A multi-time scale energy management method for a light storage grid-connected system

Abstract

The present application relates to the field of new energy power system and optimal scheduling, in particular to a kind of multi-time scale energy management method of light storage grid-connected system, comprising S1, based on day-ahead operation forecast data, under the condition that energy storage operation constraint, grid-connected constraint and system power balance constraint are satisfied, to establish day-ahead scheduling model with light storage grid-connected system operation cost, S2, introduce uncertainty modeling mechanism in day-ahead scheduling model, generate day-ahead scheduling scheme considering uncertainty, S3, construct intraday operation optimization model under shorter time scale, and combine system real-time operation state, to the rolling correction of day-ahead scheduling scheme, S4, construct penalty function to the deviation of intraday operation decision relative to day-ahead scheduling scheme, combine day-ahead scheduling scheme and output the final operation control instruction of light storage grid-connected system, the present application system effectively suppresses minute-level power fluctuation, while ensuring the safe operation of system, reduces the transaction cost with power market and prolongs the service life of energy storage.

Description

Technical Field

[0001] This invention relates to the field of new energy power systems and optimized dispatching technology, specifically to a multi-timescale energy management method for a photovoltaic-storage grid-connected system. Background Technology

[0002] With the large-scale integration of renewable energy sources such as photovoltaics into distribution networks and microgrids, photovoltaic-storage grid-connected systems have become an important component of the new power system. However, photovoltaic output is affected by factors such as irradiance and weather conditions, exhibiting significant randomness and intermittency. At the same time, user load demand is affected by electricity consumption behavior and price signals, exhibiting obvious uncertainty. These factors make the photovoltaic-storage grid-connected system face significant scheduling uncertainty during operation.

[0003] In existing technologies, energy management of photovoltaic-storage grid-connected systems mostly adopts deterministic optimization methods. Because deterministic optimization methods assume that future information is completely certain, the scheduling schemes they formulate are very fragile and difficult to cope with the inherent photovoltaic and load fluctuations in photovoltaic-storage systems. This kind of "one-off" optimization lacks dynamic adjustment capabilities. Once the actual operation deviates from the prediction, both economic efficiency and safety will decrease significantly. Therefore, in order to cope with uncertainty, existing research has proposed stochastic optimization or robust optimization methods. However, robust optimization models under a single time scale are often too conservative and cannot take into account both the economic efficiency and flexibility of the system.

[0004] Furthermore, photovoltaic power output and load power fluctuate significantly on a minute-level timescale, while traditional day-ahead dispatching typically operates on an hourly timescale, making it difficult to respond promptly to real-time operational deviations and resulting in a mismatch between the dispatching scheme and the actual operating status. Therefore, there is an urgent need for an energy management method for photovoltaic-storage grid-connected systems that can balance economy, robustness, and real-time performance across multiple timescales. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-timescale energy management method for photovoltaic-storage grid-connected systems, in order to solve the problems of excessively conservative scheduling and insufficient real-time adaptability caused by the uncertainty of photovoltaic output and load forecasting in the prior art, and to achieve synergistic optimization of system economy and operational robustness.

[0006] This invention solves the above-mentioned technical problems through the following technical solution: a multi-timescale energy management method for a photovoltaic-storage grid-connected system, comprising the following steps:

[0007] Step S1: Based on the day-ahead operation forecast data, and under the conditions of satisfying the energy storage operation constraints, grid connection constraints and system power balance constraints, establish a day-ahead scheduling model with the operating cost of the photovoltaic-storage grid-connected system.

[0008] Step S2: Introduce an uncertainty modeling mechanism into the day-ahead scheduling model to describe the uncertain fluctuations in photovoltaic power generation output and load demand, and generate a day-ahead scheduling scheme that takes uncertainty into account.

[0009] Step S3: Using the day-ahead scheduling scheme as a reference, construct an intraday operation optimization model on a shorter time scale, and combine the real-time operating status of the system to make rolling corrections to the day-ahead scheduling scheme;

[0010] Step S4: During the intraday operation optimization process, a penalty function is constructed for the deviation of the intraday operation decision from the day-ahead scheduling scheme, and the final operation control command of the photovoltaic-storage grid-connected system is output in combination with the day-ahead scheduling scheme.

[0011] Preferably, in step S1, the energy storage operation constraints include:

[0012] ;

[0013] ;

[0014] ;

[0015] ;

[0016] In the formula, P C (t) represents the charging power of the energy storage battery at time t, P C.max P represents the maximum charging power of the energy storage battery. D (t) represents the discharge power of the energy storage battery at time t, P D.max State of Charge (SOC) represents the maximum discharge power of an energy storage battery. t State of Charge (SOC) represents the charge capacity of the energy storage battery at time t. min State of Charge (SOC) represents the lower limit of the energy storage system's capacitance. max η represents the upper limit of the energy storage system's capacitance. C η represents the charging efficiency of an energy storage battery. D This represents the discharge efficiency of the energy storage battery, where Δt is the time interval.

[0017] Preferably, in step S1, the grid connection constraints include:

[0018] ;

[0019] ;

[0020] In the formula, P buy (t) represents the power that the system cannot meet the load demand at time t, and needs to purchase electricity from the grid. sell P(t) represents the surplus electrical energy in the system at time t, which is sold to the grid.buymax P represents the maximum power purchased. sellmax This indicates the maximum power output.

[0021] Preferably, in step S1, the power balance constraints of the photovoltaic-storage grid-connected system include:

[0022] ;

[0023] In the formula, P pv (t) represents the photovoltaic power generation at time t, P load (t) represents the local load power connected to the system at time t.

[0024] Preferably, in step S1, the cost F of the energy storage battery during the investment payback period at time t is... bat (t) is represented as:

[0025] ;

[0026] In the formula, K bat K is the total investment cost coefficient per unit capacity of energy storage batteries. B This is the coefficient for the operation and maintenance costs of energy storage batteries;

[0027] The cost F of energy interaction between the photovoltaic-storage grid-connected system and the grid at time t grid (t) is represented as:

[0028] ;

[0029] In the formula, C buy (t) represents the electricity price at time t, C sell (t) represents the electricity price at time t.

[0030] Preferably, in step S1, under constraints, a day-ahead scheduling model is established with the minimum cost within a 24-hour scheduling cycle as the optimization objective:

[0031] .

[0032] Preferably, in the day-ahead scheduling model, P within the same scheduling period buy (t) and P sell (t), P C (t) and P D (t) Complementary, introducing a state flag U g and U b To determine the direction of electricity purchase and sale, as well as the direction of charging and discharging;

[0033] When U g When the value is 1, it indicates that the internal power of the photovoltaic-storage grid-connected system is insufficient and it needs to purchase power from external sources;

[0034] When U g When the value is 0, it indicates that the photovoltaic-storage grid-connected system has surplus energy and can sell electricity to the external power grid.

[0035] Similarly, when U b When the value is 1, the energy storage battery is in charging mode;

[0036] When U b When the value is 0, the energy storage battery is in discharge mode.

[0037] Preferably, step S2 includes:

[0038] Step S21: Introduce a box-type uncertainty set G to construct an uncertainty modeling mechanism. The uncertainty set G includes the fluctuation range of photovoltaic power output and load.

[0039] ;

[0040] In the formula, g is an uncertain variable that includes both photovoltaic power and load. Let be the predicted photovoltaic power at time t. Let be the predicted load power at time t. For the largest fluctuation in photovoltaic power, This represents the maximum fluctuation in load.

[0041] Step S22, introduce binary variable β=(β pv ,β load ) T and uncertain adjustment parameter Γ pv and Γ load Describe the degree of uncertainty and volatility:

[0042] ;

[0043] β=(β pv ,β load ) T For boundary indicator factors in uncertain scenarios;

[0044] Γ pv and Γ load All are uncertain adjustment parameters, with values ​​ranging from 0 to 24.

[0045] Preferably, step S3 includes:

[0046] Step S31: Collect real-time operating parameters of the photovoltaic-storage grid-connected system within one cycle, with each cycle being 15 minutes. The real-time operating parameters include the energy storage state of charge and tie-line power. Call the power prediction curve for the next 2 hours. Based on the real-time operating parameters and the power prediction curve, use mixed integer linear programming to solve the optimal output sequence of each device for the next 2 hours, and obtain the optimization result of the first 15-minute scheduling cycle in the optimal output sequence.

[0047] Step S32: When the next scheduling cycle is triggered, the daily scheduling plan is revised again based on the updated forecast data and the actual operating status to form the actual intraday operating decision.

[0048] Preferably, step S4 includes:

[0049] Step S41: Based on the electricity purchase and sale deviation and energy storage charging and discharging deviation generated by the intraday operation decision relative to the day-ahead scheduling plan, a penalty function is constructed by quantifying the impact of the deviation on electricity purchase and sale costs and energy storage lifetime.

[0050] ;

[0051] In the formula, ζ is the penalty coefficient for deviation in electricity purchase and sale, ψ is the penalty coefficient for deviation in energy storage charging and discharging, and P * buy For the electricity purchase plan submitted recently, P * sell For the electricity sales plan submitted recently, P C* For the daytime energy storage charging plan, P D* This is the day-ahead energy storage discharge plan;

[0052] Step S42: Add the penalty function to the system operating cost function to obtain the intraday rolling optimization model, and output the final operation control command of the photovoltaic-storage grid-connected system in combination with the day-ahead scheduling scheme.

[0053] The beneficial effects of this invention are as follows:

[0054] This invention introduces an uncertainty modeling and correction mechanism during the day-ahead scheduling phase, enabling the generated scheduling scheme to strictly meet power balance and grid connection constraints even in extreme cases where photovoltaic power generation output and load demand fluctuate. This effectively reduces the day-ahead plan failure rate and achieves a balance between system operational robustness and economy. Combined with intraday rolling correction and deviation linearization processing, it effectively smooths out minute-level power fluctuations, ensuring safe system operation while reducing transaction costs with the electricity market and extending the lifespan of energy storage. Attached Figure Description

[0055] Figure 1 This is a flowchart illustrating the overall process of the energy optimization method under multiple time scales in this embodiment of the invention.

[0056] Figure 2 This is a schematic diagram of the overall structure of the photovoltaic-storage grid-connected system in an embodiment of the present invention;

[0057] Figure 3 This is a schematic diagram of the intraday rolling optimization strategy in an embodiment of the present invention;

[0058] Figure 4 This is a graph showing the day-ahead SOC change of energy storage in an embodiment of the present invention;

[0059] Figure 5 This is a graph showing the intraday rolling optimization results in an embodiment of the present invention;

[0060] Figure 6 This is a comparison chart of the day-to-day electricity purchase and sale plans in an embodiment of the present invention;

[0061] Figure 7 This is a comparison chart of the daytime and daytime energy storage charging and discharging plans in an embodiment of the present invention; Detailed Implementation

[0062] The embodiments of the present invention are described in detail below. These embodiments are implemented based on the technical solution of the present invention, and provide detailed implementation methods and specific operation processes. However, the scope of protection of the present invention is not limited to the following embodiments.

[0063] Example 1

[0064] like Figures 1-3 As shown, based on day-ahead operational forecast data, a day-ahead scheduling model is established with the operating cost of the photovoltaic-storage grid-connected system, under the conditions of satisfying energy storage operation constraints, grid connection constraints, and system power balance constraints.

[0065] Energy storage operation constraints include:

[0066] ;

[0067] ;

[0068] ;

[0069] ;

[0070] In the formula, P C (t) represents the charging power of the energy storage battery at time t, P C.max P represents the maximum charging power of the energy storage battery. D (t) represents the discharge power of the energy storage battery at time t, P D.max State of Charge (SOC) represents the maximum discharge power of an energy storage battery. t State of Charge (SOC) represents the charge capacity of the energy storage battery at time t.min State of Charge (SOC) represents the lower limit of the energy storage system's capacitance. max η represents the upper limit of the energy storage system's capacitance. C η represents the charging efficiency of an energy storage battery. D This represents the discharge efficiency of the energy storage battery, where Δt is the time interval.

[0071] Grid connection constraints include:

[0072] ;

[0073] ;

[0074] In the formula, P buy (t) represents the power that the system cannot meet the load demand at time t, and needs to purchase electricity from the grid. sell P(t) represents the surplus electrical energy in the system at time t, which is sold to the grid. buymax P represents the maximum power purchased. sellmax This indicates the maximum power output for electricity sales.

[0075] The power balance constraints of the photovoltaic-storage grid-connected system include:

[0076] ;

[0077] In the formula, P pv (t) represents the photovoltaic power generation at time t, P load (t) represents the local load power connected to the system at time t;

[0078] The cost F of the energy storage battery during its payback period at time t bat (t) is represented as:

[0079] ;

[0080] In the formula, K bat K is the total investment cost coefficient per unit capacity of energy storage batteries. B This is the coefficient for the operation and maintenance costs of energy storage batteries;

[0081] The cost F of energy interaction between the photovoltaic-storage grid-connected system and the grid at time t grid (t) is represented as:

[0082] ;

[0083] In the formula, C buy (t) represents the electricity price at time t, C sell (t) represents the electricity price at time t;

[0084] Under constraints, a day-ahead scheduling model is established with the goal of minimizing the cost within a 24-hour scheduling cycle:

[0085] ;

[0086] In the day-ahead scheduling model, P within the same scheduling period buy (t) and P sell (t), P C (t) and P D (t) Complementary, introducing a state flag U g and U b To determine the direction of electricity purchase and sale, as well as the direction of charging and discharging;

[0087] When U g When the value is 1, it indicates that the internal power of the photovoltaic-storage grid-connected system is insufficient and it needs to purchase power from external sources;

[0088] When U g When the value is 0, it indicates that the photovoltaic-storage grid-connected system has surplus energy and can sell electricity to the external power grid.

[0089] Similarly, when U b When the value is 1, the energy storage battery is in charging mode;

[0090] When U b When the value is 0, the energy storage battery is in discharge mode.

[0091] An uncertainty modeling mechanism is introduced into the day-ahead dispatching model to describe the uncertain fluctuations in photovoltaic power output and load demand, generating a day-ahead dispatching scheme that considers uncertainty. A box-type uncertainty set G is introduced to construct the uncertainty modeling mechanism, which includes the fluctuation range of photovoltaic power output and load.

[0092] ;

[0093] In the formula, g is an uncertain variable that includes both photovoltaic power and load. Let be the predicted photovoltaic power at time t. Let be the predicted load power at time t. For the largest fluctuation in photovoltaic power, This represents the maximum fluctuation in load.

[0094] Introducing binary variable β=(β pv ,β load ) T and uncertain adjustment parameter Γ pv and Γ load Describe the degree of uncertainty and volatility:

[0095] ;

[0096] β=(β pv ,β load ) T For boundary indicator factors in uncertain scenarios;

[0097] Γ pv and Γ load All are uncertain adjustment parameters, taking values ​​from 0 to 24 as integers. An uncertain adjustment parameter Γ is introduced. pv and Γ load This approach achieves a balance between robustness and economy: when the parameter values ​​approach the upper limit, the resulting scheduling scheme is highly robust, but its economy deteriorates due to its greater conservatism; conversely, when the parameters approach the lower limit, the scheme's robustness weakens while its economy improves, but it may be difficult to guarantee the reliability of the scheduling scheme in the face of fluctuations in uncertain variables.

[0098] Based on the day-ahead scheduling scheme, an intraday operation optimization model is constructed on a shorter time scale, and the day-ahead scheduling scheme is continuously revised in combination with the real-time operation status of the system.

[0099] The real-time operating parameters of the photovoltaic-storage grid-connected system are collected every 15 minutes. The real-time operating parameters include the energy storage state of charge and tie-line power. The power prediction curve for the next 2 hours is called up. Based on the real-time operating parameters and the power prediction curve, the optimal output sequence of each device for the next 2 hours is solved by mixed integer linear programming. The optimization result of the first 15-minute scheduling cycle in the optimal output sequence is obtained.

[0100] Step S32: When the next scheduling cycle is triggered, the daily scheduling plan is revised again based on the updated forecast data and the actual operating status to form the actual intraday operating decision.

[0101] During the intraday operation optimization process, a penalty function is constructed for the deviation of the intraday operation decision from the day-ahead scheduling scheme, and the final operation control command of the photovoltaic-storage grid-connected system is output in combination with the day-ahead scheduling scheme.

[0102] Based on the purchase and sale deviations and energy storage charging and discharging deviations generated by intraday operation decisions relative to day-ahead dispatch plans, a penalty function is constructed by quantifying the impact of these deviations on the cost of electricity purchase and sale and the lifespan of energy storage.

[0103] ;

[0104] In the formula, ζ is the penalty coefficient for deviation in electricity purchase and sale, ψ is the penalty coefficient for deviation in energy storage charging and discharging, and P * buy For the electricity purchase plan submitted recently, P * sell For the electricity sales plan submitted recently, P C*For the daytime energy storage charging plan, P D* This is the day-ahead energy storage discharge plan;

[0105] By adding the penalty function to the system operating cost function, an intraday rolling optimization model is obtained. Combined with the day-ahead scheduling scheme, the final operation control command of the photovoltaic-storage grid-connected system is output.

[0106] Example 2

[0107] In this embodiment, a typical photovoltaic-storage grid-connected power generation system is selected for analysis. In the formula, the known peak load in the system is 380kW, the maximum capacity of the photovoltaic power generation system is 400kW, the rated capacity of the energy storage device is 300kWh, and the maximum allowable charging and discharging power is 100kW. In order to prevent overcharging and over-discharging of the energy storage battery, the SOC range is set to 0.2~0.8 based on the parameters of the energy storage device itself, and the maximum power of interaction with the external power grid is 300kW.

[0108] Setting uncertain adjustment parameters Γ pv and Γ load The values ​​of 8 and 12 indicate that during the day-ahead scheduling optimization process, the photovoltaic power reached the minimum value of the prediction interval in 8 time periods, and the load power reached the maximum value of the prediction interval in 12 time periods. For the remaining time, both values ​​were predicted.

[0109] Using Matlab and the Yalmip toolbox to call the CPLEX solver, the scheduling conclusions are as follows:

[0110] Figure 4 The SOC variation curve shown indicates that the system strictly adheres to the preset safe operating range of 20%-80% throughout the entire scheduling cycle. This constraint effectively avoids situations such as deep discharge and overcharging that damage the battery's health. Furthermore, the remaining capacity of the energy storage system is equal at the beginning and end of the cycle. This helps optimize the operating strategy of the energy storage system, improves the lifespan of the energy storage battery, enhances the stability and reliability of the photovoltaic-energy storage grid-connected system, and provides a strong guarantee for the long-term and sustainable operation of the system.

[0111] Figure 5The diagram illustrates a daily rolling optimization scheduling scheme based on a 15-minute timescale. During peak periods (8:00-11:00 and 16:00-21:00), the energy storage unit primarily performs discharge operations, effectively shaving off peak loads. During off-peak periods, the energy storage unit switches to charging mode to fill the valleys. During periods of normal electricity prices, the energy storage system dynamically adjusts its charging and discharging states to achieve energy balance between photovoltaic (PV) power generation and load demand. Specifically, during the off-peak period (0:00-8:00) and due to insufficient PV output, the system ensures power supply to the load by purchasing high-power electricity from the external grid. After entering the peak period (8:00-11:00), as PV output increases and the external grid electricity price rises, the amount of electricity purchased from the external grid decreases significantly, and surplus PV power is sold back to the grid during periods of PV surplus. After 16:00, when PV output declines, the system gradually increases the amount of electricity purchased from the external grid and coordinates with energy storage discharge to ensure a continuous and stable supply to the load demand.

[0112] Figure 6 This is a comparison chart of the day-ahead and intraday power purchase and sale plans. It can be seen that the intraday optimization plan tries to track the day-ahead power purchase and sale plan as much as possible. This is because the day-ahead power purchase and sale plan needs to be reported to the external power market in advance. If there are too many changes, it will generate high change costs. Therefore, the day-ahead power purchase and sale plan needs to be tracked as much as possible when optimizing intraday.

[0113] Figure 7 This is a comparison chart of the day-ahead and intraday energy storage charging and discharging plans. It can be seen that the intraday rolling optimization is more frequent than the day-ahead plan for energy storage charging and discharging. This is due to the error in the forecasting process of the day-ahead dispatch. The cost of changing the power purchase and sale plan is greater than the cost of changing the energy storage charging and discharging plan. Therefore, energy storage devices are chosen to mitigate the error between forecast and reality.

[0114] By setting the penalty function parameters of the intraday rolling optimization model, more tracking day-ahead electricity purchase and sale plans or charging and discharging plans can be selected.

[0115] In summary, the multi-timescale energy management method for a photovoltaic-storage grid-connected system in this embodiment achieves a balance between system robustness and economy by setting uncertainty adjustment parameters. Combined with intraday rolling correction and deviation linearization processing, it effectively smooths out minute-level power fluctuations, ensuring safe system operation while reducing transaction costs with the electricity market and extending the lifespan of energy storage.

[0116] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A multi-timescale energy management method for a photovoltaic-storage grid-connected system, characterized in that, Includes the following steps: Step S1: Based on the day-ahead operation forecast data, and under the conditions of satisfying the energy storage operation constraints, grid connection constraints and system power balance constraints, establish a day-ahead scheduling model with the operating cost of the photovoltaic-storage grid-connected system. Step S2: Introduce an uncertainty modeling mechanism into the day-ahead scheduling model to describe the uncertain fluctuations in photovoltaic power generation output and load demand, and generate a day-ahead scheduling scheme that takes uncertainty into account. Step S3: Using the day-ahead scheduling scheme as a reference, construct an intraday operation optimization model on a shorter time scale, and combine the real-time operating status of the system to make rolling corrections to the day-ahead scheduling scheme; Step S4: During the intraday operation optimization process, a penalty function is constructed for the deviation of the intraday operation decision from the day-ahead scheduling scheme, and the final operation control command of the photovoltaic-storage grid-connected system is output in combination with the day-ahead scheduling scheme.

2. The multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 1, characterized in that: In step S1, the energy storage operation constraints include: ； ； ； ； In the formula, P C (t) represents the charging power of the energy storage battery at time t, P C.max P represents the maximum charging power of the energy storage battery. D (t) represents the discharge power of the energy storage battery at time t, P D.max State of Charge (SOC) represents the maximum discharge power of the energy storage battery. t State of Charge (SOC) represents the charge capacity of the energy storage battery at time t. min State of Charge (SOC) represents the lower limit of the energy storage system's capacitance. max η represents the upper limit of the energy storage system's capacitance. C Indicates the charging efficiency of energy storage batteries, η D This represents the discharge efficiency of the energy storage battery, where Δt is the time interval.

3. The multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 2, characterized in that: In step S1, the grid connection constraints include: ； ； In the formula, P buy (t) represents the power that the system cannot meet the load demand at time t, and needs to purchase electricity from the grid. sell P(t) represents the surplus electrical energy in the system at time t, which is sold to the grid. buymax P represents the maximum power purchased. sellmax This indicates the maximum power output.

4. The multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 3, characterized in that: In step S1, the power balance constraints of the photovoltaic-storage grid-connected system include: ； In the formula, P pv (t) represents the photovoltaic power generation at time t, P load (t) represents the local load power connected to the system at time t.

5. A multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 4, characterized in that: In step S1, the cost F of the energy storage battery during the investment payback period at time t is... bat (t) is represented as: ； In the formula, K bat K is the total investment cost coefficient per unit capacity of energy storage batteries. B This is the coefficient for the operation and maintenance costs of energy storage batteries; The cost F of energy interaction between the photovoltaic-storage grid-connected system and the grid at time t grid (t) is represented as: ； In the formula, C buy (t) represents the electricity price at time t, C sell (t) represents the electricity price at time t.

6. A multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 5, characterized in that: In step S1, under constraints, a day-ahead scheduling model is established with the minimum cost within a 24-hour scheduling cycle as the optimization objective: 。 7. A multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 6, characterized in that: In the day-ahead scheduling model, P within the same scheduling period buy (t) and P sell (t), P C (t) and P D (t) Complementary, introducing a state flag U g and U b To determine the direction of electricity purchase and sale, as well as the direction of charging and discharging; When U g When the value is 1, it indicates that the internal power of the photovoltaic-storage grid-connected system is insufficient and it needs to purchase power from external sources; When U g When the value is 0, it indicates that the photovoltaic-storage grid-connected system has surplus energy and can sell electricity to the external power grid. Similarly, when U b When the value is 1, the energy storage battery is in charging mode; When U b When the value is 0, the energy storage battery is in discharge mode.

8. The multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 1, characterized in that: Step S2 includes: Step S21: Introduce a box-type uncertainty set G to construct an uncertainty modeling mechanism. The uncertainty set G includes the fluctuation range of photovoltaic power output and load. ； In the formula, g is an uncertain variable that includes both photovoltaic power and load. Let be the predicted photovoltaic power at time t. Let be the predicted load power at time t. For the largest fluctuation in photovoltaic power, This represents the maximum fluctuation in load. Step S22, introduce binary variable β=(β pv ,β load ) T and uncertain adjustment parameter Γ pv and Γ load Describe the degree of uncertainty and volatility: ； β=(β pv ,β load ) T For boundary indicator factors in uncertain scenarios; Γ pv and Γ load All are uncertain adjustment parameters, with values ​​ranging from 0 to 24.

9. A multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 1, characterized in that: Step S3 includes: Step S31: Collect real-time operating parameters of the photovoltaic-storage grid-connected system within one cycle, with each cycle being 15 minutes. The real-time operating parameters include the energy storage state of charge and tie-line power. Call the power prediction curve for the next 2 hours. Based on the real-time operating parameters and the power prediction curve, use mixed integer linear programming to solve the optimal output sequence of each device for the next 2 hours, and obtain the optimization result of the first 15-minute scheduling cycle in the optimal output sequence. Step S32: When the next scheduling cycle is triggered, the daily scheduling plan is revised again based on the updated forecast data and the actual operating status to form the actual intraday operating decision.

10. A multi-timescale energy management method for a photovoltaic-storage grid-connected system according to claim 1, characterized in that: Step S4 includes: Step S41: Based on the purchase and sale deviation and energy storage charging and discharging deviation generated by the intraday operation decision relative to the day-ahead scheduling plan, a penalty function is constructed by quantifying the impact of the deviation on the purchase and sale cost and energy storage lifetime. ； In the formula, ζ is the penalty coefficient for deviation in electricity purchase and sale, ψ is the penalty coefficient for deviation in energy storage charging and discharging, and P * buy For the electricity purchase plan submitted recently, P * sell For the electricity sales plan submitted recently, P C* For the daytime energy storage charging plan, P D* This is the day-ahead energy storage discharge plan; Step S42: Add the penalty function to the system operating cost function to obtain the intraday rolling optimization model, and output the final operation control command of the photovoltaic-storage grid-connected system in combination with the day-ahead scheduling scheme.

Images