Power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm

Abstract

The invention provides a power grid multi-source data fusion investment benefit post-evaluation platform based on a quantum hybrid algorithm, and belongs to the technical field of power grid investment evaluation. The core defect of'simple combination and lack of deep improvement 'existing in the combination of quantum and a classical algorithm in the prior art is overcome; the platform adopts a full-flow closed-loop architecture of quantum enhanced hybrid algorithm driving and three-dimensional data deep fusion, and the whole architecture is divided into five core levels, namely a quantization data preprocessing layer, a quantum-space-time diagram fusion layer, a quantum-reinforcement learning evaluation optimization layer, a quantum causal traceability feedback layer and a visual output layer from bottom to top in sequence; according to the method, a set of multi-algorithm collaborative quantum enhancement hybrid algorithm system is designed for the actual demand of fusion evaluation of three types of core data of power grid operation, equipment state and financial investment, a combination mode of quantum and a classical algorithm is reconstructed from underlying logic, and power grid investment benefit post-evaluation can be perfectly adapted.

Description

Technical Field

[0001] This application relates to the field of power grid investment assessment technology, and in particular to a power grid multi-source data fusion post-assessment platform based on quantum hybrid algorithms for investment benefit evaluation. Background Technology

[0002] Currently, post-investment benefit evaluation of power grids mainly relies on classical single algorithms to process multi-source heterogeneous data. This presents significant bottlenecks in terms of data processing computing power and evaluation result accuracy, making it difficult to adapt to the evaluation needs of massive multi-source data in power grids. Although intelligent algorithms such as spatiotemporal graph neural networks and reinforcement learning have been developed in the field of multi-source data fusion, demonstrating advantages in feature extraction and data fusion, they have not yet broken through the performance limits of classical algorithms by combining quantum computing technology. While quantum computing technology has achieved preliminary applications of hybrid algorithms in energy finance sub-fields such as power dispatch optimization and financial portfolio optimization, verifying the feasibility of quantum computing in this field, a complete technical solution that deeply couples "quantum computing + multiple types of intelligent algorithms" and is suitable for comprehensive evaluation of power grid investment benefits has not yet been formed.

[0003] Specifically, the existing technology has the following drawbacks:

[0004] 1. Lack of deep integration in hybrid algorithms: In current technologies, the combination of quantum and classical algorithms is only a superficial application of the "quantum encoding + traditional algorithm" model. The algorithms are not reconstructed and optimized at the core logic level, and the natural advantages of quantum computing in parallel operation and high-dimensional space processing cannot be fully utilized. The actual effect of algorithm integration is limited.

[0005] 2. Insufficient capture of spatiotemporal correlation: Power grid operation data has significant spatiotemporal distribution characteristics, equipment status data has complex topological correlation characteristics, and financial data shows obvious time-series change characteristics. The multidimensional correlation of these three types of data is extremely complex. Traditional classical algorithms are difficult to accurately model and effectively capture such complex multidimensional correlations, which can easily lead to distortion of fusion features and affect the accuracy of subsequent evaluation results.

[0006] 3. Weak dynamic optimization capability: Most existing power grid investment benefit post-evaluation technologies can only complete the static output of evaluation results. They lack dynamic feedback and optimization adjustment mechanisms based on real-time power grid operation data, and cannot adapt to the dynamic changes in power grid operation status and investment benefits in real time. The timeliness and practicality of the evaluation results are insufficient, making it difficult to provide dynamic guidance for investment decisions.

[0007] 4. Imbalance between robustness and interpretability: When faced with external interference factors such as noise generated during the power grid data acquisition process and fluctuations in new energy output, the existing data fusion algorithm is not robust enough and is prone to bias in the evaluation results. At the same time, the quantum algorithm itself has certain "black box characteristics". The hybrid algorithm after combination further exacerbates the problem of interpretability of the evaluation results, making it difficult to clarify the causes and influencing factors of the evaluation results.

[0008] 5. Low efficiency of multi-objective optimization: The evaluation of power grid investment benefits involves multiple core objectives such as economic benefits, technical reliability, and social environmental protection. It is a typical multi-objective optimization problem. When solving such multi-objective optimization problems, traditional quantum algorithms are prone to getting trapped in local optima and are difficult to obtain global optima. Moreover, the computational complexity of the algorithms is high and the solution time is too long, which cannot meet the high efficiency evaluation requirements of actual engineering. Summary of the Invention

[0009] To overcome the core shortcomings of existing technologies that combine quantum and classical algorithms, namely "simple combination and lack of in-depth improvement," and to meet the practical needs of fusion evaluation of three core data types—grid operation, equipment status, and financial investment—this application proposes a power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithms. It designs a multi-algorithm collaborative quantum-enhanced hybrid algorithm system, reconstructing the combination of quantum and classical algorithms from the underlying logic, which can perfectly adapt to the post-evaluation of power grid investment benefits.

[0010] The technical solution adopted in this application is: a power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithms, comprising:

[0011] Quantum data preprocessing layer: Based on quantum principal component analysis (QPCA) dimensionality reduction, quantum generative adversarial network (Q-GAN) is introduced to achieve quantum-principal component adversarial dimensionality reduction, and output robust dimensionality reduction data after anti-interference processing;

[0012] Quantum-Spatiotemporal Graph Fusion Layer: By constructing a power grid quantum knowledge graph GQ-KG, quantum entanglement is achieved on multi-source data of the power grid, realizing semantic association and knowledge fusion of multi-source data of the power grid. The convolution kernel of the spatiotemporal graph neural network ST-GNN is reconstructed by introducing quantum graph convolution operation to obtain quantum-enhanced spatiotemporal graph neural network QE-ST-GNN. The quantum features of the power grid quantum knowledge graph GQ-KG are converted into classical fusion features through quantum-enhanced spatiotemporal graph neural network QE-ST-GNN.

[0013] Quantum-Reinforcement Learning Evaluation and Optimization Layer: This layer constructs a dynamic closed loop of "evaluation-feedback-optimization" for power grid investment benefits through the quantum reinforcement learning model QE-DRL, realizing efficient solution of multi-objective optimization and dynamic adjustment of evaluation strategy;

[0014] Quantum Causal Source Tracing Feedback Layer: The quantum causal adversarial fusion method (Q-CAF) is obtained by deeply integrating quantum correlation analysis, causal inference and adversarial fusion technology, which can achieve accurate location and quantitative attribution of assessment anomalies;

[0015] Visualization output layer: used to realize the classical conversion and multi-dimensional visualization of quantum evaluation results.

[0016] Furthermore, the implementation steps of the quantized data preprocessing layer are as follows:

[0017] Data integration: Standardize and integrate three core data categories—power grid operation data, equipment status data, and financial data—to construct a three-dimensional data matrix. , where R is the real number field, and n represents the total number of samples / number of data records of the multi-source data of the power grid;

[0018] Quantum encoding: Using angle / phase encoding, the integrated three-dimensional data is mapped to quantum states.

[0019] The angular parameters of the quantum state are jointly determined by the normalized values ​​of the three-dimensional data, thereby achieving efficient conversion from classical data to quantum state.

[0020] Counter-dimensionality reduction: Extracting principal component quantum states from quantum states using quantum principal component analysis (QPCA) The principal component quantum state is input into the quantum generative adversarial network (Q-GAN) for feature defense training, and finally outputs robust dimensionality reduction data after anti-interference processing.

[0021] Furthermore, the quantum generative adversarial network (Q-GAN) includes a generator and a discriminator. The generator uses the characteristics of quantum superposition to simulate the distribution features of power grid data and generate power grid data samples with strong anti-interference capabilities. The discriminator is constructed using a quantum neural network (QNN) and leverages the parallel advantages of quantum computing to accurately distinguish between real power grid data and noisy interference data.

[0022] Furthermore, the construction process of the power grid quantum knowledge graph GQ-KG is as follows:

[0023] First, the power grid operation data, equipment status data, and financial data are mapped to three types of core nodes in the power grid quantum knowledge graph GQ-KG, where the power grid operation data corresponds to the power grid topology node, the equipment status data corresponds to the power equipment node, and the financial data corresponds to the financial node.

[0024] Furthermore, by leveraging the strong correlation characteristics of quantum entanglement, attribute associations can be established between different types of nodes, thereby achieving semantic association and knowledge fusion of multi-source data.

[0025] Furthermore, the implementation steps of the quantum-spacetime graph fusion layer are as follows:

[0026] Spatial dimension feature extraction: A quantum computing model is constructed using quantum gate circuits to calculate the quantum entanglement strength between any two nodes in the power grid quantum knowledge graph GQ-KG, and this quantum entanglement strength is used as the spatial correlation weight between nodes to achieve accurate quantification of the topological correlation features of power grid data;

[0027] Temporal feature fusion: The temporal features of power grid data are mapped to the phase parameters of quantum states using quantum phase encoding. Then, the phase features of the temporal dimension are deeply fused with the entangled weight features of the spatial dimension through inverse quantum Fourier transform (IQFT) to achieve integrated extraction of spatiotemporal features.

[0028] Fusion feature output: The quantum state evolution calculation of the quantum knowledge graph of the power grid is performed by quantum graph convolution unitary transformation, and then the evolved quantum state is converted into classical fusion feature data through quantum measurement operation, realizing efficient conversion from quantum features to classical features.

[0029] Furthermore, the implementation steps of the quantum-reinforcement learning evaluation optimization layer are as follows:

[0030] Quantum state space construction: The fusion characteristics output by the quantum-spacetime graph fusion layer and the multi-objective evaluation index of power grid investment benefits are jointly quantum-encoded and converted into quantum superposition states to construct a high-dimensional quantum state space.

[0031] ;

[0032] in, For a high-dimensional quantum state space, For quantum amplitude, To evaluate the state basis vectors, k is the index of the evaluation state basis vector in the high-dimensional quantum state space, and m is the total number of evaluation state basis vectors in the high-dimensional quantum state space.

[0033] Quantum reinforcement policy optimization: The quantum approximation optimization algorithm QAOA is used to reconstruct the policy network of deep reinforcement learning to obtain the quantum reinforcement learning model QE-DRL. By leveraging the advantages of quantum parallel computing, the optimal evaluation policy can be quickly searched and optimized.

[0034] Furthermore, the implementation steps for quantum strategy optimization are as follows:

[0035] Multi-objective reward function design: Construct a weighted linear reward function, with financial return rate, equipment availability rate, and carbon emission reduction as the core evaluation indicators for economic, technological, and social objectives, respectively. The weights of each evaluation indicator are determined through a quantum game equilibrium algorithm to achieve a scientific allocation of multi-objective weights.

[0036] Quantum-enhanced policy update: The parameters of the reconstructed policy network are iteratively optimized using a quantum variational algorithm. Leveraging the advantages of quantum parallel computing, multiple policy directions are searched simultaneously, significantly improving the efficiency of policy optimization and achieving synergistic optimization of economic, technological, and social objectives.

[0037] Dynamic feedback optimization: A quantum-classical data interface is built to feed the real-time evaluation results back to the policy network through this data interface. At the same time, the real-time operation data of the power grid is connected. The policy network automatically and dynamically adjusts the evaluation parameters and policies based on the evaluation results and the real-time operation data of the power grid, forming a complete dynamic closed loop of "evaluation-feedback-optimization".

[0038] Furthermore, the implementation steps of the quantum causal attribution feedback layer are as follows:

[0039] Quantum causal graph construction: Based on the power grid quantum knowledge graph GQ-KG, by calculating the quantum entanglement strength between nodes, key causal links that have a significant impact on investment benefit assessment results are screened out. Based on these key causal links, a quantum causal graph of "data characteristics-assessment indicators-benefit results" is constructed. The nodes of this quantum causal graph are quantized power grid data characteristics and assessment indicators, and the edges are the quantum causal correlation strength between nodes.

[0040] Quantum-enhanced causal inference: The logic of causal inference is reconstructed using quantum Bayesian networks (QBNs). Leveraging the advantages of quantum parallel computing, the search and analysis process of causal links is accelerated, enabling precise location and quantitative attribution of anomalies.

[0041] Optimize feedback output: Based on the anomaly location results and quantitative attribution data obtained from quantum causal inference, combined with the actual operation of the power grid, generate targeted investment optimization suggestions, and push the optimization suggestions to the power grid investment decision-making department to complete the entire closed loop of assessment-source tracing-optimization.

[0042] Furthermore, the specific steps for implementing quantum-enhanced causal inference are as follows:

[0043] Anomaly localization: When the investment benefit assessment result deviates from the preset threshold, the quantum backpropagation algorithm is activated. Starting from the abnormal node in the assessment result, the algorithm traces back along the causal link of the quantum causal graph to accurately locate the core cause node that caused the assessment anomaly and form a complete abnormal causal path.

[0044] Quantitative attribution: Through quantum measurement operations, the inner product modulus square of the quantum state and the anomalous quantum state of each causal node is calculated and used as the contribution of each node to the assessment of the anomaly, thereby achieving accurate quantification of the cause of the anomaly.

[0045] Furthermore, the visualization output layer can also be adapted to the interface of the power grid system.

[0046] The advantages of this application over the prior art are as follows:

[0047] 1. Breakthrough in Algorithm Integration Depth: Addressing the shortcomings of "superficial integration of hybrid algorithms," this application comprehensively reconstructs the core logic of the three core algorithms, QE-ST-GNN, QE-DRL, and Q-CAF, deeply embedding the technological advantages of quantum computing into the entire process of classical intelligent algorithms, including underlying computation, feature extraction, and decision optimization. This achieves deep coupling between quantum computing and intelligent algorithms, rather than simply replacing the input / output layers of the algorithms, fundamentally breaking through the limitations of superficial integration of quantum and classical algorithms.

[0048] 2. Significantly Improved Accuracy of Integrated Assessment: Addressing the deficiency of insufficient spatiotemporal correlation capture, this application constructs a power grid quantum knowledge graph (GQ-KG) to achieve semantic correlation and topological modeling of multi-source power grid data. Then, through quantum graph convolution operations, it achieves deep fusion of data features from both spatial and temporal dimensions, accurately modeling the spatiotemporal-topological-temporal multidimensional correlations of power grid operation data, equipment status data, and financial data. This significantly improves the effectiveness of the integrated features and achieves an extremely high accuracy rate, greatly reducing the error of power grid investment benefit assessment results and significantly improving the assessment accuracy compared to traditional assessment techniques.

[0049] 3. Significantly Improved Dynamic Optimization Efficiency: Addressing the deficiency of "weak dynamic optimization capability," this application constructs a complete "evaluation-feedback-optimization" dynamic closed loop using the QE-DRL model. Real-time grid operation data is integrated into the QE-DRL model, enabling dynamic adjustment of evaluation strategies and parameters. This allows the evaluation results to adapt in real-time to the dynamic changes in grid operation status and investment benefits, significantly improving the dynamic optimization capability of the technical solution. Furthermore, the dynamic evaluation closed loop constructed in this application achieves highly efficient solutions for multi-objective optimization, drastically reducing the solution time from several hours in traditional techniques to just over ten minutes. Simultaneously, it keeps the dynamic feedback delay at an extremely low level, meeting the practical needs of real-time evaluation and adjustment of grid investment benefits.

[0050] 4. Achieving a balance between robustness and interpretability: Addressing the deficiency of "imbalance between robustness and interpretability", this application improves the algorithm's anti-interference ability and robustness through quantum-adversarial training joint preprocessing at the quantum data preprocessing layer, achieves efficient suppression of data noise through quantum-adversarial training, and achieves accurate positioning and quantitative attribution of evaluation results through quantum causal inference at the quantum causal tracing feedback layer. This allows the technical solution to have high robustness and extremely high interpretability, while simultaneously meeting the reliability requirements of power engineering and the interpretability requirements of investment decisions, thus achieving a balance between the two.

[0051] 5. High practicality and adaptability: This application achieves seamless integration with various existing power grid systems through standardized API interfaces, without the need for large-scale modifications to existing power grid systems. It has low deployment costs and is easy to implement, and can be directly adapted to the post-evaluation scenarios of various implemented power grid investment projects, demonstrating strong engineering practicality and scenario adaptability.

[0052] 6. High optimization efficiency: To address the shortcoming of "low efficiency in multi-objective optimization", this application utilizes the inherent advantages of quantum parallel computing to accelerate the optimization process of the QAOA-DRL policy network, significantly improving the solution efficiency of multi-objective optimization problems, greatly compressing the multi-objective solution time, and achieving an extremely high optimal solution coverage rate, effectively solving the problems of low efficiency and poor solution quality in existing multi-objective optimization technologies. Attached Figure Description

[0053] The following description, in conjunction with the accompanying drawings, further illustrates this application:

[0054] Figure 1 This is a schematic diagram of the platform hierarchy provided in an embodiment of this application. Detailed Implementation

[0055] like Figure 1 As shown, this application provides a power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm. It adopts a full-process closed-loop architecture of "quantum-enhanced hybrid algorithm driving + three-dimensional data deep fusion". The overall architecture is divided into five core layers from bottom to top: quantum data preprocessing layer, quantum-spacetime graph fusion layer, quantum-reinforcement learning evaluation and optimization layer, quantum causal traceability feedback layer, and visualization output layer.

[0056] The technical terms used in this application will be explained below:

[0057] 1. Quantum Enhanced Spatiotemporal Graph Neural Network (QE-ST-GNN): This quantum-classical hybrid algorithm combines the parallel computing advantages of quantum computing with the topological modeling capabilities of spatiotemporal graph neural networks. It is specifically designed to accurately capture the spatiotemporal correlation characteristics of multi-source data from power grids and can achieve in-depth mining and correlation analysis of the spatiotemporal dimension characteristics of power grid data.

[0058] 2. Quantum-Deep Reinforcement Learning Coupled Model (QE-DRL): This model effectively expands the state space of traditional deep reinforcement learning through quantum state encoding, while combining the dynamic decision-making and autonomous learning characteristics of deep reinforcement learning to create a quantum-classical hybrid model with accurate benefit evaluation and real-time feedback optimization capabilities.

[0059] 3. Quantum Causal Adversarial Fusion (Q-CAF): An innovative method that combines the strong correlation characteristics of quantum entanglement, the logical analysis capability of causal inference, and the anti-interference characteristics of adversarial training to improve the robustness and accuracy of heterogeneous data fusion in the power grid from the underlying logic of data fusion.

[0060] 4. Power Grid Quantum Knowledge Graph (GQ-KG): Based on quantum node encoding technology and quantum graph convolution operation method, it can deeply associate various types of information such as power grid operation data, equipment status data, and financial data, and realize a professional semantic graph for multi-source data semantic association and knowledge mining.

[0061] The quantum-enhanced hybrid algorithm, proposed in this application, is a cross-level collaborative algorithm system that deeply couples quantum computing and classical intelligent algorithms for post-investment benefit evaluation of multi-source data fusion in power grids. This algorithm system uses three core algorithms—QE-ST-GNN, QE-DRL, and Q-CAF—as its framework, integrating quantum algorithms such as Q-GAN and QPCA. It deeply embeds the technological advantages of quantum computing into the entire process of classical intelligent algorithms, achieving deep coupling of their core logic. The four layers—quantum data preprocessing layer, quantum-spacetime graph fusion layer, quantum-reinforcement learning evaluation and optimization layer, and quantum causal tracing feedback layer—serve as the implementation carriers of this algorithm system in the engineering platform, broken down according to the business logic of "data preprocessing - feature fusion - benefit evaluation - causal tracing." Each layer carries the operational logic of different modules of the algorithm system. The algorithm operations of the four layers together constitute its entire operational logic, becoming the core driver of the platform's closed-loop architecture, ultimately achieving accurate, dynamic, and efficient evaluation of power grid investment benefits.

[0062] As the core driver of the platform's closed-loop architecture, this algorithm system is implemented across four technical levels: quantum data preprocessing, quantum-spatiotemporal graph fusion, quantum-reinforcement learning evaluation and optimization, and quantum causal tracing feedback. It can accurately capture the spatiotemporal-topological-temporal multidimensional correlations of power grid operation data, equipment status data, and financial data, efficiently solve multi-objective optimization problems involving economic, technical, social, and environmental goals, and achieve real-time adjustment of evaluation strategies through dynamic closed-loop. Ultimately, it outputs power grid investment benefit evaluation results that are accurate, timely, and interpretable, providing scientific support for investment decisions.

[0063] The functions and principles of each level are explained in detail below.

[0064] Quantum data preprocessing layer: used to implement joint preprocessing for quantum-adversarial training.

[0065] This layer is designed to address the core deficiency of "insufficient robustness" in existing technologies. Its core innovation lies in the deep integration and synergistic application of quantum dimensionality reduction technology and adversarial training mechanisms, rather than using the two technologies separately. By combining the underlying logic, it enhances the anti-interference capability and feature retention effect of data preprocessing, laying a high-quality data foundation for subsequent data fusion and evaluation.

[0066] The main functions of this level are as follows:

[0067] 1. Quantum-Enhanced Adversarial Noise Reduction: A quantum generative adversarial network (Q-GAN) specifically adapted to multi-source power grid data is constructed. The generator in this network simulates the distribution characteristics of power grid data through the properties of quantum superposition, generating power grid data samples with strong anti-interference capabilities. The discriminator is constructed using a quantum neural network (QNN), leveraging the parallel advantages of quantum computing to accurately distinguish real power grid data from noisy interference data. Quantum parallel computing technology significantly accelerates the iterative process of adversarial training, resulting in a substantial improvement in noise reduction efficiency compared to traditional generative adversarial networks. Simultaneously, it achieves efficient suppression of noise in power grid data, preserving the core features of the data to the greatest extent possible.

[0068] The principle behind using a quantum neural network (QNN) as a discriminator to accurately distinguish between real power grid data and noisy interference data is as follows:

[0069] The essence of the QNN discriminator is to perform "feature matching" in a quantum way: the quantum state of real power grid data is used as a "standard template". The quantum state of the input data is compared with the standard template through quantum parallel operation. If the matching degree is high, it is real data; if the matching degree is low, it is noisy data. Finally, the quantum-level discrimination result is transformed into a classical judgment conclusion through quantum measurement.

[0070] The discriminator built from a quantum neural network (QNN) within the quantum generative adversarial network (Q-GAN) of this platform leverages the parallel advantages of quantum computing and quantum state feature encoding and pattern recognition to accurately distinguish between real power grid data and noisy interference data. The core process involves first mapping the data to be discriminated against into quantum states through angle / phase encoding. Real power grid data exhibits regular distribution patterns and stable superposition weights, while noisy data displays distorted features and disordered weights, creating distinguishable quantum state differences that form the foundation for discrimination. The QNN discriminator constructs quantum circuits using quantum logic gates such as CNOT, RY, and Hadamard. Utilizing the advantages of quantum hyperparallelism and high-dimensional space processing, it performs parallel extraction and pattern matching of multi-dimensional core features and noise features from the input quantum states. Furthermore, this QNN discriminator features customized quantum circuit optimizations for three types of data characteristics: power grid operation, equipment status, and financial data. This enhances the ability to identify noise in power grid scenarios. Subsequently, through quantum measurement operations, the extracted quantum features are converted into classical probability values ​​between 0 and 1. These probability values ​​represent the degree of matching between the input data and the features of real power grid data. Binary discrimination is performed according to a preset probability threshold. Data with a matching degree higher than the threshold is considered real power grid data, while data with a matching degree lower than the threshold is considered data containing noise interference. At the same time, the QNN discriminator and the quantum generator of Q-GAN form an adversarial training closed loop. The quantum generator continuously generates anti-interference fake data that closely resembles real data and inputs it into the QNN discriminator. If the QNN discriminator makes a misjudgment, it will iteratively optimize the quantum circuit gate parameters through quantum backpropagation algorithm. After multiple rounds of adversarial training, a stable discriminant model is formed, further improving the discrimination accuracy. Moreover, quantum parallel computing greatly improves the iterative efficiency of adversarial training, allowing the QNN discriminator to retain the core features of the data to the greatest extent while efficiently suppressing power grid data noise.

[0071] 2. Quantum-Principal Component Adversarial Dimensionality Reduction: Building upon Quantum Principal Component Analysis (QPCA) dimensionality reduction technology, an adversarial training mechanism is innovatively introduced, forming a dual processing mode of "quantum feature extraction + adversarial feature defense." First, core features are extracted from multi-source power grid data using Quantum Fourier Transform (QFT). Then, an adversarial network is used to train defenses against the extracted core features, resisting feature distortion and interference. The specific implementation steps are as follows:

[0072] (1) Data integration: Standardize and integrate three types of core data, namely power grid operation data (such as PMU synchronization phasor data), equipment status data (such as insulation resistance and equipment loss rate data), and financial data (such as investment cost and operation and maintenance cost data), to construct a three-dimensional data matrix. , where R is the real number field, indicating that all elements in the data matrix are real numbers, corresponding to the standardized real values ​​of the three core data types: power grid operation, equipment status, and financial data. n represents the total number of samples from multiple power grid data sources / the number of data records, that is, the total number of valid data samples obtained after standardizing and integrating the three types of data: power grid operation data, equipment status data, and financial data. It is a core dimension parameter characterizing the data scale.

[0073] (2) Quantum encoding: Using angle encoding, the integrated three-dimensional data is mapped to quantum states:

[0074] The angular parameters of the quantum state are jointly determined by the normalized values ​​of the three-dimensional data, thereby achieving efficient conversion from classical data to quantum state.

[0075] (3) Counter-dimensionality reduction: Extract principal component quantum states from quantum states through quantum principal component analysis (QPCA). The principal component quantum state is input into the adversarial network for feature defense training, and finally outputs robust dimensionality reduction data after anti-interference processing. While reducing the data dimension, the core information of the original data is preserved to the greatest extent, and the anti-interference ability of the data features is greatly improved.

[0076] Quantum-Spacetime Graph Fusion Layer: Achieves 3D data fusion through QE-ST-GNN.

[0077] This layer addresses two core shortcomings of existing technologies: insufficient capture of spatiotemporal correlations and superficial integration of algorithms. Its core innovation lies in reconstructing the core logic of spatiotemporal graph neural networks, deeply embedding the technological advantages of quantum computing into the entire process of topology modeling, feature extraction, and spatiotemporal fusion of spatiotemporal graph neural networks, rather than simply applying quantum technology to the input layer. The QE-ST-GNN algorithm enables accurate capture and deep fusion of spatiotemporal, topological, and temporal correlation features of power grid 3D data.

[0078] The main functions of this level are as follows:

[0079] 1. Construction of the Power Grid Quantum Knowledge Graph: First, power grid operation data, equipment status data, and financial data are mapped to three types of core nodes in the power grid quantum knowledge graph (GQ-KG). Power grid operation data corresponds to power grid topology nodes such as buses and lines, equipment status data corresponds to power equipment nodes such as transformers and switches, and financial data corresponds to financial nodes such as investment projects and cost items. Then, through the strong correlation characteristics of quantum entanglement, attribute associations between different types of nodes are established. For example, "line load rate" is mapped to "equipment loss rate" and "operation and maintenance cost" through quantum entanglement, realizing semantic association and knowledge fusion of multi-source data, and providing a structured knowledge carrier for subsequent spatiotemporal graph convolution.

[0080] 2. Quantum-Enhanced Spatiotemporal Graph Convolution: This method comprehensively reconstructs the convolution kernel of the traditional Spatiotemporal Graph Neural Network (ST-GNN) and innovatively introduces quantum graph convolution operations. It achieves deep feature extraction and fusion from both spatial and temporal dimensions, and finally outputs the fused features through quantum measurement. The specific implementation steps are as follows:

[0081] (1) Spatial Dimension Feature Extraction: A quantum computing model is constructed using quantum gate circuits (CNOT+RY gates) to calculate the quantum entanglement strength between any two nodes in the power grid quantum knowledge graph. This quantum entanglement strength is then used as the spatial correlation weight between nodes, achieving accurate quantification of the topological correlation features of the power grid data. The formula is as follows:

[0082] ;

[0083] in, The quantum spatial correlation weight between node i and node j is used to quantitatively characterize the topological correlation strength between two nodes in the quantum knowledge graph of the power grid, providing a quantitative basis for spatial dimension feature extraction. Trace operation is a linear algebra operation that sums the diagonal elements of a matrix. It is used to numerically solve the quantum state density matrix and realize the classical quantization transformation of quantum state characteristics. Let be the quantum state density matrix of nodes i and j, describing the superposition and entanglement properties of the quantum states of the two nodes. This matrix is ​​the core matrix for calculating the quantum entanglement strength and spatial correlation weight. The quantum spatial correlation weight is obtained by calculating the von Neumann entropy of the node quantum state density matrix.

[0084] (2) Time dimension feature fusion: The time-series features of power grid data (such as daily load change curves and monthly operation and maintenance cost change data) are mapped to the phase parameters of quantum states by quantum phase encoding. Then, the phase features of the time dimension and the entangled weight features of the spatial dimension are deeply fused through quantum inverse Fourier transform (IQFT) to achieve integrated extraction of spatiotemporal features.

[0085] (3) Fusion Feature Output: The quantum states of the power grid quantum knowledge graph are calculated through quantum graph convolution unitary transformation. Then, through quantum measurement operations, the evolved quantum states are converted into classical fusion feature data, achieving efficient conversion from quantum features to classical features. The formula is:

[0086] ;

[0087] in, This represents the fusion feature output by the quantum-spacetime graph fusion layer. Measure represents the quantum measurement operation. The unitary transform operator for quantum graph convolution is the core operator for realizing quantum state evolution in the quantum-spacetime graph fusion layer. It completes the feature extraction and fusion of quantum states in the quantum knowledge graph of the power grid through the unitary transform of quantum gate circuits. The overall quantum state of the power grid quantum knowledge graph is formed by quantum nodes corresponding to three types of data: power grid operation data, equipment status data, and financial data, after being correlated by quantum entanglement. It contains spatiotemporal-topological-temporal multidimensional correlation features of multi-source data.

[0088] This fusion method integrates the core logic of quantum computing and spatiotemporal graph neural networks. Compared with the traditional ST-GNN algorithm, it significantly improves the accuracy of capturing spatiotemporal correlations of power grid data and achieves an extremely high level of accuracy in identifying the correlation between three-dimensional data, effectively solving the problem of insufficient spatiotemporal correlation capture in existing technologies.

[0089] Quantum-reinforcement learning evaluation optimization layer: Dynamically evaluate the closed loop through QE-DRL.

[0090] This layer addresses the two core shortcomings of existing technologies: "weak dynamic optimization capability" and "low efficiency of multi-objective optimization." Its core innovation lies in deeply coupling quantum optimization technology with deep reinforcement learning. It reconstructs the algorithm logic of deep reinforcement learning from core levels such as state space construction, policy network design, and decision logic update. Through the QE-DRL model, it constructs a dynamic closed loop of "evaluation-feedback-optimization" for power grid investment benefits, realizing efficient solution of multi-objective optimization and dynamic adjustment of evaluation strategies.

[0091] The main functions of this level are as follows:

[0092] 1. Quantum State Space Construction: Overcoming the limitations of limited dimensionality and incomplete coverage of traditional deep reinforcement learning state spaces, this method integrates the fused features output by the quantum-spacetime graph fusion layer. By jointly quantum encoding the multi-objective evaluation indicators of power grid investment benefits (economic benefits, technical reliability, social and environmental protection), these indicators are converted into quantum superposition states, thus constructing a high-dimensional quantum state space. ,in For quantum amplitude, To evaluate the state basis vectors, k is the index of the evaluation state basis vector in the high-dimensional quantum state space, used to distinguish different evaluation state basis vectors. Its value ranges from 1 to m, and it is an identifier parameter for traversing all state basis vectors. m is the total number of evaluation state basis vectors in the high-dimensional quantum state space, representing the dimensionality of the quantized state space and determining the coverage of the evaluation states. This high-dimensional quantum state space consists of multiple evaluation state basis vectors and their corresponding quantum amplitudes. The quantum amplitude represents the weight of each state basis vector. Through the characteristics of quantum superposition, full-dimensional coverage of the evaluation states is achieved, significantly improving the coverage of the state space and providing a more comprehensive state foundation for multi-objective optimization.

[0093] 2. Quantum Augmentation Policy Optimization: The traditional deep reinforcement learning policy network is reconstructed using the Quantum Approximate Optimization Algorithm (QAOA). Leveraging the advantages of quantum parallel computing, the optimal evaluation policy is rapidly searched and optimized. The specific implementation steps are as follows:

[0094] (1) Multi-objective reward function design: A weighted linear reward function is constructed, with financial return rate, equipment availability, and carbon emission reduction as the core evaluation indicators for economic, technological, and social objectives, respectively. The weights of each evaluation indicator are determined through a quantum game equilibrium algorithm, achieving a scientific allocation of multi-objective weights so that the reward function can accurately reflect the comprehensive level of power grid investment benefits; the formula is:

[0095] ;

[0096] in, It is the comprehensive reward value of power grid investment benefits, used to quantitatively characterize the comprehensive benefit level of power grid investment projects in economic, technological, social and environmental dimensions, and is the core evaluation indicator for multi-objective optimization. For financial return on investment, For equipment availability, For carbon emission reductions, , , The weights are determined by the quantum game equilibrium.

[0097] (2) Quantum Enhanced Strategy Update: The parameters of the reconstructed strategy network are iteratively optimized by quantum variational algorithm. Taking advantage of quantum parallel computing, multiple strategy directions are searched at the same time, which greatly improves the efficiency of strategy optimization and realizes the coordinated optimization of economic, technical and social goals. The solution time of multi-objective optimization is greatly compressed from several hours of traditional algorithm to more than ten minutes, while the coverage of global optimal solution reaches an extremely high level.

[0098] (3) Dynamic feedback optimization: A quantum-classical data interface is built to feed the real-time evaluation results back to the strategy network through the data interface. At the same time, real-time power grid operation data (such as fluctuations in new energy output and changes in line load) are connected. The strategy network automatically and dynamically adjusts the evaluation parameters and strategies based on the evaluation results and the real-time power grid operation data, forming a complete "evaluation-feedback-optimization" dynamic closed loop. This allows the evaluation strategy to adapt to the dynamic changes in the power grid operation status and investment benefits in real time, effectively solving the problem of weak dynamic optimization capabilities of existing technologies.

[0099] Quantum Causal Source Tracing Feedback Layer: Achieves causal inference and source tracing through Q-CAF.

[0100] This level addresses the core shortcomings of existing technologies, namely "poor interpretability" and "insufficient risk attribution." Its core innovation lies in the deep integration of three technologies—quantum correlation analysis, causal inference, and adversarial fusion—rather than their independent applications. By constructing a causal attribution system for power grid investment benefit assessment through the Q-CAF method, it enhances the algorithm's anti-interference capabilities while achieving accurate attribution and anomaly attribution of assessment results, providing a clear direction for investment optimization.

[0101] The main functions of this level are as follows:

[0102] 1. Quantum Causal Graph Construction: Based on the power grid quantum knowledge graph (GQ-KG), key causal links that significantly impact investment benefit assessment results are identified by calculating the quantum entanglement strength between nodes. A quantum causal graph is then constructed based on these key links, representing "data characteristics - assessment indicators - benefit results." The nodes of this quantum causal graph are quantized power grid data characteristics and assessment indicators, while the edges represent the quantum causal correlation strength between nodes, enabling structured and visualized modeling of the causes of assessment results.

[0103] 2. Quantum-Enhanced Causal Inference: This method reconstructs the logic of traditional causal inference using Quantum Bayesian Networks (QBNs). Leveraging the advantages of quantum parallel computing, it accelerates the search and analysis of causal links, enabling precise location and quantitative attribution of anomalies. The specific implementation steps are as follows:

[0104] (1) Anomaly localization: When the investment benefit assessment result deviates from the preset threshold, the quantum backpropagation algorithm is activated. Starting from the abnormal node of the assessment result, the algorithm traces back along the causal link of the quantum causal graph to accurately locate the core cause node that leads to the assessment anomaly and form a complete abnormal causal path. For example, the core cause chain of "low investment return rate" can be accurately traced.

[0105] (2) Quantitative Attribution: Through quantum measurement operations, the squared inner product modulus of the quantum states of each causal node and the anomalous quantum state is calculated. This modulus is used as the contribution of each node to the assessment of the anomaly, achieving precise quantification of the cause of the anomaly. This keeps the error of the attribution results at an extremely low level and clarifies the degree of influence of each factor on the assessment results. The calculation formula is:

[0106] ;

[0107] in, The quantum contribution of a causal node quantum state to the assessment anomaly is quantified to characterize the degree of influence of a single causal node quantum state on the anomaly of the investment benefit assessment result. The value ranges from 0 to 1, and the larger the value, the more significant the influence. It is an anomalous quantum state. It is a causal node quantum state. The inner product of quantum states characterizes the similarity and correlation between anomalous quantum states and causal node quantum states, and the square of its modulus is the quantum contribution.

[0108] 3. Optimize feedback output: Based on the anomaly location results and quantitative attribution data obtained from quantum causal inference, combined with the actual operation of the power grid, generate targeted investment optimization suggestions, such as adjusting the renewable energy consumption strategy, optimizing equipment operation and maintenance plans, and adjusting the investment allocation ratio. The optimization suggestions are then pushed to the power grid investment decision-making department to provide clear technical guidance for investment decisions and complete the entire closed loop of assessment-source tracing-optimization.

[0109] Visualization output layer: used for 3D fusion evaluation and display.

[0110] The visualization output layer, as the final presentation stage of the entire technical solution, is used to realize the classical conversion and multi-dimensional visualization of quantum evaluation results, while also completing interface adaptation with the existing power grid system to ensure the practical implementation and application of the technical solution. Specific implementation details are as follows:

[0111] 1. Quantum-Classical Conversion: Through standardized quantum measurement operations, all quantum state information output by the quantum-enhanced hybrid algorithm, such as quantum evaluation results, quantum causal tracing paths, and quantum optimization suggestions, is converted into classical data to meet the classical data processing needs of the existing power grid system.

[0112] 2. Multi-dimensional visualization: Based on the Vue.js front-end framework and ECharts visualization components, a dedicated visualization interface for post-evaluation of power grid investment benefits is developed to realize the visualization of multi-dimensional evaluation information, including three-dimensional data fusion weight distribution, multi-objective benefit comprehensive score, quantum causal link spectrum, investment benefit dynamic optimization curve, etc., so that investment decision-makers can intuitively and clearly grasp the evaluation results;

[0113] 3. Data Interface Adaptation: Develop standardized API data interfaces to achieve seamless integration with the existing energy management system (EMS), financial management system, and equipment management system of the power grid. Support bidirectional data interaction between systems, allowing optimization suggestions obtained from the assessment to be directly implemented in the existing power grid system, ensuring the practicality and operability of the technical solutions.

[0114] This application has the following advantages:

[0115] 1. A deeply coupled quantum-enhanced hybrid algorithm system was constructed, which breaks through the limitation of the "superficial combination" of quantum and classical algorithms at the core logic level, realizes the deep integration of the core logic of quantum computing and various intelligent algorithms, and gives full play to the technical advantages of quantum computing.

[0116] 2. A feature modeling method adapted to multi-source power grid data was developed, which accurately captures the spatiotemporal-topological-temporal multi-dimensional correlation features of power grid operation data, equipment status data, and financial data, improving the effectiveness and accuracy of multi-source data fusion features and laying a solid data foundation for subsequent evaluation.

[0117] 3. A complete dynamic closed-loop mechanism of "evaluation-feedback-optimization" has been established to realize dynamic evaluation and optimization of investment benefits based on real-time power grid operation data, so that the evaluation results can adapt to the dynamic changes of the power grid and investment benefits in real time.

[0118] 4. Achieve a dual improvement and balance between algorithm robustness and interpretability. While enhancing the algorithm's ability to resist data noise and external interference, establish a scientific mechanism for tracing the source of evaluation results, clarify the quantitative causes of evaluation results, and improve the interpretability of evaluation results.

[0119] 5. Significantly improves the efficiency and quality of solving multi-objective optimization problems, greatly compresses the solution time for multi-objective optimization of power grid investment benefits, and increases the coverage of the global optimal solution to an extremely high level, meeting the needs of efficient and accurate evaluation in actual engineering projects.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithms, characterized in that: include: Quantum data preprocessing layer: Based on quantum principal component analysis (QPCA) dimensionality reduction, quantum generative adversarial network (Q-GAN) is introduced to achieve quantum-principal component adversarial dimensionality reduction, and output robust dimensionality reduction data after anti-interference processing; Quantum-Spatiotemporal Graph Fusion Layer: By constructing a power grid quantum knowledge graph GQ-KG, quantum entanglement is achieved on multi-source data of the power grid, realizing semantic association and knowledge fusion of multi-source data of the power grid. The convolution kernel of the spatiotemporal graph neural network ST-GNN is reconstructed by introducing quantum graph convolution operation to obtain quantum-enhanced spatiotemporal graph neural network QE-ST-GNN. The quantum features of the power grid quantum knowledge graph GQ-KG are converted into classical fusion features through quantum-enhanced spatiotemporal graph neural network QE-ST-GNN. Quantum-Reinforcement Learning Evaluation and Optimization Layer: This layer constructs a dynamic closed loop of "evaluation-feedback-optimization" for power grid investment benefits through the quantum reinforcement learning model QE-DRL, realizing efficient solution of multi-objective optimization and dynamic adjustment of evaluation strategy; Quantum Causal Source Tracing Feedback Layer: The quantum causal adversarial fusion method (Q-CAF) is obtained by deeply integrating quantum correlation analysis, causal inference and adversarial fusion technology, which can achieve accurate location and quantitative attribution of assessment anomalies; Visualization output layer: used to realize the classical conversion and multi-dimensional visualization of quantum evaluation results.

2. The power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 1, characterized in that: The implementation steps of the quantized data preprocessing layer are as follows: Data integration: Standardize and integrate three core data categories—power grid operation data, equipment status data, and financial data—to construct a three-dimensional data matrix. , where R is the real number field, and n represents the total number of samples / number of data records of the multi-source data of the power grid; Quantum encoding: Using angle / phase encoding, the integrated three-dimensional data is mapped to quantum states. The angular parameters of the quantum state are jointly determined by the normalized values ​​of the three-dimensional data, thereby achieving an efficient conversion from classical data to quantum state. Counter-dimensionality reduction: Extracting principal component quantum states from quantum states using quantum principal component analysis (QPCA) The principal component quantum state is input into the quantum generative adversarial network (Q-GAN) for feature defense training, and finally outputs robust dimensionality reduction data after anti-interference processing.

3. The power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 2, characterized in that: The quantum generative adversarial network (Q-GAN) consists of a generator and a discriminator. The generator uses the characteristics of quantum superposition to simulate the distribution features of power grid data and generate power grid data samples with strong anti-interference capabilities. The discriminator is constructed using a quantum neural network (QNN) and takes advantage of the parallelism of quantum computing to accurately distinguish between real power grid data and noisy interference data.

4. The power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 1, characterized in that: The construction process of the power grid quantum knowledge graph GQ-KG is as follows: First, the power grid operation data, equipment status data, and financial data are mapped to three types of core nodes in the power grid quantum knowledge graph GQ-KG, where the power grid operation data corresponds to the power grid topology node, the equipment status data corresponds to the power equipment node, and the financial data corresponds to the financial node. Furthermore, by leveraging the strong correlation characteristics of quantum entanglement, attribute associations can be established between different types of nodes, thereby achieving semantic association and knowledge fusion of multi-source data.

5. The power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 4, characterized in that: The implementation steps of the quantum-spacetime graph fusion layer are as follows: Spatial dimension feature extraction: A quantum computing model is constructed using quantum gate circuits to calculate the quantum entanglement strength between any two nodes in the power grid quantum knowledge graph GQ-KG, and this quantum entanglement strength is used as the spatial correlation weight between nodes to achieve accurate quantification of the topological correlation features of power grid data; Temporal feature fusion: The temporal features of power grid data are mapped to the phase parameters of quantum states using quantum phase encoding. Then, the phase features of the temporal dimension are deeply fused with the entangled weight features of the spatial dimension through inverse quantum Fourier transform (IQFT) to achieve integrated extraction of spatiotemporal features. Fusion feature output: The quantum state evolution calculation of the quantum knowledge graph of the power grid is performed by quantum graph convolution unitary transformation, and then the evolved quantum state is converted into classical fusion feature data through quantum measurement operation, realizing efficient conversion from quantum features to classical features.

6. The power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 1, characterized in that: The implementation steps of the quantum-reinforcement learning evaluation optimization layer are as follows: Quantum state space construction: The fusion characteristics output by the quantum-spacetime graph fusion layer and the multi-objective evaluation index of power grid investment benefits are jointly quantum-encoded and converted into quantum superposition states to construct a high-dimensional quantum state space. ； in, For a high-dimensional quantum state space, For quantum amplitude, To evaluate the state basis vectors, k is the index of the evaluation state basis vector in the high-dimensional quantum state space, and m is the total number of evaluation state basis vectors in the high-dimensional quantum state space. Quantum reinforcement policy optimization: The quantum approximation optimization algorithm QAOA is used to reconstruct the policy network of deep reinforcement learning to obtain the quantum reinforcement learning model QE-DRL. By leveraging the advantages of quantum parallel computing, the optimal evaluation policy can be quickly searched and optimized.

7. The power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 6, characterized in that: The implementation steps for quantum strategy optimization are as follows: Multi-objective reward function design: Construct a weighted linear reward function, with financial return rate, equipment availability rate, and carbon emission reduction as the core evaluation indicators for economic, technological, and social objectives, respectively. The weights of each evaluation indicator are determined through a quantum game equilibrium algorithm to achieve a scientific allocation of multi-objective weights. Quantum-enhanced policy update: The parameters of the reconstructed policy network are iteratively optimized using a quantum variational algorithm. Leveraging the advantages of quantum parallel computing, multiple policy directions are searched simultaneously, significantly improving the efficiency of policy optimization and achieving synergistic optimization of economic, technological, and social objectives. Dynamic feedback optimization: A quantum-classical data interface is built to feed the real-time evaluation results back to the policy network through this data interface. At the same time, the real-time operation data of the power grid is connected. The policy network automatically and dynamically adjusts the evaluation parameters and policies based on the evaluation results and the real-time operation data of the power grid, forming a complete dynamic closed loop of "evaluation-feedback-optimization".

8. The power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 5, characterized in that: The implementation steps of the quantum causal origination feedback layer are as follows: Quantum causal graph construction: Based on the power grid quantum knowledge graph GQ-KG, by calculating the quantum entanglement strength between nodes, key causal links that have a significant impact on investment benefit assessment results are screened out. Based on these key causal links, a quantum causal graph of "data characteristics-assessment indicators-benefit results" is constructed. The nodes of this quantum causal graph are quantized power grid data characteristics and assessment indicators, and the edges are the quantum causal correlation strength between nodes. Quantum-enhanced causal inference: The logic of causal inference is reconstructed using quantum Bayesian networks (QBNs). Leveraging the advantages of quantum parallel computing, the search and analysis process of causal links is accelerated, enabling precise location and quantitative attribution of anomalies. Optimize feedback output: Based on the anomaly location results and quantitative attribution data obtained from quantum causal inference, combined with the actual operation of the power grid, generate targeted investment optimization suggestions, and push the optimization suggestions to the power grid investment decision-making department to complete the entire closed loop of assessment-source tracing-optimization.

9. A power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 8, characterized in that: The specific steps for implementing quantum-enhanced causal inference are as follows: Anomaly localization: When the investment benefit assessment result deviates from the preset threshold, the quantum backpropagation algorithm is activated. Starting from the abnormal node in the assessment result, the algorithm traces back along the causal link of the quantum causal graph to accurately locate the core cause node that caused the assessment anomaly and form a complete abnormal causal path. Quantitative attribution: Through quantum measurement operations, the inner product modulus square of the quantum state and the anomalous quantum state of each causal node is calculated and used as the contribution of each node to the assessment of the anomaly, thereby achieving accurate quantification of the cause of the anomaly.

10. The power grid multi-source data fusion investment benefit post-evaluation platform based on quantum hybrid algorithm according to claim 1, characterized in that: The visualization output layer can also be adapted to the interface of the power grid system.

Images