Energy storage coordination system and control method based on meteorological prediction and multi-energy complementation

Abstract

The invention relates to the technical field of new energy power systems, and discloses an energy storage coordination system and control method based on meteorological prediction and multi-energy complementation, and the system comprises a power generation end module which carries out the dynamic adjustment according to a power adjustment instruction of a power generation end; the meteorological prediction and data acquisition module is used for acquiring meteorological data in a future time period in real time and sending the meteorological data to the intelligent control and scheduling platform; the energy storage end module sends data including lithium battery SOC and hydrogen storage tank pressure to the intelligent control and scheduling platform and performs charging and discharging according to an energy storage end charging and discharging priority strategy; and the intelligent control and scheduling platform receives the electricity price data of the weather prediction and data acquisition module, the energy storage end module and the power grid in real time, generates a power generation end power regulation instruction and an energy storage end charging and discharging priority strategy, and meanwhile, improves the income and maximizes the hydrogen energy use proportion through electricity price peak-valley arbitrage and green electricity transaction premium. The power grid stability and the energy utilization rate are greatly improved, the energy storage life is prolonged, and the investment payback period is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy power systems, and particularly relates to an energy storage coordination system and a control method based on meteorological prediction and multi-energy complementarity. Background Art

[0002] Photovoltaic and wind power have significant intermittency and volatility. Photovoltaic power generation is affected by day and night, weather (such as cloud cover), and wind power depends on wind speed changes, resulting in frequent fluctuations in output power. This volatility not only affects the stability of local electricity consumption but also causes power surges during grid connection, increasing the grid frequency regulation pressure.

[0003] The existing grid frequency regulation mainly relies on traditional power sources such as thermal power and hydropower, but their response speed is slow and flexibility is insufficient. Traditional frequency regulation methods are inefficient in dealing with new energy fluctuations, and the standby power sources (such as diesel generators) relying on fossil energy have high costs and large pollution. The "time asymmetry" of photovoltaic power generation (the mismatch between the peak power generation during the day and the electricity demand at night) exacerbates the difficulty of frequency regulation.

[0004] The current mainstream energy storage technologies (such as lithium-ion batteries) are mainly used to suppress short-term power fluctuations (from seconds to hours), but they cannot cope with energy gaps across seasons or lasting for multiple days. Although the microgrid coordination controller supports peak shaving and valley filling, it still mainly relies on electrochemical energy storage and lacks long-term energy storage solutions. The complementary mode of "pumped storage + new energy storage" is restricted by geographical conditions and is difficult to be widely promoted. Existing technologies mostly focus on a single type of energy storage, and many do not consider the collaborative optimization of multiple types of energy storage or do not clarify the hierarchical scheduling strategy of energy storage media.

[0005] Traditional centralized control needs to frequently adjust the main control node when the number of energy storage modules changes, resulting in an increase in system complexity and response delay. Although existing technologies have introduced meteorological prediction, they have not deeply bound it with energy storage actions, lacking real-time dynamic adjustment capabilities and not conducting preventive scheduling in combination with long-term meteorological prediction. Summary of the Invention

[0006] Aiming at the above existing technical deficiencies, the technical problem to be solved by the present invention is to provide an energy storage coordination system and a control method based on meteorological prediction and multi-energy complementarity, aiming to solve problems such as poor grid stability, low response efficiency of the energy storage system, and insufficient energy utilization rate caused by the volatility of renewable energy.

[0007] To solve the above technical problems, the present invention adopts the following technical solutions: The present invention provides an energy storage coordination system based on meteorological prediction and multi-energy complementarity, including: A power generation end module, including a photovoltaic power generation unit and a wind power generation unit, and the photovoltaic power generation unit and the wind power generation unit are dynamically adjusted according to the power generation end power regulation instruction; Meteorological prediction and data acquisition module, which obtains solar irradiance, cloud movement trajectory, wind speed and air density data in the future time period in real time and sends them to the intelligent control and scheduling platform; Energy storage module, including lithium battery energy storage unit and hydrogen energy storage unit, the lithium battery energy storage unit and hydrogen energy storage unit send data including lithium battery SOC and hydrogen storage tank pressure to the intelligent control and scheduling platform and charge and discharge according to the energy storage end charge and discharge priority strategy; Intelligent control and scheduling platform, integrating edge computing nodes and deep reinforcement learning scheduling models, receiving meteorological prediction and data acquisition module, energy storage module and grid electricity price data in real time, dynamically generating power regulation instructions for the power generation end and energy storage end charge and discharge priority strategies. At the same time, it improves the income through electricity price peak-valley arbitrage and green electricity trading premium and maximizes the proportion of hydrogen energy use, reducing the dependence on diesel backup power supplies.

[0008] Furthermore, a photovoltaic power prediction model and a wind power prediction model are used to estimate the output power of the photovoltaic power generation unit and the wind power generation unit at a specific future time point; Among them, the specific formula of the photovoltaic power prediction model is as follows: P pv (t)= η pv × A pv × G (t)× 1 - β ×( T amb (t)- T ref )]× cosθ (t) Among them, P pv (t) is the output power of the photovoltaic panel at time t, η pv is the photoelectric conversion efficiency of the photovoltaic panel, A pv is the total area of the photovoltaic panel, G (t) is the solar irradiance at time t, T amb (t) is the ambient temperature, T ref is the reference temperature, β is the temperature power attenuation coefficient, θ (t) is the angle between the solar incident angle and the normal of the photovoltaic panel, which is dynamically adjusted by a two-axis tracking bracket to make cosθ (t) approach 1; The specific formula of the wind power prediction model is as follows: ; Among them, P wind P(t) is the output power of the wind turbine at time t, v ( t ) is the real-time wind speed at the hub height, ρ is the air density, C p is the wind energy utilization coefficient, A rotor is the swept area of the wind turbine impeller, v cut-in is the cut-in wind speed, v rated is the rated wind speed, v cut-out is the cut-out wind speed, P rated is the rated power of the wind turbine.

[0009] Furthermore, based on the photovoltaic power prediction model and the wind power prediction model, a power balance equation real-time calculation system is constructed to calculate the power generation, electricity demand, and charge and discharge states of the lithium battery energy storage unit and the hydrogen energy storage unit in the system, and at the same time, the charge and discharge depth is optimized through the AI algorithm; Among them, the specific formula of the power balance equation is as follows: ; Among them, P grid P(t) is the interaction power between the system and the main power grid, P bat P(t) is the charge and discharge power of the lithium battery, η bat is the comprehensive charge and discharge efficiency of the lithium battery, P H2 P(t) is the power of the hydrogen energy system, is the power generation efficiency of the fuel cell, P load P(t) is the local load demand power.

[0010] Furthermore, based on the photovoltaic power prediction model, the wind power prediction model, and the calculation results of the power balance equation, a dynamic allocation model for the priority of energy storage charge and discharge is constructed to calculate the corresponding thresholds and determine the priority of the lithium battery and hydrogen energy; Among them, the specific formula of the dynamic allocation model for the priority of energy storage charge and discharge is as follows: ; Among them, is the change rate of the grid interaction power, P rated is the rated power of the system, E daily-demandLet \(P_{total}\) be the total daily electricity demand. The short - term fluctuations include fluctuations from the second - level to the hour - level and the low - price period of the power grid at night. The long - term gap includes continuous rainy days, windless weather or seasonal power generation troughs.

[0011] Furthermore, calculating the corresponding thresholds according to the dynamic allocation model of energy storage charging and discharging priorities and judging the priorities of lithium batteries and hydrogen energy include: (1) When the detected power grid frequency deviation \(\geq0.1Hz\), the lithium battery is preferentially started to discharge within 50ms to make up the gap; (2) When the power generation amount is lower than the demand value for a continuously set time, the output of the lithium battery and hydrogen energy is allocated according to a set ratio; (3) When the predicted power generation amount in the future set time is less than 50%, hydrogen energy power supply is started and 20% of the capacity of the lithium battery is reserved as a backup; Among them, starting the hydrogen energy power supply includes: Hydrogen production stage: When the wind - solar power generation exceeds the power grid demand, the abandoned electricity is used to drive the electrolyzer to produce hydrogen, which is stored in the high - pressure hydrogen storage tank; Power generation stage: When it is predicted that the power generation amount in the future set time is insufficient, the fuel cell is started to convert hydrogen energy into electric energy and is incorporated into the power grid through an inverter; Thermal management coordination: The waste heat of the electrolyzer and fuel cell is recovered to the phase - change material heat dissipation system for winter heating or driving an absorption chiller; (4) When the time - of - use electricity price difference \(\geq0.5\) yuan / kWh, the lithium battery charges at the low - price period and discharges at the high - price period.

[0012] Furthermore, the operation constraints of the hydrogen energy system are as follows: 1) Hydrogen production stage: ; 2) Power generation stage: ; Among them, P H2,electrolysis \(P_{electrolyzer}(t)\) is the hydrogen production power of the electrolyzer, P bat,max-charge \(P_{lithium - battery - max - charge}\) is the maximum charging power of the lithium battery, P H2,fc \(P_{fuel - cell}(t)\) is the power generation power of the fuel cell, P H2,rated \(P_{fuel - cell - rated}\) is the rated power of the fuel cell, E H2,stored \(E_{hydrogen - tank}\) is the total energy of hydrogen in the hydrogen storage tank, \(\Delta t\) is the time interval.

[0013] Furthermore, the deep reinforcement learning scheduling model includes: (a) State - space design: ; Among them, G (t) is the real-time light intensity, v (t) is the real-time wind speed, SOC bat (t) is the state of charge of the lithium battery, V H2 (t) is the hydrogen gas volume in the hydrogen storage tank, λ sell (t) is the real-time selling electricity price, P load (t) is the local load power demand; (b) The action space is designed as follows: ; Among them, P bat,charge (t) is the lithium battery charging power, P bat,discharge (t) is the lithium battery discharging power, P H2,electrolysis (t) is the electrolyzer hydrogen production power, θ pv (t) is the pitch angle of the photovoltaic panel, β wind (t) is the blade pitch angle of the wind turbine; (c) The reward function is designed as follows: ; Among them, λ sell is the selling electricity price, P sell is the power sold to the power grid or users, C bat is the unit charge-discharge cost of the lithium battery, is the absolute value of the lithium battery charge-discharge power, f ( soc ) is the SOC-related loss coefficient, H2 ×P H2 is the hydrogen energy operation and maintenance cost, C H2 is the unit power operation and maintenance cost of the hydrogen energy system, is the electrolyzer hydrogen production power and fuel cell power generation power, is the environmental protection reward, β is the environmental protection reward coefficient, is the CO2 emission reduction amount, is the power grid fluctuation penalty, γ is the fluctuation penalty coefficient, is the power grid interaction power fluctuation amount.

[0014] Further, the objective function of the deep reinforcement learning scheduling model is as follows: ; The constraints are as follows: ; Among them, λ sell (t) is the real-time electricity selling price, λ buy (t) is the real-time electricity purchasing price, C bat-loss is the loss cost of the lithium battery, C H2-op is the operation and maintenance cost of the hydrogen energy system, SOC bat (t) is the state of charge of the lithium battery, V H2 (t) is the volume of hydrogen gas in the hydrogen storage tank, V H2,min is the lower limit of the volume of the hydrogen storage tank, V H2,max is the upper limit of the volume of the hydrogen storage tank, P bat (t) is the charge and discharge power of the lithium battery, P bat,rated is the rated power of the lithium battery, min is per minute; The dynamic weight coefficient is calculated as follows: ; Among them, α (t) is the weight coefficient of wind and solar power generation, G max is the maximum solar irradiance, k is the wind power regulation factor; Furthermore, the energy storage power distribution function is constructed according to the calculated dynamic weight coefficient as follows: ; Among them, P pV,excess (t) is the surplus power of photovoltaic power generation, P wind,excess (t) is the surplus power of wind power generation, P storage (t) is the total amount of surplus power to be stored; When α (t) → 1, the surplus photovoltaic electric energy is preferentially stored; When α (t) → 0, the surplus wind power electric energy is preferentially stored.

[0015] An energy storage control method based on meteorological prediction and multi-energy complementarity includes: Obtain meteorological data, and obtain meteorological data such as solar irradiance, cloud movement trajectory, wind speed, and air density in real time for a future time period; Monitor the energy storage status, and monitor data such as the SOC of lithium batteries and the pressure of hydrogen storage tanks; Intelligent scheduling and optimization, integrating edge computing nodes and a deep reinforcement learning scheduling model, receiving meteorological data, lithium battery SOC, hydrogen storage tank pressure data, and grid electricity price data in real time, dynamically generating power adjustment instructions for the power generation end to dynamically adjust the angles of photovoltaic panels and wind turbine blade pitches; and generating charge and discharge priority strategies for the energy storage end to dynamically allocate the priorities of lithium batteries and hydrogen energy; at the same time, increase revenue through peak-valley electricity price arbitrage and green electricity trading premiums and maximize the proportion of hydrogen energy use, reducing the dependence on diesel backup power supplies.

[0016] The beneficial effects of the present invention are as follows: 1. For the first time, meteorological prediction, multiple energy storage media, AI scheduling, and blockchain transactions are deeply integrated, breaking through the limitations of existing technologies that only focus on a single link.

[0017] 2. Through edge computing and hierarchical energy storage design, the system can suppress more than 80% of power fluctuations, significantly improving the stability of the power grid; through meteorological prediction and equipment linkage, the comprehensive utilization rate of wind and light is increased to 92% (about 75% for traditional systems).

[0018] 3. The cycle life of lithium batteries is extended by 30%, and the annual attenuation rate of the hydrogen energy system is controlled within 1.5%.

[0019] 4. Under the time-of-use electricity price mechanism, the investment payback period of the system is shortened from more than 10 years for a conventional energy storage system to 6 - 8 years. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 It is a topology diagram of an energy storage coordination system based on meteorological prediction and multi-energy complementarity provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. Embodiment

[0023] As Figure 1 shown, this embodiment provides an energy storage coordination system based on meteorological prediction and multi-energy complementarity, including: The power generation end module includes a photovoltaic power generation unit and a wind power generation unit. The photovoltaic power generation unit includes a two-axis tracking bracket, a photovoltaic array, and an inverter; the wind power generation unit includes a pitch angle controller, a wind turbine, and a frequency converter; the photovoltaic power generation unit and the wind power generation unit are dynamically adjusted according to the power generation end power adjustment instruction. The meteorological prediction and data acquisition module includes a satellite remote sensing receiver, a ground meteorological station, and an unmanned aerial vehicle inspection terminal; it obtains the solar irradiance, cloud movement trajectory, wind speed, and air density data in the future time period in real time and sends them to the intelligent control and dispatching platform. It should be noted that: the system obtains the sunshine intensity and cloud movement trajectory data in the future 1 hour to 72 hours through the meteorological prediction module, dynamically adjusts the pitch angle (0° to 90°) and horizontal angle (0° to 360°) of the two-axis tracking bracket, so that the photovoltaic panel is always perpendicular to the solar incident angle, and improves the power generation efficiency by 15% to 25%; when cloud occlusion is predicted, the energy storage system is notified 5 minutes in advance to enter the standby response state to prepare for suppressing power fluctuations.

[0024] The energy storage end module includes a lithium battery energy storage unit and a hydrogen energy storage unit. The lithium battery energy storage unit includes a lithium iron phosphate battery pack, a BMS management system, a PCM heat dissipation layer, and a water cooling circulation pipeline; the hydrogen energy storage unit includes an electrolyzer, a hydrogen storage tank, a fuel cell, and a heat recovery system; the lithium battery energy storage unit and the hydrogen energy storage unit send data including the lithium battery SOC and the hydrogen storage tank pressure to the intelligent control and dispatching platform and perform charging and discharging according to the energy storage end charging and discharging priority strategy. The intelligent control and dispatching platform includes a local server, a blockchain trading interface, and a human-computer interaction interface; it integrates an edge computing node and a deep reinforcement learning scheduling model, receives the meteorological prediction and data acquisition module, the energy storage end module, and the grid electricity price data in real time, dynamically generates the power generation end power adjustment instruction and the energy storage end charging and discharging priority strategy. At the same time, it improves the revenue and maximizes the hydrogen energy usage ratio through electricity price peak-valley arbitrage and green electricity trading premium, and reduces the dependence on diesel backup power. It should be noted that: The cooperation between the AI algorithm and edge computing includes: (1) Data input and processing: (a) Input parameters: Real-time meteorological data (light, wind speed, temperature and humidity), SOC of energy storage system (lithium battery), pressure of hydrogen storage tank (hydrogen energy), grid electricity price, user load demand.

[0025] (b) Edge computing node: Data processing and preliminary decision-making are completed locally (delay ≤ 10 ms), and only key data is uploaded to the cloud, reducing the communication bandwidth requirement by 50%; The key data mainly includes: ① Long-term meteorological trend data, such as: regional meteorological prediction over 72 hours in the future, update patch of historical meteorological data; ② High-precision meteorological anomaly events, such as: prediction of sudden extreme weather, local micro-meteorological data; ③ Real-time performance indicators of equipment, such as: recording of the angle adjustment of the photovoltaic double-axis bracket, dynamic change data of the wind turbine pitch angle, lithium battery status, hydrogen production / generation efficiency of the hydrogen energy system; ④ Fault and warning information, such as: equipment fault code, complex fault magic stone that cannot be processed locally by other edge nodes; ⑤ Grid interaction and market transaction data: electricity price prediction for the next 24 hours, real-time interaction power with the main grid, blockchain green electricity transaction hash value; ⑥ User load pattern: industrial and commercial user load curve, predicted load deviation value; ⑦ Environmental protection and energy efficiency indicators; ⑧ Algorithm model update data: parameter increment of the local AI model at the edge node, new scenario data not covered by the local model.

[0026] Preferably, a photovoltaic power prediction model and a wind power prediction model are used to estimate the output power of the photovoltaic power generation unit and the wind power generation unit at a specific future time point.

[0027] Specifically, the specific formula of the photovoltaic power prediction model is as follows: P pv (t) =η pv ×A pv ×G (t) ×[1-β×(T amb (t) -T ref )]×cosθ (t); Among them, P pv (t) is the output power of the photovoltaic panel at time t, η pv is the photoelectric conversion efficiency of the photovoltaic panel, A pv is the total area of the photovoltaic panel, G (t) is the solar irradiance at time t, T amb (t) is the ambient temperature, T refis the reference temperature, β is the temperature power attenuation coefficient, that is, for every 1°C increase in temperature, the power decreases by a certain proportion, θ θ(t) is the angle between the solar incidence angle and the normal of the PV panel, which is dynamically adjusted by a two-axis tracking bracket to make cosθ(t) approach 1.

[0028] Based on the wind speed prediction including the average wind speed and turbulence intensity, the pitch angle of the wind turbine is optimized. The specific adjustment range is -5° to 30°. When the wind speed is low, the pitch angle is increased to capture more wind energy. When the wind speed is high, the pitch angle is decreased to avoid overloading. When a sudden short-term wind speed drop is predicted, such as at the end of a gust, the lithium battery energy storage is activated for instantaneous power compensation, and the response time ≤ 50ms to prevent grid frequency fluctuations.

[0029] The wind-solar hybrid strategy is as follows: (1) Sunny day + low wind speed scenario with average wind speed < 5.5m / s and wind power density < 200W / m2: PV is the main power source, wind power is for auxiliary power supply, and the surplus electric energy is stored in the lithium battery or used for hydrogen production.

[0030] (2) Cloudy day + high wind speed scenario with average wind speed > 6.5m / s and wind power density > 300W / m2: Wind power is the main power source, PV supplements power supply, and the insufficient part is released by the energy storage system.

[0031] (3) Extreme weather such as sandstorms: Both wind and solar power generation are limited. The system automatically switches to hydrogen energy reserve power supply and activates the standby diesel generator as the final guarantee.

[0032] Specifically, the formula of the wind power prediction model is as follows: ;

[0033] Among them, P wind P(t) is the output power of the wind turbine at time t, v v(t) is the real-time wind speed at the hub height, ρ is the air density, C p Cp is the wind energy utilization coefficient, A rotor S is the swept area of the wind turbine impeller, which is πr 2 2, r is the blade radius, v cut-in vcut-in is the cut-in wind speed, that is, the lowest wind speed at which the wind turbine starts to generate electricity, v rated vrated is the rated wind speed, that is, the wind speed at which the wind turbine reaches its maximum output power, v cut-out vcut-out is the cut-out wind speed, that is, the highest wind speed at which the wind turbine shuts down for protection, P rated Prated is the rated power of the wind turbine, that is, the maximum continuous output power.

[0034] Based on the photovoltaic power prediction model and the wind power prediction model, a power balance equation is constructed to calculate the generated power, power consumption demand, and the charging and discharging states of the lithium battery energy storage unit and the hydrogen energy storage unit in the real-time calculation system. At the same time, the charging and discharging depth is optimized through the AI algorithm, and the battery cycle life is increased to more than 8000 times; Among them, the specific formula of the power balance equation is as follows: ;

[0035] Among them, P grid (t) is the interactive power between the system and the main power grid. A positive value indicates power supply to the grid, and a negative value indicates power purchase from the grid. P bat (t) is the charging and discharging power of the lithium battery. A positive value indicates discharging, and a negative value indicates charging. η bat is the comprehensive charging and discharging efficiency of the lithium battery, about 85% - 90%. P H2 (t) is the power of the hydrogen energy system. A positive value indicates power generation by the fuel cell, and a negative value indicates hydrogen production by the electrolyzer. is the power generation efficiency of the fuel cell, about 50% - 60%. P load (t) is the local load demand power.

[0036] Based on the photovoltaic power prediction model, the wind power prediction model, and the calculation results of the power balance equation, a dynamic allocation model for the charging and discharging priority of energy storage is constructed to calculate the corresponding thresholds and judge the priorities of lithium batteries and hydrogen energy; Among them, the specific formula of the dynamic allocation model for the charging and discharging priority of energy storage is as follows: ; Among them, is the change rate of the grid interactive power, reflecting the severity of short-term power fluctuations. P rated is the rated power of the system, used to normalize the fluctuation threshold. E daily-demand is the total daily power consumption demand, used to judge the triggering conditions for long-term energy storage. Short-term fluctuations include fluctuations from seconds to hours and the low grid electricity price period at night. Long-term gaps include continuous rainy, windless weather, or seasonal power generation troughs.

[0037] Calculating the corresponding thresholds according to the dynamic allocation model for the charging and discharging priority of energy storage and judging the priorities of lithium batteries and hydrogen energy includes: (1) When the detected grid frequency deviation ≥ 0.1Hz, at this time, the overall supply and demand of the grid is unbalanced, and the lithium battery is preferentially started to discharge to make up the gap within 50ms; (2) When the power generation duration is less than 1 hour for the continuously set time, the lithium battery and hydrogen energy output are distributed according to the set ratio; (3) When the predicted power generation in the next 72 hours is less than 50%, start the hydrogen energy power supply and reserve 20% of the lithium battery capacity as backup; Among them, starting the hydrogen energy power supply includes: Hydrogen production stage: When the wind-solar power generation exceeds the grid demand, the abandoned power is used to drive the electrolyzer to produce hydrogen, which is stored in the high-pressure hydrogen storage tank; Power generation stage: When it is predicted that the power generation in the next 3 days is insufficient, start the fuel cell to convert hydrogen energy into electric energy and incorporate it into the grid through an inverter; Thermal management coordination: The waste heat of the electrolyzer and fuel cell is recovered to the phase change material PCM heat dissipation system for winter heating or driving an absorption chiller, and the comprehensive energy efficiency is increased by 20%.

[0038] (4) When the time-of-use electricity price difference ≥ 0.5 yuan / kWh, the lithium battery charges at low valleys and discharges at peaks.

[0039] The operation constraints of the hydrogen energy system are as follows: 1) Hydrogen production stage ; 2) Power generation stage ; Among them, P H2,electrolysis \(P_{h}(t)\) is the hydrogen production power of the electrolyzer, which is only started when the wind-solar power generation is surplus and the lithium battery cannot absorb it, P bat,max-charge \(P_{lmax}\) is the maximum charging power of the lithium battery, P H2,fc \(P_{fc}(t)\) is the power generation power of the fuel cell, P H2,rated \(P_{fc}\) is the rated power of the fuel cell, E H2,stored \(E_{h}\) is the total energy of hydrogen in the hydrogen storage tank calculated by calorific value, Time interval.

[0040] The deep reinforcement learning scheduling model includes: (a) State space design: ; Among them, G \(I(t)\) is the real-time light intensity, v \(v(t)\) is the real-time wind speed, SOC bat \(SOC(t)\) is the state of charge of the lithium battery, V H2 \(V_{h}(t)\) is the hydrogen volume in the hydrogen storage tank, λ sell(t) is the current electricity selling price, P load (t) is the local load power demand; (b) The action space is designed as follows: ; Among them, P bat,charge (t) is the lithium battery charging power, P bat,discharge (t) is the lithium battery discharging power, P H2,electrolysis (t) is the electrolyzer hydrogen production power, θ pv (t) is the pitch angle of the photovoltaic panel, β wind (t) is the blade pitch angle of the wind turbine; (c) The reward function is designed as follows: ; Among them, is the electricity selling revenue, λ sell is the electricity selling price, P sell is the power sold to the power grid or users, is the lithium battery loss cost, C bat is the unit charge-discharge cost of the lithium battery, is the absolute value of the lithium battery charge-discharge power, f ( soc ) is the SOC-related loss coefficient. If SOC = 20%, f(SOC) = 1 + 0.5×∣20−50∣ = 2.5; is the hydrogen energy operation and maintenance cost, is the unit power operation and maintenance cost of the hydrogen energy system, is the electrolyzer hydrogen production power and fuel cell power generation power, is the environmental protection reward, is the environmental protection reward coefficient, linked to the carbon trading market price (China's carbon price in 2025 is about 100 yuan / ton → β = 0.1 yuan / kg), is the CO2 emission reduction amount, is the power grid fluctuation penalty, is the fluctuation penalty coefficient, set according to the power grid frequency modulation cost, usually 0.05~0.2 yuan / kW; is the power grid interaction power fluctuation amount; The objective function is as follows: ; The constraint conditions are as follows: ; Among them, λ sell (t) is the real-time electricity price under the time-of-use electricity price mechanism, that is, the real-time price, λ buy (t) is the electricity purchase price, C bat-loss is the loss cost of the lithium battery, C H2-op is the operation and maintenance cost of the hydrogen energy system, including the maintenance costs of the electrolyzer, hydrogen storage tank and fuel cell, SOC bat (t) is the state of charge of the lithium battery, that is, the proportion of the remaining power, V H2 (t) is the volume or mass of hydrogen gas in the hydrogen storage tank, V H2,min is the lower limit of the volume of the hydrogen storage tank, V H2,max is the upper limit of the volume of the hydrogen storage tank, P bat (t) is the charge and discharge power of the lithium battery, P bat,rated is the rated power of the lithium battery, min is per minute; The dynamic weight coefficient is calculated as follows: ; Among them, is the weight coefficient of wind and solar power generation, is the maximum solar irradiance, is the wind power regulation factor used to balance the contribution weights of wind and solar power from 0.6 to 0.8; The energy storage power distribution function is constructed as follows: ; Among them, P pV,excess (t) is the surplus power of photovoltaic power generation, P wind,excess (t) is the surplus power of wind power generation, P storage (t) is the total amount of surplus power to be stored; When α (t) → 1, give priority to storing the surplus photovoltaic power; When α (t) → 0, give priority to storing the surplus wind power.

[0041] This embodiment also provides a control method for the energy storage coordination system based on meteorological prediction and multi-energy complementarity, including: Obtain meteorological data, and obtain in real time the meteorological data of solar irradiance, cloud movement trajectory, wind speed and air density in the future time period; Monitor the energy storage status, and monitor the data of the state of charge (SOC) of the lithium battery and the pressure of the hydrogen storage tank. Intelligent scheduling and optimization: Integrate edge computing nodes and a deep reinforcement learning scheduling model, receive real-time meteorological data, data on the SOC of lithium batteries, the pressure of hydrogen storage tanks, and grid electricity price data, dynamically generate power adjustment instructions for the power generation side to dynamically adjust the angles of photovoltaic panels and wind turbine blade pitch angles; and generate charge and discharge priority strategies for the energy storage side to dynamically allocate the priorities of lithium batteries and hydrogen energy; at the same time, improve revenue through peak-valley electricity price arbitrage and green electricity trading premiums and maximize the proportion of hydrogen energy use, reducing the dependence on diesel backup power supplies. Embodiment

[0042] This embodiment is the second embodiment of the present invention. The difference between this embodiment and the first embodiment is that it provides a collaborative case in a typical scenario of a meteorological prediction and multi-energy complementary energy storage coordination system to verify and illustrate the technical effects adopted in this system.

[0043] Scenario 1: Peak photovoltaic power generation at noon in summer: (1) Meteorological prediction: The system obtains through the meteorological prediction module that there will be continuous strong sunlight and extremely low wind speed within the next 2 hours; in this case, photovoltaic power generation will become the main power source, and almost no electric energy can be generated by wind power.

[0044] (2) Actions at the power generation side: The dual-axis tracking bracket dynamically adjusts the angle of the photovoltaic panel according to the meteorological prediction data, making it always perpendicular to the solar incident angle, thereby maximizing the absorption of solar energy. Specifically, the pitch angle of the photovoltaic panel is adjusted to be close to 90°, and the horizontal angle is adjusted to the optimal angle, increasing the photovoltaic power generation by about 20%.

[0045] Since the wind speed is low, the wind turbine cannot generate electricity effectively. To reduce the idling loss and protect the equipment, the system automatically adjusts the blade pitch angle to the minimum value, making the blades in the most energy-efficient state.

[0046] (3) Actions at the energy storage side: During the peak period of photovoltaic power generation, a large amount of surplus electric energy is stored in the lithium battery. Assuming that the initial state of charge (SOC) of the lithium battery is 40%, after 2 hours of charging, the SOC increases from 40% to 90%, increasing the power by 50%.

[0047] When the SOC of the lithium battery approaches the upper limit, the excess electric energy will be used to drive the electrolyzer to produce hydrogen. This process not only improves the energy utilization rate but also provides a long-term energy storage solution. The efficiency of the electrolyzer is usually above 75%, and the generated hydrogen is stored in a high-pressure hydrogen storage tank.

[0048] (4) Dispatching decision: Based on the prediction of the low electricity price period, the AI dispatching platform decided to retain 30% of the lithium battery capacity so that it can be discharged for arbitrage during the peak electricity price period. This can not only improve economic benefits but also avoid the impact of overcharging on the life of lithium batteries.

[0049] Scenario 2: Continuous rainy weather in winter: (1) Weather forecast: Weather forecast shows that it will be cloudy and rainy in the next five days, and the sunshine intensity will be greatly reduced. The photovoltaic power generation will drop to about 10% of the rated value. At the same time, the wind speed will be lower than 3 m / s, and the wind turbine can only provide a small amount of electricity, about 5% of the rated value.

[0050] (2) At the power generation end, due to the significant decrease in solar irradiance caused by rainy weather, the photovoltaic power generation capacity is greatly limited, and the output power is only 10% of the rated value. At this time, photovoltaic power generation is mainly used to meet basic load needs; when the wind speed is lower than 3m / s, the wind turbine cannot effectively capture wind energy, and the output power can only reach 5% of the rated value. In this case, wind power is also difficult to provide sufficient power support.

[0051] (3) Energy storage action: Based on weather forecast results, the system starts the hydrogen energy system 48 hours in advance, converts the stored hydrogen into electricity through fuel cells, and connects it to the grid through inverters. This ensures that a stable power supply can be maintained in the event of insufficient photovoltaic and wind power generation. The lithium battery is used as a frequency regulation backup, that is, it responds quickly to deviations in grid frequency and provides short-term power support. At the same time, the system reserves part of the lithium battery capacity as an emergency backup power source in case of emergency.

[0052] (4) Dispatching decision: In case of continuous rainy weather, the system triggers the blockchain green electricity trading agreement through smart contracts to supply hydrogen-powered electricity to surrounding communities. As hydrogen-powered electricity has a high environmental value, its selling price is 10% higher than the conventional grid electricity price, which improves the economic benefits of the system. Through thermal management collaboration, the waste heat generated by the electrolyzer and fuel cell is recovered to the phase change material heat dissipation system for winter heating or driving absorption chillers, further improving the overall energy efficiency of the entire system.

[0053] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A storage coordination system based on meteorological prediction and multi - energy complementarity, characterized in that, Including: A power generation end module, including a photovoltaic power generation unit and a wind power generation unit, and the photovoltaic power generation unit and the wind power generation unit are dynamically adjusted according to a power generation end power adjustment instruction; A meteorological prediction and data acquisition module, which obtains solar irradiance, cloud movement trajectory, wind speed, and air density data in a future time period in real time and sends them to the intelligent control and dispatching platform; A energy storage end module, including a lithium battery energy storage unit and a hydrogen energy storage unit, and the lithium battery energy storage unit and the hydrogen energy storage unit send data including the state of charge (SOC) of the lithium battery and the pressure of the hydrogen storage tank to the intelligent control and dispatching platform and perform charging and discharging according to the energy storage end charge and discharge priority strategy; An intelligent control and dispatching platform, integrating an edge computing node and a deep reinforcement learning scheduling model, receiving meteorological prediction and data acquisition module, energy storage end module, and grid electricity price data in real time, dynamically generating a power generation end power adjustment instruction and an energy storage end charge and discharge priority strategy. At the same time, it improves the revenue through electricity price peak-valley arbitrage and green power trading premium and maximizes the hydrogen energy usage ratio, reducing the dependence on diesel backup power supplies.

2. The energy storage coordination system based on meteorological prediction and multi-energy complementarity according to claim 1, wherein, A photovoltaic power prediction model and a wind power prediction model are used to estimate the output power of the photovoltaic power generation unit and the wind power generation unit at a specific future time point; Among them, the specific formula of the photovoltaic power prediction model is as follows: P pv (t)= η pv × A pv × G (t)× 1 - β ×( T amb (t)- T ref )]× cosθ (t) Among them, P pv P(t) is the output power of the photovoltaic panel at time t, η pv η is the photoelectric conversion efficiency of the photovoltaic panel, A pv S is the total area of the photovoltaic panel, G G(t) is the solar irradiance at time t, T amb T(t) is the ambient temperature, T ref T0 is the reference temperature, β K is the temperature power attenuation coefficient, θ θ(t) is the angle between the solar incidence angle and the normal of the photovoltaic panel, which is dynamically adjusted by a two-axis tracking bracket to make cosθ θ(t) approach 1; The specific formula of the wind power prediction model is as follows: ； Among them, P wind P(t) is the output power of the wind turbine at time t, v ( t ) is the real-time wind speed at hub height, ρ is the air density, C p Cp is the wind energy utilization coefficient, A rotor A is the swept area of the wind turbine impeller, v cut-in Vci is the cut-in wind speed, v rated Vr is the rated wind speed, v cut-out Vco is the cut-out wind speed, P rated Pr is the rated power of the wind turbine.

3. The energy storage coordination system based on meteorological prediction and multi - energy complementarity according to claim 2, characterized in that, Based on the photovoltaic power prediction model and the wind power prediction model, a power balance equation is constructed to calculate the power generation, electricity demand, and charge and discharge states of the lithium battery energy storage unit and the hydrogen energy storage unit in the system in real time. At the same time, the charge and discharge depth is optimized through an AI algorithm; Among them, the specific formula of the power balance equation is as follows: ； Among them, P grid (t) is the interactive power between the system and the main power grid, P bat (t) is the charging and discharging power of the lithium battery, η bat is the comprehensive charging and discharging efficiency of the lithium battery, P H2 (t) is the power of the hydrogen energy system, is the power generation efficiency of the fuel cell, P load (t) is the local load demand power.

4. The energy storage coordination system based on meteorological prediction and multi-energy complementarity according to claim 3, wherein Based on the calculation results of the photovoltaic power prediction model, the wind power prediction model, and the power balance equation, an energy storage charge and discharge priority dynamic allocation model is constructed to calculate the corresponding thresholds and judge the priorities of lithium batteries and hydrogen energy; Among them, the specific formula of the energy storage charge and discharge priority dynamic allocation model is as follows: ； Among them, is the change rate of the grid interaction power, P rated is the rated power of the system, E daily-demand is the total daily electricity demand. The short-term fluctuations include fluctuations from the second level to the hourly level and the low grid electricity price period at night. The long-term gap includes continuous rainy days, windless weather or seasonal power generation troughs.

5. The energy storage coordination system based on meteorological prediction and multi-energy complementarity according to claim 4, characterized in that, Calculating the corresponding thresholds and judging the priorities of lithium batteries and hydrogen energy according to the energy storage charge and discharge priority dynamic allocation model includes: (1) When the detected grid frequency deviation ≥ 0.1 Hz, the lithium battery is preferentially started to discharge within 50 ms to make up the gap; (2) When the power generation is lower than the demand value for a continuously set time, the output of the lithium battery and hydrogen energy is allocated according to a set ratio; (3) When the predicted power generation in a future set time is less than 50%, start hydrogen energy power supply and reserve 20% of the lithium battery capacity as a backup; Among them, starting hydrogen energy power supply includes: Hydrogen production stage: When the wind and solar power generation exceeds the grid demand, the abandoned electricity is used to drive the electrolyzer to produce hydrogen, which is stored in a high-pressure hydrogen storage tank; Power generation stage: When it is predicted that the power generation in a future set time is insufficient, start the fuel cell to convert hydrogen energy into electric energy and connect it to the grid through an inverter; Thermal management coordination: The waste heat of the electrolyzer and the fuel cell is recovered to the phase change material heat dissipation system for winter heating or driving an absorption chiller; (4) When the time-of-use electricity price difference ≥ 0.5 yuan / kWh, the lithium battery charges at low valleys and discharges at peaks.

6. The energy storage coordination system based on meteorological prediction and multi-energy complementarity according to claim 5, characterized in that, The operation constraints of the hydrogen energy system are as follows: 1) Hydrogen production stage: ； 2) Power generation stage: ； Among them, P H2,electrolysis $(t)$ is the hydrogen production power of the electrolyzer, P bat,max-charge is the maximum charging power of the lithium battery, P H2,fc $(t)$ is the power generation power of the fuel cell, P H2,rated is the rated power of the fuel cell, E H2,stored is the total energy of hydrogen in the hydrogen storage tank, Time interval.

7. The energy storage coordination system based on meteorological prediction and multi-energy complementarity according to claim 1, characterized in that The deep reinforcement learning scheduling model includes: (a) State space design: ； Among them, G (t) is the real-time light intensity, v (t) is the real-time wind speed, SOC bat (t) is the state of charge of the lithium battery, V H2 (t) is the hydrogen gas volume of the hydrogen storage tank, λ sell (t) is the real-time electricity selling price, P load (t) is the local load power demand; (b) The action space is designed as follows: ； Among them, P bat,charge P(t) is the charging power of the lithium battery, P bat,discharge D(t) is the discharging power of the lithium battery, P H2,electrolysis H(t) is the hydrogen production power of the electrolyzer, θ pv θ(t) is the pitch angle of the photovoltaic panel, β wind β(t) is the pitch angle of the wind turbine; (c) The reward function is designed as follows: ； Among them, λ sell is the selling electricity price, P sell is the power sold to the power grid or users, C bat is the unit charge-discharge cost of the lithium battery, is the absolute value of the charge-discharge power of the lithium battery, f ( soc ) is the SOC-related loss coefficient, H2 ×P H2 is the hydrogen energy operation and maintenance cost, C H2 is the unit power operation and maintenance cost of the hydrogen energy system, is the hydrogen production power of the electrolyzer and the power generation power of the fuel cell, is the environmental protection reward, β is the environmental protection reward coefficient, is the CO2 emission reduction, is the power grid fluctuation penalty, γ is the fluctuation penalty coefficient, is the power grid interaction power fluctuation amount.

8. The energy storage coordination system based on meteorological prediction and multi-energy complementarity according to claim 7, characterized in that, The objective function of the deep reinforcement learning scheduling model is as follows: ； The constraint conditions are as follows: ； Among them, λ sell (t) is the real-time electricity selling price, λ buy (t) is the real-time electricity purchasing price, C bat-loss is the loss cost of the lithium battery, C H2-op is the operation and maintenance cost of the hydrogen energy system, SOC bat (t) is the state of charge of the lithium battery, V H2 (t) is the hydrogen gas volume in the hydrogen storage tank, V H2,min is the lower limit of the hydrogen storage tank volume, V H2,max is the upper limit of the hydrogen storage tank volume, P bat (t) is the charging and discharging power of the lithium battery, P bat,rated is the rated power of the lithium battery, min is per minute; Calculate the dynamic weight coefficient as follows: ； Among them, α (t) is the weight coefficient of wind and solar power generation, G max is the maximum solar irradiance, k is the wind power regulation factor.

9. The energy storage coordination system based on meteorological prediction and multi-energy complementarity according to claim 8, wherein Construct the energy storage power distribution function according to the calculated dynamic weight coefficient as follows: ； Among them, P pV,excess (t) is the surplus power of photovoltaic power generation, P wind,excess (t) is the surplus power of wind power generation, P storage (t) is the total amount of surplus power to be stored; When α (t) → 1, preferentially store the surplus photovoltaic electric energy; When α (t) → 0, preferentially store the surplus wind power 10. Based on the control method of the energy storage coordination system based on meteorological prediction and multi - energy complementarity according to any one of claims 1 to 9, characterized in that Obtain meteorological data, and obtain the meteorological data of solar irradiance, cloud movement trajectory, wind speed and air density in the future time period in real time; Monitor the energy storage state, and monitor the data of the SOC of the lithium battery and the pressure of the hydrogen storage tank; Intelligent scheduling and optimization, integrating edge computing nodes and a deep reinforcement learning scheduling model, receiving the data of meteorological data, lithium battery SOC, hydrogen storage tank pressure and grid electricity price in real time, dynamically generating power adjustment instructions for the power generation end to dynamically adjust the angles of photovoltaic panels and wind turbine pitch angles; and generating a charge - discharge priority strategy for the energy storage end to dynamically allocate the priorities of lithium batteries and hydrogen energy; at the same time, improve the revenue through peak - valley arbitrage of electricity prices and green electricity trading premiums and maximize the proportion of hydrogen energy use, and reduce the dependence on diesel backup power supplies.

Images