Smart park micro-grid intelligent scheduling and multi-energy complementary optimization control method

Abstract

The invention discloses a smart park micro-grid intelligent scheduling and multi-energy complementation optimization control method, and aims to solve the problems of poor multi-energy complementation coordination, single scheduling strategy and insufficient system stability in the prior art. According to the method, a'multi-energy complementation-intelligent scheduling-optimization control 'three-layer architecture is constructed, photovoltaic, wind power, energy storage and power grid electric energy are integrated, distributed power supply output, energy storage charging and discharging and power grid interaction power are dynamically adjusted based on real-time data (load requirements, meteorological data and electricity price signals) and a multi-objective optimization algorithm (NSGA-II), and the optimal control of the distributed power supply is realized. And collaborative optimization of economy, environmental protection and reliability is realized. Experimental results show that the method can improve the renewable energy consumption rate to more than 90%, reduce the operation cost by more than 20%, reduce the carbon emission by more than 15%, improve the power supply reliability to 99.9%, and significantly enhance the flexibility and stability of the micro-grid. The method is suitable for distributed energy systems in smart parks, industrial parks and the like, and has high practical value.

Description

Technical Field

[0001] The present invention belongs to the technical field of microgrid intelligent control and energy management, and specifically relates to a smart park microgrid optimization control method based on multi-energy complementarity and intelligent scheduling. The method is applicable to microgrid systems that include distributed power sources (photovoltaic, wind power, energy storage systems), load demand and grid interaction, and aims to improve the economy, reliability and renewable energy absorption capacity of microgrids. Background Art

[0002] With the rapid development of distributed energy, smart campus microgrids have become an important vehicle for achieving efficient energy utilization and low-carbon operation. However, existing technologies have the following shortcomings:

[0003] Poor coordination of multi-energy complementarity:

[0004] The output of renewable energy sources such as photovoltaics and wind power is significantly affected by natural conditions (such as light intensity and wind speed) and is characterized by intermittency and volatility. Traditional scheduling methods usually use fixed-proportion allocation or simple rule control, which makes it difficult to dynamically adjust the coordinated output of photovoltaics, wind power, energy storage, and grid power according to the real-time energy status. For example, when there is sufficient sunlight, excess photovoltaic power generation may lead to curtailment; and when the wind speed is insufficient, a sudden drop in wind power output may cause a power supply gap. In addition, the charging and discharging strategies of energy storage systems (such as lithium batteries) lack global optimization and are only simply controlled based on the SOC (state of charge) threshold. This cannot effectively smooth the volatility of renewable energy, resulting in low energy utilization or an imbalance between supply and demand.

[0005] Single scheduling strategy:

[0006] Existing scheduling methods are mostly guided by a single goal (such as minimizing operating costs) and lack the ability to dynamically respond to load demand, electricity price signals, and grid status. For example, the impact of time-of-use electricity prices on energy storage charging and discharging strategies is not considered, resulting in the inability to fully utilize the energy storage system to store low-priced electricity when electricity prices are low, and excessive reliance on high-priced electricity purchases during peak electricity prices, increasing operating costs. At the same time, the output of distributed power sources is not dynamically adjusted based on load forecast results, resulting in a low renewable energy absorption rate and even the phenomenon of "wind and solar power abandonment". In addition, the lack of monitoring of the real-time status of the power grid (such as frequency fluctuations and voltage exceeding the limit) makes it impossible to adjust the interactive power between the microgrid and the large power grid in a timely manner, affecting the economy and reliability of the system.

[0007] Insufficient system stability:

[0008] The interaction between microgrids and the larger grid is complex, and existing technologies do not fully consider the impact of bidirectional power flow on system stability. For example, when a microgrid has excess power, failing to restrict power sold to the grid can trigger power backflow, leading to increased grid voltage or line overloads. Conversely, when a microgrid is underpowered, over-reliance on grid power purchases can exacerbate the grid's burden. Furthermore, traditional control methods lack the ability to quickly respond to fault scenarios (such as islanding and load shedding strategies). Failures in distributed power sources or energy storage systems can easily lead to power fluctuations or voltage collapse, compromising power supply reliability.

[0009] Insufficient data-driven and intelligent optimization capabilities:

[0010] Existing technologies often rely on empirical rules or simple models (such as linear programming) for scheduling, lacking the ability to deeply mine and analyze massive amounts of data (such as weather, load, and electricity prices). For example, they fail to employ machine learning algorithms (such as LSTM neural networks) for high-precision load forecasting, resulting in significant deviations between scheduling strategies and actual demand. They also fail to utilize multi-objective optimization algorithms (such as NSGA-II) to balance economic efficiency, environmental protection, and reliability, leading to one-sided optimization results. Furthermore, real-time control systems suffer from slow response times and are unable to adapt to the rapid dynamic changes of microgrids. Summary of the Invention

[0011] The purpose of the present invention is to provide a smart park microgrid intelligent scheduling and multi-energy complementary optimization control method, which realizes efficient and stable operation of the microgrid through real-time data driving and multi-objective optimization.

[0012] To achieve the above objectives, the present invention provides the following technical solutions:

[0013] A method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid includes the following steps:

[0014] S1 multi-energy complementary layer construction:

[0015] Integrate distributed power sources and grid power within the park, and achieve multi-energy synergy and complementarity through energy conversion and storage;

[0016] S2 intelligent scheduling layer optimization:

[0017] Based on load forecasts, real-time electricity price signals, and grid status, a multi-objective optimization algorithm is used to generate the optimal scheduling plan for distributed generation output, energy storage charging and discharging, and grid interaction power.

[0018] S3 optimizes the control layer execution:

[0019] The optimization results are sent to the microgrid controller in real time to dynamically adjust the distributed power output, energy storage charging and discharging strategy, and grid interaction power to achieve real-time optimized operation of the microgrid.

[0020] Furthermore, the multi-objective optimization algorithm is based on the following objective function:

[0021] minF=ω1C op +ω2E em -ω3R rel

[0022] in:

[0023] C op The operating costs of the microgrid, including fuel costs, electricity purchase costs, and equipment maintenance costs;

[0024] E em Carbon emissions are calculated based on the type of distributed generation and the carbon factor of electricity purchased from the grid;

[0025] R rel For power supply reliability, the load power shortage rate or voltage qualification rate is evaluated;

[0026] ω1, ω2, and ω3 are weight coefficients, satisfying ω1+ω2+ω3=1.

[0027] Furthermore, the load forecast adopts a deep learning model based on an LSTM neural network. The input data includes historical load data, meteorological data and time characteristics, and outputs a load forecast curve for the next 24 hours. Among them, the meteorological data includes light intensity, wind speed, and temperature, and the time characteristics include hours and seasons.

[0028] Furthermore, the charging and discharging strategy of the energy storage system includes:

[0029] Charge during low electricity price periods and discharge during peak periods;

[0030] It charges when renewable energy output is in excess and discharges when output is insufficient;

[0031] Participate in the power balance and voltage support of the microgrid according to the grid frequency or voltage fluctuations.

[0032] Furthermore, the grid interaction power optimization strategy includes:

[0033] Dynamically adjust the power purchase and sales between the microgrid and the large power grid based on real-time electricity price signals;

[0034] Sell ​​electricity to the grid when the microgrid has excess power, and purchase electricity from the grid when the power is insufficient;

[0035] Give priority to the consumption of renewable energy and reduce dependence on the power grid.

[0036] Furthermore, the real-time optimization operation also includes a fault protection mechanism, which triggers protection actions when the following abnormal conditions are detected:

[0037] The voltage or frequency of the microgrid exceeds the allowable range;

[0038] Failure of the distributed power generation or energy storage system;

[0039] The interaction power between the grid and the microgrid exceeds the limit.

[0040] Furthermore, the protection action includes:

[0041] Start the island operation mode and disconnect from the main power grid;

[0042] Cut off non-critical loads to ensure power supply to important loads;

[0043] Adjust the output of distributed power sources or the charging and discharging power of energy storage to restore system balance.

[0044] Furthermore, the multi-energy complementary layer further comprises:

[0045] Coordinated output control of photovoltaic and wind power, dynamically adjusting the output ratio based on light intensity and wind speed forecasts;

[0046] Energy storage system state of charge monitoring and charge and discharge power limitation to avoid overcharging or over-discharging;

[0047] The power interaction limit between the microgrid and the large grid is set to prevent power backflow or overload.

[0048] Furthermore, the multi-objective optimization algorithm is one of the following algorithms or a combination thereof:

[0049] Genetic algorithm;

[0050] Particle swarm optimization;

[0051] Simulated annealing algorithm;

[0052] Non-dominated sorting genetic algorithm.

[0053] Furthermore, the smart park microgrid also includes:

[0054] Real-time monitoring and fault diagnosis of distributed power sources;

[0055] Dynamic classification of load demand, including interruptible load and non-interruptible load;

[0056] Data interaction and collaborative control with the park energy management system.

[0057] The intelligent dispatching and multi-energy complementary optimization control method of the smart park microgrid of the present invention has the following beneficial effects:

[0058] Deep integration of multi-energy complementarity and intelligent scheduling: Driven by real-time data, multi-energy synergy and dynamic optimization are achieved to improve energy utilization. Application of multi-objective optimization algorithms: Taking into account economic efficiency, environmental protection, and reliability, avoiding the limitations of traditional single-objective optimization.

[0059] Adaptive control strategy: Dynamically adjust the scheduling plan according to electricity price, load and grid status to adapt to complex operating environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a schematic diagram of the implementation process of the smart park microgrid intelligent scheduling and multi-energy complementary optimization control method in the present invention. DETAILED DESCRIPTION

[0061] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0062] Example:

[0063] This embodiment provides a method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid.

[0064] The main steps include:

[0065] S1 multi-energy complementary layer construction:

[0066] Integrate distributed power sources and grid power within the park, and achieve multi-energy synergy and complementarity through energy conversion and storage;

[0067] Multi-energy complementary layer: Integrates photovoltaic, wind power, energy storage and grid power to achieve multi-energy synergy through energy conversion and storage, smoothing out fluctuations in renewable energy output;

[0068] S2 intelligent scheduling layer optimization:

[0069] Based on load forecasts, real-time electricity price signals, and grid status, a multi-objective optimization algorithm is used to generate the optimal scheduling plan for distributed generation output, energy storage charging and discharging, and grid interaction power.

[0070] Intelligent dispatching layer: Based on load forecasts, electricity price signals, and grid status, a multi-objective optimization algorithm is used to dynamically adjust the output of distributed power sources and energy storage charging and discharging strategies;

[0071] S3 optimizes the control layer execution:

[0072] The optimization results are sent to the microgrid controller in real time to dynamically adjust the distributed power output, energy storage charging and discharging strategy, and grid interaction power to achieve real-time optimized operation of the microgrid.

[0073] Optimization control layer: Send the optimization results to the microgrid controller in real time, dynamically adjust the system operating parameters to ensure economy, environmental protection and reliability.

[0074] Specifically,

[0075] (1) Multi-energy complementary control strategy

[0076] Photovoltaic-wind power coordinated output: Based on the forecast of light intensity and wind speed, the output ratio of photovoltaic and wind power is dynamically adjusted to give priority to the absorption of renewable energy.

[0077] Energy storage system regulation: charging during periods of low electricity prices and discharging during peak periods; charging when renewable energy output is in excess and discharging when it is insufficient; participating in grid frequency and voltage regulation.

[0078] Grid interaction optimization: Based on real-time electricity price signals, the power purchase and sales between microgrids and large grids are dynamically adjusted to give priority to the consumption of renewable energy and reduce dependence on the grid.

[0079] (2) Intelligent scheduling algorithm

[0080] Load forecasting: Based on the LSTM neural network, historical load data, meteorological data (light, wind speed, temperature) and time characteristics are input to output the load forecast curve for the next 24 hours.

[0081] Multi-objective optimization: Taking operating costs, carbon emissions, and power supply reliability as optimization targets, construct the objective function:

[0082] minF=ω1C op +ω2E em -ω3R rel

[0083] in,

[0084] C op The operating costs of the microgrid, including fuel costs, electricity purchase costs, and equipment maintenance costs;

[0085] E em Carbon emissions are calculated based on the type of distributed generation and the carbon factor of electricity purchased from the grid;

[0086] R rel For power supply reliability, the load power shortage rate or voltage qualification rate is evaluated;

[0087] ω1, ω2, and ω3 are weight coefficients, satisfying ω1+ω2+ω3=1.

[0088] (3) Real-time control and fault protection

[0089] Dynamic adjustment: Based on the optimization results, the distributed power output, energy storage charging and discharging power, and grid interaction power are adjusted in real time.

[0090] Fault protection: Monitors parameters such as voltage and frequency, triggering protection mechanisms (islanding, load shedding, and output adjustment) to restore system balance. Specific embodiment:

[0092] The following describes in detail the implementation steps, data collection and processing, algorithm parameter setting, and optimization result verification of the method of the present invention, combined with an actual case of a smart park microgrid.

[0093] 1. Implementation steps (1) System architecture and data collection

[0094] Microgrid architecture:

[0095] Distributed power sources: photovoltaic power generation (capacity 500kW), wind power generation (capacity 300kW), energy storage system (capacity 200kWh, rated power 100kW);

[0096] Load demand: The total peak load of the park is approximately 800kW, including industrial load (60%), commercial load (30%), and residential load (10%);

[0097] Grid interaction: connected to the main grid via a 10kV line, allowing bidirectional power flow.

[0098] Data collection:

[0099] Meteorological data: Light intensity of photovoltaic power station (unit: W / m 2 ), wind speed of wind turbine (unit: m / s), ambient temperature (unit: ℃);

[0100] Electricity price signal: real-time time-of-use electricity price (valley price 0.3 yuan / kWh, flat price 0.6 yuan / kWh, peak price 1.0 yuan / kWh);

[0101] Load data: historical load curve (one sampling point every 15 minutes, a total of 96 points / day);

[0102] System status: distributed power output, energy storage SOC (state of charge), grid voltage and frequency.

[0103] (2) Data preprocessing

[0104] Load data cleaning: remove outliers (e.g., load mutations exceeding ±50%) and use linear interpolation to fill in missing data;

[0105] Normalization of meteorological data: Map light intensity, wind speed, and temperature to the [0,1] interval using the following formula:

[0106]

[0107] Temporal feature encoding: Convert hours and seasons into one-hot encoding. For example, hours 0-23 are encoded into a 24-dimensional vector.

[0108] (3) Load forecasting (based on LSTM neural network)

[0109] Model structure:

[0110] Input layer: historical load (96 points), light intensity (96 points), wind speed (96 points), temperature (96 points), time characteristics (24-dimensional hour coding + 4-dimensional season coding);

[0111] Hidden layer: 2 layers of LSTM, 128 neurons per layer, ReLU activation function;

[0112] Output layer: fully connected layer, outputs the load forecast value for the next 24 hours (96 points).

[0113] Training parameters:

[0114] Loss function: mean square error (MSE);

[0115] Optimizer: Adam, learning rate 0.001;

[0116] Training set: data from the past 30 days; test set: data from the next 7 days;

[0117] Prediction results:

[0118] Mean absolute error (MAE): 12.5kW, root mean square error (RMSE): 18.3kW.

[0119] Example of prediction curve (partial):

[0120] Time (h) Actual load (kW) Forecast load (kW) Error (kW) 8:00 200 195 -5 12:00 450 462 +12 18:00 600 598 -2

[0121] (4) Multi-objective optimization scheduling (based on NSGA-II algorithm)

[0122] Objective function:

[0123] minF=ω1C op +ω2E em -ω3R rel

[0124] Weight settings: ω1 = 0.5, ω2 = 0.3, ω3 = 0.2 (can be adjusted according to needs).

[0125] Constraints:

[0126] Energy storage SOC range: 20% ≤ SOC ≤ 90%;

[0127] Grid interaction power limit: power purchase ≤ 500kW, power sales ≤ 300kW;

[0128] Voltage fluctuation range: 380V±5%.

[0129] Algorithm parameters:

[0130] Population size: 100; number of iterations: 200; crossover probability: 0.9; mutation probability: 0.1.

[0131] Optimization results:

[0132] Operating cost: After optimization, the average daily cost is 1,200 yuan, which is 22% lower than the traditional method;

[0133] Carbon emissions: The average daily carbon emissions are 150kg, a reduction of 18%;

[0134] Power supply reliability: The load power shortage rate dropped from 0.8% to 0.1%.

[0135] (5) Real-time control and fault protection

[0136] Control strategy:

[0137] Energy storage system: Charge to SOC = 80% during low electricity price (0:00-6:00), and discharge to SOC = 30% during peak electricity price (18:00-22:00);

[0138] Grid interaction: When there is excess output from photovoltaic and wind power (such as at 12:00 noon), electricity is sold to the grid; when the load is peak (such as at 19:00 in the evening), electricity is purchased from the grid.

[0139] Failsafe:

[0140] When the voltage is lower than 360V, the island operation mode is activated and non-critical loads (commercial lighting) are cut off;

[0141] When the energy storage SOC is lower than 20%, discharging is prohibited and only charging is allowed.

[0142] 2. Implementation case verification

[0143] (1) Typical daily operating data

[0144] Weather conditions: peak light intensity 800W / m 2 (12:00), average wind speed 5m / s.

[0145] Load curve:

[0146] Off-peak period (0:00-8:00): load 200kW;

[0147] Normal period (8:00-18:00): load 400-600kW;

[0148] Peak period (18:00-24:00): load 700-800kW.

[0149] Optimized scheduling results:

[0150] Time (h) Photovoltaic output (kW) Wind power output (kW) Energy storage charging and discharging (kW) Grid interaction (kW) 8:00 100 150 -50 (charging) +200 (electricity purchase) 12:00 400 200 +50 (discharge) -300 (electricity sales) 18:00 50 100 -80 (charging) +500 (electricity purchase)

[0151] (2) Performance index comparison

[0152] index Traditional methods Method of the present invention Improvement ratio Daily operating cost (yuan) 1540 1200 -22％ Daily carbon emissions (kg) 183 150 -18％ Load power failure rate (%) 0.8 0.1 -87.5％ Renewable energy consumption rate (%) 75 90 +20％

[0153] 3. Key technical parameters

[0154] LSTM model parameters:

[0155] Input dimensions: 96 (load) + 96 (light) + 96 (wind speed) + 96 (temperature) + 28 (time) = 412;

[0156] Output dimension: 96 (load for the next 24 hours).

[0157] NSGA-II algorithm parameters:

[0158] Chromosome encoding: distributed power output (continuous variable), energy storage charging and discharging power (continuous variable), grid interaction power (continuous variable);

[0159] Mutation operator: Gaussian mutation with a standard deviation of 5% of the variable range.

[0160] 4. Conclusion

[0161] Through actual case verification, the method of the present invention can significantly improve the economy, environmental protection and reliability of the smart park microgrid, specifically:

[0162] Operating costs reduced by more than 20%;

[0163] Reduce carbon emissions by more than 15%;

[0164] Power supply reliability increased to over 99.9%;

[0165] The renewable energy consumption rate has increased to over 90%.

[0166] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. A method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid, characterized in that: The following steps are involved: S1 multi-energy complementary layer construction: Integrate distributed power sources and grid power within the park, and achieve multi-energy synergy and complementarity through energy conversion and storage; S2 intelligent scheduling layer optimization: Based on load forecasts, real-time electricity price signals, and grid status, a multi-objective optimization algorithm is used to generate the optimal scheduling plan for distributed generation output, energy storage charging and discharging, and grid interaction power. S3 optimizes the control layer execution: The optimization results are sent to the microgrid controller in real time to dynamically adjust the distributed power output, energy storage charging and discharging strategy, and grid interaction power to achieve real-time optimized operation of the microgrid.

2. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 1 is characterized by: The multi-objective optimization algorithm is based on the following objective function: minF=ω1C op +ω2E em -ω3R rel in: C op The operating costs of the microgrid, including fuel costs, electricity purchase costs, and equipment maintenance costs; E em Carbon emissions are calculated based on the type of distributed generation and the carbon factor of electricity purchased from the grid; R rel For power supply reliability, the load power shortage rate or voltage qualification rate is evaluated; ω1, ω2, and ω3 are weight coefficients, satisfying ω1+ω2+ω3=1.

3. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 1 is characterized by: The load forecast adopts a deep learning model based on LSTM neural network. The input data includes historical load data, meteorological data and time characteristics, and outputs the load forecast curve for the next 24 hours. Among them, the meteorological data includes light intensity, wind speed, and temperature, and the time characteristics include hours and seasons.

4. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 1 is characterized by: The charging and discharging strategies of the energy storage system include: Charge during low electricity price periods and discharge during peak periods; It charges when renewable energy output is in excess and discharges when output is insufficient; Participate in the power balance and voltage support of the microgrid according to grid frequency or voltage fluctuations.

5. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 1 is characterized by: The optimization strategy for the grid interaction power includes: Dynamically adjust the power purchase and sales between the microgrid and the large power grid based on real-time electricity price signals; Sell ​​electricity to the grid when the microgrid has excess power, and purchase electricity from the grid when the power is insufficient; Give priority to the consumption of renewable energy and reduce dependence on the power grid.

6. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 1 is characterized by: The real-time optimization operation also includes a fault protection mechanism that triggers protection actions when the following abnormal conditions are detected: The voltage or frequency of the microgrid exceeds the allowable range; Failure of the distributed power generation or energy storage system; The interaction power between the grid and the microgrid exceeds the limit.

7. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 6 is characterized by: The protection actions include: Start the island operation mode and disconnect from the main power grid; Cut off non-critical loads to ensure power supply to important loads; Adjust the output of distributed power sources or the charging and discharging power of energy storage to restore system balance.

8. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 1 is characterized by: The multi-energy complementary layer further comprises: Coordinated output control of photovoltaic and wind power, dynamically adjusting the output ratio based on light intensity and wind speed forecasts; Energy storage system state of charge monitoring and charge and discharge power limitation to avoid overcharging or over-discharging; The power interaction limit between the microgrid and the large grid is set to prevent power backflow or overload.

9. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 1 is characterized by: The multi-objective optimization algorithm is one of the following algorithms or a combination thereof: Genetic algorithm; Particle swarm optimization; Simulated annealing algorithm; Non-dominated sorting genetic algorithm.

10. The method for intelligent dispatching and multi-energy complementary optimization control of a smart park microgrid according to claim 1 is characterized by: The smart park microgrid also includes: Real-time monitoring and fault diagnosis of distributed power sources; Dynamic classification of load demand, including interruptible load and non-interruptible load; Data interaction and collaborative control with the park energy management system.

Images