A risk assessment system based on power trading market

Abstract

The application relates to the technical field of power system risk assessment, and discloses a risk assessment system based on a power transaction market, which realizes real-time collection and dynamic processing of power transaction price data through a price data real-time collection module combined with a real-time deviation analysis module, realizes real-time calculation of a price deviation value and capture of related influence factors, provides instant data support for market subjects, reduces the risk of decision-making errors, and realizes automatic triggering of parameter adjustment when real-time factor fluctuation exceeds a threshold value through a dynamic parameter adjustment module capable of establishing a correlation model, realizes automatic updating of calculation logic through a scene adaptation and emergency adjustment subunit, ensures that the price deviation value accurately reflects actual risks, a risk assessment module constructs a model in combination with dynamic parameters, calculates a risk index and divides grades, a risk early warning module generates early warning information in a timely manner, and a risk transmission analysis and stress test unit provides comprehensive risk prediction, helps market subjects scientifically formulate plans, and maintains the stability of the power transaction market.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system risk assessment, and particularly relates to a risk assessment system based on a power transaction market. BACKGROUND

[0002] The risk assessment system integrates market data (such as price fluctuations, supply and demand changes), transaction data (such as contract terms, performance records) and external factors (such as policy adjustments, extreme weather), uses quantitative models and artificial intelligence algorithms to identify risk points such as price fluctuations, credit defaults and market power abuse. The system can monitor market dynamics in real time, warn of abnormal transactions, assess risk loss probabilities in different scenarios, and provide decision support for market participants such as power generation companies and power selling companies, while assisting regulatory agencies in maintaining market order and improving the stability and security of power transactions.

[0003] However, in today's power transaction market, price deviation values are affected by real-time factors such as supply and demand fluctuations, changes in new energy output, and changes in load demand, and exhibit high-frequency and diverse dynamic characteristics. Existing technologies rely on fixed historical data and pre-set model parameters during the evaluation process, such as CN111612289B and CN117151882B. Neither of them establishes a real-time collection and dynamic calibration channel for price deviation values, nor do they lack a mechanism that can automatically update the price deviation calculation logic according to market real-time fluctuations. This results in the inability of the price deviation values calculated based on fixed parameters to accurately reflect the price deviation risk in actual transactions when the market price formation rules change due to real-time factors. SUMMARY

[0004] The technical problem to be solved by the present application is that the price deviation values calculated based on fixed parameters in the prior art cannot accurately reflect the price deviation risk in actual transactions. To solve this problem, we propose a risk assessment system based on a power transaction market.

[0005] To achieve the above object, the application adopts the following technical scheme: a risk assessment system based on a power transaction market, comprising a price data real-time acquisition module, configured to acquire power transaction price data in real time based on an Internet of Things and a market transaction interface, preprocess the acquired original price data, and store the processed data in a real-time database; a real-time deviation analysis module, configured to dynamically process the price data in the real-time database based on a stream computing framework, construct a price deviation real-time calculation model, calculate the deviation value of the transaction price from a market benchmark price in real time, extract the dynamic characteristics of the deviation value, and capture real-time factors affecting the deviation value; a dynamic parameter adjustment module, configured to establish a correlation model of price deviation calculation parameters and real-time influencing factors based on the analysis result of the real-time deviation analysis module, automatically trigger a parameter adjustment mechanism when the real-time factors fluctuate beyond a preset threshold, and dynamically update the price deviation calculation coefficient, the benchmark price calibration factor, and the floating interval parameter; a risk assessment module, configured to construct a risk assessment model based on the real-time parameters output by the dynamic parameter adjustment module, in combination with the dynamic characteristics of the price deviation value, calculate the risk index corresponding to the price deviation, divide the risk level, and generate a real-time risk assessment report; and a risk early warning module, configured to monitor the risk index output by the risk assessment module in real time, and automatically generate early warning information when the risk level exceeds a preset early warning threshold.

[0006] The real-time deviation analysis module comprises a deviation value real-time calculation unit, configured to calculate the price deviation value in a unit time in real time based on a sliding window algorithm, construct a deviation value time sequence, and calculate the instantaneous fluctuation amplitude, cumulative deviation amount, and deviation change rate of the deviation value; the price deviation value is obtained by the following formula:

[0007] wherein, : sliding window time length; : : real-time output of new energy at the moment; : benchmark output of new energy; : : real-time transaction price at the moment; : : market benchmark price at the moment; : : real-time supply-demand ratio at the moment; and an influencing factor capturing unit, configured to acquire real-time factor data affecting the price deviation through a multi-source data interface, the real-time factors including the real-time supply-demand ratio, the real-time output of new energy, the load of a power transmission channel, and a policy temporary control signal, normalize various factors, and construct a factor and deviation correlation feature set; a dynamic characteristic extraction unit, configured to mine the price deviation value time sequence for features by using a time sequence feature extraction algorithm, extract the periodicity feature, the mutation point feature, and the trend feature of the deviation value, combine the factor and deviation correlation feature set, and generate a deviation dynamic feature vector.

[0008] Further, the real-time deviation analysis module further comprises a deviation attribution analysis unit configured to: when the price deviation value exceeds the preset normal range, analyze the contribution of each real-time factor to the deviation value based on a feature importance evaluation algorithm, locate the dominant influencing factor, generate a deviation attribution report, and clarify the influence weight and action mechanism of each factor; and a data quality verification unit configured to: monitor the integrity and timeliness of the collected price data and influencing factor data in real time, and when the data missing rate or delay time exceeds a threshold value, start a data completion mechanism, perform data repair based on historical similar period data and a trend prediction model, and ensure the accuracy of the analysis result.

[0009] Further, the dynamic parameter adjustment module comprises: a parameter correlation modeling unit configured to: construct a mapping relationship model of the price deviation calculation parameters and the real-time influencing factors based on historical data, train a parameter adjustment rule using a machine learning algorithm, determine the optimal parameter configuration under different factor combinations, and form a parameter adjustment knowledge base; an automatic triggering unit configured to: compare the real-time influencing factor data with a preset threshold value in real time, and when the fluctuation amplitude of a single factor exceeds the threshold value or the comprehensive fluctuation index of multiple factors reaches a triggering condition, automatically activate the parameter adjustment process and send an adjustment instruction to the parameter updating unit; and the parameter updating unit configured to: dynamically correct the benchmark price coefficient, deviation weight factor, and abnormality judgment threshold in the price deviation calculation model according to the optimal parameter configuration output by the parameter correlation modeling unit, in combination with the real-time deviation characteristics, record the parameter adjustment time, adjustment amplitude, and triggering reason, and generate a parameter adjustment log.

[0010] Further, the parameter updating unit comprises: a real-time parameter calibration sub-unit configured to: based on the deviation dynamic characteristics and influencing factor data output by the real-time deviation analysis module, perform minute-level calibration on the initially determined optimal parameter configuration, and correct the parameter adjustment amplitude through real-time data feedback; and the calibration formula is:

[0011] wherein, : the parameter value at the moment of calibration, the initially optimal parameter configuration, the adjustment coefficient, : the deviation dynamic characteristic value at the moment, the feedback attenuation factor, : the influencing factor comprehensive value at the moment; when the deviation dynamic characteristic value exceeds the threshold value , secondary correction is started:

[0012] wherein, the secondary correction coefficient : a symbol function; a scenario adaptation subunit, configured to: identify an operation scenario of a current power transaction market, including a spot transaction period, a futures delivery period, and a policy regulation period, call a parameter adjustment template corresponding to the scenario according to a difference in a price formation mechanism of different scenarios, and dynamically match a calculation logic of a benchmark price coefficient; a parameter constraint verification subunit, configured to: set an upper and lower limit constraint and an adjustment frequency threshold of parameter adjustment, verify whether the parameter is out of a reasonable range in real time in a parameter updating process, and when a single adjustment amplitude exceeds a preset threshold, start a stepwise adjustment mechanism to implement parameter updating in stages to avoid a risk assessment result from being shocked due to a sudden change in the parameter; an adjustment trajectory tracking subunit, configured to: record specific values of each parameter update, a triggering factor, associated deviation characteristics, and a corresponding risk assessment result, construct a parameter adjustment trajectory database, and identify a rule of parameter adjustment and a potential optimization point through time sequence analysis to provide iterative training data for a parameter correlation modeling unit; and an emergency adjustment triggering subunit, configured to: monitor an extreme value in real-time influence factors, when the extreme value occurs, skip a regular parameter adjustment process, directly call a preset emergency parameter configuration, and send an urgent evaluation request to the adjustment effect evaluation unit to ensure the effectiveness of the parameter under extreme market conditions.

[0013] Further, the dynamic parameter adjustment module further includes: an adjustment effect evaluation unit, configured to: after parameter adjustment, evaluate an adjustment effect by comparing a price deviation calculation accuracy before and after adjustment and a risk assessment fit degree, when the effect is not as expected, start a secondary adjustment mechanism to optimize a parameter correlation model based on a feedback result; and a parameter backup and backtracking unit, configured to: backup a configuration before each parameter adjustment, support backtracking to a historical parameter state according to a time point or a risk assessment result, and when the system is abnormal, quickly recover to an optimal parameter configuration.

[0014] Further, the risk assessment module includes: a risk index calculation unit, configured to: based on a price deviation value after dynamic parameter adjustment, combine a deviation duration, a fluctuation frequency, and a cumulative deviation amount to construct a risk index calculation formula to calculate a real-time risk index, wherein the risk index is in a nonlinear correlation with the deviation characteristics; a risk level division unit, configured to: divide multiple risk levels according to a risk index size and a market subject risk bearing capacity, including low risk, medium risk, high risk, and extreme risk, and clearly define a triggering condition and an influence degree corresponding to each risk level; and an evaluation model optimization unit, configured to: periodically optimize a weight configuration of a risk assessment model based on historical risk events and evaluation results by using a reinforcement learning algorithm to improve the identification ability of the model to an extreme price deviation event and reduce evaluation lag.

[0015] Further, the risk assessment module further comprises: a risk transmission analysis unit, configured to analyze the transmission path of the price deviation risk to other market links, assess the potential influence on subsequent trading periods, related regional markets and upstream and downstream subjects, and generate a risk transmission map; and a stress test unit, configured to perform stress testing on the current risk assessment model based on historical extreme price deviation data and simulation scenarios, verify the stability and evaluation accuracy of the model under extreme market conditions, and output a test report and model optimization suggestions.

[0016] The technical effects and advantages of the present application are as follows: the price data real-time acquisition module acquires real-time power transaction price data based on the Internet of Things and market transaction interfaces, the real-time deviation analysis module dynamically processes data, calculates real-time price deviation values and captures real-time influencing factors such as supply-demand ratio and new energy output using a streaming computing framework and a sliding window algorithm, so that market subjects can master the dynamic characteristics of price deviation in real time, accurately identify price deviation risks in current transactions, provide real-time data support for adjusting transaction strategies, reduce the risk of decision-making errors caused by information lag, and improve operational stability. The dynamic parameter adjustment module establishes an association model between price deviation calculation parameters and real-time influencing factors, automatically triggers parameter adjustment when real-time factors fluctuate beyond a threshold, and combines a scenario adaptation subunit to match calculation logic for different transaction scenarios, and an emergency adjustment trigger subunit to cope with extreme market conditions, so that the price deviation calculation logic is automatically updated with real-time market fluctuations, ensuring that the calculated price deviation values can accurately reflect the actual risks when market price formation rules change, helping market subjects to effectively avoid risk misjudgments caused by parameter fixation and improving the flexibility of risk response. The risk assessment module constructs a risk assessment model based on the deviation values and dynamic characteristics adjusted by the dynamic parameter adjustment module, calculates a risk index and divides it into grades, and the risk warning module automatically generates warning information when the risk exceeds a limit, so that market subjects can clearly understand the risk grade and trigger conditions, and take timely preventive measures. At the same time, the risk transmission analysis unit and the stress test unit assess the risk transmission path and the stability of the model under extreme scenarios, provide comprehensive risk prediction for market subjects, help them to scientifically formulate transaction plans, and maintain the stable operation of the power transaction market. BRIEF DESCRIPTION OF DRAWINGS

[0017] The disclosure of the present application will be described with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of the present application. In the drawings, the same reference numerals are used to refer to the same components:

[0018] Figure 1 The module schematic diagram of the present application. DETAILED DESCRIPTION

[0019] It is easy to understand that, according to the technical solutions of the present application, those skilled in the art can propose various structures and implementation modes that can be replaced with each other without changing the essential spirit of the present application. Therefore, the following specific embodiments and drawings are only exemplary descriptions of the technical solutions of the present application, and should not be considered as the whole or as a limitation or restriction on the technical solutions of the present application.

[0020] The application discloses a risk assessment system based on a power transaction market, aiming to realize accurate assessment and timely warning of the risk of the power transaction market by collecting power transaction price data in real time, dynamically analyzing price deviation, adjusting assessment parameters, providing decision support for market subjects, and guaranteeing stable operation of the power transaction market.

[0021] Please refer to Figure 1 The system comprises a price data real-time collection module, which is the entrance of system data, responsible for providing high-quality basic data for subsequent analysis. The specific working process is as follows:

[0022] The intelligent monitoring terminal based on the Internet of Things technology and the market transaction interface connected with the power transaction platform collect power transaction price data in real time, including transaction prices, listing prices and settlement prices of each transaction period. The collection frequency is dynamically adjusted according to the market transaction activity, and the second-level collection is adopted in the transaction peak period and the minute-level collection is adopted in the non-peak period, so as to ensure the timeliness of the data.

[0023] The original price data collected is preprocessed, and the preprocessing process comprises: data cleaning, excluding obvious abnormal values; format standardization, converting data of different sources and different formats into a unified data format, and unifying the time stamp accuracy to the millisecond level; data deduplication, deleting duplicate records by comparing the time stamp and the price value. The processed data is stored in a real-time database, which supports high-concurrency writing and fast query of time-series data, meeting the needs of the system for real-time data storage.

[0024] The system further comprises a real-time deviation analysis module: the real-time deviation analysis module dynamically processes and feature mines based on the preprocessed data provided by the price data real-time collection module, and provides key analysis results for risk assessment.

[0025] The real-time deviation analysis module comprises the following contents:

[0026] (1) Deviation value real-time calculation unit: based on the sliding window algorithm, the price deviation value in a unit time is calculated in real time, and the size of the sliding window can be dynamically configured according to the transaction period. The deviation value of the transaction price in the real-time calculation window and the market benchmark price is calculated. Among them, the market benchmark price can adopt the average transaction price of the same period of the day or the reference benchmark price published by the regulatory agency.

[0027] The bias value time series is constructed, and the instantaneous fluctuation amplitude, cumulative bias amount and bias change rate of the bias value are calculated through time series analysis; the price bias value is obtained through the following formula:

[0028] , wherein, : sliding window time length; : : real-time new energy output at the moment; : new energy benchmark output; : : real-time transaction price at the moment; : : market benchmark price at the moment; : : real-time supply-demand ratio at the moment.

[0029] (2) Impact factor capturing unit: Real-time factor data affecting price deviation is collected through a multi-source data interface. Among them, the real-time supply-demand ratio comes from the real-time load and power output data of the power dispatch center; the new energy real-time output data comes from the real-time monitoring system of wind power and photovoltaic power stations; the transmission channel load data comes from the power grid operation monitoring platform; the policy temporary control signal comes from the information release system of the regulatory agency. Normalize each factor, convert factor data of different magnitudes and units to the [0, 1] interval, eliminate the dimension effect, and based on the normalized factor data, build a factor and bias correlation feature set, which includes the real-time correlation degree of each factor and the price deviation value.

[0030] (3) Dynamic feature extraction unit: Time series feature extraction algorithm is used to mine the features of the price deviation value time series. The time series feature extraction algorithm is wavelet transform or LSTM time series feature extraction, which extracts the periodicity feature, mutation point feature and trend feature of the deviation value. Fuse the extracted deviation features with the factor and bias correlation feature set to generate a bias dynamic feature vector. This vector contains the dynamic characteristics of the bias value itself and the correlation information of the influencing factors, providing multi-dimensional feature support for subsequent parameter adjustment and risk assessment.

[0031] (4) Bias attribution analysis unit: When the price deviation value exceeds the preset normal range, start bias attribution analysis. Based on the feature importance evaluation algorithm, analyze the contribution of each real-time factor to the deviation value, calculate the importance score of each factor, and the higher the score, the greater the influence of the factor on the current deviation. According to the importance score, locate the dominant influencing factor and generate a bias attribution report. The report clearly states the influence weight and mechanism of each factor, providing a basis for market participants to understand the causes of deviation.

[0032] (five) data quality checking unit: real-time monitoring of the completeness and timeliness of the collected price data and influencing factor data. The completeness is measured by the data missing rate, and the timeliness is measured by the data delay time. When the data missing rate exceeds 5% or the delay time exceeds 30 seconds, the data completion mechanism is started. Based on the historical similar period data and trend prediction model, the data is repaired.

[0033] In this embodiment, the application of the sliding window algorithm realizes real-time calculation of price deviation, which can dynamically capture price fluctuation characteristics; the collection and normalization processing of multi-source influencing factors ensure the comprehensiveness and comparability of factor analysis; the time sequence feature extraction and fusion technology improves the richness of the deviation characteristics; the deviation attribution analysis unit helps market subjects to accurately locate the deviation causes; the data quality checking unit ensures the reliability of the analysis data, avoiding the distortion of the analysis results caused by data problems. The system also includes a dynamic parameter adjustment module: based on the analysis results of the real-time deviation analysis module, the dynamic parameter adjustment module realizes the dynamic optimization of the price deviation calculation parameters, ensuring that the risk assessment model adapts to market changes.

[0034] The dynamic parameter adjustment module includes the following contents:

[0035] (1) Parameter correlation modeling unit: based on historical data, a mapping relationship model of price deviation calculation parameters and real-time influencing factors is constructed. Machine learning algorithms such as gradient boosting tree and neural network are used to train parameter adjustment rules, with normalized influencing factor data as input and optimal price deviation calculation parameters as output. Through a large amount of historical data training, the optimal parameter configuration under different factor combinations is determined to form a parameter adjustment knowledge base.

[0036] (2) Automatic triggering unit: real-time comparison of real-time influencing factor data and preset threshold. The preset threshold includes single factor threshold and multi-factor comprehensive fluctuation index threshold. When the fluctuation amplitude of a single factor exceeds the threshold or the multi-factor comprehensive fluctuation index reaches 0.6, the parameter adjustment process is automatically activated, and adjustment instructions are sent to the parameter updating unit to ensure that the parameters respond to market changes in a timely manner.

[0037] (3) Parameter updating unit: the parameter updating unit includes real-time parameter calibration sub-unit, scene adaptation sub-unit, parameter constraint checking sub-unit, adjustment trajectory tracking sub-unit, and emergency adjustment triggering sub-unit. The real-time parameter calibration sub-unit calibrates the optimal parameter configuration output by the parameter correlation modeling unit based on the deviation dynamic characteristics and influencing factor data output by the real-time deviation analysis module at the minute level. The calibration formula is:

[0038] , where : the parameter value after time calibration, : the preliminary optimal parameter configuration, Adjustment coefficient, : Dynamic characteristic value of time deviation Feedback attenuation factor : The combined value of influencing factors at any given time; the dynamic characteristic value of the deviation. Exceeding the threshold At that time, initiate a second correction:

[0039] ,in, Secondary correction coefficient The symbolic function is used to correct parameter adjustments based on real-time data feedback, ensuring a high degree of alignment between parameters and the current market conditions. The scenario adaptation subunit identifies the current operating scenarios of the electricity trading market, including spot trading sessions, futures delivery sessions, and policy regulation sessions. Based on the differences in price formation mechanisms across different scenarios, it calls the corresponding parameter adjustment template. For example, a more sensitive benchmark price coefficient calculation logic is used during spot trading sessions, while the weight of policy factors in the parameters is increased during policy regulation sessions. The parameter constraint verification subunit sets upper and lower limits for parameter adjustments and adjustment frequency thresholds. During parameter updates, it verifies in real-time whether parameters exceed reasonable ranges. When a single adjustment exceeds a preset threshold, a stepped adjustment mechanism is initiated, implementing parameter updates in 2-3 stages to avoid oscillations in risk assessment results due to sudden parameter changes. The adjustment trajectory tracking subunit records the specific values, triggering factors, associated deviation characteristics, and corresponding risk assessment results for each parameter update, constructing a parameter adjustment trajectory database. Time-series analysis identifies patterns and potential optimization points in parameter adjustments, providing iterative training data for the parameter correlation modeling unit to continuously optimize the parameter adjustment model. The emergency adjustment triggering subunit monitors extreme values ​​among real-time influencing factors. When extreme values ​​occur, the normal parameter adjustment process is skipped, and the preset emergency parameter configuration is directly invoked. At the same time, an expedited evaluation request is sent to the adjustment effect evaluation unit to ensure the effectiveness of parameters under extreme market conditions and to respond quickly to market risks.

[0040] (iv) Effect Evaluation Unit: After parameter adjustment, the effect of parameter adjustment is evaluated by comparing the accuracy of price deviation calculation and the fit of risk assessment before and after adjustment. When the effect does not meet expectations, a secondary adjustment mechanism is initiated to optimize the parameter correlation model based on the feedback results, regenerate the parameter configuration, and update it.

[0041] (v) Parameter Backup and Backtracking Unit: This unit backs up the configuration before each parameter adjustment, including the specific parameter values ​​and adjustment time. It supports backtracking to historical parameter states based on a specific time point or risk assessment result. In the event of a system anomaly, it can quickly restore to the optimal parameter configuration, ensuring stable system operation.

[0042] In this embodiment, the parameter correlation modeling unit realizes intelligent correlation of parameters and influencing factors through a machine learning algorithm, improving the scientificity of parameter adjustment; the automatic triggering unit ensures the timeliness of parameter adjustment, avoiding parameter lag after market changes; the multiple sub-units of the parameter updating unit work cooperatively to realize precise calibration of parameters, scene adaptation, risk control and trajectory tracking, improving the reliability of parameter adjustment; the adjustment effect evaluation unit and the parameter backup and backtracking unit provide closed-loop optimization and security for parameter adjustment, ensuring that the parameters are always in the optimal state.

[0043] The system also includes a risk assessment module: based on the real-time parameters output by the dynamic parameter adjustment module, the risk assessment module completes risk index calculation, level division and model optimization to provide a basis for risk early warning. The risk assessment module includes the following contents:

[0044] (I) Risk index calculation unit: based on the price deviation value after dynamic parameter adjustment, combined with the deviation duration, fluctuation frequency and cumulative deviation amount, a risk index calculation formula is constructed. The risk index is nonlinearly correlated with the deviation characteristics, and when the deviation value, duration and other characteristics exceed a certain threshold, the risk index will show exponential growth to highlight the high-risk state. The real-time risk index calculated is used as the core indicator to measure market risk.

[0045] (II) Risk level division unit: according to the size of the risk index and the risk tolerance of market participants, multiple risk levels are divided, including low risk (risk index <30), medium risk (30≤risk index <60), high risk (60≤risk index <80) and extreme risk (risk index ≥80). The trigger conditions and impact degrees corresponding to each risk level are clear, for example, low risk corresponds to small deviation value and short duration, and has negligible impact on market participants; extreme risk corresponds to large deviation value and long duration, which may lead to market disorder. At the same time, considering the differences in risk tolerance of different market participants, risk level thresholds are customized for them to improve the pertinence of risk assessment.

[0046] (III) Evaluation model optimization unit: based on historical risk events and evaluation results, the weight configuration of the risk assessment model is optimized periodically using reinforcement learning algorithm. By continuously learning the correlation between risk characteristics and actual risk events in historical data, the weight factor in the risk index calculation formula is adjusted to improve the model's ability to identify extreme price deviation events and reduce evaluation lag.

[0047] (IV) Risk transmission analysis unit: analyzes the transmission path of price deviation risk to other market links, evaluates the potential impact on subsequent trading periods, related regional markets and upstream and downstream subjects, generates a risk transmission map, and visually displays the diffusion path and impact range of the risk.

[0048] (V) Stress Testing Unit: Based on historical extreme price deviation data and simulation scenarios, the current risk assessment model is stress tested. The stability and accuracy of the model under extreme market conditions are verified, and a test report and model optimization suggestions are output.

[0049] In this embodiment, the risk index calculation unit constructs a risk index through multi-dimensional deviation characteristics, achieving quantitative assessment of market risk; the risk level division unit improves the practicality of risk assessment by considering market subject differences; the evaluation model optimization unit continuously optimizes the model through reinforcement learning, ensuring that the evaluation capability keeps pace with the times; the risk transmission analysis unit helps market subjects predict the scope of risk diffusion and take countermeasures in advance; and the stress testing unit ensures the reliability of the model under extreme market conditions, improving the robustness of the system.

[0050] The system also includes a risk warning module: the risk warning module monitors risk assessment results in real time and timely pushes warning information to market subjects, achieving early detection and early treatment of risks. The specific workflow is as follows: real-time monitoring of the risk index and risk level output by the risk assessment module, and automatic generation of warning information when the risk level exceeds the preset warning threshold. Warning information includes risk type, impact range, and suggested measures. According to the market subject's preset contact method, the push channel is selected to ensure that the warning information is quickly delivered. For high-risk and extreme-risk, multiple channels are used for simultaneous push, and a delivery confirmation mechanism is set to ensure that relevant subjects receive the information in a timely manner.

[0051] In this embodiment, the risk warning module achieves rapid transmission of risk information through real-time monitoring and multi-channel push, helping market subjects to take timely countermeasures and reduce risk losses. The detailed content of the warning information provides clear action guidance for market subjects, improving the effectiveness of risk response.

[0052] The system also includes a data storage module: the data storage module is responsible for the storage and management of various types of data in the system, ensuring the security, integrity, and accessibility of the data. The data storage module includes the following:

[0053] (I) Time Series Data Storage Unit: Configure a distributed time series database to store real-time price data, deviation value sequences, and risk indexes by time dimension partitioning.

[0054] (II) Correlation Data Storage Unit: Store parameter adjustment records, influence factor data, and risk assessment report correlation information, and use a relational database to establish the correlation between data.

[0055] (III) Data Life Cycle Management Unit: Based on the importance and time characteristics of data, set different storage periods for different types of data. Automatically clean up expired redundant data, and archive important historical data for storage, ensuring data storage efficiency and reducing storage costs.

[0056] (Four) Data encryption unit: sensitive information stored is encrypted, using field-level encryption mechanism, only sensitive fields are encrypted, non-sensitive fields are not encrypted, while ensuring data security and improving data processing efficiency. At the same time, the data transmission process is encrypted to prevent information leakage during transmission.

[0057] (Five) Access control unit: implement role-based access control strategy, assign different data access permissions to different market subjects. For example, regulatory agencies can access all data, and power selling companies can only access price data and risk assessment results related to them. Record data query and export operation logs, including operator, operation time, operation content, regularly conduct permission audit to ensure data access compliance and prevent data misuse.

[0058] In this embodiment, the application of distributed time series database ensures efficient storage and access of massive time series data; the associated data storage unit realizes the association management of various types of data, improving the convenience of data query; the data lifecycle management unit optimizes storage resource allocation and reduces storage cost; the data encryption unit and access control unit ensure the security and access compliance of data from a technical level, protecting the privacy and data rights of market subjects.

[0059] The system also includes a system adaptive module: the system adaptive module is responsible for monitoring system operation status, optimizing resource allocation, and ensuring efficient operation of the system under different loads. The system adaptive module includes the following:

[0060] (I) Performance monitoring unit: real-time monitoring of system module operation indicators, including parameter adjustment time consumption, risk assessment calculation time, CPU utilization, memory occupancy, network transmission rate, etc. Regularly generate system performance reports to analyze the trend of each indicator and identify potential performance bottlenecks.

[0061] (II) Resource dynamic allocation unit: based on the monitoring results of the performance monitoring unit, when the resource consumption of a module is too high, automatically adjust the system resource allocation strategy. Through containerization technology, dynamically adjust the CPU and memory quota of each module, prioritize resource requirements for real-time data collection, deviation calculation and risk assessment modules to ensure that core functions are not affected by resource limitations.

[0062] (III) Adaptive optimization unit: analyze the correlation between system operation indicators and market trading activity, establish a resource demand prediction model. Automatically expand system processing capacity in advance during peak trading periods, and reduce resource occupancy during low periods (such as early morning hours) to achieve dynamic scheduling of system resources, improve resource utilization, and reduce operating costs.

[0063] In the embodiment, the performance monitoring unit provides data basis for system optimization; the resource dynamic allocation unit ensures stable operation of the core module and avoids resource bottleneck from affecting system functions; and the adaptive optimization unit realizes efficient use of system resources through prediction and dynamic adjustment, and improves adaptability and economy of the system under different loads.

[0064] In summary, the power transaction market risk is realized in real time by the cooperative work of each module. The price data real-time acquisition module provides high-quality basic data, the real-time deviation analysis module deeply mines the price deviation characteristics and influencing factors, the dynamic parameter adjustment module ensures that the evaluation parameters adapt to market changes, the risk evaluation module quantifies the risk and divides the grades, the risk warning module timely delivers the risk information, the data storage module guarantees data safety and management, and the system adaptive module optimizes the system performance. The system can provide accurate risk information and decision support for market subjects, and promote stable, efficient and fair operation of the power transaction market.

[0065] The technical scope of the present application is not limited to the content in the above description, and those skilled in the art can make various modifications and changes to the above embodiments without departing from the technical idea of the present application, and these modifications and changes shall belong to the protection scope of the present application.

Claims

1. A risk assessment system based on a power trading market, characterized by, The price data real-time collection module is configured to collect power transaction price data in real time based on an Internet of Things and a market transaction interface, preprocess the collected original price data, and store the processed data in a real-time database. The real-time deviation analysis module is configured to dynamically process the price data in the real-time database based on a stream computing framework, construct a price deviation real-time calculation model, calculate the deviation value of the transaction price from a market benchmark price in real time, extract the dynamic characteristics of the deviation value, and capture real-time factors affecting the deviation value. The real-time deviation analysis module includes a deviation value real-time calculation unit configured to calculate the price deviation value in a unit time in real time based on a sliding window algorithm, construct a deviation value time series, and calculate the instantaneous fluctuation amplitude, cumulative deviation amount and deviation change rate of the deviation value. wherein, : sliding window time length; : new energy real-time output at the moment; : new energy benchmark output; : real-time transaction price at the moment; : market benchmark price at the moment; : real-time supply-demand ratio at the moment; the influence factor capturing unit is configured to: collect real-time factor data affecting the price deviation through a multi-source data interface, the real-time factors including a real-time supply-demand ratio, a new energy real-time output, a power transmission channel load, and a policy temporary regulation signal, normalize each type of factor, and construct a factor and deviation correlation feature set; the dynamic feature extraction unit is configured to: use a time series feature extraction algorithm to perform feature mining on a time series of the price deviation value, extract periodicity features, mutation point features, and trend features of the deviation value, combine the factor and deviation correlation feature set, and generate a deviation dynamic feature vector; the dynamic parameter adjustment module is configured to: based on an analysis result of the real-time deviation analysis module, establish a correlation model of a price deviation calculation parameter and a real-time influence factor, automatically trigger a parameter adjustment mechanism when the real-time factor fluctuation exceeds a preset threshold, and dynamically update a price deviation calculation coefficient, a benchmark price calibration factor, and a floating interval parameter; the risk assessment module is configured to: based on real-time parameters output by the dynamic parameter adjustment module, combine dynamic features of the price deviation value to construct a risk assessment model, calculate a risk index corresponding to the price deviation, divide a risk level, and generate a real-time risk assessment report; and the risk early warning module is configured to: monitor the risk index output by the risk assessment module in real time, and when the risk level exceeds a preset early warning threshold, automatically generate early warning information.

2. The risk assessment system based on power transaction market according to claim 1, characterized in that: The real-time deviation analysis module further includes a deviation attribution analysis unit configured to analyze the contribution of each real-time factor to the deviation value based on a feature importance evaluation algorithm when the price deviation value exceeds a preset normal range, locate the dominant influencing factor, generate a deviation attribution report, and determine the influence weight and action mechanism of each factor.

3. The risk assessment system based on power transaction market according to claim 1, characterized in that: The dynamic parameter adjustment module includes a parameter correlation modeling unit configured to construct a mapping relationship model of the price deviation calculation parameters and real-time influencing factors based on historical data, train a parameter adjustment rule using a machine learning algorithm, determine the optimal parameter configuration under different factor combinations, and form a parameter adjustment knowledge base.

4. The risk assessment system based on power transaction market according to claim 3, characterized in that: The parameter updating unit comprises a real-time parameter calibration subunit, configured to perform minute-level calibration on the preliminarily determined optimal parameter configuration based on the deviation dynamic characteristic and influence factor data output by the real-time deviation analysis module, and correct the parameter adjustment amplitude through real-time data feedback; the calibration formula is: wherein, : the parameter value at the moment of calibration, : the preliminarily optimal parameter configuration, : the adjustment coefficient, : the deviation dynamic characteristic value at the moment, : the feedback attenuation factor, : the influence factor comprehensive value at the moment; when the deviation dynamic characteristic value exceeds the threshold value , secondary correction is started: wherein, : the secondary correction coefficient, : the sign function; a scene adaptation subunit, configured to identify the running scene of the current power trading market, including the spot trading period, the futures delivery period and the policy regulation period, call the parameter adjustment template corresponding to the scene according to the price formation mechanism difference of different scenes, and dynamically match the calculation logic of the benchmark price coefficient; a parameter constraint checking subunit, configured to set the upper and lower limit constraints and the adjustment frequency threshold of parameter adjustment, and check whether the parameter exceeds the reasonable range in real time during the parameter updating process; when the single adjustment amplitude exceeds the preset threshold value, a stepwise adjustment mechanism is started to implement parameter updating in stages, so as to avoid the risk assessment result from being shocked due to sudden parameter change; an adjustment trajectory tracking subunit, configured to record the specific value of each parameter update, the triggering factor, the associated deviation characteristic and the corresponding risk assessment result, construct a parameter adjustment trajectory database, identify the law of parameter adjustment and the potential optimization point through time sequence analysis, and provide iterative training data for the parameter correlation modeling unit; an emergency adjustment triggering subunit, configured to monitor the extreme value in the real-time influence factor, when the extreme value appears, skip the regular parameter adjustment process, directly call the preset emergency parameter configuration, and send an urgent evaluation request to the adjustment effect evaluation unit, to ensure the effectiveness of the parameter under the extreme market condition.

5. The risk assessment system based on power trading market according to claim 3, wherein: The dynamic parameter adjustment module further includes an adjustment effect evaluation unit configured to evaluate the parameter adjustment effect by comparing the price deviation calculation accuracy and risk assessment fitting degree before and after the parameter adjustment, and optimize the parameter correlation model based on the feedback result when the effect does not meet the expectation. The dynamic parameter adjustment module further includes a parameter backup and backtracking unit configured to backup the configuration before each parameter adjustment, support backtracking to the historical parameter state according to the time point or risk assessment result, and quickly recover to the optimal parameter configuration when the system abnormally.

6. The risk assessment system based on power trading market according to claim 1, wherein: The risk assessment module comprises: a risk index calculation unit configured to: based on the price deviation value adjusted by the dynamic parameter, in combination with the deviation duration, the fluctuation frequency and the cumulative deviation amount, construct a risk index calculation formula to calculate a real-time risk index, wherein the risk index is in a nonlinear correlation with the deviation characteristics; a risk level division unit configured to: according to the risk index size and the market subject risk bearing capacity, divide a plurality of risk levels including low risk, medium risk, high risk and extreme risk, and clearly define the trigger conditions and the influence degree corresponding to each risk level; and an evaluation model optimization unit configured to: based on historical risk events and evaluation results, periodically optimize the weight configuration of the risk assessment model by using a reinforcement learning algorithm, improve the identification ability of the model for extreme price deviation events, and reduce the evaluation lag.

7. The risk assessment system based on power transaction market according to claim 6, characterized in that: The risk assessment module further comprises: a risk transmission analysis unit configured to: analyze the transmission path of the price deviation risk to other market links, evaluate the potential influence on the subsequent trading period, the associated regional market and the upstream and downstream subjects, and generate a risk transmission map; and a stress test unit configured to: based on historical extreme price deviation data and simulation scenarios, perform stress testing on the current risk assessment model, verify the stability and evaluation accuracy of the model under extreme market conditions, and output a test report and model optimization suggestions.

8. The risk assessment system based on power trading market according to claim 1, characterized in that: The data storage module comprises: a time series data storage unit configured to: configure a distributed time series database, store real-time price data, deviation value sequences and risk indexes in time dimension partition, support high-concurrency writing and low-latency querying, and optimize data compression algorithms to reduce storage occupation; an associated data storage unit configured to: store parameter adjustment records, influence factor data and risk assessment report associated information, establish an index association with time series data, and support multi-dimensional querying according to events, parameters and risk levels; and a data life cycle management unit configured to: based on the importance and time characteristics of data, set the storage period of different types of data, automatically clean up expired redundant data, archive important historical data, and ensure data storage efficiency.

9. The risk assessment system based on power trading market according to claim 1, wherein: The system adaptive module comprises: a performance monitoring unit configured to: monitor the running indicators of each module of the system in real time, record the parameter adjustment time consumption and risk assessment calculation time length, and generate a system performance report; a resource dynamic allocation unit configured to: based on the monitoring results of the performance monitoring unit, when the resource consumption of a certain module is too high, automatically adjust the system resource allocation strategy, and prioritize the resource demand of the real-time data acquisition, deviation calculation and risk assessment modules; and an adaptive optimization unit configured to: analyze the correlation between the system running indicators and the market transaction activity, establish a resource demand prediction model, automatically expand the system processing capacity in advance during the trading peak period, reduce resource occupation during the trough period, and realize efficient operation of the system.

Images