Panoramic situation awareness and early warning method, system and device for real-time interaction between optical storage and charging microgrid and power distribution network, and medium

Abstract

The invention discloses a panoramic situation awareness and early warning method, system, equipment and medium for real-time interaction between an optical storage and charging micro-grid and a power distribution network, and belongs to the technical field of situation awareness and early warning, and the method comprises the steps: building a data collection network covering the optical storage and charging micro-grid and the power distribution network, obtaining electrical quantity, environmental quantity and equipment state information, carrying out the preprocessing of the data, and carrying out the early warning of the real-time interaction of the optical storage and charging micro-grid and the power distribution network; pre-processed multi-source data is obtained; inputting the multi-source preprocessed data into an improved Kalman filtering algorithm to obtain fused state data; constructing a panoramic situation vector of the optical storage and charging microgrid and the power distribution network based on the fused state data, and obtaining a situation assessment result; inputting the situation assessment result into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to generate a risk assessment result; constructing a multi-objective optimization model according to a risk assessment result, and outputting an intervention control signal; and performing operation monitoring and feedback updating based on the intervention control signal. According to the invention, the accuracy and real-time performance of situation identification are improved.

Description

Technical Field

[0001] This invention relates to the field of situational awareness and early warning technology, specifically to a panoramic situational awareness and early warning method, system, equipment, and medium for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network. Background Technology

[0002] With the large-scale integration of photovoltaic, energy storage, and charging microgrids into distribution networks, traditional power grid monitoring systems are facing unprecedented challenges. Existing situational awareness technologies are mainly designed for traditional power grids and lack a deep understanding of the bidirectional power flow, stochastic load characteristics, and complex operating modes of photovoltaic, energy storage, and charging microgrids.

[0003] Existing early warning methods are mostly based on threshold judgment and statistical analysis, lacking a deep understanding and predictive ability of the system's operational status. This passive early warning mechanism often only detects anomalies after a problem occurs, failing to achieve proactive prevention and early intervention. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention provides a panoramic situational awareness and early warning method, system, equipment and medium for real-time interaction between photovoltaic, energy storage and charging microgrids and distribution networks.

[0005] Therefore, the technical problem addressed by this invention is: how to solve the problem that existing situational awareness technologies are mainly designed for traditional power grids and lack a deep understanding of the bidirectional power flow, stochastic load characteristics, and complex operating modes of photovoltaic-storage-charging microgrids. Furthermore, existing early warning methods are mostly based on threshold judgments and statistical analysis, lacking a deep understanding and predictive capability of the system's operating status.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a panoramic situational awareness and early warning method for real-time interaction between a photovoltaic-storage-charging microgrid and a distribution network, comprising: establishing a data acquisition network covering the photovoltaic-storage-charging microgrid and the distribution network to acquire electrical quantities, environmental quantities, and equipment status information; preprocessing the data to obtain preprocessed multi-source data; inputting the preprocessed multi-source data into an improved Kalman filter algorithm; achieving high-reliability data fusion through state prediction, measurement updates, and dynamic weight allocation to obtain fused state data; constructing a panoramic situational vector of the photovoltaic-storage-charging microgrid and the distribution network based on the fused state data; and extracting operational features using a deep fusion network. The system comprehensively identifies and assesses the operational status to obtain a situation assessment result. This result is then input into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict future operational status and potential risk probabilities. A comprehensive evaluation method is used to classify risk levels, generating a risk assessment result. Based on the risk assessment result, a multi-objective optimization model is constructed, and an adaptive intervention strategy is generated using a reinforcement learning algorithm. This enables dynamic adjustment of intervention intensity and strategy optimization, outputting an intervention control signal. Based on the intervention control signal, operational monitoring and feedback updates are performed, establishing a closed-loop mechanism. The situation awareness model and early warning thresholds are corrected based on monitoring feedback to achieve continuous optimization and proactive prevention and control.

[0007] As a preferred embodiment of the panoramic situational awareness and early warning method for real-time interaction between the photovoltaic-storage-charging microgrid and the distribution network described in this invention, the following steps are included: establishing a data acquisition network covering the photovoltaic-storage-charging microgrid and the distribution network to acquire electrical quantities, environmental quantities, and equipment status information, and preprocessing the data to obtain preprocessed multi-source data. This includes: establishing a data acquisition architecture oriented towards the photovoltaic-storage-charging microgrid and the distribution network; configuring communication interfaces to achieve multi-source data acquisition; performing time alignment and format conversion on the acquired data to form data in a unified format; and preprocessing the unified format data to obtain preprocessed multi-source data.

[0008] As a preferred embodiment of the panoramic situational awareness and early warning method for real-time interaction between the photovoltaic, energy storage, and charging microgrid and the distribution network described in this invention, the method involves: inputting multi-source preprocessed data into an improved Kalman filter algorithm; achieving high-reliability data fusion through state prediction, measurement updates, and dynamic weight allocation to obtain fused state data; establishing a state transition model and an observation model for the photovoltaic, energy storage, and charging microgrid and the distribution network; determining the relationship between state variables and input variables; performing state prediction on the preprocessed multi-source data to generate the predicted state of the photovoltaic, energy storage, and charging microgrid and the distribution network; correcting the predicted state based on the measurement update mechanism to obtain the updated state estimate of the photovoltaic, energy storage, and charging microgrid and the distribution network; and performing fusion calculations based on the data source weight coefficients to output the fused state data.

[0009] As a preferred embodiment of the panoramic situational awareness and early warning method for real-time interaction between the photovoltaic, energy storage, and charging microgrid and the distribution network described in this invention, the following steps are included: constructing a panoramic situational vector of the photovoltaic, energy storage, and charging microgrid and the distribution network based on the fused state data; extracting operational features using a deep fusion network; completing the comprehensive identification and level assessment of the operational status; and obtaining the situational assessment result. This includes: establishing a situational vector representation structure of the photovoltaic, energy storage, and charging microgrid and the distribution network based on the fused state data; determining feature parameters; inputting the situational vector of the photovoltaic, energy storage, and charging microgrid and the distribution network into a deep fusion network to extract operational features; establishing an identification and classification model based on the operational features to assess the operational status of the photovoltaic, energy storage, and charging microgrid and the distribution network; and outputting the situational assessment result.

[0010] As a preferred embodiment of the panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as described in this invention, the step of inputting the situational assessment results into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict future operating conditions and potential risk probabilities, and then using a comprehensive evaluation method to classify risk levels and generate risk assessment results, includes: inputting the situational assessment results into a long short-term memory network model to predict future operating condition data; inputting the operating condition data into a Bayesian network to calculate the posterior probability of potential risks based on evidence conditions; constructing an evaluation matrix based on the prediction results and performing a synthesis operation according to a preset weight vector; and outputting the synthesis result as a risk level vector.

[0011] This preferred solution, by sequentially inputting the situation assessment results into a long short-term memory network and a Bayesian network, enables joint inference of future operational trends and risk probabilities. By combining the risk level vector calculated from the weight vector and the evaluation matrix, continuous prediction results can be transformed into clear risk levels, improving the timeliness of risk identification and the accuracy of classification.

[0012] As a preferred embodiment of the panoramic situational awareness and early warning method for real-time interaction between the photovoltaic-storage-charging microgrid and the distribution network described in this invention, the following steps are included: constructing a multi-objective optimization model based on risk assessment results, generating an adaptive intervention strategy using a reinforcement learning algorithm, dynamically adjusting the intervention intensity and optimizing the strategy, and outputting an intervention control signal. This includes: constructing a multi-objective function vector composed of safety, economic, and reliability objectives; setting control variables based on risk assessment results to form an initial solution for the intervention strategy; updating the state-action based on the deviation signal between the photovoltaic-storage-charging microgrid and the distribution network using a reinforcement learning algorithm; adaptively adjusting the intervention intensity based on the maximum intervention intensity and PID control parameters; and outputting the intervention control signal.

[0013] This preferred approach, by constructing a multi-objective function vector and introducing a reinforcement learning mechanism, can dynamically adjust the intervention strategy based on the risk assessment results. It can adaptively adjust the intervention intensity while controlling the deviation signal, avoiding excessive or insufficient intervention, thereby improving the comprehensive balance of the intervention strategy in terms of safety, economy and reliability.

[0014] As a preferred embodiment of the panoramic situational awareness and early warning method for real-time interaction between the photovoltaic, energy storage, and charging microgrids and the distribution network described in this invention, the following steps are included: 1) Implementing operation monitoring and feedback updates based on intervention control signals to establish a closed-loop mechanism; 2) Correcting the situational awareness model and early warning thresholds based on monitoring feedback to achieve continuous optimization and proactive prevention and control; 3) Collecting state data of the photovoltaic, energy storage, and charging microgrids and the distribution network after intervention and analyzing the operational effects; 4) Correcting the situational awareness model parameters based on the operational results of the photovoltaic, energy storage, and charging microgrids and the distribution network; 5) Calculating the standard deviation and seasonal adjustment amount of historical data to dynamically adjust the early warning thresholds; 6) Adjusting the model learning parameters based on the loss function gradient and learning rate; 7) Putting the corrected model into the next cycle of operation to complete the closed-loop update.

[0015] This preferred solution analyzes the status data after intervention, combines historical data fluctuations and model loss information, and corrects and updates the learning parameters of the situational awareness model and early warning threshold. This forms a closed-loop optimization mechanism based on actual operating results, enabling continuous improvement of model accuracy and dynamic adjustment of the early warning response mechanism.

[0016] This invention provides a panoramic situational awareness and early warning system for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network.

[0017] To address the aforementioned technical problems, this invention provides the following technical solution: a panoramic situational awareness and early warning system for real-time interaction between a photovoltaic-storage-charging microgrid and a distribution network, comprising: a data acquisition and processing module, a data fusion module, a situational modeling and evaluation module, a risk prediction module, an intervention strategy module, and a monitoring and feedback module; the data acquisition and processing module is used to establish a data acquisition network covering the photovoltaic-storage-charging microgrid and the distribution network, acquire electrical quantities, environmental quantities, and equipment status information, and preprocess the data to obtain preprocessed multi-source data; the data fusion module is used to input the preprocessed multi-source data into an improved Kalman filter algorithm, and achieve high-reliability data fusion through state prediction, measurement update, and dynamic weight allocation to obtain fused state data; the situational modeling and evaluation module is used to construct the photovoltaic-storage-charging microgrid and distribution network based on the fused state data. The system generates a panoramic situational awareness vector, extracts operational features using a deep fusion network, and performs comprehensive identification and level assessment of the operational status to obtain a situational awareness assessment result. The risk prediction module inputs the situational awareness assessment result into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict future operational status and potential risk probabilities. It then uses a comprehensive evaluation method to classify risk levels and generate a risk assessment result. The intervention strategy module constructs a multi-objective optimization model based on the risk assessment result, combines it with reinforcement learning algorithms to generate adaptive intervention strategies, dynamically adjust the intervention intensity and optimize the strategy, and output intervention control signals. The monitoring and feedback module performs operational monitoring and feedback updates based on the intervention control signals, establishes a closed-loop mechanism, and corrects the situational awareness model and early warning thresholds based on monitoring feedback to achieve continuous optimization and proactive prevention and control.

[0018] The present invention provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network.

[0019] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage and charging microgrid and a distribution network are implemented.

[0020] The beneficial effects of this invention are as follows: The panoramic situational awareness technology enables comprehensive monitoring of the interaction between the photovoltaic-storage-charging microgrid and the distribution network. Through multi-source data fusion and deep learning algorithms, it significantly improves the accuracy and real-time performance of situational awareness. The system can accurately capture complex interaction patterns and potential risks, providing reliable support for operational decision-making.

[0021] The intelligent early warning mechanism enables a shift from passive defense to proactive early warning. Through predictive algorithms and risk assessment models, it can detect and warn of potential threats in advance. The tiered early warning system and dynamic threshold adjustment mechanism ensure the accuracy and adaptability of the warnings, significantly reducing false alarm and missed alarm rates.

[0022] The adaptive intervention strategy, based on multi-objective optimization and reinforcement learning techniques, can formulate the optimal risk control plan according to the actual situation. The intervention measures ensure system safety while also taking into account economy and reliability, achieving overall optimization of system operation.

[0023] This method has good scalability and adaptability, and can meet the needs of expanding the scale and upgrading the technology of photovoltaic-storage-charging microgrids. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 The above is a flowchart of a panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network, provided as an embodiment of the present invention.

[0026] Figure 2 The flowchart illustrates a panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network, as provided in one embodiment of the present invention. Detailed Implementation

[0027] To make the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0028] Example 1, referring to Figure 1 and Figure 2 This is one embodiment of the present invention, which provides a panoramic situational awareness and early warning method for real-time interaction between a photovoltaic-storage-charging microgrid and a distribution network, including: S1. Establish a data acquisition network covering the photovoltaic, energy storage, and charging microgrid and the distribution network to obtain electrical quantities, environmental quantities, and equipment status information, and preprocess the data to obtain preprocessed multi-source data.

[0029] S2. Input the multi-source preprocessed data into the improved Kalman filter algorithm, and achieve high-reliability data fusion through state prediction, measurement update and dynamic weight allocation to obtain the fused state data.

[0030] S3. Based on the fused state data, construct a panoramic situation vector of the photovoltaic-storage-charging microgrid and the distribution network. Utilize the deep fusion network to extract operational characteristics, complete the comprehensive identification and level assessment of the operational status, and obtain the situation assessment results.

[0031] S4. Input the situation assessment results into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict the future operating situation and the probability of potential risks. Then, use a comprehensive evaluation method to classify the risk level and generate risk assessment results.

[0032] S5. Construct a multi-objective optimization model based on the risk assessment results, combine it with reinforcement learning algorithms to generate adaptive intervention strategies, realize dynamic adjustment of intervention intensity and strategy optimization, and output intervention control signals.

[0033] S6. Based on intervention control signals, perform operation monitoring and feedback updates, establish a closed-loop mechanism, and correct the situational awareness model and early warning thresholds according to monitoring feedback to complete continuous optimization and proactive prevention and control.

[0034] It should be noted that with the widespread integration of distributed resources such as photovoltaics and electric vehicles, the operating status of photovoltaic-storage-charging microgrids and distribution networks is becoming increasingly complex, with problems such as frequent power fluctuations and uncertain operating conditions. Conventional static monitoring methods cannot achieve early risk prediction, resulting in untimely situation identification and delayed intervention. For example... Figure 1 and Figure 2 As shown.

[0035] Therefore, the present invention realizes a complete technical chain from data acquisition, fusion processing, situation modeling to risk prediction, strategy intervention, and monitoring feedback through the processing flow of S1–S6. Without relying on manual inference, the system can automatically generate early warning signals and complete strategy intervention, which significantly improves the stability and proactive prevention and control capabilities of the photovoltaic, energy storage, charging and power distribution system.

[0036] Example 2, an embodiment of the present invention, provides a panoramic situational awareness and early warning method for real-time interaction between a photovoltaic-storage-charging microgrid and a distribution network, based on the previous embodiment, including: Furthermore, in step S1, a data acquisition network covering the photovoltaic-storage-charging microgrid and the distribution network is established to acquire electrical quantities, environmental quantities, and equipment status information. The data is then preprocessed to obtain preprocessed multi-source data, including the following steps A1-A3: A1. Establish a data acquisition architecture for photovoltaic, energy storage, and charging microgrids and distribution networks, configure communication interfaces, and realize the acquisition of multi-source data.

[0037] A2. Perform time alignment and format conversion on the collected data to form data in a unified format.

[0038] A3. Preprocess the data in a uniform format to obtain preprocessed multi-source data.

[0039] In this embodiment of the application, the communication interface in step A1 can be configured in the data acquisition architecture as an interface module for connecting various monitoring devices, sensors and data acquisition terminals of the photovoltaic-storage-charging microgrid and the power distribution network. It supports data transmission between the data acquisition device and the edge processing node, has the ability to acquire synchronously and output in a unified format, and realizes the orderly aggregation of electrical quantities, environmental quantities and equipment status information.

[0040] In an alternative implementation, the communication interface can also be a data interface component based on the industrial Ethernet protocol, using Modbus TCP or IEC 61850 standards for communication, supporting bidirectional data exchange between the distributed data acquisition unit and the central processing platform, and possessing certain anti-interference capabilities and data verification mechanisms.

[0041] In another alternative implementation, the communication interface can also adopt a lightweight transmission channel built with wireless communication technology. Based on a 4G / NB-IoT communication module, the collected terminal data is transmitted to the cloud processing platform through a cellular network, which is suitable for scenarios where communication cables are difficult to deploy or devices are widely distributed.

[0042] This invention enables standardized access and stable transmission of multi-source data between the photovoltaic, energy storage, and charging microgrid and the power distribution network by configuring a unified communication interface, thus ensuring the continuity and integrity of data acquisition.

[0043] Specifically, a comprehensive data acquisition network covering the photovoltaic-storage-charging microgrid and distribution network is established to collect electrical quantities, environmental quantities, and equipment status information in real time. The collected data includes multi-dimensional information such as voltage and current at each node, power flow, equipment temperature, and environmental parameters. The data acquisition frequency is optimized according to the characteristics of different data types: electrical quantities are acquired at millisecond levels, environmental quantities at second levels, and equipment status at minute levels.

[0044] The collected raw data undergoes quality inspection and preprocessing, including outlier identification, missing data imputation, and noise filtering. Data quality evaluation criteria are determined using a combination of statistical methods and expert experience. Data standardization and normalization processes are established to ensure the comparability of data from different sources and types.

[0045] The panoramic situational awareness framework is constructed, and the situational vector of the optical-storage-charging microgrid is defined as follows: (1) in, Photovoltaic power generation capacity, For energy storage power, Charging power for electric vehicles, In the state of energy storage charge, Bus voltage For system frequency, This represents the power imbalance.

[0046] The distribution network situation vector is represented as: (2) in, For the first Node voltage amplitude, For the first Branch current, For the first Node active power For the first Node reactive power For ambient temperature, Relative humidity, These are the equipment operating status parameters.

[0047] The integrated situational awareness model employs a deep fusion network: (3) in, For the comprehensive situation vector, For fusion function, This is the weight matrix. This is the bias vector.

[0048] In this embodiment of the application, the fused state data in step S2 can be based on multi-source preprocessed data, using an improved Kalman filter algorithm, by constructing a state transition model and observation model of the photovoltaic-storage-charging microgrid and the distribution network, performing state prediction and measurement update operations, and performing fusion calculations based on the reliability of the data source by setting weight coefficients, and outputting fused state data for subsequent modeling and analysis.

[0049] In one alternative implementation, the fused state data can also be fed into a weighted average fusion algorithm to perform time-domain window weighting on different data channels, remove noise effects, and then be superimposed and fused to obtain a dataset that can be used for situation modeling.

[0050] In another alternative implementation, the fused state data can also be based on a particle filter algorithm to perform Bayesian estimation and sequence resampling fusion on the preprocessed data based on a known state space model, and output a stable sequence of state estimates.

[0051] This invention achieves unified fusion of multi-source data across time, reliability, and state variable dimensions through an improved Kalman filter, resulting in higher data accuracy and real-time performance of the subsequently constructed situation vector, thereby enhancing the effectiveness and reliability of comprehensive situation assessment and risk identification.

[0052] Furthermore, in step S2, the multi-source preprocessed data is input into the improved Kalman filter algorithm. High-reliability data fusion is achieved through state prediction, measurement update, and dynamic weight allocation to obtain the fused state data, including the following steps B1-B4: B1. Establish state transition and observation models for the photovoltaic-storage-charging microgrid and the distribution network, and determine the relationship between state variables and input variables.

[0053] B2. Perform state prediction on the preprocessed multi-source data to generate the predicted state of the photovoltaic-storage-charging microgrid and the distribution network.

[0054] B3. Based on the measurement update mechanism, the predicted state is corrected to obtain the updated state estimate of the photovoltaic-storage-charging microgrid and the distribution network.

[0055] B4. Perform fusion calculations based on the weight coefficients of the data source and output the fused status data.

[0056] Specifically, an improved Kalman filter algorithm is used to fuse multi-source data, improving data accuracy and reliability. A data fusion strategy is designed, allocating fusion weights based on the credibility and timeliness of different data sources. A data consistency verification mechanism is established to identify and handle conflicting data.

[0057] Construct a spatiotemporal data model, considering the temporal correlation and spatial distribution characteristics of the data. Design an adaptive fusion algorithm to dynamically adjust fusion parameters based on the system's operating status. Establish a fusion performance evaluation mechanism to continuously optimize the fusion strategy and parameter settings.

[0058] The multi-source data fusion algorithm, which uses an improved Kalman filter method for data fusion, is expressed as follows: (4) (5) in, This is the predicted state value at step k. Here is the state transition matrix. For the first Step state estimate, To control the input matrix, To control the input, For the prediction error covariance matrix, Let be the process noise covariance matrix.

[0059] The measurement update equation is: (6) (7) in, Here is the Kalman gain matrix. For the observation matrix, To measure the noise covariance matrix, For measured values, This is the state estimate for step k.

[0060] Furthermore, in step S3, a panoramic situational vector of the photovoltaic-storage-charging microgrid and the distribution network is constructed based on the fused state data. Operational characteristics are extracted using the deep fusion network to complete the comprehensive identification and level assessment of the operational status, obtaining the situational assessment results, including the following steps C1-C4: C1. Based on the merged state data, establish the state vector representation structure of the photovoltaic-storage-charging microgrid and the distribution network, and determine the characteristic parameters.

[0061] C2. Input the status vectors of the photovoltaic-storage-charging microgrid and the distribution network into the deep fusion network to extract operational characteristics.

[0062] C3. Establish identification and classification models based on operational characteristics to assess the operational status of photovoltaic-storage-charging microgrids and distribution networks.

[0063] C4. Output the situation assessment results.

[0064] In the embodiments of this application, the identification and classification model in step C3 can be a model structure constructed based on the operating characteristics of the photovoltaic-storage-charging microgrid and distribution network extracted by the deep fusion network. This model is used to classify and identify the situation vector. By setting classification rules and training parameters, the operating status is divided into different levels and categories, thereby realizing automatic identification and level output of the operating status.

[0065] In one alternative implementation, the recognition and classification model can also use a convolutional neural network structure to extract local patterns of operational features, classify and output feature vectors through a multilayer perceptron, construct training and validation sets for supervised learning, and achieve labeled recognition of operational status.

[0066] In another alternative implementation, the identification and classification model can also use a support vector machine to perform binary or multi-class classification training on the fused situation features, and map the input situation vector according to the boundary function of each category to achieve automatic differentiation between abnormal and normal situations.

[0067] This invention constructs an identification and classification model to structurally process and classify the operational characteristics of photovoltaic-storage-charging microgrids and distribution networks, thereby improving the accuracy and efficiency of situation assessment results and providing clear input for subsequent risk prediction and intervention decisions, and enhancing the system's state discrimination capability.

[0068] Specifically, a comprehensive situational awareness model is constructed based on the fused data to identify the current operational status and potential risks of the system. Machine learning algorithms are used to classify and identify situational patterns, establishing a situational feature database and knowledge base. A multi-level situational assessment system is designed to conduct comprehensive assessments at the device, system, and network levels.

[0069] Establish a dynamic situation assessment mechanism to update situation indicators and assessment results in real time. Design a situation visualization interface to intuitively display the system's operational status and risk distribution. Construct a situation trend analysis model to predict the development direction and evolution patterns of the situation.

[0070] The situation assessment indicator system defines security indicators as follows: (8) in, For comprehensive safety indicators, These are the weighting coefficients. For voltage safety indicators, For current safety indicators, For frequency security indicators.

[0071] The voltage safety index is calculated as follows: (9) in, For the number of nodes, For the first Node voltage, For reference voltage, This refers to the voltage deviation tolerance.

[0072] Economic indicators are expressed as follows: (10) in, As an economic indicator, As a weighting factor, For network loss costs, For operating costs, To cover maintenance costs.

[0073] In this embodiment of the application, the risk assessment results in step S4 can be sequentially input into the long short-term memory network model and the Bayesian network to predict the future operating situation and calculate the posterior probability of potential risks under evidence conditions, respectively. Then, by constructing an evaluation matrix and a preset weight vector, a fuzzy synthesis is performed to output the risk assessment results representing the risk level of the current operating state.

[0074] In an optional implementation, the risk assessment results can also be used to train and classify the situation vectors using a support vector machine model to obtain classification results representing high, medium, and low risk levels, and then matched and verified with historical operating conditions to form a quantitative risk assessment result.

[0075] In another alternative implementation, the risk assessment results can also be used to fit and predict the fused situational data using a multilayer perceptron neural network, extract the correlation between feature values ​​and failure modes, and output a risk level score by combining the analytic hierarchy process (AHP).

[0076] The risk assessment results generated by this invention through a hybrid prediction model combined with fuzzy evaluation methods can simultaneously reflect future trends and uncertainty risk levels, achieving a quantitative expression of operational risks and a clear classification of risk levels, thus providing a clear input basis for subsequent strategic decisions.

[0077] Furthermore, in step S4, the situation assessment results are input into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict future operational situations and potential risk probabilities. A comprehensive evaluation method is then used to classify risk levels and generate risk assessment results, including the following steps D1-D4: D1. Input the situation assessment results into the Long Short-Term Memory network model to predict future operational situation data.

[0078] D2. Input the operational status data into the Bayesian network and calculate the posterior probability of potential risks based on the evidence conditions.

[0079] D3. Construct an evaluation matrix based on the prediction results and perform a synthesis operation according to the preset weight vector.

[0080] D4. Output the synthesized result as a risk level vector.

[0081] Specifically, deep learning algorithms are used to construct a situation prediction model to predict the system's operational status and risk level in future periods. Multi-timescale prediction strategies are designed to cover short-term, medium-term, and long-term prediction needs. A prediction accuracy evaluation mechanism is established to continuously optimize the prediction model and parameters.

[0082] Construct an intelligent early warning system that automatically generates early warning information based on forecast results and risk assessments. Design a tiered early warning mechanism to determine the warning level based on risk level and urgency. Establish an early warning information dissemination process to ensure that relevant personnel receive early warning information in a timely manner.

[0083] Risk probability is calculated using a Bayesian network: (11) in, In the evidence Risk under conditions The posterior probability, Let be the likelihood function. This represents the prior probability.

[0084] The risk level classification is expressed using the fuzzy evaluation method as follows: (12) in, For risk level vectors, For the weight vector, For fuzzy evaluation matrix, This indicates a fuzzy synthesis operation.

[0085] The warning threshold is dynamically adjusted, and the warning threshold adopts an adaptive adjustment mechanism as follows: (13) in, For the first The warning threshold at any time, As the baseline threshold, To adjust the coefficient, The standard deviation of historical data. This is a seasonal adjustment.

[0086] The learning rule for threshold adjustment is: (14) in, For learning rate, This is the gradient of the loss function.

[0087] The situation prediction model uses a long short-term memory network for situation prediction. (15) (16) (17) in, For the Gate of Oblivion For input gate, Candidate cell state, , , This is the weight matrix. , , For bias vectors, This is the hidden state from the previous moment. This is the current input.

[0088] Cell status updated to: (18) in, The current cell state, This represents the cell state at the previous moment. This represents element-wise product.

[0089] The anomaly detection algorithm using the Isolation Forest algorithm is represented as follows: (19) in, For the sample Abnormal scores, For the sample Average path length in an isolated tree for Average path length of each sample This represents the number of samples.

[0090] The criteria for anomaly detection are: (20) In the formula, This is the threshold for anomaly detection.

[0091] The warning levels are classified into four levels: (twenty one) in, This is a comprehensive risk index.

[0092] The urgency level of the warning is calculated as follows: (twenty two) in, As an urgency index, These are the weighting coefficients. As the situation develops, To determine the degree of impact.

[0093] Furthermore, in step S5, a multi-objective optimization model is constructed based on the risk assessment results. An adaptive intervention strategy is generated using a reinforcement learning algorithm to achieve dynamic adjustment of the intervention intensity and strategy optimization, outputting intervention control signals. This includes the following steps E1-E5: E1. Construct a multi-objective function vector consisting of security objectives, economic objectives, and reliability objectives.

[0094] E2. Based on the risk assessment results, set control variables to form the initial solution of the intervention strategy.

[0095] E3. Combining reinforcement learning algorithms, state-action updates are performed based on the deviation signals between the photovoltaic-storage-charging microgrid and the distribution network.

[0096] E4. Adaptively adjust the intervention intensity based on the maximum intervention intensity and PID control parameters.

[0097] E5, Output intervention control signal.

[0098] Specifically, based on the early warning results, corresponding intervention strategies are formulated, including preventative control and emergency control measures. A multi-objective optimization framework is designed to balance economy and reliability while ensuring system safety. An intervention strategy library is established, with pre-set corresponding handling plans for different types of risks and abnormal situations.

[0099] Reinforcement learning algorithms are used to optimize intervention strategies, learning the optimal control scheme through interaction with the environment. An adaptive intervention mechanism is designed to dynamically adjust the intervention intensity and strategy based on actual results. An intervention effectiveness evaluation system is established to analyze the effectiveness of intervention measures and identify areas for improvement.

[0100] The intervention strategy, expressed using multi-objective optimization, is as follows: (twenty three) in, For a multi-objective function vector, To control variables, For security purposes, For economic purposes, For reliability objectives.

[0101] The adaptive adjustment of intervention intensity is expressed as: (twenty four) in, For the intensity of intervention, For maximum intervention intensity, , , For PID control parameters, This is a deviation signal.

[0102] Furthermore, in step S6, operational monitoring and feedback updates are performed based on intervention control signals to establish a closed-loop mechanism. The situational awareness model and early warning thresholds are corrected based on monitoring feedback to complete continuous optimization and proactive prevention and control, including the following steps F1-F5: F1. Collect status data of the photovoltaic-storage-charging microgrid and distribution network after the intervention is implemented, and analyze the operation effect.

[0103] F2. Adjust the parameters of the situational awareness model based on the operation results of the photovoltaic-storage-charging microgrid and the distribution network.

[0104] F3. Calculate the standard deviation and seasonal adjustment of historical data, and dynamically adjust the warning threshold.

[0105] F4. Adjust the model learning parameters based on the gradient of the loss function and the learning rate.

[0106] F5. Put the corrected model into the next cycle of operation to complete the closed-loop update.

[0107] Specifically, implement continuous system operation monitoring to track the effectiveness of intervention measures and changes in system status. Establish a closed-loop feedback mechanism to adjust the situational awareness model and early warning parameters based on monitoring results. Design emergency response procedures to quickly respond to emergencies and abnormal situations.

[0108] Establish a system performance evaluation mechanism to regularly analyze the operational effectiveness of the situational awareness and early warning system. Build a knowledge update mechanism to continuously improve system performance based on operational experience and new data. Ensure the system's scalability and adaptability to support the integration of new devices and scenarios.

[0109] Perception accuracy is evaluated using multiple metrics, as follows: (25) (26) (27) in, For accuracy, For accuracy, For recall rate, For a real example, For a true negative example, As a false positive example, This is a false negative.

[0110] The F1 score is calculated as follows: (28) Example 3 is an embodiment of the present invention, which provides a panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network. To verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0111] This invention was verified through simulation in MATLAB based on real-world operational data from a photovoltaic-storage-charging microgrid and a distribution network. A city distribution network was selected as the test platform. This network comprises three photovoltaic-storage-charging microgrids with a total installed capacity of 25.8 MW, an energy storage capacity of 12.5 MWh, and 200 charging piles. The simulation lasted for 30 consecutive days with a sampling interval of 1 second, resulting in 2,592,000 data samples. These are shown in Tables 1, 2, and 3.

[0112] Table 1 Performance Comparison of Different Situation Awareness Methods

[0113] Table 2 Analysis of Early Warning Effect

[0114] Table 3 Evaluation of the effectiveness of intervention strategies

[0115] Simulation results show that the panoramic situational awareness and early warning method for real-time interaction between the photovoltaic-storage-charging microgrid and the distribution network proposed in this invention significantly outperforms traditional methods in terms of detection accuracy, response time, and early warning effect. The system's detection accuracy reaches 97.8%, the false alarm rate is reduced to 2.1%, and the average response time is shortened to 2.5 seconds. The early warning system can detect potential risks 8.5-25.6 minutes in advance, providing operators with sufficient response time. After the intervention strategy is implemented, the average risk level of various types of risks is reduced by 45%-50%, effectively improving the system's safe operation level.

[0116] Example 4 is an embodiment of the present invention. This embodiment provides a panoramic situational awareness and early warning system for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network. The system includes a data acquisition and processing module, a data fusion module, a situational modeling and evaluation module, a risk prediction module, an intervention strategy module, and a monitoring and feedback module.

[0117] The data acquisition and processing module is used to establish a data acquisition network covering the photovoltaic, energy storage, and charging microgrids and the power distribution network, to acquire electrical quantities, environmental quantities, and equipment status information, and to preprocess the data to obtain preprocessed multi-source data.

[0118] The data fusion module is used to input multi-source preprocessed data into an improved Kalman filter algorithm. Through state prediction, measurement update and dynamic weight allocation, it achieves high-reliability data fusion to obtain fused state data.

[0119] The situation modeling and assessment module is used to construct a panoramic situation vector of the photovoltaic-storage-charging microgrid and distribution network based on the fused state data. It uses the deep fusion network to extract operating characteristics, completes the comprehensive identification and level assessment of the operating status, and obtains the situation assessment results.

[0120] The risk prediction module is used to input the situation assessment results into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict the future operating situation and the probability of potential risks. It also uses a comprehensive evaluation method to classify the risk level and generate risk assessment results.

[0121] The intervention strategy module is used to construct a multi-objective optimization model based on the risk assessment results, combine reinforcement learning algorithms to generate adaptive intervention strategies, realize dynamic adjustment of intervention intensity and strategy optimization, and output intervention control signals.

[0122] The monitoring and feedback module is used to perform operational monitoring and feedback updates based on intervention control signals, establish a closed-loop mechanism, and correct the situational awareness model and early warning thresholds based on monitoring feedback to achieve continuous optimization and proactive prevention and control.

[0123] This embodiment also provides an electronic device applicable to a panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as proposed in the above embodiment.

[0124] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network, as proposed in the above embodiment.

[0125] The storage medium proposed in this embodiment and the panoramic situational awareness and early warning method for realizing real-time interaction between a photovoltaic storage and charging microgrid and a distribution network proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0126] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0127] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network, characterized in that: include, Establish a data acquisition network covering photovoltaic, energy storage, and charging microgrids and distribution networks to acquire electrical quantities, environmental quantities, and equipment status information, and preprocess the data to obtain preprocessed multi-source data; The improved Kalman filter algorithm inputs multi-source preprocessed data and achieves high-reliability data fusion through state prediction, measurement update and dynamic weight allocation to obtain fused state data. Based on the fused state data, a panoramic situational vector of the photovoltaic-storage-charging microgrid and the distribution network is constructed. The operation characteristics are extracted using the deep fusion network to complete the comprehensive identification and level assessment of the operation status and obtain the situational assessment results. The situation assessment results are input into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict the future operational situation and the probability of potential risks. The risk level is classified through a comprehensive evaluation method to generate risk assessment results. A multi-objective optimization model is constructed based on the risk assessment results. An adaptive intervention strategy is generated by combining reinforcement learning algorithms to achieve dynamic adjustment of intervention intensity and strategy optimization, and output intervention control signals. Based on intervention control signals, the system performs operational monitoring and feedback updates, establishes a closed-loop mechanism, and corrects the situational awareness model and early warning thresholds based on monitoring feedback to achieve continuous optimization and proactive prevention and control.

2. The panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as described in claim 1, characterized in that: The process involves establishing a data acquisition network covering the photovoltaic-storage-charging microgrid and the distribution network to obtain electrical quantities, environmental quantities, and equipment status information. The data is then preprocessed to obtain preprocessed multi-source data, including... Establish a data acquisition architecture for photovoltaic, energy storage, and charging microgrids and distribution networks, configure communication interfaces, and realize the acquisition of multi-source data; The collected data is time-aligned and format-converted to form data in a uniform format. Data in a uniform format is preprocessed to obtain preprocessed multi-source data.

3. The panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as described in claim 2, characterized in that: The improved Kalman filter algorithm inputs multi-source preprocessed data and achieves high-reliability data fusion through state prediction, measurement update, and dynamic weight allocation to obtain fused state data, including: Establish state transition and observation models for photovoltaic-storage-charging microgrids and distribution networks, and determine the relationship between state variables and input variables; Perform state prediction on the preprocessed multi-source data to generate the predicted state of the photovoltaic-storage-charging microgrid and the distribution network; Based on the measurement update mechanism, the predicted state is corrected to obtain the updated state estimate of the photovoltaic-storage-charging microgrid and the distribution network; The system performs fusion calculations based on the weight coefficients of the data sources and outputs the fused state data.

4. The panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as described in claim 3, characterized in that: The panoramic situational awareness vector of the photovoltaic-storage-charging microgrid and distribution network is constructed based on the fused state data. Operational characteristics are extracted using a deep fusion network to complete the comprehensive identification and level assessment of the operational status, resulting in a situational awareness assessment, including: Based on the fused state data, establish the state vector representation structure of the photovoltaic-storage-charging microgrid and the distribution network, and determine the characteristic parameters; The status vectors of the photovoltaic-storage-charging microgrid and the distribution network are input into a deep fusion network to extract operational characteristics; Based on operational characteristics, an identification and classification model is established to assess the operational status of photovoltaic-storage-charging microgrids and distribution networks. Output the situation assessment results.

5. The panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as described in claim 4, characterized in that: The situation assessment results are input into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict future operational situations and potential risk probabilities. A comprehensive evaluation method is then used to classify risk levels and generate risk assessment results, including... The situation assessment results are input into a long short-term memory network model to predict future operational situation data. The operational status data is input into a Bayesian network, and the posterior probability of potential risks is calculated based on the evidence conditions. An evaluation matrix is ​​constructed based on the prediction results, and a synthesis operation is performed according to the preset weight vector. The output of the composite result is used as a risk level vector.

6. The panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as described in claim 4, characterized in that: The process involves constructing a multi-objective optimization model based on risk assessment results, combining it with reinforcement learning algorithms to generate adaptive intervention strategies, dynamically adjusting intervention intensity and optimizing strategies, and outputting intervention control signals, including... Construct a multi-objective function vector consisting of safety, economic, and reliability objectives; Based on the risk assessment results, control variables are set to form the initial solution of the intervention strategy; By combining reinforcement learning algorithms, state-action updates are performed based on the deviation signals between the photovoltaic-storage-charging microgrid and the distribution network; The intervention intensity is adaptively adjusted based on the maximum intervention intensity and the PID control parameters; Output intervention control signals.

7. The panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as described in claim 4, characterized in that: The process of performing operational monitoring and feedback updates based on intervention control signals establishes a closed-loop mechanism. This mechanism, along with adjustments to the situational awareness model and early warning thresholds based on monitoring feedback, enables continuous optimization and proactive prevention and control. Collect status data of the photovoltaic-storage-charging microgrid and distribution network after the intervention is implemented, and analyze the operational effect; The parameters of the situational awareness model are adjusted based on the operation results of the photovoltaic-storage-charging microgrid and the distribution network. Calculate the standard deviation and seasonal adjustment of historical data, and dynamically adjust the early warning threshold; Adjust the model learning parameters based on the gradient of the loss function and the learning rate; The revised model is then put into operation in the next cycle to complete the closed-loop update.

8. A panoramic situational awareness and early warning system for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network, comprising the panoramic situational awareness and early warning method for real-time interaction between a photovoltaic, energy storage, and charging microgrid and a distribution network as described in any one of claims 1 to 7, characterized in that, include: The system includes a data acquisition and processing module, a data fusion module, a situation modeling and assessment module, a risk prediction module, an intervention strategy module, and a monitoring and feedback module. The data acquisition and processing module is used to establish a data acquisition network covering the photovoltaic, energy storage, and charging microgrid and the power distribution network, acquire electrical quantities, environmental quantities, and equipment status information, and preprocess the data to obtain preprocessed multi-source data; The data fusion module is used to input multi-source preprocessed data into an improved Kalman filter algorithm, and achieves high-reliability data fusion through state prediction, measurement update and dynamic weight allocation to obtain fused state data. The situation modeling and assessment module is used to construct a panoramic situation vector of the photovoltaic-storage-charging microgrid and distribution network based on the fused state data, extract operating features using the deep fusion network, complete the comprehensive identification and level assessment of the operating status, and obtain the situation assessment results. The risk prediction module is used to input the situation assessment results into a hybrid prediction model composed of a long short-term memory network and a Bayesian network to predict the future operating situation and the probability of potential risks, and to classify the risk level through a comprehensive evaluation method to generate risk assessment results. The intervention strategy module is used to construct a multi-objective optimization model based on the risk assessment results, combine it with reinforcement learning algorithms to generate adaptive intervention strategies, realize dynamic adjustment of intervention intensity and strategy optimization, and output intervention control signals. The monitoring and feedback module is used to perform operation monitoring and feedback updates based on intervention control signals, establish a closed-loop mechanism, and correct the situational awareness model and early warning thresholds based on monitoring feedback to complete continuous optimization and proactive prevention and control.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the panoramic situational awareness and early warning method for real-time interaction between the photovoltaic, energy storage, and charging microgrid and the distribution network, as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the panoramic situational awareness and early warning method for real-time interaction between the photovoltaic, energy storage, and charging microgrid and the distribution network, as described in any one of claims 1 to 7.

Images