Source network load storage artificial intelligence scheduling method, system and equipment

Abstract

The invention relates to the technical field of micro-grid control and adjustment, in particular to a source network load storage artificial intelligence scheduling method, system and equipment, and the method comprises the steps: training and updating a neural network prediction model through employing real-time new energy power generation data and real-time electrical load data; determining a power generation power curve of new energy power generation and a power utilization load curve of user power utilization by using the neural network prediction model updated in real time; based on the generated power curve and the electrical load curve, comprehensive operation cost information of the micro-grid in the next power distribution period is determined; and determining target scheduling planning information of the micro-grid based on the comprehensive operation cost information. According to the source network load storage artificial intelligence scheduling method, model training is performed by using real-time data, and a future prediction curve is continuously updated, so that prediction errors are significantly reduced, and the accuracy and efficiency of micro-grid scheduling are improved.

Description

Technical Field

[0001] The present invention relates to the field of control and regulation technology, and in particular to a source-grid-load-storage artificial intelligence scheduling method, system and device. Background Art

[0002] With the transformation of global energy structure and the enhancement of environmental protection awareness, clean energy such as wind power and photovoltaics has achieved unprecedented development. However, the inherent intermittent and random characteristics of these clean energy sources inevitably bring challenges to the safe and stable operation of the power grid when they are directly connected to the large power grid. As a key means to improve the economic efficiency of microgrid operation and promote the consumption of clean energy, the optimization dispatching technology of microgrid has received extensive attention and research in the industry.

[0003] In existing microgrid technologies, the output prediction of new energy sources such as photovoltaic and wind power generation usually relies on mathematical models. The prediction results are used to establish the operating cost function of the microgrid, which usually covers multiple aspects such as fuel costs, grid operation and maintenance costs, and grid interaction costs. The purpose is to achieve the lowest total cost of microgrid scheduling and intelligent scheduling and control of each unit of the microgrid. However, the prediction results of the mathematical model are greatly affected by the model assumptions and parameter settings. For complex and changing environmental conditions, its prediction accuracy is often low, making it difficult to achieve the expected scheduling and operating costs of each unit of the microgrid. Summary of the invention

[0004] In order to solve the technical problem of low prediction accuracy of renewable energy output in microgrids through mathematical models, the purpose of the present invention is to provide a source-grid-load-storage artificial intelligence scheduling method, system and device. The technical solutions adopted are as follows: The present invention provides a source-grid-load-storage artificial intelligence scheduling method, the method comprising: Use real-time renewable energy generation data and real-time electricity load data to train and update the neural network prediction model; Determine the power curve of renewable energy generation and the power load curve of users using a real-time updated neural network prediction model; Determining the comprehensive operating cost information of the microgrid in the next power distribution cycle based on the power generation curve and the power load curve; Based on the comprehensive operation cost information, target dispatch planning information of the microgrid is determined.

[0005] Optionally, the new energy power generation includes: photovoltaic power generation and wind power generation; the neural network prediction model includes: photovoltaic power generation prediction model, wind power generation prediction model and power load prediction model; The step of training and updating the neural network prediction model using real-time new energy generation data and real-time electricity load data includes: Use the real-time photovoltaic power generation data and the real-time light intensity, temperature, humidity and wind speed data as the first time series data set to train and update the photovoltaic power generation prediction model; Using the real-time wind power generation data and the real-time temperature, humidity, wind speed, wind direction, and air pressure data as a second time series data set to train and update the wind power generation prediction model; The real-time power load data and the real-time temperature, humidity, holidays, electricity prices and equipment maintenance event data are used as the third time series data set to train and update the power load forecasting model.

[0006] Optionally, the power load curve includes: power load curves in peak, flat and valley time periods; After the step of determining the power curve of the new energy power generation and the power load curve of the user's power consumption by using the real-time updated neural network model, the method further includes: Based on the current user satisfaction and the pre-optimization power load curves of each peak, flat and valley period, determine the post-optimization power load curves of each peak, flat and valley period; In combination with the neural network prediction model and using the power load curve before optimization and the power load curve after optimization, the user's power satisfaction is updated; The updated electricity satisfaction is used as the current user satisfaction, and the steps of determining the optimized electricity load curves for each peak, flat and valley period are performed based on the current user satisfaction and the electricity load curves before optimization for each peak, flat and valley period.

[0007] Optionally, the step of determining the optimized power load curves for each peak, flat and valley time period based on the current user satisfaction and the power load curves for each peak, flat and valley time period before optimization includes: Determine the load transfer coefficient corresponding to the current user satisfaction; The maximum load transferable during the peak period and the maximum load transferable during the flat and valley period are calculated by using the load transferability coefficient and the pre-optimization power load curves during the peak, flat and valley periods; The optimized power load curves for each peak, flat and valley period are determined by using the maximum transferable outgoing load, the maximum transferable incoming load and the pre-optimization power load curves for each peak, flat and valley period.

[0008] Optionally, the step of determining the comprehensive operating cost information of the microgrid in the next power distribution cycle based on the power generation curve and the power load curve includes: Based on the power generation curve and the power load curve, determine the energy storage charging and discharging power, diesel fuel power generation power and interaction power between the microgrid and the distribution network tie line at each moment of the next power distribution cycle; The comprehensive operation cost information of the microgrid in the next power distribution cycle is calculated by using the energy storage charging and discharging power, the diesel fuel power generation power and the interactive power.

[0009] Optionally, the step of calculating the comprehensive operating cost information of the microgrid in the next power distribution cycle by using the energy storage charging and discharging power, the diesel fuel power generation power and the interactive power includes: The energy storage operation and maintenance cost is calculated using the energy storage charging and discharging power and the energy storage charging and discharging cost coefficient; Utilizing the diesel fuel power generation capacity and the diesel fuel cost coefficient, the diesel fuel power generation cost is calculated; The electricity purchase cost and electricity sales revenue are calculated by using the interactive power and the agreed price between the microgrid and the distribution network; The environmental maintenance cost is calculated using the diesel fuel power generation, pollutant emission coefficient and pollutant discounted cost; By combining the energy storage operation and maintenance costs, diesel fuel power generation costs, electricity purchase costs, electricity sales revenue and environmental maintenance costs, the comprehensive operating cost information of the microgrid in the next distribution cycle is calculated.

[0010] Optionally, the step of determining target dispatch planning information of the microgrid based on the comprehensive operation cost information includes: Determine the objective function of the power dispatch model corresponding to the comprehensive operating cost information; Determine the constraints of the power dispatch model based on the balance between power supply and demand, the rated power of diesel generators, the maximum power of the microgrid and distribution network interconnection line, and the energy storage limit; According to the objective function and the constraint conditions, target scheduling planning information of the microgrid is determined by a particle swarm algorithm; Among them, the target scheduling planning information includes the target output power of energy storage, the target interaction power of the microgrid and distribution network interconnection line, and the target output power of diesel generator.

[0011] Optionally, after the step of determining the power curve of renewable energy power generation and the power load curve of user power consumption by using the real-time updated neural network prediction model, the method further comprises: Determine the amount of new energy overflow by using the power generation curve and the power load curve; Based on the new energy overflow, the reserved remaining energy storage capacity of the microgrid is determined.

[0012] The present invention also provides a source-grid-load-storage artificial intelligence scheduling system, which is used to implement the source-grid-load-storage artificial intelligence scheduling method as described in any one of the above items; the system comprises: Model training module, used to train and update the neural network prediction model using real-time renewable energy generation data and real-time electricity load data; The power prediction module is used to determine the power curve of renewable energy power generation and the power load curve of users' power consumption by using a real-time updated neural network prediction model; A cost analysis module, used to determine the comprehensive operating cost information of the microgrid in the next power distribution cycle based on the power generation curve and the power load curve; The scheduling planning module is used to determine the target scheduling planning information of the microgrid based on the comprehensive operation cost information.

[0013] The present invention also provides a source-grid-load-storage artificial intelligence scheduling device, which includes a processor, a memory, and a source-grid-load-storage artificial intelligence scheduling program stored in the memory and executable by the processor, wherein when the source-grid-load-storage artificial intelligence scheduling program is executed by the processor, the steps of the source-grid-load-storage artificial intelligence scheduling method as described in any of the above items are implemented.

[0014] The present invention has the following beneficial effects: The present invention predicts the power generation of renewable energy and the user load power through artificial neural networks, which can automatically learn and identify complex patterns and rules in the data, and provide more accurate prediction results. Different from the traditional prediction method based on mathematical models, the present invention uses real-time data for model training and continuously updates the future prediction curve, which significantly reduces the prediction error and improves the accuracy and efficiency of microgrid scheduling. And the use of more accurate prediction curves can more accurately calculate the actual comprehensive operating costs, and then the power dispatch of the microgrid is more reasonable, accurate and efficient from the perspective of lowering the comprehensive operating costs, which is conducive to further reducing the comprehensive operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A flowchart of the steps of a source-grid-load-storage artificial intelligence scheduling method provided by one embodiment of the present invention; Figure 2 A detailed flow chart of step S1 in a source-grid-load-storage artificial intelligence scheduling method provided by one embodiment of the present invention; Figure 3A flowchart of the steps after step S2 in a source-grid-load-storage artificial intelligence scheduling method provided by one embodiment of the present invention; Figure 4 A schematic diagram of new energy, load output and load optimization involved in a source-grid-load-storage artificial intelligence scheduling method provided by one embodiment of the present invention; Figure 5 A schematic diagram for comparing time-of-use electricity prices and dynamic electricity prices involved in a source-grid-load-storage artificial intelligence scheduling method provided by an embodiment of the present invention; Figure 6 A schematic diagram of a daily microgrid operation plan without considering load transfer involved in a source-grid-load-storage artificial intelligence scheduling method provided by an embodiment of the present invention; Figure 7 A schematic diagram of a daily microgrid operation plan taking load transfer into consideration in a source-grid-load-storage artificial intelligence scheduling method provided by an embodiment of the present invention; Figure 8 It is a structural diagram of the hardware operating environment of the source-grid-load-storage artificial intelligence scheduling device involved in the embodiment of the present invention; Fig. 9 This is a schematic diagram of the framework structure of the source-grid-load-storage artificial intelligence scheduling system involved in the embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a source-grid-load-storage artificial intelligence scheduling method proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0019] The specific scheme of the source-grid-load-storage artificial intelligence scheduling method provided by the present invention is described in detail below with reference to the accompanying drawings.

[0020] Embodiment 1: For the artificial intelligence scheduling method for source-grid-load-storage provided by the present invention, please refer to Figure 1 , which shows a step flow chart of a source-grid-load-storage artificial intelligence scheduling method provided by an embodiment of the present invention.

[0021] The method comprises: Step S1, using real-time renewable energy power generation data and real-time power load data to train and update the neural network prediction model; This embodiment can be applied to a microgrid, which has a variety of micro sources, including energy storage systems (ES), diesel generators (DE), wind turbines (WT) and photovoltaics (PV). It is necessary to comprehensively consider the characteristics and advantages of these resources and make reasonable decisions in the optimization scheduling to ensure continuous and stable power supply under complex operating conditions.

[0022] To formulate an operation plan for a typical wind, solar, diesel and energy storage grid-connected microgrid, the most important thing is to accurately predict the output of new energy and load. The traditional photovoltaic and wind power prediction curves are predicted by mathematical models. The prediction results of mathematical models are greatly affected by model assumptions and parameter settings. For complex and changing environmental conditions, their prediction accuracy is easily limited. Artificial intelligence algorithms represented by neural network models can automatically extract and learn complex patterns and laws in data, and have better prediction capabilities and adaptability.

[0023] In order to improve the prediction accuracy of the neural network model, real-time renewable energy power generation data and real-time electricity load data are selected as training sets to train and update the neural network prediction model. Real-time here refers to collecting renewable energy power generation data and electricity load data every preset time period, for example, obtaining real-time renewable energy power generation data and real-time electricity load data every half an hour or one hour, and then continuously optimizing the neural network prediction model to make the prediction results more accurate.

[0024] In a preferred embodiment, please refer to Figure 2 , the new energy power generation includes: photovoltaic power generation and wind power generation; the neural network prediction model includes: photovoltaic power generation prediction model, wind power generation prediction model and power load prediction model; The step S1 comprises: Step S10, using the real-time photovoltaic power generation data and the real-time light intensity, temperature, humidity and wind speed data as a first time series data set to train and update a photovoltaic power generation prediction model; Step S11, using the real-time wind power generation data and the real-time temperature, humidity, wind speed, wind direction and air pressure data as a second time series data set to train and update the wind power generation prediction model; Step S12, using the real-time power load data and the real-time temperature, humidity, holidays, electricity prices and equipment maintenance event data as a third time series data set to train and update the power load prediction model.

[0025] In this embodiment, considering that the photovoltaic power generation power and microgrid user load power data in new energy sources are closely related to time, light intensity, temperature, humidity, and wind speed are used as influencing factors of photovoltaic power generation (with real-time photovoltaic power generation data as the first time series data set), and special events such as temperature, humidity, holidays, electricity prices, and equipment maintenance are used as influencing factors of user load (with real-time power load data as the third time series data set). Both can be predicted by combining the time series model Prophet and the autoregressive neural network AR-Net. The corresponding models are trained with the above-mentioned real-time data at historical 1h (hour) intervals, and the photovoltaic power generation power curve and power load curve for the next 48h can be predicted. For the prediction of wind power generation, temperature, humidity, wind speed, wind direction, and air pressure are used as influencing factors (with real-time wind power generation data as the second time series data set), and an integrated model based on the long short-term memory network LSTM is used for prediction. The model is trained with historical 1h interval data to predict the wind power generation power curve for the next 48h.

[0026] There is uncertainty in the output of new energy and power load. In this prediction module, the above models will be trained based on historical data and current data (collectively referred to as real-time data). The new energy power generation curve and power load curve for the next 48 hours will be updated every 1 hour to reduce the prediction error and improve the reliability of the prediction results. This is also extremely important for timely adjusting the microgrid operation plan within the same day.

[0027] Step S2, using a real-time updated neural network prediction model to determine a power curve of renewable energy power generation and a power load curve of user power consumption; Based on the above-mentioned neural network prediction models updated in each preset time period, the power generation curve of new energy power generation and the power load curve of users' electricity consumption within a certain period of time in the future are predicted.

[0028] In one embodiment, please refer to Figure 3 The power load curve includes: power load curves in peak, flat and valley periods; after step S2, the method further includes: Step S20, based on the current user satisfaction and the pre-optimization power load curves of the peak, flat and valley periods, determine the post-optimization power load curves of the peak, flat and valley periods; In the past, most microgrid optimization dispatching models were based on the power generation side, without considering the user demand side. In order to tap the potential of user-side regulation resources and reduce the user-side electricity cost while improving the economic operation level of the microgrid, the dispatching strategy of this embodiment also considers the user demand side response, which can not only reduce the operating cost of the microgrid and the user's electricity cost, but also increase the new energy consumption rate and reduce the abandonment of solar and wind power. The specific method is as follows: User loads in microgrids can be divided into fixed loads and transferable loads. When the total load remains unchanged, the load in one period can be moved to another period. From the time-of-use electricity price, it can be seen that moving the load in the peak period to the normal period and valley period can further reduce the user's electricity purchase cost.

[0029] Load transfer should take into account user satisfaction with electricity consumption (the parameter range can be expressed as 0~5). User satisfaction with electricity consumption includes two aspects: comfort and economy (both can be expressed as 0~5). In addition to considering improving the economy of user electricity consumption, the user's comfort with electricity consumption should also be considered. The load transfer amount in each period should have a maximum limit. If the user's comfort with electricity consumption is not considered, the results of the optimized dispatch may harm the interests of the user, making the user dissatisfied, and it will be difficult to achieve the purpose of coordinating user electricity consumption to adapt to the peak shaving and valley filling of the large power grid. Comprehensively consider this aspect and use user satisfaction as a reference indicator to prevent the electricity price response measures from being too radical and ignoring the user's electricity experience.

[0030] The load power curve of the microgrid before optimization (the power load curve before optimization), For fixed load power, is the transferable load power, The unit price of electricity purchased by microgrid users, The electricity purchase cost for microgrid users, The optimized load power curve (optimized power load curve) is used to calculate the comprehensive cost of the microgrid. The total load power after optimization should be consistent with the total load power before optimization. Load transfer aims to reduce the user's electricity purchase cost as much as possible without greatly affecting the electricity consumption experience.

[0031] Power load curve It can also be expressed as peak, valley and normal period load curves (power load curves in peak, flat and valley periods). They represent the load curves of peak, valley and normal periods before optimization. Starting from the current user satisfaction, based on the user satisfaction parameter and combined with the load curves before optimization in the peak, flat and valley periods, the load curves before optimization are optimized to determine the optimized load curves after optimization in the peak, flat and valley periods. Furthermore, the optimized load curve can be further updated in combination with the neural network prediction model to obtain new user satisfaction, so as to optimize and predict the user load curve repeatedly, reduce the user's electricity purchase cost while obtaining more accurate prediction results, and reduce the impact on the user's electricity experience. Starting from improving user satisfaction with electricity, the new energy consumption rate can be increased, and the energy waste of abandoned solar and wind power can be reduced.

[0032] In addition, user satisfaction can be set to an initial value, which can be set based on actual conditions, and subsequent user satisfaction can be obtained based on calculations.

[0033] Existing technologies mainly focus on the optimization of the power generation side, while this proposal not only considers the optimization of the power generation side, but also fully combines the user demand side response. This proposal fully considers the user demand side response, and through a reasonable load transfer strategy, the load during the peak period is transferred to the valley period and the normal period to reduce the user's electricity cost and reduce the peak pressure of the power grid. At the same time, the strategy also considers the user's comfort level in electricity use. When formulating the dispatching strategy, this proposal takes user satisfaction as an important indicator to ensure that load transfer does not harm the interests of users.

[0034] Specifically, the step S20 includes: Determine the load transfer coefficient corresponding to the current user satisfaction; The maximum load transferable during the peak period and the maximum load transferable during the flat and valley period are calculated by using the load transferability coefficient and the pre-optimization power load curves during the peak, flat and valley periods; The optimized power load curves for each peak, flat and valley period are determined by using the maximum transferable outgoing load, the maximum transferable incoming load and the pre-optimization power load curves for each peak, flat and valley period.

[0035] Maximum load that can be transferred during peak hours , the maximum load that can be transferred during normal and off-peak periods , the specific load transfer method is as follows: in It indicates the maximum percentage of the load that can be transferred in this period, that is, the limit on the maximum load transfer amount in this period. It is called the load transfer coefficient here. Its value is closely related to user satisfaction and can be the normalized value of user satisfaction. .

[0036] Pick , then the optimization curves for each period are as follows: , , They respectively represent the power load curves during peak, valley and normal periods after optimization.

[0037] Step S21, combining the neural network prediction model and using the power load curve before optimization and the power load curve after optimization to update the user's power satisfaction; Step S22, taking the updated electricity satisfaction as the current user satisfaction, executing the step of determining the optimized electricity load curves for each peak, flat and valley period based on the current user satisfaction and the pre-optimization electricity load curves for each peak, flat and valley period.

[0038] The calculation formula of user satisfaction in this embodiment is as follows: in User satisfaction, defined as the comfort level of electricity consumption and electricity economy The product of . is the user tolerance threshold, which can be set according to the actual situation. It represents the worst power consumption experience that the user can tolerate after the load transfer. That is, the comfort level of electricity consumption decreases; This shows that after load optimization, the user's electricity economy is improved and the electricity bill expenditure is reduced. This shows that the economic efficiency after load transfer is not as good as before load transfer; is the ideal state after load transfer, that is, the state of increased user satisfaction, which is difficult to achieve, that is, sacrificing some power consumption comfort but in exchange for the improvement of user's overall power consumption satisfaction. It should be pointed out that here and Because the neural network prediction model continuously predicts updated results, user satisfaction will also be continuously calculated and updated based on the updated curve.

[0039] In order to effectively improve user satisfaction, it is easier to achieve In one embodiment, the time-of-use electricity price can be replaced by a dynamic electricity price: As a price-based demand response method, time-of-use electricity prices divide the 24 hours of a day into peak, flat, and valley periods, and set different electricity prices for each period. Time-of-use electricity prices change users' electricity consumption behavior through price incentive mechanisms, dispatch demand-side response resources, and while improving the economy of the system, they are also conducive to achieving load peak shaving and valley filling, alleviating the pressure on the power grid during peak periods. The energy storage system dispatches output according to the peak shaving and valley filling strategy, outputs power to the outside for discharge during peak periods, charges and discharges according to actual needs during flat periods, and absorbs external power for charging during valley periods, thereby obtaining peak-valley electricity price difference benefits.

[0040] Please refer to Figure 5 This embodiment proposes a dynamic electricity price mechanism. To achieve a win-win situation for both parties, the distribution network and the microgrid can reach an agreement, that is, dynamically update the time-of-use electricity price based on the proportion of new energy produced by the microgrid. The dynamic electricity price is cheaper than the time-of-use electricity price, and the user's willingness to transfer load is increased, which is conducive to further improving the new energy consumption rate, reducing the user's electricity purchase cost, and increasing the benefits of the energy storage system, guiding users to actively participate in demand-side response. The dynamic electricity price calculation method in this embodiment is as follows: Among the above It is a dynamic electricity price, which is also the unit price of electricity purchased by microgrid users. The electricity price discount coefficient agreed upon between the microgrid and the distribution network can be set as required. Time-of-use electricity price; is the relative surplus power of wind and photovoltaic power, and Photovoltaic and wind power Output power at the moment, It is the user's electricity load power curve.

[0041] When , it indicates that the current wind and solar power generation output cannot fully meet the user's electricity load demand, that is, all new energy is consumed, and the time-of-use electricity price is still implemented; when It indicates that there may be a phenomenon of abandoned solar and wind power, which often occurs during periods when the output of renewable energy power generation is high but the load is at a low point. Dynamic electricity prices are implemented during these periods to enhance the willingness of microgrid users to transfer loads, thereby increasing the absorption rate of renewable energy and responding to the demand for peak shaving and valley filling in the distribution network.

[0042] The dynamic electricity price mechanism introduced in this embodiment dynamically adjusts the electricity price according to the real-time proportion of the excess renewable energy, so as to encourage users to participate in demand-side response and adjust their electricity consumption behavior, thereby balancing the grid load and improving the utilization rate of renewable energy. This mechanism can effectively reduce the user's electricity purchase cost and improve the economic benefits of the microgrid.

[0043] Step S3, determining the comprehensive operating cost information of the microgrid in the next power distribution cycle based on the power generation curve and the power load curve; The relationship between renewable energy generation and user electricity consumption can be obtained from the power generation curve and the power load curve, which determines the subsequent power configuration plan. For example, when renewable energy generation can fully meet user needs, there is no need to use diesel fuel to supplement power generation, nor is there any need to purchase additional electricity from the distribution network. However, when renewable energy generation cannot meet user needs, the cost of diesel fuel generation, electricity purchase costs, and various maintenance costs must be considered. Of course, even if renewable energy generation can fully meet user needs, it still requires a certain amount of cost. Based on this, the comprehensive operating cost information of the microgrid in the next distribution cycle is determined.

[0044] In a preferred embodiment, the step S3 comprises: Based on the power generation curve and the power load curve, determine the energy storage charging and discharging power, diesel fuel power generation power and interaction power between the microgrid and the distribution network tie line at each moment of the next power distribution cycle; The comprehensive operation cost information of the microgrid in the next power distribution cycle is calculated by using the energy storage charging and discharging power, the diesel fuel power generation power and the interactive power.

[0045] Specifically, the step of calculating the comprehensive operation cost information of the microgrid in the next power distribution cycle by using the energy storage charging and discharging power, the diesel fuel power generation power and the interactive power includes: The energy storage operation and maintenance cost is calculated using the energy storage charging and discharging power and the energy storage charging and discharging cost coefficient; Utilizing the diesel fuel power generation capacity and the diesel fuel cost coefficient, the diesel fuel power generation cost is calculated; The electricity purchase cost and electricity sales revenue are calculated by using the interactive power and the agreed price between the microgrid and the distribution network; The environmental maintenance cost is calculated using the diesel fuel power generation, pollutant emission coefficient and pollutant discounted cost; By combining the energy storage operation and maintenance costs, diesel fuel power generation costs, electricity purchase costs, electricity sales revenue and environmental maintenance costs, the comprehensive operating cost information of the microgrid in the next distribution cycle is calculated.

[0046] This embodiment aims to minimize the comprehensive operation cost of the microgrid on the next day and optimizes the energy storage system charging and discharging plan, diesel output plan, and microgrid and external network interaction plan.

[0047] Objective Function In the above formula is the next power distribution cycle in the future. For example, the comprehensive operation cost of the microgrid for 24 hours is the objective function of the power dispatching model. For economic costs, Environmental maintenance costs, Profit from peak-valley price difference of energy storage; For energy storage charging and discharging operation and maintenance costs, The cost of diesel fuel for electricity generation, The cost of purchasing electricity from the microgrid to the main grid, The diesel generator sets will have different degrees of impact on the environment, and their output determines the emission of various pollutants (CO2, SO2, NOx). is the pollutant emission coefficient, Output power for diesel generator, For the discounted cost of pollutants. The composition is specifically described as follows: It is estimated that the energy storage operation and maintenance cost It is proportional to its charge and discharge capacity. is the energy storage charging and discharging power, which is positive when discharging and negative when charging. is the energy storage charging and discharging cost coefficient; diesel fuel power generation cost Its output power is approximately a quadratic function. , , is the diesel fuel cost coefficient. It can be set as needed; is the interaction power between the microgrid and the distribution network tie line, which is positive when purchasing electricity and negative when selling electricity. The electricity sales price agreed upon between the microgrid and the distribution network.

[0048] Microgrid operation strategy that comprehensively considers environmental protection costs: When formulating the microgrid operation plan, this proposal not only considers economic costs, but also fully considers environmental protection costs, such as reducing the use of diesel fuel and increasing the absorption rate of renewable energy. This comprehensive cost optimization strategy helps to promote the development of microgrids in a greener and more environmentally friendly direction.

[0049] Step S4: determining target dispatch planning information of the microgrid based on the comprehensive operation cost information.

[0050] The comprehensive operating cost information is used as the objective function of the power dispatching model. According to the actual situation and constraints, the particle swarm algorithm can be used to determine the target dispatching planning information of the microgrid, such as the energy storage output, the power interaction between the microgrid and the distribution network, and the corresponding value of the diesel generator output.

[0051] In a preferred embodiment, the step S4 comprises: Determine the objective function of the power dispatch model corresponding to the comprehensive operating cost information; Determine the constraints of the power dispatch model based on the balance between power supply and demand, the rated power of diesel generators, the maximum power of the microgrid and distribution network interconnection line, and the energy storage limit; According to the objective function and the constraint conditions, target scheduling planning information of the microgrid is determined by a particle swarm algorithm; Among them, the target scheduling planning information includes the target output power of energy storage, the target interaction power of the microgrid and distribution network interconnection line, and the target output power of diesel generator.

[0052] The constraints of the power dispatch model include power balance, micro-power output, interconnection line transmission power and energy storage operation constraints.

[0053] (1) Power balance constraints In order to ensure the stability of the grid-connected microgrid, the power balance constraint should be met as much as possible.

[0054] (2) Diesel power generation output constraints The rated power of the diesel engine is relatively high. Diesel power generation costs are relatively high and should be started only when the energy storage output cannot meet the requirements.

[0055] (3) Transmission power constraints of tie lines is the maximum transmission power of the interconnection line between the microgrid and the distribution network.

[0056] (4) Energy storage operation constraints is the maximum charge and discharge power of the energy storage battery, , The permissible upper and lower limits of the energy storage battery's state of charge.

[0057] According to the objective function and constraints, the particle swarm algorithm is used to determine the corresponding values ​​of energy storage output, power interaction between microgrid and distribution network, and diesel generator output. The energy storage charging and discharging strategy needs to be determined according to the microgrid operation scenario. After the particle swarm algorithm calculates the values, the output can be adjusted according to the power balance constraint to provide reasonable guidance for the microgrid operation plan.

[0058] (1) Operation scenario 1: The load power is greater than the sum of PV and WT. All new energy sources are consumed, and the remaining power is provided by energy storage discharge first. .

[0059] (2) Operation scenario 2: load power is greater than the sum of PV and WT, all new energy is absorbed, and the energy storage is discharged at maximum power. From the perspective of low-carbon environmental protection, the power shortage is supplemented by purchasing electricity from the distribution network first. If the interconnection line interactive power reaches the maximum and there is still a power shortage or the distribution network is out of power, it is supplemented by diesel power generation.

[0060] (3) Operation scenario 3: load power is less than the sum of PV and WT, and there is surplus renewable energy power. The surplus renewable energy power is used to charge the energy storage first. .

[0061] (4) Operation scenario 4: The load power is less than the sum of PV and WT, and there is surplus new energy power. By reserving energy storage capacity, the overflow new energy power is used to charge the energy storage. If the energy storage has been charged at the maximum power or its SOC (state of charge) has reached the set maximum value, the surplus power is sold to the distribution network to maximize the new energy consumption rate.

[0062] In addition, in one embodiment, after step S2, the method further includes: Determine the amount of new energy overflow by using the power generation curve and the power load curve; Based on the new energy overflow, the reserved remaining energy storage capacity of the microgrid is determined.

[0063] The phenomenon of abandoning solar and wind power in microgrids often occurs during the day when the output of renewable energy power generation is high but the load is at a low point. By comparing the predicted values ​​of the next day's renewable energy power generation and load, it is predicted whether there will be energy abandonment and the time period when energy abandonment may occur. By implementing the strategy of reserving energy storage capacity when charging during the valley period and adjusting the energy storage charging plan, the new energy consumption rate can be further improved, and the overall economic benefits of the microgrid can be improved.

[0064] like: , it is predicted that there may be abandonment of solar and wind power during this period, among which is the maximum transmission power of the interconnection line between the microgrid and the distribution network.

[0065] At the same time, the total new energy overflow for the next day is predicted and calculated When it is predicted that solar and wind power may be abandoned the next day, it is recommended that you do not need to charge to the maximum SOC during the off-peak period and reserve the remaining energy storage capacity. : To smooth out fluctuations in wind and solar power generation.

[0066] In order to verify the actual effects of the above embodiments, the simulation experimental data corresponding to each embodiment are provided as follows: According to the microgrid operation data generated by simulation, it is input into the microgrid optimization scheduling plan module built by Python for comparative experiments to prove the feasibility of this scheme.

[0067] 1. Detailed calculation parameters Diesel generator operating parameters: , , , . Interaction power parameters of tie lines: . Energy storage battery operating parameters: , , , , battery capacity 300kwh. Demand side response parameters: , . Table 1: Time-of-use electricity prices and electricity sales price Table 2: Time-of-use electricity prices and electricity sales price Table 3: Simulated microgrid renewable energy generation and user load power Table 4: Comparison of user demand indicators under time-of-use electricity prices and dynamic electricity prices 5: Comparison of indicators on whether microgrids consider load transfer For a comparison of the power load curves before and after optimization in this experiment, please refer to Figure 4 ; For a comparison of the TOU price and the dynamic price in this experiment, please refer to Figure 5 ; For a comparison of the daily microgrid operation plans in this experiment without and with load transfer, please refer to Figure 6 and Figure 7 .

[0068] in conclusion (1) Through the implementation of the present invention, based on new energy and load forecasting, after adopting a strategy combining load transfer with dynamic electricity prices, the electricity purchase cost of microgrid users and the comprehensive cost of the microgrid have been reduced to a certain extent. At the same time, this strategy effectively reduces the energy abandonment phenomenon of photovoltaic power generation and wind power generation in the microgrid, reduces the power supply pressure of the power grid during peak hours, achieves the effect of peak shaving and valley filling, and helps to improve the overall stability and efficiency of the power grid.

[0069] (2) The present invention sets reasonable restrictions on load transfer in each time period, fully considers the user demand-side response, and optimizes the scheduling of microgrids based on user satisfaction. On the basis of taking into account the user's electricity comfort and economy, the dynamic electricity price mechanism proposed in the present invention can better encourage users to participate in load transfer, actively respond to the peak-shaving and valley-filling needs of the distribution network, and promote the harmonious interaction between the microgrid and the distribution network.

[0070] (3) When planning the load transfer strategy, the present invention also fully considers the environmental governance cost of diesel power generation pollutants. After the implementation of load transfer, the output power of the diesel generator is reduced during peak hours, which reduces the operating cost of the microgrid and further reduces the negative impact on the environment.

[0071] Embodiment 2: The embodiment of the present invention also provides a source-grid-load-storage artificial intelligence scheduling device. The source-grid-load-storage artificial intelligence scheduling device can be a device with data processing function, such as a server, a computer, a workstation, etc.

[0072] like Figure 8 As shown, Figure 8 It is a structural diagram of the hardware operating environment of the source-grid-load-storage artificial intelligence scheduling device involved in the embodiment of the present invention.

[0073] like Figure 8 As shown, the source-grid-load-storage artificial intelligence scheduling device may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display (Display), an input unit such as a control panel, and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WIFI interface). The memory 1005 may be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001. The memory 1005 as a computer storage medium may include a source-grid-load-storage artificial intelligence scheduling program.

[0074] Those skilled in the art will understand that Figure 8 The hardware structure shown in the figure does not constitute a limitation of the device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0075] Continue to refer to Figure 8 , Figure 8 The memory 1005 as a computer-readable storage medium may include an operating system, a user interface module, a network communication module, and a source-grid-load-storage artificial intelligence scheduling program.

[0076] exist Figure 8 In the embodiment, the network communication module is mainly used to connect to the server and can communicate data with the server; and the processor 1001 can call the source-grid-load-storage artificial intelligence scheduling program stored in the memory 1005 and execute the steps in the above embodiments.

[0077] The hardware structure of the above-mentioned source-grid-load-storage artificial intelligence scheduling device is used to implement various embodiments of the source-grid-load-storage artificial intelligence scheduling method of the present invention.

[0078] Embodiment three: In addition, the present invention also provides a source-grid-load-storage artificial intelligence dispatching system, please refer to Fig. 9 , the source-grid-load-storage artificial intelligence dispatching system includes: Model training module A10, used to train and update the neural network prediction model using real-time renewable energy generation data and real-time electricity load data; The power prediction module A20 is used to determine the power curve of the new energy power generation and the power load curve of the user's power consumption by using the real-time updated neural network prediction model; The cost analysis module A30 is used to determine the comprehensive operation cost information of the microgrid in the next power distribution cycle based on the power generation curve and the power load curve; The scheduling planning module A40 is used to determine the target scheduling planning information of the microgrid based on the comprehensive operation cost information.

[0079] Furthermore, the comparison and analysis module A20 is also used for: The step of determining interference difference values ​​of different pilot signals during transmission based on the historical channel estimation parameters includes: Based on the path information, determining the delay difference, signal gain difference and phase change difference between any two pilot signals; The interference difference value between the two pilot signals is calculated by using the delay difference, the signal gain difference and the phase change difference.

[0080] Furthermore, the model training module A10 is also used for: Use the real-time photovoltaic power generation data and the real-time light intensity, temperature, humidity and wind speed data as the first time series data set to train and update the photovoltaic power generation prediction model; Using the real-time wind power generation data and the real-time temperature, humidity, wind speed, wind direction, and air pressure data as a second time series data set to train and update the wind power generation prediction model; The real-time power load data and the real-time temperature, humidity, holidays, electricity prices and equipment maintenance event data are used as the third time series data set to train and update the power load forecasting model.

[0081] Furthermore, the power prediction module A20 is also used for: Based on the current user satisfaction and the pre-optimization power load curves of each peak, flat and valley period, determine the post-optimization power load curves of each peak, flat and valley period; In combination with the neural network prediction model and using the power load curve before optimization and the power load curve after optimization, the user's power satisfaction is updated; The updated electricity satisfaction is used as the current user satisfaction, and the steps of determining the optimized electricity load curves for each peak, flat and valley period are performed based on the current user satisfaction and the electricity load curves before optimization for each peak, flat and valley period.

[0082] Furthermore, the power prediction module A20 is also used for: Determine the load transfer coefficient corresponding to the current user satisfaction; The maximum load transferable during the peak period and the maximum load transferable during the flat and valley period are calculated by using the load transferability coefficient and the pre-optimization power load curves during the peak, flat and valley periods; The optimized power load curves for each peak, flat and valley period are determined by using the maximum transferable outgoing load, the maximum transferable incoming load and the pre-optimization power load curves for each peak, flat and valley period.

[0083] Furthermore, the cost analysis module A30 is also used for: Based on the power generation curve and the power load curve, determine the energy storage charging and discharging power, diesel fuel power generation power and interaction power between the microgrid and the distribution network tie line at each moment of the next power distribution cycle; The comprehensive operation cost information of the microgrid in the next power distribution cycle is calculated by using the energy storage charging and discharging power, the diesel fuel power generation power and the interactive power.

[0084] Furthermore, the cost analysis module A30 is also used for: The energy storage operation and maintenance cost is calculated using the energy storage charging and discharging power and the energy storage charging and discharging cost coefficient; Utilizing the diesel fuel power generation capacity and the diesel fuel cost coefficient, the diesel fuel power generation cost is calculated; The electricity purchase cost and electricity sales revenue are calculated by using the interactive power and the agreed price between the microgrid and the distribution network; The environmental maintenance cost is calculated using the diesel fuel power generation, pollutant emission coefficient and pollutant discounted cost; By combining the energy storage operation and maintenance costs, diesel fuel power generation costs, electricity purchase costs, electricity sales revenue and environmental maintenance costs, the comprehensive operating cost information of the microgrid in the next distribution cycle is calculated.

[0085] Furthermore, the scheduling planning module A40 is also used to: Determine the objective function of the power dispatch model corresponding to the comprehensive operating cost information; Determine the constraints of the power dispatch model based on the balance between power supply and demand, the rated power of diesel generators, the maximum power of the microgrid and distribution network interconnection line, and the energy storage limit; According to the objective function and the constraint conditions, target scheduling planning information of the microgrid is determined by a particle swarm algorithm; Among them, the target scheduling planning information includes the target output power of energy storage, the target interaction power of the microgrid and distribution network interconnection line, and the target output power of diesel generator.

[0086] Furthermore, the power prediction module A20 is also used for: Determine the amount of new energy overflow by using the power generation curve and the power load curve; Based on the new energy overflow, the reserved remaining energy storage capacity of the microgrid is determined.

[0087] The specific implementation methods of the source-grid-load-storage artificial intelligence scheduling system of the present invention are basically the same as the various embodiments of the above-mentioned source-grid-load-storage artificial intelligence scheduling method, and will not be repeated here.

[0088] In addition, the present invention also provides a computer-readable storage medium. The computer-readable storage medium of the present invention stores a source-grid-load-storage artificial intelligence scheduling program, wherein when the source-grid-load-storage artificial intelligence scheduling program is executed by a processor, the steps of the source-grid-load-storage artificial intelligence scheduling method as described above are implemented.

[0089] Among them, the method implemented when the source-grid-load-storage artificial intelligence scheduling program is executed can refer to the various embodiments of the source-grid-load-storage artificial intelligence scheduling method of the present invention, and will not be repeated here.

[0090] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0091] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0092] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0094] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0095] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0096] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A source-grid-load-storage artificial intelligence scheduling method, characterized in that: The method comprises: Use real-time renewable energy generation data and real-time electricity load data to train and update the neural network prediction model; Determine the power curve of renewable energy generation and the power load curve of users using a real-time updated neural network prediction model; Determining the comprehensive operating cost information of the microgrid in the next power distribution cycle based on the power generation curve and the power load curve; Based on the comprehensive operation cost information, target dispatch planning information of the microgrid is determined.

2. The source-grid-load-storage artificial intelligence scheduling method according to claim 1 is characterized in that: The new energy power generation includes: photovoltaic power generation and wind power generation; the neural network prediction model includes: photovoltaic power generation prediction model, wind power generation prediction model and power load prediction model; The step of training and updating the neural network prediction model using real-time new energy generation data and real-time electricity load data includes: Use the real-time photovoltaic power generation data and the real-time light intensity, temperature, humidity and wind speed data as the first time series data set to train and update the photovoltaic power generation prediction model; Using real-time wind power generation data and real-time temperature, humidity, wind speed, wind direction and air pressure data as a second time series data set to train and update a wind power generation prediction model; The real-time power load data and the real-time temperature, humidity, holidays, electricity prices and equipment maintenance event data are used as the third time series data set to train and update the power load forecasting model.

3. The source-grid-load-storage artificial intelligence scheduling method according to claim 1 is characterized in that: The power load curve includes: power load curves in peak, flat and valley periods; After the step of determining the power curve of the new energy power generation and the power load curve of the user's power consumption by using the real-time updated neural network model, the method further includes: Based on the current user satisfaction and the pre-optimization power load curves of each peak, flat and valley period, determine the post-optimization power load curves of each peak, flat and valley period; In combination with the neural network prediction model and using the power load curve before optimization and the power load curve after optimization, the user's power satisfaction is updated; The updated electricity satisfaction is used as the current user satisfaction, and the steps of determining the optimized electricity load curves for each peak, flat and valley period are performed based on the current user satisfaction and the electricity load curves before optimization for each peak, flat and valley period.

4. The source-grid-load-storage artificial intelligence scheduling method according to claim 3 is characterized in that: The step of determining the optimized power load curves for each peak, flat and valley period based on the current user satisfaction and the power load curves before optimization for each peak, flat and valley period includes: Determine the load transfer coefficient corresponding to the current user satisfaction; The maximum load transferable during the peak period and the maximum load transferable during the flat and valley period are calculated by using the load transferability coefficient and the pre-optimization power load curves during the peak, flat and valley periods; The optimized power load curves for each peak, flat and valley period are determined by using the maximum transferable outgoing load, the maximum transferable incoming load and the pre-optimization power load curves for each peak, flat and valley period.

5. The source-grid-load-storage artificial intelligence scheduling method according to claim 1 is characterized in that: The step of determining the comprehensive operating cost information of the microgrid in the next power distribution cycle based on the power generation curve and the power load curve includes: Based on the power generation curve and the power load curve, determine the energy storage charging and discharging power, diesel fuel power generation power and interaction power between the microgrid and the distribution network tie line at each moment of the next power distribution cycle; The comprehensive operation cost information of the microgrid in the next power distribution cycle is calculated by using the energy storage charging and discharging power, the diesel fuel power generation power and the interactive power.

6. The source-grid-load-storage artificial intelligence scheduling method according to claim 5 is characterized in that: The step of calculating the comprehensive operation cost information of the microgrid in the next power distribution cycle by using the energy storage charging and discharging power, the diesel fuel power generation power and the interactive power comprises: The energy storage operation and maintenance cost is calculated using the energy storage charging and discharging power and the energy storage charging and discharging cost coefficient; Utilizing the diesel fuel power generation capacity and the diesel fuel cost coefficient, the diesel fuel power generation cost is calculated; The electricity purchase cost and electricity sales revenue are calculated by using the interactive power and the agreed price between the microgrid and the distribution network; The environmental maintenance cost is calculated using the diesel fuel power generation, pollutant emission coefficient and pollutant discounted cost; By combining the energy storage operation and maintenance costs, diesel fuel power generation costs, electricity purchase costs, electricity sales revenue and environmental maintenance costs, the comprehensive operating cost information of the microgrid in the next distribution cycle is calculated.

7. The source-grid-load-storage artificial intelligence scheduling method according to claim 1 is characterized in that: The step of determining target dispatch planning information of the microgrid based on the comprehensive operation cost information includes: Determine the objective function of the power dispatch model corresponding to the comprehensive operating cost information; Determine the constraints of the power dispatch model based on the balance between power supply and demand, the rated power of diesel generators, the maximum power of the microgrid and distribution network interconnection line, and the energy storage limit; According to the objective function and the constraint conditions, target scheduling planning information of the microgrid is determined by a particle swarm algorithm; Among them, the target scheduling planning information includes the target output power of energy storage, the target interaction power of the microgrid and distribution network interconnection line, and the target output power of diesel generator.

8. The source-grid-load-storage artificial intelligence scheduling method according to claim 1 is characterized in that: After the step of determining the power curve of the new energy power generation and the power load curve of the user's power consumption by using the real-time updated neural network prediction model, the method further includes: Determine the amount of new energy overflow by using the power generation curve and the power load curve; Based on the new energy overflow, the reserved remaining energy storage capacity of the microgrid is determined.

9. A source-grid-load-storage artificial intelligence dispatching system, characterized in that: The system is used to implement the source-grid-load-storage artificial intelligence scheduling method according to any one of claims 1 to 8; the system comprises: Model training module, used to train and update the neural network prediction model using real-time renewable energy generation data and real-time electricity load data; The power prediction module is used to determine the power curve of renewable energy power generation and the power load curve of users' power consumption by using a real-time updated neural network prediction model; A cost analysis module, used to determine the comprehensive operating cost information of the microgrid in the next power distribution cycle based on the power generation curve and the power load curve; The scheduling planning module is used to determine the target scheduling planning information of the microgrid based on the comprehensive operation cost information.

10. A source-grid-load-storage artificial intelligence scheduling device, characterized in that: The source-grid-load-storage artificial intelligence scheduling device includes a processor, a memory, and a source-grid-load-storage artificial intelligence scheduling program stored in the memory and executable by the processor, wherein when the source-grid-load-storage artificial intelligence scheduling program is executed by the processor, the steps of the source-grid-load-storage artificial intelligence scheduling method as described in any one of claims 1 to 8 are implemented.

Images