Virtual power plant economic optimization scheduling method and system considering multiple uncertainties
By constructing wind and light output scenarios and dynamic time-sharing electricity prices in virtual power plants, combining price-based demand response and snow ablation algorithm, multiple uncertainties of virtual power plants are solved, and efficient and economical operation of virtual power plants and smooth power loads are achieved.
Patent Information
- Application Number
- CN202411032269.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-22
AI Technical Summary
The existing technology fails to effectively consider multiple uncertainties such as electricity prices, price-based demand response and scenery in virtual power plants, resulting in insufficient economic and stability in the operation of virtual power plants.
The Latin supercube sampling method is used to construct the wind and light output scenario, combine the improved iterative self-organized data analysis algorithm to reduce the scenario, formulate dynamic time-sharing electricity prices, and mobilize the electric load through price-based demand response, and build the objective function to have the lowest total operating cost within the trading day. The virtual power plant economic optimization scheduling model is used to solve the model using the snow ablation algorithm to improve the economy of the virtual power plant.
The accuracy of wind and light output prediction and virtual power plant scheduling accuracy are improved, the total operating cost of virtual power plants is reduced, and the economic operation capability of virtual power plants is improved by smoothing the electric load curve and optimizing resource real-time output.
Smart Images

Figure CN119010207A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electric power monitoring, and in particular to a method and system for economic optimization scheduling of a virtual power plant taking multiple uncertainties into consideration. Background Art
[0002] With the emergence of distributed energy such as new energy generation and flexible loads, how to aggregate and efficiently control distributed energy on a large scale has become a key research area to solve the supply and demand matching and internal coordinated operation of new power systems. Virtual power plants (VPPs) are based on advanced measurement technology, state perception technology, and low-cost and low-latency communication technology. They can aggregate large-scale distributed energy and achieve fine control, thereby improving the flexible adjustment capability and operational resilience of new power systems.
[0003] Reference [1] Li Jiasen, Wang Jin, Yang Meng, et al. Economic optimization dispatch of combined heat and power of virtual power plant based on stochastic optimization [J]. Acta Energiae Solaris Sinica, 2023, 44(09): 57-65. Establish an optimization dispatch model with the maximum expected day-ahead revenue of virtual power plant as the goal. Reference [2] Zhang Dahai, Yun Yunyun, Wang Xiaojun, et al. Multi-time scale optimization dispatch of new energy virtual power plant considering wind and solar uncertainty [J]. Acta Energiae Solaris Sinica, 2022, 43(11): 529-537. Based on the robust stochastic optimization theory, the uncertainty of wind and solar is solved and a multi-time scale optimization dispatch strategy of VPP is proposed, which can improve the economic efficiency of VPP operation. Reference [3] Chen Houhe, Wang Zixuan, Zhang Rufeng, et al. Distributed optimization dispatch modeling of wind power grid-connected system with virtual power plant [J]. Proceedings of the CSEE, 2019, 39(09): 2615-2625. Introduce DR-VPP in the wind power grid-connected system and construct the system's day-ahead economic dispatch model to achieve optimal wind power consumption. Reference [4] Liu Jicheng, Feng Shuxian, Song Yanan, et al. Optimal dispatch model of virtual power plant considering source-load uncertainty [J / OL]. Modern Electric Power, 1-11 [2024-06-15]. Propose a VPP model that considers the uncertainty of distributed energy and controllable loads to reduce penalty costs and maximize the system benefits of VPP operation. Reference [5] AXK, AJX, ADL, et al. Robust stochastic optimal dispatching method of multi-energy virtual power plant considering multiple uncertainties - ScienceDirect [J]. Applied Energy, 279 [2023-12-22]. Consider an uncertain cardinality set with robust adjustable coefficients to describe the wind and solar output on the VPP source side, and improve the economic efficiency of the two-stage operation of the virtual power plant. The above literatures have all dealt with the uncertainty of source and load in the VPP model, but there are deficiencies in the comprehensive consideration of the correlation of multiple uncertainties such as electricity price, price-based demand response and wind and solar power in VPP. Therefore, it is urgent to build a scheduling method and system that considers multiple uncertainties and can achieve efficient and economical operation of virtual power plants. Summary of the invention
[0004] The technical problem to be solved by the present invention is to provide a method and system for economic optimization scheduling of a virtual power plant taking into account multiple uncertainties, efficiently and economically utilizing the distributed energy aggregated by the virtual power plant to ensure the safe and stable operation of the new power system.
[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: The economic optimization dispatching method of virtual power plant considering multiple uncertainties includes the following steps: Step 1: Based on the wind-solar power output joint distribution function of the virtual power plant, the Latin hypercube sampling method is used to construct the wind-solar power output scenario, and the improved iterative self-organizing data analysis algorithm is used to reduce the scenario; Step 2: Formulate a dynamic time-of-use electricity price for the virtual power plant based on the predicted wind and solar power output values; Step 3: Mobilize the electricity load through price-based demand response under dynamic time-of-use electricity prices; Step 4: Determine the constraints that need to be met when scheduling the virtual power plant, and construct an economic optimization scheduling model for the virtual power plant with the objective function of minimizing the total operating cost within the trading day; Step 5. Use the snow melting algorithm to solve the constructed economic optimization scheduling model of the virtual power plant, and simulate and verify the scheduling method to improve the economic efficiency of the virtual power plant operation.
[0006] The construction process of the wind and solar power output joint distribution function of the virtual power plant in Step 1 above is as follows: Taking the history of virtual power plants Based on the wind and solar power output data of the day, Contribute to the scenery Samples collected during the sampling period; The non-parametric kernel density estimation method using the kernel function as Gaussian function is used to generate the probability density function of wind and solar power output; Cumulative distribution function and It is obtained based on the probability density function of wind and solar power output; The two-dimensional Frank-Copula function constructs the joint distribution function of wind and solar power output in each time period.
[0007] The calculation methods of the above-mentioned wind-solar output probability density function and wind-solar output joint distribution function are as follows: ; ; In the formula, The first Sample values; is bandwidth; express and Relevance, and , A positive value indicates a positive correlation, whereas a negative correlation.
[0008] Compared with the original algorithm, the improved iterative self-organizing data analysis algorithm in the above Step 1 reduces the scenarios by using Mahalanobis distance instead of Euclidean distance as the measurement when selecting cluster centers and clusters that are far apart.
[0009] In the above Step 2, the dynamic time-of-use electricity price of the virtual power plant is formulated based on the predicted wind and solar power output value a day ago. The dynamic time-of-use electricity price calculation formula is: ; ; ; In the formula, for Dynamic time-of-use electricity prices corresponding to the time; is the peak electricity price; To equalize electricity prices; for valley electricity prices; and Parameters for dynamic time-of-use electricity pricing; To predict the maximum total wind and solar power output in a day; To predict the minimum total wind and solar power output in a day; for The sum of all the weather and power at all times; for Wind power output at all times; for Photovoltaic output at all times.
[0010] In the above Step 3, the power load is fully mobilized through price-based demand response under dynamic time-of-use electricity prices. The sensitivity of the power load to electricity prices described by the self-elasticity coefficient and the cross-elasticity coefficient is calculated as follows: ; ; ; In the formula, is the self-elastic coefficient; For time period The load change; For the initial period The load demand; The electricity price in the time period The amount of change; For time period Initial electricity price; is the cross elastic coefficient; For time period Change in electricity prices; For time period Initial electricity price; , and are the load changes in peak, flat and valley periods after price-based demand response; , and They are the load demands during peak, flat and valley periods before price-based demand response; , and It is the change of electricity price during peak, flat and valley periods.
[0011] The objective function calculation formula in the above Step 4 is: ; Where: Proceeds from electricity sales for virtual power plants; The operation and maintenance costs of virtual power plants; is the energy consumption cost of the virtual power plant; Penalize costs for virtual power plants; Startup and shutdown costs for virtual power plants.
[0012] In the above Step 5, the snow melting algorithm is used to solve the constructed virtual power plant economic optimization dispatch model. The solution steps of the snow melting algorithm are as follows: In the initialization phase of the algorithm, the population is randomly generated; In the exploration phase, standard Brownian motion is used to simulate the irregularity of particle motion, and the elite population individuals are selected from the set formed by the sub-cycle iteration; In the development stage, the degree-day method was selected to reflect the snowmelt rate constructed by the snowmelt process; In the development phase of the snow melting algorithm, the snow melting rate was introduced to improve the position update equation during the exploration process; A dual-population mechanism is introduced, in which two populations are responsible for exploration or development respectively. When the number of iterations reaches the maximum value, the optimal solution of the economic optimization scheduling model is output.
[0013] A dispatching system using the virtual power plant economic optimization dispatching method considering multiple uncertainties includes: Distributed energy prediction module, which combines the historical data of virtual power plants to predict the wind and solar power output values on the dispatch day; The virtual power plant control center module formulates the dynamic time-of-use electricity price of the virtual power plant on the day according to the predicted wind and solar power output values, transfers the price to different units to formulate the unit output plan for the dispatch day, and completes the transaction with the power market; The unit dispatching module in the virtual power plant collects and dispatches the daily output of the virtual power plant units; The virtual power plant economic optimization scheduling solution module constructs an objective function to minimize the total operating cost of the virtual power plant during the trading day, and gives the economic optimization scheduling plan of the virtual power plant during the scheduling day based on the solution of the objective function.
[0014] The present invention provides a virtual power plant economic optimization scheduling method and system considering multiple uncertainties, which has the following beneficial effects: 1. In terms of wind and solar forecasting of virtual power plants, the present invention uses the Latin hypercube sampling method to construct typical wind and solar output scenarios based on the wind and solar output joint distribution function of the virtual power plant, and uses an improved iterative self-organizing data analysis algorithm to reduce the scenarios, which can improve the wind and solar output forecast accuracy and the scheduling accuracy of virtual power plants, optimize the real-time output of internal resources of multi-energy complementary virtual power plants, and reduce the total operating cost of virtual power plants. According to the forecast of wind and solar output a day ago, a dynamic time-of-use electricity price is constructed within the day and a price-based demand response is guided based on the electricity price, which can more effectively combine the power supply side and the load side and smooth the electric load curve.
[0015] 2. The model built by the present invention is a high-dimensional mixed integer nonlinear optimization model. The existing optimization algorithm takes a long time to solve the model and cannot guarantee accurate calculation results. The snow melting algorithm is used to solve the economic optimization scheduling model of the virtual power plant. After simulation verification, the commonly used particle swarm algorithm and gray wolf optimization algorithm have faster convergence speed in calculation. The algorithm can effectively prevent the model from converging prematurely by balancing the two stages of development and exploration. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 A schematic diagram of a virtual power plant structure of a virtual power plant economic optimization scheduling method considering multiple uncertainties according to the present invention; Figure 2 This is a schematic diagram of the workflow of a multi-objective economic dispatch system of a virtual power plant considering exergy efficiency according to the present invention; Figure 3 A typical wind and solar power output scenario generation diagram for a virtual power plant economic optimization scheduling method considering multiple uncertainties according to the present invention; Figure 4 A typical scenario reduction diagram of wind and solar power output of a virtual power plant economic optimization scheduling method considering multiple uncertainties according to the present invention; Figure 5 A schematic diagram of electric load dispatch balance of a virtual power plant economic optimization dispatch method considering multiple uncertainties according to the present invention; Figure 6 A schematic diagram of heat and gas load scheduling balance of a virtual power plant economic optimization scheduling method considering multiple uncertainties according to the present invention; Figure 7A schematic diagram of wind and solar power output, time-of-use electricity price and dynamic time-of-use electricity price in a virtual power plant economic optimization scheduling method considering multiple uncertainties of the present invention; Figure 8 This is a schematic diagram of load curve changes under different electricity prices in a multi-objective economic dispatch method for a virtual power plant considering exergy efficiency according to the present invention. DETAILED DESCRIPTION
[0017] In order to describe the present invention more clearly and completely, the implementation of the present invention will be explained in conjunction with the following embodiments through the accompanying drawings.
[0018] In the first aspect, the present invention provides a virtual power plant economic optimization scheduling method considering multiple uncertainties. The constructed virtual power plant structure is as shown in the attached figure. Figure 1 The method workflow is shown in the attached Figure 2 As shown; A virtual power plant economic optimization scheduling method considering multiple uncertainties includes the following steps: Step 1: Based on the wind-solar joint distribution function of the virtual power plant, the Latin hypercube sampling method is used to construct a typical wind-solar output scenario, and the improved iterative self-organizing data analysis algorithm is used to reduce the scenario; Step 2: Formulate a dynamic time-of-use electricity price for the virtual power plant based on the predicted wind and solar power output values; Step 3: Fully mobilize the electricity load through price-based demand response under dynamic time-of-use electricity prices; Step 4: Determine the constraints that need to be met when scheduling the virtual power plant, and construct an economic optimization scheduling model for the virtual power plant with the objective function of minimizing the total operating cost within the trading day; Step 5. Use the snow melting algorithm to solve the constructed economic optimization scheduling model of the virtual power plant, and simulate and verify that the scheduling method can improve the economic efficiency of the virtual power plant operation.
[0019] The construction process of the wind and solar power output joint distribution function of the virtual power plant described in the specific embodiment is as follows: Taking the history of virtual power plants Based on the wind and solar power output data of the day, Contribute to the scenery Samples collected during the sampling period; The non-parametric kernel density estimation method using the kernel function as Gaussian function is used to generate the probability density function of wind and solar power output; Cumulative distribution function and It is obtained based on the probability density function of wind and solar power output; The two-dimensional Frank-Copula function constructs the joint distribution function of wind and solar power output in each time period.
[0020] The calculation method of the wind-solar output probability density function and the wind-solar output joint distribution function described in the specific embodiment is as follows: ; ; In the formula, The first Sample values; is bandwidth; express and Relevance, and , A positive value indicates a positive correlation, whereas a negative correlation.
[0021] Based on the wind and solar power joint distribution function of the virtual power plant, the Latin hypercube sampling method is used to construct a typical wind and solar power output scenario, as shown in the attached figure. Figure 3 shown.
[0022] As described in the specific embodiment, an improved iterative self-organizing data analysis algorithm is used to reduce the scene. Compared with the original algorithm, the Mahalanobis distance is used instead of the Euclidean distance as the measurement when selecting cluster centers and clusters that are far apart. The simulation experiment reduces the scene to form a typical scene and an actual scene diagram, as shown in the attached figure. Figure 4 shown.
[0023] A virtual power plant is selected as an example to verify the effectiveness of a virtual power plant economic optimization scheduling method considering multiple uncertainties in the first aspect of this implementation method; the virtual power plant consists of a 135 MW pumped storage unit, three 160 MW wind farms, a 65 MW photovoltaic power station and a 120 MW gas turbine. This paper sets the time-of-use electricity purchase and sale price and the natural gas network sales price as shown in Tables 1 and 2, the virtual power plant unit parameters are shown in Table 3, and the scheduling period is set to 24 hours.
[0024] Table 1 Time-of-use electricity purchase and sale prices in the electricity market
[0025] Table 2 Natural gas network price
[0026] Table 3 Virtual power plant unit parameters
[0027] Considering the PBDR strategy of dynamic time-of-use electricity prices, the data obtained by the scenario generation method of the improved iterative self-organizing data analysis algorithm is used to perform optimal scheduling calculations. The load balance results are shown in the attached figure. Figure 5As shown, analysis of the data in the attached figure shows that: from the transmission network optimization dispatching diagram, it is known that in the time period of 01:00-07:00, the VPP system has sufficient wind and solar power output, which can sell excess electricity to the power market for profit while meeting the user load demand, and store energy during the low electricity price period through pumped storage; in the time period of 08:00-11:00, the VPP system meets the user's peak load demand through power purchase and pumped storage discharge on the basis of gas turbine output and wind and solar output; in the time period of 12: 00-17:00, the VPP system discharges intermittently through pumped storage units to ensure the sale of excess electricity and meet the needs of electrical load; in the time period of 18:00-19:00, the VPP system purchases electricity from the power grid to meet the charging of pumped storage units and user load needs; in the time period of 20:00-24:00, the pumped storage units are continuously charged and discharged to cooperate with the VPP system to sell excess gas turbine output and wind and solar output to the power market for profit.
[0028] Considering the PBDR strategy of dynamic time-of-use electricity prices, the data obtained by the scenario generation method of the improved iterative self-organizing data analysis algorithm is used for optimization scheduling calculation. The results of heat and gas load scheduling balance are shown in the attached figure. Figure 6 As shown. The virtual power plant purchases natural gas from the natural gas grid to meet the internal gas load and gas turbine operation requirements. During daily scheduling, the heat storage tank continuously switches between heat storage and heat release states to fully utilize the waste heat power of the gas turbine. The waste heat power of the heat storage tank and the gas turbine can meet the heat load demand of the virtual power plant.
[0029] The dynamic time-of-use electricity price of the virtual power plant described in the specific embodiment is formulated based on the predicted wind and solar power output value a day ago, and the dynamic time-of-use electricity price calculation formula is: ; ; ; In the formula, for Dynamic time-of-use electricity prices corresponding to the time; is the peak electricity price; To equalize electricity prices; for valley electricity prices; and Parameters for dynamic time-of-use electricity pricing; To predict the maximum total wind and solar power output in a day; To predict the minimum total wind and solar power output in a day; for The sum of all the weather and power at all times; for Wind power output at all times; for Photovoltaic output at all times.
[0030] As described in the specific embodiment, the electric load is fully mobilized by price-based demand response under dynamic time-of-use electricity prices. The sensitivity of the electric load to the electricity price described by the self-elasticity coefficient and the cross-elasticity coefficient is calculated as follows: ; ; ; In the formula, is the self-elastic coefficient; For time period The load change; For the initial period The load demand; The electricity price in the time period The amount of change; For time period Initial electricity price; is the cross elastic coefficient; For time period Change in electricity prices; For time period Initial electricity price; , and are the load changes in peak, flat and valley periods after price-based demand response; , and They are the load demands during peak, flat and valley periods before price-based demand response; , and It is the change of electricity price during peak, flat and valley periods.
[0031] The electricity price constructed by the dynamic time-of-use electricity price method proposed in this paper is shown in the attached figure. Figure 7 As shown in the figure, the dynamic time-of-use electricity price is different from the set time-of-use electricity price at different times. The dynamic time-of-use electricity price is formulated based on the wind and solar power output values predicted on the previous day. In the time periods of 01:00-11:00, 14:00 and 17:00-22:00, the wind and solar power output values are high, so the electricity price is reduced. When the VPP system is dispatched, the internal load power consumption and pumped storage units can be increased; in the time periods of 12:00-13:00, 15:00-16:00 and 23:00-24:00, the wind and solar power output values are lower than 386 MW, and a high electricity price is set to change the internal output dispatch of the VPP system. The dynamic time-of-use electricity price can adjust the load and the output of the internal units to ensure that the wind and solar power output in the VPP system are consumed first.
[0032] The load curve changes are shown in the attached figure according to the two different electricity prices. Figure 8As shown in the figure, the implementation of price-based demand response can change the original load curve, reduce the load when the electricity price is high and increase the load when the electricity price is low, and better ensure that the load can achieve the effect of "shifting peaks and filling valleys" with the change of wind and solar power output. The coefficient of variation (CV) is introduced as a measure of the effect of price-based demand response on load curve smoothing, and the value is the ratio of the standard deviation of the sample data to its mean.
[0033] The difference coefficient of the initial load curve is 0.2322; the difference coefficient of the load curve under the PBDR case considering time-of-use electricity prices is 0.2179; the difference coefficient of the load curve under the PBDR case considering dynamic time-of-use electricity prices is 0.1986. Considering different electricity prices, the smoothness of the curve is improved by 6.16% and 14.47% respectively. The price-based demand response method with dynamic time-of-use electricity prices constructed in this paper can significantly improve the smoothness of the load curve.
[0034] The calculation formula of the objective function of Step 4 described in the specific embodiment is: ; Where: Proceeds from electricity sales for virtual power plants; The operation and maintenance costs of virtual power plants; is the energy consumption cost of the virtual power plant; Penalize costs for virtual power plants; Startup and shutdown costs for virtual power plants.
[0035] A virtual power plant economic optimization scheduling method considering multiple uncertainties described in a specific embodiment is characterized in that a snow melting algorithm is used to solve the constructed virtual power plant economic optimization scheduling model. The snow melting algorithm solution steps are as follows: In the initialization phase of the algorithm, the population is randomly generated; In the exploration phase, standard Brownian motion is used to simulate the irregularity of particle motion, and the elite population individuals are selected from the set formed by the sub-cycle iteration; In the development stage, the degree-day method was selected to reflect the snowmelt rate constructed by the snowmelt process; In the development phase of the snow melting algorithm, the snow melting rate was introduced to improve the position update equation during the exploration process; A dual-population mechanism is introduced, in which two populations are responsible for exploration or development respectively. When the number of iterations reaches the maximum value, the optimal solution of the economic optimization scheduling model is output.
[0036] The particle swarm optimization (PSO), grey wolf optimization (GWO) and SAO algorithm introduced in this paper are compared and analyzed, and the maximum number of iterations is set to 180 times. The parameters of PSO and GWO are set as follows: the learning factor and of PSO are set to 1.5 and 2 respectively, the maximum moving speed of particles is set to 5, the inertia weight is 1, and the decay ratio of the inertia weight is 0.99; the initial value of the convergence factor of GWO is set to 2 and linearly reduced to 0.2, the coefficient A is a random value in the interval [-2, 2], and the coefficient B is a random value in the interval [0, 2]; the initial population size of the three algorithms is 100, and all three algorithms can solve the VPP model constructed in this paper. The comparison of the running results of PSO, GWO and SAO algorithms is shown in Table 4. The calculation time of the three algorithms to solve the model constructed in this paper is less than 20 seconds, but the running time of SAO algorithm is the shortest and it can converge at 138 iterations. In terms of solving the total operating cost of the VPP system, the SAO algorithm has the lowest cost of 39.1973 million yuan compared with the PSO and GWO algorithms. To summarize the above analysis, the SAO algorithm introduced in this paper has a significantly better convergence speed than the other two algorithms in solving the economic optimization scheduling model of the VPP system, and can find a smaller model optimal solution.
[0037] Table 4 Comparison of PSO, GWO and SAO operation results
[0038] In a second aspect, a virtual power plant economic optimization scheduling system considering multiple uncertainties is characterized by comprising: Distributed energy prediction module, which combines the historical data of virtual power plants to predict the wind and solar power output values on the dispatch day; The virtual power plant control center module formulates the dynamic time-of-use electricity price of the virtual power plant on the day according to the predicted wind and solar power output values, transfers the price to different units to formulate the unit output plan for the dispatch day, and completes the transaction with the power market; The unit dispatching module in the virtual power plant collects and dispatches the daily output of the virtual power plant units; The virtual power plant economic optimization scheduling solution module constructs an objective function to minimize the total operating cost of the virtual power plant during the trading day, and gives the economic optimization scheduling plan of the virtual power plant during the scheduling day based on the solution of the objective function.
[0039] The above-mentioned economic dispatch system includes multiple processors and memories, wherein the processor stores a computer program that can implement the above-mentioned multi-objective economic dispatch method of a virtual power plant considering exergy efficiency, and the memory is a computer-readable storage medium.
Claims
1. A virtual power plant economic optimization scheduling method considering multiple uncertainties, characterized in that: The following steps are involved: Step 1: Based on the wind-solar power output joint distribution function of the virtual power plant, the Latin hypercube sampling method is used to construct the wind-solar power output scenario, and the improved iterative self-organizing data analysis algorithm is used to reduce the scenario; Step 2: Formulate a dynamic time-of-use electricity price for the virtual power plant based on the predicted wind and solar power output values; Step 3: Mobilize the electricity load through price-based demand response under dynamic time-of-use electricity prices; Step 4: Determine the constraints that need to be met when scheduling the virtual power plant, and construct an economic optimization scheduling model for the virtual power plant with the objective function of minimizing the total operating cost within the trading day; Step 5. Use the snow melting algorithm to solve the constructed economic optimization scheduling model of the virtual power plant, and simulate and verify the scheduling method to improve the economic efficiency of the virtual power plant operation.
2. The method for economic optimization and dispatching of virtual power plants considering multiple uncertainties according to claim 1, characterized in that: The construction process of the wind and solar power output joint distribution function of the virtual power plant in Step 1 is as follows: Taking the history of virtual power plants Based on the wind and solar power output data of the day, Contribute to the scenery Samples collected during the sampling period; The non-parametric kernel density estimation method using the kernel function as Gaussian function is used to generate the probability density function of wind and solar power output; Cumulative distribution function and It is obtained based on the probability density function of wind and solar power output; The two-dimensional Frank-Copula function constructs the joint distribution function of wind and solar power output in each time period.
3. The method for economic optimization and dispatching of virtual power plants considering multiple uncertainties according to claim 2 is characterized in that: The calculation method of the wind-solar output probability density function and the wind-solar output joint distribution function is as follows: ; ; In the formula, The first Sample values; is bandwidth; express and Relevance, and , A positive value indicates a positive correlation, whereas a negative correlation.
4. The method for economic optimization and dispatching of virtual power plants considering multiple uncertainties according to claim 3 is characterized in that: Compared with the original algorithm, the improved iterative self-organizing data analysis algorithm in Step 1 reduces the scenarios by selecting cluster centers that are far apart and using Mahalanobis distance instead of Euclidean distance as a measurement when clustering.
5. The method for economic optimization scheduling of virtual power plants considering multiple uncertainties according to claim 4 is characterized in that: The dynamic time-of-use electricity price of the virtual power plant in Step 2 is formulated based on the predicted wind and solar power output value a day ago. The dynamic time-of-use electricity price calculation formula is: ; ; ; In the formula, for Dynamic time-of-use electricity prices corresponding to the time; is the peak electricity price; To equalize electricity prices; for valley electricity prices; and Parameters for dynamic time-of-use electricity pricing; To predict the maximum total wind and solar power output in a day; To predict the minimum total wind and solar power output in a day; for The sum of all the weather and power at all times; for Wind power output at all times; for Photovoltaic output at all times.
6. The method for economic optimization and dispatching of virtual power plants considering multiple uncertainties according to claim 5 is characterized in that: In the above Step 3, the power load is fully mobilized through the price-based demand response under the dynamic time-of-use electricity price. The sensitivity of the power load to the electricity price described by the self-elasticity coefficient and the cross-elasticity coefficient is calculated as follows: ; ; ; In the formula, is the self-elastic coefficient; For time period The load change; For the initial period The load demand; The electricity price in the time period The amount of change; For time period Initial electricity price; is the cross elastic coefficient; For time period Change in electricity prices; For time period Initial electricity price; , and are the load changes in peak, flat and valley periods after price-based demand response; , and They are the load demands during peak, flat and valley periods before price-based demand response; , and It is the change of electricity price during peak, flat and valley periods.
7. The method for economic optimization and dispatching of virtual power plants considering multiple uncertainties according to claim 6 is characterized in that: The objective function calculation formula in Step 4 is: ; Where: Proceeds from electricity sales for virtual power plants; The operation and maintenance costs of virtual power plants; is the energy consumption cost of the virtual power plant; Penalize costs for virtual power plants; Startup and shutdown costs for virtual power plants.
8. The method for economic optimization and dispatching of virtual power plants considering multiple uncertainties according to claim 7 is characterized in that: In the aforementioned Step 5, the constructed virtual power plant economic optimization dispatch model is solved using the snow melting algorithm. The solution steps of the snow melting algorithm are as follows: In the initialization phase of the algorithm, the population is randomly generated; In the exploration phase, standard Brownian motion is used to simulate the irregularity of particle motion, and the elite population individuals are selected from the set formed by the sub-cycle iteration; In the development stage, the degree-day method was selected to reflect the snowmelt rate constructed by the snowmelt process; In the development phase of the snow melting algorithm, the snow melting rate was introduced to improve the position update equation during the exploration process; A dual-population mechanism is introduced, in which two populations are responsible for exploration or development respectively. When the number of iterations reaches the maximum value, the optimal solution of the economic optimization scheduling model is output.
9. A dispatching system using the virtual power plant economic optimization dispatching method considering multiple uncertainties as described in claim 8, characterized in that the system include: Distributed energy prediction module, which combines the historical data of virtual power plants to predict the wind and solar power output values on the dispatch day; The virtual power plant control center module formulates the dynamic time-of-use electricity price of the virtual power plant on the day according to the predicted wind and solar power output values, transfers the price to different units to formulate the unit output plan for the dispatch day, and completes the transaction with the power market; The unit dispatching module in the virtual power plant collects and dispatches the daily output of the virtual power plant units; The virtual power plant economic optimization scheduling solution module constructs an objective function to minimize the total operating cost of the virtual power plant during the trading day, and gives the economic optimization scheduling plan of the virtual power plant during the scheduling day based on the solution of the objective function.
Citation Information
Cited By
Method and device for optimizing virtual power plant scheduling model and readable storage medium
CN120013201A
Robust optimization scheduling method and system for participation of multiple microgrids in electric auxiliary market
CN121172886A
