A decentralized dispatch optimization method and system for weakly linked island microgrids

Abstract

The present invention relates to the technical field of microgrid optimization and scheduling, and in particular to a method and system for decentralized scheduling optimization of weakly linked island microgrid groups, including using historical island-to-island multi-microgrid scenario data sets for offline centralized optimization and solution, minimizing total resource consumption and carbon emissions as offline centralized optimization goals, generating daily controllable unit output time series data for each microgrid, and finding the most similar controllable unit output time series results through a similarity measurement algorithm based on the acquired real-time wind and solar data; using the most similar controllable unit output time series results as initial scheduling values ​​for online distributed optimization and solution, and introducing a penalty function for the controllable unit output deviation to obtain an online distributed optimization scheduling scheme, thereby performing island-to-island multi-microgrid power mutual assistance optimization scheduling. The present invention adopts a decentralized scheduling optimization method for weakly linked island microgrid groups that combines offline training and online optimization, effectively improving the efficiency and real-time performance of island microgrid power mutual assistance.

Description

Technical Field

[0001] The present invention relates to the technical field of microgrid optimization and dispatching, and in particular to a method and system for optimizing the decentralized dispatching of a weakly linked island microgrid cluster. Background Art

[0002] Against the backdrop of rapid development of the marine economy, ensuring normal power supply to my country's numerous islands has become a key task. The traditional power supply method that relies on diesel generators is not only costly, but also has a serious impact on the environment. Therefore, utilizing the abundant renewable energy sources such as wind and solar energy on islands to build island microgrids has become an effective way to solve the problem of island power supply. However, there are often problems with long communication distances and poor communication conditions between island microgrids, which poses a huge challenge to the energy scheduling of island microgrid groups.

[0003] Existing research on island microgrid systems mainly focuses on economic power planning and reliability optimization, but most of them fail to fully consider the high efficiency and real-time requirements of island microgrid power scheduling, especially under long-distance communication conditions. Among them, centralized optimization methods need to collect and process all data and variables of the entire system when dealing with large-scale island microgrids, resulting in a significant increase in calculation complexity and time, making it difficult to meet the requirements of real-time scheduling; and for distributed control methods, such as the scheduling model based on the alternating direction multiplier method, although it reduces the calculation and communication burden to a certain extent, when the system scale is large or the coupling relationship is strong, the convergence speed is slow, and it is easily restricted by communication conditions, affecting the accuracy and timeliness of the scheduling results.

[0004] In response to the above problems, the existing technology has obvious limitations in the power dispatch of island microgrid groups, especially when dealing with scheduling problems under large-scale systems, strong coupling relationships and long-distance communication conditions. It not only affects the operating efficiency of island microgrids, but also restricts the full utilization of renewable energy, thereby affecting the stability and sustainability of energy supply in island areas. Therefore, there is an urgent need for an efficient decentralized scheduling optimization method that adapts to the communication conditions of weakly linked island microgrid groups to solve the problems existing in the existing technology and improve the efficiency and real-time performance of power dispatch of island microgrid groups. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for decentralized scheduling optimization of weakly linked island microgrids, which realizes efficient and real-time scheduling of electric energy mutual assistance of island microgrids by constructing an island microgrid power architecture and a microgrid unit model and combining offline training with online optimization methods.

[0006] In order to solve the above technical problems, the present invention provides a method and system for decentralized scheduling optimization of weakly linked island microgrid clusters.

[0007] In a first aspect, the present invention provides a method for optimizing decentralized scheduling of weakly linked island microgrids, the method comprising the following steps:

[0008] According to the geographical location and resource distribution data of the island microgrid group, a multi-microgrid power mutual aid power architecture between islands, including load islands and resource islands, is constructed;

[0009] Based on the inter-island multi-microgrid power energy mutual aid power architecture, an island multi-microgrid unit model is established;

[0010] In the offline training phase, the historical multi-microgrid scenario dataset between islands is used for offline centralized optimization, and the minimization of total resource consumption and carbon emissions is used as the offline centralized optimization goal to generate the daily controllable unit output time series data of each microgrid under different natural resource conditions;

[0011] According to the real-time wind and solar data obtained, the most similar controllable unit output time series results are searched and found in the daily controllable unit output time series data through the similarity measurement algorithm;

[0012] The output timing result of the most similar controllable unit is used as the initial dispatch value, and a penalty function is introduced for the output deviation of the controllable unit. Combined with the current operating state of the island microgrid group, each island multi-microgrid unit model is optimized and solved online to obtain the online distributed optimization dispatching scheme of the current microgrid.

[0013] The online distributed optimization scheduling scheme is used to optimize the mutual power scheduling of multiple microgrids between islands.

[0014] In a second aspect, the present invention provides a decentralized scheduling optimization system for weakly linked island microgrids, the system comprising:

[0015] The power architecture building module is used to build a multi-microgrid power architecture among islands, including load islands and resource islands, based on the geographical location and resource distribution data of the island microgrid group;

[0016] A unit model building module, used to build an island multi-microgrid unit model based on the island multi-microgrid electric energy mutual aid power architecture;

[0017] The offline training optimization module is used to perform offline centralized optimization solutions using the historical multi-microgrid scenario dataset between islands during the offline training phase, and to generate daily controllable unit output time series data for each microgrid under different natural resource conditions, taking minimizing total resource consumption and carbon emissions as the offline centralized optimization goal;

[0018] A similarity measurement module is used to search and find the most similar controllable unit output time series results in the daily controllable unit output time series data based on the acquired real-time wind and solar data through a similarity measurement algorithm;

[0019] An online optimization solution module is used to use the output timing result of the most similar controllable unit as the initial scheduling value, introduce a penalty function for the output deviation of the controllable unit, and perform online distributed optimization solution on each of the island multi-microgrid unit models in combination with the current operating status of the island microgrid group to obtain an online distributed optimization scheduling plan for the current microgrid;

[0020] The microgrid optimization scheduling module is used to perform the multi-microgrid power mutual optimization scheduling between islands according to the online distributed optimization scheduling scheme.

[0021] The present invention provides a method and system for optimizing decentralized scheduling of weakly linked island microgrids. The method establishes an island multi-microgrid unit model by constructing an inter-island multi-microgrid electric energy mutual aid power architecture including a load island and a resource island; uses a historical inter-island multi-microgrid scenario data set to perform offline centralized optimization and solve, and takes minimizing total resource consumption and carbon emissions as offline centralized optimization goals to generate daily controllable unit output time series data of each microgrid under different natural resource conditions; searches and finds the most similar controllable unit output time series results in the daily controllable unit output time series data based on the acquired real-time wind and solar data through a similarity measurement algorithm; uses the most similar controllable unit output time series results as the initial scheduling value, introduces a penalty function for the controllable unit output deviation, and performs online distributed optimization and solves each island multi-microgrid unit model in combination with the current operating state of the island microgrid group to obtain an online distributed optimization scheduling plan for the current microgrid; and performs inter-island multi-microgrid electric energy mutual aid optimization scheduling according to the online distributed optimization scheduling plan. Compared with the existing technology, this method constructs an inter-island multi-microgrid power mutual assistance power architecture and its unit model, and combines offline training and online optimization methods to effectively improve the efficiency and real-time scheduling of island microgrid power mutual assistance, reduce the dependence on long-distance island-land communication, provide strong support for the stable operation of island microgrids, and improve the system's operating efficiency and real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of a flow chart of a decentralized scheduling optimization method for a weakly linked island microgrid group provided by an embodiment of the present invention;

[0023] Figure 2 It is a schematic diagram of a distributed scheduling process framework of a microgrid group provided by an embodiment of the present invention;

[0024] Figure 3 This is a schematic diagram of the electric energy interaction framework of the island microgrid group provided by an embodiment of the present invention;

[0025] Figure 4 is a schematic diagram of renewable energy output of multiple microgrids provided by an embodiment of the present invention;

[0026] Figure 5is a schematic diagram of a load curve of a load island provided by an embodiment of the present invention;

[0027] Figure 6 is a schematic diagram of interactive power between microgrids provided by an embodiment of the present invention;

[0028] Figure 7 This is a schematic diagram of wind island scheduling provided by an embodiment of the present invention;

[0029] Figure 8 is a schematic diagram of photovoltaic island scheduling provided by an embodiment of the present invention;

[0030] Fig. 9 is a schematic diagram of load-dominated scheduling provided by an embodiment of the present invention;

[0031] Fig.10 is a schematic diagram of load island scheduling provided by an embodiment of the present invention;

[0032] Fig.11 It is a block diagram of a decentralized scheduling optimization system for a weakly linked island microgrid cluster provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following specifically illustrates the implementation mode of the present invention in conjunction with the accompanying drawings. The embodiments are provided for illustrative purposes only and cannot be understood as limiting the present invention. The accompanying drawings are provided for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0034] refer to Figure 1 , the embodiment of the present invention provides a method for optimizing decentralized scheduling of weakly linked island microgrids, such as Figure 1 As shown, the method comprises the following steps:

[0035] S1. According to the geographical location and resource distribution data of the island microgrid group, a multi-microgrid power mutual aid power architecture between islands including load islands and resource islands is constructed.

[0036] Since traditional island power systems often have an island effect, that is, the power supplies between islands are independent of each other and lack interconnection, this embodiment fully considers the differences in resource distribution and electricity demand of each island, and divides the islands into two categories: load islands and resource islands by using the geographical location, resource distribution data and load demand of the island microgrid group; among them, load islands are mainly equipped with new energy power generation devices, energy storage systems and micro gas turbines for power supply due to their concentrated population, large electricity demand and diverse functions; resource islands have abundant renewable energy and are mainly equipped with a large number of wind turbines and photovoltaic power generation devices. The load islands constructed in this embodiment are mainly equipped with a large number of wind turbines and photovoltaic power generation devices. The inter-island microgrid power mutual aid power architecture makes full use of the renewable energy resources on the islands (such as wind power, photovoltaics, etc.) to break the island effect. Through microgrid interconnection and optimization algorithms, flexible scheduling and optimal configuration of island electricity are achieved, thereby realizing efficient and real-time power mutual aid between island microgrids. At the same time, the power mutual aid between microgrids reduces the dependence on external power grids or long-distance power transmission. When the natural factors of an island are insufficient (such as insufficient wind power or insufficient light), electricity can be quickly obtained from other islands, thereby ensuring the continuity and stability of the island's power supply and improving the stability and reliability of the entire power grid.

[0037] In this embodiment, the multi-microgrid electric energy mutual aid power architecture between islands includes load islands, resource islands and an information interaction system. This embodiment can evaluate the geographical location and resource distribution of the island group, determine the islands that can serve as resource islands and the islands that can serve as load islands, and configure wind turbines and photovoltaic panels on the resource islands, and configure energy storage devices and necessary new energy power generation facilities on the load islands. At the same time, a submarine cable connecting the resource islands and the load islands is established to achieve efficient transmission of electric energy through the submarine cable, realize electric energy mutual aid between the load islands and the resource islands, or realize real-time transmission of power supply and demand information between the islands through the information interaction system, wherein the information interaction The system refers to a communication network connecting each microgrid and resource island and load island, which is used to transmit power supply and demand information, control instructions, etc. The information interaction system may include communication base stations, optical fiber networks, wireless communication equipment, etc., to realize real-time transmission of power supply and demand information between islands, and provide data support for power dispatching and mutual assistance. In this embodiment, each load island and resource island is configured with an independent microgrid to achieve local power self-sufficiency. Each microgrid has the ability to operate in an isolated island and in grid-connected operation, that is, it can still supply power independently when operating in an isolated island or when the main power grid fails. When operating in grid-connected operation, it can carry out power mutual assistance with other microgrids to achieve optimal distribution and dispatch of electric energy.

[0038] Specifically, if Figure 3As shown, the microgrid system in this embodiment includes microgrid 1 (load main island), microgrid 2 (wind resource island), microgrid 3 (photovoltaic resource island) and microgrid 4 (load branch island). In this embodiment, microgrid 1 needs to be connected to other microgrids (such as microgrid 2, microgrid 3 and microgrid 4) for electric energy mutual assistance, and connected to the large power grid. At the same time, this embodiment establishes an information interaction system between the load main island and the wind resource island, photovoltaic resource island and load branch island to achieve power scheduling and optimization. Residential loads are connected to each microgrid through power cables. The power generated by each load island and resource island (wind resource island and photovoltaic resource island) or the power supply and demand information is collected in real time through sensors and monitoring equipment in its microgrid. The information interaction system or power lines are used to transmit the power supply and demand information or power to the microgrid system of the resource island and load island connected to it. Each resource island is based on After receiving the power supply and demand information, the submarine cable and the energy storage power station are started to transmit the excess power to the load island. The load island receives the power from the resource island and distributes the power to each power load through the distribution facilities in its microgrid system. For example, the load main island serves as the main power load center, receiving and distributing power from different microgrids; the wind resource island generates electricity using wind turbines, and is connected to the microgrid 1 through an information interaction system to transmit the wind power generation situation in real time and adjust the power generation according to system requirements; the photovoltaic resource island is the area where the solar photovoltaic power generation facilities are located, and is connected to the microgrid 1 through an information interaction system to transmit the photovoltaic power generation situation in real time. The island group power architecture established in this embodiment fully considers the differences in resource distribution and power demand of each island, realizes the reasonable allocation and effective utilization of resources, and improves the power supply reliability and economy of the entire archipelago.

[0039] S2. Based on the inter-island multi-microgrid power mutual aid power architecture, an island multi-microgrid unit model is established.

[0040] This embodiment establishes an island multi-microgrid unit model based on the island multi-microgrid electric energy mutual aid power architecture. The island multi-microgrid unit model includes an offshore wind power unit model, a photovoltaic unit model, a micro gas turbine unit model, a battery unit model, a seawater desalination load unit model, an electric vehicle load unit model and a distribution network model. The steps of establishing the island multi-microgrid unit model based on the island multi-microgrid electric energy mutual aid power architecture include:

[0041] According to the inter-island multi-microgrid electric energy mutual aid power architecture, the wind and solar resource generation and electric energy transmission process between different islands are simulated to obtain island power simulation data; the island power simulation data includes but is not limited to basic data of offshore wind farms, basic photovoltaic data, energy storage requirements of island power systems, and island power requirements; the field can obtain corresponding data from the simulation process, and is not limited to the embodiments of the present invention;

[0042] The wind speed at the offshore wind turbine hub height is calculated based on the basic data of the offshore wind farm; the basic data of the offshore wind farm includes the wind speed at the reference height of the offshore wind farm, the height of the offshore wind turbine hub and the surface roughness of the actual ocean state;

[0043] In different wind speed ranges, the relationship between wind speed at the hub height of offshore wind turbine and wind turbine output power is fitted using piecewise function to obtain the offshore wind power unit model.

[0044] According to the photovoltaic basic data, the photovoltaic power generation efficiency under the standard state is adjusted by using a preset photovoltaic temperature coefficient to obtain the current actual photovoltaic power generation efficiency; the photovoltaic basic data includes the current ambient temperature and the global horizontal irradiance;

[0045] Calculate the photovoltaic output power based on the photovoltaic plane area, the actual efficiency of photovoltaic power generation and the global horizontal irradiance, and build a photovoltaic unit model;

[0046] According to the power demand and resource conditions of the island, the operating parameters of the micro gas turbine are determined, and the relationship between the output power and efficiency of the micro gas turbine is fitted according to the operating parameters of the micro gas turbine to obtain a micro gas turbine unit model;

[0047] Determine the battery capacity and charging and discharging parameters according to the energy storage demand of the island power system, and use the charge state of the battery as a state variable, and update the charge state of each time step by considering the charging and discharging parameters and the time variation to obtain a battery unit model; wherein the charging and discharging parameters include charging efficiency and discharging efficiency;

[0048] The interaction power between different microgrids is obtained, and the distribution network model is constructed according to the sum of the interaction power of all microgrids in the distribution network and the distributed resource power within the microgrid.

[0049] In this embodiment, for the offshore wind power unit model, since the offshore wind power output is greatly affected by the sea breeze conditions, considering the differences in the influence of terrain, climate, and seasonal factors on the sea breeze compared to the land breeze, this embodiment uses non-parametric distribution to fit the offshore wind speed. This method has high requirements on the quality of historical wind speed data. Therefore, the offshore wind power unit model is obtained according to the relationship between the offshore wind speed and the wind turbine output. The offshore wind power unit model is specifically:

[0050]

[0051]

[0052] In the formula, is the fan output power; is the wind speed at the height of the offshore wind turbine hub; is the rated power of the fan; is the cut-in wind speed; To cut out the wind speed; is the rated wind speed; is the wind speed at the reference height of the offshore wind farm; It is the reference height of offshore wind farm; is the surface roughness of the actual ocean state; is the height of the offshore wind turbine hub; , , To fit the parameters, they can be obtained by fitting the historical data of offshore wind speed and the wind turbine power characteristic curve.

[0053] For photovoltaic devices, the output power of the photovoltaic power station is determined by the photovoltaic power generation efficiency and the global horizontal irradiance. Therefore, the photovoltaic unit model is specifically:

[0054]

[0055]

[0056] In the formula, is the photovoltaic output power; is the photovoltaic plane area; is the actual efficiency of photovoltaic power generation; is the global horizontal irradiance; is the reference global horizontal irradiance; is the photovoltaic power generation efficiency under standard conditions; is the photovoltaic temperature coefficient; is the ambient temperature; is the rated operating temperature of the photovoltaic cell; is the photovoltaic reference temperature.

[0057] For the gas turbine device, the micro gas turbine needs to consider its upper and lower output constraints, climbing constraints and efficiency equation. The micro gas turbine unit model is specifically as follows:

[0058]

[0059]

[0060]

[0061] In the formula, is the lower limit power of the micro gas turbine; is the upper limit power of the micro gas turbine; is the sequential output power of the micro gas turbine at time t; The climbing rate for output reduction; for The sequential output power of the micro gas turbine at the moment; The climbing rate for increased output; The efficiency of the combustion engine; , and They are all quadratic fitting coefficients between efficiency and output power.

[0062] For the storage battery, this embodiment uses a high-efficiency, high-energy-density lithium battery as an energy storage element. The charging state of the electric energy storage system determines its ability to absorb or provide electric energy. At any given moment, the charging state of the electric energy storage system is determined by the previous charging state and the charge and discharge amount, and needs to be limited to a certain range. In addition, the energy storage battery cannot be charged and discharged at the same time, and the charging and discharging power is limited to the maximum power range. The battery unit model is specifically:

[0063]

[0064]

[0065]

[0066]

[0067]

[0068] In the formula, is the charge state of the battery energy storage system at time t; For battery energy storage systems The charge state at the moment; It is the hour-level self-discharge efficiency; is the charging power; For charging efficiency; is the discharge power; is the discharge efficiency; is the time variation; It is the maximum power storage state; is the minimum charge state; is the maximum charge state.

[0069] This embodiment uses the desalination load as a controllable load to assist in adjusting the microgrid power. According to the water demand of island residents and the capacity limit of the reservoir, the power range of the desalination device required to meet the water demand of residents in a certain period in the future is calculated. The desalination load unit model is specifically as follows:

[0070]

[0071]

[0072] In the formula, is the electric power required by the seawater desalination system at time t; The amount of electricity required to produce a unit volume of fresh water; is the freshwater production rate at time t; is the total time period; The total amount of fresh water required per day.

[0073] For electric vehicle load, the prediction of electric vehicle charging load is the basis for the day-ahead dispatch of the microgrid. This embodiment extracts different numbers of samples to generate the start time of grid-connected charging of each sample electric vehicle according to the fitted probability distribution function. The electric vehicle load unit model is specifically as follows:

[0074]

[0075] In the formula, The probability density function of the daily charging probability of electric vehicles and the starting time; , are the parameters of the fitted normal distribution function.

[0076] Therefore, the total load of electric vehicle charging is determined by the number of electric vehicles charged at each moment and the charging time of the electric vehicles. The charging time is proportional to the daily driving distance of the electric vehicle. The calculation model is:

[0077]

[0078]

[0079]

[0080] In the formula, is the probability density function of the daily driving distance of electric vehicles; L is the driving distance of electric vehicles; Charging time for electric vehicles; and They are the power consumption per kilometer and rated charging power of electric vehicles; is the total load curve of electric vehicles; is the number of electric vehicles charged at time t.

[0081] In this embodiment, the distribution network can provide additional electric power for the multi-microgrid group, and the distribution network model is:

[0082]

[0083]

[0084] In the formula, The electrical power provided to the distribution network; The minimum value of electric power provided to the distribution network; The maximum value of electric power provided to the distribution network; is the total load curve of electric vehicles; The electrical power required for the desalination system; is the sequential output power of the micro gas turbine; is the photovoltaic output power; is the fan output power; is the interaction power between microgrid i and microgrid j; is the total load power in the distribution network.

[0085] S3. In the offline training stage, the historical multi-microgrid scenario dataset between islands is used for offline centralized optimization solution, and the minimization of total resource consumption and carbon emissions is used as the offline centralized optimization goal to generate the daily controllable unit output time series data of each microgrid under different natural resource conditions.

[0086] Due to the reliance of island microgrids on long-distance communications between islands and the mainland, as well as the variability of natural resource conditions, existing scheduling methods often find it difficult to achieve real-time and high efficiency. Figure 2 As shown, this embodiment integrates offline training and real-time online optimization of the decentralized dispatch optimization strategy of the island microgrid group to weaken the dependence on long-distance communication between the island and the mainland, and improve the real-time and high efficiency of electric energy mutual assistance. In the offline training stage, the system makes full use of the massive historical data accumulated by the microgrid group for centralized optimization and solution, obtains the daily controllable unit output time series data, and stores the output status of the adjustable unit and the corresponding natural resource conditions to form an offline training database. The offline training database provides a reference benchmark and rapid response capability for subsequent online optimization. This embodiment realizes each The optimal planning of the daily controllable unit output sequence of each microgrid is carried out, while ensuring that the optimization goal takes into account the minimization of total resource consumption and the effective reduction of carbon emissions. When entering the online optimization stage, the system first performs a similarity assessment based on the real-time data of the wind resources and photovoltaic resources of each microgrid, and performs online optimization and adjustment based on the controllable unit output in the daily controllable unit output sequence as the initial scheduling plan, ensuring the timeliness and adaptability of the scheduling plan, and being able to effectively cope with the changeable natural environment and power demand faced by the island microgrid group. In this embodiment, the construction process of the offline centralized optimization objective function in the offline training stage is:

[0087] According to the unit operating cost and output power of each energy type in each microgrid in each time period, the total resource consumption is obtained; the energy types include wind turbines, photovoltaics, micro gas turbines, energy storage batteries and distribution networks;

[0088] Calculate the total carbon emissions based on the unit carbon emission coefficient and output power of each energy type in each microgrid in each time period;

[0089] The power fluctuation variance of each microgrid is calculated based on the total output power and average power of each microgrid in each time period, and the power fluctuation variances of all time periods are summed and averaged to calculate the average power fluctuation;

[0090] Determine the power fluctuation range of each microgrid according to the maximum and minimum total output power of each microgrid, and obtain a power fluctuation stability index according to the average power fluctuation and the power fluctuation range;

[0091] The ramp stability of each microgrid is calculated based on the total output power change and the maximum allowable ramp rate of each microgrid in adjacent time periods;

[0092] Obtaining a system stability index according to the power fluctuation stability index and the climbing stability;

[0093] By minimizing the total resource consumption, the total carbon emissions and the system stability index, an offline centralized optimization objective function is constructed. In this embodiment, the offline centralized optimization objective function in the offline training phase is specifically:

[0094]

[0095] in,

[0096]

[0097]

[0098]

[0099]

[0100] In the formula, is the total resource consumption; is the total carbon emissions; It is the system stability indicator; is the carbon emission scaling factor; is the weight factor of the stability index; T is the total time period; N is the number of microgrids; is the unit operating cost of the energy storage battery; is the energy storage battery discharge power of microgrid i at time t; is the unit operating cost of the fan; is the wind turbine output power of microgrid i at time t; is the unit operating cost of the photovoltaic equipment; is the photovoltaic output power of microgrid i at time t; is the unit operating cost of the microturbine; is the output power of the micro gas turbine of microgrid i at time t; Unit operating costs of purchasing electricity for the distribution network; The electric power provided by the distribution network of microgrid i at time t; is the unit carbon emission coefficient of the energy storage battery; is the average discharge power; Adjust parameters for energy storage batteries; is the unit carbon emission coefficient of the fan; is the unit carbon emission coefficient of photovoltaic equipment; is the unit carbon emission coefficient of the micro gas turbine; is the unit carbon emission coefficient of the distribution network; is the sine function; cos is the cosine function; is the hyperbolic tangent function; is the total output power of microgrid i at time t; is the average power of microgrid i; is the power fluctuation variance; Indicates the average power fluctuation; the first summation item of the system stability index is the power fluctuation stability index; is climbing stability; For microgrid Total output power at the moment; is the time step; is the maximum allowable ramp rate of microgrid i.

[0101] The offline centralized optimization objective function adopted in the offline training stage of this embodiment finds a balance between resource consumption, carbon emissions and system stability. At the same time, the system stability index takes into account the fluctuations and climbing stability of multiple microgrids. It can more accurately simulate the operating characteristics of the microgrid group, thereby generating daily controllable unit output time series data that is more in line with actual operating conditions. This method helps to optimize the scheduling strategy of the microgrid group and achieve more efficient and environmentally friendly energy management.

[0102] S4. Based on the wind and solar data acquired in real time, the similarity measurement algorithm is used to search and find the most similar controllable unit output time series results in the daily controllable unit output time series data.

[0103] In this embodiment, the step of searching and finding the most similar controllable unit output time series result in the daily controllable unit output time series data by a similarity measurement algorithm based on the acquired real-time wind and solar data includes:

[0104] The characteristic patterns of wind and solar data between real-time wind and solar data and daily controllable unit output time series data are analyzed to obtain the potential correlation patterns between real-time wind and solar data and daily controllable unit output time series data;

[0105] A similarity index based on a potential association pattern is set, and the similarity of each controllable unit output time series data is calculated by using a similarity measurement method through the similarity index based on the potential association pattern; wherein the similarity index is determined according to the intersection of the discretized potential association pattern and the discretized controllable unit output time series data, and the attribute dimension of the discretized controllable unit output time series data;

[0106] Based on the similarity of each controllable unit output time series data, the controllable unit output time series data are sorted according to their similarity in ascending order using a sorting method, and the controllable unit output time series data with the highest similarity is used as the most similar controllable unit output time series result; wherein the similarity measurement algorithm is specifically:

[0107]

[0108]

[0109] In the formula, is the similarity of the controllable unit output time series data d, which integrates the differences of all microgrids i and all time points t, and is used to evaluate the similarity between the time series data and the real-time wind and solar data; N is the number of microgrids; T is the total time period; is a similarity index based on the underlying association pattern; is the similarity index adjustment parameter, which is a positive real number. and When they are not similar, this embodiment amplifies the difference through an exponential term; is the real-time output power of the wind turbine in microgrid i at time t; is the wind turbine output power in the output time series data of the dth controllable unit at time t of microgrid i; is the real-time output power of photovoltaic power in microgrid i at time t; is the photovoltaic output power in the output time series data of the dth controllable unit at time t of microgrid i; is the maximum output power of the wind turbine in microgrid i; is the maximum output power of photovoltaic in microgrid i, which is Used to normalize the real-time and reference output power so that the power data of different microgrids or different time periods can be compared on the same scale. This embodiment calculates the sum of the squares of the differences between the actual output power of the wind turbine and the photovoltaic power plant at time t and the corresponding values ​​in the output time series data of the controllable unit, and normalizes them by dividing by the maximum output power. The smaller the value obtained, the closer the actual output power is to the output time series data of the controllable unit. is the xth discretized potential association pattern; is the discrete controllable unit output time series data of the d-th controllable unit on microgrid i; To discretize the potential association pattern Discrete controllable unit output time series data of similar length; To discretize the potential association pattern Length; Discrete controllable unit output time series data Length; Discrete controllable unit output time series data Attribute dimension; , These are adjustment parameters.

[0110] It should be noted that wind and solar data and controllable unit output time series data usually contain a large amount of multi-dimensional information, and there are complex interactions and associations between these data. The existing similarity measurement calculation method based on Euclidean distance is only based on the similarity on the surface of the data, and cannot sensitively respond to small changes in the data. In particular, the subtle differences between real-time wind and solar data and controllable unit output time series data have a particularly important impact on the accuracy of subsequent online optimization. If the most similar results are directly obtained in wind and solar data and controllable unit output time series data through the existing similarity measurement method, the deep-level associations and characteristic patterns between the data are ignored, thereby affecting the accuracy of the final online distributed optimization results. Therefore, this embodiment ensures the accuracy of similarity calculation through feature analysis and the construction of similarity index, and avoids the ambiguity and uncertainty brought by direct measurement. Among them, when constructing the similarity index, this embodiment needs to discretize the potential association pattern and the controllable unit output time series data. The time series is divided into a series of time periods, and the data in each time period is taken as an element of the set to obtain a discretized potential correlation pattern and a discretized controllable unit output time series data, so as to quantify the similarity between the two sets by the ratio of the intersection similarity length to the square root of the product of their respective lengths. In this embodiment, a similarity index based on the potential correlation pattern is set in the decentralized scheduling optimization method of the weakly linked island microgrid group, and the similarity index is used to measure the similarity of the output time series data of each controllable unit to identify the controllable unit output pattern that is most similar to the real-time wind and solar data. The similarity measurement algorithm used in this embodiment can accurately capture subtle differences and has high flexibility, especially when it is necessary to process data with extreme values ​​or nonlinear characteristics, it can better predict the uncertainty factors such as the fluctuation of wind and solar resources in the system, so that these units can be given priority in the scheduling process, and then better match the real-time changes of wind and solar resources, thereby improving the operation efficiency and stability of the entire microgrid group.

[0111] S5. The output timing result of the most similar controllable unit is used as the initial scheduling value, and a penalty function is introduced for the output deviation of the controllable unit. Combined with the current operating status of the island microgrid group, each island multi-microgrid unit model is optimized and solved online to obtain the online distributed optimization scheduling plan of the current microgrid.

[0112] This embodiment realizes efficient dispatch and optimization of electric energy mutual assistance of island microgrid groups in offline training and online optimization stages according to the established power architecture and unit model. In actual operation, this embodiment performs online optimization on each microgrid separately. In order to make the optimization result tend to the result of lumped optimization, based on the output status of the most similar controllable unit output timing results, a penalty function is introduced into the optimization objective function for the output deviation of adjustable units (such as energy storage equipment and micro gas turbines), and combined with the specific load and operation status of the current microgrid, online distributed optimization solution is performed to adjust the output of the controllable units, so as to realize the efficiency and real-time performance of electric energy mutual assistance of island microgrid groups. In this embodiment, the process of obtaining the penalty function of the output deviation of the adjustable unit is specifically as follows:

[0113] According to the deviation between the actual output power of each adjustable unit and the reference output power, the corresponding relative deviation value is obtained, and the deviation rate is calculated according to the rated capacity of the adjustable unit and the relative deviation value;

[0114] Determine the deviation direction penalty value according to the actual output power of the adjustable unit and the reference output power of the output timing result of the most similar controllable unit;

[0115] According to the deviation rate and the deviation direction penalty value, a basic deviation penalty of the adjustable unit is obtained by using an output deviation penalty weight;

[0116] The continuous deviation penalty of the adjustable unit is calculated according to the continuous deviation time of the adjustable unit continuously deviating from the output timing result of the most similar controllable unit and the preset maximum continuous deviation time threshold;

[0117] The basic deviation penalty and continuous deviation penalty of the adjustable unit are integrated to obtain the output deviation penalty function of the adjustable unit.

[0118] Through the above steps, the adjustable unit output deviation penalty function can be obtained. In this embodiment, the controllable unit output deviation penalty function is introduced into the online distributed optimization objective function. The online distributed optimization objective function of each microgrid is specifically:

[0119]

[0120]

[0121]

[0122] In the formula, is the total resource consumption of microgrid i; is the carbon emission of microgrid i; is the carbon emission scaling factor; is the penalty function for the adjustable unit output deviation of microgrid i; is the function of electric energy exchange; N is the number of microgrids; T is the total time period; is the output power of microgrid i at time t; is the input power of microgrid j at time t; is the power grid efficiency fluctuation weight factor; is the time frequency of grid efficiency fluctuations; is the phase relationship of the grid efficiency fluctuation between microgrid i and microgrid j. This embodiment simulates the periodic fluctuation of grid efficiency through the time frequency and phase relationship, thereby reflecting the efficiency change of the grid under different time and conditions; is the actual transmission power from microgrid i to microgrid j at time t; is the time step; is the maximum capacity limit of the line from microgrid i to microgrid j at time t; is the rated capacity of the line from microgrid i to microgrid j; Relative deviation value. is the deviation rate; k is the index of the adjustable unit; is the set of all adjustable units in microgrid i; is the output deviation penalty weight of the adjustable unit k; is the actual output power of adjustable unit k at time t in microgrid i; is the rated capacity of the adjustable unit k; is the reference output power of the most similar controllable unit output timing result of adjustable unit k at time t in microgrid i; is the deviation direction penalty value; is the deviation threshold between the actual output power and the reference output power; is the deviation direction penalty coefficient of the adjustable unit k; is a sign function; the first sum term of the controllable unit output deviation penalty function represents the basic deviation penalty, and the second sum term represents the continuous deviation penalty; is the continuous deviation penalty coefficient of the adjustable unit k; is the preset maximum continuous deviation time threshold; It is the continuous deviation time of the adjustable unit k that continuously deviates from the output timing result of the most similar controllable unit before time t.

[0123] This embodiment achieves efficient and real-time scheduling of island microgrid groups through centralized offline training and online optimization scheduling, reduces dependence on long-distance island-to-land communications, and improves the efficiency and reliability of power mutual assistance. Among them, the online distributed optimization of this embodiment achieves distributed optimization of microgrids by optimizing all microgrids (referred to as microgrids) at the same time, and the optimization of each microgrid is independent. At the same time, based on the deviation size, duration, direction and unit characteristics of the adjustable unit deviation, penalties for power exchange and adjustable unit output deviation are designed to better simulate different situations in actual operation, so as to obtain a more comprehensive online distributed optimization objective function to simulate and optimize the operation of the microgrid, so that the system can respond quickly according to real-time data, and improve the flexibility and adaptability of scheduling.

[0124] S6. Perform inter-island multi-microgrid power mutual optimization scheduling according to the online distributed optimization scheduling scheme.

[0125] This embodiment verifies the effectiveness of the proposed method through actual examples, and proves the superiority of the method proposed in this embodiment in improving the efficiency and real-time performance of the electric energy mutual assistance of the island microgrid group. Figure 3 As shown, this embodiment is described by way of example by optimizing the dispatch of an island microgrid system consisting of four microgrids, and its optimization dispatch strategy is verified by example simulation. This embodiment sets the optimization period to 24 hours, the time interval to 1 hour, and sets the maximum power limit of the interconnection line between microgrids to 1400kW. In the offline training stage, this embodiment uses a data set of one year (365 days), and solves it through MATLAB programming combined with the solver Gurobi, and stores the solved time series output of the micro gas turbine and the electric energy storage device and the corresponding wind and solar resource data; in the real-time optimization stage, as shown in FIG. Figure 4 , Figure 5 As shown, this embodiment sets a penalty coefficient of 0.1 for further optimization based on the renewable energy output of the day and the load demand of each load island. The optimized interaction power between each microgrid is as follows: Figure 6 As shown in the figure, the results show that the interactive power between microgrids significantly reflects the advantages of resource complementarity: Microgrid 3 with abundant photovoltaic resources transmits a large amount of electricity to Microgrid 1 during the day (10:00 to 18:00), and uses the energy storage system to release the photovoltaic power stored during the day in the evening; while Microgrid 2 with a more balanced distribution of wind resources supplies power to Microgrid 1 throughout the day. This mutual assistance mode between microgrids allows Microgrid 1 to meet its load demand during most of the daytime hours (8:00 to 22:00) without purchasing electricity from the large power grid.

[0126] Regarding the scheduling analysis of resource islands and load islands, such as Figure 7 , Figure 8As shown in Figure 2, the dispatch strategies of resource-rich microgrids 2 and 3 show that the energy storage batteries are mainly charged during the photovoltaic peak period during the day (11:00 to 15:00) and discharged in the evening and at night (17:00 to 21:00), which effectively reduces the phenomenon of wind and solar power abandonment. In contrast, Fig. 9 , Fig.10 As shown, microgrids 1 and 4 with heavier loads reduce their dependence on micro gas turbines when photovoltaic power is sufficient during the day, and balance load demand through energy storage systems at night to ensure stable operation of the system. By comparing the results of independent optimization and distributed collaborative scheduling in this embodiment, it can be found that distributed collaborative scheduling significantly reduces the carbon emissions of the microgrid group (reduced by 3.0t), and reduces the total system cost by 28.9% by reducing the amount of interactive electricity with the large power grid; at the same time, under the same wind and light resources and load conditions, the distributed optimization method is compared with the centralized optimization method. Although the total cost deviation is extremely small (about 0.17%), the solution time is greatly shortened (from 1.899 seconds to 0.746 seconds), which proves the high efficiency and real-time advantages of distributed optimization in weakly connected island microgrid groups. In summary, the decentralized scheduling optimization strategy proposed in this embodiment can ensure the high efficiency and rapid response capability of scheduling in weakly connected island microgrid groups.

[0127] The embodiment of the present invention provides a method for decentralized scheduling optimization of weakly linked island microgrid groups. The method establishes an island multi-microgrid unit model by constructing an inter-island multi-microgrid electric energy mutual aid power architecture including a load island and a resource island; uses a historical inter-island multi-microgrid scenario data set to perform offline centralized optimization and solve, and takes minimizing total resource consumption and carbon emissions as offline centralized optimization goals to generate daily controllable unit output time series data for each microgrid under different natural resource conditions; searches and finds the most similar controllable unit output time series results in the daily controllable unit output time series data based on the acquired real-time wind and solar data through a similarity measurement algorithm; uses the most similar controllable unit output time series results as the initial scheduling value, introduces a penalty function for the controllable unit output deviation, and performs online distributed optimization and solves each island multi-microgrid unit model in combination with the current operating status of the island microgrid group to obtain an online distributed optimization scheduling plan for the current microgrid; and performs inter-island multi-microgrid electric energy mutual aid optimization scheduling according to the online distributed optimization scheduling plan. Compared with the prior art, this embodiment achieves efficient and real-time scheduling of island microgrids through a combination of offline training and online optimization, reduces dependence on long-distance island-land communications, and improves the efficiency and reliability of power mutual assistance.

[0128] It should be noted that the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0129] In one embodiment, Fig.11 As shown, an embodiment of the present invention provides a decentralized scheduling optimization system for weakly linked island microgrids, the system comprising:

[0130] The power architecture building module 101 is used to build an inter-island multi-microgrid power energy mutual aid power architecture including load islands and resource islands according to the geographical location and resource distribution data of the island microgrid group;

[0131] The unit model building module 102 is used to build an island multi-microgrid unit model based on the island multi-microgrid electric energy mutual aid power architecture;

[0132] The offline training optimization module 103 is used to perform offline centralized optimization solution using the historical multi-microgrid scenario data set between islands during the offline training phase, and to minimize total resource consumption and carbon emissions as offline centralized optimization goals, thereby generating daily controllable unit output time series data for each microgrid under different natural resource conditions;

[0133] The similarity measurement module 104 is used to search and find the most similar controllable unit output time series results in the daily controllable unit output time series data according to the acquired real-time wind and solar data through a similarity measurement algorithm;

[0134] The online optimization solution module 105 is used to use the output timing result of the most similar controllable unit as the initial scheduling value, introduce a penalty function for the output deviation of the controllable unit, and perform online distributed optimization solution on each of the island multi-microgrid unit models in combination with the current operating state of the island microgrid group to obtain an online distributed optimization scheduling plan for the current microgrid;

[0135] The microgrid optimization scheduling module 106 is used to perform the multi-microgrid power mutual optimization scheduling between islands according to the online distributed optimization scheduling scheme.

[0136] For the specific limitations of a decentralized scheduling optimization system for a weakly linked island microgrid group, please refer to the above-mentioned limitations on a decentralized scheduling optimization method for a weakly linked island microgrid group, which will not be repeated here. A person of ordinary skill in the art will appreciate that the various modules and steps described in conjunction with the embodiments disclosed in this application can be implemented in hardware, software, or a combination of both. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0137] The embodiment of the present invention provides a decentralized dispatch optimization system for weakly linked island microgrids. The system constructs an inter-island multi-microgrid electric energy mutual aid power architecture including load islands and resource islands through a power architecture construction module; establishes an island multi-microgrid unit model through a unit model establishment module; an offline training optimization module uses a historical inter-island multi-microgrid scenario data set to perform offline centralized optimization and solve, and takes minimizing total resource consumption and carbon emissions as offline centralized optimization goals to generate daily controllable unit output time series data for each microgrid under different natural resource conditions; a similarity measurement module is based on The real-time wind and solar data obtained are searched and found in the daily controllable unit output time series data through the similarity measurement algorithm; the online optimization solution module uses the most similar controllable unit output time series result as the initial scheduling value, and introduces a penalty function for the controllable unit output deviation, and combines the current operating status of the island microgrid group to perform online distributed optimization and solution on each island multi-microgrid unit model to obtain the online distributed optimization scheduling plan of the current microgrid; the microgrid optimization scheduling module performs multi-microgrid power mutual assistance optimization scheduling between islands according to the online distributed optimization scheduling plan. Compared with the prior art, this embodiment realizes efficient and real-time scheduling of island microgrid groups through the combination of offline training and online optimization, reduces the dependence on long-distance communication between islands and land, and improves the efficiency and reliability of power mutual assistance.

[0138] The above-mentioned embodiments only express several preferred implementation modes of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in the technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be based on the protection scope of the claims.

Claims

1. A decentralized scheduling optimization method for weakly linked island microgrids, characterized in that: The following steps are involved: According to the geographical location and resource distribution data of the island microgrid group, a multi-microgrid power mutual aid power architecture between islands, including load islands and resource islands, is constructed; Based on the inter-island multi-microgrid power energy mutual aid power architecture, an island multi-microgrid unit model is established; In the offline training phase, the historical multi-microgrid scenario dataset between islands is used for offline centralized optimization, and the minimization of total resource consumption and carbon emissions is used as the offline centralized optimization goal to generate the daily controllable unit output time series data of each microgrid under different natural resource conditions; According to the real-time wind and solar data obtained, the most similar controllable unit output time series results are searched and found in the daily controllable unit output time series data through the similarity measurement algorithm; The output timing result of the most similar controllable unit is used as the initial dispatch value, and a penalty function is introduced for the output deviation of the controllable unit. Combined with the current operating state of the island microgrid group, each island multi-microgrid unit model is optimized and solved online to obtain the online distributed optimization dispatching scheme of the current microgrid. Performing inter-island multi-microgrid power mutual optimization scheduling according to the online distributed optimization scheduling scheme; The step of searching and finding the most similar controllable unit output time series results in the daily controllable unit output time series data by a similarity measurement algorithm based on the acquired real-time wind and solar data includes: The characteristic patterns of wind and solar data between real-time wind and solar data and daily controllable unit output time series data are analyzed to obtain the potential correlation patterns between real-time wind and solar data and daily controllable unit output time series data; A similarity index based on a potential correlation pattern is set, and the similarity degree of the output time series data of each controllable unit is calculated by using a similarity measurement algorithm through the similarity index based on the potential correlation pattern; Based on the similarity of each controllable unit output time series data, the controllable unit output time series data are sorted according to their similarity in ascending order using a sorting method, and the controllable unit output time series data with the highest similarity is used as the most similar controllable unit output time series result; wherein the similarity measurement algorithm is specifically: In the formula, is the similarity of the controllable unit output time series data d; N is the number of microgrids; T is the total time period; is a similarity index based on the underlying association pattern; Tuning parameters for similarity index; is the real-time output power of the wind turbine in microgrid i at time t; is the wind turbine output power in the output time series data of the dth controllable unit at time t of microgrid i; is the real-time output power of photovoltaic power in microgrid i at time t; is the photovoltaic output power in the output time series data of the dth controllable unit at time t of microgrid i; is the maximum output power of the wind turbine in microgrid i; is the maximum output power of photovoltaic in microgrid i; is the xth discretized potential association pattern; is the discrete controllable unit output time series data of the d-th controllable unit on microgrid i; To discretize the potential association pattern Discrete controllable unit output time series data of similar length; To discretize the potential association pattern Length; Discrete controllable unit output time series data Length; Discrete controllable unit output time series data Attribute dimension; , These are adjustment parameters.

2. A method for decentralized scheduling optimization of weakly linked island microgrids as claimed in claim 1, characterized in that: The island multi-microgrid unit model includes an offshore wind power unit model and a photovoltaic unit model. The steps of establishing the island multi-microgrid unit model based on the island multi-microgrid electric energy mutual aid power architecture include: According to the inter-island multi-microgrid electric energy mutual aid power architecture, the wind and solar resource generation and electric energy transmission process between different islands are simulated to obtain island power simulation data; the island power simulation data includes basic data of offshore wind farms and basic data of photovoltaics; The wind speed at the offshore wind turbine hub height is calculated based on the basic data of the offshore wind farm; the basic data of the offshore wind farm includes the wind speed at the reference height of the offshore wind farm, the height of the offshore wind turbine hub and the surface roughness of the actual ocean state; In different wind speed ranges, the relationship between wind speed at the hub height of offshore wind turbine and wind turbine output power is fitted using piecewise function to obtain the offshore wind power unit model. According to the photovoltaic basic data, the photovoltaic power generation efficiency under the standard state is adjusted by using a preset photovoltaic temperature coefficient to obtain the current actual photovoltaic power generation efficiency; the photovoltaic basic data includes the current ambient temperature and the global horizontal irradiance; The photovoltaic output power is calculated based on the photovoltaic plane area, the actual photovoltaic power generation efficiency and the global horizontal irradiance, and the photovoltaic unit model is constructed.

3. A method for optimizing decentralized scheduling of weakly linked island microgrids as claimed in claim 2, characterized in that: The island multi-microgrid unit model also includes a micro gas turbine unit model, a battery unit model and a distribution network model, and the island power simulation data includes the island power system energy storage demand and the island power demand; The step of establishing an island multi-microgrid unit model based on the island multi-microgrid electric energy mutual aid power architecture further includes: According to the power demand and resource conditions of the island, the operating parameters of the micro gas turbine are determined, and the relationship between the output power and efficiency of the micro gas turbine is fitted according to the operating parameters of the micro gas turbine to obtain a micro gas turbine unit model; Determine the battery capacity and charging and discharging parameters according to the energy storage demand of the island power system, and use the charge state of the battery as a state variable, and update the charge state of each time step by considering the charging and discharging parameters and the time variation to obtain a battery unit model; wherein the charging and discharging parameters include charging efficiency and discharging efficiency; The interaction power between different microgrids is obtained, and the distribution network model is constructed according to the sum of the interaction power of all microgrids in the distribution network and the distributed resource power within the microgrid.

4. A method for decentralized scheduling optimization of weakly linked island microgrids as claimed in claim 1, characterized in that: The construction process of the offline centralized optimization objective function in the offline training phase is: The total resource consumption is obtained based on the unit operating cost and output power of each energy type in each microgrid in each time period; Calculate the total carbon emissions based on the unit carbon emission coefficient and output power of each energy type in each microgrid in each time period; The power fluctuation variance of each microgrid is calculated based on the total output power and average power of each microgrid in each time period, and the power fluctuation variances of all time periods are summed and averaged to calculate the average power fluctuation; Determine the power fluctuation range of each microgrid according to the maximum and minimum total output power of each microgrid, and obtain a power fluctuation stability index according to the average power fluctuation and the power fluctuation range; The ramp stability of each microgrid is calculated based on the total output power change and the maximum allowable ramp rate of each microgrid in adjacent time periods; Obtaining a system stability index according to the power fluctuation stability index and the climbing stability; By minimizing the total resource consumption, the total carbon emissions and the system stability index, an offline centralized optimization objective function is constructed.

5. A method for optimizing decentralized scheduling of weakly linked island microgrids as claimed in claim 4, characterized in that: The energy types include wind turbines, photovoltaics, micro gas turbines, energy storage batteries and distribution networks.

6. A method for optimizing decentralized scheduling of weakly linked island microgrids as claimed in claim 1, characterized in that: The specific objective function of online distributed optimization is: In the formula, is the total resource consumption of microgrid i; is the carbon emission of microgrid i; is the carbon emission scaling factor; is the penalty function for the adjustable unit output deviation of microgrid i; is the function of electric energy exchange; N is the number of microgrids; T is the total time period; is the output power of microgrid i at time t; is the input power of microgrid j at time t; is the power grid efficiency fluctuation weight factor; is the time frequency of grid efficiency fluctuations; is the phase relationship of grid efficiency fluctuation between microgrid i and microgrid j; is the actual transmission power from microgrid i to microgrid j at time t; is the time step; is the maximum capacity limit of the line from microgrid i to microgrid j at time t; is the rated capacity of the line from microgrid i to microgrid j.

7. A method for optimizing decentralized scheduling of weakly linked island microgrids as claimed in claim 6, characterized in that: The process of obtaining the adjustable unit output deviation penalty function is as follows: According to the deviation between the actual output power of each adjustable unit and the reference output power, the corresponding relative deviation value is obtained, and the deviation rate is calculated according to the rated capacity of the adjustable unit and the relative deviation value; Determine the deviation direction penalty value according to the actual output power of the adjustable unit and the reference output power of the output timing result of the most similar controllable unit; According to the deviation rate and the deviation direction penalty value, a basic deviation penalty of the adjustable unit is obtained by using an output deviation penalty weight; The continuous deviation penalty of the adjustable unit is calculated according to the continuous deviation time of the adjustable unit continuously deviating from the output timing result of the most similar controllable unit and the preset maximum continuous deviation time threshold; The basic deviation penalty and continuous deviation penalty of the adjustable unit are integrated to obtain the output deviation penalty function of the adjustable unit.

8. A decentralized dispatch optimization system for weakly linked island microgrids, characterized in that: The system comprises: The power architecture building module is used to build a multi-microgrid power architecture among islands, including load islands and resource islands, based on the geographical location and resource distribution data of the island microgrid group; A unit model building module, used to build an island multi-microgrid unit model based on the island multi-microgrid electric energy mutual aid power architecture; The offline training optimization module is used to perform offline centralized optimization solutions using the historical multi-microgrid scenario dataset between islands during the offline training phase, and to generate daily controllable unit output time series data for each microgrid under different natural resource conditions, taking minimizing total resource consumption and carbon emissions as the offline centralized optimization goal; A similarity measurement module is used to search and find the most similar controllable unit output time series results in the daily controllable unit output time series data based on the acquired real-time wind and solar data through a similarity measurement algorithm; An online optimization solution module is used to use the output timing result of the most similar controllable unit as the initial scheduling value, introduce a penalty function for the output deviation of the controllable unit, and perform online distributed optimization solution on each of the island multi-microgrid unit models in combination with the current operating status of the island microgrid group to obtain an online distributed optimization scheduling plan for the current microgrid; A microgrid optimization scheduling module, used for performing multi-microgrid power mutual optimization scheduling among islands according to the online distributed optimization scheduling scheme; Wherein, the similarity measurement module is specifically used for: The characteristic patterns of wind and solar data between real-time wind and solar data and daily controllable unit output time series data are analyzed to obtain the potential correlation patterns between real-time wind and solar data and daily controllable unit output time series data; A similarity index based on a potential correlation pattern is set, and the similarity degree of the output time series data of each controllable unit is calculated by using a similarity measurement algorithm through the similarity index based on the potential correlation pattern; Based on the similarity of each controllable unit output time series data, the controllable unit output time series data are sorted according to their similarity in ascending order using a sorting method, and the controllable unit output time series data with the highest similarity is used as the most similar controllable unit output time series result; wherein the similarity measurement algorithm is specifically: In the formula, is the similarity of the controllable unit output time series data d; N is the number of microgrids; T is the total time period; is a similarity index based on the underlying association pattern; Tuning parameters for similarity index; is the real-time output power of the wind turbine in microgrid i at time t; is the wind turbine output power in the output time series data of the dth controllable unit at time t of microgrid i; is the real-time output power of photovoltaic power in microgrid i at time t; is the photovoltaic output power in the output time series data of the dth controllable unit at time t of microgrid i; is the maximum output power of the wind turbine in microgrid i; is the maximum output power of photovoltaic in microgrid i; is the xth discretized potential association pattern; is the discrete controllable unit output time series data of the d-th controllable unit on microgrid i; To discretize the potential association pattern Discrete controllable unit output time series data of similar length; To discretize the potential association pattern Length; Discrete controllable unit output time series data Length; Discrete controllable unit output time series data Attribute dimension; , These are adjustment parameters.

Images