An intelligent risk prevention and control system for power market subjects suitable for high-proportion new energy provinces

Abstract

The application discloses a power market main body intelligent risk prevention and control system suitable for high-proportion new energy provinces and belongs to the technical field of power market risk control. The system comprises six modules, namely, a multi-source time sequence data base, a dynamic risk rule library, a multi-model parallel risk calculation engine, a hierarchical early warning output, a station-level risk tracing and cross-market income joint defense. The application solves the time domain mismatch, coarse supervision granularity, distortion of congestion allocation and multi-market risk fragmentation defects of the prior art, realizes station penetrating risk identification, early warning, accurate responsibility division and cross-market risk blocking, and is suitable for wind and light high-penetration spot market full-process closed-loop risk control.

Description

Technical Field

[0001] This invention belongs to the field of intelligent risk control in the power market and big data analysis technology of new energy, specifically involving a multi-entity integrated intelligent risk prevention and control system adapted to provincial power spot markets with high wind and solar power penetration. Background Technology

[0002] With the full implementation of provincial-level electricity spot markets in China, the penetration rate of wind and solar power capacity in renewable energy-rich provinces such as Gansu, Qinghai, and Inner Mongolia continues to rise, forming a new electricity market operation system with multiple trading instruments coupled together. Renewable energy possesses unique physical characteristics such as second-level random fluctuations, simultaneous declarations by power plant clusters, and limited ramp-up adjustment capabilities, giving rise to four unique market risks: significant price disorder, coordinated manipulation by market participants, unfair cost sharing due to transmission congestion, and the collapse of operating profits for renewable energy companies. Traditional electricity trading monitoring tools designed for thermal power-dominated markets are inadequate for the refined and penetrating risk control needs of high-proportion renewable energy scenarios, necessitating a completely new intelligent risk prevention and control system. Currently, risk control tools in the domestic and international power markets can be divided into two main categories: international commercial platforms and domestic power grid-supporting monitoring systems. Representative international products include PJM's PRISM price warning platform (USA), AEMO's MARS market power monitoring system (Australia), and NordPool congestion management platform (Norway). These primarily rely on static statistical thresholds, the HHI market concentration index, and the postage stamp method for average allocation to achieve basic monitoring. Existing domestic solutions include the State Grid D5000 intelligent early warning module, the China Southern Power Grid EMS negative electricity price circuit breaker system, and the Electric Power Research Institute's virtual machine group monitoring platform. These employ basic LSTM price forecasting, unified proportional congestion cost allocation, and aggregator-level data reporting monitoring, and can only provide simple anomaly alerts under conditions where thermal power accounts for a high proportion and output is stable. The existing technology suffers from four irreconcilable underlying technical flaws, each with a clear technical root cause: 1. When photovoltaic (PV) power is obstructed by clouds, the output of the generating units will drop sharply in a short period of time. The electricity spot market uses a fixed 15-minute clearing cycle. Existing tools lack cross-scale fluctuation conversion algorithms and rely solely on whole-cycle average data for analysis. The activation of the existing circuit breaker mechanism in China has a long delay, and the trigger may only occur after the price anomaly has persisted for some time. The root cause is the lack of an equivalent calculation model that converts the second-level instantaneous wind and solar disturbances into the market clearing cycle, making it impossible to incorporate short-term sharp fluctuations into risk prediction. 2. Existing monitoring systems can only identify transaction behavior at the aggregator level and cannot penetrate to subordinate wind farms; the correlation between wind and solar power output and natural power generation can easily lead to synchronized pricing; the traditional single HHI index cannot distinguish between natural fluctuations and human-coordinated manipulation; there are a large number of missed detections for illegal declarations of regional wind power clusters; existing identification algorithms require complete private operation data from wind farms, which cannot be implemented due to communication and data privacy barriers. 3. Traditional stamp-based and equal-power-allocation models do not consider the ramp-up limits of new energy sources and only allocate congestion costs based on power generation. In the actual test scenarios of the Qinghai and Zhangjiakou projects, the determination of congestion responsibility deviates significantly from the actual physical contribution of the power grid, and the calculation results of cost allocation for a single power station differ significantly from the actual contribution. Existing tools do not have a game theory quantification framework and cannot calculate the marginal contribution of a single power station to cross-sectional congestion, which can easily lead to cost disputes among market participants. 4. The risks of the spot market, green certificates, and carbon market are interconnected and transmitted, but the existing system operates independently for each module, and there is no unified quantitative model for the correlation of risks in multiple markets; the return-risk indicators are updated statically on a daily basis, and the update of the cost-per-kilowatt-hour coverage ratio (DCR) indicator has a significant delay, which cannot cope with the rapid decline in spot electricity prices in the short term, makes it difficult to predict the risk of bankruptcy of new energy companies in advance, and lacks the ability to suppress cross-market fluctuations. Summary of the Invention

[0003] To address the four major technical shortcomings of existing power market risk control systems—time-domain scale mismatch, lack of power plant penetration supervision, distorted quantification of congestion responsibility, and fragmented transmission of risks across multiple markets—this invention provides an intelligent risk prevention and control system for power market participants in provinces with a high proportion of renewable energy. Through a multi-scale data foundation, a dynamic rule base, four parallel theoretical models, tiered circuit breaker early warning, full-cycle traceability, and cross-market hedging, this system achieves coupled calculation of instantaneous fluctuations in wind and solar power with the spot market, precise identification through power plant-level collaborative operation, accurate allocation of congestion responsibility driven by differential game theory, and joint prevention of risks and benefits across multiple markets. This constructs a closed-loop risk prevention and control system encompassing pre-event prediction, in-event monitoring, and post-event traceability. The objective of this invention can be achieved through the following technical solutions: A smart risk prevention and control system for power market entities in provinces with a high proportion of renewable energy includes a multi-source time-series data base module, a dynamic risk rule base management module, a multi-model parallel risk calculation engine module, a hierarchical early warning output module, a station-level risk tracing and analysis module, and a cross-market revenue joint prevention and control module. The multi-source time-series data base module is used to collect multi-dimensional time-series data from a high proportion of the new energy power market and construct a cross-timescale data mapping channel; the dynamic risk rule base management module is used to maintain new energy-specific risk indicators, dual-layer thresholds, and five-level risk judgment rules and adaptively calibrate them; the multi-model parallel risk calculation engine module integrates four types of theoretical models for parallel operation and outputs four types of quantitative risk indicators: price, market power, congestion, and return; the graded early warning output module matches the risk level, outputs graded alarms, and automatically executes price circuit breaker control logic; the station-level risk tracing and analysis module completes risk source positioning and responsibility quantification before, during, and after the event; the cross-market return joint prevention and control module dynamically generates financial hedging schemes based on multi-market correlation models to block cross-market risk transmission. As a further aspect of the present invention, the data collected by the multi-source time-series data base module includes minute-level wind and solar power forecasts, unit ramp-up rates, weather warnings, market declaration and clearing, cross-sectional tidal currents, green certificates, and carbon trading time-series data. The module has a built-in fluctuation energy equivalent conversion algorithm to perform integral calculations on second-level wind and solar power output fluctuations, converting them into equivalent power deviations that adapt to a fifteen-minute market clearing cycle, establishing a second-level to fifteen-minute cross-scale data mapping channel, and eliminating the time-domain mismatch between instantaneous fluctuations of new energy and the market clearing cycle. As a further aspect of the present invention, the dynamic risk rule base management module incorporates three new energy-specific indicators: wind and solar forecast deviation, new energy cluster synergy index (CBI), and cost-per-kilowatt-hour coverage ratio (DCR); it establishes a two-layer threshold system of warning yellow light and severe red light, as well as five-level risk mapping rules: no risk, low risk, medium risk, high risk, and severe risk; and it is equipped with a threshold dynamic calibration algorithm that iteratively updates thresholds based on three historical typical scenarios: wind and solar peak, stable, and low-end, to achieve adaptive adjustment according to season and new energy penetration rate. As a further embodiment of the present invention, the multi-model parallel risk calculation engine module comprises four independent parallel computing units: The price disorder prediction unit adopts an improved Ornstein-Uhlenbeck nonstationary stochastic process model that introduces cloud cover / sudden wind jump terms to quantify the impact of wind and solar fluctuations on electricity prices. The market power identification unit constructs a three-party multi-agent evolutionary game model involving wind and solar aggregators, power grids, and users. It identifies cluster collaborative manipulation behavior through two criteria: Nash equilibrium deviation, power station output correlation coefficient ρ greater than 0.8, and bid similarity δ greater than 0.75. The congestion responsibility calculation unit, based on the differential Stackelberg game framework, coupled the cross-sectional power flow sensitivity and the unit ramp deviation to solve the station congestion contribution rate, and adopted the Shapley-entropy weight composite algorithm to allocate the congestion cost; The cross-market return quantification unit uses a time-varying Copula function to establish a joint distribution of spot electricity price and green certificate price, and combines it with conditional value at risk (ES) to complete the return risk quantification. As a further embodiment of the present invention, the congestion responsibility calculation unit is configured with a penalty apportionment rule. When the congestion contribution rate of a station exceeds the benchmark threshold, the station is automatically determined to bear 1.5 times the standard congestion cost. As a further embodiment of the present invention, the five levels of risk of the graded early warning output module correspond one-to-one with the wind and solar forecast deviation: the deviation in the first interval is a risk-free green light, the deviation in the second interval is a low-risk blue light, the deviation in the third interval is a medium-risk yellow light, the deviation in the fourth interval is a high-risk orange light, and the deviation in the fifth interval is a severe-risk red light; the module has a built-in dynamic circuit breaker logic, which automatically suspends market trading and activates the backup trading channel when the electricity price fluctuation exceeds the preset fluctuation boundary within fifteen minutes. As a further embodiment of the present invention, the station-level risk tracing analysis module includes a three-stage tracing logic: pre-event tracking of reserve gaps caused by sudden changes in wind and solar power output, real-time monitoring of abnormal deviations in the pricing of new energy entities during the event, and post-event quantification of economic losses caused by events such as negative electricity prices; it is equipped with two sub-units: special tracing of blocked sections and attribution of abnormal revenue, and outputs a visualized report on the dual responsibility ratio at the unit level and section level. As a further embodiment of the present invention, the cross-market return joint prevention and control module is equipped with a mean-CVaR derivative hedging optimization model and a green certificate trading return simulator; when the conditional value of risk ES output by the cross-market return quantification unit is lower than the safety threshold, the hedging optimization calculation is automatically started, and suitable futures and options hedging combination schemes are pushed to market participants to reduce the impact of the linkage fluctuations in the spot, green certificate and carbon markets. As a further embodiment of the present invention, the system hardware carrier is a multi-core high-performance server cluster, equipped with a distributed time-series database and a multi-GPU parallel computing power scheduling unit. The time taken for a single round of synchronous calculation of four types of risks in the entire market does not exceed sixty seconds, supporting the minute-level real-time risk monitoring calculation needs. Beneficial effects This system incorporates a fluctuation energy equivalent conversion algorithm into its data base, performing integral calculations on second-level wind and solar power output fluctuations. This converts instantaneous and severe disturbances into equivalent power deviations over a 15-minute clearing cycle, overcoming the shortcomings of traditional tools that only use periodic averages and ignore short-term fluctuations. It also features an improved Ornstein-Uhlenbeck process model, adding cloud cover and sudden wind jump terms to characterize extreme wind and solar disturbances, and proactively anticipating the risk of price collapses caused by insufficient reserve margins. This invention can incorporate second-level fluctuations into risk calculations in real time, enabling the prediction of negative electricity price events over a longer period. In real-world testing scenarios of severe wind and solar power generation in the Hexi Corridor, the system's accuracy in detecting price anomalies far surpasses that of traditional monitoring tools. Employing a federated learning three-party game model, the system reconstructs station behavior solely through gradient exchange, eliminating the need to access private station data and breaking down data communication barriers. It innovates with a dual-judgment standard for the CBI cluster collaboration index, distinguishing between natural synchronous wind and solar power output and manipulative behaviors such as artificially inflated output and price manipulation. This allows for regulatory granularity down to the individual station unit, coupled with a tiered handling logic. The invention can accurately identify various types of strategic behaviors in the new energy sector, significantly improving the coverage of illegal manipulation identification and eliminating the need for manual verification of application curves at each station, thus significantly reducing the workload of human-based operational analysis. A differential Stackelberg game framework was constructed to solve a dual-objective problem involving optimizing the social welfare of the power grid and optimizing the power output strategy of power plants. The marginal congestion contribution rate of power plants was calculated by coupling cross-sectional power flow sensitivity and unit ramp deviation. The Shapley-entropy weighted composite algorithm was used to replace the stamp method, matching congestion costs according to the power plant's regulation capacity and setting a 1.5 times penalty apportionment coefficient for high-contribution power plants. This invention takes into account both physical regulation capacity and cross-sectional contribution, significantly improving the alignment of congestion responsibility determination results with actual power grid operating conditions, and controlling the difference between the calculated congestion cost of a single power plant and the actual contribution within a very small range. In pilot scenarios in Haixi, Qinghai and Zhangjiakou, the number of market disputes and complaints related to congestion and imbalance funds was significantly reduced, and overall market balancing expenditures were effectively reduced. The system employs a time-varying Copula function to construct a dynamic joint distribution of spot electricity prices and green certificate prices, introducing seasonal and policy-varying parameters to characterize their linkage. It quantifies the overall return risk through Conditional Value at Risk (ES). Equipped with a mean-CVaR hedging optimization model, it automatically generates derivative hedging portfolios when ES falls below a safety threshold, creating transmission damping in the early stages of risk diffusion. The system synchronously links multiple market data at the minute-level, significantly reducing the impact of risk linkages between different markets and markedly decreasing the probability of operational pressure on new energy enterprises. The DCR (Digital Cost Reduction) coverage rate warning has no long time delay and can proactively push hedging solutions to stabilize the main entity's returns. Attached Figure Description The invention will now be further described with reference to the accompanying drawings. Figure 1 This is a schematic diagram of a smart risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in this invention. Detailed Implementation The present invention will be further described in detail below with reference to specific embodiments. These embodiments are intended to fully disclose the technical solutions of the present invention, enabling those skilled in the art to implement the present invention, and are not intended to limit the scope of protection of the present invention. Example 1: Complete Implementation of the Scenery of the Hexi Corridor in Gansu Province 1. Basic operating parameters Test scenario: The Hexi Corridor in Gansu Province is a region where wind and solar power are concentrated and transmitted. The installed capacity of wind and solar power in the region far exceeds the local load. During the off-peak hours, the total predicted output of wind and solar power exceeds the regional electricity demand, which can easily lead to negative electricity prices and the risk of cross-section blockage. Hardware configuration: Multiple multi-core high-performance servers, multiple GPU parallel computing units, and a distributed time-series database storing nearly twelve months of minute-level time-series data on wind and solar power output, market reporting, power grid flow, and green certificate transactions. 2. System Module Operation Flow Step 1: The multi-source time-series data base collects minute-level power output change data of photovoltaic power stations within 15 consecutive minutes. A single power station is affected by cloud cover, and its power output drops sharply in a short period of time. The fluctuation energy equivalent conversion algorithm is used to convert it into the equivalent power deviation of the current 15-minute cycle, and a standardized time-series dataset is generated and sent to the computing engine. Step 2: The dynamic risk rule base reads the adaptive calibration threshold for the wind and solar power scenario, including the warning threshold and severe threshold for wind and solar power prediction deviation, the warning threshold and severe threshold for CBI, and the safety threshold for DCR. Step 3: The multi-model parallel risk calculation engine simultaneously performs four-way model calculations: ① Price Disorder Unit: The improved Ornstein-Uhlenbeck model shows that the probability of negative electricity prices occurring in this cycle is extremely high; ② Market Power Identification Unit: The output correlation coefficient and price similarity of multiple photovoltaic power stations in the region both exceeded the critical value of the dual criteria, and the CBI index reached the level 2 risk range, indicating that there is a level 2 market power risk; ③ Congestion Responsibility Unit: When the congestion contribution rate of a designated photovoltaic collection station exceeds the baseline threshold, the 1.5 times congestion cost sharing rule is triggered; ④ Return Quantification Unit: The time-varying Copula model calculates that the ES value is lower than the safety threshold, indicating a high level of return risk. Step 4: The graded early warning output module matches the current wind and solar forecast deviation amount to the fifth interval, triggering a red light severe risk alarm; if it is predicted that the electricity price fluctuation will exceed the preset boundary in the next fifteen minutes, the circuit breaker mechanism will be automatically activated, suspending the spot market application channel and opening the backup capacity trading channel. Step 5: The risk tracing module at the site level outputs a complete tracing report: The root cause of the negative electricity price risk is that the predicted output of wind and solar power far exceeds the load demand and the load is in a low range; the core responsible party for the blockage is the aforementioned photovoltaic aggregation site; the source of abnormal market power is multiple photovoltaic sites that simultaneously adjusted their application curves. Step 6: The cross-market profit joint prevention and control module starts the mean-CVaR hedging model to generate a hedging combination scheme for forward green electricity options to hedge against the main profit loss caused by the decline in electricity prices. Predicting the duration of negative electricity price events Warning signals can be issued a considerable period in advance. It can only provide a brief warning when the risk is imminent. Cluster market manipulation identification coverage It can cover the vast majority of covert collaborative behaviors. A large number of station-level coordinated manipulation behaviors could not be identified. Difference between the calculated responsibility for station congestion and the actual contribution The difference is controlled within a very small range. The calculated results differ significantly from the actual physical contributions. Time consumed in single-round full-market risk synchronous calculation No more than sixty seconds Completing a full-market risk calculation takes significantly longer. Cross-market linkage risk mitigation effect It can significantly reduce the intensity of risk transmission across multiple markets. Non-targeted cross-market risk mitigation capabilities Changes in overall market balance fund expenditures Achieve significant pressure reduction No optimization or adjustment effect Example 2: Wind Power Cluster Collaborative Operation and Supervision Scenario Operating conditions: In the contiguous wind power cluster in Jiuquan, Gansu, multiple wind farms simultaneously lowered their bid electricity prices, aiming to reduce the market clearing price across the region in order to seize power generation space; This system uses a federated learning game model to simultaneously calculate the power output correlation coefficient and price similarity of the stations. When both values ​​exceed the judgment threshold, the CBI index falls into the secondary disposal range. The system automatically splits the aggregator's merged transaction unit and simultaneously generates a warning report for interviews and pushes it to the regulators. Traditional monitoring systems only read the overall transaction data of the aggregator and cannot identify abnormal behavior such as subordinate stations coordinating to adjust prices. There are no risk alarms output throughout the process. Example 3: Comparative Implementation of Cross-Sectional Blockage Responsibility Apportionment Operating condition: The power limit of the photovoltaic power collection and transmission section in the Chaidamu Basin of Qinghai Province has been exceeded, and the single photovoltaic power station is the main contributor to the blockage of the section; The traditional postage stamp method of cost allocation: the congestion cost is allocated only according to the amount of electricity generated. The allocation amount of this station is far higher than its actual contribution to the cross section. The result of the responsibility determination does not match the actual operating conditions of the power grid. The differential game calculation logic of this invention is as follows: based on power flow sensitivity and ramping constraints, the actual congestion contribution rate of the station is solved. After matching the 1.5 times penalty sharing rule, the final sharing amount is highly consistent with the actual physical contribution of the station, and the calculation result deviates very little from the actual operating conditions. Implementation Principle Explanation In the three sets of embodiments described above, the performance of this system is comprehensively superior to that of existing technologies, and the core technical logic support is divided into four layers: 1. The volatility energy equivalent conversion algorithm breaks down the data barrier between second-level volatility and fifteen-minute market clearing cycles, enabling it to capture early warning signals of price collapses and solving the problem that traditional tools can only use average data over the entire period and have delayed predictions. 2. The federated learning game framework, combined with the CBI dual-dimensional judgment criteria, extends the monitoring granularity down to the independent site level. It can identify human collusion without retrieving the site's private operation data, eliminating the blind spots in the traditional aggregator-level monitoring. 3. The differential Stackelberg game framework, combined with the Shapley-entropy weight composite allocation algorithm, simultaneously takes into account the power flow constraints of the grid and the ramp adjustment limits of new energy units, so as to realize the fine quantification of the responsibility for station congestion and avoid the responsibility mismatch caused by the traditional average allocation mode. 4. The time-varying Copula-ES model uniformly depicts the dynamic correlation between spot and green certificate market prices, updates return and risk indicators minute by minute, and automatically generates hedging schemes, thereby mitigating the operational impact caused by the superposition and transmission of risks from multiple markets from the source. In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. The above description is merely an example and illustration of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

Claims

1. A smart risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, characterized in that, include: Multi-source time-series data base module, dynamic risk rule base management module, multi-model parallel risk calculation engine module, hierarchical early warning output module, site-level risk tracing and analysis module, and cross-market revenue joint prevention and control module; The multi-source time-series data base module is used to collect multi-dimensional time-series data of a high proportion of new energy power market and construct a cross-time scale data mapping channel. The dynamic risk rule base management module is used to maintain new energy-specific risk indicators, dual-layer thresholds and five-level risk judgment rules and adaptively calibrate them. The multi-model parallel risk calculation engine module integrates four types of theoretical models for parallel computation, outputting four types of quantitative risk indicators: price, market power, congestion, and return. The tiered early warning output module matches the risk level, outputs tiered alarms, and automatically executes price circuit breaker control logic. The site-level risk tracing and analysis module completes risk source location and responsibility quantification before, during, and after the event. The cross-market return joint prevention and control module dynamically generates financial hedging schemes based on multi-market correlation models to block cross-market risk transmission.

2. The intelligent risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in claim 1, is characterized in that... The multi-source time-series data base module collects data including minute-level wind and solar power forecasts, unit ramp-up rates, weather warnings, market declaration and clearing, cross-sectional tidal currents, green certificates, and carbon trading time-series data. The module has a built-in fluctuation energy equivalent conversion algorithm to perform integral calculations on second-level wind and solar power output fluctuations, converting them into equivalent power deviations that adapt to a fifteen-minute market clearing cycle, establishing a second-level to fifteen-minute cross-scale data mapping channel, and eliminating the time-domain mismatch between instantaneous fluctuations in new energy and the market clearing cycle.

3. The intelligent risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in claim 1, is characterized in that... The dynamic risk rule base management module incorporates three new energy-specific indicators: wind and solar forecast deviation, new energy cluster synergy index (CBI), and cost-per-kilowatt-hour coverage ratio (DCR). It establishes a two-layer threshold system of warning yellow light and severe red light, as well as five-level risk mapping rules: no risk, low risk, medium risk, high risk, and severe risk. It is equipped with a threshold dynamic calibration algorithm that iteratively updates thresholds based on three historical typical scenarios: peak wind and solar activity, stable conditions, and off-peak periods, to achieve adaptive adjustment based on seasons and new energy penetration rates.

4. The intelligent risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in claim 1, is characterized in that... The multi-model parallel risk calculation engine module contains four independent parallel computing units: The price disorder prediction unit uses an improved Ornstein-Uhlenbeck nonstationary stochastic process model that incorporates cloud cover / sudden wind jump terms to quantify the impact of wind and solar fluctuations on electricity prices. The market power identification unit constructs a three-party multi-agent evolutionary game model involving wind and solar aggregators, power grids, and users. It identifies cluster collaborative manipulation behavior through two criteria: Nash equilibrium deviation, power station output correlation coefficient ρ greater than 0.8, and bid similarity δ greater than 0.

75. The congestion responsibility calculation unit, based on the differential Stackelberg game framework, coupled the cross-sectional power flow sensitivity and the unit ramp deviation to solve the station congestion contribution rate, and adopted the Shapley-entropy weight composite algorithm to allocate the congestion cost; The cross-market return quantification unit uses a time-varying Copula function to establish a joint distribution of spot electricity price and green certificate price, and combines it with conditional value at risk (ES) to complete the return risk quantification.

5. The intelligent risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in claim 4, is characterized in that... The congestion responsibility calculation unit is configured with penalty apportionment rules. When the congestion contribution rate of a station exceeds the benchmark threshold, it is automatically determined that the station shall bear 1.5 times the standard congestion cost.

6. The intelligent risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in claim 1, is characterized in that... The five risk levels of the graded early warning output module correspond one-to-one with the wind and solar prediction deviation: the deviation in the first interval is a no-risk green light, the deviation in the second interval is a low-risk blue light, the deviation in the third interval is a medium-risk yellow light, the deviation in the fourth interval is a high-risk orange light, and the deviation in the fifth interval is a severe-risk red light. The module has built-in dynamic circuit breaker logic. When the price fluctuation exceeds the preset fluctuation boundary within 15 minutes, it automatically suspends market trading and activates the backup trading channel.

7. The intelligent risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in claim 1, is characterized in that... The station-level risk tracing and analysis module includes a three-stage tracing logic: pre-event tracking of reserve gaps caused by sudden changes in wind and solar power output, real-time monitoring of abnormal deviations in new energy entity quotations during the event, and post-event quantification of economic losses caused by events such as negative electricity prices; it is equipped with two sub-units: special tracing of blocked sections and attribution of abnormal revenue, and outputs a visualized report on the dual responsibility ratio at the unit level and section level.

8. The intelligent risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in claim 4, is characterized in that... The cross-market return joint prevention and control module is equipped with a mean-CVaR derivative hedging optimization model and a green certificate trading return simulator. When the conditional value of risk (ES) output by the cross-market return quantification unit is lower than the safety threshold, the hedging optimization calculation is automatically started, and suitable futures and options hedging combination schemes are pushed to market participants to reduce the impact of the linkage fluctuations in the spot, green certificate and carbon markets.

9. The intelligent risk prevention and control system for power market entities in provinces with a high proportion of renewable energy, as described in claim 1, is characterized in that... The system hardware is a multi-core high-performance server cluster, equipped with a distributed time-series database and a multi-GPU parallel computing scheduling unit. The time taken for a single round of simultaneous calculation of four types of risks in the entire market does not exceed sixty seconds, supporting the minute-level real-time risk monitoring computing needs.

Images