System with a data center, as well as a network with the system and at least one data center

Abstract

The invention relates to a system (1100) comprising a local data center (200), a local storage device (300), and at least one local power generation plant (100). The data center (200), the storage device (300), and the power generation plant (100) form an island grid. The data center (200) is designed as an electrical consumer such that, during continuous operation, it requires a maximum of 100% of the average annual electrical energy generated by the power generation plant (100). The storage device (300) has an electrical power output of at least that of the data center and a capacity sufficient to supply the data center (200) with electrical power locally for at least 6 hours, particularly under full load, and / or at least a capacity sufficient to reliably bridge a maximum foreseeable interruption of the local power supply by the power generation plant (100). The invention further relates to a network (2000) of decentrally distributed data centers (210) which are coupled to at least one system (1100) according to one of the preceding claims.

Description

State of the art

[0001] The invention relates to a system with a data center. Furthermore, the invention relates to a network comprising the system and at least one data center.

[0002] It is known that data centers generally require an operational power supply from the general supply grid, that this power supply is generally only implemented ecologically, if at all, as an energy balance or only partially locally with renewable energies, and that, for example, considerable power losses occur when power is transported over long distances.

[0003] In contrast to smaller IT installations, the IT connection power of data centers for their computer components is more than 40 kW, and in particular more than 1 MW.

[0004] It's also well known that data centers often require short response times in the event of disruptions. For example, in the event of a power outage, emergency power is typically provided by diesel generators, which require additional fuel for extended operation. This, in turn, requires the availability of fuel and regular personnel deployment at the data center site, even in crisis or disaster situations.

[0005] Furthermore, it is known that due to the classic, so-called IT density in the data center, each rack with computer components of typically 3-8 kW requires a significant amount of space for the data center and a considerable amount of resources for the computer components and auxiliary components.

[0006] The effective utilization of the computer components is often limited. To cool the computer components, an air-cooled or indirectly water-cooled system consisting of cooling units, side cooling devices, and on-board fans is typically used. These, along with other electrical devices in the data center, typically require between 25% and 100% additional power, depending on the season and the data center's utilization, in addition to the computer components themselves, especially their operational computer components without on-board fans. Disclosure of the invention

[0007] One object of the invention is to create a system with a data center that can be operated autonomously.

[0008] A further object of the invention is to create a network with such a system with improved efficiency.

[0009] The objects are achieved by the features of the independent claims. Advantageous embodiments and advantages of the invention emerge from the further claims, the description, and the drawings.

[0010] A system is proposed comprising at least one local data center, at least one local storage device, and at least one local energy generation plant. The at least one local data center, the at least one local storage device, and the at least one local energy generation plant form an island grid, in particular a closed island grid. The at least one data center is designed such that, as an electrical consumer in continuous operation, it requires a maximum of 100% of the annual average electrical energy generated by the at least one energy generation plant.

[0011] The at least one storage device has an electrical power of at least the electrical data center power and a capacity that is sufficient to supply the at least one data center, in particular under full load, with electrical power for at least 6 hours, preferably at least 12 hours, particularly preferably at least 24 hours, very particularly preferably at least 48 hours, most preferably at least 96 hours, and / or at least a capacity to safely bridge the largest foreseeable interruption in the local power supply by the at least one energy generation plant locally.

[0012] The largest foreseeable interruption can occur, for example, during a so-called dark period, during maintenance, or similar. If the at least one storage device can bridge this, even with a potential reduction in storage quality over the life cycle of the at least one storage device, the local supply of the at least one data center can be guaranteed depending on the specific site conditions. A very large storage device with a power and capacity in the classic power plant size of, for example, 500 MW with 500 MWh or even 1 GW with 1 GWh and more can be advantageous. Such a scale can also offer added value, among other things.to ensure general security of supply at the level of distribution power grids and in particular at the level of transmission power grids, as well as generally for the energy industry via so-called grid services and also to reduce electricity price fluctuations.

[0013] In particular, a local supply of the system with at least one data center and possibly other local power consumers is thus permanently secured, and these together become locally self-sufficient and, due to the local computing capability in at least one data center, potentially also locally autonomous.

[0014] Due to the locally self-sufficient operation, at least one data center can thus perform its function essentially unhindered even in crisis or disaster situations with potentially long-term or permanent failure of the general power supply, the general transport and communication routes and / or lack of personnel availability as well as possibly other individual or associated impairments.

[0015] This allows the creation of a system with at least one data center that, through the local coupling of decentralized energy and digital economy, can achieve the highest possible overall efficiency for energy and data, thus significantly reducing current losses and inefficiencies and contributing to security of supply and cost reduction in the energy transition.

[0016] It is particularly advantageous if the at least one energy generation plant comprises one or more renewable energy sources or even exclusively renewable energy sources, which in turn have a natural, possibly dynamically predictable, possibly seasonally varying size energy storage facility, extended over preceding solar radiation, weather and / or inflow of water, from which the system can permanently obtain sufficient energy in a predictable manner and within the framework of regular natural fluctuation ranges.

[0017] If at least one local energy generation plant is renewable, the at least one local data center can be completely supplied by the locally generated and locally buffered renewable energy, so-called “24 / 7 physically locally green”.

[0018] Renewable energy generation plants can utilize renewable energy sources such as wind energy, solar energy, hydropower (including, for example, run-of-river energy, hydroelectric storage, pumped hydro storage, tidal energy, and wave energy), or even geothermal energy, among others. A wind farm or a solar power plant, in particular, can consist of several individual plants.

[0019] The energy generation plant can, in particular, comprise at least one renewable energy source that provides energy in an intermittent manner, in particular at least one renewable energy source in the form of a wind turbine and / or solar energy system. Wind energy and / or solar energy and / or hydropower and / or geothermal energy can advantageously be used. Problems caused by inconsistent energy source availability can be avoided by dimensioning the at least one data center and the at least one storage device and / or by dynamically controlling at least part of the computing power in the data center.

[0020] Combinations of different energy generation plants or different types of energy sources are also advantageous, especially a combination of different renewable energy sources. For example, combinations of wind energy, solar energy, energy generation from biogas, wood pellets, geothermal energy, and / or energy generation from a local hydroelectric power plant are particularly advantageous because their respective characteristics allow them to balance out the natural fluctuations of individual generation types, thus requiring less local storage capacity or allowing for smaller dimensions.

[0021] The design, in particular of the connected load of the at least one data center, can expediently be based on the power provided by the at least one energy generation plant. For a run-of-river power plant that is generally capable of operating at almost base load and, for example, delivers 4 MW as a nearly permanent peak power, this power can be available for the at least one data center as a dimensioning parameter and during operation.

[0022] In a wind and / or solar farm, only a portion of the peak power is generally available on average, as these energy sources can only deliver energy intermittently. If the peak power is 10 MW, approximately 1 MW can generally be available to at least one data center, depending on the technology and location of the power generation plant. If the peak power is 25 MW, approximately 2.5 MW of connected capacity can be available to at least one data center.

[0023] The power generation plants can be spatially distributed within a limited area, but are part of the local island grid. At least one power generation plant can optionally be dedicated to the data center.

[0024] “Local” means that the at least one data center and / or the at least one storage device in the system are coupled to the at least one energy generation plant without an external supply network having to be present or available on the system side.

[0025] In an isolated grid without an external supply grid, the at least one data center and the at least one storage device can be supplied with energy directly from the at least one local power generation plant. If an external supply grid is available, the at least one storage device can also be used to stabilize an external supply grid and can thus also absorb energy from the general supply grid.

[0026] The term "local" is therefore intended to refer to the island grid. The at least one energy generation facility, or combinations of several energy sources, can certainly have a spatial extent of, for example, 10-20 km, particularly in the case of renewable energy sources such as locally distributed wind turbines and / or locally distributed solar systems and / or several connected hydroelectric dams, and even considerably more in the case of offshore wind farms with long coastal and even inland connections.

[0027] A locally advantageous realization of the at least one data center and the at least one storage device can be, in particular jointly, either in the vicinity of a local substation, if present, and / or in the vicinity of a local waste heat re-user, if present, who obtains the "green" waste heat from the at least one data center.

[0028] Accordingly, a system with at least one self-sufficient and, if appropriate, autonomous data center can be created, which can be operated in continuous operation by means of local, advantageously renewable, in particular permanently locally renewable energy supply by the at least one energy generation plant and by means of the at least one local storage device automatically and sufficiently recharged therefrom.

[0029] In addition to the possible temporary direct supply of the at least one data center by the at least one energy generation plant, the operational residual supply in the event of foreseeable potential interruptions in the availability or limited partial power provision of the at least one energy generation plant, for example in the event of a dark period in a wind-solar hybrid park, can be provided via the at least one storage device.

[0030] This system according to the invention not only provides a partial residual operational supply for at least one data center, but also allows the requirements of staggered, permanent emergency power provision placed on data centers in so-called availability classes 2 to 4, for example, in the event of a permanent power failure of the general supply grid, in this case even lasting for months or years, to be permanently met by the locally automatically recharged at least one local storage device. Furthermore, this also makes it possible to avoid the need for fossil-fuel-powered emergency power generators that require regular maintenance and, if necessary, refilling.

[0031] The at least one storage device can optionally also comprise a combination of storage elements. For example, in a first expansion stage of the system, a storage device can be implemented in the form of a so-called megabattery; a second expansion stage can, for example, utilize a hydrogen chain typically consisting of an electrolyzer, storage, fuel cell, and / or gas turbine, and / or a hydroelectric power plant with a water reservoir.

[0032] Technical specifications for the four availability classes of data centers include the standards EN 50600 from 2012 and ISO / IEC 22237 from 2021. In availability class 2, for example, availability is increased through, among other things, redundant components. Maintenance-related operational interruptions can be planned. Technical malfunctions can lead to unforeseen outages. Furthermore, in availability class 3, the data center can be serviced without maintenance-related operational interruptions. Even in the event of fault-related malfunctions of the IT equipment or technical failures, operations can be maintained in most cases. In the highest availability class 4, operations can be maintained in all cases through appropriate preventive measures, even in the event of faults and technical failures, so that operational interruptions are virtually impossible.

[0033] The dimensioning of at least one storage facility can be adapted to the specific location and also dynamically control or make controllable capacity in the locally maintained emergency power reserve, e.g., depending on the size of power generation plants and data centers, as well as potentially other local consumers, i.e., expected generation and load curves of the system, as well as a potential local prioritization of the data center supply over other local consumers during normal or abnormal operation, e.g., a blackout. Local constraints, such as operational maintenance times and, if applicable, mandatory shutdown times, can also be considered and bridged.Through an intelligent local combination of various renewable energy generation plants and storage facilities in the same island grid, the required overall dimensioning of at least one storage facility can be optimized relative to the possible data center capacity of at least one data center, e.g., a system consisting of a wind-solar hybrid park with both a mega-battery and a local hydrogen chain, as well as geothermal energy, even across multiple system expansion stages.

[0034] In addition to being self-sufficient, the system can also be operated autonomously if the basic control of data centers, storage facilities, and power generation plants takes place directly in at least one local data center. Nevertheless, a higher-level, possibly more complex, and better-adapted control system can of course also be used if communication facilities are available.

[0035] Even in the case of existing energy generation plants, retrofitting with at least one data center and at least one storage device as an upgrade to a system according to the invention has the beneficial effect that the previously existing energy generation plants can thus also become self-sufficient and autonomous.

[0036] For example, the self-supply requirements of a local wind-solar hybrid farm, such as rotor heating required for the wind turbines in winter, can now be met locally from buffered, locally renewable energy via the at least one storage device of such a system, even during availability interruptions such as periods of low light or maintenance periods. This allows normal operation even in the event of a permanent power grid failure, despite the generally lack of communication capability after several hours or days, and despite the otherwise generally unavailable black start capability of a farm's grid. The local island grid remains permanently available. Furthermore, it can advantageously be equipped with a black start capability for further security.Furthermore, local overproduction of existing or new, particularly renewable energy generation plants can be buffered locally with such an upgrade, instead of having to shut down the plants, for example, in the event of negative market prices, which will also be compensated less and less as a loss of earnings in the future.

[0037] Optionally, the at least one data center of the system can also have the standard redundant general utility grid connections, depending on the selected availability class of the at least one data center, which can be used to further protect the self-sufficient, autonomous at least one data center of the system against abnormal operating conditions. A prepared storage area with a connection for temporary emergency power generators, if required, can also further increase resilience and make the operation of the system, especially the data center within it, more flexible.

[0038] According to a favorable embodiment of the system, at least one local substation can be present, to which one or more system components are connected. In particular, the at least one power generation plant and / or the at least one data center and / or the at least one storage device on the island grid side can be connected to the at least one substation.

[0039] The at least one substation can optionally also form a grid transfer node to which a supply grid is connected, which is supplied with electrical energy by the at least one energy generation plant or can provide the local demand in case of unavailability or low generation capacity of the generation plants or for potential grid services of the storage facilities.

[0040] According to a favorable system design, a supply grid, particularly at the high-voltage or extra-high-voltage level, can be connected to the at least one substation. The at least one substation separates the isolated grid from the supply grid.

[0041] Typical voltage levels for a high-voltage grid are 60 kV–150 kV, and for an extra-high-voltage grid, 220 kV–400 kV. These values ​​can vary depending on the country. High-voltage and extra-high-voltage grids are used to distribute electrical energy over long distances of well over 100 km.

[0042] Optionally, a low-voltage or medium-voltage supply network can be connected. Typical voltage levels for a low-voltage network are 400 V–1 kV, and for a medium-voltage network, 10 kV–30 kV. These values ​​can vary depending on the country. Low-voltage networks are used to distribute electrical energy over distances of up to a few kilometers, while medium-voltage networks are used for distances of up to 100 km.

[0043] Locating data centers and storage facilities near such a substation can be advantageous, as the required cable connection can be kept short. Since the cable harnesses that enable good coupling between the system's storage facilities and the supply grid have a relatively large cross-section in a practical design, cost advantages are possible when located adjacent to a local system substation.

[0044] According to a favorable embodiment of the system, the at least one energy generation plant can have at least one renewable energy source. Advantageously, the at least one data center can be operated exclusively with renewable energy. The at least one energy generation plant can, in particular, comprise at least one renewable energy source that provides energy in an intermittent manner; in particular, the at least one energy generation plant can have at least one renewable energy source in the form of a wind turbine and / or solar system. Advantageously, the use of wind energy and / or solar energy and / or hydropower and / or geothermal energy can be carried out. The dimensioning of the at least one data center and the at least one storage device can avoid problems caused by inconsistently available energy sources.

[0045] Advantageously, a location selection for the realization of a system according to the invention with at least one data center can be made primarily according to the availability of an existing or producible local energy generation plant, the producibility of at least one suitable local storage device and the availability of sufficient telecommunications connections, without restricting further parameters of a favorable location with regard to the desired availability class of the data center.

[0046] Compared to conventional data centers, the at least one data center in the system is designed to be small relative to the at least one storage device and local power generation systems and can be located at the local site of, in particular, existing power generation systems. Furthermore, the possibility of self-sufficient and autonomous operation of the at least one data center eliminates the need for easy accessibility and reachability of the data center.

[0047] Furthermore, by cleverly selecting sufficiently small standardized data center modules of, for example, 1 MW IT power, the size of the data center power in the system can be flexibly dimensioned to the local availability of energy generation plants and suitable storage facilities.

[0048] In contrast to the current standard site selection and dimensioning of data centers, primarily based on proximity to the national internet exchange and, due to local operational scaling efficiencies, relatively large dimensions at one location, often with connected capacities of 5 MW to well over 100 MW, a decentralized and modular approach can be advantageous after energy cost optimization and thus particularly after the above site selection in order to optimize life cycle costs including electricity and telecommunications for operators and customers, reduce latency to distributed data sources and optimize the overall energy efficiency of decentralized energy and digital economy.

[0049] According to a favorable embodiment of the system, the at least one data center can be selectively connectable or connected to the at least one energy generation plant and / or to the at least one storage device for its electrical supply. Switching between the at least one energy generation plant and the at least one storage device can advantageously be effected via a control and / or regulating device in the at least one data center itself. In particular, it can be calculated in the at least one data center whether such a switch is advantageous. Forecasts regarding the availability of, in particular, renewable energy can be usefully taken into account.

[0050] According to a favorable embodiment of the system, the at least one data center can have liquid cooling, in particular direct warm water cooling or hot water cooling to at least 35°C, in particular to at least 55°C, preferably to at least 60°C, for cooling computer components, optionally including several closed heat circuits. This allows for effective dissipation of the heat load directly at the high temperature at the chip level of the various computer components as well as the RAM, i.e. a very large proportion of the heat generated in the data center, currently up to approximately 90% depending on usage. Not all computer components need to be cooled advantageously with direct liquid cooling; for example, the storage and network components, which have a lower thermal quality and heat load, can be further cooled via effective air and room cooling or via indirect liquid cooling.

[0051] This results in high waste heat quality, while also reducing the considerable energy expenditure required to first cool the computer components to room temperature and then raise the heat load again to counteract the potentially higher outside temperature, depending on the availability and quality of heat sinks for the data center. This means that at least one data center is highly efficient and, even in the face of advancing climate change, can be directly cooled without increasing the temperature, presumably for significantly more than half the time, or, depending on the IT density, type of use, location, and heat sinks of the data center, up to over 95% of the year. Such high waste heat quality between 50-60°C is also significantly more attractive for residual heat utilization and can lead to its more frequent use.

[0052] If the waste heat is actually used, this can mean further advantages for the efficiency of data center operations, as an additional, possibly temporarily more efficient heat sink is created and the heat user may even take over the final heat quality optimization themselves, thus relieving the burden on the data center operator.

[0053] Due to the high temperature level, hot water cooling allows the waste heat of at least one data center to be fed into local heating networks and / or district heating networks with relatively little additional energy expenditure, used for municipal purposes such as swimming pools, retirement homes, chemical process heat from local industry or, if the location is in a rural area, for local tasks such as greenhouses, hay drying or similar.

[0054] In contrast, today's typical data centers are usually air-cooled or, if necessary, indirectly liquid-cooled via so-called side coolers. The temperature level is typically between 18°C ​​and 24°C, which requires cooling units and onboard fans depending on the site conditions. Waste heat can be used internally to a limited extent, but is generally unattractive for further use at this temperature level and is rarely used in practice. Heat up to approximately 60°C can, in principle, be generated using a heat pump stage, with additional electrical input.

[0055] For use at higher temperature levels of 90°C-120°C, as required in current local or district heating networks, two heat pump stages are required today and are practically non-existent.

[0056] Such liquid cooling also has the advantage of allowing a very compact design of IT systems in the computer cabinets, so-called “racks”, as a much higher heat load can be dissipated.

[0057] Advantageously, the at least one data center can be designed for key areas with high IT densities per rack through a compact arrangement of power components such as CPUs, GPUs, memory, and networking of at least 20 kW per rack, preferably at least 40 kW per rack, particularly preferably at least 80 kW per rack or more. This advantageously allows compact, energy-intensive applications such as high-performance computing (HPC), for example, for simulations and modeling, and / or artificial intelligence and / or cloud infrastructure and / or a combination of computing power to be served.

[0058] According to a favorable embodiment of the system, the at least one data center can provide a heat and / or cooling supply. In particular, the at least one data center can cover the heat and / or cooling requirements of the system and / or other local consumers as needed. Alternatively or additionally, the system can provide a heat and / or cooling supply for users in a local environment of the system. Advantageously, the at least one data center can be connected to a heat and / or cold storage unit, in particular an anticyclical one, or can have a heat and cold storage unit, in particular an anticyclical one.

[0059] Heat and cold storage, heat and cold transport and the respective raising or lowering of the required temperature level can be provided or seasonally buffered, for example, by anticyclical geothermal energy, heat pumps, adsorption chillers or other processes, such as sand storage, liquid storage and ice storage or combinations thereof.

[0060] In this way, efficient heating and cooling supply to the data center and, if applicable, other local consumers can always be achieved using locally generated electrical energy, particularly from renewable sources.

[0061] According to a favorable embodiment of the system, at least one of the at least one storage device can comprise at least one electrochemical storage device, in particular a battery, and / or a hydrogen storage device and / or a water storage device with or without a natural water inflow. Advantageously, the electrochemical storage device can be modular in design and flexibly reduced or enlarged by removing or adding storage modules. A combination of two or more types of energy storage devices is also possible.

[0062] According to a favorable embodiment of the system, at least one additional local consumer can be provided for the electrical supply by the system. For example, an electrolyzer system, in particular a modulatable one, can be provided as a local consumer. In particular, the at least one local consumer can be arranged in a manner that can be prioritized or deprioritized as needed, preferably deprioritized, particularly preferably dynamically deprioritized, with respect to the data center in the electrical supply by the power generation system, even differently depending on the normal or abnormal operating state. The system can thus advantageously be designed to reliably supply one or more local large consumers with energy.By supplying additional local consumers with electricity from the system, the system can enable locally generated and secured electricity tariffs that are ideally locally renewable, and especially ideally so-called "24 / 7 physically locally renewable." This also allows local self-sufficiency and autonomy to be extended to additional local consumers.

[0063] According to a favorable design of the system, the at least one data center can be arranged with priority over the at least one storage device and possibly other local consumers when it comes to the electrical supply by the energy generation plant.

[0064] In this way, a high level of direct supply to the at least one data center can be achieved by the at least one energy generation plant, in particular in abnormal operating conditions such as a permanent power failure.

[0065] According to a favorable embodiment of the system, the at least one storage device and / or the at least one energy generation plant can be designed to stabilize the supply network connected to the at least one transformer substation and / or can be connected or used to feed into and / or store energy from the supply network.

[0066] In addition to providing a reserve for the at least one data center and the at least one power generation plant, as well as an emergency power backup reserve for the at least one data center, the at least one storage facility can thus be used in particular for so-called grid services in the energy industry, for example, in the energy markets: intraday trading, day-ahead trading, grid frequency stabilization using primary control power and secondary control power, and / or in free electricity trading over several days and / or for capacity marketing and / or a security capacity reserve for critical conditions of the general supply grids. Advantageously, the system's very large storage facility can thus generate considerable added value for the supply grid.

[0067] Advantageously, the forecast of the local, possibly dynamic, availability of the energy generation plant, particularly in the case of intermittent renewable energy sources, as well as the forecast of the charge or storage state of at least one storage device and the general situation in the possibly connected supply networks can contribute to coordinating the energy supply from the various sources and storage devices of the system and thus dynamically optimizing the optimal performance of the energy generation plants and storage devices, while at the same time ensuring the local supply to the consumers.

[0068] For example, in the case of long-term forecast high power availability of the at least one energy generation plant, in particular in the case of renewable energy sources, more capacity of the at least one storage device can be dynamically used for other operational purposes of the system, such as grid services for the general supply networks or free electricity trading.

[0069] On the other hand, if long periods of low power availability are forecast, a corresponding amount of the locally generated energy can be retained in order to safely bridge these periods, even if this may mean a lower availability of at least one storage facility for potential grid services.

[0070] In the event of predictably excessive generation power in the general supply grid, a state of the at least one storage device can be selected such that locally generated energy from the at least one energy generation plant of the system can still be produced and stored locally, thus avoiding a forced shutdown. This can advantageously lead to greater efficiency of the at least one local energy generation plant. In the opposite case of a failure of the general supply grid, no grid services can be provided for the general supply grid anyway, so that the full capacity of the at least one storage device is entirely available to the local system.

[0071] According to a favorable embodiment of the system, in the event of a failure or temporary unavailability or limited partial availability of the at least one local power generation plant or one or more of its energy sources, the at least one storage device for electrical supply can be connected or connected to the at least one data center, in particular via a priority circuit. This allows the at least one data center of the system to be reliably supplied with electrical energy even over an extended period of time.

[0072] Advantageously, the at least one data center can be a self-contained or autonomously operated data center. In particular, the at least one data center can have a remote control and / or remote monitoring facility. Advantageously, on-site personnel deployment can be reduced or largely eliminated. This allows the at least one data center to be located even in remote locations.

[0073] According to a favorable design of the system, the at least one data center can also have a, particularly permanent, communication link for satellite communication. A communication link to the outside of the system is advantageous in the event of a permanent failure of the general supply networks and thus also the expected failure of normal fiber optic communication after several hours. For example, satellite communication can be provided. If this is available, the system also remains permanently remotely controllable, so that the at least one data center can continue to be supplied with computing jobs via remote access and can thus fulfill its function for customers even in this abnormal operating state, provided the potentially higher latencies of satellite communication sufficiently support the respective types of use.This allows for satellite-based data backup and / or, with sufficient bandwidth, regular input-output data transport and normal data processing to continue.

[0074] According to a favorable embodiment of the system, the at least one data center can have an uninterruptible power supply and / or emergency power supply for maintaining a certain quality, in particular a certain frequency and / or voltage, and for temporarily and / or permanently continuing the at least one power supply of the at least one data center. This can be a conventional emergency power generator. In contrast to conventional emergency power supplies with emergency power generators, this can be significantly smaller and require less maintenance, as the emergency power supply only needs to be designed for the minimum necessary supply.

[0075] The at least one data center can have a reduced uninterruptible power supply (UPS) capacity and can also forgo conventional fossil-fuel emergency power generators. Nevertheless, the requirements of the desired availability class can be met if the at least one local storage device meets the requirements of a security system of the at least one data center, for example, with regard to shared site security, and has a priority circuit for its supply, thus assuming the function of the emergency power generators required depending on the availability class.

[0076] Basically, the UPS is used to maintain a sufficiently high power quality for the computer components, in particular a frequency and / or voltage with very low fluctuation ranges, as well as to maintain the power supply in the event of a sudden failure of the general power grid until the emergency power generators, which are to be provided depending on the availability class, take over the power supply.

[0077] Thanks to the system's at least one local storage device, the UPS capacity can now be reduced from the usual 15 minutes of IT power plus a margin for multiple, possibly manual, starts of diesel-based emergency generators to, for example, less than 1 minute, depending on the storage technology used. On the other hand, fossil-fuel emergency generators can be completely eliminated in the data center, as a continuous power supply is ensured within the system.

[0078] This can reduce the significant investment and maintenance costs for the data center for these rarely, if ever, needed units. It can also reduce the data center's location requirements in terms of noise, fire loads, pollutants, operational traffic, etc., sufficiently to allow for remote, favorable locations, particularly near a decentralized renewable energy source, or even in or near nature reserves.

[0079] According to a favorable design of the system, the at least one data center can have at least one continuous DC power supply for the primary internal power supply. The primary internal power supply can dispense with AC / DC converters from an external power supply connection up to the computer components, e.g., through the at least one storage device, i.e., in this core supply of the data center.

[0080] Advantageously, the at least one data center can thus be supplied on-site with direct current sources, in particular locally available direct current sources, advantageously for all consumers in the at least one data center, in particular the computer components. Furthermore, advantageously, a conversion of the electrical energy present in the at least one storage device to supply the at least one data center can be omitted, as can the conversions within the UPS and the conversions for each computer component, be it CPU, GPU, memory, network or telecommunications devices. This can also eliminate the need for the numerous power supplies usually found in each rack and thus, due to their high number, also prevent their regularly occurring failures and reduce the associated operational expenditure as well as the temporary functional limitations of the affected computer components.

[0081] Depending on the system's favorable configuration, the system can be designed for temporary and / or permanent island operation and / or black start capability. Advantageously, the system's local island power grid can be kept permanently usable and / or quickly restored as needed from potentially connected general power grids using technical devices such as a microgrid controller and a prepared system disconnection device. Thus, in the event of a temporary, long-term, and / or permanent failure of the general power grid, the functioning of the power supply in the island grid for the at least one data center can be ensured by the automatic recharging of the at least one storage device via the at least one power generation plant.

[0082] According to a favorable design of the system, at least a portion of the computing power of the at least one local data center can be dynamically controlled or controllable as needed, in particular depending on the local and general environment. For example, in the event of an abnormal shortage of electrical supply in the system or in the event of an economic opportunity, in particular for a high energy feed-in to the general supply grid or for a maintained capacity reserve, the computing power can be appropriately shifted in time. For example, a forecast regarding the availability of renewable energy and / or a prioritization of one or more local consumers can also be taken into account. Furthermore, as much heat as possible can be provided for local heat requirements through high utilization of the local computer components.

[0083] According to a further aspect of the invention, a network of decentralized data centers is proposed, which are coupled to at least one system according to the invention.

[0084] The system comprises at least one local data center, at least one local energy generation plant for supplying a supply grid, and at least one local storage device. The at least one data center is designed such that, as an electrical consumer, it requires a maximum of 100% of the annual average electrical energy generated by the at least one energy generation plant in continuous operation each year. The at least one storage device has an electrical output of at least the electrical data center output and a capacity to supply the at least one data center with electrical power for at least 6 hours, in particular under full load, and / or at least a capacity to reliably bridge a largest foreseeable interruption in the local power supply by the at least one energy generation plant (100).

[0085] Advantageously, the at least one storage device can have an electrical power of at least the electrical data center power and a capacity to supply the at least one data center, in particular under full load, with electrical power for at least 6 hours, preferably at least 12 hours, particularly preferably 24 hours, very preferably at least 48 hours, most particularly preferably at least 96 hours.

[0086] This allows the creation of a system with at least one data center that, through the local coupling of decentralized energy and digital economy, can achieve the highest possible overall efficiency for energy and data, thus significantly reducing current losses and inefficiencies and contributing to security of supply and cost reduction in the energy transition.

[0087] To avoid unnecessary repetition of the description of the system and its advantageous embodiments, reference is made to the preceding sections of the disclosure of the invention.

[0088] According to a favorable configuration of the network, the computing power of at least one data center of the system can be dynamically controlled as needed; in particular, the computing power of at least one data center can be shifted temporally and / or spatially. Conversely, the at least one data center of the system can also temporarily take over computing power for other data centers of the network if necessary. drawing

[0089] Further advantages will become apparent from the following description of the drawings. The figures illustrate exemplary embodiments of the invention. The figures, the description, and the claims contain numerous features in combination. Those skilled in the art will also expediently consider the features individually and combine them into useful further combinations.

[0090] Examples include: Fig. 1 an embodiment of a system according to the invention, with a power generation plant, a data center and a storage device; Fig. 2 shows an embodiment of a system according to the invention, comprising a power generation plant, a data center and a storage device, as well as a supply network connected to a local substation of the system; Fig. 3 a block diagram of the system according to Fig. 2; Fig. 4 an embodiment of a system combination according to Fig. 1 or Fig. 2 with decentralized data centers. Embodiments of the invention

[0091] In the figures, components of the same type or function similarly are designated by the same reference numerals. The figures are merely examples and are not to be construed as limiting.

[0092] Before describing the invention in detail, it should be noted that it is not limited to the specific components of the device and the specific method steps, as these components and methods may vary. The terms used herein are intended only to describe particular embodiments and are not intended to be limiting. Furthermore, when the singular or indefinite articles are used in the description or claims, this also refers to the plural of these elements, unless the overall context clearly indicates otherwise.

[0093] The directional terminology used below, including terms such as "left," "right," "top," "bottom," "before," "behind," "after," and the like, is intended solely to enhance understanding of the figures and is in no way intended to limit the scope of the invention. The components and elements depicted, as well as their design and use, may vary according to the considerations of a person skilled in the art and may be adapted to specific applications.

[0094] Fig. Figure 1 schematically shows a system 1100 according to an embodiment of the invention, comprising a local energy generation plant 100 with renewable energy sources. Optionally, multiple local energy generation plants 100 are also possible.

[0095] In this embodiment, the system 1100 is designed as an island network 1000 and comprises the local energy generation plant 100, a local data center 200 and a local storage device 300.

[0096] The energy generation plant 100 is designed as a wind-solar hybrid park with a wind turbine 150 with wind turbines and a solar power plant 160 with solar panels as renewable energy sources 110. The Fig. The data center 200 shown in Figure 1 comprises, for example, four data center modules not designated in more detail. The storage device 300 is designed, for example, as a battery, in particular as a so-called mega battery, which in this exemplary embodiment comprises eight storage modules not designated in more detail. The dashed arrow lines in Fig. 1 shows by way of example that the data center 200 and the storage device 300 are generated with energy by the energy generation plant 100.

[0097] The dotted line between storage device 300 and data center 200 indicates a power supply of the data center 200 by the energy storage unit 300.

[0098] It can be provided that the local storage device 300 also provides a supply of energy for the energy generation plant 100 or the wind turbine 150, for example for rotor heating in winter, and the solar system 150, which is not particularly emphasized.

[0099] In the island grid 1000, two additional local consumers 170, for example electrolysers, are arranged.

[0100] It is understood that each of the components energy generation plant 100, data center 200, storage device 300 and local additional consumers 170 can, of course, have their own conversion devices (not explicitly shown) for converting the respective voltage level and / or voltage type and / or the respective frequency.

[0101] Fig. Figure 2 schematically shows a system 1100 according to a further embodiment of the invention. The island network 1000 present in the system 1100 is Fig. 1, to which reference is made to avoid unnecessary repetition.

[0102] In addition to the island grid, the system has 1000 (see Fig. 2) a substation 120. This forms a grid transfer node to which a supply network 1500 is connected externally.

[0103] Substation 120 of system 1000 separates the island grid 1000 from the supply grid 1500.

[0104] A power supply from substation 120 to data center 200 and storage facility 300 and from storage facility 300 to data center 200 is shown with dotted lines in Fig. 2 indicated.

[0105] The dotted line between energy storage unit 300 and data center 200 indicates that the data center 200 is supplied with power by the energy storage unit 300.

[0106] It can be provided that the local storage device 300 also provides a supply of energy for the energy generation plant 100 or the wind turbine 150, for example for rotor heating in winter, and the solar system 150, which is not particularly emphasized.

[0107] Data center 200 and storage facility 300 are located adjacent to the substation 120 of the power generation plant 100. Optionally, a separate backup substation 122 (see Fig. 3) be present.

[0108] From the substation 120, electrical energy is transferred to the supply grid 1500 at a voltage level, for example, at the high-voltage or extra-high-voltage level, or optionally taken from it. An arrow to the right of the substation 120 indicates that the energy from the power generation plant 100 is delivered to the supply grid 1500 via the substation 120 at a corresponding voltage level. The components of the system 1100, in particular the power generation plant 100, the data center 200, and the energy storage system 300, are located upstream of the substation 120 and form an island network separate from the supply grid 1500. During normal operation, the data center 200 does not receive any energy from the supply grid 1500.

[0109] Here, too, it is understood that each of the components energy generation plant 100, data center 200, storage device 300 and local additional consumers 170 can, of course, have their own conversion devices (not explicitly shown) for converting the respective voltage level and / or voltage type and / or the respective frequency.

[0110] The Fig. The separate backup substation 122 shown in Figure 3 can advantageously supply the data center 200 with energy as needed. The substation 120 of the power generation plant 100 and the power generation plant 100 charge the storage device 300, which is designed as a battery, which can, in turn, supply energy to the substation 120 and to the power generation plant 100 as needed. The storage device 300, in turn, can supply energy to the data center 200 as needed.

[0111] In Fig. 3, the thin solid lines from power generation plant 100 to substation 120, storage facility 300, and data center 200 indicate the primary energy supply in system 1100 as a direct operational supply by power generation plant 100. In this case, energy is supplied from substation 120 to a supply network 1500 (see Fig. 2) handed over.

[0112] Any conversion devices that may be present for converting the respective voltage level and / or voltage type and / or the respective frequency of the power generation plant 100, the data center 200 and the storage device 300 are not shown.

[0113] The thick solid lines between storage device 300 and data center 200 as well as between storage device 300 and power generation plant 100 indicate a temporary, partial or permanent residual power supply of the data center 200 or the internal power supply of the power generation plant 100 from the local storage device 300 if the data center 200 or the internal power supply of the power generation plant 100 cannot be sufficiently supplied via the direct supply from the power generation plant 100.

[0114] The dotted lines from storage facility 300 to substation 120 and to the separate substation 122 indicate the use of storage facility 300 for grid services, capacity marketing, and free electricity trading via the supply networks 1500 connected to substations 120 and 122.

[0115] The dashed lines from the separate substation 122 to the data center 200 and from the substation 120 to the storage facility 300 indicate optional, conventionally usually existing power grid connections, which can represent an additional increase in supply security for the system 1100, for example, if the local power generation plant 100 should fail permanently.

[0116] The further dashed line from the storage device 300 to the data center 200 indicates the provision of a locally stored emergency power reserve in order to meet the specifications of the selected availability class for a temporary or permanent emergency power supply of the data center 200, if necessary dynamically extended via the forecast of the performance and availability of the energy generation plant 100, in particular in the case of renewable energy sources with their natural energy storage, for example wind or sun.

[0117] The power generation plant 100 typically has, purely by way of example and without limitation, an electrical power peak of at least 10 MW, particularly in the case of intermittent renewable energy sources, advantageously 100 MW, particularly advantageously 1 GW and more. The power generation plant 100 provides energy to the supply grid 1500. The supply grid 1500 can be a transmission grid at high-voltage or extra-high-voltage level, or a distribution grid at medium-voltage or low-voltage level.

[0118] The data center 200 is designed so that, as an electrical consumer in continuous operation, it requires a maximum of 100% of the peak electrical power of the annual average electrical energy generated by the power generation plant 100.The storage device 300 has an electrical power of at least the power of the data center 200 and a capacity to supply the data center 200 with electrical power, in particular under full load, for at least 6 hours, preferably at least 12 hours, particularly preferably at least 24 hours, very particularly preferably at least 48 hours, and most particularly preferably at least 96 hours, and / or at least a capacity to safely bridge locally the largest foreseeable interruption in the local power supply by the energy generation plant 100, for example during a so-called dark period, during maintenance, or the like, even in the event of a potential reduction in storage quality over the life cycle of the storage device 300.

[0119] The local system 1100 is represented by an example of a renewable energy source, a small data center 200, and a large energy storage device 300 with sufficient stored energy for an ecologically favorable, self-sufficient, autonomous data center 200.

[0120] The computer components and data processing are simply called IT for short.

[0121] For simplicity, the following analysis is based on an energy efficiency of data center 200 very close to, or rounded to, 1.0, i.e., almost completely free cooling and minimal other power requirements. For example, a wind-solar hybrid park as an energy generation plant 100 has power peaks of 54 MW from wind energy and 9 MW from solar energy. Data center 200, for example, has a power requirement of 1 MW for its IT. The storage facility 300, in the form of a mega-battery, has an energy capacity of 100 MWh.

[0122] In addition, this exemplary wind-solar hybrid farm, based on real-life conditions, features a substation 120 with a high-voltage supply network 1500, as well as a local network with a separate second substation 122 for grid services, etc.

[0123] For the classification of system 1100, it is also taken into account that the data center 200 can enjoy priority in the wind-solar hybrid park, or the two consumers 170 designed as electrolyzers (see Fig. 1 and Fig. 2), with a respective peak power of 2 MW, can only be operated in a negligible basic mode if the power generation plant 100 can provide no or only low power.

[0124] With this supply priority, the data center 200 is directly supplied by the energy generation plant 100 at over 70% of the annual power consumption at this specific location. The actually measured or estimated durations of periods of low power and other interruptions and mandatory shutdowns, or the periods with less generation capacity from the energy generation plant 100 than consumption in the data center 200, have a maximum duration of approximately 55 hours at a time. With an appropriate forecast of the generation capacity of the energy generation plant 100 and appropriate capacity control of the storage facility 300, any foreseeable insufficient direct supply to the data center 200 can be bridged locally by the storage facility 300 with virtually no problem.

[0125] The generation capacity of this wind-solar hybrid park is more than fifty times greater than the IT payload of this data center 200, so that even with existing energy generation technology, the data center 200 can be fully supplied directly in the island grid 1000 of the system 1100 most of the time.

[0126] With further technological advances in renewable generation technology, baseload generation can continue to grow and enable larger, locally self-sufficient data centers.

[0127] The mega battery selected here as a storage device 300 can, depending on its life cycle, bridge the full data center load for approximately 100 hours, i.e., more than four full days, until sufficient power is automatically available locally via the renewable energies of the power generation system 100. Typically, the mega battery is sufficiently recharged every 24 hours using local solar energy, even during a prolonged period of wind calm.

[0128] If these consumers 170 are prioritized differently in the wind-solar hybrid park, for example, if the data center 200 is given lower priority than the two local consumers 170 designed as electrolyzers (see Fig. 1, Fig. 2) with 2 MW peak power each, the data center 200 would only be directly partially supplied from a local generation capacity of 4 MW, and the mega battery would only be served secondarily from 5 MW.

[0129] In this case, local generation gaps of less than 5 MW of generation capacity lasting up to 90 hours can occur, but even these can be safely bridged thanks to the size of the 300 storage facility. However, prioritizing data center 200 and sufficient battery recharging, at least during a permanent blackout or periods of darkness lasting more than 24 hours, is recommended and increases resilience, since only statistical generation profiles or expected values ​​are available for these intermittent renewable energy sources in the 1100 system.

[0130] If, for example, a hydrogen chain is planned in an expansion stage of the data center 200, the additional hydrogen storage could be operated directly with the existing consumers 170 designed as electrolyzers in order to bridge particularly long potential periods of darkness, especially in the event of a simultaneous and permanent failure of the general supply networks.

[0131] However, the values ​​presented here are highly site-specific and must be interpreted accordingly. For example, at another location, the direct supply for a potential data center 200 from the existing wind-solar hybrid park in the island grid 1000 would be more than 92% per year, with correspondingly shorter periods of dark periods or other interruptions. Here, the size of the storage facility 300 would be used specifically to meet the emergency power requirements of the selected availability class of the data center 200. In reality, more capacity of the storage facility 300 would be available for energy-related use.

[0132] In contrast to a traditional diesel emergency power supply, where the diesel storage units would generally have to be refueled after about 48 hours in the event of a prolonged blackout, assuming this is actually practically possible in such an emergency. This means that personnel, communications, traffic routes, means of transport, and fuel in the form of diesel should be given sufficient priority over the supply of power plants, hospitals, and other civilian and military emergency supplies. In such a case, this data center 200 would automatically be sufficiently supplied even in the event of a permanent blackout, thus remaining permanently self-sufficient.

[0133] A sustainable, direct design of the data center 200 can be advantageous and sustainable for use cases with high energy utilization, hardware utilization, and IT density. The data center 200 can be designed in modules that directly address the respective use case, e.g., for cloud infrastructure or specific, typically batch-operated computing services such as artificial intelligence (AI) or supercomputing / high-performance computing (HPC).

[0134] As a direct cloud data center or direct AI / HPC data center designed in this way, more than 50 kW or even more than 100 kW per rack can be provided and cooled for the compute nodes, for example, instead of the traditional 3 kW to 8 kW per rack.

[0135] Likewise, a “computing load flexibility” is possible for sustainable data center customers: In batch-operated applications such as AI or HPC, the computing load can be shifted temporally or spatially if an unfavorable power supply situation is present. This is exemplified in Fig. 4, which shows a network 2000 of decentrally distributed external, in particular island-free data centers 210, which are coupled to a system 1100.

[0136] In many HPC clusters, the calculations performed there take several hours or days anyway, so a time shift is only marginally problematic. Even with today's cloud services, a local shift often occurs automatically if a region experiences a blackout.

[0137] The data center 200 can achieve energy-efficient operational cooling of its high-performance IT components by means of hot water cooling, in particular by means of “direct hot water cooling,” and thus realistically enable the reuse of heat from the data center 200 in local heating or district heating networks.

[0138] Primary warm or hot water cooling of key IT components to, for example, 60°C results in optimal energy efficiency, e.g., by eliminating the otherwise usual onboard air coolers, as well as reducing or even completely eliminating traditional cooling units. This does not necessarily mean that all data center and IT components in the IT room operate at this temperature; rather, the room generally remains at a seasonally acceptable operating temperature of, for example, 15°C to 30°C for components and any occasional personnel deployment. Only the computing chips (CPU, GPU, etc.) and RAM in the racks, which are required depending on the above application, and thus have the greatest heat load and heat quality, are cooled directly with hot water.

[0139] The residual heat load in the room, which, depending on the application, amounts to less than 5% to 25% and comes from network and storage components, data center technology, which are typically air-cooled or indirectly liquid-cooled via side coolers, as well as from remaining insulation / heat losses from direct hot water cooling, is raised to the hot water temperature, e.g., using a heat pump, depending on the dynamic optimum. This allows for continuous free cooling of the entire data center 200 in Central European latitudes, even with progressively rising temperatures, or to make the heat available to a potential subsequent user without further heat pump increases. This can be advantageous, for example, for agricultural greenhouses, hay drying, etc., or other process heat, for local community purposes such as swimming pools and outdoor pools, or as partial load or base load in local heating or modern district heating systems.

[0140] Heat and cold storage can also be advantageous, particularly countercyclical heat and cold storage. This can further improve energy efficiency by utilizing local seasonal buffering of excess heat or cold. Depending on the location, for example, using surface geothermal energy via boreholes or sand-based heat storage, in addition to short- and medium-term electricity storage in the mega battery, the heat and cold supply can also be designed and used sustainably both in the short term during the day and seasonally. This applies both to the seasonal environment of Data Center 200 in System 1100, e.g., via rooftop solar thermal energy in summer, and to the aforementioned direct hot water cooling.

[0141] The data center 200 of the island grid 1000 of the system 1100 allows for largely unmanned operation through full remote control and monitoring without on-site personnel or regular on-site customers. For the modern use cases mentioned above, the direct data center can easily operate without permanent on-site personnel, allowing customers to fully control their workloads remotely.

[0142] In these use cases, even a partial hardware failure, e.g., of hard drives or compute nodes, is generally not problematic; rather, these failures are simply logically ignored and thus "self-healed," even if the overall capacity decreases marginally. This means that defective components can be replaced in a clustered manner or delayed for weeks, months, or even years. In the event of a permanent blackout and potentially rapidly following downstream failures, e.g., due to a lack of on-site personnel and transport logistics, no relevant operational restrictions arise for Data Center 200.

[0143] In addition to the usual redundant fiber optic connections, a "satellite backup" can be provided, i.e., ensuring self-sufficiency through satellite backup communication. Modern satellite communication, such as that provided by SpaceX's Starlink, enables the autonomous data center 200 to be supplied with computing jobs via remote access and to fulfill its function for customers even in the event of a permanent regional or continental blackout. Conventional fiber optic connections are generally not permanently available without power, or a receiving fiber optic endpoint may no longer be available in the event of a continental problem. However, this emergency operation may cause latency problems for certain types of use.

[0144] When using a storage device 300, such as the megabattery, the capacity of the uninterruptible emergency power supply (UPS) can be reduced due to the elimination of diesel-powered emergency power generators, which could potentially cause false starts in the event of a call. The megabattery can take over the power supply to the data center 200 or the replenishment of power to the UPS in just a few milliseconds. Thus, the UPS's task focuses on the quality of the power for IT and not on bridging multiple diesel start attempts.

[0145] When using a storage device 300 in the form of a mega battery and direct data centers, where no customers but only selectively trained personnel have to go to the premises of the data center 200, it is possible to optionally switch completely to a direct current supply, so that the usual multiple AC-DC converters and their power losses can be eliminated.

[0146] When using a storage device 300 in the form of the mega battery, the capacity that does not have to be reserved for the data center 200 as a backup system can be used for so-called grid services such as power grid frequency stabilization using primary and secondary control energy, intraday and day-ahead trading, other capacity marketing, or for free energy trading with green and gray energy.

[0147] In addition to the energy generation system 100, which here is based on wind and / or solar energy sources, other energy sources, particularly renewable ones, and storage technologies, including run-of-river, storage, pumped-storage, or combinations thereof, can also be used. Here, too, it can be ensured that the local island grid 1000 of the system 1100 remains permanently available even during a permanent blackout. Reference symbol 100 power generation plants 110 renewable energy sources 120 substation 122 substation 150 wind turbines 160 solar power plants 170 local consumers 200 data center 210 external data center 300 storage device 1000 island grid 1100 System 1500 supply network 2000 Association QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited non-patent literature

[0000] ISO / IEC 22237

[0032]

Claims

[1] System (1100) with at least one local data center (200), as well as with at least one local storage device (300) and with at least one local energy generation plant (100), wherein the at least one local data center (200), the at least one local storage device (300) and the at least one local energy generation plant (100) form a, in particular closed, island network (1000), the at least one data center (200) is designed as an electrical consumer such that in continuous operation it requires a maximum of 100% of the annual average electrical energy generated by the at least one power generation plant (100), and the at least one storage device (300) has an electrical power of at least the electrical data center power and a capacity sufficient to supply the at least one data center (200) with electrical power locally for at least 6 hours, in particular under full load, and / or at least a capacity to safely bridge locally a largest foreseeable interruption in the local power supply by the at least one energy generation plant (100). [2] System according to claim 1, further comprising at least one local substation (120) to which one or more components of the system are connected, in particular wherein the at least one power generation plant (100) and / or the at least one data center (200) and / or the at least one storage device (300) on the side of the island grid (1000) are connected indirectly or directly. [3] System according to claim 2, wherein a supply network (1500), in particular at high voltage level or extra-high voltage level, is connected to the at least one substation (120), wherein the at least one substation (120) separates the island network (1000) from the supply network (1500). [4] System according to one of the preceding claims, wherein the at least one energy generation plant (100) comprises at least one renewable energy source (110), in particular wherein the at least one energy generation plant (100) comprises at least one discontinuous renewable energy source (110), in particular in the form of a wind turbine (150) and / or solar plant (160). [5] System according to one of the preceding claims, wherein the at least one data center (200) is selectively connectable or connected to the at least one energy generation plant (100) and / or to the at least one storage device (300) for its electrical supply. [6] System according to one of the preceding claims, wherein the at least one data center (200) has a liquid cooling system, in particular a direct warm water cooling system or hot water cooling system to at least 35°C for cooling computer components. [7] System according to one of the preceding claims, wherein the at least one data center (200) provides a heat and / or cooling supply, in particular a heat and / or cooling supply for users in a local environment of the system. [8] System according to one of the preceding claims, wherein at least one of the at least one storage device (300) comprises at least one electrochemical storage device, in particular a battery and / or a hydrogen storage device, and / or a water storage device. [9] System according to one of the preceding claims, wherein at least one further local consumer (170) is present for the electrical supply by the at least one energy generation plant (100), in particular wherein the at least one further local consumer (170) is arranged in a manner that can be prioritized or deprioritized, preferably deprioritized, particularly preferably dynamically deprioritized, in the electrical supply by the at least one energy generation plant (100) relative to the at least one data center (200) as required. [10] System according to one of the preceding claims, wherein the at least one data center (200) is arranged with priority over the at least one storage device (300) in the electrical supply by the at least one power generation plant (100). [11] System according to one of the preceding claims, wherein in the event of failure or temporary unavailability or limited partial availability of the at least one local energy generation plant (100) or one or more of its energy sources, the at least one storage device (300) for electrical supply is connectable or connected to the at least one data center (200), in particular via a priority circuit. [12] System according to one of the preceding claims, wherein the at least one data center (200) has at least one, in particular permanent, communication connection for satellite communication. [13] System according to one of the preceding claims, wherein the at least one data center (200) has an uninterruptible power supply and / or emergency power supply for maintaining a quality, in particular a frequency and / or an electrical voltage, and a temporary and / or permanent continuation of the power supply of the at least one data center (200). [14] System according to one of the preceding claims, wherein the at least one data center (200) has at least one continuous DC power supply for the primary internal power supply. [15] System according to one of the preceding claims, which is designed for temporary and / or permanent island operation and / or black start capability. [16] System according to one of the preceding claims, wherein at least a part of a computing power of the at least one data center (200) is dynamically controlled or controllable in time as required. [17] Network (2000) of decentralized distributed data centers (210) which are coupled to at least one system (1100) according to one of the preceding claims. [18] Network according to claim 17, wherein at least a part of a computing power of the at least one data center (200) of the system (1100) is dynamically controlled as needed, in particular wherein a computing power of the at least one data center (200) is temporally and / or spatially shiftable.

Images