A method and device for large-scale distributed computing power and power coordination

Abstract

This invention relates to the field of new energy power generation technology, and in particular to a method and apparatus for large-scale distributed computing power and power coordination. Based on preset conditions of the new energy power plant, this invention predicts the power generation of each power generation panel within a preset time period, and combines this prediction with the energy storage status of each micro-energy storage unit to form prediction data. Based on the prediction data, the energy storage capacity of each micro-energy storage unit is adjusted to reserve corresponding power, and computing tasks are allocated to each computing unit to avoid interruptions in computing due to insufficient power. This achieves a stable DC power supply from distributed micro-energy storage units to the computing units, fully utilizing the power generation potential of the new energy power plant, ensuring the stable and reliable operation of the computing units, and enabling dynamic allocation and scheduling of computing tasks according to the power generation and energy storage status of the micro-energy storage units and power generation panels, thereby achieving optimal system computing power performance.

Description

Technical Field

[0001] This invention relates to the field of new energy power generation technology, and in particular to a method and apparatus for large-scale distributed computing power and power coordination. Background Technology

[0002] The combination of renewable energy power plants and distributed computing centers can effectively improve power consumption while addressing the energy challenges and insufficient computing power of current computing centers. Traditional distributed computing centers require AC-DC conversion for their computing units. In renewable energy power plants, the output of the power generation panels is DC, so a direct DC-DC power supply method can be considered to power the distributed computing center.

[0003] In photovoltaic power plants, the power generation of photovoltaic panels is affected by factors such as weather and time, resulting in natural fluctuations in output power. If the allocation of computing power tasks in computing units is not combined with power generation forecasts and energy storage status, mismatches may easily occur, such as "insufficient computing power when there is a surplus of electricity" or "power shortage when the demand for computing power is high." This can lead to power waste or, in severe cases, computing interruptions due to insufficient power, thereby reducing the operating efficiency of the distributed computing center.

[0004] Therefore, overcoming the shortcomings of the existing technology is an urgent problem to be solved in this technical field. Summary of the Invention

[0005] The technical problem to be solved by this invention is: how to provide the necessary stable power supply to distributed computing units on demand.

[0006] The present invention adopts the following technical solution: Firstly, a method for large-scale distributed computing power and power coordination is provided, including: Based on the preset conditions of the new energy power station, the power generation of each power generation panel in the future preset time period is predicted, and the predicted data is formed by combining the energy storage of each micro energy storage unit. The energy storage capacity of each micro-energy storage unit is adjusted according to the predicted data to reserve the corresponding power, and the computing tasks are allocated to each computing unit to avoid the computing unit's calculation being interrupted due to insufficient power.

[0007] Preferably, when the new energy power station is a photovoltaic power station, the preset conditions include the photovoltaic deployment direction, solar angle, time information, historical power generation data, and weather forecast data.

[0008] Preferably, the method further includes: The system collects photovoltaic deployment direction, solar angle, time information, historical power generation data, and weather forecast data of photovoltaic power stations. Combined with real-time collected photovoltaic panel operating parameters, the system uses software to calculate and predict the power generation of each photovoltaic panel within a preset time period in the future. Each micro-energy storage unit collects its own energy storage capacity and charging / discharging efficiency parameters in real time, and predicts the power supply capacity it can provide to the corresponding computing unit within a preset time period. The power supply capacity includes the maximum output power, the duration of continuous power supply, and the power demand threshold of the computing unit that it can meet.

[0009] Preferably, the method further includes: The control center receives the forecast data reported by all micro energy storage units, and, in conjunction with the current computing load status of each computing unit and the computing power demand parameters of the computing tasks to be assigned, formulates and executes a dual adjustment strategy to coordinate the power supply from the micro energy storage units to the computing units.

[0010] Preferably, the dual adjustment strategy includes: Based on the power generation forecast of the photovoltaic panel and the power supply capacity forecast of the micro energy storage unit, an energy storage capacity adjustment command is issued to the corresponding micro energy storage unit to control the micro energy storage unit to adjust the charging and discharging rhythm, thereby storing enough electrical energy to meet the power demand of the corresponding computing unit in the future preset time period.

[0011] Preferably, the dual adjustment strategy further includes: Based on the predicted upper limit of the power supply capacity of each micro energy storage unit, the computing tasks to be assigned are decomposed according to the computing power demand and allocated to different distributed computing power units. This ensures that the computing power load of each distributed computing power unit does not exceed the computing power capacity limit corresponding to the predicted power supply capacity of its supporting micro energy storage unit, thus avoiding computing interruption due to insufficient power. At the same time, it enables each computing power unit to make full use of the power supply provided by the supporting micro energy storage unit.

[0012] Preferably, the method further includes: The control center receives heartbeat signals sent by each node through a preset network at preset time intervals. If no heartbeat signal is received from a node within N consecutive preset time intervals, it is determined that the micro energy storage unit of that node has failed, and / or the computing power unit has failed, where N is a positive integer greater than or equal to 1.

[0013] Preferably, the method further includes: When the control center determines that a node is faulty, it sends a route update instruction to other normal computing power nodes in the distributed computing network that have data interaction with the faulty node. The route update instruction contains the identification information of the faulty node and the bypass path.

[0014] Preferably, the method further includes: Each normal computing node, according to the routing update instruction, switches the data or computing tasks that were originally to be transmitted to the faulty node to a preset bypass path, and completes the data interaction and computing task takeover through other normal nodes, so as to realize the bypass of the faulty node and the continuous operation of the overall computing network function.

[0015] Secondly, a device for large-scale distributed computing power and power coordination is provided, comprising: Solar panels are used to convert solar energy into direct current (DC) electricity. The micro energy storage unit is connected to the power generation panel via MPPT (Maximum Power Point Tracking) and DC-DC. It is configured to dynamically detect fluctuations in the power generation of the power generation panel, perform surplus power storage, excess power absorption, and power output operations when photovoltaic power generation decreases, and provide a stable DC voltage for the computing unit. The computing unit is connected to the micro-energy storage unit via DC-DC, receives stable DC power, and performs computing tasks. The communication module includes a mesh network fiber optic component and a WIoTa (Wide-range Internet of Things communication protocol) wireless communication component. The mesh network fiber optic component is used to realize the connection between computing nodes, and the WIoTa wireless communication component is used to realize the connection between the micro energy storage unit and the control center. The control center is configured to detect node faults via heartbeat signals, perform bypass operations on faulty nodes, receive power generation prediction data from power generation panels and power supply prediction data from micro-energy storage units, adjust the energy storage capacity of micro-energy storage units, and allocate computing power tasks.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention predicts the power generation of each power generation panel within a preset time period based on the pre-defined conditions of a new energy power station, and combines this prediction with the energy storage status of each micro-energy storage unit to form prediction data. Based on this prediction data, the energy storage capacity of each micro-energy storage unit is adjusted to reserve corresponding power, and computational tasks are allocated to each computing unit to avoid computational interruptions due to insufficient power. This achieves a stable DC power supply from distributed micro-energy storage units to the computing units, fully utilizing the power generation potential of the new energy power station, ensuring the stable and reliable operation of the computing units, and enabling dynamic allocation and scheduling of computing tasks according to the power generation and energy storage status of the micro-energy storage units and power generation panels, thereby achieving optimal system computing performance. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating a method for large-scale distributed computing power and power coordination provided in an embodiment of the present invention; Figure 2 This is a signaling diagram of a method for large-scale distributed computing power and power coordination provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a process for a micro energy storage unit to obtain power supply capability according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a device for large-scale distributed computing power and power coordination provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] Unless the context otherwise requires, throughout the specification and claims, the term "comprising" is interpreted as openly inclusive, meaning "including, but not limited to." In the description of the specification, terms such as "one embodiment," "some embodiments," "exemplary embodiment," "example," "specific example," or "some examples" are intended to indicate that a particular feature, structure, material, or characteristic associated with that embodiment or example is included in at least one embodiment or example of this disclosure. The illustrative representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics mentioned may be included in any suitable manner in any one or more embodiments or examples; that is, although they may be incorporated into embodiments or examples using the above terms for reasons such as order and position, it does not limit them to be incorporated in combination by a single embodiment or example.

[0021] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of embodiments of this disclosure, unless otherwise stated, "a plurality of" means two or more. Furthermore, for example, the description may use the prefix "A" or "B" to describe the same type of nouns as two independent entities. In this case, the corresponding features defined with "A" and "B" are used only to distinguish between similar entities and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features.

[0022] In describing some embodiments, the terms "coupled," "coupled," and "connected," and their derivative expressions, may be used. For example, the term "connected" may be used in describing some embodiments to indicate that two or more components have direct physical or electrical contact with each other. Similarly, the term "coupled" may be used in describing some embodiments to indicate that two or more components have direct physical or electrical contact. However, the terms "connected" or "coupled" may also refer to two or more components that do not have direct contact with each other but still cooperate or interact with each other, such as "optical coupling," "wireless connection," etc. The embodiments disclosed herein are not necessarily limited to the scope of this invention.

[0023] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0024] Example 1: To address the problems of existing technologies, this embodiment proposes a method for large-scale distributed computing power and power coordination. In one embodiment, such as... Figure 1 and Figure 2 As shown, it includes: Step 101: Based on the preset conditions of the new energy power station, predict the power generation of each power generation panel in the future preset time period, and combine the power storage of each micro energy storage unit to form prediction data.

[0025] The new energy power station can be a hydroelectric power station, a wind power station, or a photovoltaic power station. This embodiment uses a photovoltaic power station as an example, and the power generation panel is a photovoltaic panel. In one embodiment, when the new energy power station is a photovoltaic power station, the preset conditions include the photovoltaic deployment direction, solar angle, time information, historical power generation data, and weather forecast data (such as light intensity, cloud cover, and probability of rainfall).

[0026] Based on the above conditions, the hourly power generation of each solar panel within a preset future time period (which can be set according to the power generation fluctuation pattern or the calculation task cycle, such as 1 hour, 4 hours, etc.) is calculated through a preset algorithm (such as a power generation prediction model combined with machine learning), and the power output potential of a single solar panel is determined.

[0027] Each micro energy storage unit collects in real time parameters such as its current power storage, remaining energy storage capacity, charge-discharge efficiency, and health status, and combines the predicted power generation of the corresponding solar panel to calculate its stable power supply capacity (such as maximum power supply, continuous power supply duration, upper limit of computational load that can be supported) within a preset future time period. Finally, the prediction data including the predicted power generation of a single solar panel and the predicted power supply capacity of the corresponding micro energy storage unit is integrated.

[0028] Step 102: Adjust the energy storage capacity of each micro energy storage unit according to the prediction data to reserve the corresponding power, and allocate calculation tasks to each computing unit to avoid interruption of the calculation of the computing unit due to insufficient power.

[0029] Among them, after the control center receives the prediction data reported by all micro energy storage units, a corresponding charge-discharge strategy is formulated for each micro energy storage unit. In one embodiment, if it is predicted that the future power generation of the corresponding solar panel is sufficient, an instruction is sent to the micro energy storage unit to reserve enough energy storage space in advance to absorb excess power during the power generation peak period; if it is predicted that the power generation is insufficient, the micro energy storage unit is instructed to reserve sufficient power in advance to ensure that the power supply is not interrupted during the subsequent computing tasks.

[0030] In one embodiment, the control center splits and allocates tasks according to the predicted power supply capacity of the micro energy storage units supporting each computing unit (especially the upper limit of the computable load that can be supported), combined with the computing power requirements of the calculation tasks to be allocated (such as computational complexity, required computing resources, task execution duration). High-computing-power-demand tasks are allocated to nodes with sufficient power supply capabilities, and low-computing-power-demand tasks or short-duration tasks are allocated to nodes with limited power supply capabilities, ensuring that the load of each computing unit does not exceed the power supply capacity of its supporting micro energy storage unit, and at the same time enabling each node to make full use of its own power resources to avoid power waste or calculation interruption.

[0031] Compared with the prior art, the beneficial effects of this embodiment are as follows: This embodiment predicts the power generation of each power generation panel within a preset time period based on the pre-defined conditions of the new energy power station, and combines this prediction with the energy storage status of each micro-energy storage unit to form prediction data. Based on this prediction data, the energy storage capacity of each micro-energy storage unit is adjusted to reserve corresponding power, and computational tasks are allocated to each computing unit to avoid computational interruptions due to insufficient power. This achieves a stable DC power supply from distributed micro-energy storage units to the computing units, fully utilizing the power generation potential of the new energy power station, ensuring the stable and reliable operation of the computing units, and enabling dynamic allocation and scheduling of computing tasks according to the power generation and energy storage status of the micro-energy storage units and power generation panels, thereby achieving optimal system computing performance.

[0032] In one embodiment, such as Figure 3 As shown, the method further includes: Step 201: Collect photovoltaic deployment direction, solar angle, time information, historical power generation data and weather forecast data of the photovoltaic power station, and combine them with the real-time collected photovoltaic panel operating parameters to calculate and predict the power generation of each photovoltaic panel in the future preset time period through software.

[0033] Data collection is based on static deployment data, dynamic environmental data, and real-time operational data. Static data includes the photovoltaic deployment direction (e.g., south-facing / east-southeast) and the tilt angle of the photovoltaic panels; dynamic data includes real-time solar altitude angle / azimuth angle, current time information, historical power generation data for the same period (same season, same weather type), and authoritative weather forecast data (solar intensity, cloud cover, rainfall / dust storms, and other weather conditions affecting sunlight); real-time operational data includes the temperature, output voltage, and output current of the photovoltaic panels.

[0034] All the above data are input into a preset software model (such as a machine learning model or a multiple linear regression model), and the data is weighted and analyzed by an algorithm. For example, historical data is combined to correct the error of weather forecasts, and real-time voltage / current data is used to calibrate the power generation baseline. Finally, the predicted power generation value of each photovoltaic panel in the future preset time period (such as 1 hour or 3 hours, which can be flexibly set according to the computing power task cycle) is output, clarifying the upper limit and trend of power output of a single photovoltaic panel.

[0035] Step 202: Each micro energy storage unit collects its own energy storage capacity and charging / discharging efficiency parameters in real time, and predicts the power supply capacity it can provide to the corresponding computing unit within a preset time period. The power supply capacity includes the maximum output power, the duration of continuous power supply, and the power demand threshold of the computing unit that it can meet.

[0036] Among them, the micro energy storage unit collects two types of core parameters in real time through built-in sensors, including energy storage status data (current remaining power and remaining energy storage capacity percentage) and charge and discharge performance data (charge and discharge conversion efficiency, maximum charge and discharge rate, and continuous charge and discharge stability parameters).

[0037] Based on its own collected parameters and the predicted power generation value of the corresponding photovoltaic panel, the micro-energy storage unit calculates its power supply capacity within a preset time period. The maximum output power refers to the peak power that the energy storage unit can stably output under safe operation conditions; the continuous power supply duration refers to the time that the power supply can be maintained at the preset power output, combined with remaining power and photovoltaic supplemental power; and the computing unit power demand threshold refers to the minimum power and voltage range that can ensure the normal operation of the computing unit (without lag or interruption). The final result is a power supply capacity prediction that includes these three types of indicators.

[0038] The above process resolves the mismatch between power supply and computing power demand caused by the volatility and unpredictability of photovoltaic (PV) power generation. PV power generation is greatly affected by natural factors such as sunlight and weather, resulting in random fluctuations in output power. Traditional solutions cannot predict changes in power generation in advance, easily leading to contradictions such as insufficient power generation when computing power demand is high or idle computing power when power generation is sufficient. This process solves the problem of unclear power supply capacity of micro-energy storage units. Traditional solutions only know the current remaining power of the energy storage unit and cannot accurately determine the upper limit and duration of the computing power load it can support in the future, leading to blind allocation of computing power and easy computational interruptions due to insufficient power supply.

[0039] In one embodiment, the method further includes: the control center receiving the prediction data reported by all micro energy storage units, and combining the current computing load status of each computing unit with the computing power demand parameters of the computing tasks to be assigned, formulating and executing a dual adjustment strategy to coordinate the power supply of the micro energy storage units to the computing units.

[0040] The dual adjustment strategy includes: issuing energy storage capacity adjustment instructions to the corresponding micro energy storage units based on the photovoltaic power generation forecast results and the power supply capacity forecast results of the micro energy storage units, so as to control the micro energy storage units to adjust the charging and discharging rhythm, thereby storing enough electrical energy to meet the power demand of the corresponding computing units in a future preset time period.

[0041] The control center first aggregates two types of core forecast data: the predicted future power generation of each photovoltaic panel (e.g., time-period output power curve) and the predicted power supply capacity of each micro-energy storage unit (including maximum power, continuous power supply duration, and computing power demand threshold). Simultaneously, it acquires real-time operational data, including the current computing power load status of each computing unit (e.g., the percentage of occupied computing resources and the remaining maximum load capacity) and the computing power demand parameters for tasks to be assigned (e.g., the power required for the task, execution time, and minimum power supply requirements).

[0042] Then, a data association mapping is established, binding photovoltaic panels, micro-energy storage units, and corresponding computing units as independent nodes, clarifying the corresponding relationships between the power generation potential, power supply capacity, and computing power requirements of each node. Based on the data matching status of each node, the control center formulates differentiated charging and discharging rhythm instructions. The core content is mainly divided into three scenarios: Scenario 1: When photovoltaic power generation is sufficient and computing power demand is high, if it is predicted that the future power generation of photovoltaic panels will be sufficient and the power demand of the corresponding computing power unit to be assigned tasks is large, the control center will issue an instruction to reserve energy storage space, so that the micro energy storage unit can reduce the current charge and discharge ratio and reserve enough capacity to absorb excess power during peak power generation, thus avoiding waste of photovoltaic power.

[0043] Scenario 2: When photovoltaic power generation is insufficient and computing power demand is high, if it is predicted that the future power generation of photovoltaic panels will be insufficient, but the corresponding computing power unit has a high load task demand, an advance charging reserve instruction is issued to control the micro energy storage unit to charge at the maximum safe charging and discharging efficiency to make up for the future power supply gap and ensure that the power supply is not interrupted when the task is executed.

[0044] Scenario 3: When photovoltaic power generation fluctuates and computing power demand is stable, if the predicted photovoltaic power generation fluctuates (such as intermittent output caused by cloud cover), a dynamic adjustment command for the charging and discharging rate is issued, allowing the micro-energy storage unit to quickly absorb power during peak power generation and release it smoothly during off-peak periods, maintaining stable output voltage / power and adapting to the stable operation requirements of the computing unit.

[0045] After receiving instructions, the micro-energy storage unit adjusts the charging and discharging current, voltage, and duty cycle through its built-in DC-DC conversion module and charging and discharging control circuit, strictly adhering to the energy storage capacity target set in the instructions (such as "increase the energy storage capacity to 80% within 1 hour" or "dynamically adjust the energy storage capacity within the 50%-70% range"). Finally, the micro-energy storage unit feeds back the charging and discharging status (such as the current energy storage capacity and actual charging and discharging power) to the control center in real time. The control center continuously monitors and dynamically corrects the instructions to ensure that the energy storage adjustment is consistent with the predicted power generation and computing power demands.

[0046] The above process resolves the contradiction between the volatility of photovoltaic (PV) power generation and the rigidity of computing power demands. PV power generation is susceptible to fluctuations due to natural factors, while computing tasks require a high degree of power supply stability. By adjusting the charging and discharging rhythm of energy storage in advance, power generation fluctuations can be smoothed out, providing continuous and stable power support for computing units. This avoids situations where energy storage is full and unable to absorb power when generation is sufficient, or empty and unable to be replenished when generation is insufficient, allowing energy storage capacity to precisely match future computing power demands and improving energy utilization efficiency. By storing corresponding power in advance, it ensures that high-load or long-duration computing tasks will not be interrupted due to power shortages, guaranteeing the continuity of computing tasks.

[0047] In one embodiment, the dual adjustment strategy further includes: based on the predicted upper limit of the power supply capacity of each micro energy storage unit, decomposing the computing tasks to be allocated according to the computing power demand and allocating them to different distributed computing power units, ensuring that the computing power load of each distributed computing power unit does not exceed the computing power capacity limit corresponding to the predicted power supply capacity of its supporting micro energy storage unit, avoiding computing interruption due to insufficient power, and at the same time enabling each computing power unit to make full use of the power supply provided by the supporting micro energy storage unit.

[0048] The control center first breaks down all computing tasks to be assigned and extracts core computing power requirements parameters, including the computing power resources required for a single task, such as the utilization rate of CPU (Central Processing Unit) and GPU (Graphics Processing Unit), the corresponding power consumption threshold (such as the power required per unit of computing power), task execution time, priority (high / medium / low), and whether it can be split (such as data processing tasks can be split, while real-time transaction tasks cannot be split).

[0049] Then, based on the predicted upper limit of the power supply capacity of each micro energy storage unit (maximum power and continuous power supply duration), combined with the hardware performance of the supporting computing power unit (such as computing power density and power-to-computing power conversion coefficient), the computing power carrying capacity upper limit of each computing power unit is calculated, that is, the maximum computing power load that the node can carry under the premise of stable power supply in the future preset time period.

[0050] In one embodiment, the control center formulates differentiated allocation rules based on the principles of supply and demand matching and optimal efficiency, as follows: Rule 1: Priority First + Power Adaptation. High-priority tasks are assigned to computing units with sufficient power supply capacity (high capacity limit, stable power supply) to ensure that core tasks are not affected by power fluctuations; low-priority tasks are assigned to nodes with limited power supply capacity but sufficient to meet the needs to avoid resource waste.

[0051] Rule 2: For large tasks whose computing power requirements exceed the capacity limit of a single node, they are split into multiple sub-tasks according to computing power requirements and allocated to multiple nodes with matching power supply capacity. They are then completed through distributed parallel computing, while ensuring that the total power requirements of all sub-tasks do not exceed the sum of the capacity limits of the corresponding nodes.

[0052] Rule 3: During the allocation process, the current load of each computing unit is monitored in real time. If a node has sufficient remaining capacity, low-priority subtasks can be added as appropriate. If the load of a node is close to its capacity limit, the allocation of new tasks is suspended to avoid power supply overload.

[0053] The control center synchronizes the actual load changes of each computing unit (such as current computing power utilization and real-time power consumption) and the actual power supply status of each micro-energy storage unit (such as current output power and remaining energy storage capacity) in real time. If a deviation occurs (such as photovoltaic power generation being lower than the predicted value, resulting in a decrease in power supply capacity), dynamic adjustment is immediately initiated, which involves migrating some low-priority tasks of overloaded nodes to nodes with less load, or suspending non-core tasks to ensure that the power supply of core tasks is not affected.

[0054] The above process solves the problem of the disconnect between computing power task allocation and power supply. Traditional allocation methods only consider the hardware performance of computing units and ignore power supply constraints, which can easily lead to computing power load exceeding power capacity and causing task interruption. This strategy takes power supply as the core constraint and avoids the risk of power shortage from the source.

[0055] In one embodiment, the method further includes: the control center receiving heartbeat signals sent by each node through a preset network (e.g., mesh network fiber optic and WIoTa wireless connection) at preset time intervals; if no heartbeat signal is received from a certain node within N consecutive preset time intervals, it is determined that the micro energy storage unit of that node has failed, and / or the computing power unit has failed, where N is a positive integer greater than or equal to 1.

[0056] Each node, consisting of a micro-energy storage unit and a computing unit, transmits its heartbeat signal via a mesh network fiber optic channel and a WIoTa wireless channel. The heartbeat signal contains core information such as a unique node identifier, current operating status code, and signal transmission timestamp. The control center continuously receives dual-channel heartbeat signals from all nodes at preset fixed time intervals (e.g., 10 seconds, 30 seconds, adjustable according to system response requirements), establishing a linked ledger of node identifiers, signal reception times, and operating status.

[0057] The control center verifies the heartbeat signal reception status of each node in real time. If no heartbeat signal is received from the same node within N consecutive preset time intervals (N≥1, e.g., N=2 means two consecutive intervals without reception), a fault determination is triggered. Once the control center determines a fault, it immediately generates a fault report, including the fault node identifier, fault type (single device / dual device fault), determination time, channel type where the heartbeat signal was not received, and the number of consecutive times the signal was not received. Simultaneously, the fault information is synchronized to the system log, providing a basis for subsequent fault investigation and recovery.

[0058] In one embodiment, the method further includes: when the control center determines that a node is faulty, sending a route update instruction to other normal computing power nodes in the distributed computing network that have data interaction with the faulty node, wherein the route update instruction contains the identification information of the faulty node and a bypass path.

[0059] Once the control center determines the fault (based on the absence of a heartbeat signal), it immediately initiates a route update process to ensure the impact of the fault does not spread. Specifically, the control center quickly retrieves two types of core data: the unique identifier of the faulty node (such as node ID and physical address) and the topology of the distributed computing network (including the connection relationships of all normal nodes and data transmission links). Simultaneously, it identifies normal nodes that are interacting with the faulty node in real time (such as nodes sending or receiving data to, receiving data from, or collaborating with the faulty node).

[0060] In one embodiment, the control center plans bypass paths based on the network topology, following the principles of shortest path, load balancing, and low latency. Specifically, prioritize selecting normal nodes directly adjacent to the faulty node as bypass relay points to shorten the data transmission path and reduce latency; if the adjacent node is currently under high load (e.g., computing power utilization exceeds 70%), select the next closest low-load node to avoid the bypass node from causing new failures due to overload; the bypass path should avoid other potentially unstable nodes (e.g., nodes that have recently experienced heartbeat signal fluctuations) to ensure path reliability.

[0061] In one embodiment, the routing update instruction includes three key pieces of information: fault node identifier (clearly identifying the node to be isolated), bypass path details (e.g., "Node A → Node C → Node D" replacing the original "Node A → Faulty Node B → Node D"), and data forwarding priority (core data is forwarded first, non-core data is forwarded with normal priority). Instructions are sent in batches to the target healthy nodes (nodes that have data interaction with the faulty node) via the mesh network's Fibre Channel, ensuring that all relevant nodes receive and execute the instructions synchronously, thus avoiding data transmission conflicts.

[0062] In one embodiment, the method further includes: each normal computing node, according to the routing update instruction, switches the data or computing tasks that were originally to be transmitted to the faulty node to a preset bypass path, and completes the data interaction and computing task takeover through other normal nodes, so as to realize the bypass of the faulty node and the continuous operation of the overall computing network function.

[0063] The normal node that finally receives the route update instruction immediately updates its local routing table, switches the data flow that originally pointed to the faulty node to the bypass path, and forwards data or undertakes collaborative computing tasks according to the new path. After each execution node completes the route update, it sends a route update success confirmation signal to the control center. After the control center verifies that all relevant nodes have completed the update, the fault isolation and bypass deployment are completed.

[0064] In summary, when a node's micro-energy storage unit and computing unit fail, the control center can determine that the node has failed because it does not receive the corresponding heartbeat signal. By bypassing the node in the distributed computing network, fault traversal can be achieved, thus ensuring that the overall computing network function is not affected even when there are a few faulty nodes.

[0065] Example 2: This embodiment proposes a device for large-scale distributed computing power and power coordination. In one embodiment, such as... Figure 4 As shown, the device includes: a power generation panel for converting solar energy into DC power; a micro energy storage unit connected to the power generation panel via MPPT and DC-DC converter, configured to dynamically detect power generation fluctuations of the power generation panel, perform surplus power storage, excess power absorption, and power output operations when photovoltaic power generation decreases, and provide a stable DC voltage for the computing unit; a computing unit connected to the micro energy storage unit via DC-DC converter, receiving a stable DC power supply and performing computing tasks; a communication module (not shown in the figure), including a mesh network fiber optic component and a WIoTa wireless communication component, the mesh network fiber optic component being used to realize the connection between computing nodes, and the WIoTa wireless communication component being used to realize the connection between the micro energy storage unit and the control center; a control center configured to detect node faults through heartbeat signals, perform bypass operations on faulty nodes, and receive power generation prediction data from the power generation panel and power supply prediction data from the micro energy storage unit, adjust the energy storage capacity of the micro energy storage unit, and allocate computing tasks.

[0066] The specific process of the method for large-scale distributed computing power and power coordination is described in Example 1, and will not be repeated in this example.

[0067] This embodiment predicts the power generation of each power generation panel within a preset time period based on the pre-defined conditions of the new energy power station, and combines this prediction with the energy storage status of each micro-energy storage unit to form prediction data. Based on this prediction data, the energy storage capacity of each micro-energy storage unit is adjusted to reserve corresponding power, and computational tasks are allocated to each computing unit to avoid computational interruptions due to insufficient power. This achieves a stable DC power supply from distributed micro-energy storage units to the computing units, fully utilizing the power generation potential of the new energy power station, ensuring the stable and reliable operation of the computing units, and enabling dynamic allocation and scheduling of computing tasks according to the power generation and energy storage status of the micro-energy storage units and power generation panels, thereby achieving optimal system computing performance.

[0068] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for large-scale distributed power and electricity collaboration, characterized in that, include: Based on the preset conditions of the new energy power station, the power generation of each power generation panel in the future preset time period is predicted, and the predicted data is formed by combining the energy storage of each micro energy storage unit. The energy storage capacity of each micro-energy storage unit is adjusted according to the predicted data to reserve the corresponding power, and the computing tasks are allocated to each computing unit to avoid the computing unit's calculation being interrupted due to insufficient power.

2. The method for large-scale distributed algorithm power and electricity collaboration according to claim 1, characterized in that, When the new energy power station is a photovoltaic power station, the preset conditions include the photovoltaic deployment direction, solar angle, time information, historical power generation data, and weather forecast data.

3. The method of claim 2, wherein, The method further includes: The system collects information on the photovoltaic deployment direction, solar angle, time, historical power generation data, and weather forecast data of photovoltaic power stations. Combined with real-time collected photovoltaic panel operating parameters, the system uses software to calculate and predict the power generation of each photovoltaic panel within a preset time period in the future. Each micro-energy storage unit collects its own energy storage capacity and charging / discharging efficiency parameters in real time, and predicts the power supply capacity it can provide to the corresponding computing unit within a preset time period. The power supply capacity includes the maximum output power, the duration of continuous power supply, and the power demand threshold of the computing unit that it can meet.

4. The method of claim 2, wherein, The method further includes: The control center receives the forecast data reported by all micro energy storage units, and, in conjunction with the current computing load status of each computing unit and the computing power demand parameters of the computing tasks to be assigned, formulates and executes a dual adjustment strategy to coordinate the power supply from the micro energy storage units to the computing units.

5. The method of claim 4, wherein, The dual adjustment strategy includes: Based on the power generation forecast of the photovoltaic panel and the power supply capacity forecast of the micro energy storage unit, an energy storage capacity adjustment command is issued to the corresponding micro energy storage unit to control the micro energy storage unit to adjust the charging and discharging rhythm, thereby storing enough electrical energy to meet the power demand of the corresponding computing unit in the future preset time period.

6. The method of claim 4, wherein, The dual adjustment strategy also includes: Based on the predicted upper limit of the power supply capacity of each micro energy storage unit, the computing tasks to be assigned are decomposed according to the computing power demand and allocated to different distributed computing power units. This ensures that the computing power load of each distributed computing power unit does not exceed the computing power capacity limit corresponding to the predicted power supply capacity of its supporting micro energy storage unit, thus avoiding computing interruption due to insufficient power. At the same time, it enables each computing power unit to make full use of the power supply provided by the supporting micro energy storage unit.

7. The method of claim 1, wherein, The method further includes: The control center receives heartbeat signals sent by each node through a preset network connection at preset time intervals. If no heartbeat signal is received from a node within N consecutive preset time intervals, it is determined that the micro energy storage unit of that node has failed, and / or the computing power unit has failed, where N is a positive integer greater than or equal to 1.

8. The method for large-scale distributed algorithm power and electricity collaboration according to claim 7, characterized in that, The method further includes: When the control center determines that a node is faulty, it sends a route update instruction to other normal computing power nodes in the distributed computing network that have data interaction with the faulty node. The route update instruction contains the identification information of the faulty node and the bypass path.

9. The method of claim 8, wherein, The method further includes: Each normal computing node, according to the routing update instruction, switches the data or computing tasks that were originally to be transmitted to the faulty node to a preset bypass path, and completes the data interaction and computing task takeover through other normal nodes, so as to realize the bypass of the faulty node and the continuous operation of the overall computing network function.

10. An apparatus for large-scale distributed computing power and electricity coordination, characterized in that, For implementing the method of large-scale distributed computing power and power coordination as described in any one of claims 1-9, the apparatus comprises: Solar panels are used to convert solar energy into direct current (DC) electricity. The micro energy storage unit is connected to the power generation panel via MPPT and DC-DC converter. It is configured to dynamically detect fluctuations in power generation from the power generation panel, perform surplus power storage, excess power absorption, and power output operations when photovoltaic power generation decreases, and provide a stable DC voltage for the computing unit. The computing unit is connected to the micro-energy storage unit via DC-DC, receives stable DC power, and performs computing tasks. The communication module includes a mesh network fiber optic component and a WIoTa wireless communication component. The mesh network fiber optic component is used to realize the connection between computing nodes, and the WIoTa wireless communication component is used to realize the connection between the micro energy storage unit and the control center. The control center is configured to detect node faults via heartbeat signals, perform bypass operations on faulty nodes, receive power generation prediction data from power generation panels and power supply prediction data from micro-energy storage units, adjust the energy storage capacity of micro-energy storage units, and allocate computing power tasks.

Images