Building energy-saving optimization system, method and equipment based on photovoltaic power generation

Abstract

The invention discloses a building energy-saving optimization system, method and equipment based on photovoltaic power generation, and relates to the field of energy-saving system management.The building energy-saving optimization system comprises a data acquisition module, a data uploading module, a data processing module and a data response module. The data flow of each edge node is asynchronously uploaded to the data processing module at the maximum uploading rate, the photovoltaic equipment data is disassembled and segmented, the system splits a data processing task into a plurality of sub-tasks, and the sub-tasks are processed by adopting a multi-channel task scheduling architecture. The load state of the channel is monitored in real time, task allocation is adjusted in time, the energy use efficiency is effectively improved, building energy consumption is reduced, and meanwhile the stability of the system is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of energy-saving system management, and in particular to a building energy-saving optimization system, method and equipment based on photovoltaic power generation. Background Art

[0002] Photovoltaic power generation, as a clean, distributed energy source, is now widely used in areas such as building roofs and facades. Optimizing photovoltaic energy efficiency requires efficient and coordinated processing of large amounts of data, event identification, scheduling decisions, and execution feedback. Traditional systems are plagued by channel congestion, data lags, and uneven resource scheduling, which severely restrict system efficiency and response speed. Therefore, an intelligent scheduling system is urgently needed to comprehensively optimize building energy consumption, maximize photovoltaic efficiency, and dynamically regulate carbon emissions, thereby building a green building energy system.

[0003] Existing technology, such as the invention patent with announcement number: CN117371624B, is a method for optimizing and managing building energy-saving transformation, including: collecting hourly temperature data within one month, recorded as a temperature change sequence Tem; collecting various indicators in the building to obtain the actual total energy consumption; obtaining the theoretical usable area based on the theoretical and actual number of residents and area; obtaining the daily energy consumption per unit area and obtaining the sample households; obtaining the hourly energy consumption sequence Em; obtaining the sorted temperature change sequence Tem1 and the hourly energy consumption sequence Em1; obtaining the high temperature segment and the low temperature segment sequence according to the temperature threshold; obtaining the degree of difference between the low temperature segment and the high temperature segment according to the high temperature segment sequence and the low temperature segment sequence; and obtaining the optimal temperature threshold for distinguishing, and obtaining new high temperature segment and low temperature segment sequences, obtaining the functional relationship between temperature and energy consumption, and obtaining the theoretical energy consumption of the building, obtaining the energy consumption optimization potential of the building, and optimizing management.

[0004] Existing technologies, such as the invention patent with announcement number CN118297359B, are a green energy-saving design optimization method for construction projects, which belongs to the field of energy-saving optimization technology. The method includes step 1: equipping the building area with a photovoltaic energy station according to the planning content of the building area; step 2: collecting relevant parameter information of the building area after the building area is completed; step 3: analyzing the collected parameter information to determine whether the electricity demand in the building area exceeds the estimated electricity demand. When it is determined to exceed the estimated electricity demand, the collected information is analyzed to generate an optimization strategy; step 4: making corresponding optimization adjustments to the building area according to the generated optimization strategy.

[0005] Based on the above technical solutions, it can be seen that existing technologies often focus on building energy consumption modeling and energy-saving strategy formulation, and mostly rely on long-term periodic data collection for static modeling, which is difficult to meet the rapid regulation needs under real-time fluctuations in building loads. In addition, the above technical solutions only construct energy consumption evaluation models based on the functional relationship between temperature and energy consumption, ignoring the dynamic impact of multidimensional factors on building energy efficiency. Therefore, a more comprehensive energy-saving optimization system with dynamic response capabilities is needed. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the present invention provides a building energy-saving optimization system, method and equipment based on photovoltaic power generation. To achieve the above objectives, the present invention is implemented through the following technical solutions: A building energy-saving optimization system, method and equipment based on photovoltaic power generation, comprising:

[0007] The data acquisition module is used to obtain the core data of energy operation in the building. Based on the current load of the building, the data sampling frequency of each node is dynamically adjusted. Each edge node collects local data of each connected photovoltaic device in real time based on the data sampling frequency.

[0008] The data upload module is used to record the collection nodes of the photovoltaic equipment as edge nodes, obtain the data priority of each edge node, set the maximum upload rate of each edge node data, and each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate.

[0009] The data processing module is used to decompose the data received from each edge node into several tasks, analyze and process them through a multi-channel task scheduling architecture, and each channel writes the processing results into a unified result queue for fusion to generate system response instructions.

[0010] The data response module is used to receive the system response instruction from the data processing module and send the system response instruction to each photovoltaic device for execution.

[0011] As a preferred technical solution, local data of each connected photovoltaic device is collected in real time based on the data sampling frequency. The specific process is as follows:

[0012] Obtain core data on energy operation within the building, including photovoltaic power output power, building load power consumption, grid power output power and energy storage capacity.

[0013] The photovoltaic power generation output power is compared with the building load power consumption to obtain the current photovoltaic power generation coverage result. When the photovoltaic power generation output power is greater than the building load power consumption, the photovoltaic power generation coverage result is full coverage. When the photovoltaic power generation output power is equal to the building load power consumption, the photovoltaic power generation coverage result is balanced coverage. When the photovoltaic power generation output power is less than the building load power consumption, the photovoltaic power generation coverage result is no coverage.

[0014] If the photovoltaic power generation coverage result is full coverage, the energy storage device is started to store surplus electricity and it is determined that energy-saving optimization operations are required.

[0015] If the photovoltaic power generation coverage result is not possible, energy storage power will be called in for supplementary power.

[0016] If the photovoltaic power generation coverage result is balanced coverage, it is determined that no energy-saving optimization operation is required.

[0017] When it is determined that energy-saving optimization operations are required, the data sampling frequency of each edge node is dynamically adjusted based on the current load level of the building. Each edge node collects local data of each connected photovoltaic device in real time based on the data sampling frequency.

[0018] As a preferred technical solution, each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate. The specific process is as follows:

[0019] The number of data output destinations, the number of associated components, and the load support capacity value of each photovoltaic device are obtained, and the data priority of each photovoltaic device is obtained by correcting the coupling process.

[0020] The hardware performance parameters of the data processing module are obtained, including CPU occupancy, memory utilization, I / O bandwidth occupancy and average data processing rate. The performance characteristic values of the data processing module are obtained after correcting the coupling processing.

[0021] The total amount of collected data of each photovoltaic device is obtained, the check value of the total amount of collected data is extracted from the database, the data weight corresponding to the data priority of each photovoltaic device is extracted from the database, and the maximum upload rate matching factor is obtained after correction coupling processing with the performance characteristic value of the data processing module. The maximum upload rate of each photovoltaic device data is obtained by mapping and matching with the mapping set corresponding to the maximum upload rate matching factor and the maximum upload rate preset in the database.

[0022] As a preferred technical solution, after receiving data from each edge node, it is broken down into several tasks, including:

[0023] The data access type of each photovoltaic device data is obtained, and the data synchronization frequency of each photovoltaic device data is extracted. The data is combined with the total amount of collected data of each photovoltaic device data, and the coupling analysis is modified to obtain the lock granularity matching factor of each photovoltaic device data. The lock granularity matching factor of each photovoltaic device data is mapped and matched with the mapping set of lock granularity matching factors and lock granularities pre-stored in the database to obtain the lock granularity of each photovoltaic device data. The data of each photovoltaic device is then segmented and disassembled based on the lock granularity to obtain several tasks.

[0024] As the preferred technical solution, the analysis and processing is performed through a multi-channel task scheduling architecture, specifically including:

[0025] Determine the identifier of each task, including the task's event ID, source device ID, timestamp combination, and payload type. The identifier is used to uniquely identify each task.

[0026] Based on a preset hash function, a hash function is performed on the task identifier to generate a hash value.

[0027] Get the number of channels, perform a modulo operation on the hash value based on the number of channels, and map the task to the corresponding target channel.

[0028] Add each task to its corresponding target channel task queue for data processing.

[0029] As the preferred technical solution, the analysis and processing is performed through a multi-channel task scheduling architecture, specifically including:

[0030] Based on the monitoring agents deployed on each channel, load monitoring indicators are collected, including QPS, CPU utilization, memory occupancy, and queue length. After correcting the coupling processing, the load monitoring characteristic values of each channel are obtained.

[0031] Based on the load monitoring characteristic value of each channel, it is compared with the load monitoring characteristic threshold to obtain the load judgment result of each channel. If the load monitoring characteristic value of a channel is greater than or equal to the load monitoring characteristic threshold, the load judgment result provided is judged to be overloaded. If the load monitoring characteristic value of a channel is less than the load monitoring characteristic threshold, the load judgment result provided is judged to be not overloaded.

[0032] The channels whose load statistics result shows that they are overloaded are recorded as overload channels, and an overload alarm signal is sent to the channel scheduling controller.

[0033] After receiving the overload alarm signal, the channel scheduling controller triggers the task migration operation, which specifically includes: subtracting the load monitoring characteristic value of each overload channel from the load monitoring characteristic threshold to obtain the load monitoring characteristic difference of each overload channel; based on the load monitoring characteristic difference of each overload channel, mapping and matching the mapping set corresponding to the load monitoring characteristic difference and the task migration ratio pre-stored in the database to obtain the task migration ratio of each overload channel.

[0034] Based on the task migration ratio of each overloaded channel, the demand migration tasks of each overloaded channel are obtained. After introducing the disturbance factor, the secondary hash value of each demand migration task is calculated again. According to the secondary hash value, each demand migration task is assigned to a new channel for analysis and processing.

[0035] As a preferred technical solution, triggering the task migration operation also includes setting a migration cool-off period, specifically including:

[0036] If the number of demand migration tasks for a channel is greater than 1, it is determined that multiple task migration operations are required. Before triggering multiple migration operations, the load monitoring characteristic value of the channel is calculated again and compared with the demand migration threshold stored in the database to obtain the demand migration cool-down result.

[0037] If the load monitoring characteristic value of the channel is less than or equal to the demand migration threshold, the demand migration cooling-off result of the channel is a no-demand migration cooling-off period.

[0038] If the load monitoring characteristic value of the channel is greater than the demand migration threshold, the demand migration cooling-off result of the channel is the demand migration cooling-off period.

[0039] The statistical demand migration cooling-off result is the channel of the demand migration cooling-off period, which is recorded as each demand migration cooling-off period channel. The overall occupancy of each demand migration cooling-off period channel is extracted, and a difference analysis is performed with the demand migration threshold to obtain the migration time window matching factor of each demand migration cooling-off period channel.

[0040] The migration time window matching factor of each demand migration cooling-off period channel is mapped and matched with a mapping set of migration time window matching factors and migration time windows pre-stored in the database to obtain the migration time window of each demand migration cooling-off period channel.

[0041] When a migration cooling-off period channel for a certain requirement continuously triggers migration events, task migration is performed based on the migration time window.

[0042] As a preferred technical solution, each channel writes the processing results into a unified result queue, which is then integrated to generate system response instructions, specifically including:

[0043] After the channel completes task processing, it generates structured processing result data and encapsulates the processing result data into a message object in a standard format, including the task number, channel number, processing result type, and result generation time.

[0044] The encapsulated processing result data is asynchronously written into the unified result queue through the message middleware. The obtained results are classified and stored according to the task number, and fused. The system response instruction is generated according to the fusion result. The content of the system response instruction includes: control type, control object and control parameters.

[0045] In addition, the method of building energy-saving optimization system based on photovoltaic power generation includes:

[0046] Obtain the core data of energy operation in the building, dynamically adjust the data sampling frequency of each node based on the current load of the building, and each edge node collects local data of each connected photovoltaic device in real time based on the data sampling frequency.

[0047] The collection nodes of photovoltaic equipment are recorded as edge nodes, the data priority of each edge node is obtained, and the maximum upload rate of each edge node data is set. Each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate.

[0048] After receiving data from each edge node, it is broken down into several tasks and analyzed and processed through a multi-channel task scheduling architecture. Each channel writes the processing results into a unified result queue, which is then integrated to generate system response instructions.

[0049] Receive the system response instructions from the data processing module and send the system response instructions to each photovoltaic device for execution.

[0050] The device of the building energy-saving optimization system based on photovoltaic power generation has one or more programs, and the one or more programs are executed by one or more processors to realize the above system.

[0051] Compared with the prior art, the embodiments of the present invention have at least the following beneficial effects:

[0052] (1) The present invention provides a building energy-saving optimization system based on photovoltaic power generation, which dynamically adjusts the data sampling frequency according to the current load status of the building. This ensures that the system operates efficiently under different workloads, avoids unnecessary data redundancy and bandwidth waste, and can also monitor and optimize energy use in real time. Based on the comparison of photovoltaic power generation output power and building load power consumption, the system can perform corresponding operations based on different photovoltaic power generation coverage results, thereby improving energy utilization efficiency.

[0053] (2) The present invention asynchronously uploads the data stream of each edge node to the data processing module at the maximum upload rate. This approach avoids the performance bottleneck that may be caused by traditional synchronous upload and improves the data transmission efficiency of the system. By obtaining relevant data of photovoltaic equipment, the system can allocate appropriate upload rates to different devices. This not only improves the stability and efficiency of data upload, but also ensures the system's processing priority for key devices.

[0054] (3) The present invention disassembles and segments photovoltaic equipment data, splitting the data processing task into multiple subtasks and processing them using a multi-channel task scheduling architecture. This parallel processing approach significantly improves the system's processing power and response speed, adapting to large-scale data processing needs. Tasks are mapped to different processing channels using a hash function, achieving uniform distribution of tasks, avoiding overloading of any one channel, and improving the system's load balancing capabilities.

[0055] (4) The present invention monitors the load status of each channel in real time based on the load monitoring indicators of each channel. If a channel is overloaded, the system can trigger a task migration operation and adjust the task allocation in time to avoid system performance bottlenecks. During the task migration process, the system introduces a cool-down period mechanism to avoid the additional burden on the system caused by frequent migration operations. By calculating the load monitoring feature difference and the task migration ratio, the task migration strategy is adjusted more intelligently, further improving the efficiency and stability of task scheduling.

[0056] Of course, any product implementing the present invention does not necessarily need to achieve all of the above advantages at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 Schematic diagram of the system module of the present invention.

[0058] Figure 2 Schematic diagram of the method of the present invention.

[0059] Figure 3 The figure is a schematic diagram of the core data collection and analysis process involved in an embodiment of the present invention.

[0060] Figure 4 The figure is a schematic diagram of the edge node data collection and upload process involved in an embodiment of the present invention.

[0061] Figure 5 The figure is a flowchart of task decomposition and multi-channel scheduling involved in an embodiment of the present invention.

[0062] Figure 6 The figure is a schematic diagram of the load monitoring and task migration process involved in an embodiment of the present invention.

[0063] Figure 7 This is a flowchart of the result fusion and response instruction issuance process involved in an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0065] In the description of the present invention, it should be understood that the terms "opening", "upper", "lower", "thickness", "top", "middle", "length", "inside", "around" and the like indicating orientation or positional relationship are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0066] See also Figure 1 As shown, the embodiment of the present invention provides a building energy-saving optimization system based on photovoltaic power generation, specifically including:

[0067] See also Figure 3 The figure shows a schematic diagram of the core data collection and analysis process involved in an embodiment of the present invention. The process mainly includes obtaining building energy data, analyzing photovoltaic power generation and load power consumption, and deciding whether to activate energy storage or supplement.

[0068] The data acquisition module is used to obtain the core data of energy operation in the building. Based on the current load of the building, the data sampling frequency of each node is dynamically adjusted. Each edge node collects local data of each connected photovoltaic device in real time based on the data sampling frequency.

[0069] Based on the data sampling frequency, local data of each connected photovoltaic device is collected in real time. The specific process is as follows:

[0070] Obtain core data on energy operation within the building, including photovoltaic power output power, building load power consumption, grid power output power and energy storage capacity.

[0071] The photovoltaic power generation output power is compared with the building load power consumption to obtain the current photovoltaic power generation coverage result. When the photovoltaic power generation output power is greater than the building load power consumption, the photovoltaic power generation coverage result is full coverage. When the photovoltaic power generation output power is equal to the building load power consumption, the photovoltaic power generation coverage result is balanced coverage. When the photovoltaic power generation output power is less than the building load power consumption, the photovoltaic power generation coverage result is no coverage.

[0072] If the photovoltaic power generation coverage result is full coverage, the energy storage device is started to store surplus electricity and it is determined that energy-saving optimization operations are required.

[0073] If the photovoltaic power generation coverage result is not available, the energy storage power will be called upon to supplement. If the energy storage power is insufficient to supplement, it will be connected to the power grid.

[0074] If the photovoltaic power generation coverage result is balanced coverage, it is determined that no energy-saving optimization operation is required.

[0075] When energy-saving optimization operations are determined to be necessary, the data sampling frequency of each edge node is dynamically adjusted based on the current load level of the building. Each edge node collects local data from each connected photovoltaic device in real time based on the data sampling frequency, specifically including:

[0076] The building load power is extracted, and the building load power is mapped and matched with the mapping set corresponding to the building load power and data sampling frequency pre-stored in the database to obtain the data sampling frequency of each edge node.

[0077] See also Figure 4 FIG2 is a schematic diagram of the edge node data collection and upload process involved in an embodiment of the present invention. The process mainly includes dynamically adjusting the data sampling frequency according to the load situation and asynchronously uploading the collected data to the processing module.

[0078] The data upload module is used to record the collection nodes of the photovoltaic equipment as edge nodes, obtain the data priority of each edge node, set the maximum upload rate of each edge node data, and each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate.

[0079] It should be noted that there is a one-to-one correspondence between photovoltaic devices and edge nodes.

[0080] Each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate. The specific process is as follows:

[0081] The number of data output destinations, the number of associated components, and the load support capacity value of each photovoltaic device are obtained, and the data priority of each photovoltaic device is obtained by correcting the coupling process.

[0082] The load support capacity value of each photovoltaic device is calculated as follows:

[0083] The average output power and daily effective power generation time of each photovoltaic device are obtained, and the average daily power generation of each photovoltaic device is calculated and processed. The specific calculation formula is: E PV =P PV ×T PV , where E PV is the average daily power generation of the PV-th photovoltaic device, P PV is the average output power of the PV-th photovoltaic device, T PV is the average daily effective power generation time of the PV-th photovoltaic device, PV is the photovoltaic device number, PV = 1, 2, 3, ..., n, and n is the total number of photovoltaic devices.

[0084] The average daily power generation of each photovoltaic device is mapped and matched with the load support capacity value corresponding to the average daily power generation pre-stored in the database to obtain the load support capacity value of each photovoltaic device. It should be noted that the load support capacity value of each photovoltaic device is used to evaluate the power generation capacity of each photovoltaic device.

[0085] The data priority calculation process for each photovoltaic device includes:

[0086]

[0087] Among them, Lev PV The data priority of the PV photovoltaic device, goal PV The number of destinations for data output of the PV-th photovoltaic device, rel PV is the number of associated components of the PV-th photovoltaic device, hol PV is the load support capability value of the PV-th photovoltaic device, ε is the quantity unit elimination factor, γ1 is the weighted constraint value of the number of data output destinations, γ2 is the weighted constraint value of the number of associated components, γ3 is the weighted constraint value of the load support capability value, PV is the photovoltaic device number, PV=1,2,3,...,n, n is the total number of photovoltaic devices, For any real number x, the rounding function returns the largest integer less than or equal to x.

[0088] It should be noted that the weighted constraint value of the number of data output destinations, the weighted constraint value of the number of associated components, and the weighted constraint value of the load supporting capacity value all have value ranges between 0 and 1, and satisfy γ1+γ2+γ3=1. The weighted constraint value of the number of data output destinations is an influencing factor of the number of data output destinations pre-stored in the database, indicating the degree to which the number of data output destinations affects the data priority of each photovoltaic device; the weighted constraint value of the number of associated components is an influencing factor of the number of associated components pre-stored in the database, indicating the degree to which the number of associated components affects the data priority of each photovoltaic device; the weighted constraint value of the load supporting capacity value is an influencing factor of the load supporting capacity value pre-stored in the database, indicating the degree to which the load supporting capacity value affects the data priority of each photovoltaic device. When used, they are directly extracted from the database. For example, after the number of data output destinations, the number of associated components, and the load supporting capacity value of each photovoltaic device are input into a preset mapping set, the weighted constraint value of the number of data output destinations, the weighted constraint value of the number of associated components, and the weighted constraint value of the load supporting capacity value are obtained, and the corresponding mapping relationships are one-to-one.

[0089] It should also be noted that there is a correlation between the number of data output destinations, the number of associated components, and the load support capacity. The number of data output destinations refers to the number of target devices to which a photovoltaic device's collected data needs to be sent. This number generally determines the data flow and data distribution complexity between devices. The number of data output destinations directly affects data transmission and processing capabilities. If a device has more output destinations, the complexity and load of its data transmission will be higher, requiring higher priority and more computing resources to process this data. The number of associated components refers to the number of other devices or components associated with the photovoltaic device. For example, a photovoltaic system may integrate and coordinate with multiple sensors, controllers, energy storage systems, inverters, and other devices. The greater the number of associated components, the more complex the control and coordination tasks of the photovoltaic system. For example, if a photovoltaic device controls multiple battery packs and energy storage units simultaneously, the importance of their data in the system becomes relatively high, as this data will affect the overall operation of the system. Therefore, devices with a large number of associated components may be given higher data priority. The load support capacity value refers to the maximum load or output power that the photovoltaic device can support. This parameter reflects the photovoltaic device's power output capacity over a given period of time and its ability to support the building's energy needs. A higher load-supporting capacity value for a device indicates a more important role in building energy management, especially under conditions of high load or significant load fluctuations. If a PV device has a strong load-supporting capacity, its data priority may be increased due to its greater impact on building load management and energy-saving optimization. The greater the number of data output destinations a device has, the greater the data transmission burden within the system, potentially requiring more bandwidth and computing resources. Furthermore, a greater number of associated components requires more complex scheduling and management for the PV device. A device with a strong load-supporting capacity can handle a greater amount of data traffic and thus prioritize more tasks. The number of data output destinations, the number of associated components, and the load-supporting capacity jointly determine the device's responsiveness. When any of these parameters increases, the device's load also increases. Therefore, its data priority should be appropriately increased to ensure timely response to building energy management needs. In particular, devices with high load-supporting capacity are more responsive and can handle more complex tasks, thus receiving a higher data priority. When correcting the coupling relationship between these three factors, the goal is to dynamically adjust the data upload frequency of each PV device based on its comprehensive capabilities. For example, a device with a stronger load-bearing capacity and more associated components can support a higher frequency of data upload, while a device with many data output destinations requires more time to process and transmit data.

[0090] The hardware performance parameters of the data processing module are obtained, including CPU usage, memory usage, I / O bandwidth usage, and average data processing rate. These parameters are compared with the hardware performance standard parameters to obtain the deviation values of the hardware performance parameters. After the correction coupling process, the performance characteristic values of the data processing module are obtained, including:

[0091]

[0092] Among them, XN is the performance characteristic value of the data processing module, U CPU is the CPU usage of the data processing module, U RAM is the memory usage of the data processing module, B IO is the I / O bandwidth occupancy rate of the data processing module, E rate is the average data processing rate of the data processing module, U CPU0 is the standard value of CPU usage, U RAM0 is the standard value of memory usage, B IO0 is the standard value of I / O bandwidth usage, E rate0 is the standard value of the average data processing rate, α1 is the weighted limiting factor of CPU occupancy, α2 is the weighted limiting factor of memory occupancy, α3 is the weighted limiting factor of I / O bandwidth occupancy, and α4 is the weighted limiting factor of the average data processing rate.

[0093] It should be noted that the CPU occupancy weighted limiting factor, memory occupancy weighted limiting factor, I / O bandwidth occupancy weighted limiting factor and data average processing rate weighted limiting factor all have value ranges between 0 and 1 and satisfy α1+α2+α3+α4=1. The CPU occupancy weighted limiting factor is an influencing factor of the CPU occupancy pre-stored in the database, indicating the degree to which the CPU occupancy affects the performance characteristic value of the data processing module; the memory occupancy weighted limiting factor is an influencing factor of the memory occupancy pre-stored in the database, indicating the degree to which the memory occupancy affects the performance characteristic value of the data processing module; the I / O bandwidth occupancy weighted limiting factor is an influencing factor of the I / O bandwidth occupancy pre-stored in the database. The influence factor indicates the degree to which the I / O bandwidth occupancy rate affects the performance characteristic value of the data processing module; the weighted limiting factor of the average data processing rate is an influence factor of the average data processing rate pre-stored in the database, indicating the degree to which the average data processing rate affects the performance characteristic value of the data processing module. When used, it is directly extracted from the database. For example, after the CPU occupancy rate, memory occupancy rate, I / O bandwidth occupancy rate and average data processing rate of the data processing module are input into a preset mapping set in the database, the CPU occupancy weighted limiting factor, memory occupancy weighted limiting factor, I / O bandwidth occupancy weighted limiting factor and average data processing rate weighted limiting factor are obtained, and the corresponding mapping relationships are one-to-one.

[0094] It's also important to note that there's a correlation between CPU utilization, memory utilization, I / O bandwidth utilization, and average data processing rate. CPU utilization refers to the workload on the CPU when processing tasks, representing the percentage of CPU processing capacity consumed. A high CPU utilization typically indicates that the data processing module is performing complex computational tasks, potentially involving large amounts of data processing, calculations, or running complex algorithms. Memory utilization indicates the percentage of memory occupied and is typically related to the size of the task being processed and the number of concurrent tasks. When memory utilization is high, the data processing module may need to access memory more frequently, impacting data access and computation speed. CPU and memory utilization are often interrelated. Certain computationally intensive tasks, especially those processing large amounts of data, also consume more memory because data must be loaded into memory for computation. High CPU utilization can slow processing, especially when memory consumption is excessive, as the CPU may require more time to handle data exchange and computation tasks. When memory utilization exceeds the available memory in the data processing module, the CPU burden increases, and the module begins frequently swapping memory, significantly reducing overall performance and impacting processing speed. I / O bandwidth utilization refers to the proportion of the I / O bandwidth used by the data processing module for data transmission. It generally reflects the load on the data processing module when performing external data exchange (for example, read and write operations with external devices such as hard disks and networks). A high I / O bandwidth utilization may indicate that the data processing module is performing a large number of data input and output operations (such as reading or writing data from disk), which may require CPU cooperation for calculations and data processing. Therefore, when I / O bandwidth demand increases, CPU utilization may also increase, especially when large amounts of data are being processed or transmitted. If I / O bandwidth is insufficient, the data processing module may face an I / O bottleneck, forcing the data processing module to wait for I / O operations to complete, thereby affecting CPU utilization. Even if the CPU has sufficient computing power, if the I / O speed is too slow, the CPU will be in a waiting state and unable to perform calculations efficiently. The average data processing rate refers to the amount of data processed by the data processing module in a certain period of time, usually expressed in bytes or requests processed per second. It is a key indicator of the overall processing capacity and response speed of the data processing module. The data processing rate is directly affected by CPU and memory utilization. If CPU and memory resources are limited, processing speed will decrease, as these resources are fundamental to data processing. Excessive computing tasks or insufficient memory can slow data processing, reducing the data processing rate. Furthermore, I / O bandwidth usage can also affect data processing speed. If I / O bandwidth is overloaded, data read and write speeds are limited, impacting data inflow and outflow, and thus the overall data processing rate. Therefore, I / O bandwidth bottlenecks can become limiting factors in data processing.When a data processing module performs a large number of I / O operations, data is often first loaded into memory before being processed or output. Therefore, memory usage is often closely related to the I / O bandwidth requirements. If I / O operations require frequent reading of data from disk into memory, memory usage will increase, affecting the overall performance of the data processing module. High I / O bandwidth usage usually means that the data processing module has a large amount of external data to process, and this data usually requires a large amount of memory for caching or transfer. Therefore, I / O bandwidth usage and memory usage are often linked, especially when frequent memory reading and writing are required during data processing.

[0095] The total amount of collected data of each photovoltaic device is obtained, a check value of the total amount of collected data is extracted from the database, and a data weight corresponding to the data priority of each photovoltaic device is extracted from the database. The data weight is corrected and coupled with the performance characteristic value of the data processing module to obtain the maximum upload rate matching factor. The maximum upload rate of each photovoltaic device data is obtained by mapping and matching the mapping set corresponding to the maximum upload rate matching factor and the maximum upload rate preset in the database. Specifically, the maximum upload rate of each photovoltaic device data is obtained.

[0096]

[0097] Among them, δ PV is the maximum upload rate matching factor of the PV-th photovoltaic device data, N PV is the total amount of data collected for the PV-th photovoltaic device, N0 is the check value of the total amount of collected data, Lev PV is the data priority of the PV-th photovoltaic device, XN is the performance characteristic value of the data processing module, τ is the data weight corresponding to the data priority of the photovoltaic device, β1 is the weighting coefficient of the total amount of collected data, β2 is the weighting coefficient of the data priority, β3 is the weighting coefficient of the performance characteristic value, PV is the photovoltaic device number, PV = 1, 2, 3, ..., n, where n is the total number of photovoltaic devices.

[0098] It should be noted that the weighting coefficient of the total amount of collected data, the weighting coefficient of the data priority and the weighting coefficient of the performance characteristic value all have a value range between 0 and 1, and satisfy β1+β2+β3=1. The weighting coefficient of the total amount of collected data is an influencing factor of the total amount of collected data pre-stored in the database, which indicates the degree of influence of the total amount of collected data on the maximum upload rate matching factor of each photovoltaic device; the weighting coefficient of the data priority is an influencing factor of the data priority pre-stored in the database, which indicates the degree of influence of the data priority on the maximum upload rate matching factor of each photovoltaic device; the weighting coefficient of the performance characteristic value is an influencing factor of the performance characteristic value pre-stored in the database, which indicates the degree of influence of the performance characteristic value on the maximum upload rate matching factor of each photovoltaic device. When used, they are directly extracted from the database. For example, the total amount of collected data, data priority and performance characteristic value of each photovoltaic device are input into the preset mapping set to obtain the weighting coefficient of the total amount of collected data, the data priority weighting coefficient and the performance characteristic value weighting coefficient, and the corresponding mapping relationship is one-to-one.

[0099] See also Figure 5 The figure shows a schematic diagram of the task decomposition and multi-channel scheduling process involved in an embodiment of the present invention. The process mainly includes processing data received from edge nodes, decomposing it into tasks, and allocating the tasks to different processing channels based on hash values.

[0100] The data processing module is used to decompose the data received from each edge node into several tasks, analyze and process them through a multi-channel task scheduling architecture, and each channel writes the processing results into a unified result queue for fusion to generate system response instructions.

[0101] After receiving the data from each photovoltaic device, it is broken down into several tasks, including:

[0102] The data access type of each photovoltaic device data is obtained, including read access, write access, and read-write access. Based on the data access type of each photovoltaic device data, a mapping is performed with a pre-stored database mapping of data access type and data access type-lock granularity matching value to obtain the data access type-lock granularity matching value for each photovoltaic device data. A reasonable data lock range number is selected based on different access behaviors. In an embodiment of the present invention, the relationship between data access type and lock granularity matching value is such that the lock granularity matching value corresponding to read-write access is greater than the lock granularity matching value corresponding to write access, and the lock granularity matching value corresponding to write access is greater than the lock granularity matching value corresponding to read access, with the data access type increasing in the order of read access, write access, and read-write access.

[0103] At the same time, the data synchronization frequency of each photovoltaic device data is extracted and combined with the total amount of collected data of each photovoltaic device data, and the lock granularity matching factor of each photovoltaic device data is obtained by correcting the coupling analysis. Specifically, it includes:

[0104]

[0105] in, is the lock granularity matching factor of the PV-th photovoltaic device data, y PV The data access type-lock granularity matching value for the PV-th photovoltaic device data, f PV is the data synchronization frequency of the PV-th photovoltaic device data, N PV is the total amount of collected data for the PV-th photovoltaic device, τ f is the data synchronization frequency-lock granularity matching weight, τ N is the total amount of collected data-lock granularity matching weight, ∈1 is the data access type coupling coefficient, ∈2 is the data synchronization frequency coupling coefficient, ∈3 is the total amount of collected data coupling coefficient, PV is the photovoltaic device number, PV = 1, 2, 3, ..., n, where n is the total number of photovoltaic devices.

[0106] It should be noted that the data access type coupling coefficient, the data synchronization frequency coupling coefficient and the total amount of collected data coupling coefficient all have value ranges between 0 and 1 and satisfy ∈1+∈2+∈3=1. The data access type coupling coefficient is an influencing factor of the data access type pre-stored in the database, indicating the degree to which the data access type affects the lock granularity matching factor of each photovoltaic device data; the data synchronization frequency coupling coefficient is an influencing factor of the data synchronization frequency pre-stored in the database, indicating the degree to which the data synchronization frequency affects the lock granularity matching factor of each photovoltaic device data; the total amount of collected data coupling coefficient is an influencing factor of the total amount of collected data pre-stored in the database, indicating the degree to which the total amount of collected data affects the lock granularity matching factor of each photovoltaic device data. When used, it is directly extracted from the database. For example, the data access type, data synchronization frequency and total amount of collected data of each photovoltaic device are input into a preset mapping set to obtain the data access type coupling coefficient, data synchronization frequency coupling coefficient and total amount of collected data coupling coefficient, and the corresponding mapping relationship is one-to-one.

[0107] The lock granularity matching factor of each photovoltaic device data is mapped and matched with the mapping set of lock granularity matching factors and lock granularity pre-stored in the database to obtain the lock granularity of each photovoltaic device data, and the photovoltaic device data is split and disassembled based on the lock granularity to obtain several tasks.

[0108] In the embodiment of the present invention, the lock granularity is divided into the following levels, specifically including:

[0109] Full data lock, the largest granularity, means that the entire data set needs to be locked when accessing that data.

[0110] Field-level locking, locking only certain fields or subsets.

[0111] Row-level locking locks a row in a data table.

[0112] Based on the lock granularity of each photovoltaic device data, the photovoltaic device data is disassembled. The specific disassembly process includes:

[0113] If the lock granularity is full data lock, each access requires the complete lock of the device data. In this case, the disassembly process is relatively simple, and the data is not split into multiple tasks.

[0114] If the lock granularity is field-level, each task will be split based on the data of different fields. For example, a photovoltaic device may include multiple parameters (such as power, temperature, voltage, etc.), and the data of each field may be split into independent tasks for processing.

[0115] If the lock granularity is row-level, the data is broken down into independent tasks for parallel processing. Each row represents a point in time or a measurement result.

[0116] Analyze and process through a multi-channel task scheduling architecture, including:

[0117] Determine the identifier of each task, including the task's event ID, source device ID, timestamp combination, and payload type. The identifier is used to uniquely identify each task.

[0118] Based on a preset hash function, a hash function is performed on the task identifier to generate a hash value.

[0119] Get the number of channels, perform a modulo operation on the hash value based on the number of channels, and map the task to the corresponding target channel, including:

[0120] A hash operation is performed on the task identifier string based on a preset hash function to generate a hash value. In the embodiment of the present invention, the hash function adopts MD5 to ensure that the generated hash value has good randomness and distribution.

[0121] Based on the number of processing channels currently configured, the hash value is modulo-calculated to obtain a mapping channel index. This index is used to assign tasks to the corresponding target processing channel, achieving even distribution of tasks. Specifically, if the number of channels is M, the task will be assigned to the channel numbered "hash value mod M".

[0122] There are multiple parallel processing channels in the system for concurrently executing different tasks. According to the current load capacity or static configuration of the system, the total number of currently available channels is set to M.

[0123] A hash function is performed on the task identifier to generate an integer hash value. The hash value should have good distribution and uniqueness to avoid tasks being mapped to certain channels.

[0124] Performing modulo calculation on the above hash value, the calculation code is:

[0125] channel_index=hash_code mod M

[0126] The hash value modulo the number of channels is used to obtain an integer index value in the range [0, M-1].

[0127] Assign the task to the channel numbered channel_index. This channel is the target processing channel for the task. In this way, each task is evenly mapped to a certain channel number in a computable way.

[0128] Add each task to its corresponding target channel task queue for data processing.

[0129] See also Figure 6 FIG2 is a flow chart of load monitoring and task migration according to an embodiment of the present invention. The flow mainly includes real-time monitoring of the load of each channel and deciding whether to perform task migration or conduct a cool-down period according to the load situation.

[0130] Analyze and process through a multi-channel task scheduling architecture, including:

[0131] Based on the monitoring agents deployed on each channel, load monitoring indicators are collected, including QPS, CPU utilization, memory utilization, and queue length. The load monitoring indicator threshold set is extracted from the database, including QPS threshold, CPU utilization threshold, memory utilization threshold, and queue length threshold.

[0132] It's important to note that QPS (Queries Per Second) represents the number of tasks or requests processed by a channel per unit of time. It measures the channel's request processing throughput. A higher QPS indicates a stronger channel's processing capability, but it can also indicate a heavier channel load. If the QPS exceeds the channel's processing capacity, it can lead to performance bottlenecks or overload.

[0133] CPU utilization refers to the degree to which a channel's CPU resources are being used, typically expressed as a percentage. It reflects the channel's current computing load. High CPU utilization (close to 100%) means the CPU's computing resources are almost fully utilized, potentially leading to degraded channel performance, delayed responses, or even channel crashes. Low CPU utilization indicates that channel resources are idle and may be wasting resources.

[0134] Memory utilization refers to channel memory usage, typically expressed as a percentage. It measures the channel's use of memory resources allocated to tasks or processes. High memory utilization may indicate that the channel is processing a large amount of data or tasks. If memory utilization approaches 100%, the channel may be at risk of running out of memory, resulting in degraded performance or even a channel crash. Low memory utilization indicates that channel resources are idle and potentially wasting resources.

[0135] Queue length refers to the number of pending tasks in a channel, typically the number of tasks waiting in the task queue. Queue length is an important indicator of task processing backlogs. Excessive queue lengths indicate insufficient processing capacity and a significant backlog of tasks, potentially leading to processing delays or failures. Short queue lengths indicate that the system is able to process tasks promptly.

[0136] The load monitoring indicator of each channel is compared with the load monitoring indicator threshold set, and then the load monitoring characteristic value of each channel is obtained after correction and coupling processing, which includes:

[0137]

[0138] Among them, FU i is the load monitoring characteristic value of the i-th channel, QPS i is the QPS of the i-th channel, CPU i is the CPU utilization of the i-th channel, RAM i is the memory utilization of the i-th channel, L i is the queue length of the i-th channel, QPS0 is the QPS threshold, CPU0 is the CPU utilization threshold, RAM0 is the memory utilization threshold, L0 is the queue length threshold, d1 is the QPS coupling factor, d2 is the CPU utilization coupling factor, d3 is the memory utilization coupling factor, d4 is the queue length coupling factor, i is the channel number, i = 1, 2, 3, ..., D, and D is the total number of channels.

[0139] It should be noted that the QPS coupling factor, CPU utilization coupling factor, memory utilization coupling factor and queue length coupling factor all have a value range between 0 and 1 and satisfy d1+d2+d3+d4=1. The QPS coupling factor is the QPS influencing factor pre-stored in the database, which indicates the degree of influence of QPS on the load monitoring characteristic value of each channel; the CPU occupancy coupling factor is the CPU occupancy influencing factor pre-stored in the database, which indicates the degree of influence of CPU occupancy on the load monitoring characteristic value of each channel; the memory utilization coupling factor is the QPS influencing factor pre-stored in the database, which indicates the degree of influence of CPU occupancy on the load monitoring characteristic value of each channel; The memory utilization impact factor indicates the degree to which memory utilization affects the load monitoring characteristic value of each channel. The queue length coupling factor is a queue length impact factor pre-stored in the database, indicating the degree to which the queue length affects the load monitoring characteristic value of each channel. When used, it is directly extracted from the database. For example, after entering the QPS, CPU utilization, memory utilization, and queue length of each channel into the preset mapping set in the database, the QPS coupling factor, CPU utilization coupling factor, memory utilization coupling factor, and queue length coupling factor are obtained. The corresponding mapping relationships are one-to-one.

[0140] It's also important to note that QPS, CPU utilization, memory utilization, and queue length are correlated. QPS is a key indicator of the density of tasks a channel is currently receiving and processing. A higher QPS indicates a greater number of tasks being processed per unit time, resulting in a greater overall load. High QPS typically leads to increased CPU utilization, as the system requires more computing resources to process tasks. CPU utilization reflects the channel's utilization of computing resources. When QPS is high, CPU utilization will also increase if the task is compute-intensive. Conversely, high CPU utilization with low QPS may indicate complex tasks, low processing efficiency, or a computing bottleneck in the system. Memory utilization indicates the channel's use of memory resources while executing tasks. Large tasks, data-intensive tasks, or the need to cache large amounts of intermediate data will lead to increased memory utilization. High QPS also increases memory usage, but the rate of increase depends on the data characteristics of the tasks. Queue length reflects the number of tasks awaiting processing in the channel and is often a direct reflection of a mismatch between QPS and actual processing capacity. When QPS continues to increase while the CPU and memory are nearing full capacity, the task processing rate decreases, leading to task accumulation and increasing queue length. A continuously growing queue length indicates that the channel is gradually approaching overload.

[0141] Based on the load monitoring characteristic value of each channel, it is compared with the load monitoring characteristic threshold to obtain the load judgment result of each channel. If the load monitoring characteristic value of a channel is greater than or equal to the load monitoring characteristic threshold, the load judgment result provided is judged to be overloaded. If the load monitoring characteristic value of a channel is less than the load monitoring characteristic threshold, the load judgment result provided is judged to be not overloaded.

[0142] The channels whose load statistics result shows that they are overloaded are recorded as overload channels, and an overload alarm signal is sent to the channel scheduling controller.

[0143] After receiving the overload alarm signal, the channel scheduling controller triggers the task migration operation, which specifically includes: subtracting the load monitoring characteristic value of each overload channel from the load monitoring characteristic threshold to obtain the load monitoring characteristic difference of each overload channel; based on the load monitoring characteristic difference of each overload channel, mapping and matching the mapping set corresponding to the load monitoring characteristic difference and the task migration ratio pre-stored in the database to obtain the task migration ratio of each overload channel.

[0144] Based on the task migration ratio of each overloaded channel, the demand migration tasks of each overloaded channel are obtained. After introducing the disturbance factor, the secondary hash value of each demand migration task is calculated again. According to the secondary hash value, each demand migration task is assigned to a new channel for analysis and processing.

[0145] Get the task feature string of each demand migration task of each overload channel. The specific code is:

[0146]

[0147] After introducing the perturbation factor, the corresponding perturbation factor string is generated. The specific code is:

[0148]

[0149] Combine the task feature string of each demand migration task of each overload channel with the disturbance factor to obtain the secondary hash input string. The specific code is: A hash algorithm is executed on the secondary hash input string. In this embodiment, MD5 is used. The specific code is: Secondary_Hash_Value = MD5(Secondary_Hash_Input). The number of available new channels is set to NM. The secondary hash input value is converted to an integer and then modulo the channel number. The specific code is: New_Channel_ID = Secondary_Hash_Value mod Channel_Num.

[0150] The hash value modulo the number of channels is used to get an integer index value in the range [0, NM-1].

[0151] Assign a task to the channel numbered New_Channel_ID. This channel is the target processing channel for the task. In this way, each task is evenly mapped to a specific channel number in a calculable way.

[0152] Triggering a task migration also includes setting a migration cool-off period, which includes:

[0153] If the number of demand migration tasks for a channel is greater than 1, it is determined that multiple task migration operations are required. Before triggering multiple migration operations, the load monitoring characteristic value of the channel is calculated again and compared with the demand migration threshold stored in the database to obtain the demand migration cool-down result.

[0154] If the load monitoring characteristic value of the channel is less than or equal to the demand migration threshold, the demand migration cooling-off result of the channel is a no-demand migration cooling-off period.

[0155] If the load monitoring characteristic value of the channel is greater than the demand migration threshold, the demand migration cooling-off result of the channel is the demand migration cooling-off period.

[0156] The statistical demand migration cooling-off result is the channel of the demand migration cooling-off period, which is recorded as each demand migration cooling-off period channel. The overall occupancy of each demand migration cooling-off period channel is extracted, and a difference analysis is performed with the demand migration threshold to obtain the migration time window matching factor of each demand migration cooling-off period channel.

[0157] The migration time window matching factor of each demand migration cooling-off period channel is mapped and matched with a mapping set of migration time window matching factors and migration time windows pre-stored in the database to obtain the migration time window of each demand migration cooling-off period channel.

[0158] When a migration cooling-off period channel for a certain requirement continuously triggers migration events, task migration is performed based on the migration time window.

[0159] See also Figure 7 The figure shows a schematic diagram of the process of result fusion and response instruction issuance involved in an embodiment of the present invention. The process mainly includes fusing all processing results to generate a unified system response instruction to control the execution of each photovoltaic device.

[0160] Each channel writes the processing results into a unified result queue, and the processing results are integrated to generate system response instructions, including:

[0161] After the channel completes task processing, it generates structured processing result data and encapsulates the processing result data into a message object in a standard format, including the task number, channel number, processing result type, and result generation time.

[0162] The encapsulated processing result data is asynchronously written into the unified result queue through the message middleware. The obtained results are classified and stored according to the task number, and fused. The system response instruction is generated according to the fusion result. The content of the system response instruction includes: instruction ID, target control device ID and control parameters.

[0163] The results are sorted according to the task number in each processing result data to ensure that the sub-processing results are grouped together. If a task involves multiple channel processing, all the scattered processing results under the same task number need to be aggregated together.

[0164] Based on the processing result type of each task, the demand fusion result type of each photovoltaic device data is obtained. According to the fusion strategy corresponding to the demand fusion result type, the classified processing results are fused. As shown in Table 1, in this embodiment of the present invention, the fusion strategies corresponding to different demand fusion result types include:

[0165] Table 1 Correspondence between demand fusion result type and fusion strategy

[0166]

[0167]

[0168] It should be noted that, in the embodiment of the present invention, the fusion results of energy load adjustment include but are not limited to power generation power adjustment, energy storage scheduling, etc., the fusion results of equipment start-stop control include but are not limited to starting the backup power supply, turning on the air-conditioning load, etc., the fusion results of status synchronization include but are not limited to equipment health status, abnormal alarms, etc., and the fusion results of emergency abnormality handling include but are not limited to load power off, equipment fault handling, etc.

[0169] Based on the fusion strategy corresponding to the demand fusion result type of each photovoltaic equipment data, the processing result data of each task is fused and processed to generate system response instructions.

[0170] The data response module is used to receive the system response instruction from the data processing module and send the system response instruction to each photovoltaic device for execution.

[0171] Receive the system response instructions from the data processing module and send the system response instructions to each photovoltaic device for execution. The specific processing conditions are:

[0172] The data response module parses the received system response instruction and extracts the content of the system response instruction, including the instruction ID, target control device ID and control parameters. Based on the content of the system response instruction, it extracts the instruction transmission interface of the target control device and converts the system response instruction into a control command set that can be recognized by the target control device according to the instruction transmission interface type of the target control device.

[0173] Obtaining the target control device's command transmission interface type involves parsing the target control device ID after receiving a system response command. This involves querying the device information database to extract the device's command transmission interface type. Interface types typically include standard Modbus, MQTT, HTTP / REST, custom socket protocols, and serial command line interfaces (UART / RS485).

[0174] Based on the interface type of the target control device, a predefined command adaptation template is matched. For example, in this embodiment of the present invention, Modbus specifies function codes, register addresses, and read and write instruction formats; MQTT specifies topics and message body structures (JSON / XML / custom); HTTP / REST specifies URLs, request methods (POST / PUT), and payload formats; and Socket specifies command start and end characters, frame headers, and frame trailers. Based on the interface type, the corresponding adaptation template is loaded.

[0175] Fill the field mapping extracted from the system response instruction into the field position of the adaptation template. Perform parameter verification and format encoding.

[0176] In an embodiment of the present invention, for example:

[0177] The response instruction of a system is power on, the target control device ID is Device_01, the target control device interface is Modbus, and the converted command (hexadecimal) is 0x01 0x05 0x00 0x01 0xFF 0x00CRC.

[0178] After the conversion is completed, the final control command set is formed which can be directly executed by the target control device.

[0179] The target control device performs the corresponding response operation after receiving the control command set.

[0180] See also Figure 2 As shown, in this embodiment, the present invention provides a building energy-saving optimization method based on photovoltaic power generation, specifically including:

[0181] Obtain the core data of energy operation in the building, dynamically adjust the data sampling frequency of each node based on the current load of the building, and each edge node collects local data of each connected photovoltaic device in real time based on the data sampling frequency.

[0182] The collection nodes of photovoltaic equipment are recorded as edge nodes, the data priority of each edge node is obtained, and the maximum upload rate of each edge node data is set. Each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate.

[0183] After receiving data from each edge node, it is broken down into several tasks and analyzed and processed through a multi-channel task scheduling architecture. Each channel writes the processing results into a unified result queue, which is then integrated to generate system response instructions.

[0184] Receive the system response instructions from the data processing module and send the system response instructions to each photovoltaic device for execution.

[0185] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0186] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all details in detail, nor do they limit the invention to specific implementation methods. Obviously, many modifications and changes can be made based on the contents of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can better understand and utilize the present invention. As long as they do not deviate from the structure of the present invention or exceed the scope defined by the present invention, they should fall within the scope of protection of the present invention.

Claims

1. Building energy-saving optimization system based on photovoltaic power generation, characterized by: include: The data acquisition module is used to obtain the core data of energy operation in the building. Based on the current load of the building, the data sampling frequency of each node is dynamically adjusted. Each edge node collects local data of each connected photovoltaic device in real time based on the data sampling frequency; The data upload module is used to record the collection nodes of the photovoltaic equipment as edge nodes, obtain the data priority of each edge node, set the maximum upload rate of each edge node data, and each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate; The data processing module is used to decompose the data received from each edge node into several tasks, analyze and process them through a multi-channel task scheduling architecture, and each channel writes the processing results into a unified result queue, which is then integrated to generate system response instructions; The data response module is used to receive the system response instruction from the data processing module and send the system response instruction to each photovoltaic device for execution.

2. The building energy-saving optimization system based on photovoltaic power generation according to claim 1, characterized in that: The local data of each connected photovoltaic device is collected in real time based on the data sampling frequency. The specific process is as follows: Obtain core data on energy operation within the building, including photovoltaic power output, building load power consumption, grid power output power, and energy storage capacity; The current photovoltaic power generation coverage result is obtained by comparing the photovoltaic power generation output power with the building load power. When the photovoltaic power generation output power is greater than the building load power, the photovoltaic power generation coverage result is full coverage. When the photovoltaic power generation output power is equal to the building load power, the photovoltaic power generation coverage result is balanced coverage. When the photovoltaic power generation output power is less than the building load power, the photovoltaic power generation coverage result is no coverage. If the photovoltaic power generation coverage result is full coverage, the energy storage device is activated to store excess electricity and it is determined that energy-saving optimization operations are required; If the photovoltaic power generation coverage result is not available, energy storage power will be used for supplementary power; If the photovoltaic power generation coverage result is balanced coverage, it is determined that no energy-saving optimization operation is required; When it is determined that energy-saving optimization operations are required, the data sampling frequency of each edge node is dynamically adjusted based on the current load level of the building. Each edge node collects local data of each connected photovoltaic device in real time based on the data sampling frequency.

3. The building energy-saving optimization system based on photovoltaic power generation according to claim 1, characterized in that: Each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate. The specific process is as follows: Obtain the number of data output destinations, the number of associated components, and the load support capacity value of each photovoltaic device, and correct the coupling processing to obtain the data priority of each photovoltaic device; Obtain the hardware performance parameters of the data processing module, including CPU occupancy, memory utilization, I / O bandwidth occupancy, and average data processing rate, and obtain the performance characteristic values of the data processing module after correcting the coupling process; The total amount of collected data of each photovoltaic device is obtained, the check value of the total amount of collected data is extracted from the database, the data weight corresponding to the data priority of each photovoltaic device is extracted from the database, and the maximum upload rate matching factor is obtained after correction coupling processing with the performance characteristic value of the data processing module. The maximum upload rate of each photovoltaic device data is obtained by mapping and matching with the mapping set corresponding to the maximum upload rate matching factor and the maximum upload rate preset in the database.

4. The building energy-saving optimization system based on photovoltaic power generation according to claim 1, characterized in that: After receiving the data from each edge node, it is broken down into several tasks, including: The data access type of each photovoltaic device data is obtained, and the data synchronization frequency of each photovoltaic device data is extracted. The data is combined with the total amount of collected data of each photovoltaic device data, and the coupling analysis is modified to obtain the lock granularity matching factor of each photovoltaic device data. The lock granularity matching factor of each photovoltaic device data is mapped and matched with the mapping set of lock granularity matching factors and lock granularities pre-stored in the database to obtain the lock granularity of each photovoltaic device data. The data of each photovoltaic device is then segmented and disassembled based on the lock granularity to obtain several tasks.

5. The building energy-saving optimization system based on photovoltaic power generation according to claim 1, characterized in that: The analysis and processing through the multi-channel task scheduling architecture specifically includes: Determine an identifier for each task, including the task's event ID, source device ID, timestamp combination, and payload type, where the identifier is used to uniquely identify each task; Based on a preset hash function, a hash function is performed on the task identifier to generate a hash value; Get the number of channels, perform a modulo operation on the hash value based on the number of channels, and map the task to the corresponding target channel; Add each task to its corresponding target channel task queue for data processing.

6. The building energy-saving optimization system based on photovoltaic power generation according to claim 1, characterized in that: The analysis and processing through the multi-channel task scheduling architecture specifically includes: Based on the monitoring agent deployed on each channel, load monitoring indicators are collected, including QPS, CPU utilization, memory usage, and queue length. After correction and coupling processing, the load monitoring characteristic value of each channel is obtained; Based on the load monitoring characteristic value of each channel, the load monitoring characteristic value is compared with the load monitoring characteristic threshold to obtain the load determination result of each channel. If the load monitoring characteristic value of a channel is greater than or equal to the load monitoring characteristic threshold, the load determination result provided is determined to be overloaded. If the load monitoring characteristic value of a channel is less than the load monitoring characteristic threshold, the load determination result provided is determined to be not overloaded. The channels whose load statistics are judged to be overloaded are recorded as overload channels, and an overload alarm signal is sent to the channel scheduling controller; After receiving the overload alarm signal, the channel scheduling controller triggers the task migration operation, which specifically includes: performing a subtraction process on the load monitoring characteristic value of each overload channel and the load monitoring characteristic threshold to obtain the load monitoring characteristic difference value of each overload channel; mapping and matching the load monitoring characteristic difference value of each overload channel with the mapping set corresponding to the load monitoring characteristic difference value and the task migration ratio pre-stored in the database to obtain the task migration ratio of each overload channel; Based on the task migration ratio of each overloaded channel, the demand migration tasks of each overloaded channel are obtained. After introducing the disturbance factor, the secondary hash value of each demand migration task is calculated again. According to the secondary hash value, each demand migration task is assigned to a new channel for analysis and processing.

7. The building energy-saving optimization system based on photovoltaic power generation according to claim 6, characterized in that: The triggering of task migration operation also includes setting a migration cool-off period, specifically including: If the number of demand migration tasks for a channel is greater than 1, it is determined that multiple task migration operations are required. Before triggering multiple migration operations, the load monitoring characteristic value of the channel is calculated again and compared with the demand migration threshold stored in the database to obtain the demand migration result; If the load monitoring characteristic value of the channel is less than or equal to the demand migration threshold, the demand migration cooling-off result of the channel is no demand migration cooling-off period; If the load monitoring characteristic value of the channel is greater than the demand migration threshold, the demand migration cooling-off result of the channel is the demand migration cooling-off period; The demand migration cooling-off results are statistically analyzed as channels during the demand migration cooling-off period, which are recorded as each demand migration cooling-off period channel. The overall occupancy of each demand migration cooling-off period channel is extracted and difference analysis is performed with the demand migration threshold to obtain the migration time window matching factor of each demand migration cooling-off period channel. The migration time window matching factor of each demand migration cooling-off channel is mapped and matched with a pre-stored mapping set of migration time window matching factors and migration time windows in the database to obtain the migration time window of each demand migration cooling-off channel. When a migration cooling-off period channel for a certain requirement continuously triggers migration events, task migration is performed based on the migration time window.

8. The building energy-saving optimization system based on photovoltaic power generation according to claim 1, characterized in that: Each channel writes the processing results into a unified result queue, and performs fusion to generate a system response instruction, specifically including: After the channel completes task processing, it generates structured processing result data and encapsulates the processing result data into a message object in a standard format, including the task number, channel number, processing result type, and result generation time; The encapsulated processing result data is asynchronously written into the unified result queue through the message middleware. The obtained results are classified and stored according to the task number, and fused. The system response instruction is generated according to the fusion result. The content of the system response instruction includes: control type, control object and control parameters.

9. The method for applying the photovoltaic power generation-based building energy-saving optimization system according to any one of claims 1 to 8, characterized in that: Obtain the core data of energy operation in the building, dynamically adjust the data sampling frequency of each node based on the current load of the building, and each edge node collects local data of each connected photovoltaic device in real time based on the data sampling frequency; The collection nodes of the photovoltaic equipment are recorded as edge nodes, the data priority of each edge node is obtained, and the maximum upload rate of each edge node data is set. Each edge node asynchronously reports the collected data stream to the data processing module at the maximum upload rate; After receiving data from each edge node, it is broken down into several tasks and analyzed and processed through a multi-channel task scheduling architecture. Each channel writes the processing results into a unified result queue, which is then integrated to generate system response instructions. Receive the system response instructions from the data processing module and send the system response instructions to each photovoltaic device for execution.

10. A device for applying the photovoltaic power generation-based building energy-saving optimization system according to any one of claims 1 to 8, characterized in that: The device has one or more programs, and the one or more programs are executed by one or more processors to implement the above system.

Images