FI20225279A1

FI20225279A1 - Controlling distributed energy storage system

Info

Publication number: FI20225279A1
Application number: FI20225279A
Authority: FI
Inventors: Simon Holmbacka; Jukka-Pekka Salmenkaita; Esko Heinonen
Original assignee: Elisa Oyj
Priority date: 2022-03-31
Filing date: 2022-03-31
Publication date: 2023-10-13
Also published as: WO2023187254A1

Abstract

A computer implemented method for controlling a virtual power plant, VPP, comprising a plurality of spatially distributed energy storage, DES, devices. The method is performed by obtaining historical reference data and DES infrastructure data; identifying an operating objective for the virtual power plant; simulating operation of the virtual power plant over a first time period based on the historical reference data and the DES infrastructure data; selecting an operating plan for the first time period based on the simulation in view of the operating objective, wherein the operating plan concerns utilization of the virtual power plant for grid balancing; using the selected operating plan and responsively monitoring operation of the virtual power plant over the first time period; and using results from the monitoring to adjust future simulations.

Description

CONTROLLING DISTRIBUTED ENERGY STORAGE SYSTEM

TECHNICAL FIELD

The present disclosure generally relates to controlling a distributed energy storage system.

The present disclosure further relates to controlling a virtual power plant comprising distributed energy storage devices.

BACKGROUND

This section illustrates useful background information without admission of any technique described herein representative of the state of the art.

A distributed energy storage (DES) system is a pool of battery resources controlled by a centralized control system. A DES system can be used for forming a virtual power plant (VPP) comprising a plurality of spatially distributed energy storage (DES) devices. In this way a larger capacity may be built by pooling together smaller scale resources. The DES devices may be resources maintained for example for emergency energy backup purposes, such as backup batteries of a wireless communication network. Additionally or alternatively, the DES devices may be resources owned by households or small and medium sized companies or other smaller scaler operators. As backup batteries are not constantly used, the resources can be used for further optimization purposes e.g. through the VPP.

Such VPPs may participate in balancing of electric grid or in intraday trading market.

Transmission system operators (TSO) offer reserve markets where reserve providers, such as VPP, can offer energy capacity for grid balancing purposes. In order to participate in the grid balancing, the reserve provider needs to submit bids to the reserve market in advance, e.g. the day before (in Finland by 7.30 CET the previous day).

N

N Now, there are provided some new considerations concerning controlling virtual power plant

N formed of a plurality of spatially distributed energy storage devices. o ™ 25 SUMMARY

I a. The appended claims define the scope of protection. Any examples and technical

O descriptions of apparatuses, products and/or methods in the description and/or drawings io not covered by the claims are presented not as embodiments of the invention but as

O background art or examples useful for understanding the invention.

According to a first example aspect there is provided a computer implemented method for controlling a virtual power plant, VPP, comprising a plurality of spatially distributed energy storage, DES, devices. The method comprises obtaining historical reference data and DES infrastructure data; identifying an operating objective for the virtual power plant; simulating operation of the virtual power plant over a first time period based on the historical reference data and the DES infrastructure data; selecting an operating plan for the first time period based on the simulation in view of the operating objective, wherein the operating plan concerns utilization of the virtual power plant for grid balancing; using the selected operating plan and responsively monitoring operation of the virtual — power plant over the first time period; and using results from the monitoring to adjust future simulations.

In some embodiments, the operating objective comprises one or more of. optimizing utilization of resources of the virtual power plant, minimizing wear of the DES devices, minimizing failures to comply with grid balancing activations, and minimizing failures in — operation of the DES infrastructure.

In some embodiments, the historical reference data comprises historical information about one or more of: up or down regulation reguests, up or down regulation activations, Up or down regulation pricing, energy pricing.

In some embodiments, the DES infrastructure data comprises one or more of the following: power consumption, capacity, wear, and physical properties of the DES devices of the virtual power plant.

In some embodiments, the method comprises performing the simulation for plurality of different operating plans and/or performing the simulation for plurality of different combinations of historical reference data and DES infrastructure data.

N

N 25 In some embodiments, the method further comprises using the selected operating plan for

N

O participating in intraday trading market. o 5 In some embodiments, the simulation comprises iteratively optimizing the operating plan in = view of the operating objective based on feedback from earlier simulation. a o In some embodiments, the iteration is performed in the scope of hours or even less.

N a 30 In some embodiments, the simulation comprises a machine learning model configured to

N predict future operation of the virtual power plant.

In some embodiments, the machine learning model is configured to predict plurality of possible operating scenarios.

In some embodiments, the machine learning model is configured to predict one or more of: up or down regulation requests, up or down regulation activations, up or down regulation pricing.

In some embodiments, the results from the monitoring are used to adjust future simulations in the scope of days, weeks or months.

According to a second example aspect of the present invention, there is provided an apparatus comprising a processor and a memory including computer program code; the memory and the computer program code configured to, with the processor, cause the — apparatus to perform the method of the first aspect or any related embodiment.

According to a third example aspect of the present invention, there is provided a computer program comprising computer executable program code which when executed by a processor causes an apparatus to perform the method of the first aspect or any related embodiment.

According to a fourth example aspect there is provided a computer program product comprising a non-transitory computer readable medium having the computer program of the third example aspect stored thereon.

According to a fifth example aspect there is provided an apparatus comprising means for performing the method of any preceding aspect.

Any foregoing memory medium may comprise a digital data storage such as a data disc or diskette; optical storage; magnetic storage; holographic storage; opto-magnetic storage; phase-change memory; resistive random-access memory; magnetic random-access memory; solid-electrolyte memory; ferroelectric random-access memory; organic memory;

N or polymer memory. The memory medium may be formed into a device without other

N 25 substantial functions than storing memory or it may be formed as part of a device with other ? functions, including but not limited to a memory of a computer; a chip set; and a sub

O assembly of an electronic device.

I a = Different non-binding example aspects and embodiments have been illustrated in the = foregoing. The embodiments in the foregoing are used merely to explain selected aspects a 30 or steps that may be utilized in different implementations. Some embodiments may be

N presented only with reference to certain example aspects. It should be appreciated that corresponding embodiments may apply to other example aspects as well.

BRIEF DESCRIPTION OF THE FIGURES

Some example embodiments will be described with reference to the accompanying figures, in which:

Fig. 1 schematically shows a system according to an example embodiment;

Fig. 2 shows a block diagram of an apparatus according to an example embodiment; and

Fig. 3 shows a flow chart according to example embodiments; and

Fig. 4 shows logical components of an arrangement according to an example embodiment.

DETAILED DESCRIPTION

In the following description, like reference signs denote like elements or steps. — Various embodiments of present disclosure provide mechanisms to control a virtual power plant (VPP) that comprises a plurality of spatially distributed energy storage (DES) devices.

The DES devices may be individually owned resources of households or small and medium sized companies or other smaller scaler operators. Alternatively or additionally, the DES devices may be energy assets owned by the operator of the virtual power plant or otherwise centrally owned energy assets. The DES devices may be intended for emergency backup purposes, but this is not mandatory. In an example embodiment, the DES devices are backup batteries of a wireless communication network. In another example embodiment, the DES devices are battery units of households or battery units of buildings. In an example embodiment, the DES devices are co-located with an energy production unit. As an alternative non-limiting example, the DES devices may be intended for storing energy from renewable sources such as solar panels and/or wind generators or even from fuel cell or other type of fuel-operated genset. Yet another additional or alternative intended use of the

DES devices is optimization of self-consumption. The DES device may be a hybrid system using multiple energy sources.

Ql

O 25 Ingeneral, the DES devices in this disclosure refer to storage devices that are able to handle se regular charge and discharge cycles. For example, lithium-ion batteries are such devices. — In more detail, one or more of the following battery technologies may be used: lithium-nickel- 2 cobalt, NCA, lithium-iron-phosphate, LFP, lithium-nickel-manganese-cobalt, NMC, solid-

E state batteries, and flow batteries. The DES devices may have different properties with = 30 regard to price, durability, physical size and chemical wear depending for example on the a battery technology and storage capacity.

N In general, lithium-based batteries should not regularly exceed extreme low or high charge values. For example, state of charge below 5% or above 95% should be avoided. Such limitations should be taken into account in usage of the lithium-based batteries to avoid increased wear of the batteries.

One aim is to achieve with presently disclosed solutions is optimization of usage of the virtual power plant for grid balancing. Grid balancing may be arranged for example using 5 automatic Frequency Restoration Reserve, aFRR, or Frequency Containment Reserve,

FCR, capacity market. aFRR is a centralized automatically activated reserve. Its activation is based on a power change signal calculated on the base of the frequency deviation in the Nordic synchronized area. Its purpose is to return the frequency to the nominal value.

FCR is an active power reserve that is automatically controlled based on the frequency deviation. FCR may be Frequency Containment Reserve for Normal Operation, FCR-N, or

Frequency Containment Reserve for Disturbances, FCR-D. Their purpose is to contain the frequency during normal operation and disturbances.

The frequency balancing may comprise up regulation and/or down regulation. Up regulation means increasing power production or decreasing consumption. Down regulation means decreasing power production or increasing consumption. The up regulation and down regulation may be symmetric or asymmetric.

When the operator of the virtual power plant wants to participate in the grid balancing, bids need to be submitted to the reserve market in advance, e.g. the day before (in Finland by 7.30CET the previous day). The bidding may be a challenging task for the operator of the virtual power plant as the resources are distributed and both local energy source and aggregated energy levels for the grid balancing need to be ensured simultaneously.

The following examples illustrate complexity of the task of determining participation in grid

N balancing. In many cases, all of the examples may be relevant at the same time.

O

N 25 1)In an example case, there are 1000 DES sites, and the sites have a battery storage ? with a capacity between 5 and 50 kWh, each. In such a setup, optimal amount of

O energy needs to be allocated from each site so that the total energy consumption of

E an individual site does not exceed the respective battery capacity, but the o aggregated energy levels meet the activation requirements.

N a 30 2)In an example case, there are 1000 DES sites, and the sites have a downstream

Q rectifier capacity between 12 and 26 kW, each. In such a setup, optimal amount of power needs to be allocated from each site so that the down regulation matches the activation signal from the grid coordinator.

3)In an example case, there are 1000 DES sites, and the sites have an upstream inverter capacity between 0 and 10 kW, each. Additionally, the upstream activation can be alleviated by activating battery powered equipment (staggering) instead of drawing power from the grid. In such a setup, optimal amount of power needs to be allocated from each inverter site and staggering needs to be coordinated such that the up regulation matches the activation signal from the grid coordinator. 4)In an example case, there are 1000 DES sites, and the sites have a mains connection capacity between 11 and 25 kW, each. In such a setup, there is a need to coordinate the mechanisms in example cases 2 and 3 so that the mains capacity is not exceeded in combination with the equipment power used at the moment on each site.

Staggering herein refers to method of using energy stored in backup battery as an energy source in normal operating conditions. This may be done e.g. during time periods when electricity is expensive and the backup battery may be recharged during time periods when — the electricity is cheaper. Staggering could be referred to as load shifting, too

Various embodiments of present disclosure provide mechanisms for determining operating plan for a virtual power plant formed of a plurality of DES devices. The operating plan defines for example bids for grid balancing and may as well define other planned resource usage such as time periods for charging and/or discharging the DES devices for staggering — purposes. The embodiments solve the technical problem of determining how to control the virtual power plant to achieve efficient participation in grid balancing. Further aim is to optimize utilization of the resources of the virtual power plant.

At least one operating objective is identified and taken into account in determining the operating plan and in controlling the virtual power plant so that desired operation is

N 25 achieved. The operating objectives may be selected from optimization of utilization of the

N virtual power plant, minimizing wear of the DES devices, minimizing failures to comply with 3 grid balancing activations, and minimizing failures in operation of the DES infrastructure. In n this way, more efficient usage of DES devices may be achieved. Still further, grid balancing = may be improved, and more stable energy source may be achieved without additional > 30 environmental burden as already existing energy assets may be used. Further, the operator

N of the virtual power plant may for example ensure appropriate operation of the DES

N infrastructure whilst at the same time the operator of the virtual power plant may achieve

N maximal benefits from the grid balancing.

Fig. 1 schematically shows an example scenario according to an embodiment. The scenario shows a pool of DES devices 121-125. The DES devices 121-125 may be located at different geographical locations, but equally there may be plurality of DES devices at the same location. Fig. 1 shows the DES devices 123-125 at the same location and DES devices 121 and 122 individually at different locations. It is to be noted that this is only a non-limiting illustrative example and in practical implementations many different setups are possible.

Further, the scenario shows a control system 111. The control system 111 and the DES devices 121-125 form a DES system that may operate as a virtual power plant. Still further,

Fig. 1 shows an electric grid 151.

The control system 111 is configured to implement at least some example embodiments of present disclosure to control the virtual power plant. For this purpose, the control system 111 is operable to interact with the DES devices 121-125 or equipment associated thereto.

Additionally, the control system 111 is operable to interact with the electric grid 151 or equipment associated thereto to coordinate participation in grid balancing and/or intraday — trading market.

The operator of the virtual power plant may receive compensation based on the freguency balancing carried out for the electric grid. The compensation may depend on actual activation of freguency balancing and/or on reserving capacity for the possible freguency balancing needs.

Fig. 2 shows a block diagram of an apparatus 20 according to an embodiment. The apparatus 20 is for example a general purpose computer, cloud computing environment or some other electronic data processing apparatus. The apparatus 20 can be used for implementing at least some embodiments of the invention. That is, with suitable configuration the apparatus 20 is suited for operating for example as the control system 111 3 25 of Fig. 1.

O The apparatus 20 comprises a communication interface 25; a processor 21; a user interface ? 24; and a memory 22. The apparatus 20 further comprises software 23 stored in the memory 9 22 and operable to be loaded into and executed in the processor 21. The software 23 may = comprise one or more software modules and can be in the form of a computer program

R 30 product.

N The processor 21 may comprise a central processing unit (CPU), a microprocessor, a digital

N signal processor (DSP), a graphics processing unit, or the like. Fig. 2 shows one processor 21, but the apparatus 20 may comprise a plurality of processors.

The user interface 24 is configured for providing interaction with a user of the apparatus.

Additionally or alternatively, the user interaction may be implemented through the communication interface 25. The user interface 24 may comprise a circuitry for receiving input from a user of the apparatus 20, e.g., via a keyboard, graphical user interface shown on the display of the apparatus 20, speech recognition circuitry, or an accessory device, such as a headset, and for providing output to the user via, e.g., a graphical user interface or a loudspeaker.

The memory 22 may comprise for example a non-volatile or a volatile memory, such as a read-only memory (ROM), a programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), a random-access memory (RAM), a flash memory, a data disk, an optical storage, a magnetic storage, a smart card, or the like. The apparatus 20 may comprise a plurality of memories. The memory 22 may serve the sole purpose of storing data, or be constructed as a part of an apparatus 20 serving other purposes, such as processing data.

The communication interface 25 may comprise communication modules that implement data transmission to and from the apparatus 20. The communication modules may comprise a wireless or a wired interface module(s) or both. The wireless interface may comprise such as a WLAN, Bluetooth, infrared (IR), radio frequency identification (RF ID),

GSM/GPRS, CDMA, WCDMA, LTE (Long Term Evolution) or 5G radio module. The wired interface may comprise such as Ethernet or universal serial bus (USB), for example. The communication interface 25 may support one or more different communication technologies. The apparatus 20 may additionally or alternatively comprise more than one of the communication interfaces 25.

A skilled person appreciates that in addition to the elements shown in Fig. 2, the apparatus

N 25 20may comprise other elements, such as displays, as well as additional circuitry such as

O memory chips, application-specific integrated circuits (ASIC), other processing circuitry for se specific purposes and the like. ™ Fig. 3 shows a flow chart according to example embodiments. The illustrated processes = comprise various possible steps including some optional steps while also further steps can > 30 be included and/or some of the steps can be performed more than once. The processes

N may be implemented in the control system 111 of Fig. 1 and/or in the apparatus 20 of Fig.

N 2. The processes are implemented in a computer program code and do not require human

N interaction unless otherwise expressly stated. It is to be noted that the processes may however provide output that may be further processed by humans and/or the processes may require user input to start.

The process of Fig. 3 comprises the following steps: 301: Historical reference data and DES infrastructure data are obtained.

In an embodiment, the historical reference data comprises historical information about grid balancing actions. In an embodiment, the historical reference data comprises historical information about one or more of: up or down regulation requests, up or down regulation activations, up or down regulation pricing. The historical reference data may be data that has been collected in the control system over a past time period or the historical data may be obtained from a suitable data source.

In an embodiment, the DES infrastructure data comprises information about operation and/or status of the DES devices of the virtual power plant. In an embodiment, the DES infrastructure data comprises one or more of the following: power consumption, capacity, wear, and physical properties of the DES devices of the virtual power plant. The DES infrastructure data may be data that has been collected in the control system over a past — time period or the DES infrastructure data may be obtained from a suitable data source or the DES infrastructure data may be collected from the DES devices. 302: An operating objective for the virtual power plant is identified. The operating objective may be a preselected setting that defines which aspects of operation of the virtual power plant shall be optimized. The operating objective may define one or more of: optimizing utilization of the resources of the virtual power plant, minimizing wear of the DES devices, minimizing failures to comply with grid balancing activations, and minimizing failures in operation of the DES infrastructure. If more than one operating objectives are used, the operating objectives may be arranged in order of preference or in order of importance.

A 303: Operation of the virtual power plant is simulated over a first time period (e.g. next day)

O 25 based on the historical reference data and the DES infrastructure data. The first time period 6 is a future time period at this phase of the process. It is to be noted that more than one o - operating objective may be taken into account at the same time.

O

= The historical reference data, the DES infrastructure data and the operating objective may > be collected into a model modelling operation of the virtual power plant. The model is used

N 30 in the simulation.

LO

N

S 304: An operating plan is selected based on the simulation. The operating plan is selected in view of the operating objective. That is, the selection is performed with the aim of fulfilling the settings or targets defined in the operating objective. The operating plan concerns utilization of the virtual power plant e.g. for grid balancing and other purposes. The operating plan defines for example how much energy, during which time periods can be offered for grid balancing. Additionally, the operating plan may define how much energy, during which time periods can be used for staggering.

In an embodiment, the simulation is performed by iteratively optimizing the operating plan based on feedback from earlier simulation(s). First, an initial operating plan may be defined for the first time period and the simulator simulates behaviour of the system with such operating plan. One or more simulation runs may be performed with the same operating plan in order to simulate different combinations of historical reference data, DES infrastructure data and the operating objective. Feedback from the simulation(s) is used for defining improved operating plan, which is in turn simulated and so on. The iterative optimization may be performed in the scope of hours (or even minutes) and the iterative optimization may be referred to as a fast feedback loop.

Simulation of the operation of the virtual power plant provides that a large set of operating — plans and their effects can be rapidly tested and evaluated.

Examples of parameters that may be considered in the simulation are: - Acceptance ratio of grid balancing bids. - Activation levels, which are highly stochastic elements and determine both the magnitude of the activation as well as the duration of and activation in up and down direction. The simulation can for example analyze the risk level of bid levels according to generated activation scenarios. - The simulation can insert adhoc strategies like e.g. mid-day charging or staggering to assess the energy level in the DES devices and increase the bid levels or bidding hours.

N

O

N 25 - The simulation can assume various parameter changes in the DES infrastructure like 3 battery capacity, rectifier or inverter capacity or other infrastructure related n parameters affecting the bid levels. In this way, for example a selected area of = assets can be simulated for step-wise rollout or capacity planning or related system a » development purposes.

N io 30 In an embodiment, the simulation comprises a machine learning model configured to predict

N or generate data of future operation of the virtual power plant. The machine learning model

N may be configured to predict one or more of: up or down regulation reguests, up or down regulation activations, up or down regulation pricing, energy pricing. The energy price may include spot price and/or intraday price of energy. Additionally or alternatively, the machine learning model configured to predict or generate different possible combinations of the historical data and the DES infrastructure data.

There may be one or more different models and the model(s) may provide plurality of alternative predictions of plausible scenarios. In an embodiment, the prediction herein refers to artificially generated data of one or more possible operating scenarios. Such possible scenarios may then be used in the simulation. 305: The selected operating plan is then used in real system during the first time period (e.g. the next day). That is, the bid defined in the selected operating plan is placed in the energy reserve market. 306: Responsively, operation of the virtual power plant is monitored. It is monitored e.g. whether up or down regulation requests are actually accepted and/or activated. Further pricing of the up or down regulation may be monitored. Still further, state of charge and other properties of the DES devices of the virtual power plant may be monitored. 307: The results from the monitoring are used to adjust future simulations. Such adjustment — may be performed in the scope of days, weeks or months. This may be referred to as a slow feedback loop.

The adjustment in step 307 may concern updating models used in the simulation by reinforcement learning or other suitable learning methods.

In an embodiment, the selected operating plan further concerns participation in intraday trading market.

The logical components are historical reference data 401, DES infrastructure data 402, a model 403, an operating plan 404, a simulator 405, a first feedback 406, a selected

N operating plan 407, areal system 408, and a second feedback 409.

N

6 25 The first feedback 406 from the simulator 405 to the operating plan 404 forms a first ? feedback loop 420. The second feedback 409 from the real system 408 to the model 403 9 forms a second feedback loop 430.

I a * In an embodiment, the arrangement of Fig. 4 operates as follows: o

N

N The historical reference data 401 and the DES infrastructure data 402 are collected to the

N

O 30 model 403 that models operation of the virtual power plant. An operating plan 404 is generated including grid balancing bid levels for each time granularity (e.g. 1h) for the next day. The model 403 may be used as a basis for generating the operating plan 404. The bid levels are for either up or down regulation or for both. Up and down regulation bids might or might not be of similar magnitude.

The operating plan 404 is fed into the simulator 405, which forecasts operation of the real system and the grid balancing operations for next day based on the model 403. The task of the simulator is to reflect the real operations as closely as possible and to forecast the result of the operating plan 404. The same operating plan may be run in the simulator multiple times (e.g. tens or hundreds of times or even much more). The simulator has stochastic elements that may happen differently in different simulation rounds. There may be variation e.g. in how grid balancing bids happen to be activated in different simulation runs. In this way, a range of possible outcomes are obtained and a more comprehensive view on probable outcome can be obtained.

The first feedback 406 from one or more simulation runs of the simulated next day are then fed back to adjusting the operating plan 404. The first feedback (i.e. the results from the simulation runs) may be compared to the operating objective to determine how to adjust the — operating plan. Adjustment of the operating plan may be performed by adjusting the planned grid balancing bids, but equally also other adjustments may be performed in the operating plan. This is iterated until desired simulation result is reached and the corresponding operating plan is selected. In this way the operating plan is selected in view of the operating objective. At this stage, the first feedback loop 420 is completed.

In an example embodiment, the operating objective is minimizing failures to comply with grid balancing activations and minimizing failures in operation of the DES infrastructure.

Now, if the initial operating plan defines bidding SMW at 12:00 and the simulation results indicate that the system will run out of energy in the battery in 50% of the cases, the operating plan could be adjusted by reducing the bid level at 12:00 to 4MW. If the adjusted

N 25 — operating plan results in simulation results indicating substantial failure rate, a further

O adjustment can be made and simulated again until acceptable failure rate (e.g. close to 0%) 0 is reached. <Q ™ The selected operating plan 407 is used in the real system 408 for the next day. The

E selected operating plan 407 may be saved for example in a cloud storage solution. = 30 As the selected operating plan 407 is in use, the operation of the real system (i.e. the

O operation of the virtual power plant) is monitored. The second feedback 409 is obtained as

O a result of the monitoring. For example, operation and/or status of the DES devices and the virtual power plant, such as energy usage, power capacity, failure flags etc. may be monitored. The second feedback is fed to the model 403 to adjust the model and thereby the future simulations of the simulator 405. The second feedback (i.e. the results from the monitoring of the real system) may be compared to the operating objective to determine how to adjust the model. At this stage, the second feedback loop 430 is completed.

The second feedback loop provides that the model may be adjusted according to trends from the real system. The second feedback loop can for example signal if the up or down regulation is over represented over a longer time period, and the future operating plans may thereby be adjusted through the model 403 to adapt to real system operation.

The model 403 may include one or more models. In the simplest case, a model is for example a copy of the historical reference data, and in more complex examples the model includes a forecast based on machine learning methods to predict future behavior.

Examples are power forecasts, price forecasts, activation forecasts etc. The models 403 can be regularly updated using reinforcement learning or other appropriate learning methods based on the second feedback 409 from the real system 408.

In the following, an example of the simulation is discussed. The simulation may relate to the — step 303 of Fig. 3 or to operation of the simulator 405 of Fig. 4. - The historical reference data and the DES infrastructure data are obtained. These define the operating context. - Operating objective is identified. This may be predefined by an expert. - Initial operating plan is obtained. This may be preselected by an expert. - Grid balancing bids of the operating plan are considered in the operating context. The bids can be placed e.g. in up and/or down direction and each bid shall contain a magnitude e.g. 5 MW and time period for the bid(s) e.g. 13:00.

AN - The simulator simulates the next day to determine what would happen with such

N - - - - -

S operating plan. Resulting feedback or output may comprise for example information se 25 about acceptance of bids, activation of bids, battery condition of the DES devices, = failures to comply with activations or other parameters.

E - Based on the feedback or output and the operating objective, it is decided whether to o select the operating plan or whether to further optimize the operating plan. There

N may be for example some threshold values or other criteria for the feedback that

N 30 need to be fulfilled in view of the operating objective to select the operating plan.

N Additionally or alternatively, the simulation may be run for different operating plans and the results may be compared to find most suitable operating plan out of the simulated operating plans. The comparison may be performed in view of the operating objective, i.e. to optimize the aspects defined in the operating objective.

In this way, the operating plan may be selected for example based on most efficient utilization of the resources of the virtual power plant, least number of failures or any other suitable logic. - If another simulation is to be run to further optimize the operating plan, the system updates inputs to the simulator and the simulator is run again. - If no more simulations are to be run, the selected operating plan is used for the next day real world bidding.

In an example case, the second feedback 409 from the real system of Fig. 4 is used as follows: - Firstly, the second feedback is used for adjusting the operating plans that are simulated in the simulator 405. The second feedback and the second feedback loop are based on real data rather than simulated data, and the second feedback reflects the operation of the actual real world system. For example, if the trend in the second feedback shows an unbalance between up and down bids, the ratio between up and down bids can be tuned in future operating plans. In another example, there is a trend of overbidding on Fridays at 12:00 and underbidding on Mondays at 10:00, then the future operating plans can be tuned according to these observations. - Secondly, the second feedback is used for updating the model 403. Updating the model may result in more accurate operating plans and may keep the model up-to- date with real-world situations that affect the operation of the system. Updating the model can be performed for example by re-executing training phase of machine learning methods responsible for the model or by updating parameters in the model 3 25 training. Updating the model can be done for example once a day, once a week,

O once a month, or even more infreguently. o = - Clearly, if the second feedback does not indicate need to adjust the operating plans = or the model, adjustments are not performed and the simulations for the following a > operating plan continue based on previous settings.

E 30 Without in any way limiting the scope, interpretation, or application of the appended claims,

N a technical effect of one or more of the example embodiments disclosed herein is improved

N control of a virtual power plant. Further, one or more of the example embodiments disclosed herein provide possibility to improve operating and bidding strategies in the context of virtual power plants. The automated simulations provided by various embodiments allows easier determination of suitable operating plans, whilst ensuring desired operating objectives of the virtual power plant.

Any of the afore described methods, method steps, or combinations thereof, may be controlled or performed using hardware; software; firmware; or any combination thereof.

The software and/or hardware may be local; distributed; centralised; virtualised; or any combination thereof. Moreover, any form of computing, including computational intelligence, may be used for controlling or performing any of the afore described methods, method steps, or combinations thereof. Computational intelligence may refer to, for example, any of artificial intelligence; neural networks; fuzzy logics, machine learning; genetic algorithms; evolutionary computation; or any combination thereof.

Various embodiments have been presented. It should be appreciated that in this document, words comprise; include; and contain are each used as open-ended expressions with no intended exclusivity.

The foregoing description has provided by way of non-limiting examples of particular implementations and embodiments a full and informative description of the best mode presently contemplated by the inventors for carrying out the invention. It is however clear to a person skilled in the art that the invention is not restricted to details of the embodiments presented in the foregoing, but that it can be implemented in other embodiments using equivalent means or in different combinations of embodiments without deviating from the characteristics of the invention.

Furthermore, some of the features of the afore-disclosed example embodiments may be used to advantage without the corresponding use of other features. As such, the foregoing description shall be considered as merely illustrative of the principles of the present

N 25 invention, and not in limitation thereof. Hence, the scope of the invention is only restricted a by the appended patent claims.

O

<Q n

I a a o

N

LO

N

O

N

Claims

1. A computer implemented method for controlling a virtual power plant, VPP, comprising a plurality of spatially distributed energy storage, DES, devices, the method comprising obtaining (301) historical reference data and DES infrastructure data; identifying (302) an operating objective for the virtual power plant; simulating (303) operation of the virtual power plant over a first time period based on the historical reference data and the DES infrastructure data; selecting (304) an operating plan for the first time period based on the simulation in — view of the operating objective, wherein the operating plan concerns utilization of the virtual power plant for grid balancing; using (305) the selected operating plan and responsively monitoring (306) operation of the virtual power plant over the first time period; and using (307) results from the monitoring to adjust future simulations.

2. The method of any preceding claim, wherein the operating objective comprises one or more of: optimizing utilization of resources of the virtual power plant, minimizing wear of the DES devices, minimizing failures to comply with grid balancing activations, and minimizing failures in operation of the DES infrastructure.

3. The method of any preceding claim, wherein the historical reference data comprises historical information about one or more of: up or down regulation requests, up or down regulation activations, up or down regulation pricing, energy pricing. N 25

4. The method of any preceding claim, wherein the DES infrastructure data N comprises one or more of the following: power consumption, capacity, wear, and physical 3 properties of the DES devices of the virtual power plant. n j

5. The method of any preceding claim, further comprising performing the R 30 simulation for plurality of different operating plans and/or performing the simulation for N plurality of different combinations of historical reference data and DES infrastructure data. O

N

6. The method of any preceding claim, further comprising using the selected operating plan for participating in intraday trading market.

7. The method of any preceding claim, wherein the simulation comprises iteratively optimizing the operating plan in view of the operating objective based on feedback from earlier simulation.

8. The method of claim 6, wherein said iteration is performed in the scope of hours.

9. The method of any preceding claim, wherein said simulation comprises a — machine learning model configured to predict future operation of the virtual power plant.

10. The method of claim 9, wherein the machine learning model is configured to predict plurality of possible operating scenarios.

11. The method of claim 9 or 10, wherein said machine learning model is configured to predict one or more of: up or down regulation requests, up or down regulation activations, up or down regulation pricing, energy pricing.

12. The method of any preceding claim, wherein said results from the monitoring are used to adjust future simulations in the scope of days, weeks or months.

13. An apparatus (20, 111) comprising means for performing the method of any one a of claims 1-12. O N 2 - 25

14. The apparatus (20, 111) of claim 13, wherein the means comprise a processor O - and a memory including computer program code, and wherein the memory and the n. computer program code are configured to, with the processor, cause the performance of NG the apparatus. N LO N O N 30

15. A computer program comprising computer executable program code (23) for causing an apparatus to perform the method of any one of claims 1-12.