154]

[0155] in, The average value represents the relative rate of change, where n is the number of time steps in the prediction, and represents the number of predicted values ​​contained in the predicted trajectory vector. The relative rate of change between adjacent time steps, where i is the time step index, representing a discrete time point;

[0156] Volatility is defined as the standard deviation of a relative rate of change.

[0157] Calculation of photovoltaic power output volatility:

[0158]

[0159] in, This represents the photovoltaic power output volatility, where n is the prediction time step. The relative rate of change between adjacent time steps The average value of the relative rate of change is represented by the formula, which is directly based on the relative fluctuations of the original power sequence and reflects the force stability. i is the time step index, representing a discrete time point.

[0160] Step S4: Collect real-time load demand and fuel reserve data, construct a fuel output prediction model based on the modified virtual power plant model, input the real-time load demand and fuel reserve data into the fuel output prediction model, output the fuel output trajectory, calculate the fuel output stability based on the fuel output trajectory, and calculate the power generation side reliability index based on the photovoltaic output volatility and fuel output stability.

[0161] Collect real-time load requirements:

[0162] With the help of smart meters, distribution automation systems or energy management systems (EMS) on the power grid side, the electricity load of each node in the power grid can be monitored in real time. These systems can directly output the current power grid load value.

[0163] Collect fuel reserve data:

[0164] Sensors (such as level sensors, pressure sensors, and weighing sensors) are installed on fuel storage equipment (such as oil tanks, gas tanks, and fuel silos) to measure fuel reserves in real time. The sensors transmit the data to an industrial control system (such as a PLC) or management information system to extract the current fuel reserve data.

[0165] Construct a fuel output prediction model based on the Model Predictive Control (MPC) algorithm:

[0166] Input layer:

[0167] Real-time load requirements: (Current grid load, unit: kW, external disturbances)

[0168] Control input: (Fuel adjustment amount, such as fuel injection rate, unit: tons / hour, decision variable)

[0169] Output layer: Predicted output sequence for the next N steps: ,in This represents the predicted fuel output at time t+N from time t, providing future output information for grid dispatch and energy management.

[0170] State variables :set up For fuel reserves, These are the unit operating parameters (such as temperature and pressure) that affect the output.

[0171] enter Fuel supply rate and real-time load demand.

[0172] Equations of state:

[0173]

[0174] in, It represents the fuel reserves at the next time step t. This refers to the fuel supply rate and real-time load demand at time t. This is the fuel consumption coefficient.

[0175]

[0176] in, These are the unit operating parameters that affect output at the next moment (t). This refers to the fuel supply rate and real-time load demand at time t. These are the unit operating parameters that affect output at time t;

[0177] ( ) is a function describing changes in the unit's state, which can be determined through mechanistic analysis or data fitting.

[0178] Output equation:

[0179]

[0180] in, It is the output value at time t. , For unit characteristic coefficients, Let be the fuel reserve at time t. These are the unit operating parameters that affect output at time t.

[0181] Define the objective function and let the prediction time domain be... Reference output is ;

[0182] The objective function is:

[0183]

[0184] in, It is the objective function, providing optimization direction for model predictive control. Predict the time domain, For reference purposes, based on Time information The predicted output value at any given time; In the prediction time domain Inside, every step Upward predicted output (based on Time information (Predicted output at time) and reference output The sum of squared errors is used to calculate the predicted output. This part forces the predicted output to track the reference output as closely as possible. By minimizing the sum of squared errors, it improves the fit between the predicted output and the expected reference value, ensuring that the system output meets the expected target. These are control quantities (fuel supply rate, real-time load demand). Weighting coefficients, suppression control quantities Dramatic changes This is the control variable. This part penalizes the control variable, suppressing its drastic changes. If the control variable changes too frequently or too much, it will increase system wear or energy consumption. This is addressed through weighting. Smoothness of the balance control quantity. The larger the value, the stronger the penalty for changes in the control quantity, and the more stable the control tends to be; The smaller the value, the more emphasis is placed on the predicted output tracking the reference value.

[0185] In short, this objective function provides an optimization direction for model predictive control by balancing prediction accuracy and control smoothness, ensuring that the system meets the output expectation while the control process is stable and reasonable.

[0186] Set constraints:

[0187] Control constraints: (The feasible range of fuel supply rate and unit adjustment parameters).

[0188] Output constraints: (Upper and lower limits of fuel output).

[0189] State constraints: (Minimum fuel reserves).

[0190] Fuel output trajectory generation (rolling optimization solution):

[0191] At each sampling time Utilize the current state Solve the optimization problem of the above objective function (such as quadratic partitioning) to obtain the future Step control sequence .

[0192] Only the first step of control quantity is executed. Waiting for the next moment Update status Repeat the above process.

[0193] Calculate using the state equation and output equation That is, the future Predict fuel output trajectory step by step.

[0194] This model generates a smooth fuel output trajectory by using the rolling optimization mechanism of MPC, under the premise of satisfying fuel reserve and output constraints, and quantifies its stability by standard deviation, avoiding normalization preprocessing and directly modeling and calculating based on the original physical quantities.

[0195] Fuel output stability calculation:

[0196] Based on predicted trajectory The stability index is defined as the standard deviation of output fluctuation (reflecting the smoothness of the trajectory):

[0197] =

[0198] in, The smaller the value, the more stable the output and the smaller the fluctuation. Is Always Predicted fuel output at any given time (based on prediction results from model predictive control); Is Always Predicted fuel output at any given time. It is the total number of steps in the prediction time domain, i.e., predicting the future. Fuel output at any given moment (e.g., predicting the next 24 hours, one point per hour). ).

[0199] Based on the photovoltaic power output volatility and fuel output stability, the power generation side reliability index is calculated:

[0200]

[0201] in, It is the reliability index of the power generation side. and These are the weights for photovoltaic power output volatility and fuel power output stability, respectively. =1, It is the fluctuation rate of photovoltaic power output. It refers to fuel output stability.

[0202] Step S5: Calculate the energy storage stability based on the energy storage charging and discharging power in the modified virtual power plant model, calculate the controllable load stability based on the controllable load adjustment amount in the modified virtual power plant model, and calculate the grid-side stability index by combining the energy storage stability and the controllable load stability.

[0203] Based on a modified virtual power plant model, the impact of energy storage and controllable load regulation capabilities on grid stability is quantitatively evaluated, and regulation capability indicators are defined.

[0204] Energy storage charging and discharging power: denoted as { } , which represents the charging and discharging power of the energy storage system at each moment within N time steps (unit: kW);

[0205] Controllable load adjustment amount: denoted as { } This indicates the amount of controllable load adjustment at any given time (such as a decrease or increase in electrical power, unit: kW).

[0206] Calculate the mean absolute deviation (MAD) to measure stability.

[0207] Energy storage stability:

[0208] First, calculate the average energy storage:

[0209]

[0210] in, This represents the average value of the energy storage charging and discharging power, where N is the total number of time steps. Let t be the energy storage charging and discharging power at time t. This formula obtains the mean value of the energy storage charging and discharging power by summing the energy storage power over N time steps and taking the average value, which reflects the overall level of energy storage power, where t is the identifier of the discrete time step.

[0211] Calculate the mean absolute deviation of energy storage as a measure of energy storage stability:

[0212]

[0213] in, The stability of energy storage charging and discharging power is measured by the degree of deviation of the energy storage power from the mean. The smaller the deviation, the more stable the energy storage charging and discharging. This represents the average value of the energy storage charging and discharging power. Let N be the energy storage charging and discharging power at time t, N be the total number of time steps, and t be the identifier of the discrete time step.

[0214] Controllable load stability:

[0215] First, calculate the average value of the controllable load adjustment:

[0216]

[0217] in, This represents the average value of the controllable load adjustment. The total number of time steps. For a moment The controllable load adjustment amount (unit: kW, such as reducing or increasing electrical power consumption), where t is the identifier of the discrete time step.

[0218] This formula is derived from... Controllable load adjustment at each time step The summation and averaging yields the mean value of the controllable load adjustment, reflecting the overall level of controllable load adjustment.

[0219] Calculate the mean absolute deviation of the controllable load adjustment as a measure of controllable load stability:

[0220]

[0221] in, The stability of controllable load regulation is measured by the degree of deviation of the controllable load regulation from the mean. The smaller the deviation, the more stable the controllable load regulation. This represents the average value of the controllable load adjustment. Controllable load adjustment at each moment. First, through the average value. Characterize the overall level of controllable load adjustment, and then utilize To quantify its stability, the smaller the value, the more stable the controllable load adjustment.

[0222] Combining the energy storage stability and controllable load stability, the grid-side stability index is calculated as follows:

[0223] Energy storage charging and discharging power and controllable load adjustment Calculate the normalized fluctuation intensity separately, and amplify the impact of extreme fluctuations using a nonlinear function (such as an exponential function):

[0224] Energy storage fluctuation intensity:

[0225]

[0226] in, This represents the fluctuation intensity of energy storage at time t, normalized by the ratio of the absolute change in energy storage charging and discharging power at adjacent times to the maximum charging and discharging power of the energy storage. The charging and discharging power of the stored energy at time t. It is the charging and discharging power of the stored energy at time t-1; It is the maximum charging and discharging power of the energy storage system.

[0227] Average energy storage fluctuation intensity:

[0228]

[0229] in, Average energy storage fluctuation intensity, through the analysis of... The result is obtained by summing the exponents (to amplify the impact of extreme fluctuations) and then averaging them. This represents the fluctuation intensity of energy storage at time t, expressed as the absolute change in energy storage charging and discharging power between adjacent times and the maximum charging and discharging power of the energy storage. The ratio is normalized, where N is the total number of time steps.

[0230] Controllable load fluctuation intensity:

[0231]

[0232] in, It indicates at time. The fluctuation intensity of the controllable load is normalized by the ratio of the absolute change in the controllable load adjustment at adjacent time points to the maximum controllable load adjustment. It is a moment The adjustment amount of controllable load, :time Adjustment amount of controllable load; It is the maximum adjustable amount of the controllable load (unit: kW), used to normalize the fluctuation intensity.

[0233]

[0234] in, It is the average controllable load fluctuation intensity. It indicates at time. The fluctuation intensity of the controllable load is normalized by the ratio of the absolute change in the controllable load adjustment at adjacent time points to the maximum adjustment of the controllable load. t represents the total number of time steps, and t represents the identifier of the discrete time step.

[0235] Introduce a nonlinear function (such as the sigmoid function) that dynamically changes the weights with the adjustment amount to reflect the characteristic that "the larger the adjustment amount, the more significant the impact on stability":

[0236] Energy storage weight: ( (To adjust parameters) when Increase (strong energy storage fluctuations), Approaching 1, highlighting the dominant influence of energy storage on stability;

[0237] Controllable load weighting: ;

[0238] Considering energy storage and controllable load regulation (If both fluctuate drastically at the same time, the risk to grid stability is nonlinearly superimposed.) A quadratic interaction term is introduced: ;

[0239] Grid-side stability index (nonlinear comprehensive model)

[0240]

[0241] in, and These are the average energy storage fluctuation intensity and the average controllable load fluctuation intensity, respectively. These are the weights of the average energy storage fluctuation intensity and the average controllable load fluctuation intensity, respectively. Weights of interaction items ( This reflects the amplification effect of collaborative risks. It is an introduced secondary interaction term; when the fluctuation of a single adjustment term is strong, it is through... Dynamic weighting amplifies its influence; when both fluctuate simultaneously... Further increase the index value to reflect the risk superposition of "1+1>2".

[0242] Step S6: Combine the power generation side reliability index and the power grid side stability index to calculate the distribution network status assessment index.

[0243] The reliability index of the power generation side was determined using an expert evaluation method. Weight of grid-side stability index :

[0244] Assemble an expert team: Invite senior experts in power system planning, operation and maintenance, reliability analysis and other fields, covering professionals from both the generation and grid sides.

[0245] Background information: To provide experts with a detailed explanation of the objectives of the distribution network condition assessment, the specific impacts of generation-side reliability and grid-side stability, and the implications of the formulas, ensuring that experts fully understand the assessment background.

[0246] Independent scoring: Each expert provides an independent score based on their own experience and professional judgment. , The value of (satisfying) ).

[0247] Results Summary and Discussion: Collect expert scores and calculate the average or median. If there are significant differences of opinion among experts, organize multiple rounds of discussions to encourage experts to refine their judgments through exchange and gradually reach a consensus.

[0248] Determining weights: This is based on the final converged expert opinions. For example, if experts, after discussion, believe that generation-side reliability has a more significant impact on the distribution network status, then weights can be determined. , This ensures that the weight allocation aligns with professional experience and judgment, guaranteeing the rationality of the evaluation model.

[0249] Combining the generation-side reliability index and the grid-side stability index, the distribution network condition assessment index is calculated:

[0250] The distribution network condition assessment index is calculated using a weighted summation method, and the formula is as follows:

[0251]

[0252] in, It is a power distribution network condition assessment index. and These are the generation-side reliability index and the grid-side stability index. These are the weights of the generation-side reliability index and the grid-side stability index, respectively.

[0253] This formula integrates the assessment results from the generation side and the grid side, quantitatively reflecting the overall state of the distribution network and providing a basis for subsequent comprehensive analysis.

[0254] Step S7: Collect weather disturbance data and power distribution equipment health index data, and calculate the power distribution network impact coefficient based on the weather disturbance data and power distribution equipment health index data.

[0255] Collection and preprocessing of weather disturbance data and power distribution equipment health index data:

[0256] Collect weather disturbance data (such as temperature, wind speed, rainfall) and health indicators of power distribution equipment (e.g., equipment aging rate, insulation resistance). Normalize the data:

[0257]

[0258] in, This represents the normalized weather disturbance index data, used to map the original data to the [0,1] interval for easier subsequent processing; This represents the original weather disturbance index data, such as measured values ​​of environmental factors like temperature and humidity; This represents the minimum value in the weather disturbance index data, used to determine the lower limit of the data range; This represents the maximum value in the weather disturbance index data, used to determine the upper limit of the data range;

[0259]

[0260] in, This represents the normalized health index data of power distribution equipment, and the original data is also mapped to the [0,1] interval; This represents the original health indicators of the power distribution equipment, such as equipment aging rate and insulation resistance. This represents the minimum value in the health indicator data of power distribution equipment; This represents the maximum value in the health indicator data of the power distribution equipment.

[0261] Weighting and calculation, setting weather disturbance weights and the health weight of power distribution equipment ( And assign sub-weights to each sub-indicator. ( )and ( ).

[0262] Calculate the influence coefficient of the distribution network

[0263]

[0264] in, This represents the distribution network impact coefficient, which comprehensively reflects the degree of impact of weather disturbances and the health status of power distribution equipment on the operation of the distribution network. and These represent the weather disturbance weight and the power distribution equipment health weight, respectively, used to quantify the relative importance of weather factors and equipment health status on the power distribution network; This represents the sub-weights of the weather disturbance index, corresponding to different weather disturbance factors. This represents the sub-weights of the health indicators of power distribution equipment, corresponding to different equipment health indicators; This represents the normalized weather disturbance index data. This represents the normalized health index data for power distribution equipment. It is a comprehensive indicator that integrates weather and equipment health factors through a weighted summation method to assess the overall operating status and risk level of the power distribution network.

[0265] Step S8: Combine the distribution network status assessment index and the distribution network impact coefficient to calculate the comprehensive status assessment index. Based on the comprehensive status assessment index, identify the current distribution network status and take corresponding measures for potential risk status of the distribution network.

[0266] Combining the aforementioned distribution network condition assessment index and distribution network shadow coefficient, the comprehensive condition assessment index is calculated as follows:

[0267] .

[0268] in, It is a comprehensive status assessment index. It is a power distribution network condition assessment index. It is the distribution network impact coefficient. It is the influence coefficient amplification factor.

[0269] Based on the comprehensive condition assessment index, the current state of the distribution network is identified, and corresponding measures are taken for potential risks to the distribution network.

[0270] Thresholds are determined through statistical analysis (such as calculating the mean and standard deviation) combined with practical operational experience. For example, the mean of normal operating data plus 1.5 times the standard deviation can be set as the threshold. The mean plus 2.5 times the standard deviation is set as .

[0271] Normal state ( :

[0272] Meaning: The generation side of the distribution network has high reliability, the grid side operates stably, and weather disturbances and the health status of distribution equipment have minimal impact on the system. The system as a whole is in a safe and stable operating state, with all indicators within normal ranges, capable of meeting users' electricity needs without requiring special measures.

[0273] Performance: The power generation equipment provides stable power supply, with voltage and frequency fluctuations within the allowable range. There are no abnormal operating signals from the power distribution equipment, and users have a good power experience.

[0274] Countermeasures: Continuously conduct daily equipment inspections and maintenance, and carry out preventive tests and maintenance as planned; optimize the data monitoring and analysis system, accumulate normal operation data, and provide a basis for dynamic adjustment of thresholds; organize skills training for operation and maintenance personnel to ensure the ability to respond to emergencies.

[0275] Potential risk status ( ):

[0276] Meaning: The distribution network exhibits certain instability factors, such as decreased reliability on the generation side, minor fluctuations on the grid side, or weather disturbances and equipment health issues beginning to impact the system. Although the system can currently maintain operation, it may develop into a serious fault if not addressed promptly.

[0277] Symptoms: Some equipment may experience minor signal anomalies (such as increased equipment temperature or local voltage fluctuations), and the power output fluctuation rate on the generation side or the regulation fluctuation on the grid side may begin to increase. Close monitoring of the system's operating status is required.

[0278] Countermeasures: Enhanced monitoring: Implement high-frequency real-time monitoring, and add specialized monitoring such as infrared thermography and partial discharge detection for key equipment; Fault cause diagnosis: Utilize big data analysis and AI algorithms to model and predict abnormal data, and locate potential fault points; Resource allocation: Deploy emergency repair materials to key areas in advance, coordinate backup power or adjustable load resources, and make emergency preparations; Load optimization: Guide users to use electricity during off-peak hours through demand response strategies to mitigate the risk of local overload.

[0279] High-risk status ( ):

[0280] Meaning: The power distribution network faces serious safety hazards. The reliability of the generation side is significantly reduced, and the stability of the grid side is compromised, or it is strongly affected by factors such as extreme weather disturbances and serious equipment failures. The system may fail at any time, even leading to large-scale power outages, requiring immediate emergency measures.

[0281] Symptoms: Frequent equipment malfunction alarms, voltage and frequency deviating significantly from normal values, potential power outages in some areas, and system collapse risks.

[0282] Response measures: Activate Level 1 emergency response: Establish an on-site command center to coordinate repair teams, materials, and expert resources; Emergency load control: Prioritize cutting off non-essential loads to ensure power supply to critical users such as hospitals and transportation, preventing system collapse; Fault isolation and repair: Utilize smart switches to quickly isolate faulty sections and employ technologies such as live-line working and mobile energy storage vehicles to shorten outage time; Dynamic dispatch: Coordinate with the power generation side to adjust output and activate distributed power sources or diesel generators to supplement power shortages; Information dissemination: Announce the outage area and repair progress in real time through official channels to reassure users and solicit emergency support.

[0283] This method collects and analyzes basic data, historical and real-time equipment operation data of the distribution network. Utilizing virtual power plant model correction, photovoltaic output prediction, and fuel output prediction, it accurately assesses the reliability index of the generation side and the stability index of the grid side, quantifies the impact of energy storage and controllable load regulation capabilities, and calculates a comprehensive state assessment index by combining weather disturbance data and distribution equipment health index data. This allows for a comprehensive and accurate identification of the current state of the distribution network. This process not only improves the reliability and stability of the distribution network but also optimizes operation and maintenance, effectively assesses and addresses potential risks, supports distribution network planning and decision-making, promotes the effective integration of distributed energy resources, and ultimately achieves safe, stable, and economical operation of the distribution network, while also promoting the efficient utilization and sustainable development of new energy sources.

[0284] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, the phrase "comprising an element defined as..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0285] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their likenesses.

Claims

1. A power system state analysis method based on large-scale distributed energy resource integration, characterized in that, Includes the following steps: Step S1: Collect basic data of the power distribution network and historical equipment operation data, and construct a virtual power plant model based on the basic data of the power distribution network and historical equipment operation data; Step S2: Collect real-time equipment operation data, and correct the virtual power plant model based on the real-time equipment operation data to obtain a corrected virtual power plant model; Step S3: Collect real-time photovoltaic power generation data, construct a photovoltaic output prediction model based on the modified virtual power plant model, input the real-time photovoltaic power generation data into the photovoltaic output prediction model, and output the photovoltaic power generation prediction trajectory. Based on the photovoltaic power generation prediction trajectory, calculate the photovoltaic output volatility, specifically including: Output predicted power trajectory generation: Input the current real-time raw data at k time steps The model directly outputs the predicted force trajectory for the next n steps at the original scale: ,in, Indicates at time The collected measured values ​​of photovoltaic power generation; Indicates time Predicted photovoltaic power generation output; The predicted power trajectory Calculate the relative rate of change between adjacent time steps: ； in, It is in the corresponding trajectory In time points and trajectories The relative rate of change between points in time It is in the corresponding trajectory The predicted value at a given time point is used as the benchmark value for calculating the relative rate of change. It is in the corresponding trajectory Predicted values ​​at a given time point This is a time step index, representing discrete time points; Calculate the average of the relative rates of change based on the relative rates of change at adjacent time steps: ； in, The average value represents the relative rate of change, where n is the number of time steps in the prediction, and represents the number of predicted values ​​contained in the predicted trajectory vector. The relative rate of change between adjacent time steps, where i is the time step index, representing a discrete time point; Volatility is defined as the standard deviation of a relative rate of change. Calculation of photovoltaic power output volatility: ； in, This represents the photovoltaic power output volatility, where n is the prediction time step. The relative rate of change between adjacent time steps The average value of the relative rate of change is represented by the formula, which is directly based on the relative fluctuations of the original power sequence and reflects the force stability. i is the time step index, representing a discrete time point. Step S4: Collect real-time load demand and fuel reserve data, construct a fuel output prediction model based on the modified virtual power plant model, input the real-time load demand and fuel reserve data into the fuel output prediction model, output the fuel output trajectory, calculate the fuel output stability based on the fuel output trajectory, and calculate the power generation side reliability index based on the photovoltaic output volatility and fuel output stability. Step S5: Calculate the energy storage stability based on the energy storage charging and discharging power in the modified virtual power plant model, calculate the controllable load stability based on the controllable load adjustment amount in the modified virtual power plant model, and calculate the grid-side stability index by combining the energy storage stability and the controllable load stability. Step S6: Calculate the distribution network status assessment index by combining the generation side reliability index and the grid side stability index.

2. The power system state analysis method based on large-scale distributed energy access according to claim 1, characterized in that, The construction of the virtual power plant model includes: Building a virtual power plant model: Using a 5km radius as a standard, and leveraging GIS spatial analysis capabilities, distributed energy resources within the same geographical area are grouped into a single aggregation unit. Dispatch homogeneity aggregation: Distributed energy resources belonging to the same operator or supporting unified command issuance are aggregated into a single unit regardless of their geographical location. Response consistency aggregation: Aggregation is based on the response characteristics of the equipment. For energy storage units, energy storage devices with charge / discharge rate deviations within 10% are aggregated together. For power generation equipment, photovoltaic power plants and micro gas turbines with active power regulation times of similar magnitude are aggregated to ensure that equipment within the aggregation unit can operate collaboratively when participating in grid regulation, thereby improving regulation efficiency. Equipment list compilation: Real-time operating status: Current output power of photovoltaic, SOC and charge / discharge status of energy storage, and current fuel consumption rate of micro gas turbine; Communication status: timestamp of the most recent data update, average communication latency; Initialize model parameters: For each aggregation unit, pre-calculate: Photovoltaics: installed capacity share, historical output volatility baseline; Micro gas turbines: rated output share, fuel reserve safety threshold; Energy storage: available adjustable capacity.

3. The power system state analysis method based on large-scale distributed energy access according to claim 2, characterized in that, The process of collecting real-time equipment operation data and correcting the virtual power plant model based on the real-time equipment operation data to obtain a corrected virtual power plant model includes: Data access: Collect real-time measurement data and device operation logs from edge devices; Dynamic deviation detection: A 1-minute sliding window is used to calculate the deviation between the model's predicted value and the measured value. If the photovoltaic output prediction error is greater than 5%, a correction is triggered. The source of the deviation is located through residual analysis. Adaptive correction strategy execution: Control strategy refitting: For equipment with excessive deviation, the power response curve is refitted based on the latest 1-hour data; Influence path weight adjustment: Dynamically update the influence coefficients of virtual power plant behavior on state variables using recursive least squares method; Temporary decoupling: Devices with communication interruptions exceeding 5 minutes are marked as "offline" and temporarily removed from the model; Model validation after correction: The prediction error after correction should be less than 3%, and the output should be a corrected virtual power plant model with the latest parameters; The latest parameters include energy storage charging and discharging power: the charging and discharging power of the energy storage system at each moment within N time steps; and controllable load adjustment amount: the adjustment amount of the controllable load at each moment.

4. The power system state analysis method based on large-scale distributed energy access according to claim 3, characterized in that, The calculation of fuel output stability based on the fuel output trajectory includes: Calculate using the state equation and output equation That is, the future Predicting fuel output trajectory step by step It refers to Predicted fuel output at any given moment; The model generates a smooth fuel output trajectory by using the rolling optimization mechanism of MPC, under the premise of satisfying fuel storage and output constraints, and quantifies its stability by standard deviation, avoiding normalization preprocessing and directly modeling and calculating based on the original physical quantities. Fuel output stability calculation: Based on predicted trajectory The stability index is defined as the standard deviation of output fluctuation: = ； in, This represents fuel output stability; the smaller the value, the more stable the output and the smaller the fluctuation. Is Always Predicted fuel output at any given moment; Is Always Predicted fuel output at any given moment; It is the total number of steps in the prediction time domain, i.e., predicting the future. Fuel output at any given moment.

5. A power system state analysis method based on large-scale distributed energy access according to claim 4, characterized in that, The generation-side reliability index is calculated based on the photovoltaic power output volatility and fuel power output stability, including: Based on the photovoltaic power output volatility and fuel output stability, the power generation side reliability index is calculated: ； in, It is the reliability index of the power generation side. and These are the weights for photovoltaic power output volatility and fuel power output stability, respectively. =1, It is the fluctuation rate of photovoltaic power output. It refers to fuel output stability.

6. The power system state analysis method based on large-scale distributed energy access according to claim 5, characterized in that, The step of calculating energy storage stability based on the energy storage charging and discharging power in the modified virtual power plant model includes: Based on a modified virtual power plant model, the impact of energy storage and controllable load regulation capabilities on grid stability is quantitatively evaluated, and regulation capability indicators are defined. Energy storage charging and discharging power: denoted as { } represents the charging and discharging power of the energy storage system at each moment within N time steps; Calculating the mean absolute deviation measures stability Energy storage stability: First, calculate the average energy storage: ； in, This represents the average value of the energy storage charging and discharging power, where N is the total number of time steps. Let t be the energy storage charging and discharging power at time t. This formula obtains the average value of the energy storage charging and discharging power by summing the energy storage power over N time steps and taking the average value, which reflects the overall level of energy storage power. t is the identifier of the discrete time step. Calculate the mean absolute deviation of energy storage as a measure of energy storage stability: ； in, As a measure of the stability of energy storage charging and discharging power, it reflects the average deviation of the energy storage power from the mean. The smaller the deviation, the more stable the energy storage charging and discharging. This represents the average value of the energy storage charging and discharging power. Let N be the energy storage charging and discharging power at time t, N be the total number of time steps, and t be the identifier of the discrete time step.

7. A power system state analysis method based on large-scale distributed energy access according to claim 6, characterized in that, The step of calculating the controllable load stability based on the controllable load adjustment amount in the modified virtual power plant model includes: Controllable load adjustment amount: denoted as { } represents the adjustment amount of the controllable load at each moment; Find the mean value of the controllable load adjustment: ； in, This represents the average value of the controllable load adjustment. The total number of time steps. For a moment The controllable load adjustment amount, where t is the identifier of the discrete time step; This formula is derived from... Controllable load adjustment at each time step The summation and averaging yields the mean value of the controllable load adjustment, reflecting the overall level of controllable load adjustment. Calculate the mean absolute deviation of the controllable load adjustment as a measure of controllable load stability: ； in, As a measure of the stability of controllable load regulation, it reflects the average deviation of the controllable load regulation from its mean. The smaller the deviation, the more stable the controllable load regulation. This represents the average value of the controllable load adjustment. The controllable load adjustment at each moment is first determined by the average value. Characterize the overall level of controllable load adjustment, and then utilize To quantify its stability, the smaller the value, the more stable the controllable load adjustment.

8. A power system state analysis method based on large-scale distributed energy access according to claim 7, characterized in that, The grid-side stability index is calculated by combining the energy storage stability and the controllable load stability, including: A nonlinear function is introduced to dynamically change the weights with the adjustment amount, reflecting the characteristic that "the larger the adjustment amount, the more significant the impact on stability": Energy storage weight: ,when Increase Approaching 1, highlighting the dominant influence of energy storage on stability; Controllable load weighting: ; Considering the synergistic effect of energy storage and controllable load regulation, a second-order interaction term is introduced: ; Grid-side stability index: ； in, It is the grid-side stability index. and These are the average energy storage fluctuation intensity and the average controllable load fluctuation intensity, respectively. and These are the weights of the average energy storage fluctuation intensity and the average controllable load fluctuation intensity, respectively. Weights for interaction items; It is an introduced secondary interaction term; when the fluctuation of a single adjustment term is strong, it is through... and Dynamic weighting amplifies its influence; when both fluctuate simultaneously... Further increase the index value to reflect the risk superposition of "1+1>2".

9. A power system state analysis method based on large-scale distributed energy access according to claim 8, characterized in that, The distribution network condition assessment index is calculated by combining the generation-side reliability index and the grid-side stability index, including: The reliability index of the power generation side was determined using an expert evaluation method. Weight of grid-side stability index ; Combining the generation-side reliability index and the grid-side stability index, the distribution network condition assessment index is calculated: The distribution network condition assessment index is calculated using a weighted summation method, and the formula is as follows: ； in, It is a power distribution network condition assessment index. and These are the generation-side reliability index and the grid-side stability index. These are the weights of the generation-side reliability index and the grid-side stability index, respectively. This formula integrates the evaluation results from the generation side and the grid side, quantitatively reflecting the overall state of the distribution network and providing a basis for subsequent comprehensive analysis.

Images