A method for measuring the income and risk of hydropower stations based on the value of hydropower

Abstract

The present invention discloses a hydropower station revenue risk measurement method based on hydropower value, including the following steps: S1: constructing a dynamic hydropower value assessment model, combining a distributed hydrological model to calculate the available hydropower in different spatiotemporal scenarios, converting the available hydropower into economic value, and obtaining dynamic hydropower value; S2: establishing a multi-risk factor coupling network and calculating the synergistic effect weights between risk factors; S3: constructing a dynamic revenue risk prediction model, updating the probability distribution curve of power generation revenue within a preset future time period in real time, and identifying key risk sources; S4: constructing a risk-return trade-off decision matrix based on the prediction results and key risk sources output by the dynamic revenue risk prediction model, and generating revenue optimization solutions under different risk response strategies; S5: using a visual interface to display the dynamic hydropower value, revenue risk prediction results, and revenue optimization solutions. This method achieves accurate assessment of dynamic hydropower value and real-time prediction of power generation revenue risk.

Description

Technical Field

[0001] The present invention relates to the technical field of energy benefit risk measurement, and in particular to a method for measuring the benefit risk of a hydropower station based on water energy value. Background Art

[0002] Against the backdrop of the global energy structure accelerating its transition to clean energy, hydropower stations, as an important component of renewable energy, face crucial challenges in terms of their revenue stability and risk prevention and control capabilities, which are crucial to the security of the energy system. Existing hydropower stations face complex and dynamic environmental challenges: climate change has significantly increased the uncertainty of river basin hydrological characteristics; electricity market reforms have made electricity price fluctuations a norm; and ecological protection policies have placed increasingly stringent constraints on hydropower development. The interweaving of these multiple factors has made it difficult for traditional revenue assessment methods to meet actual needs.

[0003] Traditional hydropower station revenue assessment techniques are primarily based on statistical analysis of historical data, achieving revenue forecasts by establishing a static mapping relationship between runoff and power generation. For example, early methods used only multi-year average flow to calculate theoretical power generation, combined with fixed on-grid electricity prices to estimate revenue, without considering the spatiotemporal variability of hydrological processes. The advantages of this type of technology lie in its simple calculation logic and low data requirements, but its flaws are also significant: it cannot capture the impact of extreme climate events (such as droughts and floods) on the available amount of hydropower, ignores the direct impact of electricity market price fluctuations on revenue, and is even more difficult to quantify changes in compliance costs caused by policy factors such as ecological flow constraints. As the operating environment of hydropower stations becomes more complex, traditional static assessment models have gradually exposed problems of insufficient timeliness and incomplete risk coverage, making it difficult for operators to respond to real-time changes in market and natural risks.

[0004] Although existing technologies have introduced some dynamic elements based on traditional methods, they still have significant limitations. For example, some studies have attempted to combine hydrological models to predict runoff changes, or use time series analysis methods to capture the patterns of electricity price fluctuations, but most of them remain at the independent analysis level of a single risk factor. Typical solutions, such as power generation forecasting based on distributed hydrological models, can characterize the spatial distribution characteristics of runoff, but fail to combine the physical properties of water resources with the economic properties of the power market, and cannot achieve the transition from "water volume assessment" to "value assessment"; for example, electricity price forecasting models based on machine learning can improve the prediction accuracy of short-term price fluctuations, but lack the coupled analysis of multidimensional factors such as hydrological risks and equipment risks, making it difficult to construct a complete benefit-risk assessment framework. Existing technologies generally lack in-depth utilization of real-time monitoring data, and risk warning and decision support functions lag behind, which cannot meet the actual needs of dynamic regulation of hydropower stations. Summary of the Invention

[0005] Based on the above technical problems, this application discloses a method for measuring the risk of hydropower station revenue based on the value of hydropower, including:

[0006] S1: Build a dynamic water energy value assessment model. Collect real-time meteorological data, satellite remote sensing runoff data, and hydropower station equipment operation data within the basin. Combined with a distributed hydrological model, calculate the available water energy under different spatiotemporal scenarios. Based on real-time electricity prices in the power market, medium- and long-term contract prices, and ecological flow constraints, convert the available water energy into economic value to obtain dynamic water energy value.

[0007] S2: Establish a multi-risk factor coupling network, identify risk sources, construct risk transmission paths, quantify the marginal impact of each risk factor on hydropower value and power generation revenue, and calculate the synergistic effect weights among risk factors;

[0008] S3: By using the reinforcement learning algorithm and taking the dynamic hydropower value and multi-risk factor coupled network data as input, a dynamic profit-risk prediction model is constructed to update the probability distribution curve of power generation profit within a preset time period in the future in real time and identify key risk sources;

[0009] S4: Based on the prediction results and key risk sources output by the dynamic benefit-risk prediction model, a risk-benefit trade-off decision matrix is ​​constructed. Combined with the peak-shaving and frequency-regulating capabilities of the hydropower station and the water storage status of the reservoir, benefit optimization plans under different risk response strategies are generated;

[0010] S5: Use a visual interface to display dynamic hydropower value, benefit-risk prediction results, and benefit optimization plans, support users to conduct multi-scenario simulation interactive analysis, and output visual risk curves and benefit comparison charts.

[0011] Preferably, the distributed hydrological model is combined in S1 to calculate the available water energy under different spatiotemporal scenarios, specifically: the precipitation in the real-time meteorological data is converted into , evaporation , initial runoff of the basin in satellite remote sensing runoff data , and turbine efficiency in hydropower station equipment operation data , substituted into the space-time coupling equation of the distributed hydrological model, the formula is: ,in, is the regional precipitation-runoff conversion coefficient, is the temporal and spatial correction parameter of the terrain slope; according to the height difference of the hydropower station , the available water energy calculation formula is , calculate each spatiotemporal scene The available water energy under is the density of water, is the acceleration due to gravity, represents the spatial grid number, Represents a time node.

[0012] Preferably, in S1, the available water energy is converted into economic value through a dynamic water energy value evaluation model based on the real-time electricity price in the power market, the medium- and long-term contract price, and the ecological flow constraint condition, specifically:

[0013] Get real-time electricity prices in the power market , medium- and long-term contract electricity prices , determine the threshold of water energy availability for power generation based on ecological flow constraints ; Construct the price weight distribution function, the formula is: ,in is the real-time electricity price priority adjustment coefficient, through the mixed pricing formula , calculate each time node Dynamic water energy value ,in The available water energy at a time node is used to obtain the dynamic water energy value.

[0014] Preferably, the risk sources determined in the multi-risk factor coupling network in S2 include hydrological risk, market risk, equipment risk and policy risk; among them, hydrological risk includes seasonal runoff fluctuation risk, abnormal water volume risk caused by extreme climate and hydrological change risk caused by human activities in the basin; market risk covers the risk of real-time electricity price fluctuation in the power market, medium- and long-term contract default risk, and consumption competition risk brought about by new energy grid connection; equipment risk includes the risk of failure of hydro-turbine generator sets, the risk of aging of power transmission and transformation equipment and the risk of failure of intelligent monitoring system; policy risk involves the risk of increased compliance costs caused by environmental protection policies and the risk of changes in the income structure brought about by power system reform.

[0015] Preferably, in S2, the marginal impact of each risk factor on the value of hydropower and power generation income is quantified by constructing a risk transfer path, and the synergistic effect weights between risk factors are calculated. Specifically, the risk factor set is established. ,in Representative Class risk, the risk transfer coefficient matrix is ,in Indicates risk factors right The transmission intensity of the risk factor is calculated using the marginal impact quantification formula The value of hydropower The marginal impact of is: ,in The value of hydropower to risk factors The partial derivative of the risk factor is calculated by the synergistic effect weight formula and The synergistic effect weight between them is: ,in Risk factors and The impact on the value of water energy when they work together, They are The value of water energy when acting alone The amount of influence, Risk factors and The impact on the value of water energy when they work together, They are The value of water energy when acting alone The amount of influence, For the maximum impact among all risk factor combinations, a quantitative analysis of the impact and synergistic effect of risk factors is conducted.

[0016] Preferably, the dynamic prediction model of return risk in S3 includes a deep learning architecture of an input layer, an attention mechanism layer, a bidirectional long short-term memory network layer Bi-LSTM and an output layer; the obtained dynamic hydropower value sequence and risk factor transfer intensity matrix and synergy effect weight matrix are used as input layer data; the weights of different risk factors and time series data are adaptively allocated through the attention mechanism layer to highlight the key risk information that has a greater impact on the return; the bidirectional long short-term memory network layer is used to capture the time dependence and bidirectional dynamic characteristics of risk factors and hydropower value; in the output layer, a probability density function fitting module is used to output the probability distribution curve of power generation income in a preset time period in the future, wherein the probability density function adopts a mixed Gaussian model, and the model parameters are optimized through the expectation maximization algorithm to realize the dynamic prediction of power generation return risk.

[0017] Preferably, in S3, the probability distribution curve of power generation revenue within a preset future time period is predicted by a revenue risk dynamic prediction model to identify key risk sources, specifically:

[0018] The data acquisition frequency is , each time new dynamic water energy value data is collected , risk transfer coefficient matrix and synergy weight matrix Then, it is combined with the historical data sequence to form a sliding time window data , input the return-risk dynamic prediction model, use the bidirectional long short-term memory network layer in the return-risk dynamic prediction model to extract features of the time window data, combine the attention mechanism layer to redistribute the weights of each risk factor at the current moment, and calculate the future preset time period through the mixed Gaussian model of the output layer Internal power generation income The probability density function of ,in is the number of Gaussian distributions, 、 、 Respectively Gaussian distribution in The mixing coefficient, mean and variance at each moment are used to update the probability distribution curve of power generation income. By calculating the contribution of each risk factor to the variance of the probability distribution curve, the contribution exceeding the set threshold is The risk factors identified as key risk sources are as follows: , For power generation income The variance of For risk factors.

[0019] Preferably, the prediction results and key risk sources output by the dynamic prediction model for profit and risk in S4 are used to construct a risk-benefit trade-off decision matrix, specifically: the power generation profit probability distribution curve in the prediction result is divided into quantiles to form a set of high, medium and low profit scenarios; for each profit scenario, combined with the type and impact of the key risk source, an evaluation index system including four dimensions of risk response strategy, implementation cost, profit improvement potential and risk mitigation effect is constructed; the hierarchical analysis method is used to determine the relative weight of each dimension, and the objective weight of each evaluation index is calculated to form a combined weight vector; the peak-shaving and frequency-regulating capacity of the hydropower station and the water storage status constraints of the reservoir are converted into boundary conditions of the decision space, and a multi-objective planning algorithm is used to generate a Pareto optimal solution set in the decision space, each solution corresponds to a risk-benefit trade-off solution; the Pareto optimal solution set is clustered and analyzed to form a finite number of representative decision solution clusters, each solution cluster corresponds to the optimal decision strategy under different risk preferences, and a risk-benefit trade-off decision matrix is ​​constructed.

[0020] Preferably, in said S4, the peak-shaving and frequency-regulating capacity of the hydropower station and the water storage state constraint conditions of the reservoir are converted into the boundary conditions of the decision space, and a multi-objective programming algorithm is used to generate a Pareto optimal solution set, and then a decision solution cluster is formed by cluster analysis. The specific method is: define a decision variable set ,in Represents the implementation intensity of different risk response strategies; the peak and frequency regulation capacity constraints of hydropower stations are expressed as , the reservoir water storage state constraint is expressed as ,in 、 The upper and lower limits of peak and frequency regulation power, 、 are the upper and lower limits of the reservoir’s water storage capacity, 、 The decision space is constructed for the influence coefficient of each strategy on the peak load power and water storage status. The multi-objective optimization is performed in the decision space by the non-dominated sorting genetic algorithm. The objective function is to maximize the power generation income and minimize the risk level. The Pareto optimal solution set is obtained through iterative update. For each solution in the Pareto optimal solution set, the Pareto optimal solution set is obtained. , calculate the Euclidean distance to other solutions, the formula is: ,in is the number of objective functions, For the The objective function value is set, the density threshold and neighborhood radius are set, and the distances less than the neighborhood radius and the density greater than the density threshold are disaggregated into the same cluster, forming a finite number of representative decision plan clusters, each cluster corresponding to decision plans with similar risk-return characteristics.

[0021] Preferably, in S5, the user can input custom scenario variables through the parameter adjustment panel, including but not limited to the amplitude of electricity price fluctuations and the probability of occurrence of extreme hydrological events, call the benefit-risk dynamic prediction model and the risk-benefit trade-off decision matrix in real time, recalculate and generate visualization charts for the corresponding scenarios, associate the various visualization components, and perform linkage analysis.

[0022] Compared with the prior art, the technical solution of this application has the following technical effects:

[0023] This invention realizes the dynamic transformation of water energy resources from "physical quantity" to "economic value" by integrating real-time meteorological data, satellite remote sensing runoff and distributed hydrological models. It characterizes the runoff changes in different grids and time periods through time-space coupling equations, and combines the hybrid pricing mechanism of real-time electricity prices and medium- and long-term contract prices to accurately calculate the water energy value in various time-space scenarios.

[0024] The present invention constructs a multi-risk factor coupling network, incorporating four types of risks, namely hydrology, market, equipment, and policy, and their sub-risks into a unified framework. Through the risk transfer coefficient matrix and synergistic effect weight algorithm, the transmission paths and superimposed impacts between risks are revealed. Multi-dimensional analysis enables risk assessment to shift from "fragmentation" to "systematization", avoiding the failure of prevention and control strategies caused by traditional methods due to ignoring risk linkage.

[0025] This invention adopts the deep learning architecture of "attention mechanism + bidirectional long short-term memory network", combines the sliding time window data to update the profit risk prediction model in real time, automatically captures real-time runoff data and electricity price fluctuation information, and quickly updates the profit probability distribution curve through the mixed Gaussian model. It can identify the compound risk scenario of "flood risk leading to equipment failure + sudden drop in electricity price" in advance, and calculate the contribution of each risk factor to the profit variance, accurately locate the key risk sources, and the prediction result update frequency can reach minutes.

[0026] This invention transforms complex hydropower value, risk distribution, and decision-making scenarios into intuitive charts. Users can customize the input of electricity price fluctuations and probability parameters of extreme hydrological events, and generate multi-scenario comparative analysis results in real time. Through the linkage analysis function, by clicking on a risk area, the optimal power generation plan for the corresponding time period can be displayed.

[0027] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application so that it can be implemented in accordance with the contents of the specification, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following is a detailed description of the preferred embodiment of the present application in conjunction with the accompanying drawings.

[0028] Based on the detailed description of the specific embodiments of the present application in conjunction with the accompanying drawings below, those skilled in the art will become more aware of the above and other objects, advantages and features of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without inventive work. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0030] Figure 1 This is a flow chart of a method for measuring the risk of a hydropower station's revenue based on the value of hydropower;

[0031] Figure 2 A graph comparing the typical daily water energy value and the real-time electricity price in the embodiment;

[0032] Figure 3 This is a comparison diagram of the income distribution in the embodiment;

[0033] Figure 4 This is a radar chart showing the accuracy of identifying key risk sources in the embodiment. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. In the following description, specific details such as specific configurations and components are provided only to help fully understand the embodiments of the present application. Therefore, it should be clear to those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, for clarity and brevity, the description of known functions and structures has been omitted in the embodiments.

[0035] It should be understood that references throughout this specification to "one embodiment" or "this embodiment" mean that a particular feature, structure, or characteristic associated with the embodiment is included in at least one embodiment of the present application. Therefore, the appearance of "one embodiment" or "this embodiment" throughout this specification does not necessarily refer to the same embodiment. Furthermore, these particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0036] In addition, the present application may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplicity and clarity and does not in itself indicate the relationship between the various embodiments and / or settings discussed.

[0037] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" in this article describes another type of association object relationship, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " in this article generally indicates that the previous and subsequent associated objects are in an "or" relationship.

[0038] The term "at least one" in this article is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, at least one of A and B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0039] It should also be noted that, in this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include," "comprises," or any other variations thereof are intended to cover non-exclusive inclusion.

[0040] Example 1

[0041] This embodiment mainly describes a method for measuring the risk of hydropower station income based on the value of hydropower. Figure 1 As shown, specifically including:

[0042] S1: Build a dynamic water energy value assessment model. Collect real-time meteorological data, satellite remote sensing runoff data, and hydropower station equipment operation data within the basin. Combined with a distributed hydrological model, calculate the available water energy under different spatiotemporal scenarios. Based on real-time electricity prices in the power market, medium- and long-term contract prices, and ecological flow constraints, convert the available water energy into economic value to obtain dynamic water energy value.

[0043] S2: Establish a multi-risk factor coupling network, identify risk sources, construct risk transmission paths, quantify the marginal impact of each risk factor on hydropower value and power generation revenue, and calculate the synergistic effect weights among risk factors;

[0044] S3: By using the reinforcement learning algorithm and taking the dynamic hydropower value and multi-risk factor coupled network data as input, a dynamic profit-risk prediction model is constructed to update the probability distribution curve of power generation profit within a preset time period in the future in real time and identify key risk sources;

[0045] S4: Based on the prediction results and key risk sources output by the dynamic benefit-risk prediction model, a risk-benefit trade-off decision matrix is ​​constructed. Combined with the peak-shaving and frequency-regulating capabilities of the hydropower station and the water storage status of the reservoir, benefit optimization plans under different risk response strategies are generated;

[0046] S5: Use a visual interface to display dynamic hydropower value, benefit-risk prediction results, and benefit optimization plans, support users to conduct multi-scenario simulation interactive analysis, and output visual risk curves and benefit comparison charts.

[0047] Furthermore, the distributed hydrological model is combined in S1 to calculate the available water energy in different time and space scenarios, specifically: the precipitation in the real-time meteorological data is converted into , evaporation , initial runoff of the basin in satellite remote sensing runoff data , and turbine efficiency in hydropower station equipment operation data , substituted into the space-time coupling equation of the distributed hydrological model, the formula is: ,in, is the regional precipitation-runoff conversion coefficient, is the temporal and spatial correction parameter of the terrain slope; according to the height difference of the hydropower station , the available water energy calculation formula is , calculate each spatiotemporal scene The available water energy under is the density of water, is the acceleration due to gravity, represents the spatial grid number, Represents a time node.

[0048] Furthermore, in S1, based on the real-time electricity price in the electricity market, the price of medium- and long-term contracts, and the ecological flow constraint, the available water energy is converted into economic value through a dynamic water energy value evaluation model, specifically:

[0049] Get real-time electricity prices in the power market , medium- and long-term contract electricity prices , determine the threshold of water energy availability for power generation based on ecological flow constraints ; Construct the price weight distribution function, the formula is: ,in is the real-time electricity price priority adjustment coefficient, through the mixed pricing formula , calculate each time node Dynamic water energy value ,in The available water energy at a time node is used to obtain the dynamic water energy value.

[0050] Furthermore, the risk sources identified in the multi-risk factor coupling network in S2 include hydrological risk, market risk, equipment risk and policy risk; among them, hydrological risk includes seasonal runoff fluctuation risk, abnormal water volume risk caused by extreme climate and hydrological change risk caused by human activities in the basin; market risk covers the risk of real-time electricity price fluctuation in the power market, the risk of default in medium- and long-term contracts, and the risk of competition for consumption brought about by the grid connection of new energy; equipment risk includes the risk of failure of hydro-turbine generator sets, the risk of aging of power transmission and transformation equipment and the risk of failure of intelligent monitoring systems; policy risk involves the risk of increased compliance costs caused by environmental protection policies and the risk of changes in the income structure brought about by power system reform.

[0051] Furthermore, in S2, the marginal impact of each risk factor on the value of hydropower and power generation income is quantified by constructing a risk transmission path, and the synergistic effect weights between risk factors are calculated. Specifically, the risk factor set is established. ,in Representative Class risk, the risk transfer coefficient matrix is ,in Indicates risk factors right The transmission intensity of the risk factor is calculated using the marginal impact quantification formula The value of hydropower The marginal impact of is: ,in The value of hydropower to risk factors The partial derivative of the risk factor is calculated by the synergistic effect weight formula and The synergistic effect weight between them is: ,in Risk factors and The impact on the value of water energy when they work together, They are The value of water energy when acting alone The amount of influence, Risk factors and The impact on the value of water energy when they work together, They are The value of water energy when acting alone The amount of influence, For the maximum impact among all risk factor combinations, a quantitative analysis of the impact and synergistic effect of risk factors is conducted.

[0052] Furthermore, the dynamic prediction model of return risk in S3 includes a deep learning architecture of an input layer, an attention mechanism layer, a bidirectional long short-term memory network layer Bi-LSTM, and an output layer; the obtained dynamic hydropower value sequence and risk factor transfer intensity matrix and synergy effect weight matrix are used as input layer data; the weights of different risk factors and time series data are adaptively allocated through the attention mechanism layer to highlight the key risk information that has a greater impact on the return; the bidirectional long short-term memory network layer is used to capture the time dependence and bidirectional dynamic characteristics of risk factors and hydropower value; in the output layer, a probability density function fitting module is used to output the probability distribution curve of power generation revenue in a preset time period in the future, wherein the probability density function adopts a mixed Gaussian model, and the model parameters are optimized through the expectation maximization algorithm to realize the dynamic prediction of power generation return risk.

[0053] Furthermore, in S3, the probability distribution curve of power generation revenue within a preset time period in the future is predicted by the revenue risk dynamic prediction model to identify key risk sources, specifically:

[0054] The data acquisition frequency is , each time new dynamic water energy value data is collected , risk transfer coefficient matrix and synergy weight matrix Then, it is combined with the historical data sequence to form a sliding time window data , input the return-risk dynamic prediction model, use the bidirectional long short-term memory network layer in the return-risk dynamic prediction model to extract features of the time window data, combine the attention mechanism layer to redistribute the weights of each risk factor at the current moment, and calculate the future preset time period through the mixed Gaussian model of the output layer Internal power generation income The probability density function of ,in is the number of Gaussian distributions, 、 、 Respectively Gaussian distribution in The mixing coefficient, mean and variance at each moment are used to update the probability distribution curve of power generation income. By calculating the contribution of each risk factor to the variance of the probability distribution curve, the contribution exceeding the set threshold is The risk factors identified as key risk sources are as follows: , For power generation income The variance of For risk factors.

[0055] Furthermore, the prediction results and key risk sources output by the dynamic profit-risk prediction model in S4 are used to construct a risk-benefit trade-off decision matrix, specifically: the power generation profit probability distribution curve in the prediction result is divided into quantiles to form a set of high, medium and low profit scenarios; for each profit scenario, combined with the type and impact of the key risk source, an evaluation index system including four dimensions of risk response strategy, implementation cost, profit improvement potential and risk mitigation effect is constructed; the hierarchical analysis method is used to determine the relative weight of each dimension, and the objective weight of each evaluation index is calculated to form a combined weight vector; the peak-shaving and frequency-regulating capacity of the hydropower station and the water storage status constraints of the reservoir are converted into boundary conditions of the decision space, and a multi-objective planning algorithm is used to generate a Pareto optimal solution set in the decision space, each solution corresponds to a risk-benefit trade-off solution; the Pareto optimal solution set is clustered and analyzed to form a finite number of representative decision solution clusters, each solution cluster corresponds to the optimal decision strategy under different risk preferences, and a risk-benefit trade-off decision matrix is ​​constructed.

[0056] Furthermore, in S4, the peak-shaving and frequency-regulating capacity of the hydropower station and the water storage state constraints of the reservoir are transformed into the boundary conditions of the decision space, and the multi-objective programming algorithm is used to generate the Pareto optimal solution set, and then the decision solution cluster is formed through cluster analysis. The specific method is: define the decision variable set ,in Represents the implementation intensity of different risk response strategies; the peak and frequency regulation capacity constraints of hydropower stations are expressed as , the reservoir water storage state constraint is expressed as ,in 、 The upper and lower limits of peak and frequency regulation power, 、 are the upper and lower limits of the reservoir’s water storage capacity, 、 The decision space is constructed for the influence coefficient of each strategy on the peak load power and water storage status. The multi-objective optimization is performed in the decision space by the non-dominated sorting genetic algorithm. The objective function is to maximize the power generation income and minimize the risk level. The Pareto optimal solution set is obtained through iterative update. For each solution in the Pareto optimal solution set, the Pareto optimal solution set is obtained. , calculate the Euclidean distance to other solutions, the formula is: ,in is the number of objective functions, For the The objective function value is set, the density threshold and neighborhood radius are set, and the distances less than the neighborhood radius and the density greater than the density threshold are disaggregated into the same cluster, forming a finite number of representative decision plan clusters, each cluster corresponding to decision plans with similar risk-return characteristics.

[0057] Furthermore, in S5, the user can input custom scenario variables through the parameter adjustment panel, including but not limited to the amplitude of electricity price fluctuations and the probability of extreme hydrological events, and call the profit-risk dynamic prediction model and the risk-profit trade-off decision matrix in real time, recalculate and generate visualization charts for the corresponding scenarios, and associate the various visualization components for linkage analysis.

[0058] This embodiment describes in detail how to achieve spatiotemporal dynamic calculation of available hydropower by integrating real-time meteorological data, satellite remote sensing runoff data, and distributed hydrological models. It also combines the multi-dimensional price mechanism of the electricity market with ecological constraints to convert physical quantities into economic value, solving the problem of "lack of value assessment" in traditional methods. It establishes a multi-risk factor coupling network, quantifies the marginal impact and synergistic effect of risk sources such as hydrology, market, equipment, and policy, constructs a dynamic return-risk prediction model through reinforcement learning algorithms and mixed Gaussian models, and uses real-time data to roll over prediction results, significantly improving the timeliness and accuracy of risk prediction.

[0059] Based on Example 1, this example describes in detail the specific implementation effects of this application, specifically:

[0060] The study takes a cascade hydropower station group in a river basin in southwest my country as the object, referred to as "Power Station Group A". The area includes three cascade hydropower stations with a total installed capacity of 1200MW, which undertake the tasks of flood control in the river basin, ecological scheduling and system peak regulation. The flood season from May to August 2024 is selected as the research period. Combined with real-time meteorological data, market transaction information and equipment operation status, the effectiveness of the return-risk measurement method based on hydropower value is verified.

[0061] The basin area of ​​Power Station Group A is approximately 18,000 square kilometers, with an average annual precipitation of 1,400 mm. The runoff is mainly supplied by precipitation, with the flood season (May-September) accounting for about 70% of the water volume. The three power stations are all dam-type developments, with a total reservoir capacity of approximately 2.8 billion cubic meters. The upstream power stations have annual regulation capabilities, while the mid- and downstream power stations have seasonal regulation capabilities. The main parameters are shown in the following table:

[0062] Power station name Installed capacity (MW) Adjust performance Upstream power station 400 Annual adjustment midstream power station 500 Seasonal Adjustment Downstream power station 300 Seasonal Adjustment

[0063] Hydrological data were obtained through 12 meteorological stations and three Doppler radars in the basin to obtain hourly precipitation, evaporation, temperature, etc. The average precipitation from May to August 2024 was 3.8 mm / h, and the maximum hourly rainfall intensity reached 20 mm / h (occurring on July 12); the initial runoff of the basin was inverted using the Sentinel-1 radar satellite. The average initial runoff during the flood season was 350 m³ / s, and the peak value reached 1200 m³ / s during extreme rainstorms (such as July 12); equipment operation data was collected, and the average turbine efficiency was 88%. The reservoir water level fluctuation range during the flood season was 85%-100% of the normal water storage level.

[0064] Obtain market data and real-time electricity prices: Participate in the provincial electricity spot market. From May to August 2024, the real-time electricity price range is 0.25-0.62 yuan / kWh, with an average price of 0.55 yuan / kWh during peak hours (18:00-22:00) and 0.30 yuan / kWh during off-peak hours (00:00-06:00). Medium- and long-term contracts: Sign annual contracts for difference (CFDs), with the contracted electricity volume accounting for 60% of the forecast annual power generation, at a contract price of 0.38 yuan / kWh, and a default rate set at 5% (simulating market risk).

[0065] Risk factor settings include seasonal runoff fluctuations (30% probability), extreme rainstorms (10% probability), and reduced runoff due to upstream land reclamation and cultivation (15% probability) for hydrological risks; real-time electricity price volatility of 18%, contract default risk (trigger condition: large-scale development of new energy leads to oversupply in the market), and competition for consumption (triggered when the penetration rate of new energy exceeds 35%) for market risks; turbine failure (annual average occurrence rate of 3 times), main transformer tripping (2% probability), and monitoring system abnormality (1% probability) for equipment risks; and a 10% increase in ecological flow due to environmental protection inspections (20% probability) and adjustments to time-of-use electricity price policies (re-division of peak and valley periods, 15% probability) for policy risks.

[0066] Construct four typical scenarios based on historical data:

[0067] Baseline scenario (50%): normal hydrological conditions, real-time electricity price fluctuations of ±10%, no equipment failures and policy changes.

[0068] Hydrological risk scenario (25%): Extreme rainstorms combined with runoff fluctuations will tighten ecological flow constraints.

[0069] Market risk scenario (15%): Real-time electricity prices plummet by 20%, and contract customers default.

[0070] Compound risk scenario (10%): heavy rain + falling electricity prices + unit failure + increased ecological flow.

[0071] Conduct dynamic water energy value assessment, calculate available water energy, and analyze water balance in different spatiotemporal scenarios using a distributed hydrological model:

[0072] During normal periods (e.g. mid-May), average precipitation is 3 mm / h, evaporation is 1.2 mm / h, and initial runoff is 280 m³ / s. The model calculates that the available hydropower is approximately 2.2 million kW·h / hour, meeting the ecological flow requirements (the downstream power station must ensure a minimum of 20 m³ / s, corresponding to a hydropower threshold of 1.8 million kW·h / hour).

[0073] During the extreme rainstorm period (2:00 PM to 4:00 PM, July 12), hourly rainfall reached 20 mm / h, and the initial runoff surged to 1,200 m³ / s. However, due to ecological flow constraints (which needed to be maintained above 30 m³ / s), the actual hydropower available for electricity generation was 5.5 million kW·h / h (a 25% reduction compared to the unconstrained scenario).

[0074] Transform economic value: adopt a hybrid pricing mechanism and calculate the revenue weight of each period based on the real-time electricity price priority coefficient γ=0.6:

[0075] During peak hours (18:00-22:00): the real-time electricity price is 0.58 yuan / kWh, the contract electricity price is 0.38 yuan / kWh, and the weight ω=0.6×0.58 / (0.58+0.38)=0.36, that is, 36% of the water energy value is determined by the real-time market and 64% is guaranteed by the contract.

[0076] Normal period (06:00-18:00): real-time electricity price is 0.42 yuan / kWh, weight ω=0.6×0.42 / (0.42+0.38)=0.315, and the income structure tends to be balanced.

[0077] Valley period (00:00-06:00): real-time electricity price is 0.28 yuan / kWh, weight ω=0.6×0.28 / (0.28+0.38)=0.255, and the contract income accounts for more than 70%.

[0078] The water energy value curve for a typical day (July 15) is as follows: Figure 2As shown, the daily hydropower value fluctuates between 620,000 and 1.15 million yuan per hour, with a peak at 7:00 PM (real-time electricity price of 0.62 yuan per kWh, available hydropower of 3.2 million kW·h), and a trough at 3:00 AM (electricity price of 0.25 yuan per kWh, available hydropower of 2.1 million kW·h). Under extreme rainstorm scenarios, despite a surge in water volume, ecological constraints result in a mere 15% increase in value (from a baseline of 900,000 yuan per hour to 1.035 million yuan per hour).

[0079] Through the coupling analysis of multiple risk factors, the identification of risk sources and transmission paths, a coupling network consisting of 12 risk factors in 4 categories was constructed:

[0080] Hydrological risk chain: extreme rainstorms → surge in runoff → reservoir water levels exceeding limits → increased water abandonment → decreased water energy utilization.

[0081] Market risk chain: New energy boom → real-time electricity prices fall → spot profits decrease; contract breach → power shortage requires high-priced electricity purchases → costs rise.

[0082] Equipment risk chain: turbine failure → shutdown for maintenance → reduced output → missed high-price period; monitoring system abnormality → scheduling delay → water waste.

[0083] Policy risk chain: increased ecological flow → reduced water volume for power generation; adjustment of peak and valley periods → reconstruction of returns during high-value periods.

[0084] The risk transfer coefficient matrix T shows the transmission intensity of hydrological risk to market risk. (e.g., heavy rain causing unit shutdown, affecting spot market declaration), feedback intensity of equipment risk on hydrological risk (For example, monitoring failures lead to delays in flood control).

[0085] Quantify the synergistic effects and take hydrological risks into consideration and market risks For example, the impact on water energy value when acting alone Ten thousand yuan, Ten thousand yuan, when working together Ten thousand yuan, synergy effect weight (All combined synergy effects) = 0.33, indicating that the risk amplification effect is significant when the two are superimposed.

[0086] Identify key risk sources and analyze the contribution of each factor to the return variance through the return risk dynamic prediction model:

[0087] Real-time electricity price fluctuations ( =0.38), extreme rainstorm ( =0.32) is the primary risk source, contributing more than 70% of return fluctuations;

[0088] Equipment failure ( =0.15) and ecological policy adjustments ( =0.10) is a secondary risk source, and its impact is prominent in complex scenarios.

[0089] Through the return-risk dynamic prediction model, the "input layer-attention mechanism layer-Bi-LSTM layer-output layer" architecture is adopted. The input data includes: dynamic hydropower value sequence (hourly data for the past 7 days), risk transfer coefficient matrix T (real-time update), synergy effect weight matrix Ω (based on the latest risk event correction). Taking the data on July 10, 2024 as an example, the sliding time window includes the previous 168 hours (7 days) After inputting the model, the probability distribution curve of returns for the next 72 hours is output.

[0090] like Figure 3 As shown in the figure, the forecast results analysis shows that the baseline scenario forecast shows a normal distribution of the profit probability, with a mean of 9.2 million yuan / day, a standard deviation of 850,000 yuan, and a 95% confidence interval of 7.5-10.9 million yuan; the key risk sources are real-time electricity price fluctuations (affecting 35% of the standard deviation) and runoff fluctuations (28%); the hydrological risk scenario forecast shows that extreme rainstorms cause the profit distribution to be right-skewed, with the mean increased to 10.5 million yuan / day (increased water volume), but the standard deviation widens to 1.2 million yuan (increased water abandonment risk and scheduling costs), and the probability of tail risk (profit <8 million yuan) increases from 5% to 12%; the composite risk scenario forecast shows that under the combination of heavy rain, falling electricity prices, and unit failures, the mean profit drops sharply to 7.8 million yuan / day, with a standard deviation of 1.8 million yuan, and a high-risk probability of profit <6 million yuan reaches 25%, requiring the activation of an emergency scheduling plan.

[0091] like Figure 4 As shown, through backtesting of historical data, the model's prediction error for the mean return is ±6.5%, the standard deviation prediction error is ±8.2%, and the accuracy rate of identifying key risk sources is 91.7%, which is significantly better than the traditional ARIMA model (error ±12%).

[0092] Generate risk-return trade-off decisions and plans, divide the return distribution into three scenarios: high (>10 million yuan / day), medium (8-10 million yuan / day), and low (<8 million yuan / day). Combined with key risk sources, a four-dimensional assessment system is constructed, as shown in the following table:

[0093] Risk Response Strategies Implementation cost (10,000 yuan / day) Revenue increase potential (10,000 yuan / day) Risk mitigation effect (standard deviation reduction) The spot ratio is reduced by 10% 20 -30 (conservative scenario) 15% Increase flood control storage capacity by 5% 15 -18 (water volume reduced) 22% (hydrological risk reduction) Increased frequency of equipment inspections 25 +12 (reduced downtime losses) 10% (equipment risk reduction) Purchase electricity price insurance 30 +25 (market risk hedging) 20%

[0094] Considering the peak load regulation capacity (at least 800MW output is required during peak load periods) and the reservoir water storage status (1.5 billion cubic meters of flood control storage capacity must be reserved at the end of the flood season), a Pareto solution set is generated through multi-objective planning:

[0095] Plan A (conservative): The spot ratio is 30%, and contracts are executed first. The average profit is 8.5 million yuan / day, and the standard deviation is 700,000 yuan. It is suitable for risk-averse decision makers.

[0096] Plan B (Balanced): The spot ratio is 50%, and the water allocation is adjusted dynamically. The average income is 9.5 million yuan / day, the standard deviation is 1 million yuan, and it is suitable for medium risk preferences.

[0097] Plan C (aggressive): The spot ratio is 70%, focusing on power generation during high-price periods, with an average income of 10.8 million yuan / day and a standard deviation of 1.5 million yuan, suitable for risk-seeking entities.

[0098] Input extreme scenario parameters through the visual interface, and the system automatically generates risk curves, benefit comparison tables, and sensitivity analysis;

[0099] Therefore, for the 2024 flood season, the A power station group performed the following optimized scheduling using this method:

[0100] Risk control: During extreme rainstorms, the spot ratio was reduced to 20% in advance to avoid spot revenue losses due to water abandonment. Actual revenue decreased by only 9% compared to the baseline forecast (traditional methods predicted an 18% decrease).

[0101] Increased revenue: During periods when the real-time electricity price is higher than 0.5 yuan / kWh, the output is dynamically increased by 15%, and spot revenue increases by 12% year-on-year;

[0102] Decision-making efficiency: Through a visual interface, multi-department collaborative analysis is achieved, and the scheduling decision-making time is shortened from an average of 4 times a day to 2 times, and the emergency response speed is increased by 50%.

[0103] This embodiment describes the specific implementation technology in detail, realizing that dynamic evaluation based on water energy value can accurately characterize the temporal and spatial differences in hydropower station revenue. The hybrid pricing mechanism effectively balances spot market opportunities and contract risks. The risk factor coupling network can reveal the transmission path between hydrology, market, equipment and policy. The quantification of synergy effects provides a scientific basis for complex risk assessment. Through the combination of deep learning models and attention mechanisms, the dynamic characteristics of risks can be captured in real time. The prediction accuracy meets actual scheduling needs. The risk-return trade-off decision matrix provides customized solutions for different risk preference entities, thereby improving the flexibility of power stations in dealing with complex scenarios.

[0104] The above are only preferred embodiments of the present invention, which do not limit the scope of protection of the present invention. For those skilled in the art, the present invention can be modified and varied in various ways. Any changes, modifications, replacements, integrations and parameter changes to these embodiments through conventional substitutions or that can achieve the same functions without departing from the principles and spirit of the present invention fall within the scope of protection of the present invention.

Claims

1. A method for measuring the risk of hydropower station revenue based on the value of hydropower, characterized by: The following steps are involved: S1: Build a dynamic water energy value assessment model. Collect real-time meteorological data, satellite remote sensing runoff data, and hydropower station equipment operation data within the basin. Combined with a distributed hydrological model, calculate the available water energy under different spatiotemporal scenarios. Based on real-time electricity prices in the power market, medium- and long-term contract prices, and ecological flow constraints, convert the available water energy into economic value to obtain dynamic water energy value. Specifically including: the precipitation in real-time meteorological data , evaporation , initial runoff of the basin in satellite remote sensing runoff data , and turbine efficiency in hydropower station equipment operation data , substituted into the space-time coupling equation of the distributed hydrological model, the formula is: ,in, is the regional precipitation-runoff conversion coefficient, is the temporal and spatial correction parameter of the terrain slope; according to the height difference of the hydropower station , the available water energy calculation formula is , calculate each spatiotemporal scene The available water energy under is the density of water, is the acceleration due to gravity, represents the spatial grid number, Represents a time node; Specifically including: obtaining real-time electricity prices in the electricity market , medium- and long-term contract electricity prices , determine the threshold of water energy availability for power generation based on ecological flow constraints ; Construct the price weight distribution function, the formula is: ,in is the real-time electricity price priority adjustment coefficient, through the mixed pricing formula , calculate each time node Dynamic water energy value ,in The available water energy at a time node is used to obtain the dynamic water energy value; S2: Establish a multi-risk factor coupling network, identify risk sources, construct risk transmission paths, quantify the marginal impact of each risk factor on hydropower value and power generation revenue, and calculate the synergistic effect weights among risk factors; Specifically include: establishing a risk factor set ,in Representative Class risk, the risk transfer coefficient matrix is ,in Indicates risk factors right The transmission intensity of the risk factor is calculated using the marginal impact quantification formula The value of hydropower The marginal impact of is: ,in The value of hydropower to risk factors The partial derivative of the risk factor is calculated by the synergistic effect weight formula and The synergistic effect weight between them is: ,in Risk factors and The impact on the value of water energy when they work together, They are The value of water energy when acting alone The amount of influence, Risk factors and The impact on the value of water energy when they work together, They are The value of water energy when acting alone The amount of influence, Conduct quantitative analysis of risk factor impacts and synergistic effects for the maximum impact among all risk factor combinations; S3: By using the reinforcement learning algorithm and taking the dynamic hydropower value and multi-risk factor coupled network data as input, a dynamic profit-risk prediction model is constructed to update the probability distribution curve of power generation profit within a preset time period in the future in real time and identify key risk sources; S4: Based on the prediction results and key risk sources output by the dynamic benefit-risk prediction model, a risk-benefit trade-off decision matrix is ​​constructed. Combined with the peak-shaving and frequency-regulating capabilities of the hydropower station and the water storage status of the reservoir, benefit optimization plans under different risk response strategies are generated; S5: Use a visual interface to display dynamic hydropower value, benefit-risk prediction results, and benefit optimization plans, support users to conduct multi-scenario simulation interactive analysis, and output visual risk curves and benefit comparison charts.

2. A method for measuring the risk of a hydropower station's revenue based on the value of hydropower according to claim 1, characterized in that: The risk sources identified in the multi-risk factor coupling network in S2 include hydrological risk, market risk, equipment risk and policy risk; among them, hydrological risk includes the risk of seasonal runoff fluctuation, the risk of abnormal water volume caused by extreme climate and the risk of hydrological changes caused by human activities in the basin; market risk covers the risk of real-time electricity price fluctuation in the power market, the risk of default in medium- and long-term contracts, and the risk of competition for consumption brought about by the grid connection of new energy; equipment risk includes the risk of failure of hydro-turbine generator sets, the risk of aging of power transmission and transformation equipment and the risk of failure of intelligent monitoring systems; policy risk involves the risk of increased compliance costs caused by environmental protection policies and the risk of changes in the income structure brought about by power system reform.

3. The method for measuring the risk of a hydropower station's revenue based on the value of hydropower according to claim 1, characterized in that: The dynamic prediction model of return risk in S3 includes a deep learning architecture of an input layer, an attention mechanism layer, a bidirectional long short-term memory network layer Bi-LSTM and an output layer; the obtained dynamic hydropower value sequence and risk factor transmission intensity matrix and synergy effect weight matrix are used as input layer data; the weights of different risk factors and time series data are adaptively allocated through the attention mechanism layer to highlight the key risk information that has a greater impact on the return; the bidirectional long short-term memory network layer is used to capture the time dependence and bidirectional dynamic characteristics of risk factors and hydropower value; in the output layer, a probability density function fitting module is used to output the probability distribution curve of power generation revenue in a preset time period in the future, where the probability density function adopts a mixed Gaussian model, and the model parameters are optimized through the expectation maximization algorithm to realize the dynamic prediction of power generation return risk.

4. A method for measuring the risk of a hydropower station's revenue based on the value of hydropower according to claim 1 or 3, characterized in that: In S3, the probability distribution curve of power generation revenue within a preset time period is predicted by the revenue risk dynamic prediction model, and key risk sources are identified, specifically: The data acquisition frequency is , each time new dynamic water energy value data is collected , risk transfer coefficient matrix and synergy weight matrix Then, it is combined with the historical data sequence to form a sliding time window data , input the return-risk dynamic prediction model, use the bidirectional long short-term memory network layer in the return-risk dynamic prediction model to extract features of the time window data, combine the attention mechanism layer to redistribute the weights of each risk factor at the current moment, and calculate the future preset time period through the mixed Gaussian model of the output layer Internal power generation income The probability density function of ,in is the number of Gaussian distributions, 、 、 Respectively Gaussian distribution in The mixing coefficient, mean and variance at each moment are used to update the probability distribution curve of power generation income. By calculating the contribution of each risk factor to the variance of the probability distribution curve, the contribution exceeding the set threshold is The risk factors identified as key risk sources are as follows: , For power generation income The variance of For risk factors.

5. The method for measuring the risk of a hydropower station's revenue based on the value of hydropower according to claim 1, characterized in that: In the said S4, the prediction results output by the dynamic prediction model of profit and risk and the key risk sources are used to construct a risk-benefit trade-off decision matrix, specifically: the power generation profit probability distribution curve in the prediction results is divided into quantiles to form a set of high, medium and low profit scenarios; for each profit scenario, the type and impact degree of the key risk source are combined to construct an evaluation index system including four dimensions: risk response strategy, implementation cost, profit improvement potential and risk mitigation effect; the hierarchical analysis method is used to determine the relative weight of each dimension, and the objective weight of each evaluation index is calculated to form a combined weight vector; the peak-shaving and frequency-regulating capacity of the hydropower station and the water storage status constraints of the reservoir are converted into boundary conditions of the decision space, and a multi-objective planning algorithm is used to generate a Pareto optimal solution set in the decision space, and each solution corresponds to a risk-benefit trade-off solution; the Pareto optimal solution set is clustered and analyzed to form a finite number of representative decision solution clusters, and each solution cluster corresponds to the optimal decision strategy under different risk preferences, so as to construct a risk-benefit trade-off decision matrix.

6. A method for measuring the risk of a hydropower station's revenue based on the value of hydropower according to claim 5, characterized in that: In S4, the peak-shaving and frequency-regulating capacity of the hydropower station and the water storage state constraints of the reservoir are converted into the boundary conditions of the decision space, and the multi-objective programming algorithm is used to generate the Pareto optimal solution set, and then the decision solution cluster is formed through cluster analysis. The specific method is: define the decision variable set ,in Represents the implementation intensity of different risk response strategies; the peak and frequency regulation capacity constraints of hydropower stations are expressed as , the reservoir water storage state constraint is expressed as ,in 、 The upper and lower limits of peak and frequency regulation power, 、 are the upper and lower limits of the reservoir’s water storage capacity, 、 The decision space is constructed for the influence coefficient of each strategy on the peak load power and water storage status. The multi-objective optimization is performed in the decision space by the non-dominated sorting genetic algorithm. The objective function is to maximize the power generation income and minimize the risk level. The Pareto optimal solution set is obtained through iterative update. For each solution in the Pareto optimal solution set, the Pareto optimal solution set is obtained. , calculate the Euclidean distance to other solutions, the formula is: ,in is the number of objective functions, For the The objective function value is set, the density threshold and neighborhood radius are set, and the distances less than the neighborhood radius and the density greater than the density threshold are disaggregated into the same cluster, forming a finite number of representative decision plan clusters, each cluster corresponding to decision plans with similar risk-return characteristics.

7. The method for measuring the risk of a hydropower station's revenue based on the value of hydropower according to claim 1, characterized in that: In S5, users can input custom scenario variables through the parameter adjustment panel, including but not limited to the amplitude of electricity price fluctuations and the probability of extreme hydrological events, call the profit-risk dynamic prediction model and the risk-profit trade-off decision matrix in real time, recalculate and generate visualization charts for the corresponding scenarios, and associate the various visualization components for linkage analysis.

Images