Power transaction market risk dynamic assessment system

Abstract

The invention discloses a power transaction market risk dynamic assessment system, which comprises a multi-modal sensing module used for collecting physical equipment operation data, market transaction data and environmental parameters in real time, and performing space-time alignment processing through edge computing nodes to generate a fusion data stream; the self-adaptive evaluation engine is in data connection with the multi-modal sensing module and is used for analyzing multi-dimensional risk association in the fused data stream; the invention belongs to the technical field of electricity markets, and aims to solve the problems that in the prior art, historical data static analysis is relied on, spatial-temporal scales of multi-source data are not uniform, and data cannot be given in time. The technical effects are that a closed loop is formed through sensor data, a multi-mode sensing module (alignment processing), an adaptive evaluation engine (risk analysis), a digital twin system (simulation strategy) and a dynamic feedback module (optimization system), automatic calibration and learning are realized, and the risk prediction accuracy and the reliability of the scheme are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power market, and in particular to a dynamic risk assessment system for a power trading market. Background Art

[0002] 1. In recent years, with the deepening of power market reform, transaction types and participants have become increasingly complex, and traditional risk monitoring systems are facing the following challenges:

[0003] 2. Risk assessment in the power trading market mostly relies on static analysis of historical data. The time and space scales of multi-source data such as equipment operation, market prices and environmental parameters are not uniform, which makes it difficult to integrate features and provide accurate data that combines the situations of all parties in a timely manner.

[0004] 3. Independent monitoring modules are usually used to collect grid operation data and market transaction information respectively, which leads to a significant lag in the correlation analysis between physical equipment failures and market price fluctuations. It is difficult to quantify the transmission effect of equipment failures on market supply and demand in a timely manner, which can easily lead to misjudgment.

[0005] Therefore, the existing needs are not met, so we propose a dynamic risk assessment system for the electricity trading market. Summary of the invention

[0006] To this end, the present invention provides a dynamic assessment system for power trading market risks to solve the above-mentioned problems in the prior art.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] The power trading market risk dynamic assessment system includes a multimodal perception module, which is used to collect physical equipment operation data, market transaction data, and environmental parameters in real time, and perform spatiotemporal alignment processing through edge computing nodes to generate a fused data stream;

[0009] An adaptive assessment engine, connected to the multimodal perception module data, for parsing the multi-dimensional risk associations in the fused data stream, generating a risk assessment matrix that complies with the risk assessment and triggering a strategy preview instruction;

[0010] A numerical twin decision system, connected to the adaptive evaluation engine control, is used to preview the risk intervention strategy in a virtual trading environment and select the optimal solution through a two-word genetic algorithm;

[0011] A dynamic feedback control module forms a closed-loop connection with the digital twin decision-making system and the multimodal perception module, and is used to correct model parameters and dynamically adjust data acquisition strategies;

[0012] Each module realizes closed-loop operation of risk identification, strategy optimization and system adaptation through the interaction of data flow and control instructions.

[0013] Furthermore, the multimodal perception module includes:

[0014] A power transmission and transformation joint node detection unit that deploys a multi-physical quantity sensor array to collect in real-time the voltage, current harmonic distortion rate, and equipment temperature data of key grid nodes, and detects the conduction risk of physical equipment abnormalities to market supply and demand;

[0015] A microtransaction behavior capture unit that is embedded in the core of the power trading platform, records the quotation trajectories and order cancellation patterns of traders at a sampling interval of 50 ms, and constructs a trader association graph to identify market manipulation behaviors;

[0016] An environmental risk monitoring unit that, based on meteorological radar and satellite cloud images, outputs the probability value of regional meteorological disasters within the next 2 hours, and quantifies the potential impact of extreme weather on the power market.

[0017] Furthermore, the spatio-temporal alignment processing includes:

[0018] Using the interpolation method to fill in the missing values of physical equipment data, and calculating the local mean and standard deviation through a sliding window;

[0019] Performing spatio-temporal correlation analysis on environmental parameters and equipment status, and constructing a correlation matrix.

[0020] Furthermore, the adaptive evaluation engine includes: a market correlation analysis unit, a game behavior recognition unit, and a risk visualization unit.

[0021] Furthermore, the market correlation analysis unit uses a spatio-temporal graph convolutional network (ST-GCN) to model the coupling relationship between the power grid topology and trading behaviors, and calculates the conduction coefficient of equipment failures on market prices.

[0022] Furthermore, the game behavior recognition unit detects abnormal strategies that deviate from the Nash equilibrium based on the reinforcement learning game tree model.

[0023] Furthermore, the risk visualization unit: generates a three-dimensional heat map to dynamically mark the risk conduction path and key nodes.

[0024] Furthermore, the strategy rehearsal of the digital twin decision system includes:

[0025] Encoding the intervention strategy into a qubit chromosome, and adjusting the superposition state of the strategy combination through a quantum rotation gate;

[0026] The fitness function is defined as the revenue-cost ratio of strategy simulation in the virtual environment;

[0027] Outputting the Pareto front solution set, and selecting strategies with a revenue-cost ratio greater than the threshold to generate control instructions.

[0028] The present invention has the following advantages:

[0029] The power trading market risk dynamic assessment system senses various signals in the power market through a multi-modal perception module, packs and transmits the data to an adaptive assessment engine for risk prediction, generates a composite risk assessment matrix and triggers a strategy rehearsal instruction, and sends it to the digital twin decision system for rehearsing risk intervention strategies in a virtual trading environment and selecting the optimal solution through a quantum genetic algorithm. A closed-loop connection is formed by the dynamic feedback control module, the digital twin decision system, and the multi-modal perception module, which is used to correct model parameters and dynamically adjust the data acquisition strategy, realizing automatic calibration and learning, and improving the accuracy of risk prediction and the reliability of the solution. Brief Description of the Drawings

[0030] Figure 1 It is the overall schematic diagram of the power trading market risk dynamic assessment system proposed by the present invention;

[0031] Figure 2 It is the system schematic diagram of the power trading market risk dynamic assessment system proposed by the present invention;

[0032] Figure 3 It is the schematic diagram of the multi-modal perception module of the power trading market risk dynamic assessment system proposed by the present invention;

[0033] Figure 4 It is the schematic diagram of the adaptive assessment module of the power trading market risk dynamic assessment system proposed by the present invention;

[0034] Figure 5 It is the flowchart of the power trading market risk dynamic assessment system proposed by the present invention. Detailed Embodiment

[0035] The following specific embodiments illustrate the implementation manners of the present invention. Those skilled in the art can easily understand the other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0036] Referring to Figures 1-5 , the power trading market risk dynamic assessment system includes:

[0037] Multi-modal Sensing Module: 1. Power Grid Point Detection Unit, which collects the transformer temperature in real time through precision sensors to detect overload risks; 2. Transaction Behavior Capture Unit, which records high-frequency order cancellation behaviors at 50ms intervals to construct an association graph for early warning of collusive manipulation; 3. Environmental Prediction Unit, which combines micro cloud maps to predict typhoon paths and quantify the probability of power transmission corridor interruption; 4. Temporal and Spatial Alignment, which unifies device data and market prices to the minute level, fills in missing values and calculates local statistics;

[0038] Adaptive Evaluation Engine: 1. Market Association Analysis, which models the conduction coefficient of transformer tripping on regional electricity prices using ST-GCN; 2. Game Behavior Identification, which detects abnormal strategies of false order placement inducing market fluctuations; 3. Risk Visualization, which generates 3D heat maps to mark the propagation chain of risks from equipment failures to market imbalances;

[0039] Digital Twin Decision System: 1. Virtual Environment Construction, which synchronizes real market rules and power grid topologies and injects historical extreme event data; 2. Strategy Rehearsal, which encodes "limit price 5%" as a quantum chromosome {0101}, and selects a plan with a benefit-cost ratio > 205 after simulating 100 times; 3. Multi-department Collaboration, when a typhoon risk is triggered, synchronously adjusts the power grid topology and activates the market circuit breaker;

[0040] Dynamic Feedback Control Module: 1. Model Self-evolution, which compares the rehearsal results with the actual effects and adjusts the weight parameters of the deep reinforcement learning model; 2. Sensing Resource Scheduling, which activates infrared thermal imaging sensors in high-risk areas and switches to the basic monitoring mode in low-risk areas;

[0041] Specifically: A. Multi-modal Sensing Module:

[0042] It is used to collect physical device operation data, market transaction data, and environmental parameters in real time, and perform temporal and spatial alignment processing through edge computing nodes to generate a fusion data stream containing device state characteristics, price fluctuation characteristics, and environmental risk characteristics;

[0043] It also includes:

[0044] a1. Power Transmission and Transformation Key Node Monitoring Unit: Deploy a multi-physical quantity sensor array to collect voltage, current harmonic distortion rate, and device temperature data in real time, which is used to capture the risk of market supply and demand imbalance caused by abnormal power grid physical states. It monitors the operation status of key power grid nodes in real time through a high-precision sensor network (±0.1% accuracy), solves the misjudgment problem caused by lagging device failure information in traditional systems, and at the same time detects the correlation between power equipment abnormalities (such as transformer overload) and market price fluctuations, and early warns of chain market risks caused by physical failures;

[0045] a2. The micro trading behavior capture unit is embedded in the core of the power trading platform and records the quotation trajectories, order cancellation patterns, and position changes of traders at a sampling interval of 50 ms, which is used to identify market manipulation behaviors; it is used to capture abnormal behavior patterns in high-frequency trading (such as flash order cancellation and false quotations), identify hidden manipulation strategies that cannot be detected by traditional regulatory means, construct a trader association map through behavioral data analysis, and warn of the risk of collusive manipulation;

[0046] a3. The environmental risk prediction unit, based on meteorological radar and satellite cloud images, outputs the probability value of regional meteorological disasters within the next 2 hours, quantifies the influence probability of extreme weather (such as typhoons and ice disasters) on power generation capacity and transmission corridors, and converts environmental risks into market risk estimates; it provides forward-looking data support for cross-regional power dispatching decisions;

[0047] Among them, the spatio-temporal alignment processing includes: synchronizing the timestamps of sensor time series signals in the physical device operation data, and filling in missing data using the interpolation method; aligning the price fluctuation sequences in the market transaction data in the spatial dimension, and calculating the local mean and standard deviation through a sliding window; analyzing the correlation between meteorological data and device operation status in environmental parameters to construct a spatio-temporal correlation matrix;

[0048] B. Adaptive evaluation engine: Data-connected to the multi-modal perception module, used to receive the fusion data stream and perform the following processing:

[0049] b1. Dynamically analyze the multi-dimensional risk associations in the data stream through a deep reinforcement learning model, and generate a composite risk assessment matrix including price fluctuation risk values, supply-demand imbalance risk values, and game behavior risk values, where:

[0050] The deep reinforcement learning model adopts a double deep Q-network architecture;

[0051] The generation of the composite risk assessment matrix includes:

[0052] The market association analysis unit uses a spatio-temporal graph convolutional network (ST-GCN) to model the coupling relationship between the power grid topology and trading behaviors, calculates the conduction coefficient of equipment failures on market prices, reveals the quantitative relationship between power equipment anomalies (such as line tripping) and market price fluctuations, and realizes cross-domain association analysis of physical layer risks and market layer risks;

[0053] The game behavior recognition unit applies a reinforcement learning game tree model to simulate the strategy choices of traders in different market states, detects abnormal game behaviors that deviate from the Nash equilibrium, discovers the behavior patterns of market participants creating liquidity illusions through false quotations, warns of the risk of market failure caused by game strategy conflicts, and calculates the price fluctuation risk value at the same time;

[0054] Price volatility risk value: Calculate the risk value by simulating the future price path through Markov Chain Monte Carlo (MCMC);

[0055] The risk visualization unit generates a three-dimensional heat map including time, space, and risk level, dynamically annotates the risk transmission path and key impact nodes, and intuitively displays the transmission chain of risk events from physical equipment failure to market transaction imbalance;

[0056] b2. Automatically trigger the strategy rehearsal instructions of the digital twin decision system based on the abnormal pattern recognition results of the risk assessment matrix;

[0057] C. Digital twin decision system: connected to the adaptive evaluation engine control system, used to:

[0058] c1. Build a virtual trading environment that is updated synchronously with the physical power market, and load the corresponding risk intervention strategy set after receiving the strategy preview instruction. The purpose is to safely simulate extreme scenarios such as market collapse and liquidity depletion in the virtual environment, avoid the trial and error costs of the real market, and verify the effectiveness of intervention measures such as circuit breakers and liquidity injections;

[0059] c2. Execute multiple rounds of strategy effect simulation in a virtual environment, select the optimal intervention plan and generate control instructions based on quantum genetic algorithm, balance risk control costs and market efficiency losses, and achieve dual-objective optimization of risk prevention and control and market operation efficiency;

[0060] c3. Connect the power dispatching system, financial supervision platform and meteorological warning center to automatically generate a multi-department collaborative disposal plan. When a chain risk caused by a typhoon is detected, it will simultaneously trigger the grid topology adjustment, market fuse and disaster emergency response, solving the problem of asynchronous response of various departments in traditional disposal;

[0061] D. Dynamic feedback control module: forms a closed-loop connection with the digital twin decision-making system and the multimodal perception module, and is used to:

[0062] d1. Reversely transmit the actual implementation effect data of the intervention plan to the adaptive evaluation engine for model parameter correction;

[0063] d2. Adjust the data collection frequency and feature extraction dimension of the multimodal perception module according to the real-time changes of the risk assessment matrix;

[0064] The dynamic feedback module includes:

[0065] Model self-evolution unit: By comparing the virtual environment preview results with the real market response data, the weight parameters of the assessment model are automatically corrected, so that the risk assessment model is iteratively updated every week to maintain adaptability to changes in the market environment, and the prediction accuracy of the risk transmission path is continuously optimized through online learning;

[0066] Perception Resource Scheduling Unit: Dynamically configure sensor resources according to the distribution of risk hotspots. In high-risk areas, activate advanced sensing modes such as infrared thermal imaging and partial discharge detection. In low-risk areas, switch to the basic monitoring mode to reduce energy consumption and prioritize ensuring the data acquisition quality of key risk points.

[0067] Working Principle: Collect data of each transformer through sensors, transmit the data to the multi-modal perception module, align data of different frequencies to the "minute level", fill in missing values, and finally package it into a fused data report and transmit it to the adaptive evaluation engine in real time; conduct risk diagnosis on the received data; perform market correlation analysis, use the AI model (ST-GCN) to analyze how power grid failures affect electricity prices, and identify game behaviors: through the game theory model, detect abnormal trading behaviors. For example, if someone places a large number of fake orders to create an illusion of "market prosperity", the model will mark it as "suspected manipulation". Risk visualization: Generate a dynamic heat map, mark high-risk areas with colors (such as red representing possible electricity price crashes), and mark the risk propagation path; if the risk value exceeds the threshold range, immediately send an instruction to the digital twin decision system; generate a pre-plan and simulate strategies through the digital twin decision system to actively respond when an accident occurs. The dynamic feedback module adjusts the plan according to the actual effect, and model calibration: Compare the expected effect and the actual result of the strategy. For example, if it is expected that price limits can reduce the risk by 20%, but actually only reduce it by 10%, automatically lower the weight of "price fluctuation" in the model. Data acquisition optimization: In high-risk areas (such as substations on the typhoon path), turn on advanced monitoring such as infrared thermal imaging; switch to the basic mode in low-risk areas to save power, forming a closed loop among sensor data → multi-modal perception module (alignment processing) → adaptive evaluation engine (risk analysis) → digital twin system (simulation strategy) → dynamic feedback module (system optimization), and adjust the previous strategy through subsequent feedback to make the model prediction more accurate and data acquisition more efficient.

Claims

1. The power trading market risk dynamic assessment system is characterized by: include: The multimodal perception module is used to collect real-time data on physical equipment operation, market transaction data, and environmental parameters, and perform spatiotemporal alignment processing through edge computing nodes to generate a fused data stream; An adaptive assessment engine, connected to the multimodal perception module data, for parsing the multi-dimensional risk associations in the fused data stream, generating a risk assessment matrix that complies with the risk assessment and triggering a strategy preview instruction; A numerical twin decision system, connected to the adaptive evaluation engine control, is used to preview the risk intervention strategy in a virtual trading environment and select the optimal solution through a two-word genetic algorithm; The dynamic feedback control module forms a closed-loop connection with the digital twin decision-making system and the multimodal perception module to correct model parameters and dynamically adjust data acquisition strategies; Each module realizes closed-loop operation of risk identification, strategy optimization and system adaptation through the interaction of data flow and control instructions.

2. The power trading market risk dynamic assessment system according to claim 1 is characterized in that: The multimodal perception module comprises: The power transmission and transformation node detection unit deploys a multi-physical quantity sensor array to collect voltage, current harmonic distortion rate and equipment temperature data at key power grid nodes in real time, and detect the transmission risk of physical equipment abnormalities to market supply and demand; The micro-trading behavior capture unit is embedded in the core of the power trading platform, records the quote trajectory and order withdrawal pattern of traders at a sampling interval of 50ms, and constructs a trader association map to identify market manipulation behavior; The environmental risk monitoring unit, based on meteorological radar and satellite cloud images, outputs the probability value of regional meteorological disasters within the next 2 hours and quantifies the potential impact of extreme weather on the power market.

3. The power trading market risk dynamic assessment system according to claim 2 is characterized in that: The spatiotemporal alignment process includes: Interpolation is used to fill missing values ​​in physical device data, and the local mean and standard deviation are calculated through a sliding window; Conduct spatiotemporal correlation analysis on environmental parameters and equipment status and construct a correlation matrix.

4. The power trading market risk dynamic assessment system according to claim 3 is characterized in that: The adaptive evaluation engine includes: a market association analysis unit, a game behavior identification unit, and a risk visualization unit.

5. The power trading market risk dynamic assessment system according to claim 4 is characterized in that: The market correlation analysis unit adopts a spatiotemporal graph convolutional network (ST-GCN) to model the coupling relationship between power grid topology and transaction behavior, and calculates the transmission coefficient of equipment failure to market price.

6. The power trading market risk dynamic assessment system according to claim 5 is characterized in that: The game behavior identification unit detects abnormal strategies that deviate from Nash equilibrium based on a reinforcement learning game tree model.

7. The power trading market risk dynamic assessment system according to claim 6 is characterized in that: The risk visualization unit generates a three-dimensional heat map to dynamically mark the risk transmission path and key nodes.

8. The power trading market risk dynamic assessment system according to claim 7 is characterized in that: The strategy preview of the digital twin decision-making system includes: Encode the intervention strategy into a quantum bit chromosome and adjust the superposition state of the strategy combination through a quantum rotating gate; The fitness function is defined as the benefit-cost ratio of the strategy simulation in the virtual environment; Output the Pareto frontier solution set and select the strategy with benefit-cost ratio greater than the threshold to generate control instructions.

Images