Power consumption cost calculation method based on energy storage configuration point

Abstract

The invention relates to the technical field of control and adjustment, in particular to an energy storage configuration point-based power consumption cost calculation method, which comprises the following steps of: acquiring power supply data, renewable energy output fluctuation data and computing power load fluctuation data, and generating a coupling influence feature vector; splicing the electricity price features and the coupling features to form configuration optimization input features; the input features are processed through a dual-objective optimization model, and robustness configuration parameters and economical efficiency operation parameters are output; calculating the reserve cost compensation amount of the extreme scene based on the robustness parameter, and calculating the power purchase cost and maintenance cost of the power grid based on the economical efficiency parameter; and aggregating the cost component and outputting a power consumption cost calculation result. According to the method, the interactive influence of wind-light fluctuation and load mutation is quantified through coupling features, the target conflict between robustness and economy is coordinated by adopting the Pareto algorithm, the risk compensation amount is generated in combination with a pressure test, a feature updating mechanism is fed back in a closed loop mode, and the problem that the operation cost deviates from expectation due to quantization deviation of an existing model is solved.

Description

Technical Field

[0001] This invention relates to the field of control and regulation technology, and in particular to a method for calculating electricity costs based on energy storage configuration points. Background Technology

[0002] In hybrid power supply modes, the robust configuration and economic optimization of energy storage systems are optimized by calculating electricity costs based on dynamic reasoning involving power source diversity, load fluctuations, and energy storage interactions. By integrating the randomness of renewable energy output, the time-of-day differences in grid electricity prices, and energy storage charging and discharging strategies, the system optimizes both robustness and economic objectives. Energy storage devices absorb energy during off-peak hours and release energy during peak hours, reducing electricity procurement costs. At the same time, robust configuration enhances the system's reliability in responding to faults or sudden changes in demand, avoiding additional cost losses. Cost reasoning quantifies fixed investments such as equipment depreciation and maintenance expenses. Combined with variable electricity pricing and energy conversion loss compensation, historical data and predictive models are used to analyze the cost-benefit balance point under multi-source collaboration, deriving the economically optimal solution to minimize overall expenditure.

[0003] In the scenario of optimizing energy storage configuration in intelligent computing centers, electricity cost predictions deviate when there are fluctuations in renewable energy output and sudden load changes. Renewable energy output exhibits strong randomness and unpredictability, such as wind and solar resources. Coupled with the sudden surges or drops in computing load tasks, existing cost models cannot cover the interactive effects of all risky operating conditions. At the same time, robust optimization requires verification in extreme scenarios to ensure system reliability, which conflicts with economic objectives such as minimizing the life-cycle cost of energy storage and grid purchase costs. The model struggles to accurately allocate the weights of green electricity consumption penalty costs and operational risks with limited data. For example, in a scenario of continuous low wind speeds in winter, if the model underestimates the probability of low renewable energy output, energy storage configuration may overemphasize initial investment savings and ignore the emergency power purchase demand during periods of high grid electricity prices. In actual operation, insufficient energy storage capacity triggers computing task delays, resulting in high default compensation costs. The pre-calculated cost model does not fully capture the coupling effect of such long-term fluctuations and sudden load changes, leading to actual expenditures far exceeding the expected optimization plan. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for calculating the electricity cost of energy storage configuration points. This method solves the problem that the coupling effect of wind and solar resource fluctuations and sudden changes in computing load cannot be accurately quantified by existing cost models, which exacerbates the conflict between economic efficiency and robustness objectives in energy storage configuration schemes and causes actual operating costs to deviate significantly from the optimization expectations.

[0005] To solve the above-mentioned technical problems, the specific details of the present invention are as follows: The present invention provides a method for calculating electricity costs based on energy storage deployment points, comprising: Step 1: Obtain power supply data, renewable energy output fluctuation data, and computing load fluctuation data for the target area; identify the time-of-use electricity price segmentation characteristics from the power supply data; and extract the time-of-use electricity price status feature vector. Step 2: Using the renewable energy output fluctuation data and computing load fluctuation data as input, perform wind-solar-load coupling correlation modeling, generate wind-solar-load coupling influence feature vector, and concatenate the grid time-of-use price status feature vector with the wind-solar-load coupling influence feature vector to generate configuration optimization input features; Step 3: Input the configuration optimization input features into the preset dual-objective optimization model, output the robust configuration parameters and economic operation parameters of the energy storage device, calculate the backup cost compensation amount under extreme scenarios based on the robust configuration parameters, and calculate the grid purchase cost and equipment maintenance cost under the charging and discharging strategy based on the economic operation parameters. Step 4: Combine the reserve cost compensation amount, the power grid purchase cost, and the equipment maintenance cost to output the electricity cost calculation result.

[0006] Furthermore, in the method for calculating electricity costs based on energy storage deployment points described in this invention, step 1 includes: Power supply data is read from the power grid metering system, and historical electricity price data and real-time load data in the read power supply data are used as a subset of power supply data; Raw data on renewable energy output is obtained from the meteorological monitoring system. Wind speed and solar intensity sequences are extracted from the raw data to form renewable energy output fluctuation data. Real-time load records are collected from the computing power task scheduling system, and the instantaneous load rate and task queue length are parsed from the real-time load records to form computing power load fluctuation data.

[0007] Furthermore, the method for calculating electricity costs based on energy storage deployment points described in this invention uses renewable energy output fluctuation data and computing load fluctuation data as inputs to perform wind-solar-load coupling correlation modeling, generating a feature vector of wind-solar-load coupling influence, including: The wind speed sequence, the light intensity sequence, and the instantaneous load rate are segmented using a time sliding window to obtain a segmented time series dataset; Using the segmented time series dataset as input, calculate the correlation weights between the standard deviation of wind speed, the coefficient of variation of light intensity, and the peak-to-valley difference of load rate within each time window; The correlation weights are converted into coupling risk coefficients, and the coupling risk coefficients are stored in the feature vector of the coupling influence between wind and solar power and load.

[0008] Furthermore, in the energy storage configuration point-based electricity cost calculation method of the present invention, the step of inputting the configuration optimization input features into a preset bi-objective optimization model and outputting the robust configuration parameters and economic operation parameters of the energy storage device includes: Based on the preset failure rate threshold and load mutation buffer capacity, the constraint boundary of the robustness configuration parameters is defined as the first constraint condition. Based on the upper limit of the unit charge and discharge cost throughout the entire life cycle of energy storage, the constraint boundary of the economic operation parameters is defined as the second constraint condition. The Pareto front algorithm is used to adjust the conflict between the first constraint and the second constraint, and outputs the robust configuration parameters and the economic operating parameters that satisfy Pareto optimality.

[0009] Furthermore, in the method for calculating electricity costs based on energy storage configuration points described in this invention, the calculation of standby cost compensation under extreme scenarios based on the robust configuration parameters includes: Extract the maximum charge / discharge power parameter from the robustness configuration parameters; Using the maximum charge and discharge power parameters as boundary conditions, a continuous low wind and solar power output scenario simulation environment is constructed, and a computing load stress test is performed in the simulation environment. If the computing power load demand in the stress test results exceeds the energy storage power supply capacity, then the economic compensation data for the default delay event will be measured. The economic compensation data for all default and delay events are summed up, and the reserve cost compensation amount is output.

[0010] Furthermore, in the method for calculating electricity consumption costs based on energy storage deployment points described in this invention, the grid purchase cost includes: Parse the peak period identifier from the power grid time-of-use electricity price status feature vector and extract the corresponding peak period start and end time points; Acquire the discharge record data of the energy storage device during the start and end times of the peak period; Calculate the equivalent released charge based on the discharge record data; The equivalent released electricity volume is multiplied by the peak-hour unit price in the electricity price state feature vector to output the grid purchase cost.

[0011] Furthermore, the method for calculating electricity costs based on energy storage deployment points according to the present invention also includes: Extract historical electricity cost data within a preset period; The historical electricity cost data and the electricity cost calculation results are input into the deviation analyzer, which outputs a cost deviation curve. Calculate the peak-to-valley difference rate of the cost deviation curve; If the peak-valley difference rate exceeds a preset threshold, the feature engineering module is invoked to update the feature vector of the coupling effect between wind and solar power and load.

[0012] Furthermore, in the method for calculating electricity consumption costs based on energy storage deployment points described in this invention, the training of the dual-objective optimization model includes: Multiple sets of configuration optimization input features and corresponding measured energy storage operation cost data for different wind and solar resource scenarios were collected to form a training sample set; Using the training sample set as input, calculate the robustness loss function value and the economic loss function value; The weighted sum of the robustness loss function value and the economic loss function value is set as the overall optimization objective; The weight parameters of the bi-objective optimization model are iteratively adjusted using the backpropagation algorithm until the output of the overall optimization objective converges on the validation set.

[0013] Furthermore, in the method for calculating electricity costs based on energy storage deployment points described in this invention, the calculation of the robustness loss function value includes: Within a preset monitoring period, the cumulative duration of actual fault response times exceeding a threshold is collected; Multiply the cumulative duration by a preset unit time penalty coefficient to output the robustness loss function value; The calculation of the economic loss function value includes: Receive the predicted total cost output by the dual-objective optimization model and the measured total cost from the measured energy storage operation cost data; Calculate the root mean square error between the predicted total cost and the measured total cost, and output the economic loss function value.

[0014] Furthermore, the method for calculating electricity costs based on energy storage deployment points according to the present invention is characterized in that the electricity cost calculation results include: If the calculated electricity cost exceeds the preset economic threshold, the configuration optimization module is invoked to adjust the capacity configuration scheme of the energy storage device. If the electricity cost calculation results reflect an abnormal proportion of computing power load cost, the scheduling task controller dynamically optimizes the priority of computing power tasks; If the rate of change in the power grid purchase cost exceeds the limit in the power cost calculation results, an early warning report on the power grid purchase strategy will be generated and sent to the monitoring terminal.

[0015] Beneficial effects of this invention; This invention accurately quantifies the interaction intensity of resource fluctuations and load mutations by constructing a feature vector of the coupling effects of wind and solar power and load. It uses the Pareto front algorithm to dynamically coordinate the conflict boundary between robustness constraints and economic constraints. Combined with stress testing, it generates the backup cost compensation amount for risk scenarios and feeds it back to the feature update mechanism in a closed loop. This effectively overcomes the shortcomings of existing models in quantifying coupling effects, significantly improves the accuracy of electricity cost calculation, and enables energy storage configuration schemes to reduce the comprehensive cost of the entire life cycle while meeting fault response requirements, thus achieving convergence of the deviation between actual operating expenditures and optimization expectations. Attached Figure Description

[0016] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the drawings without creative effort.

[0017] Figure 1 A flowchart illustrating the method for calculating electricity costs based on energy storage configuration points, provided in an embodiment of the present invention. Detailed Implementation

[0018] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. The present invention provided by various embodiments will be described in detail below with reference to the accompanying drawings. To better understand the purpose of the present invention, the present invention will be described in further detail below.

[0019] Please see Figure 1 The present invention provides a method for calculating electricity costs based on energy storage deployment points, comprising: Step 1: Obtain power supply data, renewable energy output fluctuation data, and computing load fluctuation data for the target area; identify the time-of-use electricity price segmentation characteristics from the power supply data; and extract the time-of-use electricity price status feature vector. Step 2: Using the renewable energy output fluctuation data and computing load fluctuation data as input, perform wind-solar-load coupling correlation modeling, generate wind-solar-load coupling influence feature vector, and concatenate the grid time-of-use price status feature vector with the wind-solar-load coupling influence feature vector to generate configuration optimization input features; Step 3: Input the configuration optimization input features into the preset dual-objective optimization model, output the robust configuration parameters and economic operation parameters of the energy storage device, calculate the backup cost compensation amount under extreme scenarios based on the robust configuration parameters, and calculate the grid purchase cost and equipment maintenance cost under the charging and discharging strategy based on the economic operation parameters. Step 4: Combine the reserve cost compensation amount, the power grid purchase cost, and the equipment maintenance cost to output the electricity cost calculation result.

[0020] This invention first collects real-time power supply data for the target area from the power grid metering system. This dataset includes historical electricity price curves and real-time load monitoring values. Simultaneously, it acquires raw data on renewable energy output fluctuations from the meteorological monitoring system, focusing on extracting wind speed time series and solar intensity change trajectories. At the same time, it captures instantaneous load rate fluctuations and task queue length changes from the computing power task scheduling system. The power supply data undergoes time-period feature analysis, and a price pattern recognition algorithm is used to divide peak, flat, and valley price intervals. The recognition results are quantified into a power grid time-of-use price status feature vector, which includes the price weight coefficient and duration stamp for each time period.

[0021] Wind speed sequence, solar intensity sequence, and computing power load fluctuation data are input into the coupling analysis module. The time series data is segmented into continuous segments using a sliding time window mechanism. Within the window, the standard deviation of wind speed fluctuation, the coefficient of variation of solar intensity, and the peak-valley difference of load rate are calculated respectively. The dynamic correlation of the three is analyzed through the correlation coefficient matrix to generate a feature vector of wind-solar-load coupling influence. After dimensional alignment of this vector with the feature vector of grid time-of-use electricity price status, a tensor splicing operation is performed to form a configuration optimization input feature matrix that integrates energy characteristics and load characteristics.

[0022] The configuration optimization input feature matrix is ​​input into a pre-trained bi-objective optimization model, which includes a robustness branch and an economic branch for parallel processing. The robustness branch defines a safety boundary for the energy storage charging and discharging power as a constraint condition based on a preset system failure rate threshold and load surge buffer capacity requirements. The economic branch sets a unit charging and discharging cost threshold based on the energy storage device's lifecycle cost model, forming an economic constraint. The Pareto front optimization algorithm is used to coordinate the two sets of constraints across multiple objectives, outputting a set of configuration parameters that satisfy both reliability and economic balance. Based on the maximum power parameter in the robust configuration parameters, a simulation environment for a low-wind / solar continuous output scenario is constructed. A computing load stress test is performed in the simulation. When the load demand exceeds the energy storage power supply capacity threshold, the penalty for task delay is measured and accumulated as a reserve cost compensation. Simultaneously, based on the charging and discharging strategy in the economic operation parameters and combined with peak-hour discharge records, the equivalent power supply is calculated, and the grid purchase cost is calculated according to the corresponding time-period electricity price.

[0023] The reserve cost compensation, grid purchase cost, and equipment maintenance cost are input into the cost aggregator. The reserve cost compensation is discounted according to the risk probability, the grid purchase cost is added with the transmission and distribution price surcharge, and the equipment maintenance cost is included in the residual value adjustment coefficient. The weighted fusion algorithm generates the electricity cost calculation result including detailed breakdowns. This result reflects both real-time operating costs and potential risk costs. When the total cost value triggers a threshold, the optimization iteration process is automatically started.

[0024] Specifically, in the method for calculating electricity costs based on energy storage deployment points described in this invention, step 1 includes: Power supply data is read from the power grid metering system, and historical electricity price data and real-time load data in the read power supply data are used as a subset of power supply data; Raw data on renewable energy output is obtained from the meteorological monitoring system. Wind speed and solar intensity sequences are extracted from the raw data to form renewable energy output fluctuation data. Real-time load records are collected from the computing power task scheduling system, and the instantaneous load rate and task queue length are parsed from the real-time load records to form computing power load fluctuation data.

[0025] The power supply data collected by the power grid metering system includes historical electricity price records and real-time load monitoring values. The historical electricity price data is archived by timestamp to form an electricity price sequence, and the real-time load data is updated at a frequency of minutes. The wind speed sequence and light intensity sequence obtained from the meteorological monitoring system are cleaned and outliers are removed to form renewable energy output fluctuation data. The real-time load records provided by the computing power task scheduling system separate the instantaneous load rate and the task queue length through a parser. The load rate is quantified as a percentage to calculate the resource occupancy status, and the task queue length reflects the number of tasks to be processed. Together, they constitute the computing power load fluctuation data.

[0026] Specifically, the method for calculating electricity costs based on energy storage deployment points described in this invention uses renewable energy output fluctuation data and computing load fluctuation data as inputs to perform wind-solar-load coupling correlation modeling, generating a feature vector of wind-solar-load coupling influence, including: The wind speed sequence, the light intensity sequence, and the instantaneous load rate are segmented using a time sliding window to obtain a segmented time series dataset; Using the segmented time series dataset as input, calculate the correlation weights between the standard deviation of wind speed, the coefficient of variation of light intensity, and the peak-to-valley difference of load rate within each time window; The correlation weights are converted into coupling risk coefficients, and the coupling risk coefficients are stored in the feature vector of the coupling influence between wind and solar power and load.

[0027] A time-sliding window is used to segment wind speed, light intensity, and instantaneous load rate into a fixed duration, generating a segmented time-series dataset that includes time-aligned data. For each time window, the standard deviation of wind speed is calculated to characterize the fluctuation range, the coefficient of variation of light intensity reflects the stability of energy output, and the peak-to-valley difference of load rate indicates the degree of change in demand. The dynamic correlation weights of the three are calculated using the Pearson correlation coefficient, and the weight values ​​are linearly transformed and mapped to a coupling risk coefficient in the 0-1 interval. Finally, this coefficient is stored in a specified dimension of the feature vector.

[0028] Specifically, the method for calculating electricity costs based on energy storage configuration points according to the present invention, wherein the step of inputting the configuration optimization input features into a preset bi-objective optimization model and outputting robust configuration parameters and economic operating parameters of the energy storage device includes: Based on the preset failure rate threshold and load mutation buffer capacity, the constraint boundary of the robustness configuration parameters is defined as the first constraint condition. Based on the upper limit of the unit charge and discharge cost throughout the entire life cycle of energy storage, the constraint boundary of the economic operation parameters is defined as the second constraint condition. The Pareto front algorithm is used to adjust the conflict between the first constraint and the second constraint, and outputs the robust configuration parameters and the economic operating parameters that satisfy Pareto optimality.

[0029] The failure rate threshold is derived from the equipment reliability verification report, and the load mutation buffer capacity is set according to the historical maximum load mutation amplitude. Together, they define the minimum power margin and response time constraint of the robustness configuration parameters. The energy storage life cycle cost model is based on equipment depreciation rate, maintenance frequency and cycle life data to derive the upper limit of unit charge and discharge cost, forming the boundary conditions of economic operation parameters. The Pareto front algorithm takes the constraint boundary as input, selects the parameter combination that simultaneously satisfies the robustness constraint and the economic constraint through non-dominated sorting, and outputs the optimal compromise solution in the solution set.

[0030] Specifically, the method for calculating electricity costs based on energy storage configuration points according to the present invention includes the following steps: calculating the backup cost compensation amount under extreme scenarios based on the robust configuration parameters. Extract the maximum charge / discharge power parameter from the robustness configuration parameters; Using the maximum charge and discharge power parameters as boundary conditions, a continuous low wind and solar power output scenario simulation environment is constructed, and a computing load stress test is performed in the simulation environment. If the computing power load demand in the stress test results exceeds the energy storage power supply capacity, then the economic compensation data for the default delay event will be measured. The economic compensation data for all default and delay events are summed up, and the reserve cost compensation amount is output.

[0031] The maximum charge and discharge power parameters are extracted from the robust configuration parameters as the simulation boundary. In the continuous low wind and solar power output scenario, a simulation environment is set where the wind and solar power output is continuously lower than 20% of the installed capacity. The stress test uses 120% of the historical peak computing load as the load benchmark. When the duration of the test load exceeds the energy storage power supply duration threshold, a single default delay event is recorded and the compensation amount is calculated according to the service level agreement. After accumulating the compensation amounts of all events, a risk probability discount factor is introduced to calculate the final standby cost compensation amount.

[0032] Specifically, in the method for calculating electricity consumption costs based on energy storage deployment points described in this invention, the grid electricity purchase cost includes: Parse the peak period identifier from the power grid time-of-use electricity price status feature vector and extract the corresponding peak period start and end time points; Acquire the discharge record data of the energy storage device during the start and end times of the peak period; Calculate the equivalent released charge based on the discharge record data; The equivalent released electricity volume is multiplied by the peak-hour unit price in the electricity price state feature vector to output the grid purchase cost.

[0033] The peak period identifier code in the time-of-use electricity price status feature vector of the power grid is parsed, and the precise start and end times of the corresponding time period are obtained by associating with the electricity price database; the discharge log of the energy management system of the energy storage device during the specified peak period is queried to extract the discharge power and duration data; the equivalent released electricity is calculated based on the integral of the discharge power over time, and the power grid purchase cost is calculated by combining the peak period unit price parameters stored in the electricity price status feature vector.

[0034] Specifically, the method for calculating electricity costs based on energy storage deployment points according to the present invention further includes: Extract historical electricity cost data within a preset period; The historical electricity cost data and the electricity cost calculation results are input into the deviation analyzer, which outputs a cost deviation curve. Calculate the peak-to-valley difference rate of the cost deviation curve; If the peak-valley difference rate exceeds a preset threshold, the feature engineering module is invoked to update the feature vector of the coupling effect between wind and solar power and load.

[0035] Historical electricity cost data for the past 12 months is extracted from the energy management database and input together with the current electricity cost calculation results into a deviation analyzer based on the least squares method. The deviation analyzer outputs a cost deviation curve including the cost deviation magnitude and fluctuation frequency, and calculates the absolute difference ratio between adjacent peaks and troughs of the curve as the difference rate. When the difference rate exceeds the system's preset 5% threshold, a feature update command is sent to the feature engineering module, triggering the remodeling of the feature vector of the coupling effect between wind and solar power and load.

[0036] Specifically, the method for calculating electricity costs based on energy storage deployment points according to the present invention includes the following steps in training the dual-objective optimization model: Multiple sets of configuration optimization input features and corresponding measured energy storage operation cost data for different wind and solar resource scenarios were collected to form a training sample set; Using the training sample set as input, calculate the robustness loss function value and the economic loss function value; The weighted sum of the robustness loss function value and the economic loss function value is set as the overall optimization objective; The weight parameters of the bi-objective optimization model are iteratively adjusted using the backpropagation algorithm until the output of the overall optimization objective converges on the validation set.

[0037] We collect configuration optimization input feature samples under typical wind and solar resource scenarios in different seasons, and simultaneously record the actual operation and maintenance costs of the energy storage system and the cost of purchased electricity to form the measured energy storage operation cost data. The two together form a labeled training sample set. The robustness loss function value is obtained by multiplying the total system fault response timeout duration and the unit time penalty within the statistical validation period, and the economic loss function value is calculated by the root mean square error of the predicted cost and the measured cost. The two types of loss function values ​​are superimposed with a weight ratio of 7:3 to form the overall optimization objective. The weight parameters of the fully connected layer of the model are adjusted by the backpropagation algorithm until the loss value on the validation set fluctuates by less than 1% for three consecutive epochs.

[0038] Specifically, in the method for calculating electricity costs based on energy storage deployment points described in this invention, the calculation of the robustness loss function value includes: Within a preset monitoring period, the cumulative duration of actual fault response times exceeding a threshold is collected; Multiply the cumulative duration by a preset unit time penalty coefficient to output the robustness loss function value; The calculation of the economic loss function value includes: Receive the predicted total cost output by the dual-objective optimization model and the measured total cost from the measured energy storage operation cost data; Calculate the root mean square error between the predicted total cost and the measured total cost, and output the economic loss function value.

[0039] When calculating the robustness loss function, the cumulative number of minutes of timeout in the energy storage system's response to fault commands within the preset 30-day monitoring period is used to generate the loss value in combination with the compensation standard per minute stipulated in the contract. In calculating the economic loss function, the predicted total cost curve output by the bi-objective optimization model is received and compared with the monthly total cost bill in the measured energy storage operation cost data. The model prediction deviation is quantified by the root mean square error formula.

[0040] Specifically, the method for calculating electricity costs based on energy storage deployment points according to the present invention is characterized in that the electricity cost calculation results include: If the calculated electricity cost exceeds the preset economic threshold, the configuration optimization module is invoked to adjust the capacity configuration scheme of the energy storage device. If the electricity cost calculation results reflect an abnormal proportion of computing power load cost, the scheduling task controller dynamically optimizes the priority of computing power tasks; If the rate of change in the power grid purchase cost exceeds the limit in the power cost calculation results, an early warning report on the power grid purchase strategy will be generated and sent to the monitoring terminal.

[0041] When the calculated electricity cost exceeds the preset economic threshold, a capacity adjustment command is sent to the configuration optimization module. The module then recalculates the optimal energy storage capacity configuration based on historical load patterns. When the computing power load cost is detected to account for more than 40% of the total cost, the task controller dynamically reduces the priority of non-real-time tasks. If the fluctuation rate of the grid purchase cost exceeds 15% in adjacent cycles, an early warning report including risk source analysis is generated and pushed to the monitoring terminal display interface through the message queue service.

[0042] This invention addresses the cost quantification deviation problem caused by the coupling effect of wind and solar resource fluctuations and load abrupt changes. First, a feature vector of the coupling influence between wind and solar resources and load is constructed. A time-sliding window is used to segment wind speed sequences, solar intensity sequences, and computing power load fluctuation data. Within the window, the correlation weights of the wind speed standard deviation, solar intensity coefficient of variation, and load rate peak-to-valley difference are dynamically calculated. These correlation weights are converted into feature coefficients that quantify the coupling risk. This feature vector accurately characterizes the intensity of the interaction between the randomness of wind and solar resources and the suddenness of computing power tasks. Second, a dual-objective optimization model is established to handle the coupling features. Robustness constraints are defined based on a preset failure rate threshold and load abrupt change buffer capacity. Simultaneously, economic constraints are defined based on the total life-cycle cost of energy storage. The Pareto front algorithm is used to perform non-dominated sorting of the two sets of conflicting constraints, outputting a set of configuration parameters that simultaneously meet the minimum failure response requirements and the upper limit of unit charge-discharge cost, achieving a dynamic balance between economic and robust objectives. Finally, a feedback correction mechanism is deployed. When the peak-valley difference rate of the deviation curve generated by the historical electricity cost data and the prediction results exceeds the threshold, the feature engineering module is triggered to update the coupling effect feature vector. Through stress testing, the load change conditions in the continuous low wind and solar output scenario are simulated, the default delay compensation cost is measured and fed back to the model training process, forming a data closed-loop optimization link, thereby eliminating the defect of the existing model in insufficient quantification of the coupling effect of long-term fluctuations and sudden loads.

Claims

1. A method for calculating electricity costs based on energy storage deployment points, characterized in that, include: Step 1: Obtain power supply data, renewable energy output fluctuation data, and computing load fluctuation data for the target area; identify the time-of-use electricity price segmentation characteristics from the power supply data; and extract the time-of-use electricity price status feature vector. Step 2: Using the renewable energy output fluctuation data and computing load fluctuation data as input, perform wind-solar-load coupling correlation modeling, generate wind-solar-load coupling influence feature vector, and concatenate the grid time-of-use price status feature vector with the wind-solar-load coupling influence feature vector to generate configuration optimization input features; Step 3: Input the configuration optimization input features into the preset dual-objective optimization model, output the robust configuration parameters and economic operation parameters of the energy storage device, calculate the backup cost compensation amount under extreme scenarios based on the robust configuration parameters, and calculate the grid purchase cost and equipment maintenance cost under the charging and discharging strategy based on the economic operation parameters. Step 4: Combine the reserve cost compensation amount, the power grid purchase cost, and the equipment maintenance cost to output the electricity cost calculation result.

2. The method for calculating electricity costs based on energy storage deployment points according to claim 1, characterized in that, Step 1 includes: Power supply data is read from the power grid metering system, and historical electricity price data and real-time load data in the read power supply data are used as a subset of power supply data; Raw data on renewable energy output is obtained from the meteorological monitoring system. Wind speed and solar intensity sequences are extracted from the raw data to form renewable energy output fluctuation data. Real-time load records are collected from the computing power task scheduling system, and the instantaneous load rate and task queue length are parsed from the real-time load records to form computing power load fluctuation data.

3. The method for calculating electricity costs based on energy storage deployment points according to claim 2, characterized in that, The renewable energy output fluctuation data and computing load fluctuation data are used as inputs to model the correlation between wind and solar power and load coupling, generating a feature vector of the influence of wind and solar power and load coupling, including: The wind speed sequence, the light intensity sequence, and the instantaneous load rate are segmented using a time sliding window to obtain a segmented time series dataset; Using the segmented time series dataset as input, calculate the correlation weights between the standard deviation of wind speed, the coefficient of variation of light intensity, and the peak-to-valley difference of load rate within each time window; The correlation weights are converted into coupling risk coefficients, and the coupling risk coefficients are stored in the feature vector of the coupling influence between wind and solar power and load.

4. The method for calculating electricity costs based on energy storage deployment points according to claim 3, characterized in that, The step of inputting the configuration optimization input features into a preset bi-objective optimization model and outputting robust configuration parameters and economic operating parameters of the energy storage device includes: Based on the preset failure rate threshold and load mutation buffer capacity, the constraint boundary of the robustness configuration parameters is defined as the first constraint condition. Based on the upper limit of the unit charge and discharge cost throughout the entire life cycle of energy storage, the constraint boundary of the economic operation parameters is defined as the second constraint condition. The Pareto front algorithm is used to adjust the conflict between the first constraint and the second constraint, and outputs the robust configuration parameters and the economic operating parameters that satisfy Pareto optimality.

5. The method for calculating electricity costs based on energy storage deployment points according to claim 4, characterized in that, The calculation of backup cost compensation under extreme scenarios based on the robust configuration parameters includes: Extract the maximum charge / discharge power parameter from the robustness configuration parameters; Using the maximum charge and discharge power parameters as boundary conditions, a continuous low wind and solar power output scenario simulation environment is constructed, and a computing load stress test is performed in the simulation environment. If the computing power load demand in the stress test results exceeds the energy storage power supply capacity, then the economic compensation data for the default delay event will be measured. The economic compensation data for all default and delay events are summed up, and the reserve cost compensation amount is output.

6. The method for calculating electricity costs based on energy storage deployment points according to claim 5, characterized in that, The electricity purchase cost of the power grid includes: Parse the peak period identifier from the power grid time-of-use electricity price status feature vector and extract the corresponding peak period start and end time points; Acquire the discharge record data of the energy storage device during the start and end times of the peak period; Calculate the equivalent released charge based on the discharge record data; The equivalent released electricity volume is multiplied by the peak-hour unit price in the electricity price state feature vector to output the grid purchase cost.

7. The method for calculating electricity costs based on energy storage deployment points according to claim 6, characterized in that, Also includes: Extract historical electricity cost data within a preset period; The historical electricity cost data and the electricity cost calculation results are input into the deviation analyzer, which outputs a cost deviation curve. Calculate the peak-to-valley difference rate of the cost deviation curve; If the peak-valley difference rate exceeds a preset threshold, the feature engineering module is invoked to update the feature vector of the coupling effect between wind and solar power and load.

8. The method for calculating electricity costs based on energy storage deployment points according to claim 7, characterized in that, The training of the dual-objective optimization model includes: Multiple sets of configuration optimization input features and corresponding measured energy storage operation cost data for different wind and solar resource scenarios were collected to form a training sample set; Using the training sample set as input, calculate the robustness loss function value and the economic loss function value; The weighted sum of the robustness loss function value and the economic loss function value is set as the overall optimization objective; The weight parameters of the bi-objective optimization model are iteratively adjusted using the backpropagation algorithm until the output of the overall optimization objective converges on the validation set.

9. The method for calculating electricity costs based on energy storage deployment points according to claim 8, characterized in that, The calculation of the robustness loss function value includes: Within a preset monitoring period, the cumulative duration of actual fault response times exceeding a threshold is collected; Multiply the cumulative duration by a preset unit time penalty coefficient to output the robustness loss function value; The calculation of the economic loss function value includes: Receive the predicted total cost output by the dual-objective optimization model and the measured total cost from the measured energy storage operation cost data; Calculate the root mean square error between the predicted total cost and the measured total cost, and output the economic loss function value.

10. The method for calculating electricity costs based on energy storage deployment points according to claim 9, characterized in that, The electricity cost calculation results include: If the calculated electricity cost exceeds the preset economic threshold, the configuration optimization module is invoked to adjust the capacity configuration scheme of the energy storage device. If the electricity cost calculation results reflect an abnormal proportion of computing power load cost, the scheduling task controller dynamically optimizes the priority of computing power tasks; If the rate of change in the power grid purchase cost exceeds the limit in the power cost calculation results, an early warning report on the power grid purchase strategy will be generated and sent to the monitoring terminal.

Images