A distributed inverter marshalling scheduling method based on cloud edge collaboration and user behavior prediction, a cloud server and a scheduling system

Abstract

The application relates to a distributed inverter grouping scheduling method based on cloud edge cooperation and user behavior prediction, a cloud server and a scheduling system. The method comprises the following steps: based on historical power consumption data of a plurality of inverter nodes, determining user portrait information corresponding to each inverter node, and based on geographical coordinates of the plurality of inverter nodes and power suppliers, dividing the plurality of inverter nodes into at least two scheduling groups; based on the plurality of user portrait information, weather forecast data of a future period and power price prediction data of a regional power grid corresponding to each scheduling group in the future period, determining a load power consumption probability prediction interval and a photovoltaic output probability prediction interval of each inverter node in the future period; and based on the load power consumption probability prediction interval of each inverter node in the future period, the photovoltaic output probability prediction interval of each inverter node in the future period and power grid load information of the regional power grid corresponding to each scheduling group, the operation mode of each inverter node can be accurately determined.

Description

Technical Field

[0001] This application relates to the field of inverter control technology, and in particular to a distributed inverter grouping and scheduling method, cloud server and scheduling system based on cloud-edge collaboration and user behavior prediction. Background Technology

[0002] With the popularization of distributed photovoltaic technology and home energy storage systems, the number of inverters used is also increasing, and a large number of distributed inverters are starting to be connected to the grid.

[0003] As the number of inverters connected to the grid continues to increase, and different inverters correspond to different regional power grids, and users of different inverters also have different electricity consumption behaviors, all these factors have led to different actual operating conditions for different inverters. This has resulted in significant difficulties in regulating the operating modes of massive inverters. Summary of the Invention

[0004] Based on this, it is necessary to provide a distributed inverter grouping and scheduling method, cloud server, scheduling system, device, computer equipment, computer-readable storage medium, and computer program product that can determine an operating mode for inverter nodes with high adaptability to actual needs based on cloud-edge collaboration and user behavior prediction, in order to address the above-mentioned technical problems.

[0005] Firstly, this application provides a distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction, applied in a cloud server. The method includes:

[0006] Obtain historical electricity consumption data, geographical coordinates, and power suppliers for multiple inverter nodes;

[0007] Based on historical electricity consumption data of multiple inverter nodes, determine the user profile information corresponding to each inverter node, and divide the multiple inverter nodes into at least two scheduling groups based on the geographical coordinates of multiple inverter nodes and power suppliers.

[0008] Based on multiple user profiles, weather forecast data for future periods, and electricity price prediction data for the corresponding regional power grids of each dispatching group for future periods, the probability prediction range of load electricity consumption and photovoltaic output for each inverter node in future periods are determined.

[0009] Based on the predicted load power consumption range, the predicted photovoltaic power output range, and the grid load information of the corresponding regional power grid for each inverter node in the future time period, the operating mode of each inverter node is determined. The operating mode is used to control the corresponding inverter node to adjust its power to the baseline power corresponding to the operating mode in the future time period.

[0010] In an exemplary embodiment, determining the operating mode of each inverter node based on the predicted load consumption probability range for each inverter node in the future time period, the predicted photovoltaic output probability range for the future time period, and the grid load information of the corresponding regional power grid of each scheduling group includes:

[0011] Based on the predicted load power consumption probability range of each inverter node in the future time period, the predicted photovoltaic power output probability range in the future time period, and the grid load information corresponding to each scheduling group, at least two scheduling groups are divided into at least one first scheduling group and / or at least one second scheduling group.

[0012] The operating mode of each inverter node in each first scheduling group is determined to be the balanced grid mode. The balanced grid mode is used to control multiple inverter nodes in at least one first scheduling group to adjust the power in the same power adjustment direction in the future time period, so as to adjust the power to the baseline power corresponding to each inverter node, and the baseline power of each inverter node corresponds to the energy storage capacity.

[0013] Based on the predicted load power consumption probability range of each inverter node in the future period, the predicted photovoltaic power output probability range in the future period, and the electricity price prediction data of the corresponding regional power grid of each second dispatch group in the future period, the operating mode of each inverter node in each second dispatch group is determined to be either self-consumption mode or power generation priority mode.

[0014] In an exemplary embodiment, determining the operating mode of each inverter node in each second dispatch group as either a self-consumption mode or a generation priority mode based on the load consumption probability prediction range of each inverter node in the future time period, the photovoltaic output probability prediction range in the future time period, and the electricity price prediction data of the regional power grid corresponding to each second dispatch group in the future time period includes:

[0015] Based on the predicted load electricity consumption probability range and the predicted photovoltaic output probability range for each inverter node in the future time period, the net load uncertainty range for each inverter node is determined. The minimum net load of the net load uncertainty range is determined by the difference between the minimum value of the predicted load electricity consumption probability range and the maximum value of the predicted photovoltaic output probability range, and the maximum net load of the net load uncertainty range is determined by the difference between the maximum value of the predicted load electricity consumption probability range and the minimum value of the predicted photovoltaic output probability range.

[0016] Based on the electricity price forecast data corresponding to the second dispatch group, the minimum and maximum net load of each inverter node, the operating mode of each inverter node in each second dispatch group is determined to be either self-consumption mode or generation priority mode.

[0017] In an exemplary embodiment, determining whether the operating mode of each inverter node in each second scheduling group is a self-consumption mode or a generation priority mode based on the electricity price forecast data corresponding to the second scheduling group, the minimum net load of each inverter node, and the maximum net load of each inverter node includes:

[0018] If the electricity price forecast data corresponding to the second dispatch group indicates that the grid sales price is greater than or equal to the preset sales price threshold, and the maximum net load of the corresponding inverter node is greater than 0, then the operating mode of the corresponding inverter node is determined to be the self-consumption mode; in the self-consumption mode, the baseline power of each inverter node corresponds to the maximum net load.

[0019] If the electricity price forecast data corresponding to the second dispatch group indicates that the grid purchase price is greater than or equal to the preset purchase price threshold, and the minimum net load of the corresponding inverter node is less than 0, then the operating mode of the corresponding inverter node is determined to be the generation priority mode; in the generation priority mode, the baseline power of each inverter node corresponds to the minimum net load.

[0020] In an exemplary embodiment, determining the net load uncertainty interval for each inverter node based on the predicted load consumption probability interval and the predicted photovoltaic output probability interval for each inverter node in the future time period includes:

[0021] Based on the risk preference coefficient of each inverter node, the load power consumption probability prediction interval for the future time period of each inverter node is narrowed to obtain the narrowed load power consumption probability prediction interval for the future time period. The maximum value of the narrowed load power consumption probability prediction interval is less than the maximum value of the original load power consumption probability prediction interval, and the minimum value of the narrowed load power consumption probability prediction interval is greater than the minimum value of the original load power consumption probability prediction interval.

[0022] Based on the risk preference coefficient of each inverter node, the probability prediction interval of photovoltaic output for the future time period of each inverter node is narrowed to obtain the narrowed probability prediction interval of photovoltaic output for the future time period. The maximum value of the narrowed probability prediction interval of photovoltaic output is less than the maximum value of the original photovoltaic output probability prediction interval, and the minimum value of the narrowed probability prediction interval of photovoltaic output is greater than the minimum value of the original photovoltaic output probability prediction interval.

[0023] Based on the predicted range of load electricity consumption probability after the contraction of the future time period and the predicted range of photovoltaic power output probability after the contraction of the future time period, the net load uncertainty range of each inverter node is determined.

[0024] In an exemplary embodiment, determining the user profile information corresponding to each inverter node based on historical electricity consumption data of multiple inverter nodes includes:

[0025] Based on historical electricity consumption data from multiple inverter nodes, the daily load curve and multiple electricity consumption characteristics of each inverter node are determined.

[0026] Based on the daily load curves of each inverter node, multiple electricity consumption characteristics, and the characteristic thresholds corresponding to each electricity consumption characteristic, the user profile information corresponding to each inverter node is determined.

[0027] In an exemplary embodiment, determining the user profile information corresponding to each inverter node based on the daily load curve of each inverter node, multiple electricity consumption characteristics, and characteristic thresholds corresponding to each electricity consumption characteristic includes:

[0028] Obtain the energy storage capacity corresponding to each inverter node, as well as the rated power of each electrical load connected to each inverter node;

[0029] Based on the energy storage capacity corresponding to each inverter node, the rated power of each corresponding electrical load, multiple electricity consumption characteristics, and the characteristic thresholds corresponding to each electricity consumption characteristic, the user tag of each inverter node is determined.

[0030] Based on the daily load curves and user tags corresponding to each inverter node, user profile information corresponding to each inverter node is determined.

[0031] In one exemplary embodiment, the inverter node is a home energy storage inverter node.

[0032] Secondly, this application also provides a cloud server for performing the methods described in any of the first aspects of this application.

[0033] Thirdly, this application also provides a scheduling system, including a cloud server as described in any of the second aspects of this application and multiple inverter nodes;

[0034] Inverter nodes are used to switch from historical mode to operating mode when entering a future time period, so as to adjust the power to the baseline power corresponding to the operating mode in the future time period.

[0035] Fourthly, this application also provides a distributed inverter grouping and scheduling device based on cloud-edge collaboration and user behavior prediction, applied in a cloud server. The device includes:

[0036] The acquisition module is used to acquire historical electricity consumption data, geographical coordinates, and power suppliers for multiple inverter nodes.

[0037] The first determining module is used to determine the user profile information corresponding to each inverter node based on the historical electricity consumption data of multiple inverter nodes, and to divide the multiple inverter nodes into at least two scheduling groups based on the geographical coordinates of the multiple inverter nodes and the power supplier.

[0038] The second determining module is used to determine the load power consumption probability prediction range and photovoltaic output probability prediction range of each inverter node in the future period based on multiple user profile information, weather forecast data for the future period and electricity price prediction data of the regional power grid corresponding to each dispatching group in the future period.

[0039] The third determining module is used to determine the operating mode of each inverter node based on the predicted load power consumption probability range of each inverter node in the future time period, the predicted photovoltaic power output probability range in the future time period, and the grid load information of the grid in the corresponding area of ​​each dispatch group; the operating mode is used to control the corresponding inverter node to adjust the power to the baseline power corresponding to the operating mode in the future time period.

[0040] Fifthly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement some or all of the steps described in any method of the first aspect of the embodiments of this application.

[0041] Sixthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements some or all of the steps described in any method of the first aspect of the embodiments of this application.

[0042] In a seventh aspect, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements some or all of the steps described in any method of the first aspect of the embodiments of this application.

[0043] The aforementioned distributed inverter grouping and scheduling method, cloud server, scheduling system, device, computer equipment, computer-readable storage medium, and computer program product based on cloud-edge collaboration and user behavior prediction have the following advantages. On the one hand, the cloud server can determine the operating mode of each edge-side inverter node, avoiding multiple inverter nodes operating in an "island" state. Through the cloud-edge collaboration architecture, it can perform cluster collaborative response based on the grid load information of the corresponding regional power grid. On the other hand, it can consider both the user profile information corresponding to each inverter node and the grid load information of the regional power grid where the inverter node is located. Therefore, it achieves a comprehensive consideration of user electricity consumption behavior and regional power grid load balance, and can accurately determine the operating mode of the inverter node with high adaptability to actual needs. Based on this, this embodiment achieves accurate and rapid scheduling of distributed inverter groups based on cloud-edge system architecture and user behavior prediction. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A structural block diagram of a scheduling system provided in an embodiment of this application;

[0046] Figure 2 A flowchart illustrating a distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction, provided for an embodiment of this application;

[0047] Figure 3 A flowchart illustrating another distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction provided for an embodiment of this application;

[0048] Figure 4 A structural block diagram of a distributed inverter grouping and scheduling device based on cloud-edge collaboration and user behavior prediction is provided for an embodiment of this application.

[0049] Figure 5 This is an internal structural diagram of a computer device in one embodiment;

[0050] Figure 6 This is a diagram of the internal structure of a computer device in another embodiment. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0052] The distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction provided in this application can be applied to, for example... Figure 1In the application environment shown, the scheduling system 100 includes a cloud server 102 and multiple inverter nodes 104. The cloud server 102 communicates with the multiple inverter nodes 104. For the cloud server 102, the multiple inverter nodes 104 are distributed in different locations; therefore, the multiple inverter nodes 104 can be understood as multiple distributed inverter nodes on the edge side, i.e., the inverter nodes 104 are edge inverter nodes. It can be seen that the cloud server and multiple inverter nodes in this embodiment constitute a cloud-edge collaborative architecture. Based on this cloud-edge system architecture, it is possible to avoid multiple inverter nodes operating in an "island" state, thereby enabling accurate and rapid scheduling of multiple inverter groups.

[0053] In one exemplary embodiment, such as Figure 2 As shown, a distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction is provided. This method is applied to... Figure 1 The cloud server in the process includes the following steps 202 to 208. Wherein:

[0054] Step 202: Obtain historical electricity consumption data, geographical coordinates, and power suppliers for multiple inverter nodes.

[0055] Inverter nodes are nodes that are connected to the regional power grid of a power supplier, capable of purchasing electricity from the power supplier to power electrical loads or to power energy storage batteries, or selling electricity to the power supplier to earn profits.

[0056] An inverter node can be understood as an inverter.

[0057] The electrical load corresponding to an inverter node refers to the electrical equipment connected to the corresponding inverter node that consumes electrical energy to operate; it is the object to which the inverter node supplies power. Optionally, the electrical load corresponding to an inverter node can be the electrical appliances in the user's home. For example, electrical appliances can be air conditioners, refrigerators, water heaters, etc.

[0058] In one exemplary embodiment, the inverter node described above is a home energy storage inverter node.

[0059] In the case where the inverter node is a home energy storage inverter node, the user corresponding to the inverter node can be understood as a family residence.

[0060] Optionally, the rated output power of the inverter node can be 5000W, 6000W, 8000W or other values.

[0061] Historical electricity consumption data of inverter nodes refers to time-series electricity consumption data collected by inverter nodes within a historical period. Optionally, the historical electricity consumption data of inverter nodes may include photovoltaic power generation, load power consumption, energy storage capacity, or other electricity consumption data that can characterize user electricity consumption behavior. For example, the historical electricity consumption data of inverter nodes may also include active power, reactive power, voltage, current, and the remaining percentage of energy storage battery capacity within a historical period, and the event granularity of the historical electricity consumption data of inverter nodes is not less than 1 hour.

[0062] The geographic coordinates of an inverter refer to the spatial location information used to characterize the installation location of the inverter node. Optionally, the geographic coordinates of the inverter may include the latitude and longitude information, administrative region information, or other information that can characterize the installation location of the inverter node.

[0063] The power supplier for an inverter node refers to a grid operator that has dispatch and control rights over the power grid in the area where the inverter node is located. The users corresponding to the inverter node have signed electricity agreements with the power supplier. Based on this, the power supplier can sell electricity to the inverter node or purchase electricity from the inverter node.

[0064] Alternatively, the power supplier could be, for example, Nord Pool (the Nordic power exchange) or Tibber (a German power supplier).

[0065] Step 204: Based on the historical electricity consumption data of multiple inverter nodes, determine the user profile information corresponding to each inverter node, and based on the geographical coordinates of multiple inverter nodes and the power supplier, divide the multiple inverter nodes into at least two scheduling groups.

[0066] User profile information refers to information used to characterize the electricity consumption behavior of users corresponding to each inverter node. This can be understood as determining the user profile information corresponding to each inverter node's points in order to classify the historical electricity consumption data of multiple inverter nodes, thereby extracting the patterns in the electricity consumption behavior characteristics of users corresponding to multiple inverter nodes.

[0067] Optionally, user profile information can be represented using user attribute tags, electricity consumption behavior tags, and load operation tags. User attribute tags are tags used to characterize the basic static characteristics of the user corresponding to the inverter node. Electricity consumption behavior tags are tags used to characterize the electricity consumption habits and load characteristics of the user corresponding to the inverter node. Load operation tags are tags used to characterize the specific composition of the electrical load connected to the inverter node and the actual operating status of the user's energy storage system.

[0068] A user's energy storage system refers to a distributed photovoltaic energy storage system installed in a user's home. The user's energy storage system includes inverters corresponding to inverter nodes and energy storage batteries. Optionally, the inverters in the user's energy storage system integrate energy storage batteries; that is, the user's corresponding energy storage batteries and inverters can be installed as a single unit. The cloud server in the dispatch system communicates with the inverter nodes in the user's energy storage system.

[0069] Optionally, user attribute tags may include information such as user housing type (small apartment / large apartment / villa), photovoltaic panel configuration (with photovoltaic panels / without photovoltaic panels, photovoltaic panel power), and energy storage battery configuration (energy storage capacity of energy storage batteries); electricity consumption behavior tags may include electricity consumption level (extremely low / low / medium / high / extremely high), peak and valley characteristics (single-peak type with high electricity consumption during the day or night / double-peak type with high electricity consumption during both the day and night / night peak type with high electricity consumption only at night / flat type with low electricity consumption during both the day and night), regularity characteristics (highly regular / generally regular / irregular), and dispatchability (mainly rigid load with no dispatch flexibility / some dispatch flexibility / great dispatch flexibility); load operation tags may include the user's main electrical equipment (whether there is an air conditioner / electric heater / electric vehicle / electric water heater), photovoltaic self-consumption rate (self-consumption ratio / grid connection ratio), and energy storage battery operation mode (daily cycle use / cycle use every few days / almost no use), etc.

[0070] Scheduling and grouping refers to the grouping of multiple inverter nodes based on their geographical coordinates and the power supplier. In other words, users belonging to multiple inverter nodes within the same scheduling group have signed electricity agreements with the same power supplier; furthermore, these inverter nodes are geographically close to each other.

[0071] In one exemplary embodiment, the above-mentioned division of multiple inverter nodes into at least two scheduling groups based on the geographical coordinates of multiple inverter nodes and power suppliers includes:

[0072] For power suppliers with multiple inverter nodes, the multiple inverter nodes are divided into at least two initial groups;

[0073] Based on the geographical coordinates of multiple inverter nodes, the multiple inverter nodes are divided into at least two geographical regions;

[0074] Multiple inverter nodes with the same geographical area in each initial group are assigned to the same scheduling group to obtain at least two scheduling groups.

[0075] The geographical region refers to the area obtained by dividing the geographical space based on the geographical coordinates of each inverter node.

[0076] In another exemplary embodiment, the above-described method of dividing multiple inverter nodes into at least two geographical regions based on their geographical coordinates includes:

[0077] Based on the geographic coordinates of multiple inverter nodes, multiple inverter nodes whose geographic coordinates are less than or equal to a distance threshold are divided into the same geographic region, resulting in at least two geographic regions.

[0078] Alternatively, based on the geographical coordinates of multiple inverter nodes, multiple inverter nodes belonging to the same administrative region can be divided into the same geographical region, resulting in at least two geographical regions.

[0079] Optionally, the distance threshold can be 500 meters, 1 km, 2 km, or other numerical values.

[0080] Optionally, belonging to the same administrative region can mean belonging to the same jurisdiction, the same village, or the same residential community.

[0081] Since the scheduling group is determined based on the geographical coordinates of each inverter node and the power supplier, it can be understood that the scheduling group is the intersection of the power supplier and geographical coordinates of each inverter node.

[0082] Step 206: Based on multiple user profile information, weather forecast data for future periods, and electricity price prediction data for the regional power grid corresponding to each dispatching group for future periods, determine the load power consumption probability prediction range and photovoltaic output probability prediction range for each inverter node for future periods.

[0083] The future period refers to the maintenance cycle of a complete operating mode in the future.

[0084] Optionally, the future time period can be the next 4 hours, the next 12 hours, the next 24 hours, or other durations.

[0085] Weather forecast data for future periods refers to predicted data obtained from public meteorological service platforms to characterize the weather conditions for future periods.

[0086] The weather forecast data for the future period covers the geographical areas where multiple scheduling groups are located during the future period.

[0087] The electricity price forecast data for the future period refers to the unit price of electricity that needs to be paid for purchasing electricity from the corresponding regional power grid in the future period, as well as the unit price of electricity that can be obtained by selling electricity to the corresponding regional power grid.

[0088] Optionally, electricity price forecasts for future periods can be obtained from the power suppliers of multiple inverters.

[0089] The predicted load power consumption range for an inverter node in the future refers to the predicted power consumption of the load supplied by the inverter node in the future, representing the possible range of values ​​for the load power consumption in the future. The predicted load power consumption range consists of the minimum and maximum values ​​of the load power consumption, indicating the range within which the load power consumption of the inverter node is likely to fall in the future.

[0090] The photovoltaic (PV) output probability prediction range for an inverter node in the future refers to the prediction result that characterizes the possible range of PV power output values ​​for the PV panels connected to the inverter node in the future. The future PV output probability prediction range consists of the minimum and maximum values ​​of the PV power output, representing the range within which the PV power output of the inverter node is likely to fall in the future.

[0091] Photovoltaic output can be understood as the power generated by photovoltaic power generation.

[0092] Optionally, the probability prediction range of load consumption and photovoltaic output for each inverter node in the future period can be determined by using an LSTM (Long Short-Term Memory) model or a Transformer model based on multiple user profile information, weather forecast data for the future period, and electricity price prediction data of the regional power grid corresponding to each scheduling group in the future period.

[0093] The obtained probability prediction intervals for load consumption and photovoltaic power output for each inverter node in the future time period are determined to have a confidence level. That is, the probability that the actual load consumption power of each inverter node in the future time period falls within the probability prediction interval is greater than or equal to the confidence level, and the probability that the actual photovoltaic power generation power in the future time period falls within the probability prediction interval is also greater than or equal to the confidence level. The confidence level is the difference between 1 and the probability of prediction miscalculation; the probability of prediction miscalculation is the probability that the actual value falls outside the predicted value.

[0094] Step 208: Based on the predicted load power consumption probability range of each inverter node in the future time period, the predicted photovoltaic power output probability range in the future time period, and the grid load information of the grid in the corresponding area of ​​each dispatch group, determine the operating mode of each inverter node; the operating mode is used to control the corresponding inverter node to adjust the power to the baseline power corresponding to the operating mode in the future time period.

[0095] Among them, the grid load information of the regional power grid corresponding to the dispatching group refers to information used to characterize the load operating status of the corresponding regional power grid. The grid load information is used to assess the load pressure and operational stability of the regional power grid corresponding to the dispatching group, in order to determine whether the dispatching group needs to participate in the grid dispatching behavior of the corresponding regional power grid to maintain grid balance.

[0096] Optionally, the grid load information of the regional power grid corresponding to the dispatching group may include the transformer load rate of the regional power grid corresponding to the dispatching group, the frequency deviation between the regional power grid frequency and the rated frequency, dispatching level instructions issued by the power supplier, or other information that affects the grid balance of the regional power grid. Optionally, the grid load information of the regional power grid corresponding to the dispatching group may also include the remaining power percentage of multiple inverter nodes connected to the corresponding regional power grid.

[0097] Baseline power refers to the reference baseline for inverter nodes to adjust power in the future. In other words, the inverter nodes will adjust their power based on the baseline power in the future, ensuring that the average power of the inverter nodes in the future closely matches the baseline power. It can be understood as the reference trajectory for inverter nodes to adjust their power in the future, enabling them to continuously perform power tracking adjustments with the baseline power as the target, so that the actual power of the inverter nodes in the future closely matches the baseline power trajectory.

[0098] Optionally, since the inverter node is connected to the corresponding regional power grid, photovoltaic panels, and energy storage batteries, the inverter can determine the charging and discharging power of the energy storage batteries based on the baseline power.

[0099] Optionally, the operating mode can also be used to control the energy storage battery connected to the corresponding inverter node to adjust the remaining power percentage to the target remaining power percentage corresponding to the operating mode in a future period.

[0100] As can be seen, in this embodiment, the operation mode of multiple inverter nodes is determined by scheduling groups. Multiple inverter nodes in the same scheduling group can have the same operation mode or different operation modes. Based on this, it is possible to avoid multiple inverter nodes operating in an "island" state. Instead, clustered coordinated response can be carried out based on the grid load pressure of the grid corresponding to the scheduling group. That is, when the grid load pressure is high, multiple inverter nodes in the same scheduling group can be controlled to have the same operation mode to play the role of peak shaving and valley filling of the virtual power plant. Thus, by ensuring the safety of the grid corresponding to the scheduling group, the power supply stability of the users corresponding to the inverter nodes is ensured.

[0101] In the aforementioned distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction, on the one hand, the cloud server can determine the operating mode of each edge-side inverter node, avoiding multiple inverter nodes operating in an "island" state. Through the cloud-edge collaboration architecture, cluster collaborative response can be carried out based on the grid load information of the corresponding regional power grid for scheduling and grouping. On the other hand, it can consider both the user profile information corresponding to each inverter node and the grid load information of the regional power grid where the inverter node is located. Therefore, it achieves a comprehensive consideration of user electricity consumption behavior and regional power grid load balance, and can accurately determine the operating mode of the inverter node with high adaptability to actual needs. Based on this, this embodiment realizes accurate and fast scheduling of distributed inverter grouping based on cloud-edge system architecture and user behavior prediction.

[0102] In an exemplary embodiment, the above-mentioned determination of the operating mode of each inverter node based on the predicted load consumption probability range of each inverter node in the future time period, the predicted photovoltaic output probability range in the future time period, and the grid load information of the regional power grid corresponding to each scheduling group includes:

[0103] Based on the predicted load power consumption probability range of each inverter node in the future time period, the predicted photovoltaic power output probability range in the future time period, and the grid load information corresponding to each scheduling group, at least two scheduling groups are divided into at least one first scheduling group and / or at least one second scheduling group.

[0104] The operating mode of each inverter node in each first scheduling group is determined to be the balanced grid mode. The balanced grid mode is used to control multiple inverter nodes in at least one first scheduling group to adjust the power in the same power adjustment direction in the future time period, so as to adjust the power to the baseline power corresponding to each inverter node, and the baseline power of each inverter node corresponds to the energy storage capacity.

[0105] Based on the predicted load power consumption probability range of each inverter node in the future period, the predicted photovoltaic power output probability range in the future period, and the electricity price prediction data of the corresponding regional power grid of each second dispatch group in the future period, the operating mode of each inverter node in each second dispatch group is determined to be either self-consumption mode or power generation priority mode.

[0106] The power regulation direction is characterized by the current output direction of the inverter node. That is, the power regulation direction includes both the direction in which the inverter node outputs power to the regional grid for electricity sales and the direction in which the inverter node absorbs power from the regional grid for electricity purchase. When the inverter node absorbs power from the regional grid, it may be supplying power to electrical loads or charging connected energy storage batteries.

[0107] Energy storage capacity refers to the electrical energy capacity that the energy storage battery connected to the inverter node can carry. Energy storage capacity can be understood as the battery capacity of the energy storage battery. Optionally, energy storage capacity can refer to the rated energy storage capacity of the energy storage battery or the actual energy storage capacity of the energy storage battery.

[0108] Optionally, at least one dispatching group in the corresponding regional power grid with power grid balancing needs may be identified as at least one first dispatching group, and at least one other group in the at least two dispatching groups that is not the first dispatching group may be identified as at least one second dispatching group.

[0109] Optionally, the grid load information corresponding to the scheduling group may include the remaining percentage of energy storage batteries corresponding to each inverter node in the scheduling group.

[0110] The multiple inverter nodes in the first scheduling group share the same operating mode, meaning they will all operate in a balanced grid mode in the future. This can be understood as the grid load information corresponding to the first scheduling group indicating a grid balancing demand in the region corresponding to the first scheduling group. Based on this, the multiple inverter nodes in the first scheduling group can perform clustered coordinated responses based on the corresponding grid load information.

[0111] Multiple inverter nodes in the second dispatch group may have one or two operating modes. That is, in the future, some inverter nodes may operate in self-consumption mode, while others may operate in priority power generation mode.

[0112] The baseline power and energy storage capacity correspondence of each inverter node refers to the positive correlation between the baseline power of each inverter node and the energy storage capacity of its corresponding energy storage battery under balanced grid mode. Based on this, inverter nodes with larger energy storage capacities will have greater power regulation capabilities.

[0113] Self-consumption mode refers to an operating mode in which the inverter node controls the connected energy storage battery to discharge in order to power the load when the unit price of electricity to be paid for purchasing electricity from the corresponding regional power grid in the future period is high, and the predicted range of load electricity consumption in the future period is greater than the predicted range of photovoltaic power output in the future period.

[0114] This can be understood as a self-consumption mode, an operating mode under conditions of high electricity costs and high electricity demand. In this mode, controlling the discharge of energy storage batteries prioritizes the normal operation of the local electricity load of the users corresponding to the inverter nodes, thereby reducing users' electricity costs by decreasing the amount of electricity purchased from the corresponding regional power grid. Based on this, the self-consumption mode can also be understood as a cost-saving mode.

[0115] The generation priority mode refers to the operating mode in which the inverter node charges the energy storage battery connected to it and / or sells electricity to the corresponding regional power grid when the unit price of electricity obtained from selling electricity to the corresponding regional power grid in the future period is higher, and the probability prediction range of photovoltaic power output in the future period is greater than the probability prediction range of load electricity consumption in the future period.

[0116] This can be understood as a generation-priority mode, a computational model where electricity sales profits are high and electricity demand is low. In this mode, charging energy storage batteries absorbs excess photovoltaic power, preventing its waste, and increases user revenue by selling electricity back to the corresponding regional grid. Therefore, the generation-priority mode can also be understood as a more profitable model.

[0117] In this embodiment, on the one hand, it breaks the traditional islanded operation mode of inverters. Through the dual grouping of power suppliers and geographical coordinates, a large number of scattered inverter nodes are endowed with the peak-shaving and frequency regulation capabilities of virtual power plants. In the grid balancing mode, it can effectively alleviate the distribution network overload problem of the corresponding regional power grid. On the other hand, this embodiment can consider both the user profile information corresponding to each inverter node and the grid load information of the regional power grid where the inverter node is located. Therefore, it realizes a comprehensive consideration of user electricity consumption behavior and regional power grid load balancing, and can accurately and flexibly determine the operating mode of the inverter node with high adaptability to actual needs.

[0118] In an exemplary embodiment, the above-mentioned determination of the operating mode of each inverter node in each second dispatch group as either a self-consumption mode or a generation priority mode based on the load consumption probability prediction range of each inverter node in the future time period, the photovoltaic output probability prediction range of the future time period, and the electricity price prediction data of the regional power grid corresponding to each second dispatch group in the future time period includes:

[0119] Based on the predicted load electricity consumption probability range and the predicted photovoltaic output probability range for each inverter node in the future time period, the net load uncertainty range for each inverter node is determined. The minimum net load of the net load uncertainty range is determined by the difference between the minimum value of the predicted load electricity consumption probability range and the maximum value of the predicted photovoltaic output probability range, and the maximum net load of the net load uncertainty range is determined by the difference between the maximum value of the predicted load electricity consumption probability range and the minimum value of the predicted photovoltaic output probability range.

[0120] Based on the electricity price forecast data corresponding to the second dispatch group, the minimum and maximum net load of each inverter node, the operating mode of each inverter node in each second dispatch group is determined to be either self-consumption mode or generation priority mode.

[0121] The net load uncertainty interval of the inverter node refers to the prediction result calculated based on the probability prediction interval of the inverter node's load electricity consumption and the probability prediction interval of photovoltaic power output in the future period, which is used to characterize the possible range of net load power of the inverter node in the future period.

[0122] Minimum net load is used to characterize the most extreme situation of power absorption surplus that may exist in the future period, that is, the situation where the photovoltaic power generation is at its maximum and the load power consumption is at its minimum.

[0123] In determining the operating mode of each inverter node in each second scheduling group, it is necessary to consider the minimum net load value. This is to ensure that the excess photovoltaic power can be fully absorbed even under the worst-case photovoltaic output surplus conditions, so as to avoid wasting photovoltaic power and improve user benefits.

[0124] The maximum net load is used to characterize the most extreme power shortage situation that may exist in the future period, that is, the situation where the photovoltaic power generation is at its minimum and the load power consumption is at its maximum.

[0125] In determining the operating mode of each inverter node in each second scheduling group, the maximum net load needs to be considered. This is to ensure that the inverter node can guarantee the power supply reliability and stability of the user's power load under the worst power load shortage conditions, and avoid reducing the user's power experience.

[0126] For example, assuming the load power consumption probability interval for an inverter node in a future time period is [Lmin, Lmax], where Lmin represents the minimum value of the load power consumption probability interval and Lmax represents the maximum value of the load power consumption probability interval, and the photovoltaic output probability prediction interval for a future time period is [Pmin, Pmax], where Pmin represents the minimum value of the photovoltaic output probability interval and Pmax represents the maximum value of the photovoltaic output probability interval, then based on the load power consumption probability prediction interval and the photovoltaic output probability of each inverter node in the future time period... The net load of the inverter node is obtained by subtracting the maximum value of the photovoltaic output probability prediction interval from the minimum value of the future load electricity consumption probability prediction interval. That is, the net load minimum value = [Lmin-Pmax]. The net load maximum value of the inverter node is obtained by subtracting the minimum value of the future photovoltaic output probability prediction interval from the maximum value of the future load electricity consumption probability prediction interval. That is, the net load maximum value = [Lmax-Pmin]. Based on this, the net load uncertainty interval is obtained as [Lmin-Pmax, Lmax-Pmin].

[0127] For example, assuming the probability interval for load consumption of the inverter node in the future period is [3, 8], and the probability prediction interval for photovoltaic output in the future period is [5, 9], then the minimum net load = [3-9] = [-6], and the maximum net load = [8-5] = [3]. The uncertainty interval of the inverter node's net load is determined based on the minimum and maximum net load values. The minimum net load = [-6] represents the most extreme surplus situation, that is, the maximum photovoltaic power generation is 9 kW, the minimum load power consumption is 3 kW, and there is a power surplus of 6 kW. The maximum net load = [3] represents the most extreme power shortage situation, that is, the minimum photovoltaic power generation is 5 kW, the maximum load power consumption is 8 kW, and there is a power shortage of 3 kW.

[0128] In this embodiment, the operating mode of each inverter node in each second scheduling group is determined based on the electricity price forecast data corresponding to the second scheduling group, the minimum net load and the maximum net load of each inverter node. Based on this, the operating mode determined for each inverter node can not only cope with the worst possible photovoltaic output surplus situation, but also cope with the worst possible electricity load deficit situation. That is to say, no matter how the actual net load of each inverter node is in the future, its corresponding operating mode can cover the power regulation requirements of its actual net load. Thus, this embodiment can not only improve user benefits, but also avoid reducing the user's electricity experience by maintaining the power supply reliability and stability of the electricity load. It improves the adaptability between the operating mode determined for each inverter node and the actual demand, and also improves the robustness of the operating status of each inverter node in the future.

[0129] In an exemplary embodiment, the determination of the operating mode of each inverter node in each second dispatch group as either a self-consumption mode or a generation priority mode based on the electricity price forecast data corresponding to the second dispatch group, the minimum net load of each inverter node, and the maximum net load of each inverter node includes:

[0130] If the electricity price forecast data corresponding to the second dispatch group indicates that the grid sales price is greater than or equal to the preset sales price threshold, and the maximum net load of the corresponding inverter node is greater than 0, then the operating mode of the corresponding inverter node is determined to be the self-consumption mode; in the self-consumption mode, the baseline power of each inverter node corresponds to the maximum net load.

[0131] If the electricity price forecast data corresponding to the second dispatch group indicates that the grid purchase price is greater than or equal to the preset purchase price threshold, and the minimum net load of the corresponding inverter node is less than 0, then the operating mode of the corresponding inverter node is determined to be the generation priority mode; in the generation priority mode, the baseline power of each inverter node corresponds to the minimum net load.

[0132] Among them, the grid sales price refers to the price at which power suppliers sell electricity to the corresponding regional grid. It can be understood as the unit price of electricity that inverter nodes need to pay to purchase electricity from the corresponding regional grid of the dispatching group.

[0133] The grid purchase price refers to the price at which an electricity supplier buys electricity from inverter nodes in a corresponding regional power grid. It can be understood as the price per unit of electricity that an inverter node can obtain by selling electricity to the corresponding regional power grid.

[0134] The preset sales price threshold and the preset purchase price threshold can be determined based on the historical electricity price data of the regional power grid corresponding to the second dispatching group.

[0135] Optionally, the preset sales price threshold can be the higher value of the historical sales price of the corresponding regional power grid, and the preset purchase price threshold can be the higher value of the historical purchase price.

[0136] For example, the preset electricity sales price threshold can be the upper quartile of the historical electricity sales price of the corresponding regional power grid within a historical period. The upper quartile is the 75th percentile of the power grid electricity sales price represented by the electricity sales price data of the corresponding regional power grid within a historical period. The method for determining the preset purchase price threshold can be the same as the method for determining the preset electricity sales price threshold, so it will not be described again here. Optionally, the historical period can be a historical month, historical day, or other period corresponding to the future period.

[0137] When the maximum net load of the corresponding inverter node is greater than 0, it means that there is a power supply gap in the electrical load of the inverter node. In this case, the inverter node has no room to sell electricity to the corresponding regional grid. Therefore, this situation will directly determine the operating mode of the corresponding inverter node as self-consumption mode.

[0138] When the maximum net load of the corresponding inverter node is less than 0, it means that the photovoltaic power generation of the inverter node will inevitably have a surplus after being absorbed by the power load. In this case, the inverter node is fully capable of meeting the power demand of the power load, and therefore has the capacity to sell electricity to the corresponding regional power grid. Therefore, this situation will directly determine the operating mode of the corresponding inverter node as the power generation priority mode.

[0139] In the self-consumption mode, the baseline power of each inverter node corresponds to the maximum net load. This means that the baseline power corresponding to the self-consumption mode is determined based on the maximum net load, thereby ensuring that the self-consumption mode can cope with the worst possible power load shortage in the future.

[0140] In the generation priority mode, the baseline power of each inverter node corresponds to the minimum net load. This means that the baseline power corresponding to the generation priority mode is determined based on the minimum net load, thereby ensuring that the generation priority mode can cope with the worst possible photovoltaic power output surplus in the future.

[0141] In the self-consumption mode, it is necessary to ensure power supply priority to avoid power shortage of the load. The main focus of this mode is the concern that the photovoltaic output is insufficient. Therefore, when the maximum net load of the corresponding inverter node is greater than 0, the operating mode of the corresponding inverter node is determined to be the self-consumption mode. For example, assuming that the maximum net load in the uncertainty interval of the net load is [3], it means that the inverter node needs to ensure that the energy storage battery has a reserved discharge capacity of 3kW, or needs to cope with the purchase of 3kW of power from the corresponding regional grid. Based on this, even in the most extreme case where the photovoltaic output is the smallest and the load power consumption is the largest, it can be ensured that the load will not lose power.

[0142] In the generation-priority mode, it is necessary to ensure priority consumption to avoid overload. The primary focus of this mode is to prevent the waste of surplus photovoltaic power. Therefore, if the minimum net load of the corresponding inverter node is less than 0, the operating mode of the corresponding inverter node is determined to be the generation-priority mode. For example, assuming the minimum net load in the uncertainty interval is [-6], this indicates that the inverter node needs to ensure that the energy storage battery has a reserved charging channel of at least 6 kW to ensure that it can receive all surplus photovoltaic power when the photovoltaic output is at its maximum, so as to avoid wasting electrical energy.

[0143] In this embodiment, if the power load of the inverter node will inevitably experience a power supply shortage in the future, then the operating mode of the corresponding inverter node in the future is determined to be a self-consumption mode. If the photovoltaic power generation of the inverter node will inevitably have a surplus even after being absorbed by the power load in the future, then the operating mode of the corresponding inverter node in the future is determined to be a power generation priority mode. Based on this, this embodiment can not only improve user benefits, but also avoid reducing the user's power consumption experience by maintaining the power supply reliability and stability of the power load. It improves the adaptability between the operating mode determined for each inverter node and the actual demand, and also improves the robustness of the operating state of each inverter node in the future.

[0144] When both the maximum and minimum net load values ​​are greater than 0, it indicates that even in the most extreme case where photovoltaic power generation is at its maximum and load power consumption is at its minimum, there will definitely be a power shortage. Therefore, in this case, the inverter node will control the energy storage battery to discharge directly, and the discharge power of the energy storage battery will be determined within the net load uncertainty range.

[0145] When both the maximum and minimum net load values ​​are less than 0, it indicates that even in the most extreme case where photovoltaic power generation is at its minimum and load power consumption is at its maximum, there will still be a surplus in photovoltaic output. Therefore, in this case, the inverter node will directly charge the energy storage battery, and the charging power of the energy storage battery will be determined within the uncertainty range of the net load.

[0146] In an exemplary embodiment, the determination of the net load uncertainty interval for each inverter node based on the predicted load consumption probability interval and the predicted photovoltaic output probability interval for each inverter node in the future time period includes:

[0147] Based on the risk preference coefficient of each inverter node, the load power consumption probability prediction interval for the future time period of each inverter node is narrowed to obtain the narrowed load power consumption probability prediction interval for the future time period. The maximum value of the narrowed load power consumption probability prediction interval is less than the maximum value of the original load power consumption probability prediction interval, and the minimum value of the narrowed load power consumption probability prediction interval is greater than the minimum value of the original load power consumption probability prediction interval.

[0148] Based on the risk preference coefficient of each inverter node, the probability prediction interval of photovoltaic output for the future time period of each inverter node is narrowed to obtain the narrowed probability prediction interval of photovoltaic output for the future time period. The maximum value of the narrowed probability prediction interval of photovoltaic output is less than the maximum value of the original photovoltaic output probability prediction interval, and the minimum value of the narrowed probability prediction interval of photovoltaic output is greater than the minimum value of the original photovoltaic output probability prediction interval.

[0149] Based on the predicted range of load electricity consumption probability after the contraction of the future time period and the predicted range of photovoltaic power output probability after the contraction of the future time period, the net load uncertainty range of each inverter node is determined.

[0150] The risk preference coefficient is a coefficient used to adjust the degree of contraction in the predicted ranges of future load electricity consumption and photovoltaic power output. There is a positive correlation between the risk preference coefficient and the degree of contraction in these ranges.

[0151] Optionally, the risk preference coefficient can range from 0 to 1.

[0152] Optionally, the risk preference coefficients of each inverter node can be the same or different.

[0153] Optionally, the risk preference coefficient corresponding to the inverter node can be pre-set by the user of the inverter node, automatically set based on the inverter node's operating mode, or determined based on other factors. For example, the magnitude of the risk preference coefficient can correspond to the probability of prediction error. Therefore, the lower the confidence level of the predicted load consumption probability range and photovoltaic output probability range for each inverter node in the future time period, the larger the risk preference coefficient, meaning a greater degree of contraction of the prediction range. This means more extreme cases are abandoned to reduce user costs or increase user profits.

[0154] For the prediction interval, the probability of occurrence of the value interval near the maximum value and the value interval near the minimum value is usually lower than the probability of occurrence of the middle value of the interval. Based on this, when the risk preference coefficient is larger, that is, when the maximum and minimum values ​​of the corresponding prediction interval (the prediction interval of the probability of load electricity consumption after the contraction of the future period and the prediction interval of the probability of photovoltaic power output after the contraction of the future period) are closer to the middle value of the interval, it indicates that the operation mode will give up considering the extreme cases with lower probability of occurrence. These abandoned extreme cases have little impact on the reliability and stability of power supply, but can sell more electricity to the corresponding regional power grid or buy less electricity. Therefore, the larger the risk preference coefficient, the more the operation mode will tend to consider reducing the user's electricity cost or increasing the user's income.

[0155] Correspondingly, the smaller the risk preference coefficient, the more efforts are made to avoid the omission of extreme situations. Therefore, the smaller the risk preference coefficient, the more the operating mode will tend to ensure the power supply reliability and stability of the electrical load of the user corresponding to the inverter node.

[0156] Optionally, when the risk preference coefficient corresponding to the inverter node is automatically set based on the inverter node's operating mode, the risk preference coefficient in the self-consumption mode can be greater than the risk preference coefficient in the generation priority mode.

[0157] For example, assuming the load probability interval for the inverter node in the future period is [3, 8] and the photovoltaic output probability prediction interval for the future period is [5, 9], then the net load uncertainty interval before narrowing the two prediction intervals based on the risk preference coefficient is [-6, 3]. Furthermore, assuming the load probability interval for the inverter node after narrowing the future period is [4, 7] and the photovoltaic output probability prediction interval after narrowing the future period is [6, 8], then the net load uncertainty interval after introducing the risk preference coefficient is [4-8, 7-6] = [-4, 1]. Therefore, in the self-consumption mode, the inverter node will be controlled to adjust its power with a baseline power of 1kW in the future period. Although there is a risk of exceeding the limit in power adjustment, compared to the net load uncertainty interval before introducing the risk preference coefficient, the inverter node can enable the energy storage battery to release more energy storage capacity, thereby improving the daily economic benefits for the corresponding users.

[0158] In this embodiment, based on the risk preference coefficient of each inverter node, the predicted ranges for future load consumption and photovoltaic output are narrowed to a certain extent to obtain the narrowed predicted ranges for future load consumption and photovoltaic output. Based on this, inverter nodes operating in self-consumption mode will use the maximum value of the narrowed net load as the baseline power for power regulation in the future, while inverter nodes operating in generation-priority mode will use the minimum value of the narrowed net load as the baseline power for power regulation. It can be seen that this embodiment, by introducing a risk preference coefficient, can avoid overly conservative operating modes that fail to effectively reduce user electricity costs or increase user revenue, and also avoid overly aggressive operating modes that lead to a decrease in power supply reliability and stability. Therefore, it can significantly improve the flexibility of the distributed inverter grouping and scheduling process.

[0159] In one exemplary embodiment, the method further includes:

[0160] Based on historical electricity consumption data from multiple inverter nodes, the risk preference coefficient for each inverter node is determined.

[0161] In another exemplary embodiment, the determination of the risk preference coefficient for each inverter node based on historical electricity consumption data of multiple inverter nodes includes:

[0162] Based on historical electricity consumption data from multiple inverter nodes, the cumulative cycle count of the energy storage battery corresponding to each inverter node is determined.

[0163] The first risk adjustment factor for each inverter node is determined based on the cumulative cycle number of the energy storage battery corresponding to each inverter node.

[0164] The risk preference coefficient of each inverter node is determined based on the initial risk preference coefficient of each inverter node and the corresponding first risk adjustment factor.

[0165] The cumulative cycle count of an energy storage battery refers to the total number of complete charge-discharge cycles it has completed. The cumulative cycle count characterizes the degree of aging of the battery. Batteries with a higher degree of aging are less able to withstand wider charge-discharge depths due to wear and tear. Therefore, when the cumulative cycle count of an energy storage battery is high, extreme situations requiring deep charging or deep discharging should be avoided as much as possible.

[0166] Risk preference coefficient = initial risk preference coefficient × first risk adjustment factor.

[0167] When the risk adjustment coefficient is between 0 and 1, the first risk adjustment factor is negatively correlated with the cumulative number of cycles, and positively correlated with the risk preference coefficient.

[0168] Optionally, the initial risk preference coefficient corresponding to the inverter node can be pre-set by the user corresponding to the inverter node, automatically set based on the inverter node's operating mode, or determined based on other factors. For example, the magnitude of the initial risk preference coefficient can correspond to the magnitude of the probability of prediction error.

[0169] In another exemplary embodiment, the determination of the cumulative cycle count of the energy storage battery corresponding to each inverter node based on historical electricity consumption data of multiple inverter nodes includes:

[0170] Based on historical power consumption data from multiple inverter nodes, the cumulative cycle count and average operating temperature of the energy storage battery corresponding to each inverter node are determined.

[0171] The risk preference coefficient for each inverter node is determined based on its initial risk preference coefficient and the corresponding first risk adjustment factor, including:

[0172] The second risk adjustment factor is determined based on the absolute value of the temperature difference between the average operating temperature of each inverter node and the calibrated protection temperature threshold.

[0173] The risk preference coefficient of each inverter node is determined based on the initial risk preference coefficient, the corresponding first risk adjustment factor, and the corresponding second risk adjustment factor.

[0174] The average operating temperature of the inverter node refers to the average operating temperature of the inverter node over a historical period. A higher average operating temperature of the inverter node indicates poor heat dissipation of the inverter, or that the inverter is generating a lot of heat due to performance aging. In such cases, the energy storage battery corresponding to the inverter node should not withstand a wide depth of charge and discharge. Furthermore, a higher average operating temperature also indicates a greater likelihood of exceeding the calibrated protection temperature threshold. Therefore, with the same cumulative cycle count, the absolute value of the temperature difference is positively correlated with the second risk adjustment factor.

[0175] Optionally, the second risk adjustment factor is 0 to 1.

[0176] Risk preference coefficient = initial risk preference coefficient × first risk adjustment factor × second risk adjustment factor.

[0177] Optionally, the risk preference coefficient of each inverter node can be determined based on its initial risk preference coefficient, first weight, corresponding first risk adjustment factor, second weight, and corresponding second risk adjustment factor. The first weight corresponds to the corresponding first risk adjustment factor, and the second weight corresponds to the corresponding second risk adjustment factor. Since summer worsens the heat dissipation conditions of inverter nodes, the second weight is greater than the first weight if the future period falls in summer, and the first weight can be the second weight if the future period falls in a non-summer season.

[0178] Under the same first risk adjustment factor, the second risk adjustment factor is positively correlated with the initial risk preference coefficient.

[0179] In this embodiment, the average operating temperature of the energy storage battery, together with the cumulative number of cycles of the energy storage battery, is introduced as an adjustment factor for the risk preference coefficient. Based on this, both the aging degree of the energy storage battery corresponding to the inverter node and the magnitude of the thermal stress on the inverter can be considered. Thus, the energy storage battery of the inverter node can be more flexibly avoided from being subjected to excessively wide charging and discharging depths according to the actual situation of the inverter node. In other words, by appropriately giving up the low-probability extreme cases that require deep charging or deep discharging of the energy storage battery, the safety and reliability of the inverter node can be better ensured.

[0180] In one exemplary embodiment, the method further includes:

[0181] Based on the maximum net load of each inverter node and the risk preference coefficient, the maximum net load of each inverter node after contraction is determined; the self-consumption mode is used to control the corresponding inverter node in the second scheduling group to adjust the power to the baseline power corresponding to the maximum net load after contraction in the future period.

[0182] Based on the minimum net load of each inverter node and the risk preference coefficient, the minimum net load after contraction for each inverter node is determined; the generation priority mode is used to control the corresponding inverter node in the corresponding second dispatch group to adjust its power to the baseline power corresponding to the minimum net load after contraction in the future time period.

[0183] The maximum net load after contraction is less than the maximum net load before contraction, and the minimum net load after contraction is greater than the minimum net load before contraction.

[0184] In this embodiment, based on a risk preference coefficient, the maximum and minimum net load values ​​are reduced to a certain extent to obtain the reduced maximum and minimum net load values. Based on this, inverter nodes operating in self-consumption mode will use the reduced maximum net load value as the baseline power for power regulation in future periods, while inverter nodes operating in generation-priority mode will use the reduced minimum net load value as the baseline power for power regulation in future periods. It can be seen that this embodiment, by introducing a risk preference coefficient, can avoid overly conservative operating modes that fail to effectively reduce user electricity costs or increase user benefits, and also avoid overly aggressive operating modes that lead to a decrease in power supply reliability and stability. Therefore, it can significantly improve the flexibility of the distributed inverter grouping and scheduling process.

[0185] In an exemplary embodiment, the above-mentioned determination of user profile information corresponding to each inverter node based on historical electricity consumption data of multiple inverter nodes includes:

[0186] Based on historical electricity consumption data from multiple inverter nodes, the daily load curve and multiple electricity consumption characteristics of each inverter node are determined.

[0187] Based on the daily load curves of each inverter node, multiple electricity consumption characteristics, and the characteristic thresholds corresponding to each electricity consumption characteristic, the user profile information corresponding to each inverter node is determined.

[0188] The daily load curve of an inverter node refers to the curve showing how the power consumption of the load corresponding to the inverter node changes over a complete natural day. The daily load curve of an inverter node can characterize the typical time-series characteristics of the electricity consumption behavior of the users corresponding to the inverter node.

[0189] Optionally, the daily load curve of the inverter node can be a typical daily load curve for a weekday or a typical daily load curve for a weekend.

[0190] The power consumption characteristics of an inverter node refer to statistical indicators about the load power consumption extracted from the historical power consumption data of the inverter node.

[0191] Optionally, the electricity consumption characteristics of the inverter node may include the daily electricity consumption, maximum load, minimum load, peak-to-valley difference, peak-to-valley electricity consumption ratio, valley-to-valley electricity consumption ratio, night-to-night electricity consumption ratio, load fluctuation rate, regularity indicators, and other characteristics that can reflect the regularity of electricity consumption behavior. Further optionally, the electricity consumption characteristics of the inverter node may also include the correlation coefficient between load power consumption and historical weather forecast data, the correlation coefficient between load power consumption and historical electricity price data, or other characteristics that can be used to reflect the regularity of electricity consumption behavior, and may also include characteristics such as the start-stop frequency of high-power appliances.

[0192] Feature thresholds are pre-set thresholds for classifying different levels of electricity consumption characteristics. In other words, by comparing each electricity consumption characteristic of an inverter node with its corresponding feature threshold, the degree range of each electricity consumption characteristic at the inverter node can be determined.

[0193] For example, if the power consumption characteristics of an inverter node include the daily power consumption of the inverter node, then the characteristic threshold corresponding to the daily power consumption is the daily power consumption threshold.

[0194] In this embodiment, based on historical electricity consumption data of multiple inverter nodes, the daily load curve and multiple electricity consumption characteristics of each inverter node are determined. Then, based on the daily load curve, multiple electricity consumption characteristics, and the characteristic thresholds corresponding to each electricity consumption characteristic, the user profile information corresponding to each inverter node is determined. Based on this, the user profile information corresponding to each inverter node can consider both the temporal characteristics and the regularity of user electricity consumption behavior. Thus, the user profile information corresponding to each inverter node can more comprehensively and objectively represent the characteristics of the corresponding user's electricity consumption behavior, thereby improving the accuracy of the operating mode determined for each inverter node and its adaptability to actual needs.

[0195] In an exemplary embodiment, the above-mentioned determination of the daily load curve and multiple electricity consumption characteristics of each inverter node based on historical electricity consumption data of multiple inverter nodes includes:

[0196] Based on historical electricity consumption data from multiple inverter nodes, the initial daily load curve and multiple electricity consumption characteristics of each inverter node are determined.

[0197] Based on the electricity consumption level among multiple electricity consumption characteristics, the individual scaling factor of each inverter node is determined; there is a positive correlation between the electricity consumption level and the individual scaling factor;

[0198] Based on the individual scaling factor of each inverter node, the initial daily load curve corresponding to each inverter node is scaled to obtain the daily load curve of each inverter node.

[0199] The individual scaling factor is a parameter used to characterize the degree of individual difference between the electricity consumption level of each inverter node and the benchmark electricity consumption level of the user behavior cluster to which it belongs. In other words, it is the proportional relationship between the electricity consumption level of each inverter node and the benchmark electricity consumption level of the user behavior cluster.

[0200] In this embodiment, based on the individual scaling factor, the initial daily load curves of inverter nodes with similar curve shapes but different power consumption levels are corrected to obtain more objective daily load curves for each inverter node. This ensures the accuracy of the user profile information for each inverter node and ensures the adaptability of the operating mode determined for each inverter node to the actual needs.

[0201] In one exemplary embodiment, the acquisition of historical electricity consumption data, geographical coordinates, and power suppliers for multiple inverter nodes includes:

[0202] Obtain initial historical electricity consumption data, geographical coordinates, and power suppliers for multiple inverter nodes;

[0203] Data cleaning is performed on the initial historical electricity consumption data of multiple inverter nodes to obtain the historical electricity consumption data of multiple inverter nodes.

[0204] This includes data cleaning of the initial historical electricity consumption data of multiple inverter nodes. This can involve interpolating and filtering missing and outlier values ​​in the initial historical electricity consumption data to ensure that the obtained historical electricity consumption data of multiple inverter nodes is aligned with the electricity price forecast data and weather forecast data for future periods at the time granularity.

[0205] In an exemplary embodiment, the above-mentioned determination of user profile information corresponding to each inverter node based on the daily load curve of each inverter node, multiple electricity consumption characteristics, and characteristic thresholds corresponding to each electricity consumption characteristic includes:

[0206] Obtain the energy storage capacity corresponding to each inverter node, as well as the rated power of each electrical load connected to each inverter node;

[0207] Based on the energy storage capacity corresponding to each inverter node, the rated power of each corresponding electrical load, multiple electricity consumption characteristics, and the characteristic thresholds corresponding to each electricity consumption characteristic, the user tag of each inverter node is determined.

[0208] Based on the daily load curves and user tags corresponding to each inverter node, user profile information corresponding to each inverter node is determined.

[0209] Among them, the user tag of the inverter node refers to the semantic tag obtained by qualitatively classifying the electricity consumption behavior of the user corresponding to the inverter node in terms of electricity consumption dimensions such as energy storage capacity, rated power of electrical load, and electricity consumption characteristics.

[0210] Optionally, the user tags for inverter nodes can be "large peak-valley difference / small peak-valley difference", "high energy consumption at night / low energy consumption at night", "regular electricity consumption behavior / irregular electricity consumption behavior", etc.

[0211] Based on the daily load curves and user tags corresponding to each inverter node, the user behavior of multiple inverter nodes can be clustered using the K-means clustering algorithm to obtain the following user behavior clusters: daytime home users with peak load during midday, working users with peak load in both morning and evening due to early departure and late return, nighttime high-consumption users with peak load during the evening, photovoltaic self-consumption users with high photovoltaic power generation and high load during the day, or other types. Based on the user behavior clusters corresponding to each inverter node, the user profile information corresponding to each inverter node is determined.

[0212] In this embodiment, the user profile information corresponding to each inverter node is determined based on the daily load curve and user tags of each inverter node. The user tags are determined by comprehensively considering the energy storage capacity of each inverter node, the rated power of each electrical load, multiple electricity consumption characteristics, and the characteristic thresholds corresponding to each electricity consumption characteristic. Based on this, the user profile information corresponding to each inverter node can provide more targeted and objective load electricity prediction basis for each inverter node, thereby improving the accuracy of the operating mode determined for each inverter node and its adaptability to actual needs.

[0213] In one exemplary embodiment, after determining the operating mode of each inverter node, the method further includes:

[0214] The parameter set is determined based on the mode flag, power baseline, and regulation margin of the operating mode;

[0215] The parameter set is sent to each inverter node so that each inverter node can switch from the historical mode to the operating mode in the future so that the power is adjusted to the baseline power corresponding to the operating mode.

[0216] Among them, the mode flag refers to the unique code flag of the operating mode issued by the cloud server.

[0217] Optionally, when the operating modes include self-consumption mode, generation priority mode and grid balancing mode, the mode flag for self-consumption mode can be, for example, 0x01, the mode flag for generation priority mode can be, for example, 0x02, and the mode flag for grid balancing mode can be, for example, 0x03.

[0218] The adjustment margin refers to the maximum allowable fluctuation range of the actual power of the inverter node from the baseline power. It is a buffer zone that absorbs the uncertainty interval of probability prediction. Based on the adjustment margin and the baseline power, the allowable power operating range of the inverter node's power regulation process can be formed: [baseline power - adjustment margin, baseline power + adjustment margin].

[0219] After receiving the parameter set, the inverter node can determine the specific operating mode based on the mode flag and determine its own power regulation target based on the power baseline and regulation margin.

[0220] Alternatively, the parameter set can be determined based on the operating mode flag, power baseline, regulation margin, and the target remaining battery percentage for each inverter node. Based on this, after receiving the parameter set, the inverter node can determine its own power regulation target based on the power baseline and regulation margin, so as to maintain the remaining battery percentage close to the target remaining battery percentage while maintaining the power close to the baseline power.

[0221] The target remaining percentage of energy storage battery refers to the safe upper and lower limits that the remaining percentage of energy storage battery corresponding to each inverter node must maintain in the future, calculated by the motion server based on a robust optimization algorithm. This is a "physical guardrail" to prevent the energy storage battery from overcharging or over-discharging in extreme situations.

[0222] For example, suppose the cloud server predicts that the weather will be cloudy in the next hour, meaning that the photovoltaic output is highly uncertain, and the users corresponding to the inverter node are at work, meaning that the load electricity consumption is less uncertain. Based on this, the operating mode of the inverter node is determined to be self-consumption mode. Thus, assuming the power baseline is -1kW, the adjustment margin is ±1.5kW, and the target remaining battery capacity percentage is 20%~80%, the parameter set determined by the cloud server for the operating mode can be expressed as: {Mode: 0x01, Power_Base: -1.0 kW, Power_Margin: ±1.5 kW, SOC_Band: [20%, 80%]}. In this parameter set, Mode represents the operating mode, Power_Base represents the power baseline, Power_Margin represents the adjustment margin, and SOC_Band represents the target remaining battery capacity percentage.

[0223] In this embodiment, after determining the operating mode of each inverter node, the cloud server does not directly control the underlying PWM of each inverter node. Instead, it sends a set of parameters to each inverter node, including mode flag, baseline power, regulation margin, and target remaining battery charge percentage. This lightweight sending method can reduce the real-time communication between the cloud server and multiple inverter nodes, thereby improving the communication reliability between the cloud server and multiple inverter nodes.

[0224] like Figure 3 As shown, the application process of the above-mentioned distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction will be illustrated below with a detailed embodiment:

[0225] Step 302: Obtain historical electricity consumption data, geographical coordinates, and power suppliers for multiple inverter nodes;

[0226] Step 304: Based on the historical electricity consumption data of multiple inverter nodes, determine the daily load curve and multiple electricity consumption characteristics of each inverter node.

[0227] Step 306: Obtain the energy storage capacity corresponding to each inverter node, and the rated power of each electrical load connected to each inverter node.

[0228] Step 308: Determine the user tag for each inverter node based on the energy storage capacity corresponding to each inverter node, the rated power of each corresponding electrical load, multiple electricity consumption characteristics, and the characteristic threshold corresponding to each electricity consumption characteristic.

[0229] Step 310: Based on the daily load curve and user tags corresponding to each inverter node, determine the user profile information corresponding to each inverter node, and divide the multiple inverter nodes into at least two scheduling groups based on the geographical coordinates of multiple inverter nodes and power suppliers.

[0230] Step 312: Based on multiple user profile information, weather forecast data for future periods, and electricity price prediction data for the regional power grid corresponding to each dispatching group for future periods, determine the load power consumption probability prediction range and photovoltaic output probability prediction range for each inverter node in the future periods.

[0231] Step 314: Based on the probability prediction range of load power consumption of each inverter node in the future time period, the probability prediction range of photovoltaic power output in the future time period, and the grid load information corresponding to each dispatch group, determine whether each dispatch group has grid balancing demand. Based on the determination result of the grid balancing demand of each dispatch group, execute step 316 or step 318.

[0232] Step 316: Determine at least one scheduling group with grid balancing needs from at least two scheduling groups as at least one first scheduling group;

[0233] Step 318: Determine at least one dispatch group that does not have grid balancing requirements from at least two dispatch groups as at least one second dispatch group;

[0234] Step 320: Determine that the operating mode of each inverter node in each first scheduling group is the balanced grid mode; the balanced grid mode is used to control multiple inverter nodes in at least one first scheduling group to adjust the power in the same power adjustment direction in the future time period, so as to adjust the power to the baseline power corresponding to each inverter node, and the baseline power of each inverter node corresponds to the energy storage capacity.

[0235] Step 322: Based on the risk preference coefficient of each inverter node, the load power consumption probability prediction interval for the future time period of each inverter node is narrowed to obtain the narrowed load power consumption probability prediction interval for the future time period; the maximum value of the narrowed load power consumption probability prediction interval is less than the maximum value of the original load power consumption probability prediction interval, and the minimum value of the narrowed load power consumption probability prediction interval is greater than the minimum value of the original load power consumption probability prediction interval.

[0236] Step 324: Based on the risk preference coefficient of each inverter node, the photovoltaic output probability prediction interval for the future time period of each inverter node is narrowed to obtain the photovoltaic output probability prediction interval after the future time period is narrowed. The maximum value of the photovoltaic output probability prediction interval after the narrowing is less than the maximum value of the photovoltaic output probability prediction interval before the narrowing, and the minimum value of the photovoltaic output probability prediction interval after the narrowing is greater than the minimum value of the photovoltaic output probability prediction interval before the narrowing.

[0237] Step 326: Based on the predicted range of load electricity consumption probability after the contraction of the future time period and the predicted range of photovoltaic power output probability after the contraction of the future time period, determine the net load uncertainty range for each inverter node; the minimum net load of the net load uncertainty range is determined by the difference between the minimum value of the predicted range of load electricity consumption probability and the maximum value of the predicted range of photovoltaic power output probability, and the maximum net load of the net load uncertainty range is determined by the difference between the maximum value of the predicted range of load electricity consumption probability and the minimum value of the predicted range of photovoltaic power output probability.

[0238] Step 328: If the electricity price prediction data corresponding to the second dispatch group indicates that the grid sales price is greater than or equal to the preset sales price threshold, and the maximum net load of the corresponding inverter node is greater than 0, then the operating mode of the corresponding inverter node is determined to be the self-consumption mode; in the self-consumption mode, the baseline power of each inverter node corresponds to the maximum net load.

[0239] Step 330: If the electricity price prediction data corresponding to the second dispatch group indicates that the grid purchase price is greater than or equal to the preset purchase price threshold, and the minimum net load of the corresponding inverter node is less than 0, then the operating mode of the corresponding inverter node is determined to be the generation priority mode; in the generation priority mode, the baseline power of each inverter node corresponds to the minimum net load.

[0240] In this embodiment, at least one scheduling group with grid balancing needs is defined as at least one first scheduling group, and each inverter node in the first scheduling group operates in a grid balancing mode. At least one scheduling group without grid balancing needs is defined as at least one second scheduling group, and each inverter node in the second scheduling group determines whether to operate in a self-consumption mode or a generation priority mode based on electricity price forecast data and the specific situation of the net load uncertainty range. Based on this, multiple inverter nodes can avoid operating in an "island" state, and can instead perform clustered coordinated response based on the grid load information of the grid in the corresponding area of ​​the scheduling group. On the other hand, it can consider both the user profile information corresponding to each inverter node and the grid load information of the grid in the area where the inverter node is located. Therefore, it achieves a comprehensive consideration of user electricity consumption behavior and regional grid load balancing, and can accurately determine the operating mode of the inverter node with high adaptability to actual needs.

[0241] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0242] In one exemplary embodiment, a cloud server is also provided for executing any of the methods described above in the embodiments of the distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction.

[0243] In one exemplary embodiment, such as Figure 1 As shown, a scheduling system 100 is also provided, including a cloud server 102 as described in any of the above embodiments of cloud server 102 and a plurality of inverter nodes 104.

[0244] Inverter node 104 is used to switch from historical mode to operating mode when entering a future time period, so as to adjust the power to the baseline power corresponding to the operating mode in the future time period.

[0245] In an exemplary embodiment, inverter node 104 is also configured to adjust the operating priority of the photovoltaic panel branch, the energy storage battery branch, and the inverter node 104 branch based on the operating mode.

[0246] In the case of self-consumption mode, the working priority of inverter node 104 branch is higher than that of photovoltaic panel branch, and energy storage battery branch is used as a supplementary option for power supply gap of electrical load.

[0247] When the operation mode is the power generation priority mode, the working priority of the inverter node 104 branch is higher than that of the energy storage battery branch, and the photovoltaic power generation of the photovoltaic panel branch is at full capacity.

[0248] When the operating mode is the balanced grid mode, the power regulation of the inverter node 104 branch is constrained by the baseline power of the operating mode, and the working priority of the energy storage battery branch is higher than that of the photovoltaic panel branch.

[0249] Among them, the working priority can be understood as the scheduling priority in terms of power allocation.

[0250] The use of energy storage battery branches as a supplementary option for power supply gaps in electrical loads means that when the photovoltaic power generation of the photovoltaic panel branch is insufficient to meet the power demand of the electrical load, the energy storage battery branch will supplement the power supply gap between the power demand of the load and the photovoltaic power generation.

[0251] When the operating mode is the balanced grid mode, the energy storage battery branch charges and discharges in full accordance with the operating mode indicated by the cloud server.

[0252] In this embodiment, the inverter node flexibly adjusts the working priority of the photovoltaic panel branch, the energy storage battery branch, and the inverter node branch under different operating modes, so that the power regulation of the inverter node can adapt to the baseline power requirements of different operating modes. Based on this, the response speed and power regulation accuracy of the inverter node to different operating modes can be improved.

[0253] In one exemplary embodiment, the scheduling system further includes a DC bus connected to multiple inverter nodes;

[0254] In the case of entering a future time period, regarding the switch from historical mode to operating mode, the inverter node is specifically used to determine the power difference between the baseline power and the current power; perform linear interpolation on the power difference according to a preset rate of change to obtain the current inner loop reference value; determine the voltage change rate of the DC bus; determine the compensated current inner loop reference value based on the negative feedback of the voltage change rate and the current inner loop reference value; limit the compensated current inner loop reference value within the safety envelope of the inverter node's rated current to obtain the current inner loop reference value with limited amplitude; and switch from historical mode to operating mode based on the current inner loop reference value with limited amplitude.

[0255] The DC bus refers to the common DC link used to transmit DC power within the inverter node. It can be understood as the node for collecting and distributing DC power within the inverter.

[0256] The inner current loop reference value refers to the reference current value of the inner current loop that needs to be input to the inverter node.

[0257] After linearly interpolating the power difference according to a preset rate of change, the linearly interpolated power value is converted into a current value to obtain the inner current loop reference value.

[0258] Optionally, the preset rate of change can be 1 kW / s, 2 kW / s, 3 kW / s, or other values. Linear interpolation of the power difference according to the preset rate of change is used to avoid abrupt changes in the inner current loop reference value.

[0259] Optionally, a preset rate of change can be pre-stored in the control module of the inverter node. Based on this, the power difference is linearly interpolated according to the preset rate of change in the control module to obtain the inner loop reference value of the current. Optionally, this embodiment can be executed by the control module of the inverter node.

[0260] Negative feedback of voltage change rate refers to inverting the voltage change rate. Based on this, and the current inner loop reference value, the compensated current inner loop reference value is determined by inverting the voltage change amount and adding it as a feedforward compensation amount to the current inner loop reference value to obtain the compensated current inner loop reference value.

[0261] Based on the compensated inner current reference value, the duty cycle of the switching transistor can be changed in advance before the DC bus voltage drops or surges significantly, so as to achieve active damping of bus voltage fluctuations.

[0262] The safety envelope of the rated current of an inverter node is a current safety boundary line determined based on the rated current of the inverter node to ensure that the inverter node operates in a safe operating range.

[0263] Limiting the compensated inner current reference value within the safe envelope of the inverter node's rated current is to ensure that even if the feedforward compensation is too large or the dispatch system is subjected to external disturbances, the output duty cycle of the inner current loop is absolutely limited, so as to ensure that the switching stress of the power devices does not exceed the safe threshold.

[0264] Based on the current inner loop reference value after limiting the amplitude, the PWM (Pulse Width Modulation) module of the inverter node is driven to operate, and the duty cycle of each switch is adjusted. This enables a smooth transfer of energy between the photovoltaic panel branch, the energy storage battery branch and the inverter node branch during the transition from the historical mode to the operating mode. This ensures that the grid-connected current THD (Total Harmonic Distortion) meets the standard requirements, and there is no power outage at the load end.

[0265] To mitigate grid-connected current surges and DC bus voltage fluctuations that may occur during the switching from historical to operational modes when the inverter node receives the operational mode from the cloud server and transitions from historical mode to operational mode in the future, this embodiment employs a linear interpolation transition slope smoothing process. The negative feedback of the voltage change rate is used as a feedforward compensation, and a coordinated control strategy with amplitude limiting is applied to the compensated inner current reference value. This ensures that the inverter node can smoothly switch from historical mode to operational mode, achieving seamless power outages for the load. Simultaneously, it ensures the safety and stability of the inverter node and the corresponding regional power grid.

[0266] In one exemplary embodiment, the inverter node is also configured to detect the communication connection status with the motion server; if the communication connection status is detected as interrupted, and if a future time period is entered, it switches from the historical mode to the self-consumption mode.

[0267] In an exemplary embodiment, the inverter node is further specifically configured to, upon detecting a communication interruption, determine, based on historical electricity consumption data stored locally on the inverter node, a probability prediction range for load consumption in the future time period predicted locally by the inverter node, a probability prediction range for photovoltaic output in the future time period, and a net load uncertainty range predicted locally by the inverter node; in self-consumption mode, the baseline power of the inverter node corresponds to the maximum net load value in the locally predicted net load uncertainty range.

[0268] Optionally, if the inverter node detects that the communication connection has been restored, the inverter node will reacquire the operating mode determined by the cloud server in order to adjust the power to the baseline power corresponding to the operating mode determined by the cloud server in the future.

[0269] In this embodiment, if the inverter node detects a communication interruption with the cloud server, it will automatically switch from the historical mode to the self-consumption mode in the future. This ensures that the power load of the user corresponding to the inverter node will not experience power outages. Once the communication connection is restored, the cloud server will synchronize the operating mode determined for the user. It can be seen that the inverter node's network outage fault-tolerant self-determination mechanism can significantly improve the power supply stability and reliability of the inverter node.

[0270] Based on the same inventive concept, this application also provides a distributed inverter grouping and scheduling device based on cloud-edge collaboration and user behavior prediction for implementing the distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the distributed inverter grouping and scheduling device based on cloud-edge collaboration and user behavior prediction provided below can be found in the limitations of the distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction described above, and will not be repeated here.

[0271] In one exemplary embodiment, such as Figure 4 As shown, a distributed inverter grouping and scheduling device based on cloud-edge collaboration and user behavior prediction is provided, which is applied in a cloud server. The device includes: an acquisition module 402, a first determination module 404, a second determination module 406, and a third determination module 408, wherein:

[0272] The acquisition module 402 is used to acquire historical electricity consumption data, geographical coordinates, and power suppliers for multiple inverter nodes.

[0273] The first determining module 404 is used to determine the user profile information corresponding to each inverter node based on the historical electricity consumption data of multiple inverter nodes, and to divide the multiple inverter nodes into at least two scheduling groups based on the geographical coordinates of the multiple inverter nodes and the power supplier.

[0274] The second determining module 406 is used to determine the probability prediction range of load electricity consumption and the probability prediction range of photovoltaic output for each inverter node in the future period based on multiple user profile information, weather forecast data for the future period, and electricity price prediction data of the regional power grid corresponding to each dispatching group in the future period.

[0275] The third determining module 408 is used to determine the operating mode of each inverter node based on the load power consumption probability prediction range of each inverter node in the future time period, the photovoltaic power output probability prediction range in the future time period, and the grid load information of the grid in the corresponding area of ​​each dispatch group; the operating mode is used to control the corresponding inverter node to adjust the power to the baseline power corresponding to the operating mode in the future time period.

[0276] In an exemplary embodiment, the third determining module 408 is specifically used to divide at least two scheduling groups into at least one first scheduling group and / or at least one second scheduling group based on the load consumption probability prediction range of each inverter node in the future time period, the photovoltaic output probability prediction range in the future time period, and the grid load information corresponding to each scheduling group; determine that the operating mode of each inverter node in each first scheduling group is a balanced grid mode; the balanced grid mode is used to control multiple inverter nodes in at least one first scheduling group to perform power adjustment in the same power adjustment direction in the future time period, so as to adjust the power to the baseline power corresponding to each inverter node, and the baseline power of each inverter node corresponds to the energy storage capacity; and determine that the operating mode of each inverter node in each second scheduling group is a self-consumption mode or a power generation priority mode based on the load consumption probability prediction range of each inverter node in the future time period, the photovoltaic output probability prediction range in the future time period, and the electricity price prediction data of the regional grid corresponding to each second scheduling group in the future time period.

[0277] In an exemplary embodiment, the third determining module 408 is specifically used to determine the net load uncertainty interval of each inverter node based on the load power consumption probability prediction interval and the photovoltaic power output probability prediction interval of each inverter node in the future time period; the minimum net load of the net load uncertainty interval is determined by the difference between the minimum value of the load power consumption probability prediction interval and the maximum value of the photovoltaic power output probability prediction interval, and the maximum net load of the net load uncertainty interval is determined by the difference between the maximum value of the load power consumption probability prediction interval and the minimum value of the photovoltaic power output probability prediction interval; based on the electricity price prediction data corresponding to the second scheduling group, the minimum net load of each inverter node and the maximum net load, the operating mode of each inverter node in each second scheduling group is determined to be either self-consumption mode or power generation priority mode.

[0278] In an exemplary embodiment, the third determining module 408 is specifically used to determine the operating mode of the corresponding inverter node as a self-consumption mode when the electricity price forecast data corresponding to the second scheduling group indicates that the grid sales price is greater than or equal to a preset sales price threshold and the maximum net load of the corresponding inverter node is greater than 0; in the self-consumption mode, the baseline power of each inverter node corresponds to the maximum net load; when the electricity price forecast data corresponding to the second scheduling group indicates that the grid purchase price is greater than or equal to a preset purchase price threshold and the minimum net load of the corresponding inverter node is less than 0, the operating mode of the corresponding inverter node is determined as a generation priority mode; in the generation priority mode, the baseline power of each inverter node corresponds to the minimum net load.

[0279] In an exemplary embodiment, the third determining module 408 is specifically used to shrink the load power consumption probability prediction interval for each inverter node in the future time period based on the risk preference coefficient of each inverter node, to obtain a shrunken load power consumption probability prediction interval for the future time period; the maximum value of the shrunken load power consumption probability prediction interval is less than the maximum value of the load power consumption probability prediction interval before shrinkage, and the minimum value of the shrunken load power consumption probability prediction interval is greater than the minimum value of the load power consumption probability prediction interval before shrinkage; based on the risk preference coefficient of each inverter node, shrink the photovoltaic output probability prediction interval for each inverter node in the future time period, to obtain a shrunken photovoltaic output probability prediction interval for the future time period; the maximum value of the shrunken photovoltaic output probability prediction interval is less than the maximum value of the photovoltaic output probability prediction interval before shrinkage, and the minimum value of the shrunken photovoltaic output probability prediction interval is greater than the minimum value of the photovoltaic output probability prediction interval before shrinkage; based on the shrunken load power consumption probability prediction interval and the shrunken photovoltaic output probability prediction interval for the future time period, determine the net load uncertainty interval for each inverter node.

[0280] In an exemplary embodiment, the first determining module 404 is specifically used to determine the daily load curve and multiple electricity consumption characteristics of each inverter node based on the historical electricity consumption data of multiple inverter nodes; and to determine the user profile information corresponding to each inverter node based on the daily load curve, multiple electricity consumption characteristics and the characteristic thresholds corresponding to each electricity consumption characteristic.

[0281] In an exemplary embodiment, the first determining module 404 is specifically used to obtain the energy storage capacity corresponding to each inverter node and the rated power of each electrical load connected to each inverter node; determine the user tag of each inverter node based on the energy storage capacity corresponding to each inverter node, the rated power of each corresponding electrical load, multiple electricity consumption characteristics and the characteristic threshold corresponding to each electricity consumption characteristic; and determine the user profile information corresponding to each inverter node based on the daily load curve and user tag corresponding to each inverter node.

[0282] The modules in the aforementioned distributed inverter grouping and scheduling device based on cloud-edge collaboration and user behavior prediction can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0283] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data related to a distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction.

[0284] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 6As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When executed by the processor, the computer program implements a distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0285] Those skilled in the art will understand that Figure 5 or Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0286] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0287] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0288] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0289] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0290] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0291] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0292] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A distributed inverter grouping and scheduling method based on cloud-edge collaboration and user behavior prediction, applied in a cloud server, characterized in that, The method includes: Obtain historical electricity consumption data, geographical coordinates, and power suppliers for multiple inverter nodes; Based on the historical electricity consumption data of the multiple inverter nodes, user profile information corresponding to each inverter node is determined, and based on the geographical coordinates of the multiple inverter nodes and the power supplier, the multiple inverter nodes are divided into at least two scheduling groups. Based on multiple user profile information, weather forecast data for future periods, and electricity price prediction data for the regional power grid corresponding to each scheduling group for future periods, the load power consumption probability prediction range and photovoltaic output probability prediction range for each inverter node in the future period are determined. Based on the load power consumption probability prediction range of each inverter node in the future time period, the photovoltaic power output probability prediction range in the future time period, and the grid load information of the regional power grid corresponding to each scheduling group, the operating mode of each inverter node is determined; the operating mode is used to control the corresponding inverter node to adjust the power to the baseline power corresponding to the operating mode in the future time period.

2. The method according to claim 1, characterized in that, The determination of the operating mode of each inverter node based on the load consumption probability prediction interval of each inverter node in the future time period, the photovoltaic output probability prediction interval in the future time period, and the grid load information of the regional power grid corresponding to each scheduling group includes: Based on the load power consumption probability prediction interval of each inverter node in the future time period, the photovoltaic power output probability prediction interval in the future time period, and the grid load information corresponding to each scheduling group, at least two scheduling groups are divided into at least one first scheduling group and / or at least one second scheduling group. The operating mode of each inverter node in each of the first scheduling groups is determined to be the balanced grid mode; the balanced grid mode is used to control at least one of the multiple inverter nodes in the first scheduling group to perform power adjustment in the same power adjustment direction in the future time period, so as to adjust the power to the baseline power corresponding to each inverter node, and the baseline power of each inverter node corresponds to the energy storage capacity. Based on the load power consumption probability prediction range of each inverter node in the future time period, the photovoltaic power output probability prediction range in the future time period, and the electricity price prediction data of the regional power grid corresponding to each second dispatch group in the future time period, the operating mode of each inverter node in each second dispatch group is determined to be either self-consumption mode or power generation priority mode.

3. The method according to claim 2, characterized in that, The method of determining the operating mode of each inverter node in each second dispatch group as either a self-consumption mode or a generation priority mode based on the load consumption probability prediction range of each inverter node in the future time period, the photovoltaic output probability prediction range of each inverter node in the future time period, and the electricity price prediction data of the corresponding regional power grid of each second dispatch group in the future time period includes: Based on the predicted load consumption probability interval and the predicted photovoltaic output probability interval for each inverter node in the future time period, the net load uncertainty interval for each inverter node is determined; the minimum net load of the net load uncertainty interval is determined by the difference between the minimum value of the predicted load consumption probability interval and the maximum value of the predicted photovoltaic output probability interval, and the maximum net load of the net load uncertainty interval is determined by the difference between the maximum value of the predicted load consumption probability interval and the minimum value of the predicted photovoltaic output probability interval. Based on the electricity price forecast data corresponding to the second scheduling group, the minimum net load and the maximum net load of each inverter node, the operating mode of each inverter node in each of the second scheduling groups is determined to be either self-consumption mode or generation priority mode.

4. The method according to claim 3, characterized in that, The step of determining the operating mode of each inverter node in each of the second dispatch groups as either a self-consumption mode or a generation priority mode based on the electricity price forecast data corresponding to the second dispatch group, the minimum net load of each inverter node, and the maximum net load of each inverter node includes: If the electricity price prediction data corresponding to the second scheduling group indicates that the grid sales price is greater than or equal to the preset sales price threshold, and the maximum net load of the corresponding inverter node is greater than 0, then the operating mode of the corresponding inverter node is determined to be the self-consumption mode; in the self-consumption mode, the baseline power of each inverter node corresponds to the maximum net load. If the electricity price forecast data corresponding to the second scheduling group indicates that the grid purchase price is greater than or equal to the preset purchase price threshold, and the minimum net load of the corresponding inverter node is less than 0, then the operating mode of the corresponding inverter node is determined to be the generation priority mode; in the generation priority mode, the baseline power of each inverter node corresponds to the minimum net load.

5. The method according to claim 3, characterized in that, The determination of the net load uncertainty interval for each inverter node based on the load consumption probability prediction interval and the photovoltaic output probability prediction interval for each inverter node in the future time period includes: Based on the risk preference coefficient of each inverter node, the load power consumption probability prediction interval for the future time period of each inverter node is narrowed to obtain the load power consumption probability prediction interval after narrowing. The maximum value of the narrowed load power consumption probability prediction interval is less than the maximum value of the load power consumption probability prediction interval before narrowing, and the minimum value of the narrowed load power consumption probability prediction interval is greater than the minimum value of the load power consumption probability prediction interval before narrowing. Based on the risk preference coefficient of each inverter node, the photovoltaic output probability prediction interval for the future time period of each inverter node is narrowed to obtain the photovoltaic output probability prediction interval after the future time period is narrowed. The maximum value of the photovoltaic output probability prediction interval after the narrowing is less than the maximum value of the photovoltaic output probability prediction interval before the narrowing, and the minimum value of the photovoltaic output probability prediction interval after the narrowing is greater than the minimum value of the photovoltaic output probability prediction interval before the narrowing. Based on the predicted range of load electricity consumption probability after the contraction of the future time period and the predicted range of photovoltaic power output probability after the contraction of the future time period, the net load uncertainty range of each inverter node is determined.

6. The method according to claim 1, characterized in that, The process of determining user profile information corresponding to each inverter node based on historical electricity consumption data from multiple inverter nodes includes: Based on the historical electricity consumption data of multiple inverter nodes, the daily load curve and multiple electricity consumption characteristics of each inverter node are determined. Based on the daily load curve of each inverter node, multiple electricity consumption characteristics, and the characteristic thresholds corresponding to each electricity consumption characteristic, user profile information corresponding to each inverter node is determined.

7. The method according to claim 6, characterized in that, The process of determining user profile information corresponding to each inverter node based on the daily load curve of each inverter node, multiple electricity consumption characteristics, and feature thresholds corresponding to each electricity consumption characteristic includes: Obtain the energy storage capacity corresponding to each inverter node, and the rated power of each electrical load connected to each inverter node; Based on the energy storage capacity corresponding to each inverter node, the rated power of each corresponding electrical load, multiple electrical consumption characteristics, and the characteristic threshold corresponding to each electrical consumption characteristic, the user tag of each inverter node is determined. Based on the daily load curves and user tags corresponding to each inverter node, user profile information corresponding to each inverter node is determined.

8. The method according to claim 1, characterized in that, The inverter node is a home energy storage inverter node.

9. A cloud server, characterized in that, The cloud server is used to perform the method as described in any one of claims 1-8.

10. A scheduling system, characterized in that, Includes the cloud server as described in claim 9 and multiple inverter nodes; The inverter node is configured to switch from the historical mode to the operating mode when entering a future time period, so as to adjust the power to the baseline power corresponding to the operating mode during the future time period.

Images