A micro-grid operation operation scheduling method

Abstract

This application relates to the field of microgrid operation and control technology, and discloses a microgrid operation and scheduling method. The solution includes the following steps: S1: Multi-source data acquisition and high-precision prediction. The economic benefits of this application are improved: operating costs are reduced by 15%-20%, the renewable energy consumption rate is increased from 85% to over 98%, and revenue is increased through peak-valley electricity price differences and demand response; reliability is enhanced: the power supply deficit rate is reduced to below 0.5%, the voltage / frequency deviation is controlled within ±2% / ±0.5Hz, and the critical load guarantee rate is 100% in islanded mode; environmental protection is optimized: carbon emissions are reduced by 20%-30%, which meets the requirements of the "dual carbon" policy and reduces pollutant emissions; wide adaptability: it is compatible with microgrids of different scales and operating modes, the algorithm is robust, and it can cope with large fluctuations in wind and solar / load; engineering feasibility: it adopts mature algorithms and standardized processes, and can be upgraded and transformed based on existing EMS systems, with low deployment costs and quick results.

Description

Technical Field

[0001] This application relates to the technical field of microgrid operation control, and in particular to a microgrid operation scheduling method. Background Technology

[0002] With the large-scale integration of distributed energy (photovoltaic and wind power), microgrids, as the core carrier for the consumption of distributed energy, have become a hot research topic in the industry for their scheduling and control technology.

[0003] However, in practical applications, most existing microgrid dispatching methods have a single time scale design, mostly adopting only day-ahead dispatching mode, failing to form a "day-ahead-intraday-real-time" closed-loop architecture. This makes them unable to effectively cope with the uncertain fluctuations in wind and solar power output and load demand, resulting in significant deviations in dispatching plan execution. Furthermore, their optimization objectives are one-sided; traditional methods often focus on a single economic objective, neglecting environmental requirements and power supply reliability, which does not meet the multi-objective synergy requirements under the "dual-carbon" policy. Energy storage dispatching strategies are also crude, failing to incorporate refined control based on electricity prices, wind and solar power output, and energy storage lifetime constraints, leading to low energy storage utilization and rapid aging. Finally, their algorithm adaptability is insufficient; single optimization algorithms struggle to balance global optimization and real-time response, classic algorithms cannot handle nonlinear constraints, and intelligent algorithms have slow convergence speeds, affecting dispatching real-time performance.

[0004] For example, patent CN115360743A proposes a multi-objective optimization method for grid-connected microgrids, but it only uses a linear weighted summation method to transform the multi-objective problem, without considering dynamic adjustment of weights, and lacks a fast response mechanism at the real-time scheduling level; patent CN121308075A uses an AI-driven scheduling method, but it does not form a three-level time-scale closed-loop optimization, and energy storage scheduling does not incorporate multi-dimensional constraints.

[0005] Therefore, those skilled in the art provide a microgrid operation and scheduling method to solve the problems mentioned in the background art. Summary of the Invention

[0006] To address the shortcomings of most microgrid dispatching methods mentioned in the background, such as limited time scale design (most only employing day-ahead dispatching and failing to form a "day-ahead-intraday-real-time" closed-loop architecture), inability to effectively handle the uncertain fluctuations in wind and solar power output and load demand leading to significant deviations in dispatch plan execution; one-sided optimization objectives (traditional methods often focus on single economic goals while neglecting environmental requirements and power supply reliability, failing to meet the multi-objective coordination requirements under the "dual-carbon" policy); coarse-grained energy storage dispatching strategies (failing to incorporate fine-grained control based on electricity prices, wind and solar power output, and energy storage lifetime constraints, resulting in low energy storage utilization and rapid aging); and insufficient algorithm adaptability (single optimization algorithms struggle to balance global optimization and real-time response, classical algorithms cannot handle nonlinear constraints, and intelligent algorithms have slow convergence speeds, all affecting dispatch real-time performance), this application provides a microgrid operation and dispatching method.

[0007] This application provides a microgrid operation and scheduling method, which adopts the following technical solution: including the following steps:

[0008] S1: Multi-source data acquisition and high-precision forecasting: Collect data on photovoltaic / wind power output, load demand, energy storage SOC status, grid time-of-use electricity price, tie-line power, and equipment operating status within the microgrid; use an NWP+LSTM hybrid model to predict 0-24h wind and solar power output (error ≤15%), use a time series + deep learning model to predict load demand (error ≤10%), and combine electricity market data to predict day-ahead / real-time electricity prices;

[0009] S2: Source-Load-Storage-Grid Collaborative Modeling and Multi-Constraint Construction: Construct unit models of distributed power sources, energy storage systems, flexible loads, and main grid interactions, and set power balance constraints, equipment operation constraints, grid security constraints, and economic and environmental constraints. Among them, the SOC constraint of energy storage is 20%-80%, the minimum start-up and shutdown time and ramp-up rate constraints are set for diesel / gas turbines, and the grid meets the N-1 safety criterion.

[0010] S3: Construction of dynamic weighted multi-objective optimization function: With minimizing the overall operating cost as the economic objective, minimizing carbon emissions as the environmental objective, and minimizing the power supply deficit rate as the reliability objective, a dynamic weighted comprehensive objective function is constructed, and the weight coefficients are adaptively adjusted according to the microgrid operation scenario.

[0011] S4: Three-level time scale hierarchical optimization solution: The “day-to-day-real-time” closed-loop scheduling architecture is adopted. The day-to-day scheduling is based on the mixed integer linear programming (MILP) algorithm to output a 24-hour time-by-time optimization plan. The intraday scheduling is based on the MOPSO algorithm to continuously correct the prediction deviation every 15-60 minutes. The real-time scheduling is based on the PPO reinforcement learning algorithm to output fast adjustment instructions within 1-5 minutes.

[0012] S5: Source-Load-Storage Coordinated Dispatch Execution: Following the principle of prioritizing the consumption of renewable energy, energy storage implements a refined strategy of "valley charging and peak discharging + surplus charging of wind and solar power + insufficient discharging of wind and solar power", combining price-based and incentive-based demand response to achieve peak shaving and valley filling;

[0013] S6: Real-time monitoring and closed-loop correction: Real-time collection of microgrid operation data and comparison with planned values, calculation of power deviation and power quality deviation, triggering intraday / real-time dispatch correction mechanism to ensure stable operation of microgrid in grid-connected / islanded modes.

[0014] Preferably, the equipment operation constraints in step S2 specifically include:

[0015] Distributed power generation constraints: upper and lower limits of photovoltaic / wind power output, ramp rate limits;

[0016] Energy storage system constraints: charge and discharge power limits, SOC range (20%-80%), depth of discharge ≤70%, charge and discharge conversion efficiency ≥90%;

[0017] Controllable power supply constraints: upper and lower limits of diesel / gas turbine output, minimum start-stop time ≥10min, and ramp rate ≤0.1MW / min;

[0018] Flexible load constraints: adjustment response time ≤ 15 min, adjustment range ± 20% of rated load.

[0019] Preferably, the expression for the dynamically weighted synthesis objective function in step S3 is:

[0020] minF=α⋅Ceco+β⋅Cenv+γ⋅Crel

[0021] Where α+β+γ=1 (α, β, γ are dynamic weighting coefficients; α=0.5, β=0.3, γ=0.2 in grid-connected mode, and α=0.4, β=0.2, γ=0.4 in islanded mode); Ceco is the comprehensive operating cost, including the cost of purchasing and selling electricity from the main grid, the operation and maintenance cost of distributed power sources, the cost of energy storage charging and discharging losses, and the cost of demand response compensation; Cenv is the environmental cost, including CO2 emissions × carbon trading price + NO x / SO2 processing cost; Crel is the reliability cost, which includes the power shortage rate × unit power shortage loss cost.

[0022] Preferably, the specific implementation method of the three-level time-scale hierarchical optimization solution in step S4 is as follows:

[0023] Day-ahead dispatch (24-hour time scale): Based on the MILP algorithm, the global optimum is solved. The input is the predicted data and equipment parameters, and the output is the unit start-up and shutdown plan, energy storage charging and discharging curve, power purchase and sale plan and demand response plan.

[0024] Intraday scheduling (15-60 minute time scale): Based on real-time collected wind / solar / load deviation data, the MOPSO algorithm is used to make rolling corrections to the day-ahead plan, output adjustment instructions for each time period, and trigger flexible load adjustment;

[0025] Real-time scheduling (1-5 minute time scale): Based on the PPO reinforcement learning algorithm, an intelligent decision-making model is built. Input voltage / frequency deviation and power fluctuation data, and output energy storage fast adjustment command and inverter reactive power compensation command to smooth out power fluctuations.

[0026] Preferably, the refined energy storage scheduling strategy in step S5 specifically includes:

[0027] Electricity price guidance strategy: When the main grid electricity price is lower than the off-peak threshold (≤0.3 yuan / kWh), the energy storage will be charged at 80% of its rated power; when the electricity price is higher than the peak threshold (≥1.0 yuan / kWh), the energy storage will be discharged at 70% of its rated power.

[0028] Wind and solar power integration strategy: When wind and solar power output exceeds 10% of the load demand, energy storage will be charged first; when wind and solar power output is less than 10% of the load demand, energy storage will be discharged first to supplement the load.

[0029] Island mode strategy: The SOC lower limit is raised to 30%, and priority is given to ensuring power supply for critical loads (medical and communication loads), with charging and discharging power ≤ 50% of rated power.

[0030] Preferably, the closed-loop correction mechanism in step S6 includes:

[0031] Deviation detection: When the actual power deviates from the planned value by ≥5%, or the voltage deviation is ≥±2%, or the frequency deviation is ≥±0.5Hz, correction is triggered;

[0032] Correction logic: When the deviation is 5%-10%, it is quickly adjusted and corrected through energy storage; when the deviation is 10%-20%, the controllable power supply and flexible load are activated for coordinated correction; when the deviation is >20%, it switches to islanded operation mode (in grid-connected state) and activates emergency power supply.

[0033] In summary, this application includes the following beneficial technical effects:

[0034] 1. Improved economic efficiency: Operating costs are reduced by 15%-20%, the renewable energy consumption rate increases from 85% to over 98%, and revenue is increased through peak-valley electricity price differences and demand response;

[0035] 2. Enhanced reliability: Power shortage rate reduced to below 0.5%, voltage / frequency deviation controlled within ±2% / ±0.5Hz, and critical load guarantee rate of 100% in islanding mode;

[0036] 3. Environmental optimization: Carbon emissions are reduced by 20%-30%, meeting the requirements of the "dual carbon" policy and reducing pollutant emissions;

[0037] 4. Wide adaptability: It is compatible with microgrids of different sizes and operating modes. The algorithm is robust and can cope with large fluctuations in wind and solar power / load.

[0038] 5. Project Feasibility: Adopting mature algorithms and standardized processes, it can be upgraded and transformed based on the existing EMS system, with low deployment cost and quick results. Detailed Implementation

[0039] The described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0040] This application discloses a microgrid operation and scheduling method, including the following steps:

[0041] S1: Multi-source data acquisition and high-precision forecasting: Collect data on photovoltaic / wind power output, load demand, energy storage SOC status, grid time-of-use electricity price, tie-line power, and equipment operating status within the microgrid; use an NWP+LSTM hybrid model to predict 0-24h wind and solar power output (error ≤15%), use a time series + deep learning model to predict load demand (error ≤10%), and combine electricity market data to predict day-ahead / real-time electricity prices;

[0042] S2: Source-Load-Storage-Grid Collaborative Modeling and Multi-Constraint Construction: Construct unit models of distributed power sources, energy storage systems, flexible loads, and main grid interactions, and set power balance constraints, equipment operation constraints, grid security constraints, and economic and environmental constraints. Among them, the SOC constraint of energy storage is 20%-80%, the minimum start-up and shutdown time and ramp-up rate constraints are set for diesel / gas turbines, and the grid meets the N-1 safety criterion.

[0043] S3: Construction of dynamic weighted multi-objective optimization function: With minimizing the overall operating cost as the economic objective, minimizing carbon emissions as the environmental objective, and minimizing the power supply deficit rate as the reliability objective, a dynamic weighted comprehensive objective function is constructed, and the weight coefficients are adaptively adjusted according to the microgrid operation scenario.

[0044] S4: Three-level time scale hierarchical optimization solution: The “day-to-day-real-time” closed-loop scheduling architecture is adopted. The day-to-day scheduling is based on the mixed integer linear programming (MILP) algorithm to output a 24-hour time-by-time optimization plan. The intraday scheduling is based on the MOPSO algorithm to continuously correct the prediction deviation every 15-60 minutes. The real-time scheduling is based on the PPO reinforcement learning algorithm to output fast adjustment instructions within 1-5 minutes.

[0045] S5: Source-Load-Storage Coordinated Dispatch Execution: Following the principle of prioritizing the consumption of renewable energy, energy storage implements a refined strategy of "valley charging and peak discharging + surplus charging of wind and solar power + insufficient discharging of wind and solar power", combining price-based and incentive-based demand response to achieve peak shaving and valley filling;

[0046] S6: Real-time monitoring and closed-loop correction: Real-time collection of microgrid operation data and comparison with planned values, calculation of power deviation and power quality deviation, triggering intraday / real-time dispatch correction mechanism to ensure stable operation of microgrid in grid-connected / islanded modes.

[0047] In this application, the equipment operation constraints in step S2 specifically include:

[0048] Distributed power generation constraints: upper and lower limits of photovoltaic / wind power output, ramp rate limits;

[0049] Energy storage system constraints: charge and discharge power limits, SOC range (20%-80%), depth of discharge ≤70%, charge and discharge conversion efficiency ≥90%;

[0050] Controllable power supply constraints: upper and lower limits of diesel / gas turbine output, minimum start-stop time ≥10min, and ramp rate ≤0.1MW / min;

[0051] Flexible load constraints: adjustment response time ≤ 15 min, adjustment range ± 20% of rated load.

[0052] In this application, the expression for the dynamically weighted synthesis objective function in step S3 is:

[0053] minF=α⋅Ceco+β⋅Cenv+γ⋅Crel

[0054] Where α+β+γ=1 (α, β, γ are dynamic weighting coefficients; α=0.5, β=0.3, γ=0.2 in grid-connected mode, and α=0.4, β=0.2, γ=0.4 in islanded mode); Ceco is the comprehensive operating cost, including the cost of purchasing and selling electricity from the main grid, the operation and maintenance cost of distributed power sources, the cost of energy storage charging and discharging losses, and the cost of demand response compensation; Cenv is the environmental cost, including CO2 emissions × carbon trading price + NO x / SO2 processing cost; Crel is the reliability cost, which includes the power shortage rate × unit power shortage loss cost.

[0055] In this application, the specific implementation method of the three-level time-scale hierarchical optimization solution in step S4 is as follows:

[0056] Day-ahead dispatch (24-hour time scale): Based on the MILP algorithm, the global optimum is solved. The input is the predicted data and equipment parameters, and the output is the unit start-up and shutdown plan, energy storage charging and discharging curve, power purchase and sale plan and demand response plan.

[0057] Intraday scheduling (15-60 minute time scale): Based on real-time collected wind / solar / load deviation data, the MOPSO algorithm is used to make rolling corrections to the day-ahead plan, output adjustment instructions for each time period, and trigger flexible load adjustment;

[0058] Real-time scheduling (1-5 minute time scale): Based on the PPO reinforcement learning algorithm, an intelligent decision-making model is built. Input voltage / frequency deviation and power fluctuation data, and output energy storage fast adjustment command and inverter reactive power compensation command to smooth out power fluctuations.

[0059] In this application, the refined energy storage scheduling strategy in step S5 specifically includes:

[0060] Electricity price guidance strategy: When the main grid electricity price is lower than the off-peak threshold (≤0.3 yuan / kWh), the energy storage will be charged at 80% of its rated power; when the electricity price is higher than the peak threshold (≥1.0 yuan / kWh), the energy storage will be discharged at 70% of its rated power.

[0061] Wind and solar power integration strategy: When wind and solar power output exceeds 10% of the load demand, energy storage will be charged first; when wind and solar power output is less than 10% of the load demand, energy storage will be discharged first to supplement the load.

[0062] Island mode strategy: The SOC lower limit is raised to 30%, and priority is given to ensuring power supply for critical loads (medical and communication loads), with charging and discharging power ≤ 50% of rated power.

[0063] In this application, the closed-loop correction mechanism in step S6 includes:

[0064] Deviation detection: When the actual power deviates from the planned value by ≥5%, or the voltage deviation is ≥±2%, or the frequency deviation is ≥±0.5Hz, correction is triggered;

[0065] Correction logic: When the deviation is 5%-10%, it is quickly adjusted and corrected through energy storage; when the deviation is 10%-20%, the controllable power supply and flexible load are activated for coordinated correction; when the deviation is >20%, it switches to islanded operation mode (in grid-connected state) and activates emergency power supply.

[0066] All standard parts used in this invention can be purchased from the market, and irregular parts can be customized according to the description in the specification. The specific connection methods of each part adopt conventional methods such as bolts, rivets, and welding that are mature in the prior art. The machinery, parts and equipment adopt conventional models in the prior art, and the circuit connection adopts conventional connection methods in the prior art, which will not be described in detail here.

[0067] Finally, the following points should be noted: First, in the description of this application, it should be noted that, unless otherwise specified and limited, the terms "installation", "connection", and "linkage" should be interpreted broadly, and can be mechanical or electrical connections, or internal connections between two components, or direct connections. "Up", "down", "left", "right", etc. are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may change.

[0068] Secondly: In the embodiments disclosed in this invention, only the structures involved in the embodiments disclosed in this invention are involved. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

Claims

1. A microgrid operation and scheduling method, characterized in that, Includes the following steps: S1: Multi-source data acquisition and high-precision forecasting: Collect data on photovoltaic / wind power output, load demand, energy storage SOC status, grid time-of-use electricity price, tie-line power, and equipment operating status within the microgrid; use an NWP+LSTM hybrid model to predict 0-24h wind and solar power output (error ≤15%), use a time series + deep learning model to predict load demand (error ≤10%), and combine electricity market data to predict day-ahead / real-time electricity prices; S2: Source-Load-Storage-Grid Collaborative Modeling and Multi-Constraint Construction: Construct unit models of distributed power sources, energy storage systems, flexible loads, and main grid interactions, and set power balance constraints, equipment operation constraints, grid security constraints, and economic and environmental constraints. Among them, the SOC constraint of energy storage is 20%-80%, the minimum start-up and shutdown time and ramp-up rate constraints are set for diesel / gas turbines, and the grid meets the N-1 safety criterion. S3: Construction of dynamic weighted multi-objective optimization function: With minimizing the overall operating cost as the economic objective, minimizing carbon emissions as the environmental objective, and minimizing the power supply deficit rate as the reliability objective, a dynamic weighted comprehensive objective function is constructed, and the weight coefficients are adaptively adjusted according to the microgrid operation scenario. S4: Three-level time scale hierarchical optimization solution: adopts a "day-to-day-real-time" closed-loop scheduling architecture. Day-to-day scheduling is based on the mixed integer linear programming (MILP) algorithm to output a 24-hour time-by-time optimization plan. Day-to-day scheduling is based on the MOPSO algorithm to continuously correct prediction deviations every 15-60 minutes. Real-time scheduling is based on the PPO reinforcement learning algorithm to output fast adjustment instructions within 1-5 minutes. S5: Source-Load-Storage Coordinated Dispatch Execution: Following the principle of prioritizing the consumption of renewable energy, energy storage implements a refined strategy of "valley charging and peak discharging + surplus charging of wind and solar power + insufficient discharging of wind and solar power", combining price-based and incentive-based demand response to achieve peak shaving and valley filling; S6: Real-time monitoring and closed-loop correction: Real-time collection of microgrid operation data and comparison with planned values, calculation of power deviation and power quality deviation, triggering intraday / real-time dispatch correction mechanism to ensure stable operation of microgrid in grid-connected / islanded modes.

2. The method according to claim 1, characterized in that, The specific equipment operation constraints in step S2 include: Distributed power generation constraints: upper and lower limits of photovoltaic / wind power output, ramp rate limits; Energy storage system constraints: charge and discharge power limits, SOC range (20%-80%), depth of discharge ≤70%, charge and discharge conversion efficiency ≥90%; Controllable power supply constraints: upper and lower limits of diesel / gas turbine output, minimum start-stop time ≥10min, and ramp rate ≤0.1MW / min; Flexible load constraints: adjustment response time ≤ 15 min, adjustment range ± 20% of rated load.

3. The method according to claim 1, characterized in that, The expression for the dynamically weighted synthesis objective function in step S3 is: minF=α⋅Ceco+β⋅Cenv+γ⋅Crel Where α+β+γ=1 (α, β, γ are dynamic weighting coefficients; α=0.5, β=0.3, γ=0.2 in grid-connected mode, and α=0.4, β=0.2, γ=0.4 in islanded mode); Ceco is the comprehensive operating cost, including the cost of purchasing and selling electricity from the main grid, the operation and maintenance cost of distributed power sources, the cost of energy storage charging and discharging losses, and the cost of demand response compensation; Cenv is the environmental cost, including CO2 emissions × carbon trading price + NO x / SO2 processing cost; Crel is the reliability cost, which includes the power shortage rate × unit power shortage loss cost.

4. The method according to claim 1, characterized in that, The specific implementation method of the three-level time-scale hierarchical optimization solution in step S4 is as follows: Day-ahead dispatch (24-hour time scale): Based on the MILP algorithm, the global optimum is solved. The input is the predicted data and equipment parameters, and the output is the unit start-up and shutdown plan, energy storage charging and discharging curve, power purchase and sale plan and demand response plan. Intraday scheduling (15-60 minute time scale): Based on real-time collected wind / solar / load deviation data, the MOPSO algorithm is used to make rolling corrections to the day-ahead plan, output adjustment instructions for each time period, and trigger flexible load adjustment; Real-time scheduling (1-5 minute time scale): Based on the PPO reinforcement learning algorithm, an intelligent decision-making model is built. Input voltage / frequency deviation and power fluctuation data, and output energy storage fast adjustment command and inverter reactive power compensation command to smooth out power fluctuations.

5. The method according to claim 1, characterized in that, The refined energy storage scheduling strategy in step S5 specifically includes: Electricity price guidance strategy: When the main grid electricity price is lower than the off-peak threshold (≤0.3 yuan / kWh), the energy storage will be charged at 80% of its rated power; when the electricity price is higher than the peak threshold (≥1.0 yuan / kWh), the energy storage will be discharged at 70% of its rated power. Wind and solar power integration strategy: When wind and solar power output exceeds 10% of the load demand, energy storage will be charged first; when wind and solar power output is less than 10% of the load demand, energy storage will be discharged first to supplement the load. Island mode strategy: The SOC lower limit is raised to 30%, and priority is given to ensuring power supply for critical loads (medical and communication loads), with charging and discharging power ≤ 50% of rated power.

6. The method according to claim 1, characterized in that, The closed-loop correction mechanism in step S6 includes: Deviation detection: When the actual power deviates from the planned value by ≥5%, or the voltage deviation is ≥±2%, or the frequency deviation is ≥±0.5Hz, correction is triggered; Correction logic: When the deviation is 5%-10%, it is quickly adjusted and corrected through energy storage; when the deviation is 10%-20%, the controllable power supply and flexible load are activated for coordinated correction; when the deviation is >20%, it switches to islanded operation mode (in grid-connected state) and activates emergency power supply.