WO2024068730A1

WO2024068730A1 - System and method for prioritising electrical usage

Info

Publication number: WO2024068730A1
Application number: PCT/EP2023/076694
Authority: WO
Inventors: Anders SPUR; Per MADSEN; Rune Petter Domsten; Robert OTREBA
Original assignee: Watts AS
Current assignee: Watts AS
Priority date: 2022-09-27
Filing date: 2023-09-27
Publication date: 2024-04-04
Anticipated expiration: 2025-03-27
Also published as: EP4566139A1

Abstract

A master unit for use in a modular distributed power consumption system for distributing electrical power between flexible electricity units. The modular distributed power consumption system comprises a master unit adapted for directing electrical power and for prioritising the distribution of said electrical power to a plurality of outlet units. The master unit comprising a processing unit configured for on-site training of an intelligent system and determining a consumption schedule. Flexible electricity units connected to the modular distributed power consumption system via the outlet units are operated in accordance with the consumption schedule. The master unit is adapted for directing at least a subpart of the available electrical power to one or more of the outlet units based on an objective function receiving an electrical power forecast. Preferably the master unit determines the consumption schedule based on an electricity consumption forecast and/or an electricity generation forecast determined on-site by the intelligent system.

Description

SYSTEM AND METHOD FOR PRIORITISING ELECTRICAL USAGE

TECHNICAL FIELD

The present invention relates to the field of prioritising schemes for optimised electricity usage in a power grid, in particular the prioritised allocation of electricity to electrical consumers in a power grid.

BACKGROUND OF THE INVENTION

In an effort to decrease the environmental impact of emissions from carbon-based fuels, there has in recent years been a focus on developing solutions for local electricity generation and storage such as solar cells on buildings, small water or wind turbines, etc. Similarly, new technological solutions based on electricity in favour of other fuels such as electrical vehicles (EVs) are being introduced to the market. Integration of such electricity generators and electricity consumption units and storages is an important step in reducing emission, but the increased loads and conduction of power causes a significant strain in the power grid. If there is no control on when electricity is delivered to which electricity consumption unit, e.g. when the EVs are charging, it can lead to significant negative consequences such as power outages or reduced quality of the delivered power. The impacts may affect the larger regional grid and lead to power outages of entire neighbourhoods, or it may affect local micro grids such as a single household or a grid-sharing collection of households.

Some users attempt to manually compensate for this by running their dishwasher or charging their EVs at times when there is usually less strain on the power grid, e.g. at night. However, the users cannot efficiently predict when there will be a strain on the network and many users may still use electrical power the same period of time, e.g. right before they go to bed, resulting in a shift of the peak load, rather than an efficient load distribution.

Automatic systems where the electricity usage of each electricity consumption unit, such as EVs, is distributed in response to the concurrent price of electricity have been suggested in the prior art. However, such systems rely on each EV being subscribed to the system and the availability of charging stations being configured to receive the input from the EVs. This leads to only a limited selection of charging stations being accessible, or the benefits of distributing the electricity usage will be lost. Furthermore, subscription systems limit the sharing of charging stations by different users or the verifiable monitoring of the usage of different EVs belonging to the same user. In addition, incomplete knowledge of generation and consumption in the system causes the consumption schedules to be ineffective regardless of which prioritisation model is being used.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a system and a method for distributed power usage of a plurality of electricity consumption units such as EVs and household appliances where the distribution is controlled by a master unit in accordance with a dynamic consumption schedule such that the intended distribution is ensured regardless of which electricity consumption units are connected to the power consumption system, and the consumption is adjusted based on the electricity supply to the microgrid including supply from local electricity generators. A further object of the invention is to enable the maximisation of the total electrical power output from a grid such that a continuous load can be maintained.

The above object and advantages, together with numerous other objects and advantages, which will be evident from the description of the present invention, are according to a first aspect of the present invention obtained by:

A master unit for use in a modular distributed power consumption system for distributing electrical power between flexible electricity units, the modular distributed power consumption system comprising, a master unit adapted for directing electrical power and for prioritising the distribution of said electrical power to a plurality of outlet units, said master unit comprising a processing unit configured for on-site training of an intelligent system and determining a consumption schedule according to which flexible electricity units connected to said modular distributed power consumption system via said outlet units use and/or deliver electrical power, said master unit being adapted for directing at least a subpart of the available electrical power to one or more outlet units of said plurality of outlet units based on an objective function receiving an electrical power forecast.

In a variant there may provided a master unit for use in a modular distributed power consumption system for distributing electrical power between electricity consumption units, the modular distributed power consumption system comprising a master unit adapted for directing electrical power and for prioritising the distribution of said electrical power to a plurality of outlet units. The master unit comprises a processing unit for determining a consumption schedule according to which electricity consumption units, electricity generators and/or electricity reservoirs connected to the modular distributed power consumption system via the outlet units use and/or deliver electrical power. The master unit is adapted to direct at least a subpart of the available electrical power to one or more outlet units of the plurality of outlet units based on a priority scheme.

By a master unit is understood a unit which may control other units and their functionality. Such a master unit may thus be configured to communicate with various outlet units, electricity consumption units and/or electricity reservoirs in the distributed power consumption systems as well as various electricity generators. The master unit may further be configured to control the operation of those units with which it communicates, e.g. if they use electrical power and/or deliver electrical power into the local power grid. The master unit may change its operation in response to data received from other units in the distributed power system, based on conditions of the master unit, but the other units of the system may not directly control the operation of the master unit. Communication between the master unit and other units may be wired and/or wireless for the various components it controls as well as for communication with external systems or other master units.

By flexible electricity units are understood electrical units of the microgrid which are configured to have their consumption controlled by the master unit. Flexible electricity units may include deferrable loads, e.g. electricity consumption units which can be consume electrical power at a time determined by the charging schedule rather than immediately upon activation this may for example be a dishwasher or an electrical vehicle. Flexible electricity units may include electricity reservoirs which are set up to be able to store electrical power and/or supply electrical power to the microgrid in response to priority factors and the priority scheme determined by the master unit, e.g. capacity of the regional grid, market prices, phase load or other priorities set by the user or as a default of the system. Flexible electricity units may include electricity generators which may be turned on and off such that the production may be curtailed. By the flexible electricity units being set up in such a manner is understood that it may be by way of physical connection and/or it may be via permissions indicated to the master unit for how the charging schedule is determined. Most microgrids will comprise both flexible electricity units and inflexible electricity units, where inflexible electricity units will charge or generate power immediately upon connection with the microgrid regardless of the charging schedule, in which case the charging schedule will need to be updated to account for the changes caused by the inflexible units. In some variants the connection and permissions relating to the units of the microgrid enables variation of which units are flexible over time. In the following it is understood that electrical units described as controlled by the master unit and the created charging schedule are various types of flexible electricity units. Flexible electricity units may for example be electricity reservoirs such as batteries and they can also include heat pumps that may be switched on and off to control the load or they may be any household appliance which is set up to be operated in a flexible manner according to the charging schedule rather than depending on when it is switched on or off by the user, e.g. it may include a washing machine which does not need to operate at a specific time but may be run flexibly within a timeframe.

Having a master unit controlling the distribution of the electrical power between electricity consumption units of a distribution system enables the optimisation of power consumption within the distributed power consumption systems such that the amount and timing of electricity consumption may be matched to the availability and price of electricity as well as the customary usage of the microgrid, and the consumption may be distributed over time.

The master unit may thus function as a logical switch which monitors, controls and distributes electrical power among outlet units and the connected electrical devices. It is to be understood that the master unit is not itself an electrical switch, but is installed to control the switches within a local grid. The master unit thus monitors and distributes by controlling such a switch. Throughout the application any indication that the master unit receives electrical power is understood as the electrical components which the master unit controls receiving that electrical power, such that the master unit may monitor, direct and control it.

By the master unit being configured for on-site training of an intelligent system is understood that the master unit comprises a processing unit for the computer implementation of an intelligent system and the training thereof. The master unit may receive datasets and provide them as inputs for the intelligent system training it for a specific task based on the input data. In a preferred variant the intelligent system is trained for creating an electrical power forecast based on one or more forecasts datasets of other forecasts, e.g. temperature forecast, irradiation forecast, or price forecast, and on-site collected sensor datasets, e.g. temperature measurements, measurements of actual energy generation, and/or data related to conditions of the microgrid and the site of the microgrid, e.g. conversion rates of solar cells, losses of inverters, or temperature distribution within a building. Such electrical power forecasts as created by the intelligent system may be used as input for the on-site solving of an objective function which is optimized to determine how the master unit is to direct the electrical power in accordance with one or more priority factors. In some variants it is necessary that the various priority factors are accounted for and weighed relative to each other and the objective function is solved such that a suitable trade-off between the various priority factors is achieved. In some preferred variants the output of the objective function may be a pareto frontier presenting numerous possible solutions where the priority factors have been weighed differently, thus allowing the user to make an informed decision regarding how they wish for the master unit to direct the electrical power.

By an intelligent system is understood a deep learning algorithm which can be trained and will develop over time based on the inputs provided to the intelligent system. The intelligent system may thus be considered an artificial intelligence. The intelligent system may be considered a machine learning algorithm. In a preferred variant the priority scheme is calculated based on the solution of an objective function having as input an electrical power forecast created by the intelligent system for the entire duration of the prediction horizon, the relevant priority factors to be optimized with respect to and a set of initial system conditions, e.g. state-of-charge of an electricity reservoir. The priority scheme may be calculated based on the above inputs by solving the optimisation problem set up as the objective function.

By having the intelligent system trained on-site is understood that it is trained and performs the solving of the objective function on the master unit locally at the site of the microgrid. Training the intelligent system on-site has numerous benefits among these are the fast processing which is not limited by the transfer of data off-site and that the loss of internet connection would not cause the system to be unable to update calculations or malfunction. Furthermore, on-site training allows the system to handle larger datasets relating to the microgrid and the site of the microgrid as processing power is not shared and the processing does not need to account for external sites. Data relating to similar sites may be accounted for based on finished calculations and be input as part of training sets, but the master unit need only solve the objective function for the local site and microgrid. In other words the on-site training means that the master unit is an edge computing system. The master unit may receive external data, e.g. via an internet connection, but will operate without such connection. In some variant the master unit may be set up to transmit data to a server, e.g. to compare usage or to use schedules as input for other related systems, however it is not necessary for the system to be able to transmit data off site for it to function. Using and on-site trained system may also be a cheaper solution as it is not necessary to purchase server space. The onsite computing enables the calculation of larger datasets and higher granularity of the input and output data, i.e. it is possible to account for more specific conditions of the microgrid. For example, in a microgrid having a photovoltaics generator installed the system may account for variations in the irradiation detected at the site due to shadows cast during specific times of the day and may determine the conversion rate of the actual photovoltaic generator rather than of a standard solar cell. As another example the specific losses of the inverters of the microgrid may be accounted for, such that the difference of the losses of the various inverters of a microgrid may be taken into account when planning the charging schedule, e.g. to minimise the activation of the inverts with the highest losses. This in turn allows for a more precise forecast of the electricity generation and consumption and the prediction horizon may be extended.

By the distributed power consumption system controlled by such a master unit being modular is understood that the various components connected to the master unit may be operated independently of each other, as the overall system is controlled by the same master unit. Hence, a single master unit may control a plurality of slave outlet units and the slave electricity consumer and/or slave electricity generators and/or slave electricity reservoirs connected thereto. By an outlet unit is understood a connection point where an electricity consumer, electricity generator and/or electricity reservoir may be connected to the local power grid. The outlet unit may be a socket, or it may be any connection point, e.g. on an electrical cable, directly connected to an electrical appliance. The slave outlet units may be added to or removed from the distributed power consumption system without directly affecting each other. It is to be understood that there may be a consequent effect of adding or removing an outlet unit, as it may be chosen to change the settings of the master unit to prioritise the connected outlet units, however this will be controlled by the master unit and not by the connected outlet units.

The modularity of the distributed power consumption system has the benefit of making it cheap and easy to reconfigure the distributed power consumption system to comprise the number of outlet units needed at a particular site. For example, if the distributed power consumption system is installed in the microgrid of a household, an additional outlet unit may be installed without needing an additional master unit. This further contributes to making the distributed power consumption system cheap, as it is not necessary to have a separate control unit for each outlet mount. They may all benefit from the computational power of the master unit.

By the master unit being adapted to direct at least a subpart of the received electrical power to each outlet unit based on a priority scheme, is to be understood that it directs at least a subpart of the received power to each of the outlet units. The timing and/or the fraction of power being directed to each of the outlet units depends on the priority scheme. The priority scheme may in turn be based on a number of priority factors and/or forecasts including the specific requirement of each device connected in the grid; for example a freezer requires a constant supply of electrical power while other utilities, such as entertainment devices, are less critical.

By the master unit directing power is understood that it determines the amount of power whish should be made available to the various units, should be used to charge an electricity reservoir in a certain time frame or which should be directed back to a local and/or regional power grid. Thus the master unit does not need to be directly connected to any electricity lines or switches, rather it sends directions for inverters, switches or other units which follows the directions of the master unit.

In a preferred variant, the priority factors are at least in part set by the user. The user may e.g. set priority factors via a log-on to a profile on a server, for example via a browser or an app. The user may set individual priority factors or select between predefined sets of priority factors. By the user is understood any user with authorisation for setting priority factors of the master unit, e.g. by having access to the unit and/or a login for online access. Hence, it is to be understood that the setting of the master unit and priority scheme need not be restricted to a single user, but may be controlled by priority factors set by individual users of the local power grid.

In some preferred variants predefined sets of priority factors are described as operation modes allowing a user to select an operation mode without needing to know details of the related priority factors. An operation mode may be e.g. be everyday mode, vacation mode where less consumption is expected, guest mode where increased consumption is expected, maintenance mode or update mode where changes to the system are expected or any other mode indicating a period of a specific type of user action which is expected to affect the electricity consumption. Choice of operation mode may determine what the objective function is solved for and thereby what the charging schedule is optimised for. In a preferred variant the master unit is set to operate in a default mode if no user input is provided such that the user does not need to engage with the system after installation.

According to a further embodiment of the first aspect of the invention, the master unit is adapted for monitoring and distributing electrical power from a local power grid, the local power grid receiving electrical power from one or more of a regional power grid, one or more local power generators and/or one or more local energy reservoirs and the master unit further being adapted for directing electrical power to a local power grid and/or to a regional power grid.

In a variant the master unit is adapted for monitoring and distributing electrical power from a local power grid, the local power grid receiving electrical power from one or more of a regional power grid, local power generators and/or a local energy reservoir.

By a local power grid is understood a power grid confined to a specific locality. In some contexts, it may be the power grid of a single household. In other cases, it may be the power grid of a building such as an apartment complex, an office building, or a supermarket. In yet other variants, a local power grid may include multiple households of a neighbourhood being controlled by a common master unit. Allowing the master unit to control multiple households allows an improved power scheme, as more outlet units and potential electricity generators increase the flexibility of the consumption schedule. In such systems, each microgrid within the local power grid may have a control unit which controls some or all of the units within that microgrid, while the control unit itself is being directed by the master unit of the local power grid.

In the situation where a local power grid comprises multiple households or otherwise comprises multiple sub-grids within the local power grid, the sub-grids having at least two different owners, such sub-grids are considered microgrids within the local power grid. In situations where the local power grid comprises only a single household or is associated with only a single owner, the terms microgrid and local power grid may be used interchangeably. A local power grid may receive electrical power from a number of different sources. Such electricity sources may include a regional power grid, such as the power grid of a state or government, a local electricity generator, and/or an electricity reservoir.

By electricity supply is understood the combined supply of electricity to the master unit regardless of the source. Similarly, an electricity supplier is any component supplying electricity to be distributed by the master unit. This may be from a local electricity generator or by a regional power grid, i.e. electrical power purchased from a power plant.

By an electricity generator is understood any specific unit connected to the local power grid, such as a solar cell or cluster of solar cells or a local turbine for wind or water power. An electricity generator is in the context of this disclosure seen as different from the regional power grid from which electricity may be purchased by the owner of the master unit (even though electricity generators are likely to be used in the generation of electricity for the power supplier from which the user purchases the electricity).

By an outlet unit is understood a connection point for an electricity consumption unit to be connected to. An outlet unit may in some variants be a known plug for passively connecting an electricity consumption unit to a power grid. The outlet unit may in some variants comprise sensors or communication units to relay information to the master unit, but these outlet units are in themselves not involved in creating the priority scheme. An outlet unit transmits electricity allocated to it by the master unit to the electricity consumption unit connected to the power grid via that outlet unit.

It is to be understood that the master unit may receive power from multiple electrical power sources in a local power grid. This includes receiving electricity from a regional power grid, e.g. supplied by a power company or larger infrastructure. It also includes electrical power from known types of electricity generators, such as solar panels, turbines, mills and the like. Electrical power may also be supplied from energy reservoirs, such as batteries, which may be placed locally, either purely for the storage of power, or as an integrated part of equipment; for example, the system may receive electrical power from batteries on board EVs either for redistribution of power or for use in case of emergencies.

According to a further variant, the master unit is adapted for monitoring and distributing electrical power from one or more of the outlet units such that it can redistribute electrical power from the one or more electricity consumption units, electricity generators and/or electricity reservoirs.

According to a further variant, the master unit is adapted for directing electrical power to a local power grid and/or to a regional power grid.

This enables the selling of electrical power from local power generators in situations where more electricity is generated than can be consumed or stored in the local power grid. In addition, this allows the local power grid to assist the regional power grid and keep it stable in case of emergency. For example if the regional power grid is being overloaded, the local power grid may supply electrical energy from reservoirs. Additionally, the master unit may be configured to decrease the energy consumption in the local power grid in response to changes in the regional power grid.

The master unit and modular distributed power consumption system containing such a master unit is suitable for domestic electricity distribution, e.g. at villas, apartment complexes, housing co-operatives and town houses as well as company sites where a master unit may be connected to and manage the distribution of electricity to a plurality of locally connected outlets. The electricity that is being distributed by the system may be locally generated such as from solar cells, or wind or water turbines and other known electricity generators which are directly connected to the microgrid. The electricity consumption units to which the electricity is being distributed may be household appliances such as ovens, heat pumps, washing machines, electrical vehicles, etc. Electricity reservoirs, i.e. batteries, may be connected to the microgrid as well and can take part in the system both as electricity generators and electricity consumption units. In some variants, this may include the batteries incorporated in equipment, e.g. the batteries on board EVs.

The master unit and a modular distributed power consumption system containing such a master unit are likewise suitable for various types of public parking areas such as parking garages and open parking lots for cars or electrical bikes. The system is in particular suitable for charging electricity consumption units which need energy over a span of time, e.g. parked charging for EVs, i.e. charging in a period of time, where the EV is not in use as the user is otherwise engaged and not simply waiting for the charging of the EV. The system is also usable in connection with rapid charging, where the EV is stopped for the specific purpose of charging. The system may also be used in combination for rapid charging of some electricity consumption units and long-term charging other electricity consumption units by giving the rapid charging electricity consumption units priority in the priority scheme.

By the distributed power consumption system being modular is understood that a single master unit is adapted to control a plurality of slave outlet units, while said outlet units may be added to or removed from the distributed power consumption system without directly affecting each other. It is to be understood that there may be a consequent effect of adding or removing an outlet unit, as it may be chosen to change the settings of the master unit to prioritise the connected outlet units; however, this will be controlled by the master unit and not by the connected outlet unit.

Similarly, the distributed power consumption system is adapted to control one or more electricity sources, i.e. regional power grid input and input from local power generators.

The modularity of the distributed power consumption system has the benefit of making it cheap and easy to reconfigure the distributed power consumption system to comprise the number of outlet units needed at a particular site. For example, if the distributed power consumption system is installed in the microgrid of a household, an additional outlet unit may be installed without needing an additional master unit. This further contributes to making the distributed power consumption system cheap, as it is not necessary to have a separate control unit for each outlet unit, as they may all benefit from the computational power of the master unit.

Furthermore, the modularity enables the integration of local electricity generators which are being installed after the master unit is installed.

According to a further variant, the processing unit of the master unit is configured to be in communication with an external server such as a database and/or the internet.

In a preferred variant, the master unit is connected to the internet enabling retrieval of forecast and database information. Furthermore, this enables the communication with and/or the exchange of data with other master units in other distributed power consumption systems. Multiple master units communicating and cooperating to make a collective consumption schedule is considered an electricity consumption network. It is to be understood that within the same microgrid multiple master units may work together in an electricity consumption network for example to provide redundancy, to operate sub-groups in the grid separately or as a master unit with several dependent control units. The electricity consumption network may create a consumption schedule for the charging of EVs within that electricity consumption network. The electricity consumption network may in addition to distributing the charging of EVs also consider the power consumption of other electricity consumption units in the electricity consumption network, such as but not limited to household appliances such as washing machines, the charging and discharging of energy reservoirs such as batteries, and storage or routing of electrical power generated within a microgrid. The electricity consumption network may also enable the functionality of the redistribution of power, e.g. by supplying power from energy reservoirs such as batteries, which may include the batteries within the EVs. Such redistribution may be done with a threshold leaving at least a certain amount of power in each energy reservoir to be used locally. Furthermore, such energy networks may reserve emergency power which can be distributed among the various network sources within the electricity consumption network, e.g. between various households within the electricity consumption network. This enables a suitable amount of reserved emergency power to be available even though each network source may only be able to reserve a fraction of what is required in case of an emergency, e.g. to be delivered into a regional grid for stability.

In a preferred variant the master unit is set up for edge computing as previously described. It is to be understood that edge computing can take place while the master unit is also in communication with external servers via the internet. Exchange of information from other microgrids with similar configurations as the microgrid of the master unit may provide forecasts that can be used as input and further improve the training of the intelligent system. The exchange of information with other microgrids, e.g. the sharing of forecasts and/or charging schedules, further enables the charging schedules determined by a master unit related to a microgrid to account for the charging schedules of one or more other microgrids in a local grid, such that the load of the local grid is distributed between the microgrids, e.g. such that the charging of EVs of different microgrids is sequential.

The distribution of energy consumption within the electricity consumption network may be distributed both with respect to the time at which electrical power is used and with respect to phase on which electrical power is used, as both are relevant to ensure a stable power delivery within a local power grid. While phase may be a relevant factor for some systems and/or devices within a grid, it is to be understood that the present invention is equally applicable to grids and/or devices applying DC electricity.

In some preferred variants the master unit may also be configured to control AC to DC converts within the distributed power consumption system such that the master unit may control from which AC phases electrical power is converted to DC for consumption of various units, such as EVs, requiring DC power within the power consumption system. Thereby, the master unit can use the AC to DC converter when phase load balancing within the grid.

The functionality of communication of multiple distributed power consumption systems, either via the internet or via a local area network, enables the creating of consumption schedules for an increased number of electricity consumption units.

According to a second aspect of the present invention, the above objects and advantages are obtained by:

A modular distributed power consumption system for distributing electrical power between electricity consumption units, the distributed power consumption system comprising, a master unit as previously described5; and a plurality of slave electricity outlet units; and at least one sensor for collecting a sensor dataset and providing the master unit with the sensor dataset; the master unit being adapted to direct at least a subpart of said electrical power to one or more of the electricity outlet units based on a priority scheme determined by an intelligent system.

In a variant a modular distributed power consumption system for distributing electrical power between electricity consumption units, the distributed power consumption system comprising a master unit as previously described and a plurality of slave electricity outlet units. The master unit is adapted for directing at least a subpart of the electrical power to one or more of said electricity outlet units based on a priority scheme.

By the master unit being able to direct at least a subpart of the electrical power to one or more of the electricity outlet units is to be understood that the master unit is configured to be able to direct all or a subpart of the available electrical power in the local grid to any of the outlet units connected with the master unit. Depending on the specifics of the priority scheme and the need of each device connected to an outlet unit, the master unit may at any given time direct electrical power to only some of the plurality of outlet units connected to the master unit; however it is capable of directing electrical power to each of them.

By a sensor for collecting a sensor dataset is understood any detector or sensor which may supply a dataset as input for the training of the intelligent system and/or for the objective function. The sensor may for example be a power meter determining the amount of consumed or generated electrical power, it may be a sensor determining the charge state of a battery or it may be a temperature sensor providing on-site data of the ambient conditions under which electricity is being generated and/or consumed. In preferred variants a plurality of sensors provides the master unit with a plurality of sensor datasets which may be used for the correlation of the sensed data with forecasts and via training of the intelligent system for the further updating of forecasts and the priority scheme.

According to a further embodiment of the second aspect of the invention, the modular distributed power consumption system further comprises one or more electricity reservoirs and/or electricity generators.

By having the master unit direct at least a subpart of the power it has received onward to outlet units based on a priority scheme, the power distribution between the plurality of outlet units may be ensured regardless of the type of electricity consumption units connected to each of the outlet units. A benefit of the power distribution being independent of the type of electricity consumption unit is that the owner of the master unit remains in control of the outlet units regardless of which electricity consumption units are connected to the outlet units. This enables easy adaptation for the introduction of new electricity consumption units, which may in particular be relevant for electric vehicles where various models with their own characteristics may be connected to the system at various times, in particular where the distributed power consumption system is arranged to have public or semi-public outlet units adapted for the charging of electrical vehicles. For example, a household may have both private and semi-public outlet units at different locations, and the distributed power consumption system may enable prioritising the delivery of electrical power to the private outlet units over that to the semi- public outlets, such that guests or strangers buying electricity from the household do not hamper the use of electricity for other electricity consumption units such as ovens or washing machines or the charging of the owner’s electrical vehicles.

By the outlet units being slaves is to be understood that they are controlled by the master unit, delivering power to the electricity consumption units in accordance with the instructions of the master unit, and cannot affect the performance of each other or the master unit. However, it is to be understood that the outlet units can still collect information and communicate it back to the master unit, and that the master unit may react accordingly; however, this is in response to the provided data and not because of the outlet unit controlling the master unit or other of the plurality of outlet units.

According to a further variant, the modular distributed power consumption system comprises at least one submeter adapted for monitoring characteristics of electrical power delivered to and/or from devices connected to said modular distributed consumption system and transmitting such data to said master unit.

By a submeter is understood any detection means for monitoring current, voltage, phase, current waveforms and/or other characteristics of the current conducted through the system. In a preferred variant, such a submeter comprises at least a current sensor, such as a current transducer. Data detected by the submeters provide the master unit with information which may be used to determine necessary adjustments of the directing of electrical power within the system to fulfil the requirement priority scheme, as the data of the submeters enable the monitoring of consumed and/or available electrical power and thus the need for adjustments in response to changes in the consumption and/or generation.

A submeter is a type of sensor for collecting of a sensor dataset.

In a preferred variant of the second aspect of the invention, the master unit further receives data from a master detection means concerning the amount of electrical power delivered from the regional grid to the local grid.

Another object of the present invention is to provide a method for determining the priority schedule used by a master unit of a distributed power consumption system. The terms priority schedule and charging schedule may be used interchangeably throughout the application. According to a third aspect of the present invention, the above objects and advantages are obtained by:

A method of determining a charging schedule for electrical power distribution to and from a plurality of flexible electrical units of a distributed power consumption system carried out by a master unit of the distributed power consumption system, the method comprising the steps of: receiving at least one forecast dataset, receiving at least one sensor dataset from on-site data, on-site training an intelligent system for creating an electrical power forecast, the intelligent system being trained on the at least one forecast dataset and the at least one sensor dataset, obtaining an electrical power forecast for inputting to an objective function to be solved by the master unit, the electrical power forecast comprising an electricity consumption forecast, on-site solving of the objective function determining a charging schedule based on the electrical power forecast, the charging schedule defining action periods during which electrical power is transmitted between the connected flexible electrical units and/or the regional power grid and the fraction of available electrical power which is directed by the master unit to and/or from the flexible electrical units during the action periods.

In a variant of the third aspect is provided a method of determining a consumption schedule for the electrical power distribution to electricity consumption units in a distributed power consumption system carried out by a master unit of the distributed power consumption system, the method comprising the steps of: providing a list of priority factor options, for one or more of the connected electricity consumption units and/or electricity reservoirs, receiving user input ranking one or more priority factors, creating a priority scheme ranking priority factors for all connected electricity consumption units, and/or electricity reservoirs, receiving an electrical power forecast, based on the electrical power forecast and the priority scheme determining a charging schedule, said charging schedule defining usage periods during which electrical power is transmitted to said connected electricity consumption units and/or electricity reservoirs and the fraction of available electrical power which is directed by said master unit to the electrical consumption unit and/or electricity reservoirs during the usage periods.

While the user may define ranking of the priority factors, including information on specific outlet units, electricity consumption units and/or electricity reservoirs, the user may not provide input for every single appliance in the local grid. In such cases, the master unit may have base ranking for outlet units which have not been assigned a specific priority by the user.

In some preferred variants, the method further includes prioritisation of electricity generators within the local grid, e.g. prioritising usage of electricity generated by specific electricity generators. The priority for electricity generators may also include the choice of which source an electricity consumer and/or electricity reservoir receives electrical power from, e.g. to provide load balancing or to store electrical power in electricity reservoirs until a later time. Similarly, the method includes prioritisation factors and ranking for when electrical power is consumed from electricity reservoirs within the local grid.

By a forecast is understood an estimate of future energy generation and/or an estimate of future energy consumption.

A forecast may relate to factors different from electrical power which are relevant to determine other forecasts, e.g. a weather forecast may be relevant to determine electricity consumption as this may increase on cold days and a weather forecast may also be relevant to determining electricity generation as the forecast solar irradiation. A forecast may also be related to market prices of electricity.

An electrical power forecast is a forecast relating to the consumption and/or generation and/or storage of electrical power in the microgrid. By an electrical power forecast being obtained is understood that it may be supplied as a finished forecast to the master unit or that it may be calculated by the master unit based on one or more forecast datasets and one or more sensor datasets and either the trained intelligent system or statistical models implemented in the master unit. By the electrical power forecast being obtained via statistical models is understood that it is determined based on conventional mathematics and physical equations necessary in situations where there is insufficient data for the intelligent system. In a preferred variant at least part of the obtained electrical power forecast originates from the intelligent system.

In a variant one or more of the forecast datasets for electrical power consumption, the electrical power generation and the electrical power storage may be determined by the intelligent system while the remaining parts are supplied from an external source and/or from statistical models implemented calculated on-site by the master unit.

In a yet more preferred variant the priority scheme is determined based on the electrical power forecast derived by the intelligent system trained by the master unit.

By using the electrical power forecast derived by the intelligent system trained by the master unit the power forecast may be improved as the intelligent system is trained on sensor datasets collected on-site, such that the intelligent system may account for the specifics of the microgrid and the site of the microgrid. Should there be insufficient data, e.g. immediately after installation of the master unit or if an on-site sensor is malfunctioning, it is possible to calculate an electrical power forecast using statistics and physics equations separate from the intelligent system.

According to a further embodiment of the third aspect of the invention, the electrical power forecast is obtained by on-site calculation performed by the intelligent system.

According to a variant of the third aspect of the invention the electrical power forecast being obtained by receiving one or more parts of the forecast data from an external source.

Forecast and forecast dataset are used interchangeably throughout the application as it is understood that the forecast comprises data relating to this forecast.

By a forecast dataset is understood data values relating to the future relative to the time of calculation. Thus a forecast dataset may include predicted values, e.g. predicted weather conditions such as temperature or precipitation or best estimates of electrical power consumption or electrical power generation which is determined with some uncertainty. A forecast dataset may also include one or more fixed values for a future condition which has no uncertainty, e.g. market prices may be fixed for a specific period of time into the future, in such an example all of a market price forecast dataset may be deterministic values or some parts for the nearest future may be deterministic values while parts of the market forecast dataset further into the future may be predictions.

By a sensor dataset from on-site data is understood a dataset relating to data measured and/or collected by a sensor placed at the on-site location of the microgrid. The on-site location of the microgrid refers to the site supplied with electricity by the microgrid, e.g. a household. For example on-site data may this relate to the actual electricity usage in the microgrid. On-site data may also relate to other factors affecting the electricity consumption, generation and/or storage, such as temperature, humidity, solar irradiation and/or losses of components in the microgrid. Multiple types of on-site data may be used as input to the master unit.

According to a preferred variant of the third aspect of the invention the electrical power forecast used as input for the objective function to create an optimized charging schedule is processed on-site by the master unit and the objective function receives and stores at least one on-site collected sensor data-set.

By including at lest one on-site collected sensor data-set the forecast and optimized charging schedule is made more accurate to the individual system and microgrid thereby improving the optimization of the charging schedule for the chosen priority factors, e.g. making the charging schedule more efficient in regards to minimizing purchase of electricity from the regional grid or optimizing the load balancing within the microgrid.

According to a further embodiment of the third aspect of the invention, the electrical power forecast comprising an electricity generation forecast and/or an electricity storage forecast.

In a variant of the third aspect, the electrical power forecast comprises an electricity generation forecast.

According to a variant of the third aspect of the invention, the electrical power forecast comprises an electricity consumption forecast.

According to a further embodiment of the third aspect of the invention, the electrical power forecast is based at least in part on a historic forecast such that electricity consumption and/or electricity generation is estimated based on generation and/or consumption from the same system at a prior time. By a historic forecast is understood that the forecast is based on observations of what has previously occurred under the assumption of repeatability. This may for example include calendar-based historic forecasts where data relating to the energy generation and consumption on the same date in the previous year is used to forecast the energy generation and/or consumption of a specific date of the following year. Such historic data may be used directly as for the previous example, or it may be based on averages, e.g. the average generation and/or consumption over a week or month of the previous year or the average consumption on that date over a period of multiple previous years. Historic forecasts may also be based on types of days rather than dates, i.e. having the data averaged based on weekday, weekend day and/or holidays, which may be on different dates in subsequent years. An example of such a historic forecast may be that the estimated generation and consumption is determined for an average weekend of a specific month of previous years, or as another example an average Tuesday of November based on the Tuesdays of the past five years. It is noted that different amounts of time may be used for the average for of generation and consumption, for example it may be possible to determine the generation of power in the grid for ten years, while the consumption of a private household will depend on the occupants of that household, and estimates for new owners may not reach as far back.

According to a further embodiment of the third aspect of the invention, the electrical power forecast is based at least in part on a dynamic forecast, said dynamic forecast being consecutively updated.

In some variants a dynamic forecast is used. By a dynamic forecast is understood a forecast which is continuously updated based on dynamic data, i.e. data which becomes gradually available. Examples of dynamic data is weather forecasts, which may be used to estimate the energy generation based on wind speeds and amount of sunlight as well as estimation of energy consumption based on temperatures. Another example of dynamic data may be the recent consumption of the particular user, e.g. changes in consumption over the last few days or months compared to their average for days with similar temperatures.

According to a further embodiment of the third aspect of the invention, the electrical power forecast has a prediction horizon exceeding the forecast update rate of at least one external forecast dataset, such as the prediction horizon being at least twenty-four hours

According to a variant of the third aspect of the invention electrical power forecast having a prediction horizon exceeding one hour such as being at least twenty-four hours.

In a variant the charging schedule is updated at regular intervals such that the charging schedule is calculated for a twenty-four hour period following the last update. Preferably these regular intervals are determined based on at least one update rate of an external forecast dataset.

In a variant the electrical power forecast is determined for the following at least twenty- four hours. In a preferred variant at least part of the electrical power forecast is determined for the following at least twenty-four hours by the locally trained intelligent system of the master unit.

The duration for which the electrical power forecast is determined is also called the prediction horizon. In a preferred variant the charging schedule is determined for the duration of the prediction horizon. In a preferred variant the prediction horizon exceeds one hour. In a preferred variant the prediction horizon is at least twenty-four hours. In a preferred variant the prediction horizon is at least forty-eight hours. For a preferred variant the prediction horizon is five days.

The granularity of the data input to the master unit and the training of the intelligent system on both forecasts and sensor data collected on-site such that the forecasts may be adapted to the on-site conditions improves the accuracy of the electrical power forecast and hence also for the extension of the prediction horizon.

By determining an electrical power forecast with a prediction horizon exceeding one hour, preferably at least twenty-four hours, it is possible to determine a charging schedule for the same timeframe. This enables a more optimized charging schedule as it may account for future events rather than relying on the present situation of a given moment or for the near future. For example, having an electrical power forecast and corresponding charging schedule of twenty-four hours allows planning of the most cost efficient times of charging an electricity reservoir and selling to the regional power grid, such that electrical power may be sold at a higher price rather than deferring selling electrical power until an electricity reservoir has been filled.

According to a further embodiment of the third aspect of the invention, the intelligent system receives user input of an operation mode such that user expectations of changes are taken into consideration.

By an operation mode is understood a preset of one or more priority factors for which the objective function of the systems is optimized. Hence, the user may influence the optimization without needing to choose specific priority factors and can simply choose overall preferences, such as lowest cost of operation of the microgrid, steady consumption or green operation with low CO2 footprint.

In a variant of the third aspect, the electrical power forecast is based at least in part on a user forecast such that user expectations of changes are taken into consideration.

User forecasts may also be included as a type of dynamic data, wherein users may input their own expectations of change in their usage, e.g. if the users know they will be travelling and thus have less consumption in the near future, or if the owner is a mall, they may expect increased usage during a sale.

In a preferred variant, the dynamic forecast is updated consecutively. The update of the dynamic forecast may for example take place in predefined intervals, e.g. once a day, or it may take place every time a new input is received, e.g. when dynamic data is updated and transferred to the master unit. High frequency of update of the dynamic forecasts, e.g. multiple times a day, such as every hour, is preferred, as more data enables a more precise forecast and thus better energy balancing.

In preferred variants, a combination of historic and dynamic forecasts is used. In such preferred variants, the historic forecasts may be used as a baseline, which is modified based on the dynamic data in that the generation and/or usage forecasts are adjusted based on the dynamic forecasts deviation from the historic forecast. It is to be understood that in such systems both types of forecasts may be used some of the time, while at other times only a single type of forecast may be used. For example, in a system retrieving dynamic forecast data via an internet connection, in the case of a loss of such an internet connection, the priority scheme may be based on historic forecast data - preferably in combination with user priority factors - until updated dynamic forecast data is once again available and can also be taken into account. This is in particular possible, as the historic forecast may in principle be made arbitrarily far in advance, e.g. it can be made for any future date and or be made to be repeated indefinitely within a predetermined period of e.g. a year. The dynamic forecast cannot be made as far in advance as the historic forecast. The dynamic forecast depends on the dynamic data and thus cannot be made before the dynamic data is available and cannot be made/updated further into the future than the dynamic data extends. For example, the dynamic data may itself be a forecast, such as a weather forecast, and the dynamic forecast cannot be made before the weather forecast is available, e.g. from a public database, and cannot extend further than the weather forecast on which it is based, e.g. a week into the future.

According to a further embodiment of the third aspect of the invention, the sensor dataset is received with update rate exceeding the update rate of said forecast dataset.

By an update rate is understood the frequency with which the master unit receives a new dataset. For example a master unit may receive an external forecast dataset relating to a weather forecast with one time interval, e.g. hourly. In preferred variants the master unit receives multiple forecast datasets which may be received with the same update rate or with different update rates. For example an weather forecast may be updated daily while a forecast of electricity prices may be updated hourly or both such forecasts may be updated multiple times an hour. The update rate of a sensor dataset may be the rate with which new data is collected and/or transmitted to the master unit. For example, an electricity meter may be configured to take a measurement and transmit the updated dataset to the master unit every ten seconds.

In a preferred variant the master unit will receive a plurality of forecast datasets and a plurality of sensor datasets. In a preferred variant the master unit will receive a plurality of forecast datasets and a plurality of sensor datasets and the update rate of all sensor datasets will exceed the update rate of all external forecast datasets.

Having a higher update rate of the sensor datasets allows the intelligent system of the master unit to adapt the forecast data with respect the detected on-site situation and allows for correction of the calculated electrical power forecast and/or the priority scheme. The longest expected update rate for an external forecast may set the calculation time for updating of the calculated electrical power forecast and/or the priority scheme.

If connection is lost to an external server such that one or more external forecast datasets are not updated, the intelligent system may reuse the previously supplied dataset treating it as if the following forecast dataset is identical to the previously supplied and/or basing any missing time periods on historical forecasts.

According to a further embodiment of the third aspect of the invention, the intelligent system is trained in an accumulative manner, such that previous data from a dynamically updated forecast dataset affects the resulting charging schedule.

By training the intelligent system accumulatively, previous correlations between the sensor datasets and forecast dataset may be taken into account in the determination in the adaptation of subsequent forecast datasets.

According to a further embodiment of the third aspect of the invention, the intelligent system being trained for stochastic optimization such that a distribution of sensor dataset is used as the input for the objective function.

According to a further embodiment of the third aspect of the invention, at least one forecast dataset comprises a dynamic forecast of inverter losses.

Microgrids tend to comprise a plurality of inverters each of which will introduce losses when active. The losses introduced by each inverter is independent of the losses of the other inverters and the losses may vary over time depending on how they are used, e.g. how frequent they are activated or how much current is run through them. Thus, accounting for how the losses of the inverters of the microgrid dynamically varies over time enables the creation of an improved charging schedule, e.g. by minimizing the loss of the inverters and/or minimizing the use of the inverters with the highest losses.

By the forecast of inverter losses being dynamic is understood that the inverter loss forecast is determined for each individual time unit within the duration of the forecast, i.e. for each time unit up unto the prediction horizon. The forecast is updated periodically with the most recently collected sensor data used to determine the forecast, i.e. the power flow on each direction of the inverter. The higher the sampling frequency the better the forecast, the sampling frequency may e.g. be 10 Hz or 1 Hz.

In a preferred variant the at least one forecast dataset comprises a dynamic forecast of inverter losses both for conversion from AC-to-DC and for conversion from DC-to-AC.

The losses of the inverter varies depending on whether it is alternating current being converted to direct current or direct current being converted to alternating current, this may be described as the two directions of the inverter. Hence, the optimization of the charging schedule will be improved by taking into account the difference in the losses depending on the operation of the inverter such that the losses are accounted in both directions.

In a preferred variant the at least one forecast dataset comprises a dynamic forecast of inverter losses for inverters connected to electricity reservoirs both for conversion from AC-to-DC and for conversion from DC-to-AC and at least one sensor dataset comprises the power flow on both sides of the inverter.

According to a further variant of the third aspect of the invention, the method comprises performing a forecast evaluation step wherein the forecast dataset is compared to the sensor dataset and the next forecast dataset is changed to a default if the difference exceeds a predefined threshold.

Performing a forecast evaluation steps minimizes the risk of unintended results due to unforeseen changes of the on-site conditions. For example, wear of or dirt on an electricity generator such as a solar cell may cause a lower actual electrical power generation than the predicted generation or malfunction of an electricity consumption unit may cause it to draw more electrical power than expected.

In a preferred variant a warning is sent to a user if the predefined threshold is exceeded such that the user may determine the cause of the issue ro schedule maintenance. Furthermore, should an inconsistency be detected due to malfunctioning of a sensor this may also be determined.

Reverting to a default forecast setting will in this case ensure that the master unit may continue to function until the issue has been solved even though the resulting charging schedule may be less optimal than for the fully functional system. SHORT LIST OF THE DRAWINGS

In the following, examples of embodiments are described according to the invention, where:

Fig. 1 illustrates how the distributed power consumption system may be part of the local grid of a private household.

Fig. 2 illustrates how the distributed power consumption system may be integrated as part of the local grid of a neighbourhood in a regional power grid.

Figs. 3 is a schematic illustration of the electrical and communication connections between components of a distributed power consumption system.

Fig. 4 is a schematic illustration of data input and output for the trained intelligent system of the master unit.

Fig. 5 illustrates the concept of different timing of the update of forecast datasets and sensor datasets.

Fig. 6 illustrates an example of electricity generation, electricity consumption and electricity prices.

DETAILED DESCRIPTION OF THE DRAWINGS

The invention will now be explained in more detail below by means of examples with reference to the accompanying drawings.

The invention may, however, be embodied in different forms than depicted below, and should not be construed as limited to any examples set forth herein. Rather, any examples are provided so that the disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Like reference numerals refer to like elements throughout. Like elements will, thus, not be described in detail with respect to the description of each figure.

Fig. 1 illustrates one use case for a distributed power consumption system 10 for which it is connected to a local grid in the form of a private household. While the illustration shows a domestic setting of a single household, it is to be understood that the distributed power consumption system 10 may be used in various other settings, as discussed below.

The illustrated embodiment of the distributed power consumption system 10 comprises a plurality of outlet units 100 accessible outside the house. Their placement are examples of different usage options, and other systems may have few, more and/or differently located outlet units 100. In preferred embodiments the distributed power consumption system 10 will further include a plurality of outlet units inside the house, such outlet units inside the house may be differently constructed than outlet units used outside the house, as they may not need the same functionalities, e.g. they may not need to be as durable to weather conditions, and/or they may not need the same detection capabilities for identifying what is connected to the outlet unit. Outlet units 100 located on the outside of a building may be dedicated for transmitting electrical power to/from electricity consumption units 40 in the form of electrical vehicles such as an electrical car privately owned by the household, for the charging of electrical vehicles in the form of electrical bicycles having a different capacity than cars, and/or for semi-public outlets accessible to electrical vehicles external to the household, e.g. guests or owners of electrical vehicles interested in buying electricity from the microgrid of the household, while also being usable by the owner.

The distributed power consumption system 10 may be incorporated in other types of grids, such as the microgrid of a company where the distributed power consumption system 10 distributes charging of employee and costumer vehicles while they are parked. Other examples include connecting the distributed power consumption system 10 to microgrids comprising multiple households, either by connecting multiple houses in a neighbourhood or by connecting to the grid of an apartment complex. In yet other situations, the distributed power consumption system 10 may be owned by a store or a group of stores providing charging opportunities for both customers and employees.

A master unit 50 (see Fig. 3) is part of the distributed power consumption system adapted to distribute electrical power received from various electricity generators 20, such as solar cells 21 or wind turbines 22 or from a regional power grid 15 (illustrated in Fig. 1 as the point of entry into the microgrid via a master relay) delivering power generated off the site. Microgrids comprising local electricity generators 20 may also preferably comprise an electrical transformer 29 of their own. The master unit 50 may also be adapted to control the supply of electrical power to and from electricity reservoirs 30, such as batteries, which may be separate or integrated parts of electricity consumption units, e.g. batteries on board electrical vehicles. In preferable variants of the distributed power consumption system, the master unit 50 is located inside a building or is otherwise shielded, e.g. from weather or tampering. The master unit 50 is connected to all outlet units 100 of the distributed power consumption system and enables the distribution of electricity to the slave outlet units 100.

In some embodiments, at least one of the outlet units 100 comprises identifier means, such as a receiver/transmitter for sharing information with the master unit regarding the electricity consumption unit connected to that outlet unit.

In a preferred embodiment the master unit 50 comprises or receives data from a master detection meter, while one or more of the outlet units comprise a submeter. In such a preferred embodiment the submeter may be used to determine the fraction of the electricity determined by the master detection meter which has been delivered to a specific electricity consumption unit 40 in a predetermined period of time such that the relative consumption of the various electricity consumption units 40 may be determined. In some embodiments, each of the outlet units 100 may comprise an electricity submeter. In other embodiments, only some outlet units 100 comprise an electricity submeter, such as outdoors outlet units 100 intended for use with electrical vehicles. By a detection meter, i.e. a master detection meter or a submeter, is understood any detection means suitable for monitoring current, voltage, phase, current waveforms and/or other characteristics of the current conducted through the system.

In a preferred embodiment, the meters of the system comprise at least current sensors. Such current sensors may enable faster monitoring of the current in the system than the regular central electricity-meter of a grid can provide, and thus provides the master unit 50 with the necessary information to adjust the electrical power consumption of the connected outlet units 100, consumption units 40, electricity reservoirs, and/or electricity generators 20 and/or to redirect electrical power.

In various embodiments of the system, meters and/or submeters may be mounted at various locations on the grid. For example, a master electricity meter may be arranged in the power line entering the local or microgrid, to monitor the delivered current. Submeters, such as or including current sensors, may be integrated in outlet units 100, consumption units 40, electricity reservoirs, and/or electricity generators 20 and/or they may be mounted within the electrical conduits, such as cables, connecting such appliances to the grid controlled by the master unit 50.

Fig. 2 shows a regional power grid 15, illustrated as an overhead power line, receiving electrical power from a number of electricity generators 20, such as wind turbines 22 and nuclear power plants 23. The electrical power of the regional power grid is passed via transformers 29 to regional power grids 17, which may be controlled by distributed power consumption systems adapted to distribute the electrical power to electricity consumption units within the regional power grid 17.

The solid arrows in Fig. 2 indicate the direction of transmission of electrical power. It is noted that electrical power is delivered from electricity generators 20 to the regional power grid 15. This may also be the case for electricity generators 20 within the local power grids 17. Hence, if a household in a local power grid 17 comprises an electricity generator 20, such as a wind turbine or solar cells, electricity generated by these electricity generators 20 may be used in the microgrid of the owner of the electricity generator 20, e.g. the particular household, or it may be delivered back to the regional power grid 15 to be redistributed elsewhere, e.g. to different microgrids 19 within the local power grid 17.

By controlling multiple microgrids 19 within a local power grid 17 using the same distributed power consumption system, it is possible to ensure a more even power consumption distribution. Furthermore, it enables the local redistribution of locally generated electrical power when it cannot be used within the microgrid 19 where it is generated. It can be redistributed within the local power grid 17 rather than being transported further to a more distant microgrid 19. This in turn decreases the loss of electrical power and increases the efficiency of the local power grid. Furthermore, the collective distribution within the local power grid 17 further enables the possibility of simultaneously delivering more electrical power back to the regional power grid 15 with a faster response time than if only single microgrids were connected. This is in particular useful in case of a consumption spike, which might otherwise lead to overloading of the regional power grid and potentially cause power outages.

Figs. 3 schematically illustrates how various components of a distributed power consumption system 10 may be connected to receive and deliver electrical power as well as with respect to information exchange. Solid arrows indicate a connection for transmitting electrical power. Dashed arrows indicate an exchange of data used for determining the distribution of the electrical power.

The distributed power consumption system is configured to receive electrical power from a regional power grid 15 such as the grid for a state or a government controlled grid. In preferred embodiments as the one illustrated, the master unit is also configured such that it may direct electrical power back into the regional power grid 15 if a surplus of electrical power is generated by the local microgrid, i.e. more power is generated than can be consumed by electricity consumption units 40 of the microgrid. In other embodiments of the distributed power consumption system, the capability to direct electrical power back into the regional power grid 15 may be foregone, e.g. if the microgrid comprises no electricity generators.

The electrical power is delivered to the distributed power consumption system via a main fuse box 45. In a preferred embodiment, the distributed power consumption system is configured to receive electrical power from multiple sources, e.g. both from a regional power grid 15 and from local electricity generators 20, such as solar cells 21 or turbines, and electricity reservoirs 30 such as batteries. In some preferred embodiments, the master unit 50 is also configured to receive electrical power from the electricity reservoirs in the form of batteries of electrical vehicles via outlet units 100, which are the connection points of one or more the electricity consumption units 40 of the microgrid controlled by the master unit 50.

In some embodiments, the master unit 50 may be integrated in the main fuse box 45 such that it is configured to receive electrical power and directly distribute it to electricity consumption units 40 in the distributed power consumption system.

In alternative embodiments as illustrated in Fig. 3, the master unit 50 may be external to the main fuse box, the master unit 50 being configured to control the distribution of electrical power received by the main fuse box 45 to the electricity consumption units of the microgrid.

In embodiments of the invention wherein multiple microgrids of a local power grid are being controlled by the same master unit, that master unit may communicate with multiple fuse boxes receiving electrical power for each microgrid. The main fuse box 45 may comprise a plurality of groups. Further, the main fuse box 45 may comprise CS grid or standalone CS for exchanging data between the processing unit of the master unit 50 and the fuse box 45.

Within a microgrid there may also be additional control units 51 , which based on directions from the master unit 50 control the distribution of electricity to a specific subpart of the microgrid, e.g. outlet units 100 intended in particular for the charging of electrical vehicles, while other electricity consumption units are controlled by other control units or directly by the master unit 50.

Fig. 3 shows an example embodiment of a distributed power consumption system 10 wherein two outlet units 100 are serially connected to a control unit 51 receiving directions from a master unit 50. The master unit 50 determines the distribution of electrical power to the control unit 51 as well as directly to electricity consumption units 40, such as heat pumps, ovens, washing machines, and/or lighting. As illustrated, when in serial configuration, a first outlet unit 100 may be connected directly to the control unit 51 , while a second outlet unit 100 is connected to the first outlet unit, such that the connection between the second outlet unit 100 and the control unit 51 is mediated via the control unit 51. In some embodiments, further outlet units may be connected to the second outlet unit or directly to the control unit, as it is possible to have both serial and parallel connections within the same microgrid. Serial connection may be employed both for the electrical connection and the communication connection, or the outlet units may be differently connected, e.g. such that the electricity is serially connected while the communication is parallel.

The choice of electrical connections and communication connections for a particular installation embodiment depends on the physical layout of the location of the distributed power consumption system. In a large area and/or where a significant number of outlet units 100 is to be mounted and/or other connections made, serial connection may be preferrable to avoid installation of many and/or long cables, as the distance between neighbouring outlet units 100 may be smaller than the distance between an outlet unit 100 and the master unit 50. In locations with few outlet units, for example with five or fewer outlet units, it is possible that the distance between the master unit and each outlet unit is small enough so that it is preferable to use parallel connection, thereby minimising the number of connection points on each connection line. As previously mentioned, in some embodiments a combination of parallel and serial connection will be the preferable solution to accommodate the layout and mounting of the outlet units 100 relative to each other and the master unit 50. This may for example be the case if the distributed power consumption system controls a local power grid with several microgrids; for example each microgrid may be connected to a master unit 50 and the main regional grid 15 in parallel, while the outlet units, electricity consumption units and other connection of each microgrid may then be serially connected to a fuse box and a control unit of each microgrid.

The system of Fig. 3 is only an example embodiment, and microgrids with various numbers of outlet units and directly connected electricity consumption units or multiple electricity reservoirs 30 are also within the scope of the invention.

In other preferred embodiments of the distributed power consumption system 10, similar to that illustrated in Fig. 3, the master unit 50 and the control unit 51 may be a single integrated unit. In such variants the outlet units 100 may receive electrical power directly from the fuse box, while the master unit 50 with the integrated capabilities of a control unit will communicate with the outlet units 100 and control the directing of electrical power to and from those outlet units 100. As such, it is to be understood that as long as the capabilities of the master unit 50 and the control unit 51 are present in the distributed power consumption system 10, they may be separate or integrated components. In some embodiments capabilities of the control unit may also be in part integrated within the master unit 50 and in part within the outlet units 100.

In addition to the master unit 50, the distributed power consumption system may comprise various additional detectors, controllers and/or gateways specific to the presence of electricity generators 20, electricity reservoirs 30 and/or electricity consumption units 40 connected in the microgrid. For example, if a consumption unit 40 in the form of a heat pump is present, a heat signal gateway 65 may be included in the microgrid.

GROUPS AND TEMPERATURE SENSORS

In some preferred embodiments, the distributed power consumption system for use in grids with AC electricity further comprises one or more inverters 61 and/or one or more switches 63. The inverters 61 and/or one or more switches 63 determine to which phases the electricity generators 20 and/or electricity reservoirs 30 deliver electrical power. The inverters 61 and/or one or more switches 63 may also determine from which phases electrical power is delivered to electricity reservoirs 30 if such are connected to the grid. This allows the phase at which power is being delivered to various consumption units 40 of the microgrid to be a priority factor when setting up the priority scheme and consumption schedule.

INVERTER DETAIL

In some preferred embodiments, the distributed power consumption system for use in grids with AC electricity may comprise one or more AC-to-DC converters such that the delivery of electrical power of DC consumption units 40 may be further adjusted via such converters. For example, phase load balancing may be controlled by determining from which phases the AC power is converted before it is delivered to the consumption unit 40 requiring DC electrical power.

In other grids, where electricity is delivered as DC and/or where the devices are configured for DC consumption, such inverters and/or switches for controlling the phase may be foregone.

Some outlet units 100 and/or electricity consumption units 40 may also include sensors, transmitters and receivers for collecting and exchanging data with the master unit 50 used in the determination of how electricity is to be distributed within the microgrid.

One or more of the processing components of a distributed power consumption system, such as the master unit, control unit or switch may in preferred embodiments be configured to communicate with the cloud 55, i.e. with external servers being specific to the local power grid and/or with databases of the internet. Further, the master unit may receive user input via the cloud, e.g. input through a dashboard, user login to the system and/or online service.

The distributed power consumption system further comprises one or more detection meters 110, which may be submeters 110. The detection meters and/or submeters 110 may be integrated in the fuse box 45, the master unit 50 and/or in outlet units 100, consumption units 40, electricity reservoirs, and/or electricity generators 20. Alternatively, they may be arranged as separate components in the electrical connection pathways connecting said various appliances. The detection meter and submeters 110 transmit data regarding various electrical characteristics measured, such as detected current, to the master unit 50. The master detection meter and the submeters may provide the master unit 50 with data which the master device may in turn use to determine the need for adjustments in the direction of electrical power to and from the connected components in the grid. In a preferred embodiment, the master detection meter and the submeters comprise current transducers for continuously monitoring current to enable fast adjustment of the consumed current levels. The master detection meter and the submeters may also detect the phase to enable adjustment of the current based on phase load.

The master unit 50 comprises a processing unit for processing user input, determining a priority scheme based on that input, and controlling the distribution of the power based on the determined priority scheme. The master unit 50 may further include a database and/or means for communicating with the cloud 55 comprising an external database for storing identifier codes for particular electricity consumption units, determining the fraction of electricity delivered to a particular electricity consumption unit 40 and/or outlet unit 100 in a certain time frame. The master unit may further be adapted for transmitting information to the owner of the master unit or control units within a local power grid to inform the settings of the distributed power consumption systems, connected unit, amount of electricity locally generated, or similar information regarding the past, current or predicted states of the distributed power consumption system.

Directing of power to outlet units 100 may be undertaken by a fuse box 45 comprising a relay. This fuse box may be external and controlled by the master unit 50, or it may be an integrated part of the master unit 50. In some embodiments, the fuse box 45 comprises mechanical switches.

The master unit 50 distributes the power between the outlet units 100, electricity consumption units 40 and/or electricity reservoirs 30 in accordance with a priority scheme. The priority scheme determines how the electrical power is to be distributed between these outlet units 100, electricity consumption units 40 and/or electricity reservoirs 30, i.e. the priority scheme indicates which of these takes highest priority during which circumstances. For example, some outlet units 100 may be prioritised at particular times of the day. As another example, if a limited amount of electrical power is assigned to a group of outlet units 100, the priority scheme may indicate if such power should be evenly distributed between those outlet units 100 or if power must be directed to an electricity consumption unit connected to that outlet unit before electrical power is directed to other outlet units.

Based on the priority scheme, the processing unit in the master unit determines a consumption schedule for the connected electricity consumption units, determining when and how big a fraction of the available electrical power is directed to each outlet unit of the distributed power consumption system.

As previously noted, the master unit may determine a consumption schedule allotting an amount of electrical power to be directed to a group of outlet units and/or electricity consumption units, wherein a control unit is relaying the electrical power between the outlet units and/or electricity consumption units within that group. For example, such a group may be outlet units adapted for the connection of electrical vehicles. The master unit determines how much electrical power is dedicated to the charging of electrical vehicles within the microgrid, while the control unit determines a consumption schedule for the electrical vehicles connected to the outlet units at a particular time. As such, the control unit may be considered a secondary master unit or a slave master unit within the system.

In other preferred variants, the control unit may be foregone and the master unit may directly determine both how much electrical power is delivered to each outlet unit or is drawn from each outlet unit, and it determines the timing of electrical power usage and distribution between each electrical consumption units, such that the entirety of the local power grid may be controlled by the master unit.

For example, a priority scheme may lead to a consumption schedule where connected EVs are charged sequentially, i.e. a first EV connected at a first outlet unit 100 is charged in a specific time interval, and at the end of this interval the charging of the first EV ends whereafter the charging of a second EV connected to a second outlet unit 100’ begins.

Another example of a priority scheme is parallel charging, where a first EV connected to a first outlet unit 100 and a second EV connected to a second outlet unit 100’ is charging in the same or at least an overlapping time interval, each EV receiving a fraction of the total available electrical power. It is to be understood that consumption schedules may also include a combination of sequential and parallel charging. In some embodiments, the consumption schedule will be constructed such that a first group of outlet units and/or electricity consumption units will be prioritised such that they will continuously be provided with electrical power such that these units will charge parallel with any other groups of outlet units or electricity consumption units in the local power grid, while a second group of outlet units may be set up to charge sequentially with respect to each other and in parallel with the first group of outlet units. For example, lighting, heating and kitchen appliances may be a first group which has continuous priority, while the second group may comprise EVs and washing machines which may be restricted to receive electrical power only under specific conditions and/or to only receive electrical power in series, such that none of those electricity consumption units receives electrical power at the same time.

The priority scheme may be based on a number of priority factors, which may be set by the owner of the distributed power consumption system 10. In a preferred variant, the owner may set and/or adjust priority factors via an app on a smartphone, via an online browser login, and/or directly on a user interface on the master unit 50 and/or an interface on the outlet units 100. Priority factors may be set individually for various outlet units. In such cases they may be set based on which electricity consumption units are assumed to be connected to those outlet units.

Priority factors may include, but are not limited to:

The primary time period in which electrical power usage is to take place, e.g. at night and/or at other times where the owner assumes that the power consumption needed for other appliances is minimal. Time of electrical power consumption may also include intended load shifting, e.g. by increasing electrical heating at times where consumption is otherwise low and increasing heating above a normal level to decrease the necessary heating at later times where the electricity consumption is higher due to other electricity consumption units requiring electrical power.

Time periods in which no electrical power is led to a particular outlet unit 100, e.g. periods of time in which no charging of EV may takes place or in which a washing machine may not run. This may for example be relevant if it is known that the microgrid will be strained by other loads in that period of time, e.g. when the owner usually cooks dinner using numerous kitchen appliances. Such periods may also indicate times wherein the owners know that they themselves are not interested in using electrical power on a particular outlet unit, e.g. an outdoors outlet unit for the charging of EVs. By prohibiting electrical power to be directed to the outlet unit in a specific period of time, e.g. while the owner is not scheduled to be home, unintended charging may be avoided and provide a safeguard against electricity theft.

Price of the electricity delivered by the regional power grid 15. For example, a threshold may be set such that one or more of the outlet units receive electrical power only when the price of electricity is below that threshold. Such a threshold may for example be an absolute value or it may be relative to the average cost of electricity within a day, or it may be a predetermined number of hours of the day in which the price is the lowest of that day. An example of an alternative way of using price as a priority factor is to consume electricity only for a predetermined maximum cost within a period of time, such as a day. Yet another way of using price as a priority factor is to purchase electricity above the needed consumption when the price is below a certain threshold. Such purchased electrical power may be stored by local electricity reservoirs such as batteries and be used by electricity consumption units or sold back to the regional power grid at a later point.

- Availability of green electricity in the grid, e.g. using electrical power at specific outlet units only when a certain fraction of the electricity in the grid is delivered by renewable electricity sources.

Using electrical power at one or more outlet units only when electrical power generated in the local power grid exceeds the electrical power consumed, e.g. for the charging of electricity reservoirs such as batteries.

Time at which an electricity consumption unit has at the latest finished charging to a predetermined amount, e.g. the time at which an EV is fully charged, or the time at which an EV is charged to at least 80% of its battery capacity, or the time at which a dishwasher has finished a specific cleaning routine.

Even distribution of load across the phases in the case of an AC grid, such that consumption may be switched from one phase to another, or such that electricity consumption units on a phase experiencing a low load are prioritised. In grids comprising one or more AC-to-DC converters the phase load may also be balanced by prioritising conversion of electrical power from phases experiencing a lower load at that time.

Maintaining electrical consumption within a local grid within a predetermined range to achieve an approximately constant electricity consumption. Maintaining a constant level of electricity consumption by adjusting the load, e.g. by usage of electricity reservoirs, prevents fluctuations which may cause instability in the power grid and increase the risk of power outages.

Hierarchy of outlet units, wherein one or more outlet units are given priority over other outlet units of the distributed power consumption system 10, e.g. such that the electrical power usage period of the higher priority outlet unit begins earlier, or a larger fraction of the available electrical power is directed to the high priority outlet unit.

Hierarchy of electricity consumption units for embodiments in which the electricity consumption units comprise means for enabling the master unit to identify the particular electricity consumption unit connected to an outlet unit, such that one or more electricity consumption units may be given priority with respect to other electricity consumption units connected to the distributed power consumption system regardless of which outlet unit the electricity consumption unit is connected to.

The priority factors are thus intended goals of the consumption schedule. The distributed power consumption system may be set up with a set of priority factors as a default and/or the user may set specific priority factors for their distributed power consumption system. The priority factors may be weighted in relation to each other, such that some cannot be ignored while others may be set aside if it is not possible to fulfil all priority factors.

The priority scheme determines the relation between the various units of the distributed power consumption system, i.e. which outlet units and/or electricity consumption units should be provided with electrical power first and/or with the largest fraction of electrical power in case full charging cannot be ensured while also fulfilling the priority factors.

The master unit will determine a priority scheme based on the priority factors and the priority scheme.

The master unit will determine a priority scheme based on an objective function which optimises for the specified priority factors. In some preferred embodiments various standard set of priority factors are presented to the user as modes of operation which determines what the objective function will optimize for. In some preferred embodiments the user may be presented with a pareto frontier such that the user may determine which specific priority scheme the operation of the master unit will be based on. In preferred embodiments, the priority factors are weighted compared to each other and/or ordered in a hierarchy such that the consumption schedule is created such that some of the priority factors must be fulfilled, while electrical power may be supplied to one or more of the outlet units 100 even in time periods when not all of the factors are fulfilled.

In some preferred embodiment the consumption schedule is constructed such that the starting time of an electrical power usage session is determined such that as many of the priority factors as required and/or as possible are fulfilled in an uninterrupted period of electrical power usage, following the starting time of the electrical power usage session. In such consumption schedules, the supply of electrical power to each electricity consumption unit will then take place in an uninterrupted period of time, following the starting time of the electrical power usage session. In some embodiments multiple starting times may instead be set for each electricity consumption unit, while a minimum period of time for each electrical power usage period is specified as a priority factor, e.g. each electrical power usage session on a specific outlet unit may be two hours.

The terms usage period and action period may be used interchangeably.

By the starting time of an electrical power usage session is understood the time at which electricity is first supplied to an electricity consumption unit connected to an outlet unit 100, since supply of electricity to that outlet unit was last terminated. By an electrical power consumption session is understood a period where electrical power is continuously supplied to an electricity consumption unit 40 via an outlet unit 100 to which that electricity consumption unit 40 is connected.

The priority scheme may dynamically change in response to input factors such as the number of electricity consumption units 40 connected to the distributed power consumption system 10, price and/or availability of electrical power in the microgrid, time of day, and/or charging level of electricity reservoirs 30 within the distributed power consumption system 10. The processing unit of the master unit 50 may be adapted to receive outside input such as time, date, and forecasts such as weather and price forecasts. In some embodiments the master unit 50 may receive such information directly from a smart relay. In other embodiments the master unit 50 may obtain such information from external databases, e.g. from the cloud using Wi-Fi. A first example priority scheme for a distributed power consumption system with three outlet units 100 may be to statically direct 60% of the electrical power to a first outlet unit, directing 30% of the electrical power to the second outlet unit, and directing 10% of the electrical power to the third outlet unit.

In a second example priority scheme, the same base distribution of power may be used as for the first example priority scheme, while the outlet units are simultaneously arranged hierarchically such that the first outlet unit has the highest priority, followed by the second outlet unit, which in turn is followed by the third outlet unit. If no power is needed at one of the outlet units, e.g. no electricity consumption unit is connected or the connected electricity consumption unit is fully charged, then the fraction of electrical power which would normally be dedicated to that outlet unit 100 will instead be directed to the outlet unit 100’ highest in the hierarchy where power is needed by a connected electricity consumption unit.

In a third example priority scheme, the electrical power may be distribution with respect to time such that a first outlet unit receives 100% of the power until a certain charge level of the battery of a connected electricity consumption unit, such as the battery of an EV, has been achieved, whereafter 100% of the power is directed to a second outlet unit until the battery of the electricity consumption unit 40 connected to the second outlet unit has reached a certain charge level, whereafter electrical power is directed to the next outlet unit. In such a priority scheme, the battery charge level required before power is directed to another outlet unit may differ between each outlet unit and/or electricity consumption unit.

In a fourth example priority scheme, the fraction directed to each outlet unit may depend on various input factors and conditions. For example, initially 70% of the electrical power may be directed to a first outlet unit, while 30% of the electrical power is directed to another outlet unit, until a battery charge level of the electricity consumption unit connected to the first outlet unit has been reached. Once this charge level has been achieved, the distribution may change such that for example 50% of the electrical power is directed to each of the first and the second outlet units. Conditions of the charge level of the battery of the electricity consumption unit may for example be a percentage of full capacity, or in the case of an EV a minimum number of kilometres which can be driven on the charged amount. Other conditions may for example include at what time charging may at the latest commence, e.g. such that all other conditions set for various priority factors are disregarded in the time frame of 3:00-6:30 in the morning to ensure that a minimum amount of electric power is charged to an electricity consumption unit before the user needs it, e.g. to ensure that an EV battery is sufficiently charged before the user needs the EV for commuting.

These are only some example priority schemes, and it is to be understood that many other configurations are possible within the scope of the invention.

The previously described examples of priority schemes may be modified by additional conditions for when electrical power usage takes place, i.e. the priority schedules may be based on a priority scheme as described above in combination with one or more priority factors, such that no electrical power usage is taking place on one or more of the outlet units unless the price of electricity is below a certain threshold or unless the electricity consumption in other parts of the microgrid is below a certain threshold. Such priority factors in the form of conditions to be satisfied may be applied to only some of the outlet units, such that a first outlet unit will be supplied with electrical power regardless of these conditions, while electrical power usage will only commence at other outlet units once those conditions are met.

Fig. 4 illustrates the concept of the master unit 50 receiving inputs for the training of an intelligent system and optimising of the objective function. The processing unit of the master unit is adapted for local processing of the data including the training of the intelligent system, the solving of the objective function and determining of the priority scheme, i.e. this is done by edge computing. The on-site processing of the data enables the rapid processing of vast amounts of data without the need for connection to an external server. Hence, should an internet connection be temporarily down the system will continue to function. Furthermore there is a benefit to the limit of the data, i.e. the data handled by the processing unit is solely related to the grid controlled by the master unit. In addition edge computing also provides further data security as the data does not become available to third parties.

The master unit 50 comprises a processing unit training and running an intelligent system which is adapted for forecasting consumption and production of electrical power within the microgrid. The intelligent system may be based on deep learning. One or more of the priority factors may set the boundaries of what the objective function is optimized for and the priority scheme 150 results from the optimisation of the objective function performed locally by the processing unit of the master unit 50.

The intelligent system may receive input in the form of one or more forecast datasets 110 and one or more sensor datasets 120. The forecast datasets 110 may be external forecast datasets 111 which are developed outside the master unit 50, e.g. weather forecasts 113 or electricity price forecasts 115 provided by government instances or private companies transmitted to the master unit 50 via internet connection. The forecast datasets 110 may be a local forecast dataset 112 which is calculated by the processing unit of the master unit 50 on-site and may for example be based on on-site sensor datasets providing a historical forecast e.g. an electricity consumption forecast 114. The master unit 50 may receive any combination of external forecast datasets 111 and local forecast datasets 112.

The consumption data and/or generation data and/or storage data based on the previous activity of the microgrid.

In a preferred embodiment the master unit 50 receives at least one forecast dataset 119 and at least one sensor dataset 120. In a more preferred embodiment the master unit 50 receives at least one forecast dataset 119 and at least one consumption dataset 122. In a yet more preferred embodiment the master unit 50 receives at least one forecast dataset 119, at least one consumption dataset 122 and at least one generation dataset 124.

In a preferred embodiment the master unit receives at least one forecast dataset 119, a consumption dataset 122 and a generation dataset 124 which are input as training data for the intelligent system of the master unit 50, the intelligent system further uses the input data for creating a local forecast dataset 112 which is used as an input for the master unit 50 to solve an objective function and create a charging schedule 150. In a more preferred embodiment the master unit receives a plurality of forecast datasets. In a preferred embodiment the intelligent system creates local forecast datasets 112 being at least a local electricity consumption forecast 114 and a local electricity generation forecast 116. Regardless of the number and types external forecast datasets 111 and sensor datasets 120 the intelligent system of the master unit 50 may be trained to create one or more local datasets 112 being an update of an external forecast dataset 111 which has been adapted based on one or more sensor datasets 120. As an example the master unit 50 may be provided with an external forecast dataset in the form of a weather forecast as well as a sensor dataset in the form of on-site temperature measurements and by correlating the external forecast with the on-site measurements create an updated local forecast dataset 112 providing a more accurate weather forecast for the specific site. Updated locale forecast dataset 112 may in turn be used as input for other forecasts, as the intelligent system of the master unit 50 may create additional local forecast datasets 112 based on one or more forecast datasets 110 in combination with each other and/or in combination with one or more sensor datasets 120. For example an updated local weather forecast dataset may be used in combination with a historical forecast of energy consumption and a historical forecast of energy generation to calculate a local forecast of the energy generation and/or consumption.

The input forecast dataset 119 may comprise one or more of the previously mentioned forecast types, e.g. historic forecasts and dynamic forecasts. The input forecast datasets 110 may originate from external sources, e.g. market prices for electricity price forecasts 115 or weather forecasts 113 from external providers and/or they may be calculated locally by the master unit 50 based on collected data, e.g. forecast of electricity generation by local electricity generators of the microgrid or forecast of electricity consumption. In preferred embodiment forecasts of several types are used to train the intelligent system simultaneously.

ACCUMULATED TRAINING

Consumption datasets 130 and generation datasets 140 may be provided from sensor data in the microgrid, e.g. from submeters determining the amount of electrical power used by consumption units and electricity generators respectively. A consumption dataset 122 may provide a measure of the electricity consumption of the entire microgrid, it may provide subsets for groups of electricity consumption units, for single electricity consumption units or a mix thereof.

The intelligent system may be trained on further sensor data, e.g. local measurements of temperature, humidity, solar irradiation, windspeeds, etc. at the location of the microgrid, e.g. in a private household. In a preferred embodiment the intelligent system is trained on at least one forecast dataset 119 and at least one sensor dataset 120. Preferably at least on sensor dataset 120 and one forecast dataset 119 has a common data-type, e.g. temperature, irradiation or electricity consumption, such that the sensor dataset 120 may be used to evaluate and/or update the forecast. For example, if one forecast dataset 119 is an external weather forecast 111 including expected temperature and expected solar irradiation a sensor dataset 120 may be a temperature dataset 121 such that the on-site temperature may be related to the external weather forecast. The intelligent system may then create an updated local weather forecast 11 T which is adjusted based on historical correlation between the sensor dataset and the external forecast dataset, thereby allowing a better solution to the objective function. It is to be understood that the intelligent system may be configured to use the above described inputs of forecasts and sensor data and correlate them to create optimum local forecast datasets which are different from the input. The intelligent system may be optimized for forecasting electricity consumption and forecasting electricity generation, this may be based on correlating the local sensor datasets 120 with external forecasts 111 without the intelligent system being required to output intermediate local forecasts. In particular it is to be understood that the correlation which might be used to output such improved locale forecasts is part of what is training the intelligent system and the resulting output local forecasts may benefit from such training of the intelligent system without being configured to provide output relating to the correlation outside of preferred target local forecasts such as the local electricity consumption forecast and the local electricity generation forecast.

In preferred embodiments the intelligent system will evaluate the possible correlation between multiple forecast datasets and sensor datasets, such that correlation may be found for differing data types. For example correlation may be found between temperatures in a weather forecast 111 and the consumption dataset 130.

Another input for the objective function and/or for the intelligent system may be the operation mode 130. The operation mode may be set by the user and may affect local forecasts such as the consumption forecast based on historic data collected during previous use of the particular mode or it may be based on userprovided data of expectations during such a mode, e.g. vacation mode or everyday usage mode. The operation mode may affect the boundaries of the objective function, i.e. what is optimized for, e.g. the mode may be a green mode optimizing for least CO2 emission or an economic mode optimizing for the lowest cost of electricity consumption. The mode may further be used as input for the intelligent system to train it on the behaviour of the microgrid correlated to specific modes.

The intelligent system receives the the forecasts datasets 110, sensor datasets 120 and optionally operation modes 130 of the connected system and supplies the master unit with updated forecasts which in combination with datasets relating to initial values for the master unit to solve the objective function and create a charging schedule 150 for optimize the behaviour of the microgrid according to which the master device directs the usage of electricity consumption units, electricity generators, electricity reservoirs and/or returning electricity to the grid.

In a local power grid comprising multiple microgrids a consumption schedule may also be made based on priority schemes and/or priority factors for each microgrid within the local power grid. In some variants the master unit may be set up to limit consumption of each microgrid while allowing a control unit of the microgrid to balance the load between units within the microgrid. In other variants the master unit may control the distribution of electricity to the various outlet units within each microgrid of the local power grid. Having more outlet units belonging to a single master unit allows for more fine-tuning of the distribution, as the preferred time of electrical power usage is not determined only based on the power consumed within a single microgrid, but takes into account the consumption in multiple microgrids. It is noted that an average distribution may be accounted for based on forecast data such as electricity consumption forecasts, which are made for larger areas than the master unit controls, e.g. for a state. By having more outlet units within a distributed power consumption system controlled by a single master unit, it is possible to more accurately fulfil the priority factors, as more combinations are possible within the distributed power consumption system. Alternatively, the same local grid and/or microgrid may have multiple master units cooperating to ensure redundancy in the system and/or to enable the adjustment of electrical power consumption of a large number of outlet units while ensuring individual control of the outlet units.

Fig. 5 illustrates the concept of time scales of updating forecast data and sensor data. An external forecast dataset is illustrated as being updated at a forecast intervals TN. During each forecast interval TN a sensor dataset is updated multiple times with intervals tnk, such that for each forecast interval TN there is are k sensor intervals tn. The objective function may be solved once for each time interval TN outputting a solution in the form of an electrical power forecast and/or a charging schedule for a predefined prediction horizon, in Fig. 5 the prediction horizon is illustrated as three forecast intervals but it is to be understood that this is purely exemplary. The sensor dataset intervals are shorter than the forecast intervals such that a number of sensor datasets are collected and input to the intelligent system and used as input for the update of the solution calculated for the next forecast interval. The objective function may be solved for a dataset in the form of an array or on a stochastic dataset.

The forecast intervals TN may be the time units for which a specific optimized operation of the microgrid is determined. At a predetermined frequency the master unit will determine a charging schedule for the microgrid, the charging schedule is solved for the optimum operation of one or more flexible units of the microgrid during each time interval of the prediction horizon. In the illustrated example the master unit may update the charging schedule once for every time unit based on the updated forecast dataset and the collected sensor dataset and will output a charging schedule for the next prediction horizon of the following three time units determining the operation parameters within each of the time units. It is understood that the example illustrated in Fig. 5 is for illustration purposes only and is made to be visually simple. In many of the embodiments of the invention a plurality of forecast datasets and sensor datasets are included and the prediction horizon will include more time units TN than three. For example the forecast intervals may be one hour and the time unit also being one hour with a prediction horizon of twenty four hours.

In preferred embodiments the master unit will receive multiple forecast datasets and multiple sensor datasets. In such embodiments the update rate for the solving of the objective function and the time unit of the charging schedule may be set by one of the forecast intervals. In some variants it may be the longest forecast interval, in other more preferred embodiments the solving of the objective function is synchronized with a specific forecast interval having a high impact on the objective function. The choice of synchronization may be dependent on a user input operation mode 130. For example the calculation rate for solving the objective function may depend on the weather forecast when the operation mode is focused on generating and selling electrical power or it may be based on the longest forecast rates when in a maintenance mode where downtime is necessary.

In a preferred embodiment the intelligent system of the master unit is trained on accumulated forecast data, such that the results of previous forecast intervals and data intervals and their relation is taken into account when creating the updated local forecasts. As the inputs are updated the solution of the objective function will also be updated and the resulting charging schedule may be updated at the update rate of an external forecast dataset.

Fig. 6 illustrates an example of data related to operation of the master unit during a day cycle. The first axis 201 this illustrates the passing of time for a day starting at midnight and ending at midnight. The second axis 202 varies for the different datasets. The sensor data relating to the electricity consumption is shown as a solid line 210; three peaks occur, one during the night as an EV is charged, one during the morning and one during the evening where the residents are at home. A forecast dataset for the price of electricity is shown as a dashed line 230 and is seen to partially correspond to the usage of the household but with slightly shifted peaks corresponding to the accumulated use of microgrids in a regional grid. The sensor dataset of electricity generation 220 of a solar cell of the microgrid is shown as a thin line with dots and dashes. Above the graph the charge state of an electricity reservoir 30 is illustrated, where no arrow means that no change of the state is taking place while a downward pointing arrow indicates discharging of the electricity reservoir and an upward pointing arrow indicates charging of the electricity reservoir.

The illustrated example explains a use case of the functioning of the intelligent system is described for the microgrid of a household having an electricity generator in the form of a solar cell, an electricity reservoir in the form of a battery, a number of electricity consumption units being regular household appliances and a master unit for guiding electricity to the consumption units and the electricity reservoir of the microgrid as well as the purchase and selling of electricity to the local power grid and/or the regional power grid. The intelligent system is computer implemented in the processing unit of the master unit. In the illustrated example one or more priority factors are set for the objective function to optimises for low cost operation of the microgrid. The master unit receives external forecast datasets relating to weather forecast including temperature and solar irradiation. The objective function further receives as an input a local electricity consumption forecast determined locally in the processing unit by the intelligent system trained on at least historical data of electricity consumption of the microgrid. A local electricity generation forecast is calculated by the intelligent system using data relating to the specific solar cell of the microgrid based on historic data of previous generation, the conversion efficiency of the specific solar cell and the weather forecast. A forecast of the generation of electricity by the solar cell for the following twenty-four hours is determined by the intelligent system. The intelligent system also creates a forecast of the electricity consumption for the same twenty-four hour period. The forecasts are consecutively updated whenever new input is received from local sensor data relating to generation and consumption and external weather forecasts respectively. These forecasts are processed together by the intelligent system in the local processing unit of the master unit to create a charging schedule which determines how the master unit will direct the electrical power usage of the microgrid.

As shown in Fig. 6 electricity consumption units may use power generated by the solar cell during periods of peak consumption where the electricity reservoir of the microgrid may also be depleted. In the night where the cost of electricity is low, electricity may be purchased to charge a connected EV, while electrical power stored in the electricity reservoir 30 may be consumed during the morning where the microgrid consumes electrical power while the electricity prices are high. During a subsequent period of less consumption in the local microgrid while an external market forecast 230 predicts high prices of electricity, the master unit may direct electricity to be sold to the regional power grid both from newly generated electrical power 220 and depleting the electricity reservoir 30. Later while the local electricity consumption is predicted to be low the electricity reservoir may be filled, i.e. the battery may be charged, when the market price of electricity is low and electricity generation is high such that the electricity reservoir is again at full capacity to be consumed during the evening when the microgrid is predicted to have a high energy consumption while electricity prices are also high - the relative term high is used the indicate higher than in other periods during the twenty-four hour prediction horizon. Hence, the long prediction horizon enables long-term optimization of the usage and trading of electricity rather than deferring selling of electricity generated in the microgrid until the electricity reservoirs are filled, this is both more economic and provides better electricity capacity and usage of electricity reservoirs as flexible resources for the regional power grid.

REFERENCE NUMBERS

10 Distributed charging system

15 Regional power grid

17 Local power grid Microgrid

Electricity generator

Solar cells

Wind turbine

Nuclear power plant

T ransformer

Electricity reservoirs

Electricity consumption unit

Fuse box

Master unit

Control unit

Cloud

Inverter

Switch

Signal gateway

Outlet unit

Detection meter

External forecast dataset

Local forecast dataset

Weather forecast

Electricity consumption forecast

Electricity price forecast

Electricity generation forecast

Forecast dataset

Sensor dataset

Consumption dataset

Generatin dataset

Priority scheme

Electricity consuption dataset

Electricity generation dataset 230 Electricity price forecast

Now follows a set of items, which constitute aspects of the present disclosure which may be considered independently patentable and as such the following sets form basis for possible future sets of claims:

1 . A master unit for use in a modular distributed power consumption system for distributing electrical power between electricity consumption units, the modular distributed power consumption system comprising, a master unit adapted for directing electrical power and for prioritising the distribution of said electrical power to a plurality of outlet units, said master unit comprising a processing unit for determining a consumption schedule according to which electricity consumption units, electricity generators and/or electricity reservoirs connected to said modular distributed power consumption system via said outlet units use and/or deliver electrical power, said master unit being adapted to direct at least a subpart of the available electrical power to one or more outlet units of said plurality of outlet units based on a priority scheme.

2. The master unit according to item 1 , adapted for monitoring and distributing electrical power from a local power grid, said local power grid receiving electrical power from one or more of a regional power grid, local power generators and/or a local energy reservoir.

3. The master unit according to any one of the preceding items, said master unit being adapted for monitoring and distributing electrical power from one or more of said outlet units such that it can redistribute electrical power from said one or more electricity consumption units, electricity generators and/or electricity reservoirs.

4. The master unit according to any one of the preceding items, said master unit further being adapted for directing electrical power to a local power grid and/or to a regional power grid.

5. The master unit according to any one of the preceding items, said processing unit of the master unit being configured to be in communication with an external server such as a database and/or the internet. 6. A modular distributed power consumption system for distributing electrical power between electricity consumption units, the distributed power consumption system comprising, a master unit according to any one of the items 1-5; and a plurality of slave electricity outlet units; said master unit being adapted to direct at least a subpart of said electrical power to one or more of said electricity outlet units based on a priority scheme.

7. A modular distributed power consumption system according to item 6, further comprising one or more electricity reservoirs and/or electricity generators.

8. A modular distributed power consumption system according to any one of items 6-7, comprising at least one submeter adapted for monitoring characteristics of electrical power delivered to and/or from devices connected to said modular distributed consumption system and transmitting such data to said master unit.

9. A method of determining a charging schedule for electrical power distribution to a plurality of electricity consumption units in a distributed power consumption system carried out by a master unit of said distributed power consumption system, said method comprising the steps of: providing a list of priority factor options, for one or more of the connected electricity consumption units and/or electricity reservoirs, receiving user input ranking one or more priority factors, creating a priority scheme ranking priority factors for all connected electricity consumption units and/or electricity reservoirs, receiving an electrical power forecast, based on said electrical power forecast and said priority scheme determining a charging schedule, said charging schedule defining usage periods during which electrical power is transmitted to said connected electricity consumption units and/or electricity reservoirs and the fraction of available electrical power which is directed by said master unit to said electrical consumption unit and/or electricity reservoirs during said usage periods.

10. A method according to item 9, said electrical power forecast comprising an electricity generation forecast. 11. A method according to any one of the items 9-10, said electrical power forecast comprising an electricity consumption forecast.

12. A method according to any one of the items 9-11 , said electrical power forecast being based at least in part on a historic forecast such that electricity consumption and/or electricity generation is estimated based on generation and/or consumption from the same system at a prior time.

13. A method according to any one of the items 9-12, said electrical power forecast being based at least in part on a dynamic forecast, said dynamic forecast being consecutively updated.

14. A method according to item 13, said dynamic forecast being based on dynamic data such as weather forecasts, recent changes in user consumption and/or local temperature measurements.

15. A method according to any one of the items 9-14, said electrical power forecast being based at least in part on a user forecast such that user expectations of changes are taken into consideration.

Claims

1 . A master unit for use in a modular distributed power consumption system for distributing electrical power between flexible electricity units, the modular distributed power consumption system comprising, a master unit adapted for directing electrical power and for prioritising the distribution of said electrical power to a plurality of outlet units, said master unit comprising a processing unit configured for on-site training of an intelligent system and determining a consumption schedule according to which flexible electricity units connected to said modular distributed power consumption system via said outlet units use and/or deliver electrical power, said master unit being adapted for directing at least a subpart of the available electrical power to one or more outlet units of said plurality of outlet units based on an objective function receiving an electrical power forecast.

2. The master unit according to claim 1 , adapted for monitoring and distributing electrical power from a local power grid, said local power grid receiving electrical power from one or more of a regional power grid, one or more local power generators and/or one or morelocal energy reservoirs and the master unit further being adapted for directing electrical power to a local power grid and/or to a regional power grid.

3. A modular distributed power consumption system for distributing electrical power between electricity consumption units, the distributed power consumption system comprising, a master unit according to any one of the claims 1-2; and a plurality of slave electricity outlet units; and at least one sensor for collecting a sensor dataset and providing said master unit with said sensor dataset; said master unit being adapted to direct at least a subpart of said electrical power to one or more of said electricity outlet units based on a priority scheme.

4. A modular distributed power consumption system according to claim 3, further comprising one or more electricity reservoirs and/or electricity generators.

5. A method of determining a charging schedule for electrical power distribution to and from a plurality of flexible electrical units of a distributed power consumption system carried out by a master unit of said distributed power consumption system, said method comprising the steps of: receiving at least one forecast dataset, receiving at least one sensor dataset from on-site data, on-site training an intelligent system for creating an electrical power forecast, said intelligent system being trained on said at least one forecast dataset and said at least one sensor dataset, obtaining an electrical power forecast for inputting to an objective function to be solved by the master unit, said electrical power forecast comprising an electricity consumption forecast, on-site solving of said objective function determining a charging schedule based on said electrical power forecast, said charging schedule defining action periods during which electrical power is transmitted between said connected flexible electrical units and/or the regional power grid and the fraction of available electrical power which is directed by said master unit to and/or from said flexible electrical units during said action periods.

6. A method according to claim 5, said electrical power forecast comprising an electricity generation forecast, an electricity consumption forecast and/or an electricity storage forecast.

7. A method according to any one of the claims 5-6, at least part of said input electrical power forecast being obtained by on-site calculation performed by the intelligent system.

8. A method according to any one of the claims 5-7, said electrical power forecast having a prediction horizon exceeding the forecast update rate of at least one external forecast dataset, such as the prediction horizon being at least twenty-four hours.

9. A method according to any one of the claims 5-8 , wherein said sensor dataset is received with an update rate exceeding the update rate of said forecast dataset.

10. A method according to any one of the claims 5-9, said electrical power forecast being based at least in part on a historic forecast such that electricity consumption and/or electricity generation is estimated based on generation and/or consumption from the same system at a prior time.

11. A method according to any one of the claims 5-10, said electrical power forecast being based at least in part on a dynamic forecast, said dynamic forecast being consecutively updated.

12. A method according to any one of the claims 5-11, said at least one forecast dataset comprising a dynamic forecast of inverter losses.

13. A method according to any one of the claims 5-12, said master unit receiving user input of an operation mode such that user expectations of changes are taken into consideration.

14. A method according to any one of the claims 5-13, said intelligent system being trained in an accumulative manner, such that previous data from a dynamically updated forecast dataset affects the resulting charging schedule.

15. A method according to any one of the claims 5-14, said intelligent system being trained for stochastic optimization such that a distribution of sensor dataset is used as the input for the objective function.