Wind-light-electricity coordinated scheduling method and device, electronic equipment and storage medium

Abstract

The invention discloses a wind-light-electricity coordinated scheduling method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring a load demand predicted quantity of a current power consumption unit when a scheduling opportunity is reached; determining a target scheduling strategy of wind-light-electricity scheduling according to the load demand prediction quantity of the power consumption unit, a preset wind-light-electricity coordination model and constraint conditions; performing wind, light and electricity scheduling according to scheduling parameters included in the target scheduling strategy; wherein the scheduling parameters comprise one or more of the following items: the output of the self-generating unit, the purchased electric quantity and the sold electric quantity; the self-generating unit comprises a wind power generation unit and / or a photovoltaic power generation unit; an objective function included in the wind-light-electricity coordination model is a first weighted sum of one or more of the following sub-objective functions: a lowest energy consumption sub-objective function, a lowest economic cost sub-objective function and a self-generation waste rate sub-objective function.

Description

Technical Field

[0001] The present disclosure relates to, but is not limited to, the field of automated scheduling technology, and in particular to a method, device, electronic device, and storage medium for coordinated scheduling of wind and solar power. Background Art

[0002] Energy is one of the main material foundations supporting the development of modern civilized society. Every energy revolution in history has greatly promoted the development of modern civilization. With the increasing depletion of traditional mineral energy such as oil and coal, the development and utilization of new environmentally friendly renewable energy is imminent. The development and production process of oil and gas fields requires a large amount of heat and electricity, and it is urgent to reduce energy consumption and replace it with clean energy. The main working idea of ​​clean substitution in oil and gas fields is "replacing heat with heat and electricity with electricity", that is, replacing grid electricity or oil field coal-fired and gas-fired self-generated electricity with green electricity such as wind power and photovoltaic power generation. However, wind and solar power generation has the characteristics of randomness and intermittency, and it is difficult to ensure the reliability of electricity use in oil and gas fields. Therefore, how to ensure the reliability of electricity use in oil and gas fields has become an urgent problem to be solved in this field. Summary of the invention

[0003] The disclosed embodiments provide a method, device, electronic device and storage medium for coordinated dispatch of wind and solar power. According to the load demand forecast of the power consumption unit, the preset wind and solar power coordination model and the constraint conditions, the target dispatching strategy of wind and solar power dispatching is determined, and the power supply mode combining self-generated power and purchased power is scientifically and rationally coordinated for the relevant production systems to overcome the uncertainty and instability of wind and solar power self-generation, and reduce the overall energy consumption under the premise of improving the power reliability of the own system.

[0004] The embodiment of the present disclosure provides a wind-solar-electricity coordinated scheduling method, comprising:

[0005] When the dispatching time is reached, the load demand forecast of the current power consumption unit is obtained;

[0006] Determine the target dispatching strategy of wind and solar power dispatching according to the load demand forecast of the power consumption unit, the preset wind and solar power coordination model and the constraint conditions;

[0007] Perform wind and solar power dispatching according to the dispatching parameters included in the target dispatching strategy;

[0008] The dispatching parameters include one or more of the following: the output of the self-generating unit, the amount of electricity purchased and the amount of electricity sold; the self-generating unit includes: a wind power generation unit and / or a photovoltaic power generation unit;

[0009] The objective function included in the wind-solar-electricity coordination model is the first weighted sum of the following one or more sub-objective functions: a minimum energy consumption sub-objective function, a minimum economic cost sub-objective function and a self-generation abandonment rate sub-objective function.

[0010] The embodiment of the present disclosure also provides a wind-solar-electricity coordinated scheduling device, comprising:

[0011] A load forecasting module is configured to obtain a load demand forecast of a current power consumption unit when a scheduling opportunity is reached;

[0012] A dispatching strategy determination module, configured to determine a target dispatching strategy for wind and solar power dispatching according to a load demand forecast of the power consumption unit, a preset wind and solar power coordination model and constraint conditions;

[0013] A scheduling module, configured to perform wind and solar power scheduling according to the scheduling parameters included in the target scheduling strategy;

[0014] The dispatching parameters include one or more of the following: the output of the self-generating unit, the amount of electricity purchased and the amount of electricity sold; the self-generating unit includes: a wind power generation unit and / or a photovoltaic power generation unit;

[0015] The objective function included in the wind-solar-electricity coordination model is the first weighted sum of the following one or more sub-objective functions: a minimum energy consumption sub-objective function, a minimum economic cost sub-objective function and a self-generation abandonment rate sub-objective function.

[0016] The present disclosure also provides an electronic device, including:

[0017] one or more processors;

[0018] a storage device for storing one or more programs,

[0019] When the one or more programs are executed by the one or more processors, the one or more processors implement the wind-photovoltaic coordinated scheduling method as described in any embodiment of the present disclosure.

[0020] The embodiments of the present disclosure further provide a computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the wind-photovoltaic coordinated scheduling method as described in any embodiment of the present disclosure is implemented.

[0021] Other aspects will be apparent upon reading and understanding the drawings and detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be 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 the structures shown in these drawings without paying creative work.

[0023] Figure 1 This is a flow chart of a wind-solar-electricity coordinated scheduling in an embodiment of the present disclosure;

[0024] Figure 2 A schematic diagram of a process for obtaining load demand forecast in an embodiment of the present disclosure;

[0025] Figure 3 A schematic diagram of a neural network training process in an embodiment of the present disclosure;

[0026] Figure 4 This is another schematic diagram of a neural network training process in an embodiment of the present disclosure;

[0027] Figure 5 This is another schematic diagram of a neural network training process in an embodiment of the present disclosure;

[0028] Figure 6 It is a structural schematic diagram of a wind-photovoltaic coordinated scheduling device in an embodiment of the present disclosure.

[0029] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] It should be noted that in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0032] In the various embodiments of the present application, it should be understood that the size of the serial numbers of the following 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.

[0033] It should be understood that the term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0034] In the embodiments provided in the present application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined according to A. However, it should also be understood that determining B according to A does not mean determining B only according to A, and B can also be determined according to A and / or other information.

[0035] In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0036] The present disclosure provides a method for coordinated dispatching of wind and solar power. Figure 1 As shown, including

[0037] Step 110, when the scheduling opportunity is reached, obtaining the load demand forecast of the current power consumption unit;

[0038] Step 120, determining a target dispatching strategy for wind and solar power dispatching according to the load demand forecast of the power consumption unit, a preset wind and solar power coordination model and constraint conditions;

[0039] Step 130, performing wind and solar power scheduling according to the scheduling parameters included in the target scheduling strategy;

[0040] The dispatching parameters include one or more of the following: the output of the self-generating unit, the amount of electricity purchased and the amount of electricity sold; the self-generating unit includes: a wind power generation unit and / or a photovoltaic power generation unit;

[0041] The objective function included in the wind-solar-electricity coordination model is the first weighted sum of the following one or more sub-objective functions: a minimum energy consumption sub-objective function, a minimum economic cost sub-objective function and a self-generation abandonment rate sub-objective function.

[0042] The self-generating unit includes a wind power generating unit and / or a photovoltaic power generating unit. Accordingly, the output of the self-generating unit includes: the output of the wind power generating unit and / or the output of the photovoltaic power generating unit.

[0043] It can be understood that the self-generating unit is a unit in the production system that can generate electricity by itself to supply power to the power consumption units (power users) of the production system, or sell electricity to the outside.

[0044] It can be understood that the minimum energy consumption sub-objective function refers to the minimum energy consumption calculated according to the objective function; the minimum economic cost sub-objective function refers to the minimum economic cost calculated according to the objective function; the self-generation abandonment rate sub-objective function is used to calculate the self-generation abandonment rate. The self-generation abandonment rate is the ratio of the first difference obtained by subtracting the load demand forecast of the power consumption unit from the sum of the output of the wind power generation unit and the output of the photovoltaic power generation unit, and the second sum obtained by adding the output of the wind power generation unit to the output of the photovoltaic power generation unit.

[0045] In some exemplary embodiments, the method is applied to the oil and gas field production system. Since a large amount of heat and electricity is consumed in the oil and gas field production process, in order to achieve clean substitution, the oil and gas field production system is equipped with self-generating units such as wind power and photovoltaic power to replace the power grid or the oil field coal-fired and gas-fired self-generated power; when the self-generated power is insufficient, it can be supplemented by purchasing electricity from outside, and when the self-generated power exceeds the self-consumption requirements, the excess electricity can be sold to obtain additional economic benefits. Therefore, the power network of the oil and gas field system is also called the wind and solar power grid, and accordingly, its power dispatch is called wind and photovoltaic dispatch, and its power coordination model is called the wind and photovoltaic coordination model.

[0046] It can be seen that due to the randomness and intermittency of wind and solar power generation, it is difficult to ensure the reliability of electricity production in oil and gas fields. Therefore, the method described in the embodiment of the present disclosure is proposed to determine the target scheduling strategy of wind and solar power scheduling based on the load demand forecast of the power consumption unit in the oil and gas field, the preset wind and solar power coordination model and constraints to carry out wind power generation and / or photovoltaic power generation in the oil and gas field system, as well as the coordinated scheduling of purchased electricity and sold electricity, so as to ensure the reliability of electricity consumption in the oil and gas field and minimize carbon emissions.

[0047] It should be noted that this article records the wind and solar power coordination scheduling scheme by taking the self-generation, external purchase of electricity and external sales of electricity scheduling in the oil and gas field production system as an example, but it is not limited to the oil and gas field production system. If other production systems involve the coordinated scheduling of wind and solar self-generation, electricity consumption, external purchase of electricity and / or external sales of electricity, the scheme described in the embodiment of the present disclosure can be adopted. Among them, the power consumption unit (also called power user) in the oil and gas field production system includes the unit in the oil and gas field that needs to use electricity for oil and gas production and / or related auxiliary / management / operation work, which can be an oil well, a pressure station, a separation station or an oil and gas field monitoring room, etc., and is not limited to specific aspects. When the scheme of this application is applied to other production systems, the specific power consumption unit included corresponds to the power use unit therein, and is not limited to the above-mentioned specific power consumption unit of the example of this disclosure. The equipment / system range corresponding to a power consumption unit is defined according to the production scheduling needs in the production system. For example, a power consumption unit includes two oil wells, which is not limited to specific aspects. The specific demarcation method is not discussed in detail in this application.

[0048] In some exemplary embodiments, the power consumption unit includes one or more power consumption units. Accordingly, in the case of including multiple power consumption units, the load demand forecast amount of the power consumption unit includes the load demand forecast amounts of the multiple power consumption units.

[0049] It should be noted that the embodiments of the present disclosure take the wind and solar power dispatching in the oil and gas field production system as an example to record the implementation details of related aspects. For other production systems, other types of power consumption units can be used, and are not limited to the specific aspects of the examples of the present disclosure.

[0050] It can be understood that the wind and solar power scheduling scheme provided by the embodiment of the present disclosure can coordinate the self-generated electricity, purchased electricity, and sold electricity in the overall production system to meet its own electricity consumption needs and ensure electricity reliability.

[0051] In some exemplary embodiments, the wind power generation unit includes a wind turbine. Taking an oil and gas field as an example, the wind power generation unit may include one or more wind turbines, that is, an oil and gas field includes one or more wind turbines to generate electricity, and the generated electricity is used for production in the oil and gas field, or the excess electricity can be incorporated into the power grid for external sale.

[0052] It can be understood that in the above-mentioned coordinated scheduling method, the scheduling parameters include: at least one of the output of the wind power generation unit, the output of the photovoltaic power generation unit, the purchased electricity, and the sold electricity. In actual application, adaptive scheduling can be performed according to the photovoltaic power generation equipment and / or wind power generation equipment specifically configured in the oil and gas field production system and whether it is possible to sell electricity to the outside, that is, the photovoltaic output is scheduled only when there is a photovoltaic power generation equipment, the wind turbine output is scheduled only when there is a wind power generation equipment, and the sold electricity is scheduled only when the production plant can sell electricity.

[0053] In some exemplary embodiments, the photovoltaic power generation unit includes a photovoltaic power station. Taking an oil and gas field as an example, the photovoltaic power generation unit may include one or more photovoltaic power stations, that is, an oil and gas field includes one or more photovoltaic power stations for self-generation, and the generated electricity is used for production in the oil and gas field, or the excess electricity can be incorporated into the power grid for external sale.

[0054] Taking an oil and gas field as an example, its self-generating unit includes a wind turbine, or a photovoltaic power station, or a wind turbine and a photovoltaic power station, without limitation to a specific aspect.

[0055] In some exemplary embodiments, the scheduling opportunity includes any of the following opportunities:

[0056] The scheduling time point of periodic scheduling arrives;

[0057] An instruction from an operator or external system to initiate a periodic schedule is received.

[0058] In some exemplary embodiments, the periodic scheduling includes one or more scheduling periods, each of which sets a scheduling period.

[0059] In some exemplary embodiments, the scheduling cycle is a set number of days, that is, scheduling is performed periodically every set number of days; this type of scheduling method is also called day-ahead scheduling, that is, the target scheduling strategy determined by this method will take effect in the subsequent set number of days.

[0060] In some exemplary embodiments, the scheduling cycle is a set number of hours, that is, scheduling is performed periodically every set number of hours; this type of scheduling method is also called intraday scheduling, that is, the target scheduling strategy determined by this method will take effect in the subsequent set hours.

[0061] In some exemplary embodiments, the scheduling period is a set number of minutes, that is, scheduling is performed periodically every set number of minutes; this type of scheduling method is also called real-time scheduling, that is, the target scheduling strategy determined by this method will take effect in the subsequent set number of minutes.

[0062] In some exemplary embodiments, each scheduling period includes a plurality of data sampling time points at set intervals, and data is collected at each sampling time point.

[0063] The collected data includes one or more of the following:

[0064] P wq (t), P vj (t),ω(t),p(t),P grid (t), P sell (t), P buy (t).

[0065] Among them, P wq (t) is the fan output at time t, P vj (t) is the output of the photovoltaic power station at time t, ω(t) is the line fluctuation price at time t, p(t) is the electricity price at time t, P grid (t) is the Internet power of all nodes at time t, P sell (t) is the electricity sold at time t, P buy (t) is the purchased electric power at time t.

[0066] In some exemplary embodiments, the collected data also includes:

[0067] F k (t);

[0068] Among them, F k (t) is the load demand forecast of the kth power consumption unit at time t.

[0069] It can be seen that the specific data collected at each sampling time point is determined according to the objective function and / or constraint conditions included in the wind-solar-electricity coordination model, and is not limited to specific aspects.

[0070] In some exemplary embodiments, when each scheduling period includes multiple time sampling points for data collection, a set sampling data processing method is used to determine data for the objective function and / or corresponding constraint conditions based on the data collected at the multiple time sampling points;

[0071] The set sampling data processing method includes any one of the following:

[0072] Mean method, median method.

[0073] Optionally, other processing methods can be used for the sampling data at multiple sampling time points to determine more representative sampling result data, so as to bring in the objective function and / or corresponding constraints to determine the scheduling parameters included in the target scheduling strategy for the corresponding wind and solar power scheduling. More processing methods are not limited to the aspects of the examples disclosed herein and can be selected according to the characteristics of the data.

[0074] For example, in the day-ahead scheduling, the sampling interval is set to 1 hour, or other durations. For example, in the day-ahead scheduling with a period of 1 day, data sampling is performed every hour. When step 120 is executed, the data to be brought into the objective function and / or corresponding constraints are determined by the sampling mean method based on the 24 groups of sampled data included in 1 day, and the scheduling parameters corresponding to the day-ahead scheduling are determined.

[0075] For example, in the intraday scheduling, the sampling interval is set to 15 minutes, or other durations. For example, in the intraday scheduling with a period of 1 hour, data sampling is performed every 15 minutes. When step 120 is executed, the data to be brought into the objective function and / or corresponding constraints are determined by the sampling mean method based on the 4 groups of sampled data included in 1 hour, and the scheduling parameters corresponding to the intraday scheduling are determined.

[0076] For example, in the real-time scheduling, the sampling interval is set to 5 minutes, or other durations. For example, in the real-time scheduling with a period of 15 minutes, data sampling is performed every 5 minutes. When executing step 120, based on the three groups of sampled data included in 15 minutes, the sampling mean method is used to determine the data to be brought into the objective function and / or the corresponding constraints, and determine the scheduling parameters corresponding to the real-time scheduling.

[0077] More examples of scheduling cycles and / or sampling cycles are not listed here one by one. The duration of the scheduling cycle and the duration of the sampling interval can be determined as needed and are not limited to specific aspects.

[0078] In some exemplary embodiments, each periodic schedule (event) includes an attribute: a schedule period.

[0079] In some exemplary embodiments, each periodic schedule (event) further includes one or more of the following attributes: the start time and end time of the effective time period, and the execution time point. The specific attribute value is a preset value and can be flexibly set according to needs.

[0080] In some exemplary embodiments, the periodic scheduling may include one or more scheduling periods, respectively set. When multiple periodic scheduling are included, the arrival of the scheduling opportunity will be determined according to the respective set scheduling periods, and then the subsequent scheduling tasks will be executed.

[0081] In some exemplary embodiments, the periodic scheduling includes: day-ahead scheduling,

[0082] The method further comprises:

[0083] When the scheduling timing of the day-ahead scheduling is reached, the grid connection point configuration data is determined, and the grid connection point is configured.

[0084] In some exemplary embodiments, the periodic scheduling includes: day-ahead scheduling and intra-day scheduling;

[0085] The method further comprises:

[0086] When the scheduling opportunity of the day-ahead scheduling is reached, determining the grid connection point configuration data and configuring the grid connection point;

[0087] The situation where the scheduling opportunity is reached includes: reaching the scheduling opportunity for the intraday scheduling; that is, when the day-ahead scheduling arrives, the grid connection point configuration data is determined, and the grid connection point is configured, and when the intraday scheduling opportunity arrives, wind and solar power scheduling is carried out.

[0088] It can be understood that in this embodiment, when the scheduling opportunity corresponding to the day-ahead scheduling arrives, the grid connection point configuration data is determined and configured, and when the scheduling opportunity corresponding to the intraday scheduling arrives, the target scheduling strategy is determined and the wind and solar power grids are scheduled.

[0089] In some exemplary embodiments, the periodic scheduling includes: day-ahead scheduling and real-time scheduling;

[0090] The method further comprises:

[0091] When the scheduling opportunity of the day-ahead scheduling is reached, determining the grid connection point configuration data and configuring the grid connection point;

[0092] The situation where the scheduling opportunity is reached includes: reaching the scheduling opportunity of the real-time scheduling; that is, when the day-ahead scheduling arrives, the grid connection point configuration data is determined, and the grid connection point is configured, and when the real-time scheduling opportunity arrives, wind and solar power scheduling is performed.

[0093] It can be understood that in this embodiment, when the scheduling opportunity corresponding to the day-ahead scheduling arrives, the grid connection point configuration data is determined and configured, and when the scheduling opportunity corresponding to the real-time scheduling arrives, the target scheduling strategy is determined and the wind and solar power grids are scheduled.

[0094] In some exemplary embodiments, the method further comprises:

[0095] Configure the grid connection points according to the preset grid connection point configuration data.

[0096] Among them, the method for determining the grid connection point configuration data, the specific content of the grid connection point configuration data, and the configuration method are executed according to relevant feasible solutions and are not discussed in detail in the embodiments of this application.

[0097] It can be seen that the determination and configuration of the grid connection point configuration data can be flexibly determined according to needs, and is not limited to being predetermined or periodically dynamically determined.

[0098] When determining that the current scheduling opportunity has been reached, the power coordination scheduling scheme provided by the embodiment of the present disclosure first obtains the load demand forecast of the power consumption unit at the current moment in order to ensure the correctness of the scheduling strategy adopted during scheduling and to ensure that it is consistent with the actual situation, wherein the load demand forecast corresponds to the preset scheduling period corresponding to the timing of the triggered scheduling. Then, the load demand forecast and the predetermined preset constraint conditions are brought into the preset wind-photovoltaic coordination model, and the target scheduling strategy including the corresponding scheduling parameters can be determined by running the preset wind-photovoltaic coordination model, and then the optimized scheduling is performed according to the target scheduling strategy.

[0099] It should be noted that the objective function in the wind-solar-electricity coordination model is the first weighted sum of multiple sub-objective functions. Different optimization requirements can be met by adjusting the weight corresponding to each sub-objective function, so that the target scheduling strategy obtained by the preset wind-solar-electricity coordination model can take into account multiple sub-objective functions, thereby realizing multi-objective coordinated optimization.

[0100] In some exemplary embodiments, the minimum energy consumption sub-objective function is the minimum value of the sum of the wind power generation cost, the photovoltaic power generation cost, the first electricity purchase cost and the negative electricity sales profit within a set scheduling period; wherein the first electricity purchase cost is the purchase cost corresponding to part of the purchased electricity.

[0101] In some exemplary embodiments, the minimum economic cost sub-objective function is the minimum value of the function value of the minimum energy consumption sub-objective function within a set scheduling period and the second electricity purchase cost; wherein the second electricity purchase cost is the purchase cost corresponding to all self-generated electricity.

[0102] In some exemplary embodiments, the self-generation abandonment rate sub-objective function is the ratio of a first difference obtained by subtracting the load demand forecast of the power consumption unit from the sum of the output of the wind power generation unit and the output of the photovoltaic power generation unit, and a second sum obtained by adding the output of the wind power generation unit to the output of the photovoltaic power generation unit; wherein, when the ratio is a negative number, the function value of the self-generation abandonment rate sub-objective function is determined to be zero.

[0103] In some exemplary embodiments, the preset wind-solar power coordination model includes an objective function which is a first weighted sum of three sub-objective functions:

[0104] f(E1(t), E2(t), E3(t))=min(λ1E1(t)+λ2E2(t)+λ3E3(t))

[0105] Among them, E1(t) is the minimum energy consumption sub-objective function, which is used to indicate low energy consumption, and λ1 is the corresponding weight; E2(t) is the minimum economic cost sub-objective function, which is used to indicate low cost and high economy, and λ2 is the corresponding weight; E3(t) is the self-generation abandonment rate sub-objective function, which is used to indicate a low self-generation abandonment rate (i.e., less abandonment of self-generation), and λ3 is the corresponding weight.

[0106] In some exemplary embodiments, the self-generation units in the oil and gas field production system include wind power generation units and / or photovoltaic power generation units. Accordingly, the self-generation abandonment rate sub-objective function is also called the wind and light abandonment rate sub-objective function. The smaller the function value, the less wind energy and / or light energy are abandoned.

[0107] In some exemplary embodiments, the wind power generation unit is a wind turbine, the photovoltaic power generation unit is a distributed photovoltaic power station, and the minimum energy consumption sub-objective function is:

[0108]

[0109] Among them, N w is the number of fans, N v is the number of photovoltaic power plants, C wq (t) is the equipment maintenance cost of the fan in the dispatching cycle corresponding to time t, C vj (t) is the equipment maintenance cost of the photovoltaic power station in the dispatching period corresponding to time t, P sell is the electricity sold in the current dispatch period, P buy is the purchased electricity in the current dispatch cycle, ω(t) is the line fluctuation price at time t, and p(t) is the electricity price at time t. Specifically,

[0110] C wq (t) = P m *T*P wq (t), q = (1, 2, ..., N)

[0111] C vj (t) = P v *T*P vj (t), q = (1, 2, ..., N)

[0112] Among them, T is the scheduling period, P m is the operation and maintenance cost per unit of electricity generated by the wind turbine, P v is the operation and maintenance cost per unit of photovoltaic power generation, P wq (t) is the fan output at time t, P vj (t) is the output of the photovoltaic power station at time t.

[0113] Specifically, is the cost of wind turbine power generation, is the power generation cost of the photovoltaic power station, P sell *ω(t) is the profit from selling electricity to other parties, P buy *p(t) is the first electricity purchase cost; it can be seen that the first electricity purchase cost is the purchase cost corresponding to the actual externally purchased electricity in the current scheduling cycle.

[0114] In some exemplary embodiments, the minimum economic cost sub-objective function is:

[0115]

[0116] Where T is the dispatch period, p(t) is the electricity price at time t, and F k (t) is the load demand forecast of the kth power consumption unit at time t, and N is the number of power consumption units.

[0117] Specifically, is the second electricity purchase cost; it can be seen that the second electricity purchase cost refers to the cost required if all self-generated electricity is purchased from outside, that is, the purchase cost corresponding to all self-generated electricity in the current scheduling cycle.

[0118] In some exemplary embodiments, the wind and solar power abandonment rate sub-objective function is:

[0119]

[0120] Among them, P wq (t) is the fan output at time t, P vj (t) is the output of the photovoltaic power station at time t, N w is the number of fans, N v is the number of photovoltaic power plants, F k (t) is the load demand forecast of the kth power consumption unit at time t, and N is the number of power consumption units.

[0121] It should also be noted that when determining the amount of electricity purchased P buy When the wind and solar power abandonment rates are calculated, first determine whether the sum of wind turbine output and photovoltaic output meets the total load demand forecast, that is, determine

[0122]

[0123] When the sum of the wind turbine output and the photovoltaic power station output is greater than or equal to the total load demand forecast, it means that there is no need to purchase electricity from outside, and electricity can be sold to the outside and wind and solar power can be abandoned. At this time, the purchased electricity can be set to zero, and the wind and solar power abandonment rate needs to be calculated through the third sub-objective function. The electricity sold is:

[0124]

[0125] Among them, α1 is a preset electricity sales conversion coefficient, which is a number greater than or equal to 0 and less than or equal to 1. In a specific embodiment, α1 is 0.25, and can also be flexibly set according to needs, without being limited to specific aspects.

[0126] When the sum of the wind turbine output and the photovoltaic power station output is less than the total load demand forecast, it means that the current wind turbine and photovoltaic power station output cannot meet the actual demand, so it is necessary to purchase electricity from outside. At this time, the amount of electricity sold outside is set to zero, the wind and solar power abandonment rate is zero, and the amount of electricity purchased is the actual amount of electricity purchased.

[0127] In some exemplary embodiments, the constraint condition includes one or more of the following sub-constraint conditions:

[0128] Line power fluctuation constraint, used to constrain the transmission power on the power grid line to fall within a preset transmission power range;

[0129] Energy constraints, used to constrain the output of each power generation unit to fall within its corresponding power range;

[0130] The load balance constraint is used to constrain the sum of the predicted load demands of all power consumption units to be equal to the sum of the power generated by all self-generating units plus the purchased power, minus the sold power.

[0131] Among them, the line power fluctuation constraint means that the transmission power on the power grid line needs to be within the preset transmission power range of the power grid line. In other words, the transmission power on the power grid line cannot exceed the load limit of the line per unit time, which is conducive to avoiding damage to the power grid line and inaccurate final results due to failure to carry the design parameters.

[0132] In some exemplary embodiments, the line power fluctuation constraint may be specifically expressed as:

[0133]

[0134] Pgrid_min ≤P grid (t)≤P grid_max

[0135] Among them, P grid (t) is the Internet power of all nodes at time t, P grid_min and P grid_max They are the minimum and maximum values ​​of the power transmission of the upper power grid line.

[0136] In some exemplary embodiments, taking the wind and solar power grid as an example, the energy constraint means that the wind turbine output needs to be within the preset wind turbine output range, and the photovoltaic power station output needs to be within the preset photovoltaic output range. In other words, the obtained wind turbine output and photovoltaic power station output cannot exceed the maximum output of the corresponding equipment, which is used to avoid damage caused by overload operation of the equipment, and inaccurate final results due to failure to achieve the design parameters.

[0137] In some exemplary embodiments, the energy constraint may be specifically expressed as:

[0138]

[0139]

[0140] in, and They are the minimum power and rated power of the fan when it is running;

[0141] and They are the minimum power and rated power of the distributed photovoltaic power station during operation.

[0142] In some exemplary embodiments,

[0143] In some exemplary embodiments, taking the wind and solar power grid as an example, the total load demand forecast is equal to the third sum of wind turbine output, photovoltaic output, externally purchased power and negative externally sold power. In one embodiment, the load balance constraint can be specifically expressed as:

[0144]

[0145] Among them, F k (t) is the load demand forecast of the kth power consumption unit at time t, N is the number of production plants, P sell (t) is the electricity sold at time t, P buy (t) is the purchased electric power at time t.

[0146] In some exemplary embodiments, Figure 2As shown, the method of obtaining the load demand forecast of the current power consumption unit includes:

[0147] Step 1110, obtaining the current actual load of the power consumption unit;

[0148] Step 1120, input the actual load into the pre-trained target neural network model to obtain the corresponding load demand forecast.

[0149] In some exemplary embodiments, the actual load of the power consumption unit can be obtained by monitoring the actual operation of each power consumption unit in the production system. The specific monitoring scheme or the monitoring system involved is not discussed in detail in this application. The actual load is input / substituted into the pre-trained target neural network model for calculation to obtain the load demand forecast that needs to be predicted. It can be understood that obtaining the load demand forecast through a neural network improves the efficiency and accuracy of the forecast.

[0150] It should be noted that the scheduling cycle for power scheduling in the production system includes at least one. When there are at least two scheduling cycles, the structure and / or parameters of the target neural network model can be adjusted to enable it to output the load demand forecast corresponding to each preset scheduling cycle.

[0151] In some exemplary embodiments, Figure 3 As shown, the target neural network model is pre-trained according to the following method:

[0152] Step 310, constructing a preset neural network model;

[0153] Step 320, determining a learning rate range and a weight correction factor range of the neural network model;

[0154] Step 330, performing a learning rate search to obtain an initial value of the learning rate and an initial value of the weight correction factor according to the learning rate range and the weight correction factor range, and updating the neural network model;

[0155] Step 340, training the updated neural network model according to the training set to obtain the target neural network model;

[0156] The training set includes actual loads of power consumption units at multiple historical moments.

[0157] First, a preset neural network model is constructed, and the required input and output are predetermined. In one embodiment, the input is the actual load of the power consumption unit, and the output is the load demand forecast of the power consumption unit; further, the preset neural network model is initialized, and the number of neurons contained in each layer and the connection weights and other related parameters are set so that the preset neural network model can perform calculations; then the learning rate range and weight correction factor range of the preset neural network model are determined, wherein the introduction of the weight correction factor is used to solve the problems of oscillation or extended convergence time that may occur during the weight learning process; the learning rate search operation is performed according to the learning rate range and the weight correction factor range to obtain the initial value of the learning rate and the initial value of the weight correction factor, and the preset neural network model is updated, that is, a neural network model with high learning efficiency and stability is obtained. Then, the actual load data at the historical moment is obtained to form a training set and the updated preset neural network model is trained to obtain a trained target neural network model.

[0158] In some exemplary embodiments, Figure 4 As shown, the updated neural network model is trained according to the training set to obtain the target neural network model, including:

[0159] Step 410, calculating a second weighted sum of actual loads of power consumption units at historical moments in the training set according to a second weight between the input layer and the hidden layer, and inputting the second weighted sum into the hidden layer to obtain a hidden layer output value;

[0160] Step 420, calculating a third weighted sum of the output values ​​of the hidden layer according to a third weight between the hidden layer and the output layer, and inputting the third weighted sum into the output layer to obtain an output value of the output layer;

[0161] Step 430, determining a second difference between the output value of the output layer and a preset expected value;

[0162] Step 440, determining error accuracy according to the second difference and a preset error criterion function;

[0163] Step 450, when the error accuracy meets the preset accuracy range, determine that the training is completed to obtain the target neural network model;

[0164] Step 460, when the error accuracy does not meet the preset accuracy range, the second weight and the third weight are respectively corrected according to the initial value of the learning rate and the initial value of the correction factor, and training is performed again until the training is completed.

[0165] It should be noted that the training is performed again using similar steps until the preset accuracy range is met, then the training is terminated to obtain a trained target neural network model.

[0166] In some exemplary embodiments, the neural network model includes a BP (back propagation) neural network model.

[0167] In some exemplary embodiments, the preset error criterion function includes: a mean square error (MSE) criterion function.

[0168] In some exemplary embodiments, Figure 5 As shown, the updated neural network model is trained according to the training set to obtain the target neural network model, including:

[0169] Step 510, the actual load data of the power consumption unit at the historical moment in the training set is selected through the Spearman rank correlation coefficient to select the data with high correlation coefficient for weighted calculation, and the second weighted sum value is used as the input value of the hidden layer to obtain the hidden layer output;

[0170] Step 520, obtaining an output value of an output layer and a second difference between the output value and a preset expected value according to the output of the hidden layer; wherein the output value of the output layer is a third weighted sum value of the output of the hidden layer;

[0171] Step 530, obtaining error accuracy according to the second difference and a preset MSE square error criterion function;

[0172] Step 540, when the error accuracy meets the preset accuracy range, determining that the neural network model at this time is a trained target neural network model;

[0173] Step 550, when the error accuracy does not meet the preset accuracy range, the first weight between the input layer and the hidden layer, the second weight between the hidden layer and the output layer, and the second difference between the output value and the preset expected value are corrected according to the initial value of the learning rate and the initial value of the weight correction factor, and training is performed again.

[0174] In some exemplary embodiments, the target neural network model includes at least an input layer, a hidden layer and an output layer, wherein the actual load data of the power consumption unit at the historical moment (referred to as the historical load data) is used to select input variables with high correlation coefficients using the Spearman rank correlation coefficient, which plays a role in dimensionality reduction, thereby improving the training speed. All training data are divided into two column vectors X and Y: the corresponding elements x h ,y hConverted to the ranking in the respective column vector, denoted as x' h , y' h The Spearman rank correlation coefficient is:

[0175]

[0176] where r s is the Spearman rank correlation coefficient; t is the number of actual load data of power consumption units at historical moments; d h is the rank difference, d h =x' h -y' h .

[0177] Through the input layer, when entering the hidden layer, it needs to be weighted by the historical load data selected by the Spearman rank correlation coefficient, and the weighted sum is used as the input of the neurons in the hidden layer, which can be specifically:

[0178]

[0179] Among them, E i (t) is the input signal of the i-th input layer neuron, i.e., the actual value selected from the historical load data by the Spearman rank correlation coefficient, n is the number of input layer neurons, λ ji is the weight value from the jth neuron in the hidden layer to the ith neuron in the input layer.

[0180] The activation function f(M j ) is preferably a Sigmoid function, the output of which, i.e., the output of the hidden layer, is:

[0181]

[0182] Then the actual output of the output layer is:

[0183]

[0184] Among them, λ kj is the weight from the kth neuron in the output layer to the jth neuron in the hidden layer, and q is the number of neurons in the hidden layer.

[0185] Furthermore, during the Pth training, the second difference between the actual output of the output layer and the expected value is:

[0186]

[0187] Where, e(P) is the second difference, y qk (P) is the actual output of the kth neuron in the output layer during the Pth training, y ek(P) is the expected output of the kth neuron in the output layer during the Pth training, that is, the expected value corresponding to the historical load data, and m is the number of output values ​​output by the output layer.

[0188] In some exemplary embodiments, the MSE error criterion function is preferably:

[0189]

[0190] Wherein, λ is the weight, and L is the obtained error accuracy: when the error accuracy meets the preset accuracy range, it is determined that the result obtained by the neural network at this time meets the preset accuracy requirement, and therefore it can be determined that the neural network model at this time is the target neural network model;

[0191] When the error accuracy does not meet the preset accuracy range, it is determined that the result obtained by the neural network at this time does not meet the preset accuracy requirement, so it is necessary to perform error correction based on the error criterion function, specifically including: using the gradient descent method to realize the weight and second difference correction of each neuron, wherein the weight correction formula between the hidden layer and the output layer is preferably:

[0192]

[0193] Among them, Δλ kj is the weight correction from the kth neuron in the output layer to the jth neuron in the hidden layer, γ is the learning rate, γ∈(0,1).

[0194] Then the corrected weight between the hidden layer and the output layer during the P+1th training is:

[0195] λ kj (P+1)=λ kj (P)+Δλ kj

[0196] The weight correction formula between the input layer and the hidden layer is preferably:

[0197]

[0198] Among them, Δλ ji is the weight correction from the jth neuron in the hidden layer to the ith neuron in the input layer.

[0199] Therefore, the revised weights between the input layer and the hidden layer during P+1 training are:

[0200] λ ji (P+1)=λ ji (P)+Δλ ji

[0201] The second difference correction formula between the actual output of the output layer and the expected value is preferably:

[0202]

[0203] Therefore, during the P+1 training, the second difference between the actual output of the output layer and the expected value is corrected to:

[0204] e(P+1)=e(P)+Δe(P)

[0205] Considering that oscillation or extended convergence time may occur during weight learning, in some exemplary embodiments, the influence of two adjacent weights is considered, and a weight correction factor is added for correction. At this time, the corrected weight between the input layer and the hidden layer is:

[0206] λ ji (P+1)=λ ji (P)+Δλ ji +δ(λ ji (P-1))

[0207] The revised weights between the output layer and the hidden layer are:

[0208] λ kj (P+1)=λ kj (P)+Δλ kj +δ(λ kj (P-1))

[0209] The second difference between the actual output of the output layer and the expected value is corrected to:

[0210] e(P+1)=e(P)+Δe(P)+δ(e(P-1))

[0211] Among them, δ is the weight correction factor, and δ∈[0,1].

[0212] By continuously correcting the weights until the error accuracy meets the requirements, the accuracy of the results obtained in actual use is guaranteed.

[0213] Preferably, if the weight correction factor used is inappropriate, it may affect the speed of the learning rate. Therefore, in one embodiment of the present application, the learning speed and the weight correction factor can be continuously corrected as the output error changes. The formula in the correction process is:

[0214] γ=γ min +γ min δ max s(L / C)

[0215] δ=δ max -γ min δ max s(L / C)

[0216] Among them, γmin is the minimum value of the learning rate, δ max is the maximum weight correction factor, C is the weight correction factor range and C∈[0,1], s is the learning rate range.

[0217] The embodiment of the present disclosure also provides a wind-solar-electricity coordinated scheduling device, such as Figure 6 As shown, including,

[0218] The load prediction module 610 is configured to obtain the load demand prediction of the current power consumption unit when the scheduling opportunity is reached;

[0219] The dispatch strategy determination module 620 is configured to determine a target dispatch strategy for wind and solar power dispatch according to the load demand forecast of the power consumption unit, a preset wind and solar power coordination model and constraint conditions;

[0220] A scheduling module 630, configured to perform wind and solar power scheduling according to the scheduling parameters included in the target scheduling strategy;

[0221] The dispatching parameters include one or more of the following: the output of the self-generating unit, the amount of electricity purchased and the amount of electricity sold; the self-generating unit includes: a wind power generation unit and / or a photovoltaic power generation unit;

[0222] The objective function included in the wind-solar-electricity coordination model is the first weighted sum of the following one or more sub-objective functions: a minimum energy consumption sub-objective function, a minimum economic cost sub-objective function and a self-generation abandonment rate sub-objective function.

[0223] In some exemplary embodiments, the minimum energy consumption sub-objective function is the minimum value of the sum of the wind power generation cost, the photovoltaic power generation cost, the first power purchase cost and the negative power sales profit within the set scheduling period; wherein the first power purchase cost is the purchase cost corresponding to the actual external power purchase within the set scheduling period;

[0224] The minimum economic cost sub-objective function is the minimum value of the function value of the minimum energy consumption sub-objective function and the second power purchase cost within the set scheduling period; wherein the second power purchase cost is the purchase cost corresponding to the actual total self-generated electricity within the set scheduling period;

[0225] The self-generation abandonment rate sub-objective function is the ratio of a first difference obtained by subtracting the load demand forecast of the power consumption unit from the sum of the output of the wind power generation unit and the output of the photovoltaic power generation unit, to a second sum obtained by adding the output of the wind power generation unit to the output of the photovoltaic power generation unit; wherein, when the ratio is a negative number, the function value of the self-generation abandonment rate sub-objective function is determined to be zero.

[0226] In some exemplary embodiments, the apparatus further comprises:

[0227] The neural network model module is configured to train a preset neural network model according to the following method to obtain a target neural network model:

[0228] Build a preset neural network model;

[0229] Determining a learning rate range and a weight correction factor range of the neural network model;

[0230] According to the learning rate range and the weight correction factor range, a learning rate search is performed to obtain an initial value of the learning rate and an initial value of the weight correction factor, and the neural network model is updated;

[0231] Training the updated neural network model according to the training set to obtain the target neural network model;

[0232] The training set includes actual loads of power consumption units at multiple historical moments.

[0233] In some exemplary embodiments, the load prediction module 610 is configured to obtain the current actual load of the power consumption unit;

[0234] The actual load is input into the pre-trained target neural network model to obtain the corresponding load demand forecast.

[0235] The present disclosure also provides a production system, including:

[0236] A self-generating unit, a power consuming unit, and a wind-solar-electricity coordinated dispatching device as described in any embodiment of the present disclosure;

[0237] Among them, the self-generating unit includes one or more; the power consumption unit includes one or more.

[0238] In some exemplary embodiments, the self-generating unit includes: a wind power generating unit and / or a photovoltaic power generating unit.

[0239] In some exemplary embodiments, the wind power generation unit includes a wind turbine, and the photovoltaic power generation unit includes a photovoltaic power station, wherein the wind turbine includes one or more wind turbines, and the photovoltaic power station includes one or more photovoltaic power stations.

[0240] In some exemplary embodiments, the production system includes an oil and gas field production system, wherein the power consumption units (also referred to as power users) include units in the oil and gas field that need to use power for oil and gas production and / or related auxiliary / management / operation work.

[0241] The present disclosure also provides an electronic device, including:

[0242] one or more processors;

[0243] a storage device for storing one or more programs,

[0244] When the one or more programs are executed by the one or more processors, the one or more processors implement the wind-photovoltaic coordinated scheduling method as described in any embodiment of the present disclosure.

[0245] The embodiments of the present disclosure further provide a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the wind-photovoltaic coordinated scheduling method as described in any embodiment of the present disclosure is implemented.

[0246] The wind and solar power coordinated scheduling solution provided by the embodiments of the present disclosure determines the target scheduling strategy based on the objective function composed of multiple sub-objective functions and the accurate electricity load demand forecast, and performs wind and solar power scheduling according to the target scheduling strategy, thereby realizing scientific scheduling of self-owned power production, purchased power and sold power for large-scale production systems with wind and solar self-generation capabilities, ensuring the power reliability of the production system and reducing overall energy consumption.

[0247] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all components may be implemented as software executed by a processor, such as a digital signal processor or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0248] 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 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 wind and solar power coordinated dispatching method, characterized in that: include, When the dispatching time is reached, the load demand forecast of the current power consumption unit is obtained; Determine the target dispatching strategy of wind and solar power dispatching according to the load demand forecast of the power consumption unit, the preset wind and solar power coordination model and the constraint conditions; Perform wind and solar power dispatching according to the dispatching parameters included in the target dispatching strategy; The dispatching parameters include one or more of the following: the output of the self-generating unit, the amount of electricity purchased and the amount of electricity sold; the self-generating unit includes: a wind power generation unit and / or a photovoltaic power generation unit; The objective function included in the wind-solar-electricity coordination model is the first weighted sum of the following one or more sub-objective functions: a minimum energy consumption sub-objective function, a minimum economic cost sub-objective function and a self-generation abandonment rate sub-objective function.

2. The method according to claim 1, characterized in that The minimum energy consumption sub-objective function is the minimum value of the sum of the wind power generation cost, the photovoltaic power generation cost, the first power purchase cost and the negative power sales profit within the set scheduling period; wherein the first power purchase cost is the purchase cost corresponding to the actual external power purchase within the set scheduling period; The minimum economic cost sub-objective function is the minimum value of the function value of the minimum energy consumption sub-objective function and the second power purchase cost within the set scheduling period; wherein the second power purchase cost is the purchase cost corresponding to all self-generated electricity within the set scheduling period; The self-generation abandonment rate sub-objective function is the ratio of a first difference obtained by subtracting the load demand forecast of the power consumption unit from the sum of the output of the wind power generation unit and the output of the photovoltaic power generation unit, to a second sum obtained by adding the output of the wind power generation unit to the output of the photovoltaic power generation unit; wherein, when the ratio is a negative number, the function value of the self-generation abandonment rate sub-objective function is determined to be zero.

3. The method according to claim 1 or 2, characterized in that The constraint condition includes one or more of the following sub-constraint conditions: Line power fluctuation constraint, used to constrain the transmission power on the power grid line to fall within a preset transmission power range; Energy constraints, used to constrain the output of each power generation unit to fall within its corresponding power range; The load balance constraint is used to constrain the sum of the predicted load demands of all power consumption units to be equal to the sum of the power generated by all self-generating units plus the purchased power, minus the sold power.

4. The method according to claim 1 or 2, characterized in that: The step of obtaining the load demand forecast of the current power consumption unit includes: Obtain the current actual load of the power consumption unit; The actual load amount is input into the pre-trained target neural network model to obtain the corresponding load demand forecast amount.

5. The method according to claim 4, characterized in that The target neural network model is pre-trained according to the following method: Build a preset neural network model; Determining a learning rate range and a weight correction factor range of the neural network model; According to the learning rate range and the weight correction factor range, a learning rate search is performed to obtain an initial value of the learning rate and an initial value of the weight correction factor, and the neural network model is updated; Training the updated neural network model according to the training set to obtain the target neural network model; The training set includes actual loads of power consumption units at multiple historical moments.

6. The method according to claim 5, characterized in that The step of training the updated neural network model according to the training set to obtain the target neural network model includes: Calculating a second weighted sum of actual loads of power consumption units at historical moments in the training set according to a second weight between the input layer and the hidden layer, and inputting the second weighted sum into the hidden layer to obtain a hidden layer output value; Calculating a third weighted sum of the output values ​​of the hidden layer according to a third weight between the hidden layer and the output layer, and inputting the third weighted sum into the output layer to obtain an output value of the output layer; Determine a second difference between the output value of the output layer and a preset expected value; Determine the error accuracy according to the second difference and a preset error criterion function; When the error accuracy meets the preset accuracy range, determining that the training is completed to obtain the target neural network model; When the error precision does not meet the preset precision range, the second weight and the third weight are respectively corrected according to the initial value of the learning rate and the initial value of the correction factor, and training is performed again until the training is completed.

7. The method according to claim 1 or 2, characterized in that: The scheduling opportunity includes any of the following opportunities: The scheduling time point of periodic scheduling arrives; An instruction from an operator or external system to initiate a periodic schedule is received.

8. A wind-photovoltaic coordinated dispatching device, characterized in that: include, A load forecasting module is configured to obtain a load demand forecast of a current power consumption unit when a scheduling opportunity is reached; A dispatching strategy determination module, configured to determine a target dispatching strategy for wind and solar power dispatching according to a load demand forecast of the power consumption unit, a preset wind and solar power coordination model and constraint conditions; A scheduling module, configured to perform wind and solar power scheduling according to the scheduling parameters included in the target scheduling strategy; The dispatching parameters include one or more of the following: the output of the self-generating unit, the amount of electricity purchased and the amount of electricity sold; the self-generating unit includes: a wind power generation unit and / or a photovoltaic power generation unit; The objective function included in the wind-solar-electricity coordination model is the first weighted sum of the following one or more sub-objective functions: a minimum energy consumption sub-objective function, a minimum economic cost sub-objective function and a self-generation abandonment rate sub-objective function.

9. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the wind-photovoltaic coordinated scheduling method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the wind-photovoltaic coordinated scheduling method as described in any one of claims 1 to 7 is implemented.

Images