A high-altitude green electricity transaction volume decomposition method, system, device and medium

Abstract

The present application relates to the technical field of power market transaction and new energy dispatching, and discloses a green electricity transaction volume decomposition method, system, equipment and medium in high-altitude areas, wherein the method comprises: constructing a probability distribution model of historical electricity price, generating an initial electricity price path reflecting the fluctuation characteristics of spot electricity price through a hierarchical sampling method, and screening K representative scenarios and corresponding probabilities through a scene similarity calculation method to form a representative scenario set; a dynamic constraint system for real-time perception of power transmission capacity is constructed based on multi-dimensional power transmission constraint conditions; the multi-dimensional power transmission constraint conditions include wind and light output prediction, icing risk of power transmission lines, spot electricity price fluctuation, total green power and time period performance constraint; based on the representative scenario set and the dynamic constraint system, a green power distribution scheme for each hour is solved through a stochastic optimization method to realize green electricity transaction volume decomposition. The present application can realize safe performance of green electricity contracts and maximize economic benefits.

Description

Technical Field

[0001] This invention relates to the fields of electricity market trading and new energy dispatch technology, and in particular to a method, system, equipment and medium for decomposing green electricity trading volume in high-altitude areas. Background Technology

[0002] Driven by the "dual carbon" goals, medium- and long-term green electricity contracts have become the main source of revenue for new energy power plants. Green electricity trading refers to the buying and selling of renewable energy electricity (such as wind power and photovoltaic power) through market mechanisms to optimize clean energy allocation, promote energy consumption, and support low-carbon transformation. my country's green electricity market mainly includes two stages: medium- and long-term contract trading and spot trading.

[0003] The medium- and long-term green electricity trading market, through multi-year contracts, offers relatively fixed prices and provides long-term revenue guarantees for wind and solar power plants, but it cannot flexibly respond to short-term electricity price fluctuations. The electricity spot market includes the day-ahead market (pricing one day in advance) and the real-time market (pricing before immediate use, with volatile prices). Its price formation mechanism reflects the temporal and spatial scarcity of electricity, and can be expressed as:

[0004]

[0005] in, The spot electricity price for time period t; The variable cost of power generation for the marginal units of the system; Regional price differences caused by power transmission network congestion; Cost of power transmission losses; Price fluctuations caused by deviations in new energy output and load forecasting.

[0006] However, high-altitude regions (such as the Qinghai-Tibet Plateau and the western Sichuan Plateau) face fundamentally different technical challenges in implementing green electricity trading compared to plains areas due to their high altitude, cold climate, and weak power grid infrastructure.

[0007] (1) The power transmission network is weak and susceptible to weather conditions.

[0008] Transmission lines in high-altitude areas are long and difficult to maintain, making them highly susceptible to severe weather conditions such as icing and snowstorms. Icing on transmission lines can reduce their current-carrying capacity and even cause line outages, posing a significant potential risk of default on green electricity contracts. Unlike in plains areas, icing in high-altitude areas tends to occur during the nighttime low-temperature period, which coincides with peak wind power output, creating a spatiotemporal coupling of high output and high risk.

[0009] (2) Strong fluctuations in wind and solar power output and negative correlation with time series

[0010] The wind and light resources in high-altitude areas have unique time characteristics. Photovoltaic output shows an extreme form of high peak in the middle of the day and zero in the morning and evening. The valley wind effect leads to wind power output concentrated in the night to early morning period. These two major green energy sources show significant negative correlation within a day. More importantly, the period of high wind power output at night coincides with the period of lowest temperature and highest risk of line icing, which poses a challenge to traditional independent power forecasting and dispatching methods.

[0011] (3) Incomplete power market mechanism and lack of flexible compliance mechanism

[0012] Medium and long-term green electricity trading contracts are usually signed in advance, and lack flexible dispatching mechanism in the face of future electricity price fluctuations. Existing green electricity trading plans often allocate electricity based on static average forecasts or subjective experience, which cannot effectively respond to the icing risk of high-altitude transmission lines, the strong volatility of wind and light output, and the temporal and spatial differences of spot electricity prices, resulting in weak price signal response, low trading revenue, and high default risk in green power contract compliance.

[0013] Currently, the time decomposition of green power contract electricity is usually based on the following two technical paths:

[0014] The first type is static equal division method. The total amount of the contract is evenly allocated by time period, or proportionally decomposed according to the historical load curve. This method does not consider the fluctuation of spot electricity price and cannot respond to market price signals. In high-altitude areas, this method is particularly unsuitable because the day-night electricity price difference in this area can reach 3-5 times, and static allocation results in a large loss of arbitrage opportunities.

[0015] The second type is scenario optimization based on historical statistics. By fitting the probability distribution of historical electricity prices, possible future electricity price scenarios are generated, and the electricity allocation is carried out to maximize expected revenue. The core assumption of this method is that the statistical characteristics of future electricity prices are consistent with history. However, in high-altitude areas, this assumption does not hold because:

[0016] (1) The formation of electricity prices in high-altitude areas is significantly affected by transmission line icing. When the temperature drops below freezing at night, the current-carrying capacity of the line may decrease by 30%-50%, at which time the system is in a state of supply and demand tension, and the electricity price triggers the congestion pricing mechanism, which can be several times higher than normal conditions. The existing method models all historical data uniformly and cannot identify this hierarchical structure, resulting in distorted scenarios in extreme weather conditions.

[0017] (2) Existing optimization models usually set the transmission capacity as a fixed parameter (such as a line capacity ≤ 500 MW), which is basically reasonable in plain areas. But in high-altitude areas, transmission capacity is not an exogenous parameter, but a function of meteorological conditions. For example, in the normal period, 500 MW, and in the icing period, it may drop to 350 MW. The dynamic change of this constraint boundary makes the feasible region of the optimization problem evolve over time, and the existing static constraint model cannot handle it.

[0018] (3) The objective function of the existing method is usually the linear summation of the revenue of each period, which implies that the decisions of each period are independent of each other. But in high-altitude scenarios, if the online plan of a certain period exceeds the transmission capacity after icing, not only will there be a negative deviation penalty in that period, but it will also trigger a power grid safety event, which may limit the trading qualifications in subsequent periods. This cross-period risk transmission makes the revenue function exhibit path-dependent characteristics, and a simple linear model cannot capture it.

[0019] In addition, high-altitude areas have the following special power operation characteristics:

[0020] (1) Time sequence reversibility of wind and light output: photovoltaic output is concentrated in 10:00-16:00 (when the temperature is high, there is no icing risk); wind power output is concentrated in 00:00-06:00 (when the temperature is low, the icing risk is high).

[0021] (2) Asymmetry of price fluctuations: under normal weather, the price presents a regular peak-valley difference (1.5-2 times); but in icing weather, the night price may be extremely high due to supply gap (3-5 times of the daily average price), and the daytime price may be extremely low due to large photovoltaic output. This asymmetric fluctuation feature cannot be described by normal distribution or conventional statistical model.

[0022] Currently, the related patents of green electricity transaction volume decomposition are as follows:

[0023] CN117540986A discloses a wind-solar-storage collaborative planning method and system based on typical scenarios, which includes: obtaining scenario data, generating a typical day scenario, generating a typical day multi-random scenario, generating a typical uncertainty scenario, constructing a collaborative planning model, and solving the wind-solar-storage collaborative planning result. K The means clustering algorithm reduces the annual operation scenarios, represents the uncertainty of wind power, photovoltaic and load by Latin hypercube sampling, obtains typical scenarios based on the Kantorovich distance reduction method, and takes the minimum system investment and operation cost and the minimum annual generation cut-off of new energy as the target of the multi-objective planning model. The planning model is simplified into a linear single-objective planning problem by using the analytic hierarchy process. However, its application scenario is limited to installed capacity planning and does not involve special environments in high-altitude areas; uniform sampling under a single statistical population and equal-weight scenario compression are adopted without meteorological pre-classification and hierarchical modeling logic; the scenario generation does not have the ability to dynamically adjust with meteorological forecasts and cannot adapt to the characteristics of the tiered electricity price structure.

[0024] CN118864147A discloses a high-energy-consuming enterprise source-load collaborative optimization operation method considering wind and light uncertainty and demand response, including the following steps: obtaining wind and light meteorological historical data set, preprocessing the data set; constructing wind and light uncertainty source output representation model based on Weibull distribution and Beta distribution, and generating wind and light output historical data set; using Latin hypercube sampling method combined with Kantorovich scene reduction method to generate and reduce the wind and light output historical data; constructing price elasticity coefficient and price elasticity coefficient matrix of electricity and heat price; classifying enterprise load side, constructing energy consumption cost model; introducing objective function and constraint condition of optimization operation, obtaining optimization operation model; inputting wind and light output and load parameters into the optimization operation model to obtain the actual result of high-energy-consuming enterprise source-load collaborative optimization operation. However, the application scenario is enterprise energy optimization and does not target the meteorological-transmission-electricity price coupling problem in high-altitude areas; uniform scenario compression is also used without considering the differentiated reservation of scenarios in high-risk periods; there is a lack of joint modeling mechanism of meteorological conditions and electricity price and transmission constraints, and the scenario generation does not have dynamic adaptability.

[0025] CN115545768A is a large hydropower cross-provincial and cross-regional day-ahead stochastic bidding method considering contract decomposition. The method first proposes a price scenario analysis method considering time period correlation from the perspective of time period volatility; then constructs a stochastic expectation model with the maximum total expected revenue of the medium and long-term contract market and the day-ahead market as the objective function; secondly, a medium and long-term contract decomposition method based on peak shaving demand classification is proposed, and a cross-provincial and cross-regional stochastic coordination optimization model that coordinates the peak shaving performance and economic benefits is constructed; finally, the model is transformed into a MILP model by using multi-objective transformation and linearization method, and the model is solved by using optimization solving software. However, the transmission constraint is a static exogenous fixed parameter, and the constraint boundary does not change with external meteorological conditions; the application scenario focuses on large hydropower cross-provincial and cross-regional day-ahead bidding, and the decision dimension is spatial power distribution, without considering time series risk avoidance in high-altitude areas; the objective function only pursues pure economic optimization without quantifying physical default risk, and there is no meteorological-physical-economic three-domain coordination mechanism.

[0026] In summary, the existing green electricity transaction volume decomposition method has the following problems when applied in high altitude areas:

[0027] (1) Unable to establish the correlation mechanism between the probability distribution of electricity price and meteorological conditions, resulting in the loss of representativeness of the generated electricity price scenario under extreme weather;

[0028] (2) Unable to dynamically adjust the power transmission constraint with the meteorological condition, resulting in an unfeasible optimization result at the physical level;

[0029] (3) Unable to quantify the reduction effect of physical default risk on economic benefits, resulting in overly aggressive decision-making and high penalties in actual performance. SUMMARY

[0030] To solve the above problems, the present application proposes a high-altitude green electricity transaction volume decomposition method, system, device and medium, which deeply couples meteorological forecasting, physical constraints and market electricity prices to model, and can seek an economically optimal green electricity transaction strategy within the dynamically changing safety boundary. The present application establishes a meteorological-electricity price joint probability modeling mechanism, a power transmission constraint dynamic reconstruction mechanism and a risk-aware robust optimization mechanism to solve the core problems of existing technology in this scenario, such as distorted electricity price scenarios, static physical constraints and missing risk assessment, and to achieve safe performance and economic benefit maximization of green electricity contracts.

[0031] The technical solution adopted by the present application is as follows:

[0032] A high-altitude green electricity transaction volume decomposition method, comprising:

[0033] Constructing a probability distribution model of historical electricity price, generating an initial electricity price path reflecting the volatility characteristics of spot electricity price through a hierarchical sampling method, and selecting K representative scenarios and corresponding probabilities through a scenario similarity calculation method to form a representative scenario set;

[0034] Based on multi-dimensional power transmission constraint conditions, a dynamic constraint system with real-time awareness of power transmission capacity is constructed; the multi-dimensional power transmission constraint conditions include wind and light output prediction, power transmission line icing risk, spot electricity price fluctuation, total green power and time period performance constraint;

[0035] Based on the representative scenario set and the dynamic constraint system, a green power distribution scheme for each hour is solved through a stochastic optimization method to realize green electricity transaction volume decomposition.

[0036] Further, the method for constructing the probability distribution model of the historical electricity price comprises the following steps: collecting historical electricity price data and synchronous meteorological data of a target area, dividing historical electricity price samples into a normal working condition electricity price set and an icing working condition electricity price set according to a power transmission line icing critical condition, and constructing the probability distribution model of the historical electricity price by using an electricity price-meteorological condition joint probability modeling method.

[0037] Further, the method for generating an initial electricity price path reflecting fluctuation characteristics of spot electricity prices by using a hierarchical sampling method comprises the following steps: constructing a non-parametric probability density function for the normal working condition electricity price set and the icing working condition electricity price set, respectively, and forming a continuous probability density curve by using a Gaussian kernel function to locally diffuse each historical price point as a probability mass center through a historical sample kernel weight accumulation method.

[0038] Further, the method for screening K representative scenes and corresponding probabilities by using a scene similarity calculation method to form a representative scene set comprises the following steps: setting a time period importance weight, including a risk indicator function and a risk weighting coefficient; calculating a weighted distance between scenes, and then determining K representative scenes through an iterative clustering compression process.

[0039] Further, the method for constructing a dynamic constraint system for real-time sensing of power transmission capacity based on multi-dimensional power transmission constraint conditions comprises the following steps: fitting a capacity reduction function based on historical operation data of a power grid in a target area, establishing a mapping relationship between temperature prediction and power transmission capacity, and generating an icing-triggered capacity reduction model; the capacity reduction function adopts a piecewise linear or exponential form.

[0040] Further, the method for constructing a dynamic constraint system for real-time sensing of power transmission capacity based on multi-dimensional power transmission constraint conditions further comprises the following steps: converting icing state judgment into a continuous differentiable constraint by using a penalty factor method, introducing an auxiliary variable to represent a time period icing state, and enabling an upper limit of power transmission capacity to be automatically calculated and adjusted inside the model according to meteorological prediction.

[0041] Further, the method for solving an hourly green power allocation scheme by using a random optimization method comprises the following steps: taking maximum multi-scenario expected revenue as an objective function, wherein the objective function comprises a green electricity contract price, a scenario spot electricity price and a negative deviation penalty cost item, and the negative deviation penalty cost item comprises a penalty coefficient and a time period actual available power generation amount; and optimizing by using a plurality of constraint types as constraint conditions, wherein the plurality of constraint types comprise an energy balance constraint, a total contract amount constraint, a decomposition amount boundary constraint, a dynamic power transmission constraint and an icing state logic constraint.

[0042] A green electricity transaction volume decomposition system in a high-altitude area comprises the following steps:

[0043] The representative scenario set construction module is configured to construct a probability distribution model of historical electricity prices, generate an initial electricity price path reflecting the fluctuation characteristics of spot electricity prices through a hierarchical sampling method, and filter K representative scenarios and corresponding probabilities through a scenario similarity calculation method to form a representative scenario set.

[0044] The dynamic constraint system construction module is configured to construct a dynamic constraint system for real-time perception of power transmission capacity based on multi-dimensional power transmission constraint conditions, including wind and solar power output prediction, power transmission line icing risk, spot electricity price fluctuation, total green power amount, and time period fulfillment constraint.

[0045] The green electricity transaction volume decomposition module is configured to solve the green power allocation scheme of each hour through a stochastic optimization method based on the representative scenario set and the dynamic constraint system, and realize the decomposition of the green electricity transaction volume.

[0046] A computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the high-altitude region green electricity transaction volume decomposition method when executing the computer program.

[0047] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the high-altitude region green electricity transaction volume decomposition method.

[0048] The beneficial effects of the present application are:

[0049] (1) The weather response capability of the electricity price scenario is improved. The existing method generates an electricity price scenario based on historical full sample statistics. When the future weather condition deviates from the historical average state (such as continuous low temperature icing), the representativeness of the scenario decreases sharply. The present application can respond to weather forecasts through meteorological hierarchical modeling, and automatically generate electricity price distribution under icing conditions when low temperature is forecasted, ensuring the fidelity of the scenario under extreme conditions.

[0050] (2) Active avoidance of physical constraint violation risk. The power transmission constraint of the existing method is an exogenous fixed parameter, and the optimization result may be theoretically optimal but physically unfeasible. The present application introduces weather forecasts into the constraint generation link to realize dynamic perception of power transmission capacity.

[0051] (3) Robustness of comprehensive income is enhanced. The existing method pursues the optimal solution under single scenario or deterministic prediction, and the income decreases significantly when the actual situation deviates from the prediction. The present application adopts multi-scenario probability weighted optimization to achieve robust optimal solution in a statistical sense.

[0052] (4) Targeted adaptation of high-altitude scenarios. Existing general methods do not consider the three coupled characteristics of "meteorological extremity, power transmission vulnerability, and asymmetric electricity price" when applied to high-altitude areas. The present invention designs a complete technical chain for this specific scenario: from meteorological-electricity price joint modeling, to dynamic constraint construction, to risk-aware optimization, each link embodies the adaptability to high-altitude characteristics, filling the gap in existing technology in this scenario.

[0053] (5) Compared with CN117540986A, the present invention adapts to the three coupled characteristics of high altitude, and can solve the problem of electricity price stratification; meteorological stratified modeling + dynamic scenario generation, rather than single uniform sampling; risk-oriented non-uniform compression, preferentially preserving high-risk period differences.

[0054] (6) Compared with CN118864147A, the present invention constructs a meteorological-electricity price joint model to adapt to the uncertainty of high-altitude electricity prices; weighted amplification of high-risk period differences can avoid the limitations of uniform compression; supports time series decomposition and risk quantification of green electricity contracts, and adapts to more accurate scenarios.

[0055] (7) Compared with CN115545768A, the present invention dynamically endogenously generates power transmission constraints, which are adjusted in real time according to weather forecasts; the objective function contains a risk penalty term, which can achieve three-domain coordinated optimization; it focuses on time dimension risk avoidance, which is suitable for high-altitude green electricity trading needs. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 is a high-altitude region green electricity transaction volume decomposition method flowchart of embodiment 1 of the present invention.

[0057] Figure 2 is an electricity price scenario reduction flowchart of embodiment 1 of the present invention.

[0058] Figure 3 is a model solution flowchart of embodiment 1 of the present invention.

[0059] Figure 4 is a wind and light output prediction chart of embodiment 3 of the present invention.

[0060] Figure 5 is a green power load chart of embodiment 3 of the present invention.

[0061] Figure 6 is a temperature chart of a certain area of embodiment 3 of the present invention.

[0062] Figure 7 is an electricity price scenario simulation result chart of embodiment 3 of the present invention.

[0063] Figure 8 is an electricity price scenario reduction result chart of embodiment 3 of the present invention.

[0064] Figure 9 is a green power decomposition result graph of embodiment 3 of the present application.

[0065] Figure 10 is a green power decomposition result and green power load comparison graph of embodiment 3 of the present application.

[0066] Figure 11 is a green power decomposition comparison result graph of embodiment 3 of the present application. DETAILED DESCRIPTION

[0067] In order to have a more clear understanding of the technical features, objectives and effects of the present application, the specific embodiments of the present application will now be described. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application, that is, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0068] Embodiment 1

[0069] As shown in Figure 1 , the present embodiment provides a green electricity transaction volume decomposition method in high-altitude areas, comprising:

[0070] A probability distribution model of historical electricity price is constructed, an initial electricity price path reflecting the fluctuation characteristics of spot electricity price is generated through a hierarchical sampling method, and K representative scenarios and corresponding probabilities are screened through a scene similarity calculation method to form a representative scenario set;

[0071] A dynamic constraint system with real-time perception of transmission capacity is constructed based on multi-dimensional transmission constraint conditions; wherein the multi-dimensional transmission constraint conditions include wind and light output prediction, icing risk of transmission line, fluctuation of spot electricity price, total amount of green electricity and time period performance constraint;

[0072] Based on the representative scenario set and the dynamic constraint system, a green electricity allocation scheme for each hour is solved through a stochastic optimization method to realize the decomposition of green electricity transaction volume.

[0073] It should be noted that the present method comprehensively considers market electricity price, new energy output and the line icing risk specific to high-altitude areas, and through an optimized decomposition process, an economic benefit optimal time-sharing performance plan under the premise of safety is formulated for the green electricity contract. Specifically, the present method first generates multiple groups of electricity price scenarios through hierarchical sampling to cope with electricity price fluctuation, and compresses them into representative scenarios through weighted distance to reduce computational complexity; then, the total amount of green electricity contract signed by the user is decomposed into each time period in combination with the uncertainty of spot market electricity price and the dynamic change of transmission capacity to form a time-sharing green electricity supply plan.

[0074] Preferably, such as Figure 2 As shown, the electricity price scenario reduction process in this embodiment is as follows: First, K initial scenarios are randomly selected as cluster centers. Then, through an iterative optimization process, the weighted distance matrix between all scenarios is calculated, and the original scenarios are assigned to the nearest cluster center to form clusters. The cluster center positions are then updated based on the weighted average of the scenarios within each cluster. This iterative optimization process is repeated until the changes in the cluster center positions tend to stabilize (reaching a preset threshold or the maximum number of iterations). Finally, a representative set of reduced scenarios and their probability weights are output. This algorithm significantly reduces computational complexity while preserving key statistical features of electricity prices by minimizing the transmission cost between the original and reduced distributions.

[0075] Preferably, such as Figure 3 The diagram illustrates the model solution process for this embodiment, which optimizes the solution for green electricity trading based on wind and solar power output forecasts, electricity market information, and key weather forecasts. This embodiment maximizes the overall benefits of green electricity in both the green electricity trading market and the spot market through an integrated optimization decision-making process that incorporates icing risk assessment.

[0076] Specifically, the model solving process in this embodiment includes: model establishment and multi-dimensional information input, integrating market economic information, new energy output forecasts, and weather forecasts; quantification of market uncertainty, using hierarchical sampling and weighted clustering to generate and reduce scenarios; dynamic optimization solution, inputting typical electricity price scenarios, output forecasts, and weather forecasts into an optimization model with embedded icing risk assessment logic, and calling the solver for time-segmented optimization; finally, outputting a set of optimal decisions, including 24-hour time-segmented green electricity trading curves, future physical risk assessment results for each time period, and expected return analysis.

[0077] Example 2

[0078] This embodiment is based on embodiment 1:

[0079] This embodiment provides a method for decomposing green electricity trading volume in high-altitude areas. It dynamically decomposes green electricity trading volume by combining the characteristics of wind and solar power output in high-altitude areas, the risk of transmission line icing, and the uncertainty of spot electricity prices. This method achieves optimal performance of green electricity contracts through the organic combination of electricity price scenario generation, probability compression, and stochastic optimization solution mechanisms. Specific details are as follows.

[0080] 1. The Basic Logic of Electricity Market Mechanism and Green Electricity Trading

[0081] Green electricity trading involves the coordinated operation of the medium- and long-term contract market and the spot market. Medium- and long-term contracts guarantee long-term returns with fixed prices, while the spot market reflects short-term supply and demand relationships through day-ahead and real-time pricing.

[0082] Day-ahead market refers to the actual electricity day-ahead 24 hours completed transaction, mainly based on the predicted load and generating capacity, through centralized bidding to form time-of-use electricity price (mostly domestic node price or regional price). Its main application scenarios include: thermal power plant in advance to arrange unit start-stop plan, new energy power station to declare the next day generating capacity, electricity sales company to buy basic electricity to hedge risks, etc. Its price fluctuation is relatively flat, but it needs to bear the prediction deviation penalty.

[0083] The transaction time of real-time market is 1 hour to 15 minutes before the actual operation period (mostly 15 minutes level transaction in China). It is mainly used to correct the prediction deviation of day-ahead market, handle unexpected factors such as sudden failure and weather changes. Its price fluctuation is violent (the highest / lowest price of domestic real-time market can reach 5 times of day-ahead electricity price), reflecting the real-time congestion cost and standby capacity value of the system.

[0084] Green power contract performance usually adopts fixed price for settlement, but in the actual power supply process, it still needs to arrange the performance electricity reasonably according to the wind and light output capacity, transmission capacity and spot electricity price fluctuation in each period. Otherwise, it may appear large performance deviation or loss of income.

[0085] The essence of spot market price mechanism is to form a dynamic price signal reflecting the time and space scarcity of electricity commodity through short-term supply and demand balance and marginal cost pricing principle. The price composition includes:

[0086] Energy Component is the fuel cost of the last dispatched unit (marginal unit) in the system, among which, the fuel cost of coal-fired unit is about 0.25~0.40 yuan / kWh, the fuel cost of gas-fired unit is about 0.50~0.80 yuan / kWh, the marginal cost of wind power / photovoltaic is zero, which preferentially clears and lowers the benchmark price.

[0087] The main reason for the generation of Congestion Component is that the regional supply and demand imbalance is caused by the insufficient transmission path capacity.

[0088]

[0089] In the formula, is the congestion cost, is the power transmission distribution factor of node i to line k; is the shadow price of line k, and n is the number of lines.

[0090] Loss Component adopts node loss factor (Loss Factor) to convert:

[0091]

[0092] In the formula, For network loss cost, For reference node marginal price; For the comprehensive network loss rate from the sending end to the reference node (usually 2%~8%).

[0093] In high-altitude areas, wind power is mainly concentrated at night, and photovoltaic output is concentrated at noon. The price fluctuation is more susceptible to icing transmission constraints, exacerbating the phenomenon of night high and day low or even jumping, which brings uncertainty risk to the execution of green power contracts.

[0094] 2. Technical path

[0095] Uncertainty quantification and representative scenario generation: Construct a probability distribution model of historical electricity prices, and generate a large number of initial electricity price paths that can fully reflect the characteristics of spot electricity price fluctuations through hierarchical sampling method. In order to reduce the computational complexity of subsequent optimization while ensuring model accuracy, the invention further adopts a scene similarity calculation method to select K most representative typical scenarios and their probabilities. The final output of this link is a simplified scenario set that can retain key price fluctuation risks and has a controllable calculation scale, providing a data basis for subsequent optimization decisions.

[0096] Dynamic decomposition optimization modeling: Under multiple electricity price scenarios, considering wind and light output prediction, transmission line icing risk, spot electricity price fluctuation, green power total amount and time period fulfillment constraints, the hourly green power allocation scheme is solved through stochastic optimization method to maximize contract fulfillment income and reduce fulfillment risk.

[0097] This method not only has strong theoretical innovation in green power trading, but also has high practical application value in high-altitude area power grid environment or new energy microgrid. By reasonably inputting data, reliable electricity price scenarios and clear optimization model, this method avoids complex algorithm design and can efficiently adapt to complex power market and power grid operation environment.

[0098] 3. Technical details

[0099] 3.1 Dynamic modeling of electricity price probability structure under meteorological condition stratification

[0100] In view of the characteristics that the electricity price formation mechanism in high-altitude areas is significantly affected by meteorological conditions, this embodiment proposes an electricity price-meteorological joint probability modeling method.

[0101] 3.1.1 Meteorological condition annotation and stratification of historical data

[0102] Collect historical electricity price data and contemporaneous meteorological data (temperature, humidity, wind speed, etc.) of the target area, and stratify the historical samples according to the critical conditions of transmission line icing. Specifically:

[0103] Identify the period when icing occurs in history (criterion: temperature is lower than threshold and relative humidity is higher than threshold );

[0104] Divide the historical price samples into two subsets: normal condition price set and icing condition price set .

[0105] 3.1.2 Construction of stratified probability density function

[0106] Construct probability density models for the two subsets respectively. To avoid distortion of parameter distribution assumptions (such as high-altitude prices often showing multi-peak, thick-tail characteristics), this embodiment adopts a non-parametric modeling approach:

[0107] For normal conditions, establish the probability density function . By kernel weight accumulation of historical samples, each historical price point is regarded as a probability mass center, and a Gaussian kernel function is used for local diffusion to form a continuous probability density curve:

[0108] (1)

[0109] where is the number of normal condition samples, is the Gaussian kernel function, is the smoothing parameter (bandwidth).

[0110] For icing conditions, establish the probability density function , also using the kernel weight method, but the samples come from the icing period:

[0111] (2)

[0112] 3.1.3 Conditional scenario generation

[0113] According to the future 24-hour weather forecast, select the corresponding probability density function for each period to generate conditional price scenarios. The specific steps are:

[0114] (1) For each period t, according to the forecast temperature to determine the working condition type:

[0115] If , select ;

[0116] If , select .

[0117] (2) For the selected probability density function, calculate its cumulative distribution function F(λ):

[0118] (3)

[0119] (3) Divide the cumulative probability interval [0, 1] into B sub-intervals (B is the initial number of scenarios), and generate a uniform random number in each sub-interval:

[0120] (4)

[0121] (4) Generate price samples by inverse function mapping:

[0122] (5)

[0123] Repeat the above process to generate B sets of price time series curves , each curve representing a possible future price evolution path.

[0124] The hierarchical modeling mechanism of this embodiment can identify and retain the impact of weather conditions on price distribution. For example, when the weather forecast shows that low temperature will occur at night, the generated night price scenarios will automatically contain more high price possibilities, reflecting the expected supply and demand tension caused by icing.

[0125] 3.2 Representative scenario screening to retain extreme risk characteristics

[0126] The B sets of scenarios generated initially (usually B = 100-200) will result in excessive computation if directly used for subsequent optimization. However, simple random sampling may lose critical information (such as extreme high or low price scenarios). This embodiment proposes a scenario compression method for risk retention.

[0127] 3.2.1 Time period importance weight setting

[0128] Not all time periods have the same impact on decision-making. In high-altitude scenarios, the icing risk period (usually the low temperature period at night) has a stronger constraint on decision-making, so the price scenarios of these periods need higher fidelity.

[0129] Define the time period weight :

[0130] (6)

[0131] where is the risk indicator function of time period :

[0132] (7)

[0133] where is the risk weighting coefficient (usually 1-2).

[0134] 3.2.2 Inter-scenario diversity measure

[0135] Define the weighted distance between two scenarios and :

[0136] (8)

[0137] The distance measure has the feature that the price difference is amplified in high-risk periods; in normal periods, the contribution of the price difference to the total distance is small.

[0138] 3.2.3 Iterative clustering compression

[0139] The following iterative procedure is used to select K representative scenarios (usually K = 8 ~ 15):

[0140] (1) Initialization: randomly select K scenarios as clustering centers ;

[0141] (2) Scenario assignment: for each original scenario , calculate its distance to each clustering center , assign to the nearest center to form a cluster ;

[0142] (3) Center update: for each cluster , calculate the probability-weighted average of the scenarios within the cluster as the new center:

[0143] (9)

[0144] At the same time, update the probability weight of the cluster: ;

[0145] (4) Convergence criterion: if the position change of all clustering centers is less than a threshold (such as ), stop iteration; otherwise, return to step (2).

[0146] By introducing the period weight , this compression method will preferentially retain scenarios that have differences in high-risk periods, i.e., icing periods. Even if some scenarios have similar prices in normal periods, as long as their prices in icing periods have significant differences, they will not be merged. This ensures that the compressed scenario set still maintains sufficient diversity in critical risk periods.

[0147] 3.3 Constraint system construction with dynamic perception of transmission capacity

[0148] The existing optimization model usually takes the transmission capacity as a fixed constraint. This embodiment proposes a mechanism for dynamically adjusting the constraint boundary according to the weather forecast.

[0149] 3.3.1 Ice-coating triggered capacity degradation model

[0150] A mapping relationship between temperature forecast and transmission capacity is established. Based on the historical operation data of the regional power grid, the capacity degradation function is fitted:

[0151] (10)

[0152] wherein, is the degradation amplitude function, which can be in the form of piecewise linear or exponential. A typical piecewise linear model is:

[0153] (11)

[0154] wherein, , is the capacity reduction coefficient, which is calibrated according to the line parameters and historical data.

[0155] 3.3.2 Continuous processing of logical constraints

[0156] This embodiment converts the ice-coating state judgment into a continuous and differentiable constraint form. An auxiliary variable is introduced to represent whether the period t is covered with ice. The correlation is established by the big M method:

[0157] (12)

[0158] wherein, M is a large positive number, is a small positive number (such as 0.01).

[0159] The transmission constraint is represented as:

[0160] (13)

[0161] wherein, is the total on-grid power of period t.

[0162] The key of this mechanism is that the upper limit of transmission capacity is no longer a fixed parameter input from outside the model, but a variable automatically calculated inside the model according to the weather forecast. When the temperature forecast of a certain period in the future is updated, the capacity upper limit of this period will be automatically adjusted, and the constraint boundary will change accordingly.

[0163] 3.3.3 Time sequence decoupling characteristics of feasible region

[0164] Unlike traditional methods that impose uniform transmission constraints on all periods, the constraints of this embodiment have time sequence independence:

[0165] (1) The constraint boundary of each time period t is determined by the weather forecast of that time period, independent from other time periods;

[0166] (2) The constraint is automatically tightened during the night low temperature time period, and remains loose during the day normal time period;

[0167] (3) This segmented dynamic characteristic enables the optimizer to make fine-tuned decision adjustments in the time dimension.

[0168] Compared with the global uniform constraint of the prior art, the dynamic constraint of the embodiment can adjust the safety boundary according to the actual physical condition, allow greater on-grid power during the safe time period to increase the income, and automatically reduce the on-grid plan to avoid default during the risk time period, thereby achieving dynamic balance between safety and economy.

[0169] 3.4 Robust optimization decision with multi-scenario probability weighting

[0170] On the basis of the foregoing representative scenario set and dynamic constraint system, an optimization model for green electricity transaction volume decomposition is constructed.

[0171] 3.4.1 Definition of decision variables

[0172] The main decision variables include:

[0173] : the electricity volume (unit: MW) used to fulfill the green electricity contract in time period t;

[0174] : the spot market transaction electricity volume in time period t under scenario k (unit: MW, positive value indicates selling electricity, and negative value indicates buying electricity);

[0175] : the icing state indication variable of time period t (0 or 1).

[0176] 3.4.2 Objective function

[0177] The target is to maximize the multi-scenario expected income:

[0178] (14)

[0179] wherein, is the probability weight of scenario k; is the green electricity contract price; is the spot electricity price of time period t under scenario k; is the negative deviation penalty cost.

[0180] Modeling of the penalty cost term: ​​​

[0181] (15)

[0182] wherein, is a penalty coefficient (usually taken as 1.5-2.0), is the actual available generation at time period t.

[0183] 3.4.3 Constraints

[0184] (1) Energy balance constraint:

[0185] (16)

[0186] wherein, and are the predicted output power of photovoltaic and wind energy at time t, respectively.

[0187] (2) Total contract constraint:

[0188] (17)

[0189] wherein, is the total green power output of the planning decomposed to dispatch cycle T; Δt is the time period length.

[0190] (3) Decomposition amount boundary constraint:

[0191] (18)

[0192] (4) Dynamic transmission constraint:

[0193] (19)

[0194] (5) Icing state logic constraint:

[0195] (20)

[0196] 3.4.4 Model solution

[0197] The above model is a mixed integer linear programming problem (containing continuous variables , and 0-1 integer variables ). An optimization solver CPLEX can be used for solution, and the optimal green power decomposition scheme of each time period is output.

[0198] Risk-aware decision adjustment: when a time period is determined to be high icing risk, the upper limit of transmission of the time period is automatically reduced, and the optimizer will actively reduce the on-grid power of the time period, even if the electricity price of the time period is high, part of the income will be given up to avoid physical default risk.

[0199] Cross-time period resource reallocation: Since the total contract amount is fixed, if internet access is reduced during certain periods due to ice restrictions, the optimizer will automatically transfer this power to other safe periods, thereby achieving time-series optimized resource allocation.

[0200] Multi-scenario probability balance: The optimization result does not depend on a single electricity price forecast, but seeks the optimal in the expected sense under multiple possible scenarios, reducing the impact of forecast error on decision-making.

[0201] Example 3

[0202] This embodiment is based on embodiment 2:

[0203] This embodiment provides a method for decomposing green electricity trading volume in high-altitude areas, taking a wind and solar energy base in a high-altitude area as an example for illustration, as detailed below.

[0204] 1. Case Background

[0205] Contract Information: Wind farms and photovoltaic power plants have signed green electricity trading contracts with users, stipulating a total green electricity trading volume of 481 MWh / day. Based on the trading situation in the pilot areas of the green electricity market, the green electricity price is set at 506.6 yuan / MWh, and the environmental premium is set at 91.3 yuan / MWh.

[0206] Market information: The spot market consists of both day-ahead and real-time markets, with electricity prices fluctuating between 200-700 yuan / MWh.

[0207] Risk parameters: Based on the power grid operation experience in this high-altitude area, a critical temperature threshold for line icing is set. The temperature is 0°C. When the model determines that icing has occurred, the transmission capacity decreases by [amount missing]. It is set at 8MW.

[0208] Input information:

[0209] The region's landscape forecasting output is as follows: Figure 4 As shown, the peak output of photovoltaic power is 25MW during the daytime (12 hours) and 0MW at night; the peak output of wind power is 20MW during the nighttime (3 hours) and an average of 5-10MW during the day.

[0210] The region's demand for green electricity is as follows Figure 5 As shown, the average daily load in this region is 20MW, with no obvious peaks or valleys.

[0211] Temperature forecast for the region as follows Figure 6As shown, the curve in the figure is the hourly temperature forecast for the next 24 hours. According to the forecast, the temperature will drop to -2°C at 2-4 am, which is lower than the set 0°C icing threshold, and the temperature is above the threshold at other times. This temperature data will be used as a key physical input for the embedded icing risk assessment model in this embodiment to dynamically determine the physical safety state of the line.

[0212] 2. Weather stratified price scenario generation

[0213] (1) Historical data stratification:

[0214] Identify icing periods (nighttime periods with temperature <0°C and relative humidity >80%) from historical price data;

[0215] Normal operating condition sample number (about 10 years of normal weather data);

[0216] Icing condition sample number (about 10 years of icing weather data).

[0217] (2) Stratified probability density fitting:

[0218] For normal operating conditions, a unimodal distribution with prices concentrated in 280-450 yuan / MWh is fitted;

[0219] For icing conditions, a right-skewed distribution with prices concentrated in 400-700 yuan / MWh is fitted, with a significant high-price tail.

[0220] (3) Conditional scenario generation:

[0221] For 2-4 am (forecast temperature -2°C), sample from the icing condition distribution ; for other periods, sample from the normal condition distribution ; generate 100 initial price scenarios as shown, with price fluctuations ranging from normal, high, to low scenarios. Figure 7

[0222] 3. Risk-oriented scenario reduction

[0223] (1) Time period weight setting:

[0224] Set the weight for 2-4 am (high-risk period) ;

[0225] Set the weight for other periods (normal period) .

[0226] (2) Iterative clustering compression:

[0227] ​From 100 initial scenarios to 10 representative scenarios, converged after 5 rounds of iteration.

[0228] (3) Compression effect verification:

[0229] The 10 compressed scenarios maintain the key characteristics of the original 100 scenarios in the night 2-4 price distribution (mean deviation <3%, standard deviation <5%);

[0230] And in the daytime normal period, the deviation of the compressed scenario and the original scenario is slightly larger (mean deviation <8%), but the impact on decision-making is small, because these periods are not the bottleneck of physical constraints.

[0231] Based on the scenario compression of the electricity price scenario, the 10 typical day-ahead settlement price scenarios are as shown in Figure 8 .

[0232] Typical scenarios include: daytime high price (8-11 hours price >500 yuan / MWh), night stable (price 280-420 yuan / MWh), etc. And the probability distribution of 10 scenarios is less than 5% different from the original set, and the calculation amount is reduced by 90%.

[0233] 4. Green electricity decomposition optimization results

[0234] Based on the wind and light forecast value and the day-ahead market settlement price green power demand, the green power decomposition results are as shown in Figure 9 .

[0235] As shown in the figure, the wind and light energy base will decompose green power to non-peak price period in the limited decomposition area according to the day-ahead price forecast, resulting in the final green power decomposition curve being negatively correlated with the day-ahead price forecast.

[0236] Table 1-Optimization results

[0237]

[0238] In the period of safe grid physical conditions and no icing risk, the model shows sharp economic arbitrage ability. For example, in the 12-16h period when the spot price is low, the model decision will prefer to use large photovoltaic power to fulfill the green power contract with higher price. Conversely, in the peak price period such as 8-11h, the model will appropriately reduce the decomposition amount of green power contract and put part of the power into the spot market to capture higher marginal revenue.

[0239] However, the most fundamental difference between the present embodiment and the prior art is that it can actively avoid risks according to the input physical conditions. As shown in Table 1, at night 2-3h, although the wind power output is at the peak, the model endogenously judges that there is a high icing risk according to the input low temperature forecast, and actively limits the on-grid power within a safe range. This decision to sacrifice short-term power generation in exchange for grid physical safety reflects the coordination ability of the present embodiment method between physical feasibility and economic optimality.

[0240] 5. Result comparison and verification

[0241] (1) Comparison of green power decomposition and green power load

[0242] After considering the uncertainty of spot electricity price, the comparison between the green power decomposition and the original green power load is as follows: as shown in Figure 10 , the green power decomposition curve obtained after decomposition for the purpose of maximizing the revenue of wind power and photovoltaic in green power market and power spot market and the comparison of the original green power load are shown.

[0243] (2) Comparison of green electricity transaction deviation

[0244] The deviation of green electricity after decomposition and without decomposition is as follows:

[0245] As shown in Figure 11 , the green electricity deviation based on the green electricity decomposition curve and the green electricity deviation based on the basic green electricity demand are shown. The deviation after decomposition: 0 to +10 MWh (excess electricity sold in spot market, no negative deviation); the deviation without decomposition: -5 to +5 MWh (fixed allocation leads to partial period of supply and demand).

[0246] It can be seen that the wind power and photovoltaic predicted output can fully meet the green power decomposition amount, and the excess part can participate in the day-ahead spot market to obtain the revenue of the spot market. The green electricity deviation obtained from the power decomposition curve considering the overall benefit of the fluctuation of spot electricity price only exists positive deviation, and the positive deviation part can obtain revenue by participating in the spot market; while the green electricity deviation caused by the original green power load without decomposition exists more negative deviation, which cannot obtain revenue and may also lead to failure to meet the specified green electricity demand and pay a penalty fee;

[0247] (3) Comparison of green electricity transaction revenue

[0248] Case 1: only considering the uncertainty of spot electricity price for green power decomposition;

[0249] Case 2: considering the uncertainty of electricity price and icing risk for green power decomposition.

[0250] Table 2-revenue comparison table

[0251]

[0252] wherein the negative deviation penalty cost is calculated according to 8MWh physical negative deviation generated by the high icing risk period of the case one and the market high price.

[0253] As shown in Table 2, the optimization result of the case one has unfeasibility in the physical layer because the icing risk of the transmission line is not considered. In the low temperature period at night, the online power arrangement fails to avoid the transmission bottleneck caused by the line capacity reduction, resulting in a large negative deviation and a high penalty cost, and finally causing a significant decrease in the overall actual income.

[0254] In comparison, the optimization method proposed in the embodiment is adopted in the case two, and the online plan is actively reduced in the period with high risk at night relying on the embedded icing risk assessment model. Although the green electricity trading volume and the spot electricity price income are slightly adjusted, the negative deviation penalty is effectively avoided. Finally, the total income of the case two is 309,881.94 yuan, which is increased by 51,008.98 yuan compared with the case one, with an increase of 19.7%. This shows the advantage of considering the economic efficiency and the physical feasibility. The result proves that the scheme proposed in the embodiment can achieve better comprehensive benefits in the complex high-altitude environment, and has economic efficiency and reliability.

[0255] 6. SUMMARY

[0256] The embodiment shows the actual application effect of the method in the green power transaction in the high-altitude area, and verifies the comprehensive ability of the method in dealing with the electricity price fluctuation and the transmission risk. By constructing the optimization model based on the electricity price scenario and introducing the dynamic assessment mechanism of the transmission icing risk, the method can flexibly adjust the online strategy under different weather conditions. In the period with obvious electricity price fluctuation, the model can accurately capture the spot transaction opportunity, and in the low temperature period with bad weather conditions, the online plan can be actively adjusted to avoid the transmission bottleneck risk, and the safety and stability of the power transaction performance are significantly improved.

[0257] Meanwhile, the method is constructed as a mixed integer linear programming model, and the scene compression technology is combined to control the complexity of the solution, and has good engineering deployment adaptability and actual application feasibility. Overall, the method provides an optimization strategy with economic efficiency, safety and implementation for the green power transaction in the high-altitude area, and has a wide engineering promotion prospect.

[0258] Embodiment 4

[0259] The embodiment provides a green electricity transaction volume decomposition system in a high-altitude area, which comprises:

[0260] The representative scenario set construction module is configured to construct a probability distribution model of historical electricity prices, generate an initial electricity price path reflecting the volatility characteristics of spot electricity prices through a hierarchical sampling method, and filter K representative scenarios and corresponding probabilities through a scenario similarity calculation method to form a representative scenario set.

[0261] The dynamic constraint system construction module is configured to construct a dynamic constraint system for real-time perception of power transmission capacity based on multi-dimensional power transmission constraint conditions, including wind and solar power output prediction, icing risk of power transmission lines, spot electricity price fluctuation, total amount of green power, and time period performance constraint.

[0262] The green electricity transaction volume decomposition module is configured to solve the green power allocation scheme for each hour based on the representative scenario set and the dynamic constraint system through a stochastic optimization method to realize decomposition of green electricity transaction volume.

[0263] Embodiment 5

[0264] This embodiment is based on Embodiment 1:

[0265] This embodiment provides a computer device including a memory and a processor, the memory stores a computer program, and the processor implements the high-altitude region green electricity transaction volume decomposition method of Embodiment 1 when executing the computer program. The computer program can be in the form of source code, object code, an executable file, or some intermediate form, etc.

[0266] Embodiment 6

[0267] This embodiment is based on Embodiment 1:

[0268] This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the high-altitude region green electricity transaction volume decomposition method of Embodiment 1. The computer program can be in the form of source code, object code, an executable file, or some intermediate form, etc. The storage medium includes any entity or device capable of carrying computer program code, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0269] The above description is only the preferred embodiments of the present application, and it should be understood that the present application is not limited to the forms disclosed herein, and should not be considered as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concepts described herein by the above teachings or related art or knowledge. Any modification and change made by those skilled in the art without departing from the spirit and scope of the present application shall be within the protection scope of the appended claims of the present application.

[0270] It is apparent that, for the method embodiments described previously, the steps of the methods have been described as being arranged in a particular order. However, it is to be understood that this is merely for the sake of description and the order of the steps of the method embodiments can be altered so long as the methods remain operable. Also, it is to be understood that the method embodiments described above can be implemented by a computer program product, which can be stored on a computer-readable medium.

Claims

1. A method for decomposing green electricity transaction volume in high altitude areas, characterized in that, The method comprises the following steps: A probability distribution model of historical electricity price is constructed, an initial electricity price path reflecting the fluctuation characteristics of spot electricity price is generated by a hierarchical sampling method, and K representative scenarios and corresponding probabilities are screened by a scenario similarity calculation method to form a representative scenario set; Based on the multi-dimensional power transmission constraint conditions, a dynamic constraint system for real-time sensing of power transmission capacity is constructed; the multi-dimensional power transmission constraint conditions include wind and light output prediction, icing risk of power transmission line, fluctuation of spot electricity price, total amount of green power and time period performance constraint; Based on the representative scenario set and the dynamic constraint system, a green power allocation scheme for each hour is solved by a random optimization method to realize decomposition of green power transaction volume; The dynamic constraint system for real-time sensing of power transmission capacity is constructed based on the multi-dimensional power transmission constraint conditions, which comprises: A capacity reduction function is fitted based on historical operation data of a target regional power grid, a mapping relationship between temperature prediction and power transmission capacity is established, and an icing-triggered capacity reduction model is generated: wherein, is a capacity derating function; is a derating magnitude function, either piecewise linear or exponential; is a maximum capacity derating; is a predicted temperature, is a temperature threshold; The icing state judgment is converted into a continuous differentiable constraint by a penalty factor method, an auxiliary variable is introduced to represent the icing state of the time period, and the upper limit of the power transmission capacity is automatically calculated and adjusted in the model according to the weather forecast.

2. The method of claim 1, wherein, The probability distribution model of historical electricity price comprises the following steps: collecting historical electricity price data and synchronous weather data of a target region, dividing historical electricity price samples into a normal working condition electricity price set and an icing working condition electricity price set according to the icing critical condition of the power transmission line, and constructing the probability distribution model of historical electricity price by using an electricity price-weather joint probability modeling method.

3. The method of claim 2, wherein, The initial electricity price path reflecting the fluctuation characteristics of spot electricity price is generated by the hierarchical sampling method, which comprises the following steps: constructing a non-parametric probability density function for the normal working condition electricity price set and the icing working condition electricity price set respectively, and forming a continuous probability density curve by locally diffusing each historical price point as a probability mass center by using a Gaussian kernel function through the way of historical sample kernel weight accumulation.

4. The method of claim 1, wherein, The K representative scenarios and corresponding probabilities are screened by the scenario similarity calculation method to form the representative scenario set, which comprises the following steps: setting a time period importance weight including a risk indicator function and a risk weighting coefficient; calculating the weighted distance between scenarios, and then determining the K representative scenarios through an iterative clustering compression process.

5. The method of claim 1, wherein, The green power allocation scheme for each hour is solved by the random optimization method, which comprises the following steps: A multi-scenario expected revenue maximization is taken as an objective function, the objective function includes a green electricity contract price, a scenario spot electricity price and a negative deviation penalty cost item, and the negative deviation penalty cost item includes a penalty coefficient and a time period actual available power generation; A plurality of constraint types are taken as constraint conditions for optimization, and the plurality of constraint types include energy balance constraint, total contract quantity constraint, decomposition quantity boundary constraint, dynamic power transmission constraint and icing state logical constraint.

6. A high altitude region green electricity transaction volume decomposition system characterized by, The method comprises the following steps: The representative scenario set construction module is configured to construct a probability distribution model of historical electricity price, generate an initial electricity price path reflecting the fluctuation characteristics of spot electricity price by a hierarchical sampling method, and screen K representative scenarios and corresponding probabilities by a scenario similarity calculation method to form a representative scenario set. The dynamic constraint system construction module is configured to construct a dynamic constraint system for real-time perception of power transmission capacity based on multi-dimensional power transmission constraint conditions, wherein the multi-dimensional power transmission constraint conditions include wind and solar power output prediction, power transmission line icing risk, spot electricity price fluctuation, total green power amount, and time period performance constraint; The green electricity transaction amount decomposition module is configured to solve an hourly green power distribution scheme by a stochastic optimization method based on the representative scenario set and the dynamic constraint system, so as to realize decomposition of the green electricity transaction amount. The dynamic constraint system for real-time perception of power transmission capacity is constructed based on multi-dimensional power transmission constraint conditions, and includes the following steps: A capacity reduction function is fitted based on historical operation data of a target regional power grid, a mapping relationship between temperature prediction and power transmission capacity is established, and an icing-triggered capacity reduction model is generated. wherein, is a capacity derating function; is a derating magnitude function, either piecewise linear or exponential; is a maximum capacity derating; is a predicted temperature, is a temperature threshold; An icing state judgment is converted into a continuous differentiable constraint by a penalty factor method, an auxiliary variable is introduced to represent a time period icing state, and an upper limit of power transmission capacity is automatically calculated and adjusted in the model according to a weather forecast. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-6 when the computer program is executed by the processor. The processor executes the computer program to implement the high-altitude region green electricity transaction amount decomposition method of any one of claims 1-5.

8. A computer readable storage medium storing a computer program, characterized in that, The computer program is executed by the processor to implement the high-altitude region green electricity transaction amount decomposition method of any one of claims 1-5.

Images