New energy consumption cost quantitative measurement and calculation method in market environment

Abstract

The invention belongs to the technical field of new energy consumption cost calculation, and relates to a new energy consumption cost quantitative calculation method in a market environment. The method comprises the following steps: S1, extracting and constructing a new energy multi-dimensional characteristic vector; s2, establishing a mapping relation between the multi-dimensional feature vectors and the clearing boundaries of the power market, and calculating the demand up-regulation amount of each clearing boundary; s3, constructing an operation model corresponding to each time scale, and forming an operation constraint set; s4, identifying and classifying additional cost to form a consumption cost list; and S5, cost quantitative measurement and calculation are carried out, an operation simulation model and a long-term planning model are constructed and solved, the total system cost of the reference scene and the actual new energy characteristic scene is calculated and compared, and the difference value is the total consumption cost caused by the new energy characteristics. According to the method, the physical characteristics of new energy can be converted into quantifiable economic signals, so that more economical, efficient and safe novel power system transformation is supported and realized in multiple key fields such as power market construction, power grid planning and operation, energy policy making and the like.

Description

Technical Field

[0001] This invention belongs to the field of new energy consumption cost calculation technology, specifically involving a method for quantitatively calculating the cost of new energy consumption in a market environment. Background Technology

[0002] In traditional electricity market operations, the lack of an effective quantitative correlation mechanism between the cost of renewable energy integration and its physical characteristics leads to market clearing decisions failing to accurately reflect the additional system costs incurred by renewable energy integration. The essence of this problem lies in the failure to establish a dynamic mapping relationship between the economic signal of integration costs and the uncertainty, volatility, and spatiotemporal distribution characteristics of renewable energy. This results in distorted market clearing prices, weakens resource allocation efficiency, and poses a potential threat to the safe and stable operation of the power grid. Furthermore, the absence of a quantitative mechanism for integration costs means that system operation constraints and investment decisions lack precise cost basis, further impacting the overall performance of new power systems in terms of economy, efficiency, and safety.

[0003] For example, in a typical scenario where a high proportion of wind power is integrated into a regional power grid, the actual output of renewable energy power plants fluctuates significantly due to sudden changes in weather conditions. The prediction error exceeds the range of conventional reserve capacity, and minute-level output changes frequently trigger frequency regulation demands. Because the market clearing model does not translate these volatility characteristics into quantifiable economic parameters, the upward adjustments to reserve capacity and frequency regulation mileage requirements are not included in the clearing price formation mechanism. Consequently, the system frequently experiences insufficient or excessive use of reserve resources, leading to increased intraday market rescheduling frequency, deviations in power flow distribution from safety margins on critical lines, and a lack of cost attribution basis for investment decisions regarding flexible power sources. Specifically, in this scenario, the net load peak-to-valley difference widens due to fluctuations in renewable energy output, but the clearing price fails to reflect the resulting unit start-up and shutdown costs and transmission equipment investment pressures, making it difficult for market participants to optimize operating strategies based on true cost signals.

[0004] If the aforementioned problems are not addressed, the electricity market mechanism will be unable to effectively guide the optimal allocation of resources, the system's safety margin will continue to narrow, and the coordination between long-term planning and short-term operation will further deteriorate. Specifically, the implicit accumulation of absorption costs will cause market clearing results to deviate from the economically optimal solution, and biased investment decisions may lead to a mismatch between power generation structure and grid architecture, thereby hindering the transformation of the new power system towards economic efficiency and safety. Furthermore, the disconnect between operational constraints and cost signals will exacerbate system operational risks, making it impossible for the market to provide sufficient guarantees for reliable power supply in environments with a high proportion of renewable energy.

[0005] Therefore, the technical problem this invention aims to solve is whether a quantitative calculation method for the cost of new energy consumption can be constructed, which can transform the physical characteristics of new energy into quantifiable economic signals, thereby supporting the transformation to a more economical, efficient, and safe new power system in several key areas such as power market construction, power grid planning and operation, and energy policy formulation. Summary of the Invention

[0006] In view of this, the present invention provides a method for quantitatively calculating the cost of renewable energy consumption in a market environment. By accurately quantifying the cost of renewable energy consumption, the electricity market price signal can more realistically reflect the impact of the physical characteristics of renewable energy, thereby improving the efficiency of market resource allocation, supporting the safe and stable operation of the power grid, and promoting the efficient consumption of renewable energy.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for quantitatively calculating the cost of new energy consumption under market conditions includes the following steps: Step S1: Based on the historical output data and predicted data of new energy power plants, extract and construct a multi-dimensional characteristic vector of new energy that characterizes its uncertainty, volatility and spatiotemporal distribution characteristics; Step S2: Quantify the impact of clearing boundaries. Establish a quantitative mapping relationship between the multidimensional feature vector and the multiple clearing boundaries of the power market. Based on this relationship, quantify the upward adjustment of demand at each clearing boundary caused by the access of new energy sources. The multiple clearing boundaries include at least: reserve capacity demand, frequency regulation mileage demand, net load peak-valley difference, power flow margin of critical lines, and power flow distribution. Step S3: Operational constraint modeling. For the medium- and long-term market, day-ahead market, intraday market, and real-time market, construct power system operation models corresponding to each time scale. The clearing boundary adjustment obtained in step S2 is used as a new constraint condition or a tightening parameter for existing constraints and embedded into the corresponding operation model, thereby forming a system operation constraint set coupled in the full time dimension. Step S4: Analyze the composition of absorption costs. Based on the upward adjustment of the clearing boundary demand in Step S2 and the tightening of operational constraints in Step S3, identify and classify the additional costs arising therefrom beyond the traditional electricity market clearing price, and construct a complete absorption cost list. The absorption costs include at least off-site costs, which include reserve adjustment costs, frequency regulation costs, unit rescheduling costs, unit start-up and shutdown costs, investment costs for supporting flexible power sources, and investment costs for transmission equipment. Step S5: Cost Quantification and Calculation. Taking a power system scenario with no new energy fluctuations or ideally controllable new energy sources as the benchmark scenario, the total system cost is calculated and compared between the benchmark scenario and the scenario with actual new energy characteristics by constructing and solving the operation simulation model and the long-term planning model that take into account the operation constraint set. The difference is the total absorption cost caused by the new energy characteristics. The operation simulation model is used to quantify the increase in operation cost, and the long-term planning model is used to quantify the increase in investment cost.

[0008] Furthermore, in step S1, the multidimensional characteristic vector includes at least: output level, output fluctuation range, prediction error described by probability distribution, and minute-level and hour-level output change rate; wherein the prediction error and output change rate are used to correlate reserve capacity demand and frequency regulation mileage demand, respectively.

[0009] Furthermore, the construction of the new energy multidimensional characteristic vector specifically includes: The power output level is obtained by calculating its statistical mean and specific quantile value after dynamically correcting the ultra-short-term power output forecast based on the historical actual power output data of the new energy power station SCADA system and combined with numerical weather forecast information to make dynamic corrections to the ultra-short-term power output forecast. The output fluctuation range is obtained by processing historical output data using a sliding time window algorithm, extracting the historical output range, and correcting it by combining physical parameters such as turbulence intensity or cloud movement. The prediction error is calculated by comparing the actual output with the predicted values ​​at different time scales, and then fitted using a probability distribution model. The output change rate is obtained by performing a difference operation on the output time series and extracting the maximum climbing rate at the minute and hour levels.

[0010] Furthermore, in step S2, the quantitative mapping relationship is established by determining the quantitative impact coefficients of changes in various characteristic parameters of new energy on various clearing boundary demands based on power system production simulation or machine learning models trained based on historical market clearing data.

[0011] Furthermore, in step S2, the calculation of the upward adjustment of the clearing boundary demand specifically includes: The upward adjustment of the reserve capacity requirement is determined based on the probability distribution of the prediction error, through confidence interval calculation or conditional value at risk model. The upward adjustment of the frequency modulation mileage requirement is calculated and determined based on the minute-level output change rate and the system frequency response characteristic model. The change in the net load peak-to-valley difference is calculated by superimposing the load curve on the output level and fluctuation range. The changes in the power flow margin and power flow distribution of the critical path are determined by power flow calculation or sensitivity analysis based on the spatiotemporal distribution characteristics of the new energy output.

[0012] Furthermore, in step S3, the modeling of operational constraints across the entire time dimension is reflected as follows: the medium-to-long-term market, the day-ahead market, the intraday market, and the real-time market need to consider source-load balance and line power flow constraints; the medium-to-long-term market also needs to consider unit start-up and shutdown; the day-ahead market needs to consider both unit start-up and shutdown and short-term ramping; the intraday market needs to consider both short-term ramping and frequency security; and the real-time market needs to consider frequency security; and the constraints of the model at each stage are tightened synchronously due to the boundary adjustment in step S2.

[0013] Furthermore, in step S5, the quantification method of the incremental operating cost is as follows: under the same load and network conditions, high-precision time-series production simulation and market clearing calculation are performed on the baseline scenario and the scenario containing actual new energy characteristics, respectively. The difference between the total system operating costs in the two scenarios is the incremental operating cost.

[0014] Furthermore, the method for quantifying the incremental investment cost is as follows: with the goal of meeting the same power supply reliability and safety standards, construct and solve the optimal extended planning model of traditional power supply corresponding to the benchmark scenario and the source-grid-storage coordinated extended planning model corresponding to the scenario with a high proportion of new energy access. The difference between the total net present value of the two planning schemes is the incremental investment cost.

[0015] Furthermore, step S5 also includes a cost decomposition step: the quantified total absorption cost is decomposed according to the contribution of the new energy source characteristics, and / or attribution analysis is performed according to cost type.

[0016] Furthermore, the new energy multidimensional characteristic vector also includes a parameter characterizing the spatial distribution and aggregation degree of new energy power stations. This parameter is specifically used to correct the upward adjustment of the clearing boundary demand related to the power flow margin and distribution of the line.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention focuses on the clearing boundaries of the electricity market, which directly determine market prices and cost allocation, making the cost quantification results naturally applicable to the market environment.

[0018] The calculation results can be directly used for: ancillary service market pricing, providing a cost basis for reasonable pricing of services such as reserve and frequency regulation; transmission and distribution price determination, quantifying the congestion costs of new energy access on specific lines, and providing support for accurate node or zone transmission pricing; and new energy assessment and cost allocation, providing a precise calculation basis for rules such as assessment costs for new energy forecast deviations and imbalance responsibility allocation. This invention clarifies the impact of new energy on the long-term reliability of the system and the required reserve capacity cost.

[0019] This application provides a method for quantitatively calculating the cost of renewable energy consumption in a market environment. By constructing a multi-dimensional characteristic vector of renewable energy, quantifying the impact of the clearing boundary, modeling the full-time dimension of operational constraints, sorting out the cost composition, and performing cost quantification calculation, it realizes the direct correlation between the physical characteristics of renewable energy and economic signals, thereby accurately quantifying the consumption cost. It has the advantages of making the electricity market price signal more realistically reflect the impact of the physical characteristics of renewable energy by accurately quantifying the cost of renewable energy consumption, thereby improving the efficiency of market resource allocation, supporting the safe and stable operation of the power grid, and promoting the efficient consumption of renewable energy. Attached Figure Description

[0020] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0021] To further illustrate the technical means and effects of the present invention in order to achieve the intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0022] It should be noted that, in this application, unless otherwise stated, directional terms such as "upper," "lower," "top," and "bottom" are generally used in relation to the direction shown in the accompanying drawings, or in relation to the vertical, perpendicular, or gravitational direction of the component itself; similarly, for ease of understanding and description, "inner" and "outer" refer to the inner and outer contours of each component itself, but the above directional terms are not intended to limit the present invention. Example

[0023] like Figure 1 As shown, this application proposes a method for quantitatively calculating the cost of new energy consumption in a market environment, including the following steps: Step S1: Based on the historical output data and predicted data of new energy power plants, extract and construct a multi-dimensional characteristic vector of new energy that characterizes its uncertainty, volatility and spatiotemporal distribution characteristics; Step S2: Quantify the impact of clearing boundaries. Establish a quantitative mapping relationship between the multidimensional feature vector and the multiple clearing boundaries of the power market. Based on this relationship, quantify the upward adjustment of demand at each clearing boundary caused by the access of new energy sources. The multiple clearing boundaries include at least: reserve capacity demand, frequency regulation mileage demand, net load peak-valley difference, power flow margin of critical lines, and power flow distribution. Step S3: Operational constraint modeling. For the medium- and long-term market, day-ahead market, intraday market, and real-time market, construct power system operation models corresponding to each time scale. The clearing boundary adjustment obtained in step S2 is used as a new constraint condition or a tightening parameter for existing constraints and embedded into the corresponding operation model, thereby forming a system operation constraint set coupled in the full time dimension. Step S4: Analyze the composition of absorption costs. Based on the upward adjustment of the clearing boundary demand in Step S2 and the tightening of operational constraints in Step S3, identify and classify the additional costs arising therefrom beyond the traditional electricity market clearing price, and construct a complete absorption cost list. The absorption costs include at least off-site costs, which include reserve adjustment costs, frequency regulation costs, unit rescheduling costs, unit start-up and shutdown costs, investment costs for supporting flexible power sources, and investment costs for transmission equipment. Step S5: Cost Quantification and Calculation. Taking a power system scenario with no new energy fluctuations or ideally controllable new energy sources as the benchmark scenario, the total system cost is calculated and compared between the benchmark scenario and the scenario with actual new energy characteristics by constructing and solving the operation simulation model and the long-term planning model that take into account the operation constraint set. The difference is the total absorption cost caused by the new energy characteristics. The operation simulation model is used to quantify the increase in operation cost, and the long-term planning model is used to quantify the increase in investment cost.

[0024] For ease of understanding, the following explains some key terms in this embodiment: Renewable energy power plants: These are facilities that generate electricity using renewable energy sources such as solar and wind power. Their output is intermittent, fluctuating, and uncertain. Reserve capacity requirement: The extra generating capacity reserved to cope with the uncertainty of power system load or generation output, in order to maintain supply and demand balance. Frequency regulation mileage requirement: The total amount of upward or downward regulation capacity that generating units need to provide to maintain power system frequency stability in response to frequency deviations. Net load peak-to-valley difference: The difference between the peak and valley values ​​of the load curve after subtracting renewable energy output from the total load, reflecting the flexibility required for system dispatch. Critical line power flow margin: The extra electrical energy that important transmission lines in the power system can transmit while meeting safe operation requirements. Power flow distribution: The specific path of electrical energy transmission along transmission lines and the distribution of electrical flow on each line in the power system.

[0025] Medium- to long-term market: This refers to the electricity market with a trading period typically ranging from several months to several years, primarily used to lock in long-term electricity energy and capacity. Day-ahead market: This refers to the electricity market with a trading period of the following day, primarily used to determine the unit combination and generation plan for the next day. Intraday market: This refers to the electricity market with a trading period between the day-ahead market and the real-time market, used for rolling adjustments to day-ahead plans. Real-time market: This refers to the electricity market with a trading period ranging from several minutes to several hours, used to balance real-time supply and demand and respond to unforeseen events and short-term fluctuations.

[0026] Off-site costs: These refer to costs arising from the characteristics of new energy sources, beyond the clearing price in the traditional electricity market, that need to be compensated through other mechanisms or investments. Examples include reserve adjustment costs, frequency regulation costs, unit rescheduling costs, unit start-up and shutdown costs, investment costs for supporting flexible power sources, and investment costs for transmission equipment. Operational simulation models: These are used to simulate the production and scheduling processes of the power system under different operating scenarios to quantify the increment of operating costs. Long-term planning models: These are used to optimize power source and grid investment decisions for the power system over long-term timescales to quantify the increment of investment costs.

[0027] This embodiment provides a method for quantitatively calculating the cost of new energy consumption in a market environment. Through a series of steps, it transforms the physical characteristics of new energy into quantifiable economic signals, thereby achieving the assessment of consumption costs.

[0028] In step S1, the multidimensional characteristic vector includes at least: output level, output fluctuation range, prediction error described by probability distribution, and minute-level and hour-level output change rate; wherein the prediction error and output change rate are used to correlate reserve capacity demand and frequency regulation mileage demand, respectively.

[0029] The construction of the new energy multidimensional characteristic vector specifically includes: The power output level is calculated based on historical actual power output data from the SCADA system of the new energy power station, combined with numerical weather prediction information to dynamically correct the ultra-short-term power output forecast, and then calculating its statistical mean and specific quantile values. The power output fluctuation range refers to the variation in new energy power output within a certain time window. This can be obtained by processing historical power output data using a sliding time window algorithm, extracting its historical power output range, and correcting it with physical parameters such as turbulence intensity or cloud movement. The prediction error is calculated by comparing the actual power output with predicted values ​​at different time scales, and then fitted using a probability distribution model. The power output change rate is obtained by performing a difference operation on the power output time series to extract the maximum ramp rate at the minute and hour levels.

[0030] Specifically, this application concretizes the multidimensional characteristic vector of renewable energy into output level, output fluctuation range, prediction error described by probability distribution, and minute-level and hourly output change rate, and clarifies the relationship between prediction error and reserve capacity demand, and between output change rate and frequency regulation mileage demand, thus making the characterization of renewable energy characteristics more refined and comprehensive. In the above-mentioned method for quantitatively calculating the cost of renewable energy consumption in the market environment, the multidimensional characteristic vector constructed in step S1 is the basis for subsequent quantitative analysis of the clearing boundary impact and operational constraint modeling. By introducing these specific characteristic parameters, this method can more accurately capture the intrinsic attributes of renewable energy. For example, the prediction error described by probability distribution can more comprehensively reflect the uncertainty of renewable energy output prediction, which is crucial for accurately assessing the reserve capacity required by the system, because the allocation of reserve capacity is directly related to the risk of prediction error. At the same time, the minute-level and hourly output change rate directly quantifies the degree of rapid change in renewable energy output, which plays a decisive role in assessing the frequency regulation mileage demand required by the system, because rapid output changes require the system to provide corresponding frequency regulation capabilities to maintain grid stability. This refined characteristic description and clear correlation mechanism make the calculation of the increase in reserve capacity requirements and frequency regulation mileage requirements in subsequent steps more accurate and physically reasonable, thus providing a more solid and reliable data foundation for the entire method of quantitative calculation of absorption costs.

[0031] In step S2, the quantitative mapping relationship is established by determining the quantitative impact coefficients of changes in various characteristic parameters of new energy on various clearing boundary demands based on power system production simulation or machine learning models trained based on historical market clearing data.

[0032] The calculation of the upward adjustment of the clearing boundary demand specifically includes: The upward adjustment of the reserve capacity demand is determined by confidence interval calculation or conditional value-at-risk model based on the probability distribution of the prediction error; the upward adjustment of the frequency regulation mileage demand is determined by calculation based on the minute-level output change rate combined with the system frequency response characteristic model; the change in the net load peak-to-valley difference is calculated by superimposing the output level and fluctuation range with the load curve; the changes in the power flow margin and power flow distribution of the critical line are determined by power flow calculation or sensitivity analysis based on the spatiotemporal distribution characteristics of the renewable energy output.

[0033] In step S3, the modeling of operational constraints across the entire time dimension is reflected as follows: the medium-to-long-term market, the day-ahead market, the intraday market, and the real-time market need to consider source-load balance and line power flow constraints; the medium-to-long-term market also needs to consider unit start-up and shutdown; the day-ahead market needs to consider both unit start-up and shutdown and short-term ramping; the intraday market needs to consider both short-term ramping and frequency security; and the real-time market needs to consider frequency security; and the constraints of the model at each stage are tightened synchronously due to the boundary adjustment in step S2.

[0034] In step S5, the quantification method of the incremental operating cost is as follows: under the same load and network conditions, high-precision time-series production simulation and market clearing calculation are performed on the baseline scenario and the scenario containing actual new energy characteristics, respectively. The difference between the total system operating costs in the two scenarios is the incremental operating cost.

[0035] The method for quantifying the incremental investment cost is as follows: with the goal of meeting the same power supply reliability and safety standards, construct and solve the optimal extended planning model of traditional power sources corresponding to the benchmark scenario and the source-grid-storage coordinated extended planning model corresponding to the scenario with a high proportion of new energy access. The difference between the total net present value of the two planning schemes is the incremental investment cost.

[0036] Step S5 also includes a cost decomposition step: decomposing the quantified total absorption cost according to the contribution of the new energy source, and / or performing attribution analysis according to cost type.

[0037] The new energy multidimensional characteristic vector also includes a parameter characterizing the spatial distribution and aggregation degree of new energy power stations. This parameter is specifically used to correct the upward adjustment of the clearing boundary demand related to the power flow margin and distribution of the line. Example

[0038] This embodiment is based on Embodiment 1 and is implemented in a power system in a region with a large number of new energy power plants such as wind farms and photovoltaic power plants. The power market in this region is facing the challenge of absorbing the high proportion of new energy sources. It is necessary to quantify the absorption cost to guide the design of market mechanisms and grid planning. The specific content is as follows.

[0039] First, in step S1, historical power output data for the past year and short-term forecast data for the next 24 hours are collected for the renewable energy power plants in the region. This data undergoes preliminary processing, such as calculating the average hourly power output and maximum power output fluctuation of wind farms, and the rate of change of the daily power output curve of photovoltaic power plants. These statistics are combined into a multi-dimensional characteristic vector to preliminarily characterize the uncertainty and volatility of renewable energy power output in the region. For example, this vector may include the daily average power output of wind farms, the peak power output time of photovoltaic power plants, and the maximum change in wind power output within a specific time period.

[0040] Subsequently, in step S2, to quantify the impact of new energy sources on the electricity market clearing boundary, a quantitative mapping relationship is established between this multidimensional characteristic vector and multiple clearing boundaries. For example, by analyzing historical data, it is found that when wind power output fluctuations exceed a certain threshold, the system reserve capacity demand will increase accordingly. A simple linear relationship can be established to link the wind power output fluctuation amplitude with the increase in reserve capacity demand. Similarly, an empirical mapping can be established between the photovoltaic output change rate and the frequency regulation mileage demand. Through these mapping relationships, the increase in demand for reserve capacity, frequency regulation mileage, net load peak-valley difference, critical line power flow margin, and power flow distribution in the regional power system due to the access of new energy sources is calculated. For example, the calculated increase in reserve capacity demand is X MW, and the increase in frequency regulation mileage demand is Y MWh.

[0041] Next, in step S3, corresponding power system operation models are constructed for the medium- and long-term, day-ahead, intraday, and real-time electricity markets in the region. These models can be based on traditional unit combination and economic dispatch algorithms. The clearing boundary adjustment amount calculated in step S2 is embedded into these operation models as a new constraint or a tightening parameter for existing constraints. For example, in the day-ahead market operation model, the reserve capacity demand adjustment amount is used as a hard lower limit constraint on the total system reserve capacity; in the real-time market operation model, the frequency regulation mileage demand adjustment amount is transformed into a more stringent limit on the generator unit regulation rate. Thus, a system operation constraint set coupled across the entire time dimension is formed, which can reflect the impact of new energy characteristics on system operation.

[0042] In step S4, the composition of absorption costs is analyzed. Based on the upward adjustment of the clearing boundary requirements in step S2 and the tightening of operational constraints in step S3, the resulting additional costs are identified and categorized. For example, due to the increased reserve capacity requirements, the system may need to start additional gas turbine units to provide backup, thereby incurring fuel costs and start-up and shutdown costs. Due to the increased frequency regulation mileage requirements, existing generator units may need to be adjusted more frequently, leading to increased wear and tear and higher maintenance costs. These costs are identified as off-site costs and included in the complete absorption cost list, such as reserve adjustment costs, frequency regulation costs, and unit rescheduling costs.

[0043] Finally, in step S5, cost quantification is performed. First, a baseline scenario is set, for example, assuming that the renewable energy output in the region is completely predictable and unfluctuating. Then, a scenario incorporating actual renewable energy characteristics is constructed, considering the multi-dimensional characteristic vector extracted in step S1 and the operational constraint set formed in step S3. By constructing and solving the operational simulation model and the long-term planning model, the total system operating cost and total investment cost under these two scenarios are calculated and compared. For example, the operational simulation model can simulate hourly power dispatch within a year, calculating operating costs such as fuel consumption and start-up / shutdown costs; the long-term planning model can optimize investment in power sources and the power grid over the next ten years, calculating the investment costs of new generating units and transmission lines. The difference in the total system cost between the two scenarios is the total absorption cost caused by the renewable energy characteristics. For example, by comparison, it is found that the total system cost in the scenario containing actual renewable energy characteristics is Z billion yuan higher than that in the baseline scenario; this Z billion yuan is the renewable energy absorption cost.

[0044] In the above implementation, when constructing the multi-dimensional characteristic vector of new energy, the output level can be specifically set as the P50 and P90 quantiles of the average daily output of a certain new energy power station over the past year, to reflect its typical and high-output scenarios. The output fluctuation range can be characterized by the difference between the maximum and minimum output within a 15-minute sliding window, and corrected by combining the wind speed turbulence intensity or cloud cover change rate during that period. The prediction error can be calculated by comparing the day-ahead predicted values ​​and actual output data of the power station over the past six months, and fitting the result using a Gaussian mixture model to obtain its probability distribution function. The minute-level output change rate can be calculated as the maximum absolute value of the difference between the output data point per minute and the data point of the previous minute, while the hourly output change rate can be calculated as the maximum absolute value of the difference between the output data point per hour and the data point of the previous hour. By analyzing the probability distribution of the prediction error, its 95% confidence interval can be used to estimate the required reserve capacity; and by analyzing the statistical characteristics of the minute-level output change rate, the frequency regulation mileage requirement that the system needs to provide can be determined.

[0045] The construction of its new energy multidimensional characteristic vector can be implemented as follows: The power output level can be constructed based on the wind farm's actual SCADA system output data for the past three years (10 minutes each), combined with 48-hour numerical weather prediction data provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). A deep learning-based prediction model, such as a Long Short-Term Memory (LSTM) network, is used to dynamically correct the power output prediction for the next four hours. The corrected power output data can then be used to calculate the daily average output as the statistical mean, and the daily P50 and P90 output values ​​can be calculated as specific quantile values.

[0046] The power output fluctuation range can be constructed by using a 5-minute sliding time window algorithm to process historical 1-minute power output data. Within each 5-minute window, the difference between the maximum and minimum power output is calculated as the range. Simultaneously, the turbulence intensity is monitored in real time by the wind farm's meteorological towers, and the calculated range is corrected based on the empirical relationship between turbulence intensity and power output fluctuation to reflect the actual intensity of the fluctuation.

[0047] The prediction error is constructed by comparing the actual 15-minute output data of the wind farm with the predicted values ​​from 24 hours and 4 hours ago. The root mean square error (RMSE) between these predicted and actual values ​​is calculated, and these RMSE data are fitted into a nonparametric probability distribution model using kernel density estimation to describe the probabilistic characteristics of the prediction error.

[0048] The power output change rate is constructed by performing a first-order difference operation on a 1-minute power output time series of a wind farm to obtain the power output change per minute. Then, in a continuous 60-minute data series, the largest positive change is identified as the hourly maximum ramp rate, and the largest negative change is identified as the hourly maximum descent rate. Similarly, the minute-level maximum ramp rate can also be extracted.

[0049] In step S2, to establish a quantitative mapping relationship between the multi-dimensional characteristic vector of new energy sources and the multiple clearing boundaries of the electricity market, a machine learning model trained based on historical market clearing data is adopted. First, operational data of the power system over the past few years is collected, including actual output data, forecast data, load data of new energy power plants, and clearing results of the electricity market at different times, such as actual reserve capacity demand, frequency regulation mileage demand, net load peak-valley difference, and critical line power flow margin. Next, based on this historical data, a multi-dimensional characteristic vector of new energy sources is constructed, including features such as output level, output fluctuation range, prediction error, and minute-level and hourly output change rates. Then, a suitable machine learning algorithm is selected, such as a gradient boosting decision tree model, using the constructed multi-dimensional characteristic vector of new energy sources as the input features of the model and the various clearing boundary requirements as the output targets. The machine learning model is trained and optimized using a large amount of historical data, enabling it to learn and capture the complex nonlinear relationship between new energy characteristics and clearing boundary requirements. After model training is complete, interpretable methods such as feature importance analysis or SHAP values ​​can be used to quantify the impact of each parameter in the multidimensional characteristic vector of new energy sources on the clearing boundary requirements, thereby obtaining specific quantitative impact coefficients. For example, it can be found that the impact coefficient of an increase in the prediction error of new energy output on reserve capacity demand is a positive value, and the impact coefficient of an increase in the minute-level output change rate on frequency regulation mileage demand is also a positive value. These determined quantitative impact coefficients will be directly used to calculate the upward adjustment of each clearing boundary demand caused by the access of new energy sources.

[0050] When calculating the increase in reserve capacity demand, we can first collect the daily output prediction error data of a certain renewable energy power station over the past year and use this data to fit a nonparametric probability density function, such as using a Gaussian kernel function for kernel density estimation. Then, based on this probability density function, we can determine the required increase in reserve capacity in most cases by calculating a 95% confidence interval. Simultaneously, to address extreme risks, we can further employ a conditional value-at-risk model, setting a 99% confidence level, to calculate the average error value when the prediction error is in the worst-case scenario of 1%, and use this as the additional reserve capacity demand. When calculating the increase in frequency regulation mileage demand, we can perform differential processing on the minute-level actual output data of the renewable energy power station to obtain a time series of its minute-level output change rate. Subsequently, these change rates are input into a pre-established system frequency response characteristic model. This model can be a simplified model based on transfer functions or state-space equations, used to simulate the deviation and recovery process of the system frequency under different output change rates, thereby calculating the required frequency regulation mileage. To assess the change in net load peak-to-valley difference, the daily output level and intraday fluctuation range of the renewable energy power station can be obtained and overlaid hourly with the typical load curve of the power grid region. By comparing the net load curves before and after overlay, the specific changes in net load peak, valley, and peak-to-valley difference caused by renewable energy access can be quantified. Finally, when determining the changes in power flow margin and power flow distribution of critical lines, power system power flow calculation software, such as PSSE, can be used to establish a power grid model including the renewable energy power station. By simulating the power flow distribution under different renewable energy output scenarios and comparing it with a baseline scenario without renewable energy access, the power flow changes of critical lines can be identified, and their impact on line margin can be assessed. In addition, sensitivity analysis can be performed to calculate the sensitivity coefficient of renewable energy power station output changes to the power flow of specific critical lines, thereby quickly assessing their impact.

[0051] In constructing full-time operational constraints in this application, the following approaches can be adopted: For the medium- to long-term market, the operational model can be a multi-stage unit combination and economic dispatch model. Here, the source-load balance constraint is represented by the total power generation equaling the total load demand in each time period, while the line power flow constraint ensures that the power of each line does not exceed limits through a DC or AC power flow model. Unit start-up and shutdown constraints are modeled by introducing binary variables and parameters such as minimum start-up and shutdown times and minimum operating / outage times. For example, the minimum operating time for a coal-fired unit is set to 4 hours, and the minimum outage time is set to 6 hours. In the day-ahead market, the operational model can be a mixed-integer linear programming model. In addition to the aforementioned source-load balance and line power flow constraints, unit start-up and shutdown constraints are also considered, but with a higher time resolution, such as making decisions on an hourly basis. Short-term ramp constraints are achieved by limiting the maximum change in unit output within adjacent hours. For example, the maximum ramp rate of a gas turbine within adjacent hours is set to 10% of its rated capacity per hour. For the intraday market, the operating model can be a rolling optimization scheduling model with further improved time resolution, such as in 15-minute increments. Besides short-term ramp-up constraints, frequency safety constraints can be reflected by setting a minimum requirement for the system's total spinning reserve capacity, for example, requiring the system to always maintain at least 500MW of spinning reserve to cope with unforeseen events. In the real-time market, the operating model can be a real-time scheduling system based on model predictive control, with a time resolution reaching the minute or even second level. Frequency safety constraints are maintained by real-time monitoring of system frequency deviations and triggering rapid response frequency regulation resources to adjust output. In all the above market models, the clearing boundary adjustment amount quantified in step S2, for example, the additional 200MW reserve capacity requirement due to new energy forecast errors, will be directly added to the original minimum reserve capacity requirement in each market model, thereby tightening reserve constraints. Similarly, the increased frequency regulation mileage requirement due to fluctuations in new energy output will lead to higher requirements for frequency regulation resource configuration in both intraday and real-time market models, for example, requiring frequency regulation units to provide higher ramp rates or larger frequency regulation ranges.

[0052] When quantifying incremental operating costs, a typical load curve of a regional power grid over the next year can be used, along with the grid's topology, line parameters, and substation capacity, as identical load and network conditions. In the baseline scenario, the output of renewable energy power plants can be set to their predicted average output, or treated as fully dispatchable conventional power sources, whose output can be adjusted according to system demand without incurring prediction errors. In scenarios involving actual renewable energy characteristics, the output data of renewable energy power plants uses an output sequence with uncertainty and volatility generated based on historical data and prediction models, considering its prediction errors, output change rates, and other characteristics. High-precision time-series production simulation and market clearing calculations can employ a unit combination model based on mixed-integer linear programming, considering operational constraints such as generator start-up and shutdown costs, minimum start-up and shutdown times, ramp rates, and fuel costs, as well as the marginal cost clearing mechanism of the electricity market. The simulation period can be set to one year, with rolling optimization using hourly time steps. Ultimately, the total operating cost of the system will include the fuel costs, start-up and shutdown costs, ancillary service procurement costs, and losses due to wind and solar curtailment caused by fluctuations in renewable energy sources.

[0053] In this embodiment, when quantifying the incremental investment cost, the planning period can be set to 20 years, and the net present value (NPV) can be calculated using an annual discount rate of 8%. When constructing the optimal extended planning model for traditional power sources in the baseline scenario, a multi-stage mixed-integer linear programming model can be used, where decision variables include the construction capacity, decommissioning capacity, and operation plan of different types of traditional generator units. The objective function of the model is to minimize the NPV of the total investment cost, fuel cost, operation and maintenance cost, and off-load penalty cost within the planning period. The reliability constraint can be set to an annual off-load probability of no more than 0.1 days / year, considering the N-1 safety criterion. When constructing a source-grid-storage coordinated extended planning model with a high proportion of renewable energy access scenarios, in addition to the decision variables for traditional power sources, the construction capacity and operation plan of wind farms, photovoltaic power plants, battery energy storage systems, and transmission lines also need to be introduced as decision variables. This model can use a stochastic programming method, considering multiple renewable energy output scenarios with different probabilities, and performing weighted optimization on these scenarios. The synergy between energy sources, grid, and energy storage is reflected in the model's coordination between energy storage charging and discharging and renewable energy output and load demand, as well as the role of transmission lines in enhancing the renewable energy absorption capacity. Similarly, the model's objective function is to minimize the net present value (NPV) of the total investment cost, operating cost, and off-load penalty cost of all power sources, grid, and energy storage facilities within the planning period, while satisfying the same LOLP and N-1 safety criteria. Finally, subtracting the NPV of the total investment obtained from the two planning models yields the investment cost increment caused by the characteristics of renewable energy.

[0054] The multi-dimensional characteristic vector of new energy also includes parameters characterizing the spatial distribution and clustering degree of new energy power stations. These parameters are specifically used to correct the upward adjustment of clearing boundary requirements related to line power flow margin and distribution. In this embodiment, this is achieved as follows: First, the geographical coordinate information of all new energy power stations in the area to be evaluated is collected. Based on this data, a spatial clustering index can be calculated. By calculating the centroid of all stations and using the centroid as a reference point, the sum of squared distances from each station to the centroid is calculated. This value can reflect the overall dispersion degree of the stations. In step S2, when it is necessary to calculate the change in the power flow margin and distribution of critical lines, the spatial distribution and clustering degree parameters constructed above can be used. When performing power system production simulation, instead of simply aggregating all new energy outputs into an equivalent power source, the outputs of each new energy power station are injected into their corresponding grid nodes according to their geographical location and grid topology. The spatial clustering index can be used to weight the renewable energy output injection points in the power flow calculation model, or to adjust the sensitivity matrix of the power flow calculation, to more precisely reflect the impact of the spatial distribution of renewable energy output on the power flow of critical lines. For example, if the spatial clustering index shows that multiple large renewable energy power plants are concentrated near the starting point of a critical transmission line, then when calculating the upward adjustment of the power flow margin demand for that line, this parameter will prompt the model to more rigorously consider its cumulative effect, thus calculating a larger margin demand. Conversely, if the power plants are more dispersed, a smaller margin demand may be calculated.

[0055] By introducing parameters characterizing the spatial distribution and clustering of renewable energy power plants into the multidimensional characteristic vector of renewable energy sources, and specifically using these parameters to correct for upward adjustments in clearing boundary demand related to line power flow margin and distribution, this method can more comprehensively and precisely capture the actual impact of the geographical location and concentration of renewable energy power plants on power flow. This allows for full consideration of the clustering effect of renewable energy power plants and the carrying capacity of local power grids when quantifying clearing boundary demand, avoiding assessment biases caused by insufficient spatial information. Therefore, this method can more accurately identify the challenges posed by renewable energy integration to the safe and stable operation of the power grid, and accordingly form a more reasonable and realistic set of operational constraints. Ultimately, this helps improve the accuracy and reliability of quantitative calculations of renewable energy consumption costs, providing stronger technical support for power system planning, market design, and operational decisions, thereby promoting the healthy and orderly development of renewable energy.

[0056] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for quantifying and calculating the cost of new energy consumption in a market environment, characterized in that, The method comprises the following steps: Step S1, based on the historical output data and predicted data of the new energy station, extracting and constructing a new energy multi-dimensional characteristic vector representing the uncertainty, volatility and spatial and temporal distribution characteristics thereof; Step S2, clearing the boundary influence quantity, establishing a quantitative mapping relationship between the multi-dimensional characteristic vector and the multiple clearing boundaries of the electricity market, and quantitatively calculating the upward amount of each clearing boundary demand caused by the access of new energy based on the relationship; The multiple clearing boundaries at least include: reserve capacity demand, frequency modulation mileage demand, net load peak-valley difference, key line flow margin and flow distribution; Step S3, modeling the operation constraints, constructing the corresponding power system operation model for each time scale for the medium and long-term market, day-ahead market, intraday market and real-time market; the upward amount of the clearing boundary obtained in step S2 is embedded into the corresponding operation model as a new constraint condition or a tightening parameter of the existing constraint, thereby forming a system operation constraint set coupled in full time dimension; Step S4, sorting out the consumption cost composition, based on the upward of the clearing boundary demand in step S2 and the tightening of the operation constraint in step S3, identifying and classifying the additional costs caused thereby, which are outside the clearing price of the traditional electric energy market, to form a complete consumption cost list; the consumption cost at least includes off-site cost, and the off-site cost includes reserve adjustment cost, frequency modulation cost, unit rescheduling cost, unit start-stop cost, supporting flexible power supply investment cost and power transmission equipment investment cost; Step S5, cost quantification and calculation, taking the power system scene without new energy fluctuation or new energy as an ideal controllable benchmark scene, constructing and solving the operation simulation model and long-term planning model considering the operation constraint set, respectively calculating and comparing the system total cost under the benchmark scene and the scene containing actual new energy characteristics, and the difference is the total consumption cost caused by the characteristics of new energy; wherein, the operation simulation model is used to quantify the incremental operation cost, and the long-term planning model is used to quantify the incremental investment cost.

2. The method for quantifying and calculating new energy consumption cost in market environment according to claim 1, characterized in that, In step S1, the multi-dimensional characteristic vector at least includes: output level, output fluctuation range, predicted error described by probability distribution, and minute and hour level output change rate; wherein, the predicted error and the output change rate are respectively used to associate the reserve capacity demand and the frequency modulation mileage demand.

3. The method for quantifying and calculating new energy consumption cost in market environment according to claim 2, characterized in that, The construction of the new energy multi-dimensional characteristic vector specifically comprises: The output level is calculated based on the historical actual output data of the new energy station SCADA system, combined with the numerical weather forecast information to dynamically correct the ultra-short-term output prediction, and the statistical mean value and specific quantile value are obtained; The output fluctuation range is obtained by processing the historical output data by using a sliding time window algorithm, extracting the historical output range, and combining the physical parameters of turbulence intensity or cloud movement for correction; The predicted error is obtained by comparing the actual output with the predicted value of different time scales, calculating the statistical error, and fitting the probability distribution model; The output change rate is obtained by performing difference operation on the output time series to extract the maximum climbing rate of minute and hour levels.

4. The method for quantifying and calculating new energy consumption cost in market environment according to claim 1, characterized in that, In step S2, the establishment of the quantitative mapping relationship is based on power system production simulation or a machine learning model trained based on historical market clearing data to determine the quantitative influence coefficient of each characteristic parameter change of the new energy on each clearing boundary demand.

5. The method for quantifying and calculating the cost of new energy consumption in the market environment according to claim 4, characterized in that, In step S2, the calculation of the upward adjustment amount of the clearing boundary demand specifically includes: The upward adjustment amount of the reserve capacity demand is determined by confidence interval calculation or conditional value at risk model based on the probability distribution of the prediction error; The upward adjustment amount of the frequency modulation mileage demand is calculated and determined based on the minute-level output change rate and the system frequency response characteristic model; The change amount of the net load peak-valley difference is calculated based on the output level and fluctuation range and superimposed load curve; The change amount of the key line power flow margin and power flow distribution is determined based on the spatiotemporal distribution characteristics of the new energy output through power flow calculation or sensitivity analysis.

6. The method for quantifying and calculating the cost of new energy consumption in the market environment according to claim 1, characterized in that, In step S3, the modeling of the full-time dimension operation constraint is embodied in that the source-load balance and line power flow constraint need to be considered in the medium and long-term market, day-ahead market, intraday market and real-time market; the unit start-stop needs to be considered in the medium and long-term market; the unit start-stop and short-time climbing need to be considered in the day-ahead market; the short-time climbing and frequency safety need to be considered in the intraday market; the frequency safety needs to be considered in the real-time market; and the constraint conditions of each stage model are simultaneously tightened due to the boundary upward adjustment in step S2.

7. The method for quantifying and calculating the cost of new energy consumption in the market environment according to claim 1, characterized in that, In step S5, the quantitative method of the operation cost increment is to perform high-precision time-series production simulation and market clearing calculation on the benchmark scenario and the scenario containing actual new energy characteristics under the same load and network conditions, and the difference between the total system operation costs in the two scenarios is the operation cost increment.

8. The method for quantifying and calculating new energy consumption cost in market environment according to claim 7, characterized in that, The quantitative method of the investment cost increment is to construct and solve the traditional power source optimal expansion planning model corresponding to the benchmark scenario and the source-grid-storage coordinated expansion planning model corresponding to the scenario containing high proportion of new energy access, respectively, with the goal of meeting the same power supply reliability and safety standards, and the difference between the total investment net present values of the two planning schemes is the investment cost increment.

9. The method for quantifying and calculating new energy consumption cost in market environment according to claim 8, characterized in that, In step S5, the cost decomposition step is also included: the total consumption cost quantified is decomposed according to the contribution degree of the new energy characteristics dimension or attributed according to the cost type.

10. The method for quantifying and calculating the cost of new energy consumption in the market environment according to claim 9, characterized in that, The new energy multi-dimensional characteristic vector also includes a parameter representing the spatial distribution and aggregation degree of the new energy station, which is specially used to correct the upward adjustment amount of the clearing boundary demand related to the line power flow margin and distribution.

Images