Intelligent energy storage method and system for a photovoltaic power plant

Abstract

The application discloses an intelligent energy storage method and system of a photovoltaic power station, relates to the technical field of intelligent energy storage, and comprises the following steps: collecting meteorological data, constructing a fusion data set by combining photovoltaic power station and geographic information data, importing a photovoltaic prediction model, and outputting a photovoltaic prediction power sequence; acquiring a time-of-use electricity price from a power transaction platform, receiving a power fluctuation limit value issued by a dispatching system, collecting a battery pack state of charge, namely an energy storage SOC state, in real time, adopting a rolling time domain optimization solving model, solving a double-target model, and outputting an optimal charging and discharging power instruction sequence; analyzing the optimal charging and discharging power instruction sequence, clearly defining the target, submitting an acceptable power range and a unit power offer by an energy storage unit, executing a market clearing algorithm, summarizing the offers of the target and the energy storage unit, sorting by price in priority, distributing the issued instruction according to a bidding proportion after cumulative matching, and distributing the instruction to a physical energy storage unit. The whole life cycle cost minimization is realized.

Description

Technical Field

[0001] This invention relates to the field of intelligent energy storage technology, specifically to an intelligent energy storage method and system for photovoltaic power plants. Background Technology

[0002] The intermittency of photovoltaic power generation and the contradiction between grid stability are becoming increasingly prominent, and there is an urgent need to support energy storage systems to smooth out power fluctuations and improve grid absorption capacity.

[0003] Currently, photovoltaic energy storage technology faces multiple challenges: First, the calendar life of lithium batteries is significantly affected by temperature and SOC, and traditional health assessments rely on static parameters, which cannot dynamically reflect the degradation mechanism; Second, the system's charging and discharging strategies are rigid and difficult to adapt to complex operating conditions, leading to equipment overload and low efficiency; Third, in large-scale applications, the individual differences of energy storage units are ignored, resulting in a shortened overall lifespan.

[0004] Chinese invention application CN120090247A discloses a microgrid hybrid energy storage scheduling method and system, including acquiring historical load curves, renewable energy forecast data, and energy storage unit parameters; using historical load curves and renewable energy forecast data to determine the charging and discharging strategy of the energy storage unit; initially setting the SOC safety range, continuously optimizing multi-objective weight coefficients, and acquiring filter frequency band thresholds; periodically performing MPC scheduling optimization, issuing power commands in real time, and forming a closed-loop control of the SOC safety range, multi-objective weight coefficients, and filter frequency band thresholds; under the premise of microgrid hybrid energy storage, extending the service life of the energy storage unit and maintaining the optimal charging and discharging frequency during the scheduling process, continuously outputting the optimal power command sequence to assist scheduling through rolling optimization; and, with the assistance of SOH estimation, performing scheduling control on each sub-device of the energy storage unit.

[0005] In the above invention applications, the initial SOC safety range is set, the multi-objective weight coefficients are continuously optimized, and the filter frequency band threshold is obtained; MPC scheduling optimization is performed periodically, and power commands are issued in real time to form a closed-loop control of the SOC safety range, multi-objective weight coefficients, and filter frequency band threshold. However, it is necessary to track the health status of the energy storage unit in real time, estimate the aging degree online through a Kalman filter, and dynamically adjust the SOC safety range, multi-objective weight coefficients, and filter frequency band threshold. This process involves complex state estimation and parameter optimization, which places extremely high demands on hardware computing power and algorithm accuracy, increasing the computational burden.

[0006] Therefore, the present invention provides an intelligent energy storage method and system for photovoltaic power plants. Summary of the Invention

[0007] (a) Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides an intelligent energy storage method and system for photovoltaic power plants. This invention analyzes the optimal charge / discharge power command sequence to identify the target asset. Energy storage units submit acceptable power ranges and unit power bids. A market clearing algorithm is executed to aggregate the bids from the target asset and energy storage units, prioritizing them by price. After cumulative matching, commands are allocated to physical energy storage units according to the winning bid ratio. Dynamic adjustments prevent individual unit overload and extend battery cycle life. This allocation strategy achieves a balance between power point tracking accuracy and equipment lifespan, ultimately minimizing the total lifecycle cost.

[0009] (II) Technical Solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: a smart energy storage method for photovoltaic power plants, comprising the following steps:

[0011] Collect meteorological data, photovoltaic power plant data and geographic information data to construct a fused dataset, import it into the photovoltaic prediction model, and output the photovoltaic predicted power sequence;

[0012] The system obtains time-of-use electricity prices from the power trading platform, receives power fluctuation limits issued by the dispatch system, collects the state of charge (SOC) of the battery pack in real time, adopts a rolling time-domain optimization solution model, solves the dual-objective model, and outputs the optimal charging and discharging power command sequence.

[0013] The optimal charging and discharging power instruction sequence is analyzed to identify the target asset. The energy storage unit submits its acceptable power range and unit power price. The market clearing algorithm is executed to summarize the bids of the target asset and the energy storage unit. The bids are sorted by price priority, and after cumulative matching, instructions are allocated and issued to physical energy storage units according to the winning bid ratio.

[0014] Furthermore, a rolling time-domain optimization model is adopted, using the time-of-use electricity price sequence, photovoltaic predicted power sequence, power fluctuation limit, and energy storage SOC state sequence as input data. The objective function is a bi-objective optimization function, specifically: min(J1+λJ2).

[0015] Wherein, λ is the weighting coefficient, balancing the economic objective J1 and the smoothness objective J2.

[0016] J1 represents the economic objective, namely minimizing the total cost of electricity purchase.

[0017]

[0018] Among them, C k The time-of-use electricity price at time k is represented by Δt, the time step is H, the prediction time domain length is H, and P is P. grid,k The power purchased by the grid at time k is:

[0019] Pgrid,k =max(0,P) load,k -P pv,k -P batt,k )

[0020] When the load power P load,k Exceeding photovoltaic power generation P pv,k and energy storage and discharge P batt,k When the sum is equal, the difference needs to be supplemented by purchasing electricity from the grid. If energy storage discharge and photovoltaic power generation have already met the load demand (i.e., P), then... load,k ≤P pv,k +P batt,k If P, then grid,k =0, meaning no electricity is sold to the grid.

[0021] J2 is the smoothness objective, i.e., minimizing the power fluctuation penalty term:

[0022]

[0023] Among them, P batt,k Let P be the energy storage power at time k, where positive represents discharging and negative represents charging. batt,k-1 Let P be the stored energy power at time k-1. At the initial time k = t, P... batt,k-1 It can be set to 0 or a historical value.

[0024] Set constraints, including power constraints, SOC constraints, grid power fluctuation constraints, and power balance constraints, solve the bi-objective model, and output the optimal charging and discharging power command sequence.

[0025] Power constraint: -P charge,max ≤P batt,k ≤P discharge,max

[0026] P charge,max The maximum charging power of the energy storage system is determined by the rated charging capacity of the PCS power conversion system, P discharge,max The maximum discharge power of the energy storage system is determined by the rated discharge capacity of the PCS.

[0027] SOC constraint: SOC min ≤SOC k ≤SOC max

[0028] SOC min The minimum permissible state of charge (SOC) for a battery. max The maximum allowable state of charge (SOC) of the battery. k Let k be the battery state of charge at time k, representing the ratio of the current remaining charge to the total capacity, which is dynamically updated.

[0029]

[0030] SOC k-1 η represents the battery's state of charge at the previous moment. charge Charging efficiency, expressed as the ratio of the energy actually stored in the battery to the input energy during charging, η. discharge Discharge efficiency is the ratio of the actual energy output during discharge to the energy stored in the battery. This refers to the charging power, i.e., P. batt,k The absolute value when <0, The discharge power, i.e., P batt,k The absolute value when >0, C batt Δt represents the total battery capacity and the time step.

[0031] Grid power fluctuation constraints: |P batt,k -P batt,k-1 |≤Δp max

[0032] Δp max This is the power fluctuation limit.

[0033] Power balance constraint: P load,k =P pv,k +P batt,k +P grid,k

[0034] P load,k Let P be the load power at time k. pv,k Let P be the photovoltaic power generation at time k. grid,k Let P be the grid interaction power at time k. batt,k Let be the energy storage power at time k.

[0035] The specific implementation of the rolling time-domain optimization is as follows:

[0036] Prediction time domain H: Select a short window, such as 4 hours, with a step size of 15 minutes.

[0037] Rolling optimization: At each time t, update the optimization problem with the latest data.

[0038] Solve for the optimal power sequence in the finite-time domain [t, t+H-1].

[0039] Only the first control quantity is implemented. It will roll forward in the next moment.

[0040] Furthermore, the expected lifetime loss for each energy storage unit is calculated in real time for a single operation, a dynamic cost mapping table is established, and the basic loss cost of each energy storage unit is obtained based on the dynamic cost mapping table.

[0041] Furthermore, the energy storage unit is considered to have reached the end of its life when its capacity decays to 80% SOH or its internal resistance increases by 150%. Based on the battery life decay model, the total number of battery cycles and the accelerated aging coefficient are obtained, and the value of unit power loss is calculated.

[0042] Based on the battery life degradation model, the total number of battery cycles and the accelerated aging coefficient are obtained, and the value of unit power loss is calculated.

[0043] Value per cycle = Battery purchase cost / Total number of cycles; Value per unit power loss = (Accelerated aging factor for this operation) × (Value per cycle) / Operation power; Accelerated aging factor = Actual attenuation rate under operating conditions / Baseline attenuation rate; Actual attenuation rate under operating conditions is the percentage of battery capacity attenuation per unit time / cycle under current operating conditions; Baseline attenuation rate is the percentage of battery capacity attenuation per unit time / cycle under standard operating conditions.

[0044] Furthermore, the energy storage unit submits an acceptable power range ([P_min,P_max]) and a unit power price Bid, where Bid = base loss cost × (1 + strategy factor);

[0045] The specific logic for generating strategy factors is as follows:

[0046]

[0047]

[0048] The sum of all price adjustments is recorded as the strategy factor.

[0049] Furthermore, the optimal charging and discharging power instruction sequence is analyzed, the bidding period is set, and the target is identified, which is the charging and discharging power that the energy storage system needs to absorb.

[0050] Furthermore, the bids for the target assets and energy storage units are aggregated, sorted by price priority, and the discharge demand is sorted in ascending order by bid, with lower prices taking priority. The charging demand is sorted in descending order by bid, with higher prices equaling lower charging willingness. After cumulative matching, instructions are issued to allocate the units to physical energy storage units according to the winning bid ratio.

[0051] A smart energy storage system for a photovoltaic power station includes:

[0052] The photovoltaic forecasting module collects meteorological data, photovoltaic power plant data, and geographic information data to build a fused dataset, imports it into the photovoltaic forecasting model, and outputs the photovoltaic forecast power sequence.

[0053] The charge / discharge command analysis module obtains time-of-use electricity prices from the power trading platform, receives power fluctuation limits issued by the dispatch system, collects the battery pack's state of charge (SOC) in real time, and uses a rolling time-domain optimization solution model to solve the dual-objective model, outputting the optimal charge / discharge power command sequence.

[0054] The allocation module parses the optimal charging and discharging power instruction sequence, identifies the target asset, and the energy storage unit submits its acceptable power range and unit power price. The module executes the market clearing algorithm, summarizes the bids of the target asset and the energy storage unit, sorts them by price priority, and after cumulative matching, allocates and issues instructions to the physical energy storage unit according to the winning bid ratio.

[0055] (III) Beneficial Effects

[0056] This invention provides an intelligent energy storage method and system for photovoltaic power plants, which has the following beneficial effects:

[0057] 1. Collect meteorological data, photovoltaic power station data and geographic information data to build a fusion dataset, import photovoltaic prediction model, output photovoltaic predicted power sequence, accurately predict and reduce grid reserve capacity demand, reduce dispatch costs, identify irradiance mutations and current anomalies in advance, avoid grid impact caused by power fluctuations, dynamically optimize string operating points, reduce mismatch losses, and reduce the operation and maintenance cost of residential photovoltaic systems by more than 30%.

[0058] 2. By acquiring time-of-use electricity prices from the power trading platform, receiving power fluctuation limits from the dispatch system, and collecting the battery pack's state of charge (SOC) in real time, a rolling time-domain optimization model is used to solve the dual-objective model and output the optimal charging and discharging power command sequence. This achieves coordinated optimization of economy, stability, and equipment lifespan. This technology provides key technical support for photovoltaic energy storage systems to participate in the electricity market and support the construction of new power systems, promoting the transformation of energy storage from "passive response" to "active optimization."

[0059] 3. Analyze the optimal charge / discharge power command sequence, identify the target, and have the energy storage unit submit its acceptable power range and unit power bid. Execute the market clearing algorithm, summarize the bids of the target and energy storage units, sort them by price priority, and after cumulative matching, allocate and issue commands according to the winning bid ratio to physical energy storage units. Through dynamic adjustment, avoid individual unit overload and extend battery cycle life. This allocation strategy enables the energy storage system to achieve a balance between power point tracking accuracy and equipment life, ultimately minimizing the total life cycle cost. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating an intelligent energy storage method for a photovoltaic power station according to the present invention.

[0061] Figure 2 This is a schematic diagram of the structure of an intelligent energy storage system for a photovoltaic power station according to the present invention;

[0062] Figure 3 This is a schematic diagram of the dynamic cost mapping table of the present invention. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. 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.

[0064] Please see Figure 1 and Figure 3 This invention provides an intelligent energy storage method for photovoltaic power plants, comprising the following steps:

[0065] Step 1: Collect meteorological data, photovoltaic power station data, and geographic information data to construct a fused dataset, import it into the photovoltaic prediction model, and output the photovoltaic predicted power sequence.

[0066] Step one includes the following steps:

[0067] Step 101: Collect meteorological data, photovoltaic power station data, and geographic information data. Clean outliers, fill in missing data, and perform standardization / normalization. Then, perform time series alignment to construct a fused dataset.

[0068] Meteorological data includes solar radiation intensity, temperature, humidity, wind speed, cloud cover, and precipitation probability. Photovoltaic power plant data includes historical photovoltaic power generation, photovoltaic module efficiency, inverter status, and equipment fault records. Geographic information data includes the photovoltaic power plant's geographical location, tilt angle, azimuth angle, and distribution of obstructions.

[0069] Step 102: Using the fused dataset as input, import the photovoltaic prediction model and output the photovoltaic predicted power sequence.

[0070] Among them, the photovoltaic prediction model can refer to the existing CN120163279A A photovoltaic prediction method based on SARIMA-LSTM hybrid method, CN119357723A Regional distributed photovoltaic prediction method and device based on spatiotemporal convolutional neural network, etc.

[0071] When using this method, refer to steps 101 and 102:

[0072] By collecting meteorological data, photovoltaic power plant data, and geographic information data to construct a fusion dataset, importing it into a photovoltaic prediction model, and outputting a photovoltaic predicted power sequence, the system can accurately predict and reduce the grid reserve capacity demand, lower dispatch costs, identify sudden changes in irradiance and current anomalies in advance, avoid grid impacts caused by power fluctuations, dynamically optimize string operating points, reduce mismatch losses, and reduce the operation and maintenance costs of residential photovoltaic systems by more than 30%.

[0073] Step 2: Obtain time-of-use electricity prices from the power trading platform, receive power fluctuation limits issued by the dispatch system, collect the state of charge (SOC) of the battery pack in real time, use a rolling time-domain optimization solution model to solve the dual-objective model, and output the optimal charging and discharging power command sequence.

[0074] Step two includes the following steps:

[0075] Step 201: Obtain the time-of-use electricity price sequence from the power trading platform, receive the power fluctuation limit issued by the dispatch system, and collect the battery pack state of charge (SOC) sequence in real time.

[0076] Step 202: Employ a rolling time-domain optimization model, using the time-of-use electricity price sequence, photovoltaic predicted power sequence, power fluctuation limit, and energy storage SOC state sequence as input data. The objective function is a bi-objective optimization function, specifically: min(J1+λJ2).

[0077] Wherein, λ is the weighting coefficient, balancing the economic objective J1 and the smoothness objective J2.

[0078] J1 represents the economic objective, namely minimizing the total cost of electricity purchase.

[0079]

[0080] Among them, C k The time-of-use electricity price at time k is represented by Δt, the time step is H, the prediction time domain length is H, and P is P. grid,k The power purchased by the grid at time k is:

[0081] P grid,k =max(0,P) load,k -P pv,k -P batt,k )

[0082] When the load power P load,k Exceeding photovoltaic power generation P pv,k and energy storage and discharge P batt,k When the sum is equal, the difference needs to be supplemented by purchasing electricity from the grid. If energy storage discharge and photovoltaic power generation have already met the load demand (i.e., P), then... load,k ≤P pv,k +P batt,k If P, then grid,k =0, meaning no electricity is sold to the grid.

[0083] J2 is the smoothness objective, i.e., minimizing the power fluctuation penalty term:

[0084]

[0085] Among them, P batt,k Let P be the energy storage power at time k, where positive represents discharging and negative represents charging.batt,k-1 Let P be the stored energy power at time k-1. At the initial time k = t, P... batt,k-1 It can be set to 0 or a historical value.

[0086] Step 203: Set constraints, including power constraints, SOC constraints, grid power fluctuation constraints, and power balance constraints; solve the dual-objective model; and output the optimal charging and discharging power command sequence.

[0087] Power constraint: -P charge,max ≤P batt,k ≤P discharge,max

[0088] P charge,max The maximum charging power of the energy storage system is determined by the rated charging capacity of the PCS power conversion system, P discharge,max The maximum discharge power of the energy storage system is determined by the rated discharge capacity of the PCS.

[0089] SOC constraint: SOC min ≤SOC k ≤SOC max

[0090] SOC min The minimum permissible state of charge (SOC) for a battery. max The maximum allowable state of charge (SOC) of the battery. k Let k be the battery state of charge at time k, representing the ratio of the current remaining charge to the total capacity, which is dynamically updated.

[0091]

[0092] SOC k-1 η represents the battery's state of charge at the previous moment. charge Charging efficiency, expressed as the ratio of the energy actually stored in the battery to the input energy during charging, η. discharge Discharge efficiency is the ratio of the actual energy output during discharge to the energy stored in the battery. This refers to the charging power, i.e., P. batt,k The absolute value when <0, The discharge power, i.e., P batt,k The absolute value when >0, C batt Δt represents the total battery capacity and the time step.

[0093] Grid power fluctuation constraints: |P batt,k -P batt,k-1 |≤Δp max

[0094] Δp max This is the power fluctuation limit.

[0095] Power balance constraint: P load,k =P pv,k +P batt,k +P grid,k

[0096] P load,k Let P be the load power at time k. pv,k Let P be the photovoltaic power generation at time k. grid,k Let P be the grid interaction power at time k. batt,k Let be the energy storage power at time k.

[0097] The specific implementation of the rolling time-domain optimization is as follows:

[0098] Prediction time domain H: Select a short window, such as 4 hours, with a step size of 15 minutes.

[0099] Rolling optimization: At each time t, update the optimization problem with the latest data.

[0100] Solve for the optimal power sequence in the finite-time domain [t, t+H-1].

[0101] Only the first control quantity is implemented. It will roll forward in the next moment.

[0102] When using this method, refer to steps 201 to 203:

[0103] By acquiring time-of-use electricity prices from the power trading platform, receiving power fluctuation limits from the dispatch system, and collecting real-time data on the battery pack's state of charge (SOC), a rolling time-domain optimization model is used to solve a dual-objective model and output the optimal charging and discharging power command sequence. This achieves coordinated optimization of economy, stability, and equipment lifespan. This technology provides key technical support for photovoltaic energy storage systems to participate in the electricity market and support the construction of new power systems, promoting the transformation of energy storage from "passive response" to "active optimization."

[0104] Step 3: Analyze the optimal charging and discharging power instruction sequence, identify the target, and have the energy storage unit submit its acceptable power range and unit power price. Execute the market clearing algorithm, summarize the bids of the target and the energy storage unit, sort them by price priority, and after cumulative matching, allocate and issue instructions to physical energy storage units according to the winning bid ratio.

[0105] Step three includes the following:

[0106] Step 301: Calculate the expected lifetime loss of each energy storage unit in real time for a single operation, establish a dynamic cost mapping table, and obtain the basic loss cost of each energy storage unit based on the dynamic cost mapping table.

[0107] The lifespan of an energy storage unit is considered to end when its capacity decays to 80% of its state of harmonics (SOH) or its internal resistance increases by 150%. Based on the battery life decay model, the total number of battery cycles and the accelerated aging factor are obtained, and the value of unit power loss is calculated.

[0108] Value per cycle = Battery purchase cost / Total number of cycles; Value per unit power loss = (Accelerated aging factor for this operation) × (Value per cycle) / Operation power; Accelerated aging factor = Actual attenuation rate under operating conditions / Baseline attenuation rate; Actual attenuation rate under operating conditions is the percentage of battery capacity attenuation per unit time / cycle under current operating conditions; Baseline attenuation rate is the percentage of battery capacity attenuation per unit time / cycle under standard operating conditions.

[0109] Step 302: The energy storage unit submits its acceptable power range ([P_min,P_max]) and unit power price Bid, where Bid = basic loss cost × (1 + strategy factor).

[0110] The specific logic for generating strategy factors is as follows:

[0111]

[0112] The sum of all price adjustments is recorded as the strategy factor. For example, if there is a high electricity price difference in the next 15 minutes, the price will increase by 30% + SOH < 80% or the temperature > 40℃, the price will increase by 100% = 130%, then the strategy factor is 1.3.

[0113] Step 303: Analyze the optimal charging and discharging power instruction sequence, set the bidding period, identify the target, which is the charging and discharging power that the energy storage system needs to absorb, execute the market clearing algorithm, summarize the bids of the target and energy storage units, sort them by price priority, sort the discharging demand by Bid in ascending order, with low price priority, sort the charging demand by Bid in descending order, with high price = low charging willingness, and after cumulative matching, allocate and issue instructions to physical energy storage units according to the winning bid ratio.

[0114] When using this method, refer to steps 301 to 303:

[0115] The optimal charge / discharge power command sequence is analyzed to identify the target. Energy storage units submit their acceptable power range and unit power bids. The market clearing algorithm is executed to aggregate the bids of the target and energy storage units, sort them by price priority, and allocate and issue commands according to the winning bid ratio after cumulative matching. The commands are allocated to physical energy storage units. Through dynamic adjustment, individual unit overload is avoided and battery cycle life is extended. This allocation strategy enables the energy storage system to achieve a balance between power point tracking accuracy and equipment life, and ultimately minimizes the total life cycle cost.

[0116] Please see Figure 2 This invention provides an intelligent energy storage system for a photovoltaic power station, comprising:

[0117] The photovoltaic forecasting module collects meteorological data, photovoltaic power plant data, and geographic information data to build a fused dataset, imports it into the photovoltaic forecasting model, and outputs the photovoltaic forecast power sequence.

[0118] The charge / discharge command analysis module obtains time-of-use electricity prices from the power trading platform, receives power fluctuation limits issued by the dispatch system, collects the battery pack's state of charge (SOC) in real time, and uses a rolling time-domain optimization solution model to solve the dual-objective model, outputting the optimal charge / discharge power command sequence.

[0119] The allocation module parses the optimal charging and discharging power instruction sequence, identifies the target asset, and the energy storage unit submits its acceptable power range and unit power price. The module executes the market clearing algorithm, summarizes the bids of the target asset and the energy storage unit, sorts them by price priority, and after cumulative matching, allocates and issues instructions to the physical energy storage unit according to the winning bid ratio.

[0120] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0121] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0122] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A smart energy storage method for a photovoltaic power station, characterized in that: Includes the following steps: Collect meteorological data, photovoltaic power plant data and geographic information data to construct a fused dataset, import it into the photovoltaic prediction model, and output the photovoltaic predicted power sequence; The system obtains time-of-use electricity prices from the power trading platform, receives power fluctuation limits issued by the dispatch system, collects the state of charge (SOC) of the battery pack in real time, adopts a rolling time-domain optimization solution model, solves the dual-objective model, and outputs the optimal charging and discharging power command sequence. The optimal charging and discharging power instruction sequence is analyzed to identify the target asset. The energy storage unit submits an acceptable power range and a unit power price. The market clearing algorithm is executed to summarize the prices of the target asset and the energy storage unit, sort them by price priority, and after cumulative matching, instructions are allocated and issued to physical energy storage units according to the winning bid ratio. The energy storage unit submits an acceptable power range [P_min, P_max] and a unit power price Bid, where Bid = base loss cost × (1 + strategy factor); The specific logic for generating the strategy factor is as follows: when an energy storage unit triggers a price adjustment condition, all price adjustments are summed and recorded as the strategy factor.

2. The intelligent energy storage method for a photovoltaic power station according to claim 1, characterized in that: The rolling time-domain optimization solution model is used to solve the dual-objective model as follows: A rolling time-domain optimization solution model is adopted, taking the time-of-use electricity price sequence, photovoltaic predicted power sequence, power fluctuation limit and energy storage SOC state sequence as input data, and the objective function is a bi-objective optimization function min(J1+λJ2); Where λ is the weighting coefficient, balancing the economic objective J1 and the smoothness objective J2. J1 is the economic objective, which is to minimize the total electricity purchase cost, and J2 is the smoothness objective, which is to minimize the power fluctuation penalty term. Set constraints, including power constraints, SOC constraints, grid power fluctuation constraints, and power balance constraints.

3. The intelligent energy storage method for a photovoltaic power station according to claim 1, characterized in that: The expected lifetime loss of each energy storage unit is calculated in real time for a single operation, a dynamic cost mapping table is established, and the basic loss cost of each energy storage unit is obtained based on the dynamic cost mapping table.

4. The intelligent energy storage method for a photovoltaic power station according to claim 3, characterized in that: The specific steps for calculating the expected lifespan loss of each energy storage unit in real time for a single operation are as follows: Based on the battery life degradation model, the total number of battery cycles and the accelerated aging coefficient are obtained, and the value of unit power loss is calculated. Value per cycle = Battery purchase cost / Total number of cycles; Value per unit power loss = (Accelerated aging factor for this operation) × (Value per cycle) / Operation power; Accelerated aging factor = Actual attenuation rate under operating conditions / Baseline attenuation rate; Actual attenuation rate under operating conditions is the percentage of battery capacity attenuation per unit time / cycle under current operating conditions; Baseline attenuation rate is the percentage of battery capacity attenuation per unit time / cycle under standard operating conditions.

5. The intelligent energy storage method for a photovoltaic power station according to claim 1, characterized in that: The analysis of the optimal charge / discharge power command sequence clarifies the target object as follows: The optimal charging and discharging power instruction sequence is analyzed, the bidding period is set, and the target is identified, which is the charging and discharging power that the energy storage system needs to absorb.

6. The intelligent energy storage method for a photovoltaic power station according to claim 1, characterized in that: The aggregated bids for the target assets and energy storage units are sorted by price priority as follows: The bids for the target assets and energy storage units are aggregated and sorted by price priority. Discharge demand is sorted in ascending order by bid number, with lower prices taking priority. Charging demand is sorted in descending order by bid number, with higher prices indicating lower charging willingness. After cumulative matching, instructions are issued to allocate physical energy storage units according to the winning bid ratio.

7. An intelligent energy storage system for a photovoltaic power station, used to implement the method described in any one of claims 1 to 6, characterized in that: include: The photovoltaic forecasting module collects meteorological data, photovoltaic power plant data, and geographic information data to build a fused dataset, imports it into the photovoltaic forecasting model, and outputs the photovoltaic forecast power sequence. The charge and discharge command analysis module obtains the time-of-use electricity price from the power trading platform, receives the power fluctuation limit issued by the dispatch system, collects the battery pack's state of charge (SOC) in real time, adopts a rolling time-domain optimization solution model, solves the dual-objective model, and outputs the optimal charge and discharge power command sequence. The allocation module parses the optimal charging and discharging power instruction sequence, identifies the target asset, and the energy storage unit submits its acceptable power range and unit power price. The module executes the market clearing algorithm, summarizes the bids of the target asset and the energy storage unit, sorts them by price priority, and after cumulative matching, allocates and issues instructions to the physical energy storage unit according to the winning bid ratio.

Images