Multi-level real-time risk control management method and device for power transaction

Abstract

The invention provides a multi-level real-time risk control management method and device for power transaction, relates to the technical field of power market transaction, and solves the technical problems that an existing power transaction risk control system is lagged in response and poor in adaptability and lacks a full-process closed-loop management mechanism. The method comprises the following steps: before transaction execution, performing simulation deduction on market conditions and transaction behaviors based on a strategy model to obtain a deduction result; setting risk limit parameters according to the deduction result; in the transaction execution process, risk index data from the power transaction platform are monitored in real time, the risk index data are compared with the risk limiting parameters, and when it is judged that a risk event occurs, a predefined intervention response strategy is executed; and after the transaction is completed, comparing the actual transaction data with the beforehand plan data, carrying out deviation attribution analysis, obtaining an analysis result, and generating a traceable audit log. The method is used in a power transaction risk control management process.

Description

Technical Field

[0001] This application relates to the field of electricity market trading technology, and in particular to a multi-level real-time risk control management method and device for electricity trading. Background Technology

[0002] As the electricity market accelerates its transition from medium- and long-term transactions to real-time spot transactions, the frequency of market transactions has increased significantly, and price volatility has surged, posing unprecedented challenges to real-time risk management for electricity retailers. Currently, the industry's commonly used risk control methods rely primarily on manual monitoring and post-event analysis, exhibiting significant lag and passivity. Existing technical solutions are mostly based on fixed threshold rules for early warning, lacking the ability to dynamically adapt to market conditions and struggling to achieve effective intervention in the millisecond-level high-frequency trading environment. Furthermore, the existing risk control process suffers from a disconnect between pre-event simulation, in-event intervention, and post-event analysis, failing to form a closed-loop management system. This results in low risk handling efficiency and reliance on manual experience in anomaly attribution, lacking scientifically quantifiable analytical methods. These limitations make it difficult for electricity retailers to achieve truly real-time, accurate, and intelligent risk control in the face of a rapidly changing electricity market. Summary of the Invention

[0003] This application provides a multi-level real-time risk control management method and device for power trading, which solves the technical problems of existing power trading risk control systems, which suffer from slow response, poor adaptability and lack of a closed-loop management mechanism for the entire process due to excessive reliance on manual monitoring and static rules.

[0004] To achieve the above objectives, this application adopts the following technical solution: Firstly, a multi-layered real-time risk control management method for electricity trading is provided, including: S1: Before the trade is executed, the market conditions and trading behavior are simulated and deduced based on the strategy model to obtain the simulation results; risk limit parameters are set according to the simulation results. S2: During the transaction execution process, the risk indicator data from the power trading platform is monitored in real time, and the risk indicator data is compared with the risk limit parameters. When a risk event is determined to occur, a predefined intervention response strategy is executed. S3: After the transaction is completed, compare the actual transaction data with the pre-planned data, perform deviation attribution analysis, obtain the analysis results, and generate a traceable audit log; Steps S1, S2, and S3 form a closed-loop process, with the analysis results of step S3 being fed back to step S1 to optimize risk limit parameters.

[0005] Based on the above technical solution, the multi-level real-time risk control management method for power trading provided in this application achieves a significant leap in power trading risk management capabilities by constructing a closed-loop risk control management method that integrates "simulation deduction - real-time monitoring - attribution analysis." This method organically links pre-emptive risk control based on model deduction, real-time comparison and automatic intervention of risk indicators during the process, and post-event deviation attribution based on data-driven analysis. It also dynamically optimizes pre-emptive parameters through post-event analysis results, forming an intelligent risk control closed loop capable of self-learning and continuous improvement. This mechanism effectively overcomes the lag and static defects of traditional risk control, significantly improving the real-time and automation levels of risk identification and handling. Furthermore, through closed-loop feedback, it achieves precise and adaptive adjustment of risk control strategies, thus providing power sales companies with forward-looking, proactive, and continuously optimized security guarantees in the high-frequency, high-volatility power spot market.

[0006] In conjunction with the first aspect above, in one possible implementation, the types of risk events include price volatility risk, forecasting bias risk, credit and performance risk, and abnormal trader behavior risk; The aforementioned price volatility risk is mitigated by identifying abnormal fluctuations through real-time market price tracking. The predicted deviation risk is identified by calculating the deviation rate between the predicted load or output value and the actual value; The aforementioned credit and performance risks are identified by monitoring margin deposits, credit lines, and electricity bill collection to identify potential defaults. The aforementioned risks of abnormal trader behavior are identified by analyzing the frequency of operations, order placement patterns, and the degree of price deviation.

[0007] In conjunction with the first aspect above, in one possible implementation, the method for identifying the risk of abnormal trader behavior includes: Construct a behavioral knowledge graph with traders and accounts as nodes and trading relationships as edges; Graph neural network clustering algorithms are used to analyze behavioral knowledge graphs and identify abnormal transaction patterns. When behavior that deviates significantly from the normal pattern is detected, a risk of abnormal trader behavior is generated.

[0008] In conjunction with the first aspect above, in one possible implementation, the risk constraint parameter is a dynamic risk threshold; wherein the dynamic adjustment of the risk threshold is modeled as a Markov decision process, including: Based on a reinforcement learning model, the risk threshold is adaptively optimized by taking the current market volatility index, risk exposure level and trader behavior score as state input, threshold adjustment action as output, and comprehensive evaluation of risk control effect and business impact as reward signal.

[0009] In conjunction with the first aspect above, in one possible implementation, the intervention response strategy, based on the severity of the risk event, includes at least one of the following: Automatic interruption: Reject or cancel high-risk trading orders; Risk mitigation: Dynamically reduce trading limits or restrict trading strategies; Early warning notification: Send alert messages to traders and risk control personnel; Multi-level approval: Submit high-risk operations to different levels of manual review; Strategy Adjustment: Automatically executes adjustments to the trading portfolio, such as reducing positions or hedging.

[0010] In conjunction with the first aspect above, in one possible implementation, the bias attribution analysis employs causal inference techniques, including: Identify potential risk factors related to risk events; Construct a causal relationship model to characterize the causal association between potential risk factors and risk indicator data; The causal contribution of potential risk factors to risk bias is quantified through intervention calculations or counterfactual simulations; wherein the risk bias is the difference between risk indicator data and risk threshold. Generate an attribution report that identifies the main causes of risk deviations and their percentage impact.

[0011] In conjunction with the first aspect above, in one possible implementation, the closed-loop process includes: Real-time transaction data triggers abnormal behavior monitoring; When a risk is detected, the risk threshold is dynamically adjusted within milliseconds. Once the risk indicator data exceeds the risk threshold, the attribution analysis of the root cause of the risk is completed within seconds, and the attribution results are obtained. Based on the attribution results, new risk control strategies are generated and implemented through multi-objective optimization. The data after execution is fed back to the monitoring end, initiating the next iteration.

[0012] In conjunction with the first aspect above, in one possible implementation, the method is executed by a dedicated risk control device deployed on the internal network of the electricity sales company. This device adopts an integrated hardware and software architecture, using a high-performance industrial server as the hardware foundation to ensure low-latency data processing in a private environment.

[0013] Secondly, a multi-level real-time risk control management device for electricity trading is provided, including: The data acquisition interface is used to connect to the power trading platform to obtain market information, trading instructions and status data in real time; The risk control analysis engine is connected to the data acquisition interface and is used to process data, calculate risk indicator data, and identify risk events. The strategy execution module is connected to the risk control analysis engine and is used to execute corresponding intervention response strategies based on risk events. The human-computer interaction interface is used to display risk information, receive parameter configurations, and issue early warning notifications.

[0014] In conjunction with the second aspect above, in one possible implementation, the risk control analysis engine integrates: The behavior graph engine is used to connect transaction entities through a knowledge graph and identify abnormal behavior patterns in real time. A reinforcement learning regulator is used to dynamically optimize risk thresholds based on market conditions. Causal attribution engine, used to analyze the root causes of risk deviations and quantify the contribution of key factors; A multi-objective optimization engine is used to generate Pareto optimal strategies that coordinate risk control, return, and liquidity objectives.

[0015] Thirdly, an electronic device is provided, comprising: a communication unit and a processing unit; the communication unit is used to acquire market information, trading instructions and status data in real time; the processing unit is used to process data, calculate risk indicator data and identify risk events; execute corresponding intervention response strategies according to risk events; display risk information, receive parameter configurations and issue early warning notifications.

[0016] Fourthly, this application provides an electronic device, including: a processor and a storage medium; the storage medium includes instructions, and the processor is configured to execute the instructions to implement the methods described in the first aspect and any possible implementation thereof. The electronic device may be an electronic device or a chip within an electronic device.

[0017] Fifthly, this application provides a multi-level real-time risk control management system for power trading, comprising: a data acquisition interface, a risk control analysis engine, a strategy execution module, and a human-computer interaction interface; wherein, the data acquisition interface is used to connect to the power trading platform to acquire market conditions, trading instructions, and status data in real time; the risk control analysis engine is connected to the data acquisition interface and is used to process data, calculate risk indicator data, and identify risk events; the strategy execution module is connected to the risk control analysis engine and is used to execute corresponding intervention response strategies based on risk events; the human-computer interaction interface is used to display risk information, receive parameter configurations, and issue early warning notifications.

[0018] In a sixth aspect, this application provides a computer-readable storage medium storing instructions that, when executed on an electronic device, cause the electronic device to perform the methods described in the first aspect and any possible implementation thereof.

[0019] In a seventh aspect, this application provides a computer program product containing instructions that, when run on an electronic device, cause the electronic device to perform the methods described in the first aspect and any possible implementation thereof.

[0020] This application provides a multi-layered real-time risk control management method and device for power trading. By constructing a multi-layered real-time risk control system with a closed-loop process covering "pre-trade, during-trade, and post-trade," it fundamentally improves power trading risk management. This method utilizes behavioral graph technology and a graph anomaly detection mechanism to achieve intelligent and correlated identification of abnormal trader behavior, significantly enhancing the ability to discover complex risk patterns. Through a reinforcement learning-driven dynamic threshold adjustment mechanism, the risk control strategy becomes adaptive, enabling real-time optimization of risk tolerance amidst market fluctuations, balancing risk control and trading efficiency. A causal inference engine traces the root causes of risk deviations, providing interpretable and quantifiable decision-making basis for strategy adjustments. The overall system is based on a privately deployed integrated hardware and software system, achieving millisecond-level risk monitoring and response, completely overcoming the lag problem of traditional manual risk control. This solution not only significantly improves the real-time performance, accuracy, and automation level of risk control but also strengthens enterprises' performance capabilities and compliance guarantees in the high-frequency power spot market, providing power sales companies with an end-to-end intelligent risk control closed-loop solution.

[0021] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0022] Figure 1 A system architecture diagram of a multi-level real-time risk control management system for power trading is provided in the embodiments of this application; Figure 2 A flowchart illustrating a multi-level real-time risk control management method for power trading, provided as an embodiment of this application; Figure 3 A flowchart illustrating a risk graph-based identification mechanism for abnormal trader behavior provided in this application embodiment; Figure 4 A flowchart illustrating another multi-level real-time risk control management method for power trading provided in this application embodiment; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application; Figure 6 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0023] In the description of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. The "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" means one or more, and "multiple" means two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0024] It should be noted that, in this application, the terms "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0025] The multi-level real-time risk control management method for power trading provided in this application embodiment can be applied to, for example... Figure 1 In the multi-level real-time risk control management system 100 for electricity trading shown, such as Figure 1 As shown, the communication system includes: a data acquisition interface 10, a risk control analysis engine 20, a strategy execution module 30, and a human-computer interaction interface 40.

[0026] Among them, data acquisition interface 10 is used to connect to the power trading platform to obtain market conditions, trading instructions and status data in real time; The risk control analysis engine 20 is connected to the data acquisition interface 10 and is used to process data, calculate risk indicator data, and identify risk events. The strategy execution module 30 is connected to the risk control analysis engine 20 and is used to execute corresponding intervention response strategies based on risk events. The human-computer interaction interface 40 is used to display risk information, receive parameter configurations, and issue early warning notifications.

[0027] To address the technical problems of existing power trading risk control systems, which rely excessively on manual monitoring and static rules, resulting in slow response times, poor adaptability, and a lack of a closed-loop management mechanism throughout the entire process, this application provides a multi-level real-time risk control management method for power trading. This method includes: S1: Before the trade is executed, the market conditions and trading behavior are simulated and deduced based on the strategy model to obtain the simulation results; risk limit parameters are set according to the simulation results. S2: During the transaction execution process, the risk indicator data from the power trading platform is monitored in real time, and the risk indicator data is compared with the risk limit parameters. When a risk event is determined to occur, a predefined intervention response strategy is executed. S3: After the transaction is completed, compare the actual transaction data with the pre-planned data, perform deviation attribution analysis, obtain the analysis results, and generate a traceable audit log; Steps S1, S2, and S3 form a closed-loop process, with the analysis results of step S3 being fed back to step S1 to optimize risk limit parameters.

[0028] Based on this, the technical problems of existing power trading risk control systems, which rely too heavily on manual monitoring and static rules, resulting in slow response, poor adaptability, and a lack of a closed-loop management mechanism for the entire process, have been solved.

[0029] like Figure 2 As shown in the embodiments of this application, the multi-level real-time risk control management method for power trading includes: S201. Before the transaction is executed, the market conditions and trading behavior are simulated and deduced based on the strategy model to obtain the deduction results; risk limit parameters are set according to the deduction results.

[0030] It should be noted that before the transaction is executed, the system simulates and evaluates possible market conditions and trading behaviors through strategy models to identify potential risks in advance. Based on this, corresponding risk control measures are established, including setting risk control parameters. These parameters include trading strategy boundaries and various limit thresholds (such as position limits, credit limits, and price ranges). Through pre-emptive strategy simulation, the power sales company can identify the worst-case loss scenarios under adverse conditions and adjust its trading plan and risk control parameters accordingly, keeping risks within an acceptable range.

[0031] As an example, the system can set risk thresholds and transaction limits for different scenarios based on stress test results, thus constraining trading behavior under high-risk conditions. Pre-emptive risk control is like putting a "safety helmet" on transactions, modeling and limiting risk exposure at the source, and preventing significant risks from being introduced during the strategy formulation stage.

[0032] In some implementations, the risk constraint parameter is a dynamic risk threshold; wherein the dynamic adjustment of the risk threshold is modeled as a Markov decision process, including: Based on a reinforcement learning model, the risk threshold is adaptively optimized by taking the current market volatility index, risk exposure level and trader behavior score as state input, threshold adjustment action as output, and comprehensive evaluation of risk control effect and business impact as reward signal.

[0033] It should be noted that the dynamic risk threshold control mechanism integrating reinforcement learning can automatically adjust risk control thresholds based on real-time risk status, achieving dynamic and adaptive risk threshold management. Its function is to endow risk restrictions (such as early warning thresholds and transaction limits) with "self-learning" capabilities: when risk levels rise or market volatility intensifies, the agent can promptly tighten the thresholds to increase risk control; conversely, when risks ease, the thresholds can be appropriately relaxed to reduce interference with normal transactions. This ensures timely and effective risk control while also considering business continuity and efficiency. Traditional risk control relies on manual setting of fixed thresholds based on experience, which suffers from lag and rigidity, failing to adapt to dynamically changing black swan events or new types of violations. This mechanism, through the introduction of reinforcement learning for continuous trial-and-error optimization, solves the problems of lag and difficulty in balancing manual adjustments. Especially in rapidly changing markets such as electricity trading, this module can adjust risk tolerance in real time according to market changes, preventing risk exposure or business interruption due to inappropriate thresholds.

[0034] As an example, this mechanism models risk threshold adjustment as a Markov decision process: a reinforcement learning agent continuously decides on threshold adjustment strategies based on the environmental state. Specifically, the state includes current risk environment characteristics, such as market volatility indicators, portfolio risk exposure, and trader behavior scores; the action is the threshold adjustment operation, such as raising or lowering a risk limit, or modifying warning trigger conditions; the strategy is provided by the agent's decision network; and the reward feedback is set based on the adjusted effect, including factors such as whether the risk event was successfully avoided, the number of false positives and false negatives, and the impact on normal trading. At each moment, the reinforcement learning agent reads the current state, calculates the optimal threshold adjustment action through the policy network, and sends it to the risk control engine for execution. After applying the new threshold, the risk control engine observes the risk outcomes over time (such as changes in subsequent risk loss rates or business success rates), calculates reward / penalty signals based on this, and feeds them back to the agent to continuously adjust its strategy.

[0035] Based on the above technical solution, through this closed-loop training of decision-feedback-optimization, the agent gradually learns how to adjust thresholds to achieve the best balance between risk and return under different risk scenarios. When the environment changes (such as the emergence of new fraud methods or increased market volatility), the agent can quickly readjust the threshold settings based on the new feedback, ensuring that risk control is always in a dynamically optimal state.

[0036] S202. During the transaction execution process, the risk indicator data from the power trading platform is monitored in real time, and the risk indicator data is compared with the risk limit parameters. When a risk event is determined to occur, a predefined intervention response strategy is executed.

[0037] It's important to note that during transaction execution, the system monitors key risk indicators in real-time, promptly identifying abnormal changes and triggering alerts or control actions. Specifically, the risk control analysis engine continuously tracks data such as electricity market conditions, transaction order execution, and its own asset output, comparing these against preset threshold rules to calculate real-time risk indicator values ​​and detect anomalies. When an indicator exceeds a threshold or exhibits an abnormal pattern, the system immediately implements in-process risk control measures, including issuing warnings to traders and risk control personnel, and automatically intercepting or adjusting relevant transaction operations according to predefined rules. For example, when a transaction node experiences sharp fluctuations in electricity prices or a significant increase in enterprise power forecast deviations, the system automatically triggers an alert and suggests reducing exposure; if the risk continues to escalate, it directly suspends new transaction orders or closes some positions to prevent further losses. Through the real-time and proactive nature of in-process risk control, the system overcomes the weakness of previous manual monitoring, significantly improving the timeliness and effectiveness of risk handling.

[0038] As an example, to achieve dynamic management of various risks, this application designs a real-time risk indicator monitoring system and triggering mechanism. The system continuously calculates and updates several key risk indicators, including but not limited to: risk exposure (position exposure), i.e., the net exposure size of an enterprise in different markets and time periods, used to measure the potential profit and loss exposure caused by price fluctuations; deviation rate, used to measure the degree of deviation between contract execution and forecast plans; overload, used to assess whether the utilization rate of a certain resource or quota (such as load capacity, credit line) is close to or exceeds the limit; operation frequency, used to monitor excessively frequent or abnormal trading operations, etc. Each risk indicator corresponds to a pre-set threshold or rule. When the indicator value exceeds the threshold or meets specific abnormal conditions, the system immediately triggers the corresponding risk control event. To improve the accuracy and intelligence of the triggering mechanism, this application supports multi-dimensional threshold management and dynamic adjustment: based on different trading instruments, market types, and time periods, differentiated limit standards and alarm thresholds are automatically adopted. For example, during peak trading hours or for instruments with drastic price fluctuations, the system can automatically tighten the risk control threshold; conversely, it can appropriately loosen it during stable periods to avoid excessive intervention in normal trading. This flexible configuration of indicator thresholds allows the risk control triggering mechanism to align with actual market conditions, improving the sensitivity and accuracy of early warnings. When the triggering conditions are met, the system will execute triggering actions according to predefined correspondences, including issuing early warning notifications, recording event logs, and invoking intervention modules to implement further risk management.

[0039] Based on the above technical solution, through real-time indicator monitoring and automatic triggering linkage, this application can achieve "responding as soon as a risk is detected", shortening the risk control response time to the millisecond level, which is far ahead of the minute or even hour delay of traditional manual monitoring, thus winning valuable time for risk prevention.

[0040] In some implementations, the types of risk events include price volatility risk, forecasting bias risk, credit and performance risk, and abnormal trader behavior risk; The aforementioned price volatility risk is mitigated by identifying abnormal fluctuations through real-time market price tracking. The predicted deviation risk is identified by calculating the deviation rate between the predicted load or output value and the actual value; The aforementioned credit and performance risks are identified by monitoring margin deposits, credit lines, and electricity bill collection to identify potential defaults. The aforementioned risks of abnormal trader behavior are identified by analyzing the frequency of operations, order placement patterns, and the degree of price deviation.

[0041] As an example, this application establishes a comprehensive risk identification mechanism to monitor and classify various risk types that may arise in power trading operations, including but not limited to the following: Price volatility risk: This refers to the risk of uncertainty in trading profits due to significant fluctuations in market electricity prices. Spot wholesale market prices change rapidly; electricity retailers may suffer losses if they lack power supply during periods of high prices or hold too many high-priced contracts during periods of sharp price drops. This system tracks real-time market prices and price forecast data to identify abnormal price fluctuations. When a sharp rise or fall in prices within a short period is detected, it is marked as a price risk warning to prompt timely adjustments to positions or hedging.

[0042] Forecast deviation risk refers to the risk arising from a mismatch between contracted and actual electricity volumes due to inaccurate load or output forecasts, including the risk of penalties for deviation assessments. In electricity trading, deviations between the hourly output curve and the contract curve create exposure risk in the spot market. Furthermore, many spot markets require generators to submit their generation plans in advance, and exceeding the allowable deviation range will result in penalties. Therefore, electricity retailers face the dual risks of "misalignment between contracted and actual curves" due to forecast errors and deviation fees. This application calculates the deviation rate by continuously comparing the forecasted load value with the real-time actual value, promptly identifying situations where forecast deviations are too large. Once the deviation rate exceeds a threshold, the system will trigger a risk warning or automatically adjust the electricity purchase plan to control the difference between actual and planned loads within an allowable range, thereby reducing the economic risks caused by forecast uncertainty.

[0043] Credit and performance risk refers to the risk of default by the parties involved in a transaction regarding contract performance and fund settlement. This includes situations where the electricity sales company itself or its customers are unable to fulfill their contractual obligations due to credit issues, or defaults due to insufficient margin or a broken capital chain. Such risks can have a cascading impact on market transactions, and regulatory agencies have established corresponding mechanisms (such as guaranteed electricity sales measures to take over the users of defaulting electricity sales companies) to prevent systemic risks. This system monitors credit indicators such as the electricity sales company's transaction margin, credit line usage, and customer electricity fee collection to promptly identify potential credit crises. For example, when it is discovered that a customer has been continuously in arrears on electricity bills or that a trading seat's margin is approaching a threshold, the system will issue a credit risk warning and can restrict further trading operations of that entity according to preset rules to prevent default events. By incorporating credit and performance factors into the risk control monitoring scope, the system ensures that transaction contracts are fulfilled and fund settlements are completed safely. Early intervention can address any abnormal signs, thereby maintaining market stability and protecting the rights and interests of all parties.

[0044] Abnormal trader behavior refers to the risks arising from excessively frequent, unconventional, or potentially illegal trading activities during manual trading. Human factors in trading can lead to emotional decision-making, erroneous operations, and even illegal profit-seeking. Differences in strategy choices among different traders can also result in vastly different outcomes. This application incorporates a trading behavior recognition module that analyzes behavioral characteristics such as trader operation frequency, order placement patterns, order cancellation frequency, and price deviation to identify abnormal trading behavior. For example, the system can detect a trader submitting a large number of high-priced buy orders and then quickly canceling them within a short period, or identify abnormal orders where the price deviates significantly from the reasonable range and classify them as suspicious. Once such anomalies are detected, the system will immediately trigger a risk control response: minor anomalies will warn the trader and be recorded; serious anomalies will automatically reject the relevant orders and notify risk control management personnel for review and processing. By monitoring and restricting inappropriate trader behavior, "human operational risks" can be effectively prevented, avoiding losses and compliance risks to the company due to operational errors or violations.

[0045] In some implementations, the methods for identifying the risk of abnormal trader behavior include: Construct a behavioral knowledge graph with traders and accounts as nodes and trading relationships as edges; Graph neural network clustering algorithms are used to analyze behavioral knowledge graphs and identify abnormal transaction patterns. When behavior that deviates significantly from the normal pattern is detected, a risk of abnormal trader behavior is generated.

[0046] As an example, the flowchart of the abnormal trader behavior risk graph identification mechanism is as follows: Figure 3As shown in the diagram, this mechanism first connects to real-time transaction data and historical behavioral data. After data cleaning and fusion, a behavioral knowledge graph covering traders and their trading relationships is constructed. In this graph, nodes represent entities such as traders and accounts, and edges represent transaction relationships or interaction attributes. Subsequently, a graph anomaly detection model (such as an algorithm based on graph neural network clustering analysis) is used to intelligently analyze the graph and uncover "abnormal" patterns and abnormal associations. For example, the model can identify abnormally high-frequency transactions between a trader and a specific account, or discover abnormally "isolated" or abnormally close subgraph structures in the trading network. Finally, when behavior that significantly deviates from the normal pattern is detected, the system generates an abnormal behavior alarm, pushes it to risk control personnel, or automatically triggers corresponding risk control measures. Compared to traditional solutions that rely on threshold rules and manual monitoring, the innovation of this mechanism lies in introducing graph analysis and machine learning to identify complex correlation risks. Traditional risk control often monitors single accounts or single indicators exceeding limits according to preset rules, making it difficult to detect hidden abnormal behaviors across accounts and trading relationships in a timely manner. This mechanism connects scattered information through a knowledge graph and uses graph algorithms to capture abnormal relationships and transaction patterns, enabling associative and intelligent identification of abnormal behavior. For example, it can automatically identify potential chains of profit transfer or abnormal profit and loss patterns among traders, breaking through the limitations of human experience from the perspective of intelligent reasoning.

[0047] Based on the above technical solution, this mechanism integrates cutting-edge graph neural network technology, which can learn implicit patterns from massive transaction relationship data, "see through" abnormal connections and behavioral characteristics that are difficult to detect by traditional methods, and significantly improve the intelligence level of monitoring.

[0048] In some implementations, the intervention response strategy, based on the severity of the risk event, includes at least one of the following: Automatic interruption: Reject or cancel high-risk trading orders; Risk mitigation: Dynamically reduce trading limits or restrict trading strategies; Early warning notification: Send alert messages to traders and risk control personnel; Multi-level approval: Submit high-risk operations to different levels of manual review; Strategy Adjustment: Automatically executes adjustments to the trading portfolio, such as reducing positions or hedging.

[0049] As an example, this application pre-defines multi-level and diversified intervention and response strategies for different risk events and their severity, ensuring that risks can be controlled in their nascent stage and handled step by step as appropriate: Automatic Interruption: When risk indicators exceed critical thresholds or an emergency risk situation occurs, the system will immediately and automatically interrupt relevant trading operations. This includes refusing to execute new trading orders, canceling unexecuted orders, and suspending the trading channel if necessary. The automatic interruption mechanism is equivalent to quickly applying the brakes when significant risks are detected. For example, when the system detects a significant deviation between spot market declaration data and actual electricity consumption, exceeding compliance limits, it can automatically block subsequent similar declarations and notify the trading institution to explain the situation. This mandatory measure prevents further escalation of risk events, especially providing immediate interception of major errors or malicious operations, thus ensuring transaction security.

[0050] Risk Downgrading: For risks that reach the warning level but have not yet reached a fatal level, the system employs a downgrading strategy, dynamically reducing trading permissions and strategy intensity based on the risk level. Specifically, when moderate risk is detected, the system automatically reduces the available trading volume and leverage level of the trader or account, for example, reducing their maximum position size by a certain percentage or restricting them to only executing conservative strategies. Risk downgrading may also manifest as switching the automated trading algorithm to a safe mode, reducing the frequency and scale of high-risk operations. Once the risk situation has eased, the original permissions can be gradually restored under manual evaluation. Through this downgrading mechanism, the system controls potential losses to a lower range, essentially "slowing down" trading in advance, preventing continued high-speed trading under high-risk conditions.

[0051] Early Warning Notification: Regardless of the risk level, real-time early warning is one of the system's fundamental response actions. Once any risk trigger condition is met, the system immediately sends alert information to relevant personnel through multiple channels, including pop-up warning windows in the trading client and SMS / email notifications to risk control managers. The early warning information details the indicators and values ​​of the risk occurrence, along with suggested countermeasures. For example, "Warning: Deviation rate reaches 15%, it is recommended to reduce intraday trading volume." Through timely early warning communication, traders can quickly understand the risk level they face in their operations, and management can simultaneously understand the overall risk situation of the enterprise, thereby taking necessary manual intervention or decision support measures. Furthermore, all early warning events are recorded for post-event analysis and continuous improvement.

[0052] Multi-level approval: For certain high-risk operations or abnormal behaviors, this application introduces a tiered approval mechanism to strengthen manual review and supervision. When the system determines that a transaction request exceeds the preset risk procedures (e.g., buy / sell orders exceeding a certain capital size, or orders with continuous abnormal quotes), the request is upgraded to a pending approval status. According to pre-set rules, different levels of risk correspond to different levels of approvers: general risks are approved by departmental risk control personnel, while higher-level risks require approval from higher management (such as the risk control director or company executives) before execution. If approval is rejected, the transaction instruction is canceled or delayed. The multi-level approval mechanism ensures that potentially high-risk operations are subject to human review and oversight, preventing significant losses to the company due to errors or overstepping of authority by individual traders, while also improving compliance—important decisions are traceable and accountable.

[0053] Strategy Adjustment: When excessive risk exposure or abnormal market fluctuations are detected, the system can not only passively defend but also proactively take strategy adjustment measures to mitigate risks. For example, based on built-in algorithms, the system can automatically suggest adjustments to trading portfolios when an alert is triggered, such as reducing positions (closing some positions to reduce risk exposure) or hedging (hedging price risk through reverse transactions, purchasing standby contracts, etc.). In certain pre-authorized situations, the system can even directly execute partial adjustment actions, such as quickly selling a certain proportion of exposed electricity contracts or calling on standby ancillary service resources to balance deviations. This automated risk correction action can proactively mitigate risks in their early stages, avoiding missed opportunities due to information delays or slow human decision-making. For example, when market prices suddenly fluctuate far beyond expectations, the system immediately suggests and executes hedging transactions related to that price to lock in price risk, thereby minimizing potential losses. Strategy adjustment measures provide enterprises with a toolbox of proactive defense, enabling the risk control system not only to "respond to events as they arise" but also to "preemptively strike."

[0054] Based on the aforementioned technical solution, various intervention and response strategies can be activated individually or in combination according to the risk scenario, forming a layered and progressive defense: from warnings and alerts to operational restrictions, and then to manual review and proactive correction, gradually escalating the control measures. This mechanism ensures that regardless of whether the risk stems from drastic market fluctuations, forecasting errors, or operational anomalies, this application can respond quickly and take appropriate measures to contain the risk within a controllable range. This not only protects the immediate interests of the electricity sales company but also safeguards the overall stability and order of market transactions.

[0055] S203. After the transaction is completed, compare the actual transaction data with the pre-planned data, perform deviation attribution analysis, obtain the analysis results, and generate a traceable audit log.

[0056] Steps S201, S202, and S203 form a closed-loop process, with the analysis results of step S203 being fed back to step S201 to optimize risk limit parameters.

[0057] It's important to note that after transaction completion and settlement, the system enters the post-transaction risk control phase, conducting deviation analysis and audit traceability of the transaction process and results. First, the system collects actual execution data and compares it with pre-planned or predicted data, calculating indicators such as deviation rates and analyzing the causes of deviations. For example, it distinguishes whether the deviation is due to load / generation forecast errors, abnormal market prices, or operational mistakes. The system performs attribution analysis on the deviations and generates risk reports to help companies identify key risk factors. For significant deviation events exceeding thresholds, the system supports post-transaction audit playback, which recreates the entire process of transaction and risk control decisions based on log records for review by management or regulatory authorities. This audit playback helps identify loopholes in existing risk control rules, provides a basis for improvement, and can be used to train trading personnel to enhance risk awareness. Through this closed-loop feedback of post-transaction risk control, the risk management system can continuously optimize itself: on the one hand, lessons learned from deviations are fed back into the pre-transaction model and rule base to improve risk control strategies; on the other hand, complete risk handling records are preserved to meet the needs of compliance audits and liability determination. In summary, post-event risk control ensures that every risk event can be seen and analyzed, thereby turning it into an opportunity to improve risk control capabilities and enabling the risk control system to continuously evolve and improve.

[0058] In some implementations, the bias attribution analysis employs causal inference techniques, such as... Figure 4 As shown, it includes: Identify potential risk factors related to risk events; Construct a causal relationship model to characterize the causal association between potential risk factors and risk indicator data; The causal contribution of potential risk factors to risk bias is quantified through intervention calculations or counterfactual simulations; wherein the risk bias is the difference between risk indicator data and risk threshold. Generate an attribution report that identifies the main causes of risk deviations and their percentage impact.

[0059] It's important to note that when risk indicators exceed thresholds or exhibit abnormal fluctuations, the system doesn't simply issue alerts. Instead, it employs causal inference methods to identify the key factors causing the risk deviation. Its role is to answer the question "Why did the risk deviation occur?", identifying the underlying risk drivers and providing a basis for risk mitigation and decision-making. Compared to traditional correlation-based analysis (which easily mistake coincidence for causation), this technology can more deeply explain whether risk changes are directly caused by specific events or factors. For example, it can distinguish whether abnormal electricity price fluctuations stem from changes in market supply and demand or trader strategy errors, thus providing scientific support for managers to develop targeted risk control strategies.

[0060] As an example, once the risk monitoring module detects an abnormal deviation in a key risk indicator, causal attribution analysis is immediately triggered. First, data on potential factors related to the risk event are collected, including market conditions, trader behavior, policy changes, and other variables that may affect the risk indicator. Next, based on domain knowledge and data, a corresponding causal relationship model is constructed (e.g., representing the causal association between each factor and the risk outcome using a directed acyclic graph (DAG)). In this model, nodes represent independent variables (risk factors) and dependent variables (risk indicators), and directed edges represent causal influence hypotheses. Then, causal inference algorithms are used to analyze and reason about the model, including intervention calculations and counterfactual simulations, to quantify the causal effect of each factor on the risk deviation. For example, the contribution of a change in a factor to the increase in the risk indicator is calculated. Finally, a risk deviation attribution report is generated, clearly indicating the main causes of the risk deviation and their proportion of influence, providing a basis for subsequent decision-making. If necessary, the system can also perform sensitivity analysis to test the robustness of the conclusions, ensuring the reliability and credibility of the attribution results.

[0061] In some implementations, the closed-loop process is implemented in the following ways: Real-time transaction data triggers abnormal behavior monitoring; When a risk is detected, the risk threshold is dynamically adjusted within milliseconds. Once the risk indicator data exceeds the risk threshold, the attribution analysis of the root cause of the risk is completed within seconds, and the attribution results are obtained. Based on the attribution results, new risk control strategies are generated and implemented through multi-objective optimization. The data after execution is fed back to the monitoring end, initiating the next iteration.

[0062] As an example, real-time trading data first triggers the behavioral graph engine to detect anomalies (such as abnormal order placements); if a risk is detected, the reinforcement learning regulator adjusts the risk control threshold within 50ms (such as tightening limits); when the risk indicator exceeds the threshold, the causal attribution engine locates the root cause within 1 second (such as attributing it to strategy errors); the multi-objective optimization engine generates a new strategy within 800ms (such as adjusting the weight of trading rules); after execution, the data is fed back to the graph engine to start the next round of iteration, forming a closed loop of "monitoring-regulation-attribution-decision".

[0063] In some implementations, the multi-level real-time risk control management method for electricity trading is executed through a dedicated risk control device deployed on the internal network of the electricity sales company. This device adopts an integrated hardware and software architecture, using a high-performance industrial server as the hardware foundation to ensure low-latency data processing in a private environment.

[0064] As an example, the risk control management device of this application adopts an integrated hardware and software system architecture, deployed as a dedicated device within the internal network of the power sales company. The device includes a high-performance industrial server as its hardware foundation, equipped with a customized risk control management software module, enabling 24 / 7 uninterrupted operation and real-time processing. The overall system architecture includes a data acquisition interface, a risk control analysis engine, a strategy execution module, and a human-machine interface. The data interface is responsible for connecting to market data, trading instructions, and operational status information from the power trading platform, importing external market data and internal business data into the risk control system in real time. The analysis engine performs high-speed data processing and risk calculation locally, while the strategy execution module is directly embedded in the trading process, capable of intercepting or adjusting trading instructions to achieve immediate implementation of risk control rules. The entire device operates within the power sales company in a private deployment manner, ensuring data security and preventing the leakage of sensitive business information, while utilizing local computing resources to achieve low-latency, high-reliability risk control response.

[0065] Based on the above technical solutions, the multi-level real-time risk control management method for power trading provided in this application fundamentally improves power trading risk management by constructing a multi-level real-time risk control system with a closed-loop "pre-event, during-event, and post-event" process. This method utilizes behavioral graph technology and graph anomaly detection mechanisms to achieve intelligent and correlated identification of abnormal trader behavior, significantly enhancing the ability to discover complex risk patterns. Leveraging a reinforcement learning-driven dynamic threshold adjustment mechanism, the risk control strategy becomes adaptive, enabling real-time optimization of risk tolerance amidst market fluctuations, balancing risk control and trading efficiency. A causal inference engine traces the root causes of risk deviations, providing interpretable and quantifiable decision-making basis for strategy adjustments. The overall system is based on a privately deployed integrated hardware and software system, achieving millisecond-level risk monitoring and response, completely overcoming the lag problem of traditional manual risk control. This solution not only significantly improves the real-time performance, accuracy, and automation level of risk control but also strengthens enterprises' performance capabilities and compliance guarantees in the high-frequency power spot market, providing power sales companies with an end-to-end intelligent risk control closed-loop solution.

[0066] In one possible implementation, this application embodiment also provides a multi-level real-time risk control management device for power trading, comprising: The data acquisition interface is used to connect to the power trading platform to obtain market information, trading instructions and status data in real time; The risk control analysis engine is connected to the data acquisition interface and is used to process data, calculate risk indicator data, and identify risk events. The strategy execution module is connected to the risk control analysis engine and is used to execute corresponding intervention response strategies based on risk events. The human-computer interaction interface is used to display risk information, receive parameter configurations, and issue early warning notifications.

[0067] In some implementations, the risk control analysis engine integrates: The behavior graph engine is used to connect transaction entities through a knowledge graph and identify abnormal behavior patterns in real time. A reinforcement learning regulator is used to dynamically optimize risk thresholds based on market conditions. Causal attribution engine, used to analyze the root causes of risk deviations and quantify the contribution of key factors; A multi-objective optimization engine is used to generate Pareto optimal strategies that coordinate risk control, return, and liquidity objectives.

[0068] In summary, the multi-level real-time risk control management method and device for power trading provided in this application can significantly improve the risk control level of the entire power trading process, bringing the following beneficial effects: Fully Automated Closed-Loop Risk Control: This application organically combines pre-event prevention, in-event monitoring, and post-event feedback to form a closed-loop risk management process, achieving proactive and continuous control over transaction activities. Compared to traditional fragmented manual risk control, this application constructs a complete, systematic, full-process, and intelligent risk management system, enabling electricity sales companies to calmly cope with risks in a highly uncertain market.

[0069] Enhancing Risk Control Efficiency and Accuracy: Real-time data processing and intelligent rule engines significantly improve risk identification and response speed, achieving millisecond-level risk detection and handling. Simultaneously, models and big data analytics improve the accuracy of risk assessment, avoiding human oversight and misjudgments. Introducing automated risk control tools increases the speed of risk monitoring and significantly reduces the error rate caused by manual operations. The risk control process thus becomes more efficient and reliable, enabling enterprises to identify risks more promptly, quantify them accurately, and make decisions and take action more quickly.

[0070] Enhancing the precision and responsiveness of risk control: This application supports flexible and diverse risk control indicators and strategies, shifting risk control from a broad, static approach to a refined, dynamic management system. For example, differentiated threshold and limit strategies can be employed for different time periods and trading instruments, improving precision and preventing the disruption of normal business operations caused by a "one-size-fits-all" approach to risk control. Furthermore, automated system responses replace manual approvals, significantly shortening the reaction time for risk handling and transforming the approach from post-event accountability to real-time prevention, eliminating many potential losses. Enterprise risk control shifts from reactive to proactive, with a significantly faster overall response speed.

[0071] Ensuring Contractual Performance and Compliance: By comprehensively monitoring credit indicators and transaction behavior, this application can promptly detect and prevent potential defaults, ensuring that electricity sales companies fulfill their contractual obligations and reducing default risks. Simultaneously, the identification and interception of abnormal transaction behavior prevents the impact of irregular operations on market order, helping companies strictly comply with electricity spot market rules and meet regulatory requirements. The system automatically records risk control measures and transaction logs, facilitating post-event audits and regulatory reports, strengthening the company's internal control and compliance management system, and ensuring that transaction behavior remains under control at all times.

[0072] Adapting to the High-Frequency Spot Market Environment: The electricity spot market features high-frequency and fast-paced transactions, making it difficult for traditional risk control methods to intervene effectively and promptly. This application is optimized for high-frequency trading scenarios, enabling the processing of massive amounts of data instantaneously and the rapid execution of complex risk control decisions within a short time window to meet the short-term risk control needs of the spot market. This gives electricity retailers a significant advantage when participating in spot transactions, allowing them to cope with rapidly changing market conditions. The system's high performance and real-time capabilities ensure that even in a rapidly changing market, the company's risk defenses are in place instantly, providing a solid guarantee for participating in electricity market competition.

[0073] In summary, this application significantly improves the efficiency, accuracy, and response speed of risk control in electricity trading through multi-level and comprehensive real-time risk control; it constructs an automated risk control closed loop for the entire trading process, ensuring that electricity sales companies' trading activities are compliant and orderly, and that contract performance is safe and reliable; it can adapt to the high-frequency and high-volatility characteristics of the electricity spot market, safeguarding electricity sales companies in a complex and ever-changing market environment, and has significant practical value and promotional significance. Through the implementation of this application, electricity sales companies will gain stronger risk prevention capabilities and faster decision-making capabilities, thus remaining invincible in the fierce competition of the electricity market.

[0074] The foregoing mainly describes the solutions of the embodiments of this application from the perspective of device implementation. It is understood that each device, such as an electronic device, includes at least one of the hardware structures and software modules corresponding to the execution of each function in order to achieve the above-mentioned functions. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in a hardware or computer software-driven hardware manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0075] This application embodiment can divide the electronic device into functional units according to the above method example. For example, each function can be divided into a separate functional unit, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0076] When using integrated units, Figure 5 A possible structural schematic diagram of the electronic device (denoted as electronic device 50) involved in the above embodiments is shown. The electronic device 50 includes a processing unit 501 and a communication unit 502, and may also include a storage unit 503. Figure 5 The structural diagram shown can be used to illustrate the structure of the electronic device involved in the above embodiments.

[0077] when Figure 5 The schematic diagram shown is used to illustrate the structure of the electronic device involved in the above embodiments. The processing unit 501 is used to control and manage the operation of the electronic device, the communication unit 502 is used for the electronic device to communicate with other devices, and the storage unit 503 is used to store the program code and data of the electronic device.

[0078] For example, communication unit 502 is used to acquire market information, trading instructions and status data in real time; The processing unit 501 is used to process data, calculate risk indicator data and identify risk events; execute corresponding intervention response strategies based on risk events; display risk information, receive parameter configurations and issue early warning notifications.

[0079] The processing unit 501 can be a processor or a controller, and the communication unit 502 can be a communication interface, transceiver, transceiver circuit, transceiver device, etc. The term "communication interface" is a general term and may include one or more interfaces. The storage unit 503 can be a memory. When the electronic device 50 is a chip, the processing unit 501 can be a processor or a controller, and the communication unit 502 can be an input interface and / or an output interface, pins, or circuits, etc. The storage unit 503 can be a storage unit within the chip (e.g., a register, cache, etc.) or a storage unit located outside the chip (e.g., read-only memory (ROM), random access memory (RAM, etc.).

[0080] The communication unit can also be called a transceiver unit. The antenna and control circuit with transceiver functions in the electronic device 50 can be considered as the communication unit 502 of the electronic device 50, and the processor with processing functions can be considered as the processing unit 501 of the electronic device 50. Optionally, the device in the communication unit 502 used to implement the receiving function can be considered as the communication unit. The communication unit is used to execute the receiving steps in the embodiments of this application, and the communication unit can be a receiver, a receiver circuit, etc. The device in the communication unit 502 used to implement the transmitting function can be considered as the transmitting unit. The transmitting unit is used to execute the transmitting steps in the embodiments of this application, and the transmitting unit can be a transmitter, a transmitter, a transmitting circuit, etc.

[0081] Figure 5 If the integrated units in the process are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. Storage media for storing computer software products include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0082] Figure 5 The units in the process can also be called modules; for example, a processing unit can be called a processing module.

[0083] This application also provides a hardware structure diagram of an electronic device (denoted as electronic device 60), see [link to diagram]. Figure 6 The electronic device 60 includes a processor 601, and optionally, a memory 602 connected to the processor 601.

[0084] In the first possible implementation, see Figure 6 The electronic device 60 also includes a transceiver 603. The processor 601, memory 602, and transceiver 603 are connected via a bus. The transceiver 603 is used to communicate with other devices or communication networks. Optionally, the transceiver 603 may include a transmitter and a receiver. The device in the transceiver 603 that implements the receiving function can be considered as a receiver, which is used to perform the receiving steps in the embodiments of this application. The device in the transceiver 603 that implements the transmitting function can be considered as a transmitter, which is used to perform the transmitting steps in the embodiments of this application.

[0085] Based on the first possible implementation method Figure 6 The structural diagram shown can be used to illustrate the structure of the electronic device involved in the above embodiments.

[0086] in, Figure 6 This can also be illustrated by a system chip in an electronic device. In this case, the actions performed by the aforementioned electronic device can be implemented by this system chip; the specific actions performed can be found above and will not be repeated here.

[0087] In implementation, each step of the method provided in this embodiment can be completed by integrated logic circuits in the processor or by instructions in software form. The steps of the method disclosed in the embodiments of this application can be directly manifested as being executed by a hardware processor, or being executed by a combination of hardware and software modules in the processor.

[0088] The processor in this application may include, but is not limited to, at least one of the following: a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), a microcontroller unit (MCU), or an artificial intelligence processor, etc., which are various computing devices that run software. Each computing device may include one or more cores for executing software instructions to perform calculations or processing. The processor may be a separate semiconductor chip or integrated with other circuits into a single semiconductor chip. For example, it may be integrated with other circuits (such as encoding / decoding circuits, hardware acceleration circuits, or various bus and interface circuits) to form a SoC (System-on-a-Chip), or it may be integrated as a built-in processor within an ASIC. The ASIC with the integrated processor may be packaged separately or together with other circuits. In addition to the cores for executing software instructions to perform calculations or processing, the processor may further include necessary hardware accelerators, such as field-programmable gate arrays (FPGAs), PLDs (programmable logic devices), or logic circuits that implement dedicated logic operations.

[0089] The memory in the embodiments of this application may include at least one of the following types: read-only memory (ROM) or other types of static storage devices capable of storing static information and instructions; random access memory (RAM) or other types of dynamic storage devices capable of storing information and instructions; or electrically erasable programmable-only memory (EEPROM). In some scenarios, the memory may also be a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0090] This application also provides a computer-readable storage medium including instructions that, when run on a computer, cause the computer to perform any of the methods described above.

[0091] This application also provides a computer program product containing instructions that, when run on a computer, cause the computer to perform any of the methods described above.

[0092] This application also provides a chip including a processor and an interface circuit. The interface circuit is coupled to the processor. The processor is used to run computer programs or instructions to implement the above-described method. The interface circuit is used to communicate with other modules outside the chip.

[0093] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software programs, implementation can be, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device containing one or more servers, data centers, etc., that can be integrated with the medium. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state disks (SSDs)).

[0094] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, disclosure, and appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude multiple instances. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0095] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of this application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from the spirit and scope of this application. Thus, if such modifications and modifications of this application fall within the scope of the claims of this application and their equivalents, this application is also intended to include such modifications and modifications.

Claims

1. A multi-level real-time risk control management method for electricity trading, characterized in that, include: S1: Before the trade is executed, the market conditions and trading behavior are simulated and deduced based on the strategy model to obtain the simulation results; Set risk limit parameters based on the simulation results; S2: During the transaction execution process, the risk indicator data from the power trading platform is monitored in real time, and the risk indicator data is compared with the risk limit parameters. When a risk event is determined to occur, a predefined intervention response strategy is executed. S3: After the transaction is completed, compare the actual transaction data with the pre-planned data, perform deviation attribution analysis, obtain the analysis results, and generate a traceable audit log; Steps S1, S2, and S3 form a closed-loop process, with the analysis results of step S3 being fed back to step S1 to optimize risk limit parameters.

2. The multi-level real-time risk control management method for power trading according to claim 1, characterized in that, The types of risk events mentioned include price volatility risk, forecasting error risk, credit and performance risk, and abnormal trader behavior risk. The aforementioned price volatility risk is mitigated by identifying abnormal fluctuations through real-time market price tracking. The predicted deviation risk is identified by calculating the deviation rate between the predicted load or output value and the actual value; The aforementioned credit and performance risks are identified by monitoring margin deposits, credit lines, and electricity bill collection to identify potential defaults. The aforementioned risks of abnormal trader behavior are identified by analyzing the frequency of operations, order placement patterns, and the degree of price deviation.

3. The multi-level real-time risk control management method for power trading according to claim 2, characterized in that, The methods for identifying abnormal trader behavior risks include: Construct a behavioral knowledge graph with traders and accounts as nodes and trading relationships as edges; Graph neural network clustering algorithms are used to analyze behavioral knowledge graphs and identify abnormal transaction patterns. When behavior that deviates significantly from the normal pattern is detected, a risk of abnormal trader behavior is generated.

4. The multi-level real-time risk control management method for power trading according to claim 1, characterized in that, The risk limit parameter is a dynamic risk threshold; wherein, the dynamic adjustment method of the risk threshold is modeled as a Markov decision process, including: Based on a reinforcement learning model, the risk threshold is adaptively optimized by taking the current market volatility index, risk exposure level and trader behavior score as state input, threshold adjustment action as output, and comprehensive evaluation of risk control effect and business impact as reward signal.

5. The multi-level real-time risk control management method for power trading according to claim 1, characterized in that, The intervention response strategy, depending on the severity of the risk event, includes at least one of the following: Automatic interruption: Reject or cancel high-risk trading orders; Risk mitigation: Dynamically reduce trading limits or restrict trading strategies; Early warning notification: Send alert messages to traders and risk control personnel; Multi-level approval: Submit high-risk operations to different levels of manual review; Strategy Adjustment: Automatically executes adjustments to the trading portfolio, such as reducing positions or hedging.

6. The multi-level real-time risk control management method for power trading according to claim 4, characterized in that, The bias attribution analysis employs causal inference techniques, including: Identify potential risk factors related to risk events; Construct a causal relationship model to characterize the causal association between potential risk factors and risk indicator data; The causal contribution of potential risk factors to risk bias is quantified through intervention calculations or counterfactual simulations; wherein the risk bias is the difference between risk indicator data and risk threshold. Generate an attribution report that identifies the main causes of risk deviations and their percentage impact.

7. The multi-level real-time risk control management method for power trading according to claim 4, characterized in that, The closed-loop process is implemented in the following ways: Real-time transaction data triggers abnormal behavior monitoring; When a risk is detected, the risk threshold is dynamically adjusted within milliseconds. Once the risk indicator data exceeds the risk threshold, the attribution analysis of the root cause of the risk is completed within seconds, and the attribution results are obtained. Based on the attribution results, new risk control strategies are generated and implemented through multi-objective optimization. The data after execution is fed back to the monitoring end, initiating the next iteration.

8. The multi-level real-time risk control management method for power trading according to claim 1, characterized in that, The method is executed through a dedicated risk control device deployed on the internal network of the electricity sales company. This device adopts an integrated hardware and software architecture, using a high-performance industrial server as the hardware foundation to ensure low-latency data processing in a private environment.

9. A multi-level real-time risk control management device for power trading according to claim 1, characterized in that, include: The data acquisition interface is used to connect to the power trading platform to obtain market information, trading instructions and status data in real time; The risk control analysis engine is connected to the data acquisition interface and is used to process data, calculate risk indicator data, and identify risk events. The strategy execution module is connected to the risk control analysis engine and is used to execute corresponding intervention response strategies based on risk events. The human-computer interaction interface is used to display risk information, receive parameter configurations, and issue early warning notifications.

10. The multi-level real-time risk control management device for power trading according to claim 9, characterized in that, The risk control analysis engine integrates: The behavior graph engine is used to connect transaction entities through a knowledge graph and identify abnormal behavior patterns in real time. A reinforcement learning regulator is used to dynamically optimize risk thresholds based on market conditions. Causal attribution engine, used to analyze the root causes of risk deviations and quantify the contribution of key factors; A multi-objective optimization engine is used to generate Pareto optimal strategies that coordinate risk control, return, and liquidity objectives.

Images