A new energy power and distributed supercomputer fusion architecture and an energy scheduling method thereof

Abstract

The present application relates to the field of new energy power generation and distributed supercomputing technology, in particular to a new energy power and distributed supercomputing fusion architecture and an energy scheduling method thereof. The present application realizes grid connection of the supercomputing energy storage array when power is in surplus through a bidirectional inverter unit, and cooperates with the power grid to take power for reverse energy compensation when power is insufficient, thereby breaking through the power supply bottleneck of a single power generation panel. At the same time, the multi-source power sharing mechanism can use the output of other power generation panel arrays or the power grid power to continue power supply when the local power supply of the supercomputing unit is insufficient or power generation is unbalanced, thereby ensuring uninterrupted and stable operation of the supercomputing unit, and improving the reliability of computing power tasks.

Description

Technical Field

[0001] This invention relates to the fields of new energy power generation and distributed supercomputing technology, and in particular to a new energy power and distributed supercomputing fusion architecture and its energy scheduling method. Background Technology

[0002] Due to the characteristics and geographical limitations of renewable energy generation, there is a significant amount of power curtailment. Meanwhile, supercomputing faces the problem of insufficient or excessively high power costs. Distributed supercomputing places a low-power (e.g., 300W) distributed supercomputing unit under each power generation panel to absorb the power generated by the power generation panel. This achieves low-cost supercomputing while also solving the problem of power curtailment from renewable energy generation.

[0003] To address the inherent instability of renewable energy generation, a battery bank was added under each power generation panel to store energy and ensure a stable power supply for the supercomputing unit. However, this still cannot fully guarantee the supercomputing unit's power stability requirements. One reason is the limited storage capacity of the batteries, which decreases over time due to aging. Another reason is the impact of environmental factors on renewable energy generation; a single power generation panel may be affected by weather, angle, equipment aging, etc., preventing a stable 24 / 7 power supply. Finally, the computing unit requires a long-term, reliable, and stable power supply that adapts to dynamic computing loads. Under the aforementioned conditions, the power from a single power generation panel and its associated batteries cannot continuously provide the power needed for the dynamic computing load for an extended period, thus failing to guarantee the stable operation of the computing unit.

[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 address the issue of insufficient power supply faced by distributed supercomputing, and the fact that existing single power generation boards and energy storage architectures cannot guarantee long-term stable power supply to supercomputing units, thereby seriously affecting the reliability of computing tasks.

[0006] The present invention adopts the following technical solution: In the first aspect, a new energy power and distributed supercomputing integrated architecture is provided, including: a supercomputing energy storage array, a bidirectional inverter unit, at least one power generation panel array and at least one unidirectional inverter unit; The supercomputing energy storage array is connected to the DC terminal of the bidirectional inverter unit, the power generation panel array is connected to the DC terminal of the unidirectional inverter unit, the AC terminals of the bidirectional inverter unit and the unidirectional inverter unit are respectively connected to the AC bus, and the AC bus is connected to the power grid through a transformer. When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit is used to invert the excess DC power into AC power and transmit it to the AC bus; when the supercomputing energy storage array has a power shortage, the bidirectional inverter unit is used to obtain AC power from the AC bus and rectify it into DC power to power the supercomputing units in the supercomputing energy storage array. The unidirectional inverter unit is used to invert the DC power generated by the corresponding power generation panel array into AC power and transmit it to the AC bus. When all the new energy power collected by the AC bus is still insufficient to meet the power demand of the supercomputing unit, the power grid replenishes the AC bus through the transformer to ensure the uninterrupted and stable operation of the supercomputing unit.

[0007] Preferably, the supercomputing energy storage array includes at least one supercomputing energy storage module, and the supercomputing energy storage module includes a power generation unit, a supercomputing unit, and a micro energy storage unit; The input terminal of the micro energy storage unit is connected to the DC output terminal of the power generation unit and the DC terminal of the bidirectional inverter unit, respectively. The micro energy storage unit is used to receive and store DC power from the power generation unit, as well as supplementary power after rectification by the bidirectional inverter unit, and continuously output stable DC power to the supercomputing unit to ensure uninterrupted operation of the supercomputing unit.

[0008] Preferably, the supercomputing energy storage module further includes a charge-discharge control unit, which is connected in series between the micro-energy storage unit and the supercomputing unit; The charging and discharging control unit is used to automatically control the charging, discharging, and voltage regulation output of the micro-energy storage unit based on the output power of the power generation unit, the remaining power of the micro-energy storage unit, and the load status of the supercomputing unit.

[0009] Preferably, the supercomputing energy storage module further includes an MPPT (Maximum Power Point Tracking) unit; the input terminal of the MPPT unit is connected to the DC output terminal of the power generation unit, and the output terminal of the MPPT unit is connected to the charging terminal of the micro energy storage unit. The MPPT unit is used to track the maximum power point of the power generation unit in real time, and dynamically adjust the working voltage and current according to the changes in light intensity and load, so as to convert the unstable new energy power output into stable DC power, thereby charging the micro energy storage unit.

[0010] Secondly, a method for energy scheduling of a new energy power and distributed supercomputing fusion architecture is provided, applied to the new energy power and distributed supercomputing fusion architecture described in the first aspect, the method comprising: When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit will invert the excess DC power into AC power and transmit it to the AC bus. When the supercomputing energy storage array is short of power, the bidirectional inverter unit obtains AC power from the AC bus and rectifies it into DC power to power the supercomputing units in the supercomputing energy storage array. The unidirectional inverter unit inverts the DC power generated by the corresponding power generation panel array into AC power and transmits it to the AC bus. When all the new energy power collected by the AC bus is still insufficient to meet the power demand of the supercomputing unit, the power grid replenishes the AC bus through the transformer to ensure the uninterrupted and stable operation of the supercomputing unit.

[0011] Thirdly, a new energy power and distributed supercomputing fusion architecture is provided, including: a supercomputing energy storage array, a bidirectional inverter unit and at least one power generation panel array; The supercomputing energy storage array and the at least one power generation panel array are respectively connected to the DC bus. The DC terminal of the bidirectional inverter unit is connected to the DC bus, and the AC terminal of the bidirectional inverter unit is connected to the power grid through a transformer. When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit is used to invert the excess DC power on the DC bus into AC power and transmit it to the power grid through a transformer; When the supercomputing energy storage array is short of power, the bidirectional inverter unit is used to obtain AC power from the grid through a transformer and rectify it into DC power to power the supercomputing units in the supercomputing energy storage array. The power generation array is used to output the generated DC power to the DC bus to provide shared power support for the supercomputing energy storage array.

[0012] Preferably, the supercomputing energy storage array includes at least one supercomputing energy storage module, and the supercomputing energy storage module includes a power generation unit, a supercomputing unit, and a micro energy storage unit; The input terminal of the micro energy storage unit is connected to the DC output terminal and the DC bus of the power generation unit, respectively. The micro energy storage unit is used to receive and store DC power from the power generation unit, as well as supplementary power after rectification by the bidirectional inverter unit, and continuously output stable DC power to the supercomputing unit to ensure uninterrupted operation of the supercomputing unit.

[0013] Preferably, the supercomputing energy storage module further includes a charge-discharge control unit, which is connected in series between the micro-energy storage unit and the supercomputing unit; The charging and discharging control unit is used to automatically control the charging, discharging, and voltage regulation output of the micro-energy storage unit based on the output power of the power generation unit, the remaining power of the micro-energy storage unit, and the load status of the supercomputing unit.

[0014] Preferably, the supercomputing energy storage module further includes a step-down unit, the input terminal of which is connected to the DC bus, and the output terminal of which is connected to the power supply terminal of the supercomputing unit. When the load on the supercomputing unit increases and the remaining power of the corresponding micro-energy storage unit is insufficient, the step-down unit is used to step down the DC power generated by the power generation array and then transmit it to the supercomputing unit for supplementary power supply, so as to ensure the stable operation and power safety of the supercomputing unit in the multi-power supply mode.

[0015] Fourthly, an energy scheduling method for a fusion architecture of new energy power and distributed supercomputing is provided, applied to the fusion architecture of new energy power and distributed supercomputing described in the third aspect, the method comprising: When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit inverts the excess DC power on the DC bus into AC power and transmits it to the power grid through a transformer; When the supercomputing energy storage array is short of power, the bidirectional inverter unit obtains AC power from the grid through a transformer and rectifies it into DC power to power the supercomputing units in the supercomputing energy storage array. At the same time, the power generation array outputs the generated DC power to the DC bus to provide shared power support for the supercomputing energy storage array.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention enables the supercomputing energy storage array to connect to the grid when there is a power surplus through a bidirectional inverter unit, and to draw power from the grid for reverse energy replenishment when there is a power shortage, thus breaking through the power supply bottleneck of a single power generation board. At the same time, the multi-source power sharing mechanism can utilize the output of other power generation board arrays or grid power to continuously supply power when the local power supply of the supercomputing unit is insufficient or the power generation is unbalanced, ensuring the uninterrupted and stable operation of the supercomputing unit, thereby improving the reliability of computing tasks. 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 schematic diagram of a new energy power and distributed supercomputing fusion architecture provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of the communication structure between a task distribution server and a supercomputing unit provided in an embodiment of the present invention; Figure 3This is a flowchart illustrating an energy scheduling method based on a new energy power and distributed supercomputing fusion architecture provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of another new energy power and distributed supercomputing fusion architecture provided by an embodiment of the present invention; Figure 5 This is a flowchart illustrating another energy scheduling method based on a new energy power and distributed supercomputing fusion architecture provided in an embodiment of the present invention. Figure 6 This is a schematic diagram of a multi-scenario computing power-electricity mapping relationship provided by 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 technical challenges of high curtailment rates in renewable energy generation and high costs and instability in distributed supercomputing power supply, this embodiment proposes a fusion architecture of renewable energy power and distributed supercomputing. In one implementation, such as... Figure 1 As shown, the architecture includes: a supercomputing energy storage array, a bidirectional inverter unit, at least one power generation panel array, and at least one unidirectional inverter unit; the supercomputing energy storage array is connected to the DC terminal of the bidirectional inverter unit, the power generation panel array is connected to the DC terminal of the unidirectional inverter unit, the AC terminals of the bidirectional inverter unit and the unidirectional inverter unit are respectively connected to an AC bus, and the AC bus is connected to the power grid through a transformer.

[0025] The supercomputing energy storage array and the bidirectional inverter unit form a branch that integrates power generation, energy storage and computing. It can directly power the supercomputing unit with the power of the local power generation panel, and can also achieve bidirectional power interaction with the AC bus through the bidirectional inverter unit.

[0026] The photovoltaic panel array and the unidirectional inverter unit constitute a pure photovoltaic power generation branch. The photovoltaic panel array does not include supercomputing units, but only includes pure photovoltaic panels, which can be a single panel or an array of multiple panels connected in series or parallel, and is used solely for power generation. The supercomputing units that bear all the computing power tasks of the new energy power and distributed supercomputing fusion architecture are deployed in the supercomputing energy storage array. The photovoltaic panel array and the unidirectional inverter unit are only used to invert the DC power from the photovoltaic panels into AC power and transmit it to the AC bus as a shared power source for the supercomputing energy storage array.

[0027] It is worth noting that the new energy power generation proposed in this embodiment is photovoltaic power generation. However, in other embodiments, other types of new energy power generation, such as wind power, hydropower, tidal energy, and biomass energy, can also provide power to the supercomputing energy storage array and power generation panel array. Accordingly, the unidirectional inverter unit and bidirectional inverter unit can be configured to adapt to wind power converters, hydropower grid-connected devices, or biomass power generation grid-connected devices according to the output characteristics of different new energy power generation forms, so as to realize the unified access and scheduling of multiple types of new energy power, thereby realizing the integrated application of distributed supercomputing and new energy power in a wider range of new energy scenarios.

[0028] In one implementation, when the supercomputing energy storage array has a power surplus, the bidirectional inverter unit converts the excess DC power into AC power and supplies it to the AC bus. When the supercomputing energy storage array has a power shortage, the bidirectional inverter unit obtains AC power from the AC bus and rectifies it into DC power to supply power to the supercomputing units in the supercomputing energy storage array. The unidirectional inverter unit converts the DC power generated by the corresponding power generation panel array into AC power and supplies it to the AC bus. When all the new energy power collected by the AC bus is still insufficient to meet the power demand of the supercomputing units, the power grid replenishes the AC bus through a transformer to ensure the uninterrupted and stable operation of the supercomputing units.

[0029] When the supercomputing unit's load is low and the output power of the local power generation panel exceeds the power demand of the supercomputing unit and the micro-energy storage unit, the excess DC power is converted into AC power through the bidirectional inverter unit and fed into the AC bus for use by other loads or connected to the grid, achieving power curtailment and absorption. When the output of the local power generation panel is insufficient, the micro-energy storage unit's power is depleted, or it reaches a preset low threshold (e.g., 5% of the total power), the bidirectional inverter unit obtains AC power from the AC bus and rectifies it into DC power to supply the supercomputing unit and the micro-energy storage unit, prioritizing the use of grid-connected power from other power generation panel arrays. When all the new energy power collected on the AC bus still cannot meet the power demand of the supercomputing unit, the grid replenishes energy to the AC bus through a transformer, providing a stable AC power supply to the supercomputing unit. After rectification by the bidirectional inverter unit, the power supply continues, achieving uninterrupted operation of the supercomputing unit.

[0030] In one implementation, refer to Figure 1The supercomputing energy storage array includes at least one supercomputing energy storage module, which comprises a power generation unit, a supercomputing unit, and a micro-energy storage unit. The input terminal of the micro-energy storage unit is connected to both the DC output terminal of the power generation unit and the DC terminal of the bidirectional inverter unit, enabling it to simultaneously receive DC power from the power generation unit and supplementary power fed back from the AC bus side after rectification by the bidirectional inverter unit. The micro-energy storage unit receives and stores DC power from the power generation unit and supplementary power after rectification by the bidirectional inverter unit, and continuously outputs a stable DC power supply to the supercomputing unit to ensure uninterrupted operation of the supercomputing unit. Figure 1 The example shown uses two supercomputing energy storage modules, but other embodiments may include more supercomputing energy storage modules.

[0031] The power generation unit is a new energy power generation panel, which can be a single power generation panel or an array of power generation panels connected in parallel or in series.

[0032] In one embodiment, the supercomputing unit is a distributed computing node with a rated power consumption of approximately 300-400W. It is configured next to the power generation unit and is used to execute synchronous or asynchronous computing tasks. It is the core power load in this architecture, and its operating status directly determines the system's energy scheduling logic. The supercomputing unit receives computing task instructions, task priorities, computing power allocation parameters, and power constraints from the task distribution server. It can adaptively adjust the task execution timing and computing power output based on its own operating status, load rate, and the remaining power of the supporting micro-energy storage unit. When the micro-energy storage unit has sufficient power, the supercomputing unit operates at its rated computing power. When the micro-energy storage unit has low power, it can dynamically adjust the task load within permissible limits, prioritizing the uninterrupted core computing process. Simultaneously, it feeds back its available status and power information to the task distribution server, enabling the server to rationally schedule and avoid low-power nodes, ensuring the stable and efficient operation of the overall distributed supercomputing system.

[0033] The micro energy storage unit can be composed of lithium batteries or supercapacitors, either individually or in combination. It has the ability to charge and discharge quickly and output voltage stably. It is used to buffer the power fluctuations of the power generation unit, store surplus energy, and continuously supply power to the supercomputing unit when power is insufficient, thereby improving the continuity and stability of power supply.

[0034] To maximize the capture of photovoltaic energy, improve the photoelectric conversion efficiency of the power generation unit, and provide a stable DC input for the micro-energy storage unit, in one embodiment, referring to... Figure 1The supercomputing energy storage module also includes an MPPT unit; the input terminal of the MPPT unit is connected to the DC output terminal of the power generation unit, and the output terminal of the MPPT unit is connected to the charging terminal of the micro-energy storage unit. The MPPT unit is used to track the maximum power point of the power generation unit in real time, and dynamically adjust the operating voltage and current according to changes in light intensity and load, converting unstable new energy power output into stable DC power, thereby charging the micro-energy storage unit.

[0035] To improve the power supply stability of micro-energy storage units, extend the lifespan of energy storage devices, and achieve precise power matching among power generation, energy storage, and computing loads, in one embodiment, refer to... Figure 1 The supercomputing energy storage module also includes a charge and discharge control unit, which is connected in series between the micro energy storage unit and the supercomputing unit. The charge and discharge control unit is used to automatically control the charging, discharging and voltage regulation output of the micro energy storage unit according to the output power of the power generation unit, the remaining power of the micro energy storage unit and the load status of the supercomputing unit.

[0036] The charging and discharging control unit is used to collect the output power of the power generation unit, the remaining power of the micro energy storage unit, and the real-time load information of the supercomputing unit in real time. According to the preset strategy, it automatically switches the charging, discharging, resting and protection states of the micro energy storage unit, stably adjusts the output voltage and output current, prevents overcharging, over-discharging, overcurrent and short circuit, ensures the safe and reliable operation of the micro energy storage unit, and provides the supercomputing unit with a continuous, stable DC power supply that meets the rated operating range, thereby improving the operational stability and service life of the supercomputing unit.

[0037] In one implementation, refer to Figure 1 Taking the 800V AC bus of photovoltaic power generation as an example, the specific working method of the new energy power and distributed supercomputing fusion architecture proposed in this embodiment includes: When there is sufficient sunlight and local power surplus, the photovoltaic panels (local power generation units) generate electricity from sunlight. The MPPT unit tracks the maximum power point in real time and outputs the electrical energy to the micro-energy storage unit. The supercomputing unit prioritizes using the local micro-energy storage unit for power. When the supercomputing unit has a low load and the micro-energy storage unit is fully charged, local power becomes surplus. The bidirectional inverter unit operates in inverter mode, converting the excess DC power into 800V AC power and sending it to the AC bus. At the same time, the power from the ordinary photovoltaic strings (i.e., the photovoltaic panel array) is also sent to the AC bus through a unidirectional inverter. After the power from multiple sources is uniformly integrated into the 800V AC bus, it can be directly transformed by a transformer to the grid or used by other loads, realizing the local consumption of new energy and the grid connection of surplus electricity.

[0038] When sunlight and local power are insufficient, the output power of the photovoltaic panels decreases, failing to meet the charging needs of the corresponding micro-energy storage units. Under the control of the charge / discharge control unit, the micro-energy storage units discharge to power the supercomputing unit. When the power of the micro-energy storage units drops to a set threshold (e.g., 5% of the total power), the bidirectional inverter unit switches to rectification mode, drawing AC power from the 800V AC bus and rectifying it into DC power to supplement the DC side of the supercomputing energy storage array. At this time, the AC bus collects power from the ordinary photovoltaic strings on the right. The local micro-energy storage units preferentially use this shared power for charging, while simultaneously powering the local supercomputing unit. In this scenario, there is no immediate reliance on the grid, achieving cross-array power sharing and extending the power supply guarantee time.

[0039] However, under extreme operating conditions (no light / continuous cloudy / rainy days, insufficient output of renewable energy across the entire station), if the combined power output of all the light-emitting panel arrays on the AC bus and the power generation of the supercomputing energy storage module itself is still insufficient to meet the total load demand of the supercomputing unit, then grid power is required. The grid supplies power to the 800V AC bus in reverse through a transformer. The bidirectional inverter unit operates in rectification mode, rectifying the AC power from the grid side into DC power to continuously power the micro-energy storage unit, ensuring uninterrupted and stable operation of the supercomputing unit and preventing computing task failure due to power outages.

[0040] In one implementation, such as Figure 2 As shown, the system also includes a task distribution server, which establishes a communication connection with each supercomputing unit to receive and parse external computing tasks. Based on the real-time computing power, load, and power status of each supercomputing unit, the server issues computing tasks, task priority instructions, and power consumption constraint information to each supercomputing unit to achieve computing power scheduling and load balancing of the entire system.

[0041] When a supercomputing unit executes two or more computing tasks simultaneously, it will monitor the remaining power of the micro-energy storage unit in real time and report its status to the task distribution server. If the remaining power of the micro-energy storage unit is detected to be sufficient (i.e., the remaining power of the micro-energy storage unit is higher than the preset high power threshold, such as ≥80%, and the current output of the power generation unit can cover the current load of the supercomputing unit, while there is still surplus power), the bidirectional inverter unit can maintain partial grid connection priority and prioritize converting the local surplus photovoltaic power into AC power through the bidirectional inverter unit and sending it into the AC bus / grid, rather than storing it all or using it for local loads. Under the premise of not affecting the execution of the supercomputing unit's tasks, the unit maximizes the absorption of surplus new energy power and reduces the curtailment rate, and absorbs surplus power without affecting the tasks.

[0042] If the remaining power of the micro-energy storage unit is detected to be low (e.g., below the preset low power threshold, such as ≤30%), the supercomputing unit will proactively lower its grid connection priority, adjust its energy dispatch strategy, and prioritize supplying all the output power of the power generation unit and the power of the micro-energy storage unit to the supercomputing unit for task execution. It will suspend or reduce the power transmission to the AC bus / grid, prioritize the power resources to ensure computing load rather than grid connection and consumption, and prioritize the use of power for task execution.

[0043] When the supercomputing unit detects that the remaining power of the micro-energy storage unit is insufficient to support new computing tasks, the supercomputing unit will send an avoidance signal to the task distribution server. The task distribution server will then avoid this node in subsequent task allocations and will no longer issue new computing tasks to it, only maintaining the power supply for its current computing tasks, in order to achieve dynamic load balancing and power supply security.

[0044] In summary, the proposed architecture for the integration of renewable energy power and distributed supercomputing in this embodiment significantly reduces renewable energy curtailment by having supercomputing units consume power from corresponding photovoltaic panels nearby, while simultaneously achieving low-cost computing power deployment. By aggregating power from various photovoltaic arrays via an AC bus, supercomputing units with insufficient power can share surplus power from other arrays, overcoming the power supply bottleneck of single photovoltaic energy storage. Furthermore, when the output of renewable energy across the entire station is insufficient or power generation is unbalanced, the grid provides reverse power supplementation, ensuring uninterrupted and stable operation of the supercomputing units and resolving the contradiction between the intermittency of renewable energy power generation and the continuity of computing tasks. This embodiment significantly improves the power supply stability and task reliability of the distributed supercomputing system while increasing the utilization rate of renewable energy power and reducing computing costs, achieving efficient integration of renewable energy power and distributed computing power.

[0045] Example 2: To further illustrate the fusion architecture of new energy power and distributed supercomputing proposed in Example 1, this example proposes an energy scheduling method for the fusion architecture of new energy power and distributed supercomputing. In one implementation, such as... Figure 3 As shown, the method includes: Step 101: When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit converts the excess DC power into AC power and transmits it to the AC bus.

[0046] When external lighting conditions are good and the power generation unit outputs sufficient power, after meeting the power consumption of the supercomputing unit's normal computing load and the full charging requirements of the micro-energy storage unit, there is still surplus DC power inside the supercomputing energy storage array. At this time, the bidirectional inverter unit operates in inverter mode, converting the excess DC power on the array side into AC power that is compatible with the AC bus voltage level, and stably delivering it to the AC bus. This power can be used to supply other electrical loads in the system, or it can be connected to the grid to realize the consumption of surplus power, effectively reducing the phenomenon of new energy curtailment and improving the utilization efficiency of power generation resources.

[0047] Step 102: When the supercomputing energy storage array is short of power, the bidirectional inverter unit obtains AC power from the AC bus and rectifies it into DC power to power the supercomputing units in the supercomputing energy storage array.

[0048] When sunlight weakens, the output of new energy power generation decreases, the output power of the local power generation unit cannot meet the operating power consumption of the supercomputing unit, and the power consumption of the micro energy storage unit is consumed to the lower limit threshold and it is difficult to continue discharging to support the load, the supercomputing energy storage array is determined to be in a power shortage state. The bidirectional inverter unit switches to the rectification working mode, draws the AC power collected by all the power generation arrays from the AC bus, and rectifies it into stable DC power to supply the DC side of the supercomputing energy storage array, continuously providing power for the supercomputing unit. At the same time, it can replenish the power of the micro energy storage unit to maintain the uninterrupted operation of the computing load.

[0049] Step 103: The unidirectional inverter unit inverts the DC power generated by the corresponding power generation panel array into AC power and transmits it to the AC bus. When all the new energy power collected by the AC bus is still insufficient to meet the power demand of the supercomputing unit, the power grid replenishes the AC bus through the transformer to ensure the uninterrupted and stable operation of the supercomputing unit.

[0050] Among them, conventional power generation arrays without supporting supercomputing units continuously invert the DC power they generate into AC power through unidirectional inverter units and feed it into the AC bus, so that the AC bus can collect multiple new energy power sources to form a shared power source; when the weather is continuously severe and the output of new energy power generation in the whole region is low, and the total amount of new energy power collected by the AC bus is still insufficient to cover the overall power load of all supercomputing units, the power grid transmits AC power to the AC bus in reverse through transformers to make up for the power gap in the system. Relying on the power grid's backup power supply mechanism, the power supply to the supercomputing units is always uninterrupted, ensuring the stable and continuous execution of distributed computing tasks.

[0051] For the specific structure of the new energy power and distributed supercomputing fusion architecture, please refer to Embodiment 3, which will not be repeated in this embodiment.

[0052] In summary, this embodiment enables the supercomputing energy storage array to connect to the grid when there is a power surplus through a bidirectional inverter unit, and to draw power from the grid for reverse energy replenishment when there is a power shortage, thus breaking through the power supply bottleneck of a single power generation board. At the same time, the multi-source power sharing mechanism can utilize the output of other power generation board arrays or grid power to continuously supply power when the local power supply of the supercomputing unit is insufficient or the power generation is unbalanced, ensuring the uninterrupted and stable operation of the supercomputing unit, thereby improving the reliability of computing tasks.

[0053] Example 3: The new energy power and distributed supercomputing fusion architecture proposed in Example 1 requires a separate unidirectional inverter unit for each power generation array to achieve independent grid connection of power from each array. However, this approach increases the number of devices and system cost, and also increases topology complexity. To eliminate the need for unidirectional inverter units, this example proposes another new energy power and distributed supercomputing fusion architecture. In one implementation, such as... Figure 4 As shown, it includes: a supercomputing energy storage array, a bidirectional inverter unit, and at least one power generation panel array; the supercomputing energy storage array and the at least one power generation panel array are respectively connected to a DC bus, the DC end of the bidirectional inverter unit is connected to the DC bus, and the AC end of the bidirectional inverter unit is connected to the power grid through a transformer.

[0054] In this architecture, the DC outputs of the supercomputing energy storage array and each power generation array are directly fed into the same DC bus, and the AC-DC conversion and grid connection are uniformly realized by a bidirectional inverter unit. The DC power of each power generation array and the surplus DC power of the supercomputing energy storage array are all collected into the DC bus, eliminating the need for a separate unidirectional inverter unit; the bidirectional inverter unit uniformly inverts the power on the DC bus into AC power, which is then connected to the grid via a transformer, realizing the unified grid connection and consumption of new energy power throughout the system.

[0055] In one implementation, when the supercomputing energy storage array has a power surplus, the bidirectional inverter unit is used to invert the excess DC power on the DC bus into AC power and transmit it to the power grid through a transformer.

[0056] When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit is used to invert the excess DC power on the DC bus into AC power and transmit it to the grid through a transformer. At this time, the supercomputing unit has a low load, the micro-energy storage unit is fully charged, and the local power generation exceeds the computing load and energy storage charging requirements. The surplus DC power is collected on the DC bus, and the bidirectional inverter unit works in inverter mode to convert the high-voltage DC power into AC power that is compatible with the grid. After being stepped up / stepped down by the transformer, it is connected to the grid, realizing the unified consumption of surplus new energy power and avoiding power curtailment.

[0057] In one implementation, when the supercomputing energy storage array is short of power, the bidirectional inverter unit is used to obtain AC power from the power grid through a transformer and rectify it into DC power to power the supercomputing units in the supercomputing energy storage array.

[0058] When the supercomputing energy storage array is short of power, the bidirectional inverter unit is used to obtain AC power from the grid through a transformer and rectify it into DC power to power the supercomputing units in the supercomputing energy storage array. At this time, if the local power generation cannot meet the power demand of the supercomputing units due to reduced sunlight or a sudden increase in load, and the power consumption of the micro-energy storage unit reaches the lower limit threshold, the bidirectional inverter unit switches to rectification mode, draws AC power from the grid side and converts it into stable DC power, injects it into the DC bus, provides continuous power supply to the supercomputing units, and ensures that the computing tasks are not interrupted.

[0059] In one embodiment, the power generation array is used to output the generated DC power to the DC bus to provide shared power support for the supercomputing energy storage array.

[0060] The power generation array is used to output the generated DC power to the DC bus to provide shared power support for the supercomputing energy storage array. Each conventional power generation array does not have an independent inverter unit. Its DC output is directly connected to the DC bus, forming a unified power pool with the DC side of the supercomputing energy storage array. When the local power of the supercomputing energy storage array is insufficient, it can directly obtain the surplus power of other power generation arrays from the DC bus, realizing cross-array power sharing. No additional grid-connected equipment is required, which greatly simplifies the system structure and reduces hardware costs.

[0061] In the centralized architecture of this embodiment, when the supercomputing energy storage array experiences local power shortage, the system adopts a two-level priority strategy: first, DC bus sharing, and then backup power from the power grid. The specific decision-making and execution process is as follows: First, the system will determine if there is insufficient local power. When the output power of the supercomputing energy storage module's power generation unit cannot simultaneously meet the real-time load of the supercomputing unit and the power replenishment needs of the micro energy storage unit, and the power of the micro energy storage unit has dropped to a preset low threshold (e.g., ≤20%), the system determines that there is insufficient local power and initiates the energy dispatch process.

[0062] Next, the system prioritizes drawing shared power from other power generation arrays on the DC bus. The system monitors the voltage and power of the DC bus in real time. If the DC bus voltage is detected to be higher than a set threshold, it indicates that other power generation arrays without supercomputing units are generating power normally and outputting surplus power to the DC bus. In this case, the DC power from the DC bus is directly allocated to the supercomputing units with insufficient power, without needing the rectification process of the bidirectional inverter unit, achieving efficient and low-cost cross-array power sharing.

[0063] When there is no surplus power on the DC bus, the system automatically switches to grid backup power mode. If, due to extreme conditions such as continuous rainy days, the total output of all generator arrays is insufficient, causing the voltage on the DC bus to continuously drop to a low threshold, the system determines that there is a shortage of renewable energy power across the entire region. At this time, the bidirectional inverter unit immediately switches to rectification mode, obtains AC power from the grid through a transformer, and rectifies it into stable DC power, injecting it into the DC bus to make up for the system's power shortage, providing uninterrupted power supply to all supercomputing units and ensuring the stable execution of computing tasks.

[0064] In one implementation, refer to Figure 4 In this embodiment, the supercomputing energy storage array includes at least one supercomputing energy storage module. The supercomputing energy storage module includes a power generation unit, a supercomputing unit, and a micro energy storage unit. The input terminal of the micro energy storage unit is connected to the DC output terminal and the DC bus of the power generation unit, respectively. The micro energy storage unit is used to receive and store DC power from the power generation unit, as well as supplementary power after rectification by the bidirectional inverter unit, and continuously outputs stable DC power to the supercomputing unit to ensure uninterrupted operation of the supercomputing unit. Figure 4 The example shown uses two supercomputing energy storage modules, but other embodiments may include more supercomputing energy storage modules.

[0065] The structure and working principle similar to those in Example 1 will not be described again in this example.

[0066] In order to obtain power shared by other power generation arrays from the DC bus and adapt the high-voltage DC bus to a low-voltage power supply that can be safely used by the supercomputing unit, in one embodiment, refer to Figure 4 The supercomputing energy storage module also includes a step-down unit. The input of the step-down unit is connected to the DC bus, and the output of the step-down unit is connected to the power supply of the supercomputing unit. When the load on the supercomputing unit increases and the remaining power of the corresponding micro-energy storage unit is insufficient, the step-down unit is used to step down the DC power generated by the power generation array before transmitting it to the supercomputing unit for supplementary power supply, thereby ensuring the stable operation and power safety of the supercomputing unit in multi-power supply mode. Through the step-down unit, both the proximity access for cross-array power sharing is realized, and the stable operation and power safety of the supercomputing unit in multi-power supply mode are ensured, avoiding damage to the computing hardware caused by direct high-voltage power supply.

[0067] In summary, this embodiment enables the supercomputing energy storage array to connect to the grid when there is a power surplus through a bidirectional inverter unit, and to draw power from the grid for reverse energy replenishment when there is a power shortage, thus breaking through the power supply bottleneck of a single power generation board. At the same time, the multi-source power sharing mechanism can utilize the output of other power generation board arrays or grid power to continuously supply power when the local power supply of the supercomputing unit is insufficient or the power generation is unbalanced, ensuring the uninterrupted and stable operation of the supercomputing unit, thereby improving the reliability of computing tasks.

[0068] Example 4: To further illustrate the fusion architecture of new energy power and distributed supercomputing proposed in Example 3, this example proposes an energy scheduling method for the fusion architecture of new energy power and distributed supercomputing. In one implementation, such as... Figure 5 As shown, the method includes: Step 201: When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit inverts the excess DC power on the DC bus into AC power and transmits it to the power grid through a transformer.

[0069] When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit converts the excess DC power on the DC bus into AC power and transmits it to the grid through a transformer. At this time, the supercomputing unit has a low load, the micro-energy storage unit is fully charged, and the local power generation exceeds the computing load and energy storage charging requirements. The surplus DC power is collected on the DC bus, and the bidirectional inverter unit operates in inverter mode, converting the high-voltage DC power into AC power that is compatible with the grid. After being stepped up / stepped down by a transformer, it is connected to the grid, realizing the unified consumption of surplus new energy power and avoiding power curtailment.

[0070] Step 202: When the supercomputing energy storage array is short of power, the bidirectional inverter unit obtains AC power from the grid through a transformer and rectifies it into DC power to power the supercomputing units in the supercomputing energy storage array.

[0071] When the supercomputing energy storage array is short of power, the bidirectional inverter unit obtains AC power from the grid through a transformer and rectifies it into DC power to power the supercomputing units in the supercomputing energy storage array. At this time, if the local power generation cannot meet the power demand of the supercomputing units due to reduced sunlight or a sudden increase in load, and the power consumption of the micro-energy storage unit reaches the lower limit threshold, the bidirectional inverter unit switches to rectification mode, draws AC power from the grid side and converts it into stable DC power, injects it into the DC bus, provides continuous power supply to the supercomputing units, and ensures that the computing tasks are not interrupted.

[0072] Step 203: The power generation array outputs the generated DC power to the DC bus to provide shared power support for the supercomputing energy storage array.

[0073] The power generation array outputs the generated DC power to the DC bus to provide shared power support for the supercomputing energy storage array. Each conventional power generation array does not have an independent inverter unit; its DC output is directly connected to the DC bus, forming a unified power pool with the DC side of the supercomputing energy storage array. When the local power of the supercomputing energy storage array is insufficient, it can directly obtain surplus power from other power generation arrays on the DC bus, realizing cross-array power sharing without the need for additional grid-connected equipment, which greatly simplifies the system structure and reduces hardware costs.

[0074] In summary, this embodiment enables the supercomputing energy storage array to connect to the grid when there is a power surplus through a bidirectional inverter unit, and to draw power from the grid for reverse energy replenishment when there is a power shortage, thus breaking through the power supply bottleneck of a single power generation board. At the same time, the multi-source power sharing mechanism can utilize the output of other power generation board arrays or grid power to continuously supply power when the local power supply of the supercomputing unit is insufficient or the power generation is unbalanced, ensuring the uninterrupted and stable operation of the supercomputing unit, thereby improving the reliability of computing tasks.

[0075] For the specific structure of the new energy power and distributed supercomputing fusion architecture, please refer to Embodiment 3, which will not be repeated in this embodiment.

[0076] Example 5: To address the technical issues of existing new energy and distributed supercomputing fusion architectures, such as the inability to provide power in one direction, respond to grid peak shaving and frequency regulation, fail to maximize revenue by combining with the real-time electricity market, and lack of coordinated scheduling of power and computing power, this embodiment, based on the architectures of Embodiments 1 and 3, introduces a grid-type bidirectional inverter unit and a real-time electricity market information interaction mechanism to form a coordinated power and computing power allocation system that can dynamically switch between grid support, electricity market revenue, and supercomputing power guarantee.

[0077] This embodiment uses a grid-type bidirectional inverter unit as the core control unit to achieve bidirectional dynamic allocation of power and computing power based on grid stability, real-time electricity price, and computing power revenue. Specifically, it includes grid priority scheduling strategy and local computing power guarantee strategy.

[0078] Grid priority dispatch strategy: When the grid-type bidirectional inverter unit detects any of the following conditions through the real-time electricity market interface unit, it will automatically trigger grid priority dispatch: the grid voltage or frequency exceeds the preset stability threshold; the grid issues peak shaving, frequency regulation or support instructions; the real-time grid electricity price is higher than the local computing power operation revenue threshold.

[0079] At this time, the grid-type bidirectional inverter unit sends a signal to the local computing node to suspend or reduce computing power. After receiving the signal, the local computing node automatically saves the current task site data to ensure that the task can be recovered. At the same time, the grid-type bidirectional inverter unit cuts off or reduces the power supply to the local computing node, integrates all available energy of photovoltaic modules and micro energy storage devices, outputs power to the grid in grid mode, and actively adjusts voltage and frequency to assist the grid in stable operation and participate in electricity market services to obtain higher returns.

[0080] Local computing power guarantee strategy: When the power grid is stable, the real-time electricity price is lower than the local computing power revenue, or the local computing node is executing a high-priority task, the system switches to computing power priority mode. In this mode, the grid-type bidirectional inverter unit prioritizes supplying photovoltaic and micro-storage power to the local computing node. If the photovoltaic output is insufficient or the micro-storage power is low, the grid-type bidirectional inverter unit can draw power from the grid in reverse, rectify it, and continuously supply power to the local computing node, ensuring uninterrupted computing tasks. Simultaneously, in winter or low-temperature scenarios, reverse power can also be used for photovoltaic module insulation and snow / ice removal, enabling the photovoltaic system to quickly resume power generation after the weather improves, thus improving the utilization rate of new energy sources.

[0081] During supercomputing power scheduling, the tasks carried by local computing nodes exhibit diverse characteristics in terms of computing power, energy consumption, and operating modes. The power requirements, node deployment numbers, and synchronous / asynchronous computing modes vary significantly across different computing scenarios, directly impacting power dispatch priority, load reduction, and grid connection benefit strategies. This system targets mainstream supercomputing task scenarios, categorizing them into high, medium, and low power consumption levels. It also adapts to both synchronous and asynchronous computing operating modes for each scenario, achieving refined collaborative scheduling of power and computing power. Synchronous computing is a real-time computing mode where task instructions are executed serially, nodes are occupied throughout the process, and there is no interruption. Asynchronous computing is a non-real-time computing mode where tasks can be fragmented and parallelized, support breakpoint pauses, and nodes can be dynamically released. Both modes can adapt to different grid dispatch requirements.

[0082] To clearly define the power consumption parameters, node configurations, and operational characteristics of each computing scenario, and to support precise system scheduling, such as... Figure 6 As shown, this embodiment constructs a multi-scenario computing power-power mapping table, covering core computing power tasks such as large model training, large model inference, industrial simulation, scientific computing, and general data processing, and clarifies the energy consumption level, computing mode, single task power consumption, required number of nodes, and scheduling adaptation characteristics of each scenario.

[0083] based on Figure 6The multi-scenario computing power-power mapping relationship shown in the diagram enables differentiated and refined scheduling: for high-energy-consuming, synchronous core computing tasks, priority is given to ensuring power supply stability, adapting to grid demand only by reducing redundant power consumption; for medium- and low-energy-consuming, asynchronous general-purpose tasks, flexible start-up and shutdown, and load reduction can be implemented to quickly release power resources, balancing the reliability of supercomputing tasks with grid peak shaving, frequency regulation, and electricity market revenue requirements. Simultaneously, the task distribution server can dynamically calculate the total local computing power consumption and revenue threshold based on the real-time task type, computing mode, and number of nodes, accurately determining the switching timing between grid-priority scheduling and local computing power guarantee mode, thus solving the problems of traditional scheduling schemes being "one-size-fits-all," having poor energy consumption adaptability, and low revenue utilization.

[0084] In one embodiment, in the AC bus architecture described in Example 1, the original bidirectional inverter unit is replaced with a grid-type bidirectional inverter unit. The grid-type bidirectional inverter unit possesses real-time grid status sensing, real-time electricity price acquisition, bidirectional energy flow switching, and computing power scheduling control capabilities. The grid-type bidirectional inverter unit is communicatively connected to both the AC bus and the task distribution server.

[0085] When the real-time electricity market price exceeds the local supercomputing unit's operating revenue threshold, or when grid voltage or frequency anomalies require peak shaving and frequency regulation support, the grid-type bidirectional inverter unit sends a computing power reduction command to the task distribution server. Upon receiving the command, the task distribution server controls the supercomputing unit to save the task status and reduce the computing power load or suspend non-core tasks. At this time, the grid-type bidirectional inverter unit prioritizes inverting and transmitting surplus power from the supercomputing energy storage array and the power generation panel array to the AC bus, and further feeds it into the grid to support grid stability and obtain higher electricity market revenue.

[0086] When the power grid stabilizes and the real-time electricity price falls below the local supercomputing revenue threshold, the system reverts to a local computing power priority mode. The grid-type bidirectional inverter unit prioritizes power supply to the supercomputing unit via the shared power supply of the power generation panel array and AC bus; if the power is still insufficient, the power grid will provide reverse power through transformers to ensure uninterrupted operation of the supercomputing unit.

[0087] In one embodiment, in the DC bus architecture described in Example 3, the bidirectional inverter unit also adopts a grid-type bidirectional inverter unit, which has the functions of grid support, real-time electricity price judgment, and coordinated scheduling of power and computing power.

[0088] The DC end of the grid-type bidirectional inverter unit is connected to the DC bus, and the AC end is connected to the power grid via a transformer, establishing communication with the task distribution server. When it detects that the power grid needs support or the real-time electricity price is higher than the local computing power revenue, the grid-type bidirectional inverter unit sends a computing power adjustment command to the task distribution server. The task distribution server controls the supercomputing unit to reduce its load, save the task status, and reduce local power consumption. At this time, all photovoltaic power collected on the DC bus is preferentially inverted by the grid-type bidirectional inverter unit and then transmitted to the power grid, maximizing both grid support and electricity market revenue.

[0089] When the local supercomputing unit is short of power, the system prioritizes power supply to the DC bus first, followed by power from the grid. If there is surplus power from other power generation arrays on the DC bus, the supercomputing unit will be supplied with power through the step-down unit first. If the DC bus power is still insufficient, the grid-type bidirectional inverter unit will draw power from the grid and rectify it into DC power to inject into the DC bus, ensuring the stable operation of the supercomputing unit.

[0090] In summary, this embodiment achieves bidirectional dynamic allocation of renewable energy power and distributed supercomputing by integrating AC and DC bus architectures, combined with grid-type bidirectional inverter units, real-time power market scheduling, and task distribution server collaborative control. Simultaneously, by subdividing computing scenarios across multiple energy consumption levels and adapting to synchronous / asynchronous dual computing modes, a standardized computing power-power mapping scheduling system is established. This enables differentiated and refined power scheduling for computing tasks. The system can flexibly switch energy flow directions based on grid status, real-time electricity prices, and supercomputing power demand, prioritizing the consumption of local renewable energy power. When power is insufficient, it utilizes array sharing and provides tiered power supplementation from the grid. When there is a power surplus or the grid needs support, it actively connects to the grid. This ensures uninterrupted and stable operation of the supercomputing unit while maximizing renewable energy utilization and power market revenue, effectively solving the technical challenges of balancing the intermittency of renewable energy, the reliability of computing power supply, and the need for grid support.

[0091] 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 fusion architecture of new energy power and distributed supercomputing, characterized in that, include: The supercomputing energy storage array, bidirectional inverter unit, at least one power generation panel array, and at least one unidirectional inverter unit; The supercomputing energy storage array is connected to the DC terminal of the bidirectional inverter unit, the power generation panel array is connected to the DC terminal of the unidirectional inverter unit, the AC terminals of the bidirectional inverter unit and the unidirectional inverter unit are respectively connected to the AC bus, and the AC bus is connected to the power grid through a transformer. When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit is used to invert the excess DC power into AC power and transmit it to the AC bus; when the supercomputing energy storage array has a power shortage, the bidirectional inverter unit is used to obtain AC power from the AC bus and rectify it into DC power to power the supercomputing units in the supercomputing energy storage array. The unidirectional inverter unit is used to invert the DC power generated by the corresponding power generation panel array into AC power and transmit it to the AC bus. When all the new energy power collected by the AC bus is still insufficient to meet the power demand of the supercomputing unit, the power grid replenishes the AC bus through the transformer to ensure the uninterrupted and stable operation of the supercomputing unit.

2. The new energy power and distributed supercomputing fusion architecture according to claim 1, characterized in that, The supercomputing energy storage array includes at least one supercomputing energy storage module, and the supercomputing energy storage module includes a power generation unit, a supercomputing unit, and a micro energy storage unit. The input terminal of the micro energy storage unit is connected to the DC output terminal of the power generation unit and the DC terminal of the bidirectional inverter unit, respectively. The micro energy storage unit is used to receive and store DC power from the power generation unit, as well as supplementary power after rectification by the bidirectional inverter unit, and continuously output stable DC power to the supercomputing unit to ensure uninterrupted operation of the supercomputing unit.

3. The new energy power and distributed supercomputing fusion architecture according to claim 2, characterized in that, The supercomputing energy storage module also includes a charge and discharge control unit, which is connected in series between the micro energy storage unit and the supercomputing unit; The charging and discharging control unit is used to automatically control the charging, discharging, and voltage regulation output of the micro-energy storage unit based on the output power of the power generation unit, the remaining power of the micro-energy storage unit, and the load status of the supercomputing unit.

4. The new energy power and distributed supercomputing fusion architecture according to claim 2, characterized in that, The supercomputing energy storage module also includes an MPPT unit; the input terminal of the MPPT unit is connected to the DC output terminal of the power generation unit, and the output terminal of the MPPT unit is connected to the charging terminal of the micro energy storage unit. The MPPT unit is used to track the maximum power point of the power generation unit in real time, and dynamically adjust the working voltage and current according to the changes in light intensity and load, so as to convert the unstable new energy power output into stable DC power, thereby charging the micro energy storage unit.

5. An energy scheduling method for a new energy power and distributed supercomputing integrated architecture, characterized in that, Applied to the new energy power and distributed supercomputing fusion architecture as described in any one of claims 1-4, the method includes: When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit will invert the excess DC power into AC power and transmit it to the AC bus. When the supercomputing energy storage array is short of power, the bidirectional inverter unit obtains AC power from the AC bus and rectifies it into DC power to power the supercomputing units in the supercomputing energy storage array. The unidirectional inverter unit inverts the DC power generated by the corresponding power generation panel array into AC power and transmits it to the AC bus. When all the new energy power collected by the AC bus is still insufficient to meet the power demand of the supercomputing unit, the power grid replenishes the AC bus through the transformer to ensure the uninterrupted and stable operation of the supercomputing unit.

6. A fusion architecture of new energy power and distributed supercomputing, characterized in that, include: Supercomputing energy storage array, bidirectional inverter unit, and at least one power generation panel array; The supercomputing energy storage array and the at least one power generation panel array are respectively connected to the DC bus. The DC terminal of the bidirectional inverter unit is connected to the DC bus, and the AC terminal of the bidirectional inverter unit is connected to the power grid through a transformer. When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit is used to invert the excess DC power on the DC bus into AC power and transmit it to the power grid through a transformer; When the supercomputing energy storage array is short of power, the bidirectional inverter unit is used to obtain AC power from the grid through a transformer and rectify it into DC power to power the supercomputing units in the supercomputing energy storage array. The power generation array is used to output the generated DC power to the DC bus to provide shared power support for the supercomputing energy storage array.

7. The new energy power and distributed supercomputing fusion architecture according to claim 6, characterized in that, The supercomputing energy storage array includes at least one supercomputing energy storage module, and the supercomputing energy storage module includes a power generation unit, a supercomputing unit, and a micro energy storage unit. The input terminal of the micro energy storage unit is connected to the DC output terminal and the DC bus of the power generation unit, respectively. The micro energy storage unit is used to receive and store DC power from the power generation unit, as well as supplementary power after rectification by the bidirectional inverter unit, and continuously output stable DC power to the supercomputing unit to ensure uninterrupted operation of the supercomputing unit.

8. The new energy power and distributed supercomputing fusion architecture according to claim 6, characterized in that, The supercomputing energy storage module also includes a charge and discharge control unit, which is connected in series between the micro energy storage unit and the supercomputing unit; The charging and discharging control unit is used to automatically control the charging, discharging, and voltage regulation output of the micro-energy storage unit based on the output power of the power generation unit, the remaining power of the micro-energy storage unit, and the load status of the supercomputing unit.

9. The new energy power and distributed supercomputing fusion architecture according to claim 6, characterized in that, The supercomputing energy storage module also includes a step-down unit, the input of which is connected to the DC bus, and the output of which is connected to the power supply of the supercomputing unit. When the load on the supercomputing unit increases and the remaining power of the corresponding micro-energy storage unit is insufficient, the step-down unit is used to step down the DC power generated by the power generation array and then transmit it to the supercomputing unit for supplementary power supply, so as to ensure the stable operation and power safety of the supercomputing unit in the multi-power supply mode.

10. An energy scheduling method for a new energy power and distributed supercomputing integrated architecture, characterized in that, Applied to the new energy power and distributed supercomputing fusion architecture as described in any one of claims 6-9, the method includes: When the supercomputing energy storage array has a power surplus, the bidirectional inverter unit inverts the excess DC power on the DC bus into AC power and transmits it to the power grid through a transformer; When the supercomputing energy storage array is short of power, the bidirectional inverter unit obtains AC power from the grid through a transformer and rectifies it into DC power to power the supercomputing units in the supercomputing energy storage array. At the same time, the power generation array outputs the generated DC power to the DC bus to provide shared power support for the supercomputing energy storage array.

Images