Index

frequenz.sdk.microgrid

A microgrid is a local electrical grid that connects a set of electrical components together. They are often built around a passive power consumer, to supplement the electricity consumed from the public grid with on-site power generation or storage systems.

Microgrids can also function in island-mode, without a grid connection, or without a local power consumer, but they have to have at least one of the two, to be meaningful.

Frequenz SDK Microgrid Model

The SDK aims to provide an abstract model of the microgrid that enables high-level interactions with microgrid components, without having to worry about (or even be aware of) location-specific details such as:

• where the meters are placed,
• how many batteries,
• whether there's a grid connection or a passive consumer,
• what models the inverters are, etc.
• whether components are having downtimes, because metrics and limits get adjusted automatically when components are having downtimes.

Users of the SDK can develop applications around this interface once and deploy anywhere, and the SDK will take care of translating the requests and instructions to correspond to the specific microgrid configurations.

flowchart LR

subgraph Left[Measurements only]
direction LR
  grid["Grid Connection"]
  consumer["Consumer"]
  pv["PV Arrays"]
  chp["CHP"]
end

junction(( ))

subgraph Right[Measurements and control]
direction LR
  bat["Batteries"]
  ev["EV Chargers"]
end

grid --- junction
consumer --- junction
pv --- junction
chp --- junction

junction --- bat
junction --- ev

Grid

This refers to a microgrid's connection to the external Grid. The power flowing through this connection can be streamed through grid_power.

In locations without a grid connection, this method remains accessible, and streams zero values.

Consumer

This is the main power consumer at the site of a microgrid, and often the load the microgrid is built to support. The power drawn by the consumer is available through consumer_power

In locations without a consumer, this method streams zero values.

Producers: PV Arrays, CHP

The total CHP production in a site can be streamed through chp_power. PV Power is available through the PV pool described below. And total producer power is available through microgrid.producer().power.

As is the case with the other methods, if PV Arrays or CHPs are not available in a microgrid, the corresponding methods stream zero values.

PV Arrays

The total PV power production is available through pv_pool's power. The PV pool by default uses all PV inverters available at a location, but PV pool instances can be created for subsets of PV inverters if necessary, by specifying the inverter ids.

The pv_pool also provides available power bounds through the power_status method.

The pv_pool also provides a control method propose_power, which accepts values in the Passive Sign Convention and supports only production.

Batteries

The total Battery power is available through the battery_pool's power. The battery pool by default uses all batteries available at a location, but battery pool instances can be created for subsets of batteries if necessary, by specifying the battery ids.

The battery_pool also provides soc, capacity, temperature and available power bounds through the power_status method.

The battery_pool also provides control methods propose_power (which accepts values in the Passive Sign Convention and supports both charging and discharging), or through propose_charge, or propose_discharge.

EV Chargers

The ev_charger_pool offers a power method that streams the total power measured for all the EV Chargers at a site.

The ev_charger_pool also provides available power bounds through the power_status method.

The ev_charger_pool also provides a control method propose_power, which accepts values in the Passive Sign Convention and supports only charging.

Component pools

The SDK provides a unified interface for interacting with sets of Batteries, EV chargers and PV arrays, through their corresponding Pools.

• Battery pool
• EV charger pool
• PV pool

All of them provide support for streaming aggregated data and for setting the power values of the components.

Streaming component data

All pools have a power property, which is a FormulaEngine that can

• provide a stream of resampled power values, which correspond to the sum of the power measured from all the components in the pool together.

• be composed with other power streams to for composite formulas.

In addition, the battery pool has some additional properties that can be used as streams for metrics specific to batteries: soc, capacity and temperature.

Setting power

All pools provide a propose_power method for setting power for the pool. This would then be distributed to the individual components in the pool, using an algorithm that's suitable for the category of the components. For example, when controlling batteries, power could be distributed based on the SoC of the individual batteries, to keep the batteries in balance.

How to work with other actors

If multiple actors are trying to control (by proposing power values) the same set of components, the power manager will aggregate their desired power values, while considering the priority of the actors and the bounds they set, to calculate the target power for the components.

The final target power can be accessed using the receiver returned from the power_status method available for all pools, which also streams the bounds that an actor should comply with, based on its priority.

Adding the power proposals of individual actors

When an actor A calls the propose_power method with a power, the proposed power of the lower priority actor will get added to actor A's power. This works as follows:

• the lower priority actor would see bounds shifted by the power proposed by actor A.
• After lower priority actor B sets a power in its shifted bounds, it will get shifted back by the power set by actor A.

This has the effect of adding the powers set by actors A and B.

Example 1: Battery bounds available for use: -100kW to 100kW

| Actor | Priority | System Bounds | Requested Bounds | Requested | Adjusted | Aggregate | | | | | | Power | Power | Power | |-------|----------|-----------------|------------------|-----------|--------------|-----------| | A | 3 | -100kW .. 100kW | None | 20kW | 20kW | 20kW | | B | 2 | -120kW .. 80kW | None | 50kW | 50kW | 70kW | | C | 1 | -170kW .. 30kW | None | 50kW | 30kW | 100kW | | | | | | | target power | 100kW |

Actor A proposes a power of 20kW, but no bounds. In this case, actor B sees bounds shifted by A's proposal. Actor B proposes a power of 50kW on this shifted range, and if this is applied on to the original bounds (aka shift the bounds back to the original range), it would be 20kW + 50kW = 70kW.

So Actor C sees bounds shifted by 70kW from the original bounds, and sets 50kW on this shifted range, but it can't exceed 30kW, so its request gets limited to 30kW. Shifting this back by 70kW, the target power is calculated to be 100kW.

Irrespective of what any actor sets, the final power won't exceed the available battery bounds.

Example 2:

| Actor | Priority | System Bounds | Requested Bounds | Requested | Adjusted | Aggregate | | | | | | Power | Power | Power | |-------|----------|-----------------|------------------|-----------|--------------|-----------| | A | 3 | -100kW .. 100kW | None | 20kW | 20kW | 20kW | | B | 2 | -120kW .. 80kW | None | -20kW | -20kW | 0kW | | | | | | | target power | 0kW |

Actors with exactly opposite requests cancel each other out.

Limiting bounds for lower priority actors

When an actor A calls the propose_power method with bounds (either both lower and upper bounds or at least one of them), lower priority actors will see their (shifted) bounds restricted and can only propose power values within that range.

Example 1: Battery bounds available for use: -100kW to 100kW

| Actor | Priority | System Bounds | Requested Bounds | Requested | Adjusted | Aggregate | | | | | | Power | Power | Power | |-------|----------|-----------------|------------------|-----------|--------------|-----------| | A | 3 | -100kW .. 100kW | -20kW .. 100kW | 50kW | 40kW | 50kW | | B | 2 | -70kW .. 50kW | -90kW .. 0kW | -10kW | -10kW | 40kW | | C | 1 | -60kW .. 10kW | None | -20kW | -20kW | 20kW | | | | | | | target power | 20kW |

Actor A with the highest priority has the entire battery bounds available to it. It sets limited bounds of -20kW .. 100kW, and proposes a power of 50kW.

Actor B sees Actor A's limit of -20kW..100kW shifted by 50kW as -70kW..50kW, and can only propose powers within this range, which will get added (shifted back) to Actor A's proposed power.

Actor B tries to limit the bounds of actor C to -90kW .. 0kW, but it can only operate in the -70kW .. 50kW range because of bounds set by actor A, so its requested bounds get restricted to -70kW .. 0kW.

Actor C sees this as -60kW .. 10kW, because it gets shifted by Actor B's proposed power of -10kW.

Actor C proposes a power within its bounds and the proposals of all the actors are added to get the target power.

Example 2:

| Actor | Priority | System Bounds | Requested Bounds | Requested | Adjusted | Aggregate | | | | | | Power | Power | Power | |-------|----------|-----------------|------------------|-----------|--------------|-----------| | A | 3 | -100kW .. 100kW | -20kW .. 100kW | 50kW | 50kW | 50kW | | B | 2 | -70kW .. 50kW | -90kW .. 0kW | -90kW | -70kW | -20kW | | | | | | | target power | -20kW |

When an actor requests a power that's outside its available bounds, the closest available power is used.

Comprehensive example

Battery bounds available for use: -100kW to 100kW

| Priority | System Bounds | Requested Bounds | Requested | Adjusted | Aggregate | | | | | Power | Power | Power | |----------|-------------------|------------------|-----------|--------------|-----------| | 7 | -100 kW .. 100 kW | None | 10 kW | 10 kW | 10 kW | | 6 | -110 kW .. 90 kW | -110 kW .. 80 kW | 10 kW | 10 kW | 20 kW | | 5 | -120 kW .. 70 kW | -100 kW .. 80 kW | 80 kW | 70 kW | 90 kW | | 4 | -170 kW .. 0 kW | None | -120 kW | -120 kW | -30 kW | | 3 | -50 kW .. 120 kW | None | 60 kW | 60 kW | 30 kW | | 2 | -110 kW .. 60 kW | -40 kW .. 30 kW | 20 kW | 20 kW | 50 kW | | 1 | -60 kW .. 10 kW | -50 kW .. 40 kW | 25 kW | 10 kW | 60 kW | | 0 | -60 kW .. 0 kW | None | 12 kW | 0 kW | 60 kW | | -1 | -60 kW .. 0 kW | -40 kW .. -10 kW | -10 kW | -10 kW | 50 kW | | | | | | Target Power | 50 kW |

Classes

Functions

frequenz.sdk.microgrid.consumer

consumer() -> Consumer

Return the Consumption measuring point.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def consumer() -> Consumer:
    """Return the [`Consumption`][frequenz.sdk.timeseries.consumer.Consumer] measuring point."""
    return _get().consumer()

frequenz.sdk.microgrid.frequency

frequency() -> GridFrequency

Return the grid frequency measuring point.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def frequency() -> GridFrequency:
    """Return the grid frequency measuring point."""
    return _get().frequency()

frequenz.sdk.microgrid.grid

grid() -> Grid

Return the grid measuring point.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def grid() -> Grid:
    """Return the grid measuring point."""
    return _get().grid()

frequenz.sdk.microgrid.initialize async

initialize(
    server_url: str,
    resampler_config: ResamplerConfig,
    *,
    api_power_request_timeout: timedelta = timedelta(
        seconds=5.0
    )
) -> None

Initialize the microgrid connection manager and the data pipeline.

PARAMETER	DESCRIPTION
server_url	The location of the microgrid API server in the form of a URL. The following format is expected: grpc://hostname{:port}{?ssl=ssl}, where the port should be an int between 0 and 65535 (defaulting to 9090) and ssl should be a boolean (defaulting to false). For example: grpc://localhost:1090?ssl=true. TYPE: str
resampler_config	Configuration for the resampling actor. TYPE: ResamplerConfig
api_power_request_timeout	Timeout to use when making power requests to the microgrid API. When requests to components timeout, they will be marked as blocked for a short duration, during which time they will be unavailable from the corresponding component pools. TYPE: timedelta DEFAULT: timedelta(seconds=5.0)

Source code in frequenz/sdk/microgrid/__init__.py

async def initialize(
    server_url: str,
    resampler_config: ResamplerConfig,
    *,
    api_power_request_timeout: timedelta = timedelta(seconds=5.0),
) -> None:
    """Initialize the microgrid connection manager and the data pipeline.

    Args:
        server_url: The location of the microgrid API server in the form of a URL.
            The following format is expected: `grpc://hostname{:port}{?ssl=ssl}`,
            where the port should be an int between `0` and `65535` (defaulting to
            `9090`) and ssl should be a boolean (defaulting to false). For example:
            `grpc://localhost:1090?ssl=true`.
        resampler_config: Configuration for the resampling actor.
        api_power_request_timeout: Timeout to use when making power requests to
            the microgrid API.  When requests to components timeout, they will
            be marked as blocked for a short duration, during which time they
            will be unavailable from the corresponding component pools.
    """
    await connection_manager.initialize(server_url)
    await _data_pipeline.initialize(
        resampler_config,
        api_power_request_timeout=api_power_request_timeout,
    )

frequenz.sdk.microgrid.logical_meter

logical_meter() -> LogicalMeter

Return the logical meter of the microgrid.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def logical_meter() -> LogicalMeter:
    """Return the logical meter of the microgrid."""
    return _get().logical_meter()

frequenz.sdk.microgrid.new_battery_pool

new_battery_pool(
    *,
    priority: int,
    component_ids: Set[int] | None = None,
    name: str | None = None
) -> BatteryPool

Return a new BatteryPool instance for the given parameters.

The priority value is used to resolve conflicts when multiple actors are trying to propose different power values for the same set of batteries.

Note

When specifying priority, bigger values indicate higher priority.

It is recommended to reuse the same instance of the BatteryPool within the same actor, unless they are managing different sets of batteries.

In deployments with multiple actors managing the same set of batteries, it is recommended to use different priorities to distinguish between them. If not, a random prioritization will be imposed on them to resolve conflicts, which may lead to unexpected behavior like longer duration to converge on the desired power.

PARAMETER	DESCRIPTION
priority	The priority of the actor making the call. TYPE: int
component_ids	Optional set of IDs of batteries to be managed by the BatteryPool. If not specified, all batteries available in the component graph are used. TYPE: Set[int] | None DEFAULT: None
name	An optional name used to identify this instance of the pool or a corresponding actor in the logs. TYPE: str | None DEFAULT: None

RETURNS	DESCRIPTION
BatteryPool	A BatteryPool instance.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def new_battery_pool(
    *,
    priority: int,
    component_ids: abc.Set[int] | None = None,
    name: str | None = None,
) -> BatteryPool:
    """Return a new `BatteryPool` instance for the given parameters.

    The priority value is used to resolve conflicts when multiple actors are trying to
    propose different power values for the same set of batteries.

    !!! note
        When specifying priority, bigger values indicate higher priority.

        It is recommended to reuse the same instance of the `BatteryPool` within the
        same actor, unless they are managing different sets of batteries.

        In deployments with multiple actors managing the same set of batteries, it is
        recommended to use different priorities to distinguish between them.  If not,
        a random prioritization will be imposed on them to resolve conflicts, which may
        lead to unexpected behavior like longer duration to converge on the desired
        power.

    Args:
        priority: The priority of the actor making the call.
        component_ids: Optional set of IDs of batteries to be managed by the
            `BatteryPool`.  If not specified, all batteries available in the component
            graph are used.
        name: An optional name used to identify this instance of the pool or a
            corresponding actor in the logs.

    Returns:
        A `BatteryPool` instance.
    """
    return _get().new_battery_pool(
        priority=priority, component_ids=component_ids, name=name
    )

frequenz.sdk.microgrid.new_ev_charger_pool

new_ev_charger_pool(
    *,
    priority: int,
    component_ids: Set[int] | None = None,
    name: str | None = None
) -> EVChargerPool

Return a new EVChargerPool instance for the given parameters.

The priority value is used to resolve conflicts when multiple actors are trying to propose different power values for the same set of EV chargers.

Note

When specifying priority, bigger values indicate higher priority.

It is recommended to reuse the same instance of the EVChargerPool within the same actor, unless they are managing different sets of EV chargers.

In deployments with multiple actors managing the same set of EV chargers, it is recommended to use different priorities to distinguish between them. If not, a random prioritization will be imposed on them to resolve conflicts, which may lead to unexpected behavior like longer duration to converge on the desired power.

PARAMETER	DESCRIPTION
priority	The priority of the actor making the call. TYPE: int
component_ids	Optional set of IDs of EV Chargers to be managed by the EVChargerPool. If not specified, all EV Chargers available in the component graph are used. TYPE: Set[int] | None DEFAULT: None
name	An optional name used to identify this instance of the pool or a corresponding actor in the logs. TYPE: str | None DEFAULT: None

RETURNS	DESCRIPTION
EVChargerPool	An EVChargerPool instance.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def new_ev_charger_pool(
    *,
    priority: int,
    component_ids: abc.Set[int] | None = None,
    name: str | None = None,
) -> EVChargerPool:
    """Return a new `EVChargerPool` instance for the given parameters.

    The priority value is used to resolve conflicts when multiple actors are trying to
    propose different power values for the same set of EV chargers.

    !!! note
        When specifying priority, bigger values indicate higher priority.

        It is recommended to reuse the same instance of the `EVChargerPool` within the
        same actor, unless they are managing different sets of EV chargers.

        In deployments with multiple actors managing the same set of EV chargers, it is
        recommended to use different priorities to distinguish between them.  If not,
        a random prioritization will be imposed on them to resolve conflicts, which may
        lead to unexpected behavior like longer duration to converge on the desired
        power.

    Args:
        priority: The priority of the actor making the call.
        component_ids: Optional set of IDs of EV Chargers to be managed by the
            EVChargerPool.  If not specified, all EV Chargers available in the
            component graph are used.
        name: An optional name used to identify this instance of the pool or a
            corresponding actor in the logs.

    Returns:
        An `EVChargerPool` instance.
    """
    return _get().new_ev_charger_pool(
        priority=priority, component_ids=component_ids, name=name
    )

frequenz.sdk.microgrid.new_pv_pool

new_pv_pool(
    *,
    priority: int,
    component_ids: Set[int] | None = None,
    name: str | None = None
) -> PVPool

Return a new PVPool instance for the given parameters.

The priority value is used to resolve conflicts when multiple actors are trying to propose different power values for the same set of PV inverters.

Note

When specifying priority, bigger values indicate higher priority.

It is recommended to reuse the same instance of the PVPool within the same actor, unless they are managing different sets of PV inverters.

In deployments with multiple actors managing the same set of PV inverters, it is recommended to use different priorities to distinguish between them. If not, a random prioritization will be imposed on them to resolve conflicts, which may lead to unexpected behavior like longer duration to converge on the desired power.

PARAMETER	DESCRIPTION
priority	The priority of the actor making the call. TYPE: int
component_ids	Optional set of IDs of PV inverters to be managed by the PVPool. If not specified, all PV inverters available in the component graph are used. TYPE: Set[int] | None DEFAULT: None
name	An optional name used to identify this instance of the pool or a corresponding actor in the logs. TYPE: str | None DEFAULT: None

RETURNS	DESCRIPTION
PVPool	A PVPool instance.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def new_pv_pool(
    *,
    priority: int,
    component_ids: abc.Set[int] | None = None,
    name: str | None = None,
) -> PVPool:
    """Return a new `PVPool` instance for the given parameters.

    The priority value is used to resolve conflicts when multiple actors are trying to
    propose different power values for the same set of PV inverters.

    !!! note
        When specifying priority, bigger values indicate higher priority.

        It is recommended to reuse the same instance of the `PVPool` within the same
        actor, unless they are managing different sets of PV inverters.

        In deployments with multiple actors managing the same set of PV inverters, it is
        recommended to use different priorities to distinguish between them.  If not,
        a random prioritization will be imposed on them to resolve conflicts, which may
        lead to unexpected behavior like longer duration to converge on the desired
        power.

    Args:
        priority: The priority of the actor making the call.
        component_ids: Optional set of IDs of PV inverters to be managed by the
            `PVPool`. If not specified, all PV inverters available in the component
            graph are used.
        name: An optional name used to identify this instance of the pool or a
            corresponding actor in the logs.

    Returns:
        A `PVPool` instance.
    """
    return _get().new_pv_pool(priority=priority, component_ids=component_ids, name=name)

frequenz.sdk.microgrid.producer

producer() -> Producer

Return the Production measuring point.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def producer() -> Producer:
    """Return the [`Production`][frequenz.sdk.timeseries.producer.Producer] measuring point."""
    return _get().producer()

frequenz.sdk.microgrid.voltage_per_phase

voltage_per_phase() -> VoltageStreamer

Return the per-phase voltage measuring point.

Source code in frequenz/sdk/microgrid/_data_pipeline.py

def voltage_per_phase() -> VoltageStreamer:
    """Return the per-phase voltage measuring point."""
    return _get().voltage_per_phase()