Optical storage layered collaborative optimization control method coupling photovoltaic priority absorption and time-of-use electricity price

Abstract

The invention discloses an optical storage layered collaborative optimization control method coupling photovoltaic priority absorption and time-of-use electricity price, and relates to the technical field of new energy power system optimization control. The method comprises the following steps: taking a photovoltaic generating capacity prediction value, a load electricity consumption prediction value and peak and valley electricity price information as input data; carrying out rolling optimization by utilizing a model prediction control framework, and generating a light storage plan table; and calculating a photovoltaic output regulation value and energy storage charging and discharging power based on the generated light storage plan table in combination with the photovoltaic power generation power and the load power consumption power which are acquired in real time, executing corresponding energy storage charging and discharging and photovoltaic output processing by applying a decision tree control mechanism, and optimizing light storage control according to a processing result. According to the method, the photovoltaic preferential consumption strategy is executed, the light abandoning amount is reduced, the photovoltaic consumption rate is improved, the time-of-use electricity price dynamic response mechanism and the energy storage efficiency compensation and light abandoning punishment mechanism are combined, peak-valley arbitrage is maximized, meanwhile, the comprehensive electricity utilization cost is reduced, and the power purchase demand of a power grid is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of new energy power system optimization control, in particular to a photovoltaic storage hierarchical collaborative optimization control method coupled with photovoltaic preferential consumption and time-of-use electricity price. BACKGROUND

[0002] Under the background of large-scale popularization and application of distributed photovoltaic power generation, the contradiction between the inherent intermittency and volatility of photovoltaic output and the randomness of power load demand is increasingly prominent, not only causing a large amount of photovoltaic output to be abandoned during the noon peak period, which seriously wastes renewable energy, but also increasing the pressure on grid peak regulation, which restricts the efficient use of distributed photovoltaic and the stable operation of the power grid.

[0003] The control strategy currently used by the photovoltaic storage system relies on static rules or simple electricity price response mechanisms, which has exposed obvious technical bottlenecks in actual application: first, the priority of photovoltaic consumption is not effectively guaranteed, and the consumption rate is generally low, with a serious problem of abandoned photovoltaic output during the noon period, which reduces the utilization rate of renewable energy and weakens the overall economy of the photovoltaic storage system; second, the coordination between the charging and discharging scheduling of the energy storage system and the time-of-use electricity price mechanism is insufficient, and the arbitrage space brought by the peak-valley electricity price difference cannot be fully tapped, which cannot effectively reduce the cost of purchasing electricity from the grid; third, the energy storage device frequently performs deep charging and discharging operations during operation, and lacks a targeted efficiency compensation mechanism, which accelerates the decay of the energy storage capacity, shortens the service life of the device, and increases the operation and maintenance cost.

[0004] Although the existing model predictive control (MPC) technology is applied to the multi-period optimization control of the photovoltaic storage system, the constraint processing capability of this technology has limitations and cannot meet the requirements of the mandatory logic of photovoltaic preferential consumption, the physical mutual exclusion constraint of energy storage charging and discharging, and the dynamic response demand of real-time electricity price, and cannot realize the collaborative optimization of the photovoltaic storage system under multiple targets such as consumption, economy, and device life.

[0005] At present, no effective solution has been proposed for the problems in the related art. SUMMARY

[0006] In view of the problems in the related art, the present application proposes a photovoltaic storage hierarchical collaborative optimization control method coupled with photovoltaic preferential consumption and time-of-use electricity price to overcome the above technical problems existing in the prior art.

[0007] To this end, the specific technical solutions adopted by the present application are as follows:

[0008] A photovoltaic storage hierarchical collaborative optimization control method coupled with photovoltaic preferential consumption and time-of-use electricity price, the method comprising:

[0009] S1, taking the photovoltaic power generation prediction value, the load power consumption prediction value and the peak-valley electricity price information as input data, using a model predictive control framework to perform rolling optimization to generate a photovoltaic storage schedule;

[0010] S2, based on the generated photovoltaic storage schedule, combining the real-time acquired photovoltaic power and load power, calculating the photovoltaic output adjustment value and the energy storage charging and discharging power, and using a decision tree control mechanism to execute corresponding energy storage charging and discharging and photovoltaic output processing, and optimizing photovoltaic storage control according to the processing result.

[0011] Further, taking the photovoltaic power generation prediction value, the load power consumption prediction value and the peak-valley electricity price information as input data, using a model predictive control framework to perform rolling optimization to generate a photovoltaic storage schedule includes:

[0012] S11, initializing energy storage system parameters and grid parameters, the energy storage system parameters including minimum capacity, maximum capacity, charging efficiency, discharging efficiency, maximum charging power and maximum discharging power, and the grid parameters including grid maximum power supply power threshold and grid minimum power supply power threshold;

[0013] S12, based on the initialization result, collecting meteorological data, historical photovoltaic power data and power consumption data, and generating a photovoltaic power generation prediction sequence according to photovoltaic module parameters, and using a load prediction algorithm to generate a load demand prediction sequence;

[0014] S13, generating an electricity price prediction sequence according to the change rule of the grid electricity price, and taking the photovoltaic power generation prediction sequence, the load demand prediction sequence and the electricity price prediction sequence as input data, and transmitting them to the control platform;

[0015] S14, the control platform receives the transmission data, uses a model predictive control framework to perform rolling optimization, obtains a charging and discharging plan, and formats the charging and discharging plan to generate a photovoltaic storage schedule.

[0016] Further, based on the initialization result, collecting meteorological data, historical photovoltaic power data and power consumption data, and generating a photovoltaic power generation prediction sequence according to photovoltaic module parameters, and using a load prediction algorithm to generate a load demand prediction sequence includes:

[0017] S121, based on the initialization result, collecting meteorological data and historical photovoltaic power data, and using a machine learning algorithm to establish an association model between the meteorological data and the historical photovoltaic power data;

[0018] S122, according to the photovoltaic module parameters, establishing a physical model, and fusing the output results of the association model and the physical model, constructing a hybrid prediction model through a fusion algorithm, and generating a photovoltaic power generation prediction sequence based on the hybrid prediction model;

[0019] S123、According to the collected power consumption data, statistics are carried out according to time units, and a load demand prediction sequence is generated through a load prediction algorithm.

[0020] Further, according to the collected power consumption data, statistics are carried out according to time units, and a load demand prediction sequence is generated through a load prediction algorithm, which includes:

[0021] S1231, according to the load data in the power consumption data, statistics are carried out according to time units, a historical load curve is generated, and the historical load curve is preprocessed;

[0022] S1232, based on the preprocessed historical load curve, a time series model is constructed by using a long short-term memory neural network, and the load behavior mode difference of each period under different scenes is identified, and the period mode of the load data is processed according to the identification result;

[0023] S1233, the collected meteorological data is taken as an input factor, combined with the load data, a load driving factor model is established;

[0024] S1234, the output results of the time series model, the period mode processing results and the output results of the load driving factor model are fused to generate a load demand prediction sequence.

[0025] Further, the control platform receives the transmission data, uses the model predictive control framework for rolling optimization, obtains the charging and discharging plan, and formats the charging and discharging plan to generate the light storage plan table, which includes:

[0026] S141, the control platform receives input data, and uses a model predictive control algorithm to construct a model predictive control framework;

[0027] S142, based on the constructed model predictive control framework, taking time units as granularity, combining the preset energy storage efficiency compensation and light abandonment penalty mechanism and constraint conditions, using a solver to perform rolling optimization on the input data;

[0028] S143, according to the rolling optimization result, the charging and discharging plan is obtained, which includes the energy storage charging and discharging plan, the photovoltaic consumption plan, the abandoned light power plan and the power grid purchase power plan;

[0029] S144, the obtained charging and discharging plan is formatted to generate a light storage plan table, which includes a charging and discharging time sequence table, a photovoltaic utilization plan table and a power purchase cost table.

[0030] Further, based on the generated light storage plan table, combining the real-time acquired photovoltaic power generation power and load power consumption power, the photovoltaic output adjustment value and the energy storage charging and discharging power are calculated, and the decision tree control mechanism is used to execute the corresponding energy storage charging and discharging and photovoltaic output processing, and the light storage control is optimized according to the processing result, which includes:

[0031] S21, based on the obtained light storage schedule, combining the preset energy storage system parameters and the real-time acquired photovoltaic power and load power, calculating the photovoltaic output adjustment value and the energy storage charging and discharging power;

[0032] S22, according to the calculated photovoltaic output adjustment value and the energy storage charging and discharging power, triggering the decision tree control mechanism, judging the relationship between the photovoltaic power and the load demand, and performing corresponding energy storage charging and discharging and photovoltaic output processing;

[0033] S23, according to the processing result, using the energy storage life extension mechanism and the real-time performance monitoring and correction mechanism to optimize the light storage control.

[0034] Further, according to the calculated photovoltaic output adjustment value and the energy storage charging and discharging power, triggering the decision tree control mechanism, judging the relationship between the photovoltaic power and the load demand, and performing corresponding energy storage charging and discharging and photovoltaic output processing includes:

[0035] When the photovoltaic power is greater than the load demand, the remaining photovoltaic power is calculated, and it is judged whether the energy storage system has charging conditions, and based on the judgment result, the energy storage system is charged or the light is abandoned;

[0036] When the photovoltaic power is less than the load demand, the shortage power of the photovoltaic power is calculated, and the time-of-use price dynamic response mechanism is used for charging and discharging processing combined with the power consumption period of the energy storage system, including peak period, flat period and valley period.

[0037] Further, when the photovoltaic power is greater than the load demand, the remaining photovoltaic power is calculated, and it is judged whether the energy storage system has charging conditions, and based on the judgment result, the energy storage system is charged or the light is abandoned, including:

[0038] When the state of charge of the energy storage system is less than the preset safe upper limit, the energy storage system has charging conditions, the remaining photovoltaic power is compared with the target charging power in the charging and discharging plan, if the remaining photovoltaic power is less than the target charging power, the energy storage system is charged according to the target charging power; if the remaining photovoltaic power is greater than the target charging power, the energy storage system is charged according to the preset safety threshold of the target charging power, and the remaining photovoltaic power is consumed, and the light is abandoned for the remaining photovoltaic power that cannot be consumed;

[0039] When the state of charge of the energy storage system is greater than the preset safe upper limit, the energy storage system does not have charging conditions, and the energy storage system enters standby state and abandons the remaining photovoltaic power.

[0040] Further, when the photovoltaic power is less than the load demand, the shortage power of the photovoltaic power is calculated, and combined with the power consumption period of the energy storage system, a time-of-use dynamic response mechanism is used for charging and discharging processing, and the power consumption period includes a peak period, a flat period and a valley period, including:

[0041] In the peak period, it is judged whether the energy storage system has a discharging condition, and corresponding charging and discharging processing is executed based on the judgment result;

[0042] In the flat period, the energy storage system charges and discharges according to the charging and discharging plan, and if there is still a shortage power after charging and discharging, the power grid supplies power;

[0043] In the valley period, if the charging and discharging plan is charging, the energy storage system charges according to the target charging power, and the grid power is less than the preset demand threshold, if the charging and discharging plan is discharging, the energy storage system discharges according to the minimum value of the target discharging power and the shortage power, and the grid supplies power to supplement the remaining shortage power.

[0044] Further, in the peak period, it is judged whether the energy storage system has a discharging condition, and corresponding charging and discharging processing is executed based on the judgment result, including:

[0045] When the state of charge of the energy storage system is greater than the preset lower limit of safety, the energy storage system has a discharging condition, the shortage power is compared with the target discharging power in the charging and discharging plan, if the shortage power is less than or equal to the target discharging power, the energy storage system discharges according to the shortage power, if the shortage power is greater than the target discharging power, the energy storage system discharges according to the maximum value of the full power and the shortage power, and if there is still a shortage power after charging and discharging, the power grid supplies power;

[0046] When the state of charge of the energy storage system is less than or equal to the preset lower limit of safety, the energy storage system does not have a discharging condition, the energy storage system enters a standby state, discharging is prohibited, and the power grid supplies power.

[0047] The beneficial effects of the present application are:

[0048] 1. The present application reduces the amount of abandoned light and improves the photovoltaic consumption rate by executing the photovoltaic preferential consumption strategy, reduces the comprehensive power consumption cost, reduces the power grid purchase demand, and enhances the system energy self-sufficiency by combining the time-of-use dynamic response mechanism and the energy storage efficiency compensation and abandoned light punishment mechanism while maximizing the peak-valley arbitrage.

[0049] 2. The present application suppresses the damage of deep charging and discharging to the energy storage system by using the charging and discharging optimization, real-time performance monitoring and correction mechanism, prolongs the service life of the system by combining the life extension mechanism, improves the response speed based on the model predictive control and decision tree mechanism, enhances the adaptability to photovoltaic fluctuations and load mutations, and reduces the operation and maintenance complexity. BRIEF DESCRIPTION OF DRAWINGS

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart of a photovoltaic-storage tiered collaborative optimization control method that couples photovoltaic priority consumption and time-of-use pricing according to an embodiment of the present invention;

[0052] Figure 2 This is a primary wiring diagram of a photovoltaic-storage-load stratified collaborative optimization control method for photovoltaic priority consumption and time-of-use pricing according to an embodiment of the present invention.

[0053] Figure 3 This is a flowchart of the hierarchical optimization control of a photovoltaic-storage hierarchical collaborative optimization control method that couples photovoltaic priority consumption and time-of-use pricing according to an embodiment of the present invention.

[0054] Figure 4 This is a schematic diagram of a real-time control decision tree for a photovoltaic-storage hierarchical collaborative optimization control method that couples photovoltaic priority consumption and time-of-use pricing according to an embodiment of the present invention.

[0055] Figure 5 This is a schematic diagram of the energy storage charge and discharge limit curve of a photovoltaic-storage hierarchical collaborative optimization control method that couples photovoltaic priority consumption and time-of-use pricing according to an embodiment of the present invention;

[0056] Figure 6 This is a functional module interaction diagram of a photovoltaic-storage hierarchical collaborative optimization control method that couples photovoltaic priority consumption and time-of-use pricing according to an embodiment of the present invention. Detailed Implementation

[0057] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention.

[0058] According to an embodiment of the present invention, a photovoltaic-storage tiered collaborative optimization control method that couples photovoltaic priority consumption and time-of-use pricing is provided.

[0059] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figures 1-2 As shown, the photovoltaic-storage tiered collaborative optimization control method according to an embodiment of the present invention, which couples photovoltaic priority consumption and time-of-use pricing, includes:

[0060] S1, taking the photovoltaic power generation prediction value, the load power consumption prediction value and the peak-valley electricity price information as input data, using a model predictive control framework to perform rolling optimization to generate a photovoltaic storage schedule.

[0061] In this optional embodiment, taking the photovoltaic power generation prediction value, the load power consumption prediction value and the peak-valley electricity price information as input data, using a model predictive control framework to perform rolling optimization to generate a photovoltaic storage schedule includes: S11, initializing the energy storage system parameters and the grid parameters, the energy storage system parameters including the minimum capacity, the maximum capacity, the charging efficiency, the discharging efficiency, the maximum charging power and the maximum discharging power, and the grid parameters including the grid maximum power supply threshold and the grid minimum power supply threshold.

[0062] It should be noted that the minimum capacity Smin and the maximum capacity Smax of the energy storage system are set, wherein Smin = 10% Smax, to ensure that the energy storage system will not be damaged due to deep discharge; the charging efficiency ηcharge and the discharging efficiency ηdischarge of the energy storage are set to 0.94, indicating that there is a 6% energy loss in the charging and discharging process of the energy storage; the maximum charging power Pmax_charge and the maximum discharging power Pmax_discharge of the energy storage system are set to meet the charging and discharging requirements and hardware capabilities of the system; the grid maximum power supply threshold Gmax and the grid minimum power supply threshold Gmin, wherein the grid maximum power supply threshold limits the upper limit of the power taken by the user side to avoid over-demand, and the grid minimum power supply threshold (preferably 0) limits the lower limit of the power sent back to the grid by the user side to prevent reverse power flow.

[0063] S12, based on the initialization result, collecting meteorological data, historical photovoltaic power data and power consumption data, and generating a photovoltaic power generation prediction sequence according to the photovoltaic module parameters, and generating a load demand prediction sequence using a load prediction algorithm.

[0064] In this optional embodiment, based on the initialization result, collecting meteorological data, historical photovoltaic power data and power consumption data, and generating a photovoltaic power generation prediction sequence according to the photovoltaic module parameters, and generating a load demand prediction sequence using a load prediction algorithm includes:

[0065] S121, based on the initialization result, collecting meteorological data and historical photovoltaic power data, and using a machine learning algorithm to establish an association model between the meteorological data and the historical photovoltaic power data.

[0066] S122, according to the photovoltaic module parameters, establishing a physical model, and fusing the output results of the association model and the physical model, constructing a hybrid prediction model through a fusion algorithm, and generating a photovoltaic power generation prediction sequence based on the hybrid prediction model.

[0067] It should be noted that, based on factors such as weather and season, the input should be the photovoltaic power generation prediction sequence for the next 24 hours [P1, P2, ..., P]. 24 (Unit: kWh), used to guide energy storage charging and discharging decisions, the steps are as follows:

[0068] 1. Meteorological data collection: By connecting to a third-party meteorological platform or installing local meteorological monitoring devices, auxiliary parameters such as solar radiation intensity, temperature, cloud cover percentage, humidity, and wind speed are obtained.

[0069] 2. Historical photovoltaic power output retrospective modeling: Establish a correlation model between historical power and meteorological data, and use machine learning methods such as regression analysis, support vector machine or random forest to model the output response of photovoltaic modules.

[0070] 3. PV module physical characteristic modeling: Based on the module parameters, such as conversion efficiency, tilt angle, area, etc., establish a physical model (such as PVWatts or PVsyst model) to calculate the ideal output under different weather conditions.

[0071] 4. Fusion of short-term forecast models: The outputs of physical models and data-driven models (i.e., models that correlate historical power with meteorological data) are fused to form a hybrid forecast model; common methods include weighted average method, LSTM neural network, etc.

[0072] 5. Time-period prediction sequence generation: Perform rolling predictions for the next 24 hours to obtain the photovoltaic power generation prediction sequence [P1,P2,……,P24].

[0073] S123. Based on the collected electricity consumption data, statistics are performed according to time units, and a load demand forecast sequence is generated through a load forecasting algorithm.

[0074] In this optional embodiment, based on the collected electricity consumption data, statistics are performed according to time units, and a load demand forecast sequence is generated using a load forecasting algorithm, including:

[0075] S1231. Based on the load data in the electricity consumption data, perform statistics according to the time unit, generate historical load curves, and preprocess the historical load curves.

[0076] S1232. Based on the preprocessed historical load curve, a time series model is constructed using a long short-term memory neural network, and the differences in load behavior patterns in different time periods under different scenarios are identified. The periodic patterns of the load data are processed according to the identification results.

[0077] S1233. Using the collected meteorological data as input factors, and combining it with load data, establish a load driving factor model;

[0078] S1234, the output result of the fusion time series model, the output result of the periodic pattern processing result, and the output result of the load driving factor model are generated to generate a load demand prediction sequence.

[0079] It should be noted that according to the load prediction algorithm, a load demand sequence [L1, L2, …, L24] (unit: kWh) of 24 hours in the future is input for calculating the load gap and the energy storage adjustment strategy, and the steps are as follows: 24

[0080] 1. Load data acquisition and preprocessing: Collect electricity data in the past 30 to 90 days through an energy management system or a smart meter; form a historical load curve by hourly statistics; perform outlier rejection and smoothing processing.

[0081] 2. Time series modeling: A long short-term memory neural network method is used to build a time series model.

[0082] 3. Periodic behavior identification: Identify the differences in load behavior between weekdays, weekends, holidays, and other scenarios for industrial parks, commercial buildings, and other scenarios; encode or classify the periodic patterns to improve prediction accuracy.

[0083] 4. Load driving factor modeling: Weather parameters (temperature, humidity), etc. are used as input factors to establish a multivariate regression model with load to improve prediction accuracy.

[0084] 5. Output hourly prediction value: The output results of the above models are integrated to generate a load demand prediction sequence [L1, L2, …, L24] of 24 hours in the future.

[0085] S13, according to the change rule of the power grid price, a price prediction sequence is generated, and the photovoltaic power generation prediction sequence, the load demand prediction sequence, and the price prediction sequence are transmitted to the control platform as input data.

[0086] It should be noted that according to the change rule of the power grid price, a price sequence [price1, price2, …, price24] (unit: yuan / kWh) of 24 hours in the future is input for price period optimization of energy storage charging and discharging decision. 24

[0087] S14, the control platform receives the transmission data, uses a model predictive control framework for rolling optimization to obtain a charging and discharging plan, and formats the charging and discharging plan to generate a photovoltaic storage plan table.

[0088] In this optional embodiment, the control platform receives the transmission data, uses a model predictive control framework for rolling optimization to obtain a charging and discharging plan, and formats the charging and discharging plan to generate a photovoltaic storage plan table, which includes:

[0089] ​​S141, the control platform receives input data and constructs a model predictive control framework using a model predictive control algorithm.

[0090] S142, based on the constructed model predictive control framework, the input data is rolled and optimized in time units, combined with the preset energy storage efficiency compensation and light abandonment penalty mechanism and constraint conditions, and a solver is used.

[0091] It should be noted that the energy storage efficiency compensation and light abandonment penalty mechanism includes dynamically adjusting the light abandonment penalty coefficient, that is:

[0092] ;

[0093] In the formula, Penalty(t) represents the light abandonment penalty coefficient; η charge represents the energy storage charging efficiency; η discharge represents the energy storage discharging efficiency; δ t represents the dynamic compensation term based on the photovoltaic prediction error.

[0094] S143, according to the rolling optimization result, a charging and discharging plan is obtained, which includes an energy storage charging and discharging plan, a photovoltaic consumption plan, a light abandonment power plan, and a power grid power purchase plan.

[0095] It should be noted that the energy storage charging and discharging plan includes outputting a sequence of energy storage charging and discharging amounts [charging and discharging amount 1, charging and discharging amount 2, …, charging and discharging amount 24 ] (unit: kWh) in the next 24 hours; the photovoltaic consumption plan includes outputting a sequence of photovoltaic consumption amounts [photovoltaic consumption amount 1, photovoltaic consumption amount 2, …, photovoltaic consumption amount 24 ] (unit: kWh) in the next 24 hours.

[0096] S144, the obtained charging and discharging plan is formatted to generate a light storage plan table, which includes a charging and discharging time sequence table, a photovoltaic utilization plan table, and a power purchase cost table.

[0097] It should be noted that the control platform receives transmission data, performs rolling optimization using the model predictive control framework, obtains a charging and discharging plan, and formats the charging and discharging plan to generate a light storage plan table, which specifically includes:

[0098] 1, the access and use location of the input data:

[0099] The photovoltaic power generation prediction data is input into the power balance constraint module of the control platform, which is used to calculate the available renewable energy in each period, and at the same time as the reference value of the light abandonment penalty term, to guide the setting of the goal of local consumption of photovoltaic power; the load prediction data is input into the power demand side module as the basic load demand in each period, which is used to construct the key item Lt ; electricity price data input to the optimization objective function module as the coefficient P of the grid electricity purchase cost term t , at the same time as the guiding factor of energy storage scheduling in the electricity price linkage adjustment module, adjust the discharge strategy to realize peak-valley arbitrage.

[0100] 2. Output data type and generation mechanism: the output data is generated by the model solver under the multi-constraint optimization framework.

[0101] (1) Energy storage charging and discharging plan:

[0102] Format: sequence [C1, C2, …, C 24 ], [D1, D2, …, D 24 ] in kWh;

[0103] Generation method: obtained by rolling optimization under multi-objective function (electricity purchase cost, curtailment penalty, SOC compensation) and constraint conditions by model predictive control (MPC) algorithm.

[0104] (2) Photovoltaic consumption plan:

[0105] Format: sequence [PV use,1 , PV use,2 , …, PV use,24 ] in kWh;

[0106] Generation method: dynamically calculated according to photovoltaic prediction and curtailment amount, and matched with energy storage chargeable power and load gap.

[0107] (3) Curtailed photovoltaic power plan:

[0108] Optional output, representing the amount of energy [PV curtail,1 , …, PV curtail,24 ] that cannot be consumed in each period due to full energy storage or low load.

[0109] (4) Grid electricity purchase plan:

[0110] Output sequence [G1, …, G 24 ], representing the energy purchased from the grid in each period, used for subsequent calculation of operation cost and implementation of scheduling strategy.

[0111] 3. Specific implementation steps:

[0112] (1) Data initialization: after system startup, import the next day's prediction data to the control platform automatically through API or data interface.

[0113] (2) Model loading and configuration: call the optimization module, load the MPC model structure, and configure optimization parameters such as objective function weight, SOC limit, electricity price model, etc.

[0114] (3) Rolling solution process: use the solver to perform 24-hour rolling optimization with hourly time granularity, and obtain the optimal charging power C of the energy storage in each time period t and discharging power D t ; actual photovoltaic utilization, calculated by electric quantity conservation; optimal power purchase of the power grid, energy storage SOC curve, light abandonment, economic indicators, etc.

[0115] (4) Output plan generation: format the solution results into a charging and discharging time sequence table, a photovoltaic utilization plan table, and a power purchase cost table, and import them into a database or a control interface.

[0116] (5) Plan result publishing: distribute the generated plan data to the real-time control module for execution control; system display interface for visualization; historical data module for comparative analysis and closed-loop correction.

[0117] (6) Verification and fault tolerance processing: perform consistency verification on the output plan, and if conflicts such as charging and discharging exclusion, SOC out-of-bounds, and power grid power over-limit are found, trigger the fault tolerance mechanism, and re-solve after adjusting the optimization weight.

[0118] 4. Rolling optimization and model predictive control framework:

[0119] (1) Under the model predictive control (MPC) framework, perform rolling optimization of the energy storage charging and discharging strategy for the next 24*N hours to minimize the overall power cost of the system, avoid the problem of light abandonment caused by excess photovoltaic power generation, and compensate for the final value of the energy storage capacity. The optimization objective is:

[0120] ; where Gt represents the power purchase of the power grid at time period t (unit: kWh), which is the amount of power obtained by the system from the power grid; Pcurtail represents the light abandonment (unit: kWh), which is the amount of photovoltaic power generation that cannot be consumed due to insufficient energy storage capacity; ST represents the energy storage capacity at time period T (unit: kWh), which represents the remaining amount of the energy storage system at the last time; S0 represents the initial capacity of the energy storage (unit: kWh), which is the initial amount of the energy storage when the system starts; and the energy storage final value compensation coefficient is used to evaluate the value of the remaining amount of the energy storage and the power cost generated by the difference between the remaining amount of the energy storage and the charging amount.

[0121] (2) Introduction of charging and discharging efficiency: in the MPC optimization process, the influence of charging and discharging efficiency on the energy flow of the energy storage system must be considered. Specifically, the actual charging power of the energy storage system during charging is affected by the charging efficiency, and the discharging amount of the energy storage system during discharging is affected by the discharging efficiency, which can be expressed as:

[0122] ;

[0123] wherein S t represents the energy storage capacity at time period t (unit: kWh); C t represents the energy storage charging power at time period t (unit: kWh); η charge represents the energy storage charging efficiency; D t represents the discharging amount at time period t (unit: kWh); η discharge represents the energy storage discharging efficiency.

[0124] (3) Constraint conditions:

[0125] A. Power balance constraint:

[0126] L t +C t =G t +P t +D t -PV curtail [t];

[0127] wherein G t represents the grid power consumption at time period t (unit: kWh); L t represents the load power consumption at time period t (unit: kWh); P t represents the photovoltaic power generation at time period t (unit: kWh); C t represents the energy storage charging power at time period t (unit: kWh); D t represents the energy storage discharging power at time period t (unit: kWh); PV curtail represents the photovoltaic power generation curtailment at time period t (unit: kWh).

[0128] B. Energy storage capacity dynamic constraint:

[0129] ;

[0130] wherein S t represents the energy storage capacity at time period t (unit: kWh); C t represents the energy storage charging power at time period t (unit: kWh); η charge represents the energy storage charging efficiency, preferably 0.94; D t represents the discharging amount at time period t (unit: kWh); η discharge represents the energy storage discharging efficiency, preferably 0.94.

[0131] C. Charging and discharging mutual exclusion constraint:

[0132] ;

[0133] ;

[0134] wherein C t represents the energy storage charging power (unit: kWh) at time period t; P max_charge represents the maximum allowed charging power of the energy storage; y t represents a binary variable, y t = 1 represents charging, y t = 0 represents discharging; D t represents the energy storage discharging power (unit: kWh) at time period t; P max_discharge represents the maximum allowed discharging power of the energy storage.

[0135] S2, based on the generated light storage schedule, combining the real-time acquired photovoltaic power and load power, calculating the photovoltaic output adjustment value and the energy storage charging and discharging power, and using the decision tree control mechanism to execute the corresponding energy storage charging and discharging and photovoltaic output processing, and optimizing the light storage control according to the processing result.

[0136] In this optional embodiment, based on the generated light storage schedule, combining the real-time acquired photovoltaic power and load power, calculating the photovoltaic output adjustment value and the energy storage charging and discharging power, and using the decision tree control mechanism to execute the corresponding energy storage charging and discharging and photovoltaic output processing, and optimizing the light storage control according to the processing result includes:

[0137] S21, based on the obtained light storage schedule, combining the preset energy storage system parameters and the real-time acquired photovoltaic power and load power, calculating the photovoltaic output adjustment value and the energy storage charging and discharging power.

[0138] It needs to be noted that, as Figure 3 shown, the energy storage system parameters include the demand threshold, the anti-backflow threshold, the energy storage capacity, the time-of-use price, and the charging and discharging power threshold.

[0139] S22, according to the calculated photovoltaic output adjustment value and the energy storage charging and discharging power, triggering the decision tree control mechanism, judging the relationship between the photovoltaic power and the load demand, and executing the corresponding energy storage charging and discharging and photovoltaic output processing.

[0140] In this optional embodiment, according to the calculated photovoltaic output adjustment value and the energy storage charging and discharging power, triggering the decision tree control mechanism, judging the relationship between the photovoltaic power and the load demand, and executing the corresponding energy storage charging and discharging and photovoltaic output processing includes:

[0141] When the photovoltaic power is greater than the load demand, the remaining photovoltaic power is calculated, and it is judged whether the energy storage system has charging conditions, and based on the judgment result, the energy storage system is charged or the light is abandoned.

[0142] In the optional embodiment, when the photovoltaic power generation power is greater than the load demand, the residual photovoltaic power generation power is calculated, and it is judged whether the energy storage system has a charging condition, and the energy storage system is charged or the photovoltaic power is abandoned based on the judgment result, including:

[0143] When the state of charge of the energy storage system is less than the preset safe upper limit, the energy storage system has a charging condition, the residual photovoltaic power generation power is compared with the target charging power in the charging and discharging plan, if the residual photovoltaic power generation power is less than the target charging power, the energy storage system is charged according to the target charging power; if the residual photovoltaic power generation power is greater than the target charging power, the energy storage system is charged according to the preset safe threshold of the target charging power, and the residual photovoltaic power generation power is consumed, and the photovoltaic power that cannot be consumed is abandoned;

[0144] When the state of charge of the energy storage system is greater than the preset safe upper limit, the energy storage system does not have a charging condition, the energy storage system enters a standby state, and the residual photovoltaic power generation power is abandoned.

[0145] It should be noted that, as shown in Figure 4 When the photovoltaic power generation power is greater than the load demand, the energy storage can only be charged or standby, and is prohibited to be discharged; the residual photovoltaic power is preferentially used for charging, and the photovoltaic power that cannot be charged is abandoned, specifically including: 1. Calculate the residual photovoltaic power (residual photovoltaic power = PV-Load), and judge whether the energy storage is allowed to be charged.

[0146] 2. If the energy storage is allowed to be charged, the SOC (state of charge) is less than the safe upper limit; if the residual power is less than the planned charging power: charge according to the planned charging power, the photovoltaic power provides part of the power, and the remaining part is provided by the power grid (planned charging power-residual power); if the residual power is greater than the planned charging power: charge according to the safe threshold of the energy storage planned charging power (the maximum of the planned charging power and the charging power safe threshold), the energy storage tries to consume the photovoltaic power generation as much as possible, and the photovoltaic power that cannot be consumed is abandoned (abandoned power = residual power-planned charging power), to avoid the energy storage sending power back.

[0147] 3. If the energy storage cannot be charged, the SOC of the energy storage is greater than the safe upper limit, the energy storage is standby, and all the residual power is abandoned.

[0148] When the photovoltaic power generation power is less than the load demand, the shortage power of the photovoltaic power generation power is calculated, and the dynamic response mechanism of time-of-use price is used for charging and discharging treatment in combination with the power consumption period in which the energy storage system is located, the power consumption period includes peak period, flat period and valley period.

[0149] In the optional embodiment, when the photovoltaic power generation power is less than the load demand, the shortage power of the photovoltaic power generation power is calculated, and the charging and discharging processing is performed by using the time-of-use electricity price dynamic response mechanism in combination with the power consumption period in which the energy storage system is located. The power consumption period includes a peak period, a flat period and a valley period.

[0150] In the peak period, it is judged whether the energy storage system has a discharging condition, and corresponding charging and discharging processing is performed based on the judgment result.

[0151] In the optional embodiment, in the peak period, it is judged whether the energy storage system has a discharging condition, and corresponding charging and discharging processing is performed based on the judgment result.

[0152] When the state of charge of the energy storage system is greater than the preset lower limit of safety, the energy storage system has a discharging condition. The shortage power is compared with the target discharging power in the charging and discharging plan. If the shortage power is less than or equal to the target discharging power, the energy storage system discharges according to the shortage power. If the shortage power is greater than the target discharging power, the energy storage system discharges according to the maximum value of the full power and the shortage power. After the charging and discharging, there is still a shortage power, which is supplemented by the power grid.

[0153] When the state of charge of the energy storage system is less than or equal to the preset lower limit of safety, the energy storage system does not have a discharging condition. The energy storage system enters a standby state, discharging is prohibited, and the power grid supplies power.

[0154] In the flat period, the energy storage system charges and discharges according to the charging and discharging plan. If there is still a shortage power after the charging and discharging, the power grid supplements the power.

[0155] In the valley period, if the charging and discharging plan is charging, the energy storage system charges according to the target charging power, and the grid power is less than the preset demand threshold. If the charging and discharging plan is discharging, the energy storage system discharges according to the minimum value of the target discharging power and the shortage power, and the power grid supplies the remaining shortage power.

[0156] It needs to be noted that, as shown in Figure 4 The photovoltaic power generation power is less than the load demand. The energy storage system discharges to supplement the shortage. When the power is insufficient, the power grid supplies power. The discharging strategy is linked with the period (peak / flat / valley). The energy storage system can discharge in the peak period, and the peak time arbitrage is maximized. The energy storage system cannot excessively discharge in the flat and valley periods, so that there is no power to discharge in the peak period. Other conditions need to be met, including the demand constraint and the anti-flow constraint.

[0157] In addition, as shown in Figure 5 When the photovoltaic power generation power is less than the load demand, the shortage power of the photovoltaic power generation power is calculated, and the charging and discharging processing is performed by using the time-of-use electricity price dynamic response mechanism in combination with the power consumption period in which the energy storage system is located. The power consumption period includes a peak period, a flat period and a valley period.

[0158] 1. Calculate the power deficit (deficit power = Load - PV) and control the charging and discharging power according to the peak and valley periods.

[0159] 2. Peak Hours: Discharge is performed according to the planned discharge power deficit (the minimum of the planned discharge power and the deficit power). During peak hours, supplementary discharge can be performed to obtain additional revenue. The energy storage charging and discharging power must meet the discharge power constraint under the current SOC. Further judgment is made on whether the energy storage is allowed to discharge (i.e., SOC > safety lower limit). If the deficit power is less than or equal to the planned discharge power, the energy storage discharges according to the deficit power. If the deficit power is greater than the planned discharge power, the energy storage discharges according to the maximum value of the full power and the deficit power, and the insufficient part is supplied by the grid. If the energy storage cannot discharge (i.e., SOC ≤ safety lower limit), the grid supplies power directly, and the energy storage does not operate.

[0160] 3. Normal period: Charge and discharge according to plan, and the remaining power shortage is provided by the power grid.

[0161] 4. Off-peak hours: If charging is planned, charging will be carried out according to the planned charging power, while ensuring that the grid power is less than the demand threshold; if discharging is planned, the energy storage will discharge according to the planned discharge power and the minimum value of the shortfall power, and the insufficient part will be supplied by the grid.

[0162] During peak electricity price periods (i.e., when the β value is relatively large, β=1.5), energy storage is prioritized for discharge to maximize the release of stored energy; during off-peak electricity price periods (i.e., when the α value is relatively large, α=1.3), energy storage is prioritized for charging to enhance the charging strategy; during flat electricity price periods (i.e., when α=β=1.0), the charging and discharging strategy is executed according to the plan.

[0163] S23. Based on the processing results, optimize the photovoltaic-storage control by using the energy storage life extension mechanism and the real-time performance monitoring and correction mechanism.

[0164] It should be further explained that the energy storage lifespan extension mechanism: when the energy storage state of health (SOH) is less than 0.7, the charging and discharging power is limited to 60% of the rated value to prevent overcharging and discharging from shortening the equipment lifespan. The real-time performance monitoring and correction mechanism: monitors the health status of the energy storage system in real time and dynamically adjusts the charging and discharging strategy according to changes in energy storage capacity; based on changes in photovoltaic power generation, load demand, energy storage capacity, and electricity prices, it adjusts the grid power purchase and energy storage charging and discharging plans in real time.

[0165] In addition, such as Figure 6 As shown, a tiered collaborative optimization control method for photovoltaic and energy storage that couples photovoltaic priority consumption and time-of-use pricing includes the following two control steps:

[0166] The day-ahead planning layer uses the next day photovoltaic power generation prediction value, load power consumption prediction value and peak-valley electricity price, combines with constraint conditions, and generates the next day charging plan through a constraint solving algorithm; the real-time control layer uses a decision tree algorithm based on the next day photovoltaic actual power generation, load real-time value, combines with the peak-valley electricity price and the real-time value of the energy storage capacity, and adjusts and controls the energy storage charging and discharging power and photovoltaic output in real time.

[0167] In specific embodiments, it is assumed that the initial capacity of the energy storage is 77 kWh, the initial capacity of the energy storage is 30 kWh, the maximum capacity of the energy storage is 100 kWh, the charging efficiency is 0.94, the discharging efficiency is 0.94, the maximum charging power of the energy storage is 50 kW, and the maximum discharging power of the energy storage is 50 kW; according to the prediction values of the next day photovoltaic power generation and load power consumption shown in Table 1, the next day charging and discharging power plan is calculated, the charging and discharging plan table is obtained as shown in Table 2, and the charging and discharging plan is sent to the real-time control module.

[0168] Table 1: Prediction values of next day photovoltaic output and load power consumption Period Photovoltaic (kWh) Load (kWh) Electricity price (yuan / kWh) 0-1 0 50 0.3 1-2 0 50 0.3 2-3 0 50 0.3 3-4 0 50 0.3 4-5 0 50 0.3 5-6 5 50 0.3 6-7 20 50 0.3 7-8 30 50 0.3 8-9 40 50 0.5 9-10 40 50 0.5 10-11 40 50 0.9 11-12 40 50 0.3 12-13 40 50 0.3 13-14 40 50 0.9 14-15 40 50 0.9 15-16 40 50 0.5 16-17 40 50 0.5 17-18 20 50 0.5 18-19 5 50 0.9 19-20 0 50 0.9 20-21 0 50 0.9 21-22 0 50 0.9 22-23 0 50 0.3 23-24 0 50 0.3

[0169] Table 2: Next day charging and discharging plan Period Abandoned light (kWh) Charging (kWh) Discharging (kWh) Grid electricity (kWh) Electricity price (yuan / kWh) Grid cost (yuan) Decision 0-1 0 50 0 100 0.3 30 Charging 1-2 0 24.47 0 74.47 0.3 22.34 Charging 2-3 0 0 0 50 0.3 15 Standby 3-4 0 0 0 50 0.3 15 Standby 4-5 0 0 0 50 0.3 15 Standby 5-6 0 0 0 45 0.3 13.5 Standby 6-7 0 0 0 30 0.3 9 Standby 7-8 0 0 0 20 0.3 6 Standby 8-9 0 0 10 0 0.5 0 Discharging 9-10 0 0 10 0 0.5 0 Discharging 10-11 0 0 10 0 0.9 0 Discharging 11-12 0 33.95 0 43.95 0.3 13.19 Charging 12-13 0 0 0 10 0.3 3 Standby 13-14 0 0 10 0 0.9 0 Discharging 14-15 0 0 10 0 0.9 0 Discharging 15-16 0 22.63 0 32.63 0.5 16.32 Charging 16-17 0 0 0 10 0.5 5 Standby 17-18 0 0 0 30 0.5 15 Standby 18-19 0 0 45 0 0.9 0 Discharging 19-20 0 0 49 1 0.9 0.9 Discharging 20-21 0 0 0 50 0.9 45 Standby 21-22 0 0 0 50 0.9 45 Standby 22-23 0 0 0 50 0.3 15 Standby 23-24 0 0 0 50 0.3 15 Standby

[0170] The real-time control module calculates the charging and discharging power of the energy storage in real time according to the charging and discharging plan, combines with the demand threshold, the anti-backflow threshold, the energy storage capacity, the time-of-use electricity price, the charging power threshold, and the real-time values of the load real-time power, the photovoltaic real-time output, the peak-valley period, etc.

[0171] Taking the actual operation of a certain industrial park on a certain day as an example: after the system is initialized, the predicted photovoltaic output (peak value of 0 kW in the afternoon), load demand (50 kW during the day and 30 kW at night), and time-of-use electricity price (0.9 yuan / kWh during peak hours and 0.3 yuan / kWh during valley hours) are input; according to the real-time control decision tree, when the photovoltaic power generation (P t ) exceeds the load (L t ) (such as P t = 80 kW and L t = 50 kW at 12:00 P t M.), the photovoltaic excess branch of the decision tree is triggered preferentially, and 30 kW of excess photovoltaic is charged into the energy storage (if the energy storage is not full); when the photovoltaic is insufficient (such as P t = 0 kW and L t = 40 kW at 19:00 P t M.) and is in a peak electricity price period, the decision tree calls the electricity price linkage mechanism, preferentially releases the energy storage (such as discharging 20 kW), and only purchases 20 kW of electricity from the power grid to reduce costs; the real-time monitoring module dynamically corrects the strategy according to the actual photovoltaic fluctuation (such as a 10% decrease in output due to cloud cover), forming a closed-loop optimization.

[0172] Another example scenario: a commercial building has 60kW PV output at 14:00, and 40kW load demand (P t ≥L t ), first calculate the net excess power 20kW, if the remaining capacity of the energy storage is 15kWh (S max =100kWh), then the charging power C t =min(20kW, 85kW)=20kW, no light waste; if the energy storage is full (S t =100kWh), then trigger the light waste 20kW; conversely, at 18:00 peak electricity price period, the PV output is 5kW, and the load demand is 35kW (P t <L t ), the lack of power is 30kW, because β=1.5 (peak time), preferentially discharge D t =min(energy storage current power 50kWh, 30kW)=30kW, and the grid purchase power G t =0, significantly reducing the cost of high electricity price period.

[0173] The economic benefits of the electricity price linkage mechanism are shown by 24-hour time sequence; taking a microgrid as an example: the system actively charges in the valley electricity price period (22:00-08:00, 11:00-13:00, electricity price 0.3 yuan / kWh, α=1.3), for example, charging 30kWh, and preferentially discharges in the peak electricity price period (10:00-11:00, 13:00-15:00, and 18:00-22:00, electricity price 0.9 yuan / kWh, β=1.5), for example, releasing 30kWh; through this strategy, 30kWh of high-price electricity purchase is reduced at peak time, saving 27 yuan, and only 9 yuan is consumed for low-price charging at valley time, with a net arbitrage of 18 yuan; at the same time, the energy storage final value compensation model ensures the final capacity of the energy storage, for example, S t =40kWh at T time, close to the initial value (S0=45kWh), avoiding excessive consumption of the energy storage for arbitrage.

[0174] As shown in Table 3, the energy storage state of health (SOH) monitoring and life protection mechanism is shown; for example, a certain energy storage system has a rated power of 50kW, and the initial SOH=0.85, which is reduced to 0.68 after long-term operation, lower than the threshold 0.7, the monitoring system immediately triggers the derating instruction, limiting the charging and discharging power to 30kW (60%×50kW); in subsequent decision-making, the decision tree adjusts the strategy in combination with the constraint: for example, when the PV is excess, the charging power C t cannot exceed 30kW, instead of the original 50kW, avoiding high-power impact; when the PV is insufficient, the discharging power D t is also limited; this mechanism combines the charging and discharging loss quantification model (η charge ×η discharge= 0.94 x 0.94 ≈ 0.88), effectively reduces the number of deep charge and discharge times, prolongs the battery life by more than 20%.

[0175] Table 3: Energy storage charge and discharge limit table Charging / discharging type Starting SOC (%) Termination SOC (%) Charging power limit (kW) Charging 0 80 150 Charging 80 95 50 Charging 95 100 0 Discharging 0 20 0 Discharging 20 100 150

[0176] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A tiered, collaborative optimization control method for photovoltaic (PV) and energy storage systems that couples PV priority consumption with time-of-use pricing, characterized in that... The method includes: S1. Using the photovoltaic power generation forecast, load power consumption forecast, and peak-valley electricity price information as input data, the model prediction and control framework is used for rolling optimization to generate a photovoltaic-storage plan table. S2. Based on the generated photovoltaic-storage plan, combined with the real-time acquired photovoltaic power generation and load power consumption, calculate the photovoltaic output adjustment value and energy storage charging and discharging power, and use the decision tree control mechanism to execute the corresponding energy storage charging and discharging and photovoltaic output processing, and optimize the photovoltaic-storage control based on the processing results.

2. The photovoltaic-storage tiered collaborative optimization control method according to claim 1, coupled with photovoltaic priority consumption and time-of-use pricing, is characterized in that... The process of using photovoltaic power generation forecasts, load power consumption forecasts, and peak-valley electricity price information as input data, and performing rolling optimization using a model prediction and control framework to generate a photovoltaic-storage plan includes: S11. Initialize the energy storage system parameters and grid parameters. The energy storage system parameters include minimum capacity, maximum capacity, charging efficiency, discharging efficiency, maximum charging power, and maximum discharging power. The grid parameters include the grid maximum power supply threshold and the grid minimum power supply threshold. S12. Based on the initialization results, collect meteorological data, historical photovoltaic power data and electricity consumption data, and generate a photovoltaic power generation prediction sequence according to the photovoltaic module parameters, and generate a load demand prediction sequence using the load forecasting algorithm. S13. Based on the changing patterns of grid electricity prices, generate an electricity price prediction sequence, and use the photovoltaic power generation prediction sequence, load demand prediction sequence, and electricity price prediction sequence as input data to transmit to the control platform; S14. The control platform receives the transmitted data, uses the model predictive control framework to perform rolling optimization, obtains the charging and discharging plan, and formats the charging and discharging plan to generate a photoelectric storage plan table.

3. The photovoltaic-storage tiered collaborative optimization control method according to claim 2, coupled with photovoltaic priority consumption and time-of-use pricing, is characterized in that... The process of collecting meteorological data, historical photovoltaic power data, and electricity consumption data based on the initialization results, generating a photovoltaic power generation prediction sequence according to the photovoltaic module parameters, and generating a load demand prediction sequence using a load forecasting algorithm includes: S121. Based on the initialization results, collect meteorological data and historical photovoltaic power data, and use machine learning algorithms to establish a correlation model between meteorological data and historical photovoltaic power data; S122. Based on the photovoltaic module parameters, establish a physical model, and integrate the output results of the correlation model and the physical model. Construct a hybrid prediction model through a fusion algorithm, and generate a photovoltaic power generation prediction sequence based on the hybrid prediction model. S123. Based on the collected electricity consumption data, statistics are performed according to time units, and a load demand forecast sequence is generated through a load forecasting algorithm.

4. The photovoltaic-storage tiered collaborative optimization control method according to claim 3, which couples photovoltaic priority consumption and time-of-use pricing, is characterized in that... The process of statistically analyzing collected electricity consumption data according to time units and generating a load demand forecast sequence using a load forecasting algorithm includes: S1231. Based on the load data in the electricity consumption data, perform statistics according to the time unit, generate historical load curves, and preprocess the historical load curves. S1232. Based on the preprocessed historical load curve, a time series model is constructed using a long short-term memory neural network, and the differences in load behavior patterns in different time periods under different scenarios are identified. The periodic patterns of the load data are processed according to the identification results. S1233. Using the collected meteorological data as input factors, and combining it with load data, establish a load driving factor model; S1234. By fusing the output results of the time series model, the periodic pattern processing results, and the output results of the load driving factor model, a load demand forecast sequence is generated.

5. The photovoltaic-storage tiered collaborative optimization control method according to claim 4, which couples photovoltaic priority consumption and time-of-use pricing, is characterized in that... The control platform receives transmitted data, performs rolling optimization using a model predictive control framework to obtain a charging and discharging plan, and formats the charging and discharging plan to generate a photovoltaic storage plan table, including: S141. The control platform receives input data and uses model predictive control algorithms to construct a model predictive control framework. S142. Based on the constructed model predictive control framework, with time units as the granularity, combined with the preset energy storage efficiency compensation and curtailment penalty mechanism and constraints, the solver is used to perform rolling optimization on the input data. S143. Based on the rolling optimization results, a charging and discharging plan is obtained, which includes an energy storage charging and discharging plan, a photovoltaic consumption plan, a curtailed photovoltaic power plan, and a grid power purchase plan. S144. The obtained charging and discharging plan is formatted to generate a photovoltaic and energy storage plan table, which includes a charging and discharging time sequence table, a photovoltaic utilization plan table, and an electricity purchase cost table.

6. The photovoltaic-storage tiered collaborative optimization control method according to claim 1, coupled with photovoltaic priority consumption and time-of-use pricing, is characterized in that... Based on the generated photovoltaic-storage plan, combined with real-time acquired photovoltaic power generation and load power consumption, the photovoltaic output adjustment value and energy storage charging and discharging power are calculated. A decision tree control mechanism is then used to execute corresponding energy storage charging and discharging and photovoltaic output processing. The photovoltaic-storage control is optimized based on the processing results, including: S21. Based on the obtained photovoltaic-storage plan, combined with the preset energy storage system parameters and the real-time photovoltaic power generation and load power consumption, calculate the photovoltaic output adjustment value and energy storage charging and discharging power. S22. Based on the calculated photovoltaic output adjustment value and energy storage charging and discharging power, trigger the decision tree control mechanism to determine the relationship between photovoltaic power generation and load demand, and execute the corresponding energy storage charging and discharging and photovoltaic output processing. S23. Based on the processing results, optimize the photovoltaic-storage control by using the energy storage life extension mechanism and the real-time performance monitoring and correction mechanism.

7. The photovoltaic-storage tiered collaborative optimization control method according to claim 6, coupled with photovoltaic priority consumption and time-of-use pricing, is characterized in that... The step of triggering a decision tree control mechanism based on the calculated photovoltaic output adjustment value and energy storage charging and discharging power to determine the relationship between photovoltaic power generation and load demand, and executing corresponding energy storage charging and discharging and photovoltaic output processing includes: When the photovoltaic power generation exceeds the load demand, the remaining photovoltaic power generation is calculated, and it is determined whether the energy storage system has the conditions for charging. Based on the determination result, the energy storage system is charged or the photovoltaic power is abandoned. When the photovoltaic power generation is less than the load demand, the shortfall in photovoltaic power generation is calculated, and the charging and discharging is processed using a time-of-use pricing dynamic response mechanism, taking into account the electricity consumption period of the energy storage system. The electricity consumption period includes peak period, normal period and valley period.

8. The photovoltaic-storage tiered collaborative optimization control method according to claim 7, coupled with photovoltaic priority consumption and time-of-use pricing, is characterized in that... When the photovoltaic power generation exceeds the load demand, the remaining photovoltaic power generation is calculated, and it is determined whether the energy storage system has charging capabilities. Based on the determination result, the energy storage system is either charged or the photovoltaic power is abandoned. When the state of charge of the energy storage system is less than the preset safety limit, the energy storage system is ready to charge. The remaining photovoltaic power generation is compared with the target charging power in the charging and discharging plan. If the remaining photovoltaic power generation is less than the target charging power, the energy storage system charges according to the target charging power. If the remaining photovoltaic power generation is greater than the target charging power, the energy storage system charges according to the preset safety threshold of the target charging power and absorbs the remaining photovoltaic power generation. The remaining photovoltaic power generation that cannot be absorbed is subject to curtailment. When the state of charge of the energy storage system exceeds the preset safety limit, the energy storage system is not ready for charging, enters standby mode, and the remaining photovoltaic power generation is subject to curtailment.

9. The photovoltaic-storage tiered collaborative optimization control method according to claim 8, which couples photovoltaic priority consumption and time-of-use pricing, is characterized in that... When the photovoltaic power generation is less than the load demand, the shortfall in photovoltaic power generation is calculated, and the charging and discharging is processed using a time-of-use pricing dynamic response mechanism, taking into account the electricity consumption period of the energy storage system. The electricity consumption period includes peak hours, normal hours, and valley hours. During peak hours, determine whether the energy storage system is ready to discharge, and perform corresponding charging and discharging processes based on the determination results; During normal periods, the energy storage system charges and discharges according to the charging and discharging plan. If there is still a power deficit after charging and discharging, the power grid will supplement the supply. During off-peak hours, if the charging / discharging plan is to charge, the energy storage system will charge according to the target charging power, and the grid power will be less than the preset demand threshold. If the charging / discharging plan is to discharge, the energy storage system will discharge according to the minimum of the target discharge power and the shortfall power, and the grid power will supplement the remaining shortfall power.

10. The photovoltaic-storage tiered collaborative optimization control method according to claim 9, characterized in that, During peak hours, determining whether the energy storage system has the conditions to discharge, and performing corresponding charging and discharging processes based on the determination result, includes: When the state of charge of the energy storage system is greater than the preset safety lower limit, the energy storage system is ready to discharge. The power deficit is compared with the target discharge power in the charge and discharge plan. If the power deficit is less than or equal to the target discharge power, the energy storage system discharges according to the power deficit. If the power deficit is greater than the target discharge power, the energy storage system discharges according to the maximum value of the full power and the power deficit. If there is still a power deficit after charging and discharging, the power grid will supplement the power supply. When the state of charge of the energy storage system is less than or equal to the preset safety lower limit, the energy storage system does not have the conditions to discharge, enters the standby state, discharge is prohibited, and the power grid provides power.

Images