EnerPlanET to MEME

config/enerplanet-to-meme.json, the mapping embedded in the package, converts the calculation payload EnerPlanET's backend sends to its simulation webservice into a MEME job, the body of MEME's POST /simulate. This page records what it maps, the decisions behind it and what the reverse direction restores.

The two documents

Source: the CalculationPayload built by enerplanet/backend/internal/payload/payload.go. Its topology is an array of connections; each connection has a from feature, and, unless the building stands alone, a to feature, a length in km and a pipe kind (lv building to transformer, mv transformer to transformer). A feature is a GeoJSON point with properties (identity, usage class, annual demands) and a techs object keyed by the technology catalogue (pv_supply, battery_storage, ...) whose entries hold Calliope 0.6 constraint names in kW, kWh and EUR/kW. The simulation engine's servicehub/schemas/calliope.json describes the shape.

Target: MEME's canonical model (internal/model in the MEME repository): nodes, technologies, transmission and trade keyed by identifiers, quantities in MW, MWh and EUR/MW, plus an experiment. MEME decodes the body with unknown fields disallowed, so the job may contain nothing MEME does not define.

Decisions

• A node per feature, an arc per connection. Every building and transformer becomes a node keyed by its feature id with the prefix n (n1, ntrafo_2585); every connection becomes a bidirectional transmission arc keyed by pipe kind and position (lv_0, mv_2) with the length as distance. The topology is therefore preserved exactly and the reverse can rebuild it. (The translator on EnerPlanET's feature/heat-integration branch instead sums buildings into one node per transformer area; that aggregation cannot be inverted, so T1K keeps the features.)
• Identifiers that Calliope accepts. Calliope 0.7 requires identifiers to match ^[^_^\d][\w]*$: no leading digit or underscore, and only letters, digits and _. EnerPlanET numbers its buildings, so node ids get the prefix n, and every generated key joins its parts with _ (demand_1, pv_supply_1, grid_trafo_2585, households_1). The node references copied into transmission arcs get the same prefix through the affix converter and are checked by a match guard, so an id that cannot be made valid fails the transformation rather than MEME's Calliope target. The reverse strips the prefix, so EnerPlanET's ids come back unchanged. This answers issue #1.
• The transformer is the grid connection. As in the simulation engine, which attaches a transformer_supply technology capped by the rated power, each transformer gets a trade entry importing at most its rating (kVA as MW) at 300 EUR/MWh, the tariff of that template technology, with unpaid export. households_supply and non_households_supply attached to a building become a trade at the building with their own tariff.
• kW to MW, divided by 1000, for capacities, storage sizes and the transformer rating; EUR/kW, EUR/kWh and EUR/kW/a are multiplied by 1000.
• Annual demand as average power. demand_energy (kWh/a) becomes a constant demand_profile in MW (divided by 8760 h and 1000), rounded to nine decimals; the reverse rounds back to whole kWh. A time series can be put in its place by pointing the rule at model.timeseries.
• Run settings from the Calliope template: mode plan, objective min_cost, solver highs, unmet demand allowed at a penalty of 1e6 EUR/MWh (ensure_feasibility with bigM: 1e6).
• Electricity only. Heat demand (demand_heat) is carried by no rule; the reverse writes 0.

Identifiers

MEME accepts letters, digits, ., _ and - in identifiers, but its Calliope target applies Calliope's stricter rule, ^[^_^\d][\w]*$. The mapping therefore generates every key and reference as follows:

EnerPlanET	MEME key	MEME references
building id 1	node n1	node: "n1" in its technologies and trades; from/to: "n1" in arcs
transformer id trafo_2585	node ntrafo_2585	node: "ntrafo_2585" in its trade; from/to in arcs
demand of building 1	technology demand_1
techs.pv_supply of building 1	technology pv_supply_1
techs.households_supply of building 1	trade households_1
transformer trafo_2585 rating	trade grid_trafo_2585
connection at position 0 with pipe lv	arc lv_0

The prefix comes from the key template model.nodes{n$k} and the templates n$k of the node references; the references copied from the topology (from.id, to.id) get it through the affix converter followed by a match guard. An id that is still invalid after the prefix, for example one containing a hyphen, fails the transformation with the rule named. The reverse strips the prefix, so EnerPlanET's ids return unchanged; a job that references nodes without the prefix is rejected by the reverse rather than guessed at.

Field mapping

kW→MW means linear with divisor 1000, ×1000 the factor 1000, and inf→absent the absent converter for "inf" with reverse_default: "inf".

Payload root

EnerPlanET	MEME	Conversion
model_id	model.metadata.name	copy
start_date, end_date	model.time.start, model.time.end	datetime from 2006-01-02T15:04:05.000Z to RFC 3339
resolution (minutes)	model.time.resolution	lookup: 15 15min, 30 30min, 60 1H
	model.metadata.description	constant
	model.carriers.electricity	constant {"name": "Electricity", "unit": "MWh"}
	experiment.mode, .objective, .solver.name, .allow_unmet_demand	constants plan, min_cost, highs, {enabled, penalty_price 1e6}
user_id, session_id, callback_url, country, lkr	none	reverse writes ""
pypsa	none	reverse writes EnerPlanET's default line and transformer types
custom_demand_time_series, parameters	none	not carried

Connections

Each topology[$i] with a to feature becomes model.transmission{${p}_$i} with $p bound to pipe:

EnerPlanET	MEME	Conversion
pipe	the key prefix (lv_, mv_)	bind
from.id, to.id	from, to	affix prefix n, then match against Calliope's identifier syntax
length (km)	distance	number
	carrier, bidirectional, capacity	constants electricity, true, {"expandable": true}

Features

Each topology[$i].from and topology[$i].to becomes model.nodes{n$k} with $k bound to the feature id; a transformer referenced by several connections is written once.

EnerPlanET	MEME	Conversion
id	the node key	bind
geometry.coordinates[0], [1]	coords.lon, coords.lat	number
properties.osm_id	name	copy
properties.demand_energy (kWh/a, when > 0)	technologies{demand_$k} with role demand, node n$k, carrier_in electricity, demand_profile in MW	÷ 8 760 000, rounded
properties.rated_power (kVA, transformers)	trade{grid_$k} with node n$k, import.limit in MW, import.price 300, export	kW→MW
type, geometry.type, techs, custom_demand_timeseries, properties.id, session_id, created_at, modified_at, f_class_demands, demand_heat	none	reverse constants: Feature, Point, null, null, the key, "", null, null, null, 0
properties.feature_type, type, f_class, demand_profile, area	none	reverse: BasePOI, residential, residential, residential, 0 for a node without a grid_* trade; TopologyNode, transformer, null area for one with

residential is EnerPlanET's own fallback when a building has no usage class.

Technologies

techs.<key> of a feature becomes technologies{<key>_$k} (or trade{..._$k} for the grid tariffs), with node set to n$k, the feature's node key. The shared rules, applied to every supply technology and the battery:

EnerPlanET (kW, kWh, EUR)	MEME (MW, MWh, EUR)	Conversion
	capacity.expandable	constant true
cont_energy_cap_max	capacity.max	inf→absent, kW→MW
cont_energy_cap_min	capacity.min	kW→MW
cont_energy_cap_max_systemwide	capacity.systemwide_max	inf→absent, kW→MW
cont_lifetime	lifetime	number
cost_interest_rate	interest_rate	number
cost_energy_cap	costs.monetary.investment_per_capacity	×1000
cost_om_annual	costs.monetary.fixed_om	×1000
cost_om_annual_investment_fraction	costs.monetary.fixed_om_fraction	0→absent (MEME allows fixed_om or the fraction, not both); reverse default 0
cost_om_prod	costs.monetary.variable_om	×1000
cost_om_con	costs.monetary.fuel_cost	×1000
cost_purchase	costs.monetary.purchase	number

Per technology:

Key	MEME	Specific fields
pv_supply	supply, performance physics pv	tilt, azimuth, losses→params.tilt, .azimuth, .loss; cont_energy_eff→efficiency; cont_parasitic_eff→native.calliope.flow_out_parasitic_eff
battery_storage	storage	cont_energy_eff→storage.charge_eff and .discharge_eff; cont_storage_cap_max/_min→storage.energy_capacity.max/.min (kWh→MWh, inf→absent); cont_storage_loss→self_discharge; cont_storage_initial→initial_soc; cont_storage_discharge_depth→depth_of_discharge; cost_storage_cap→investment_per_energy_capacity (×1000); energy_capacity.expandable true
wind_onshore	supply, performance physics wind	hub_height, rotor_diameter, nominal_power, turbine_id→params.*
biomass_supply, geothermal_supply, water_supply	supply	shared rules and efficiency only
households_supply, non_households_supply	trade{households_$k}, trade{non_households_$k}	cost_om_con→import.price (×1000); cont_energy_cap_max→import.limit (inf→absent, kW→MW)

Not carried

PySAM inputs of pv_supply (system_capacity, module_type, inverter_type, optimize_orientation, inv_eff, dc_ac_ratio) and the process parameters of biomass and geothermal exist to generate resource profiles outside the optimiser; MEME has no PySAM, so they have no counterpart. cont_resource_unit, cont_energy_cap_scale, cont_export_cap, cost_export and cont_resource_eff likewise have none. A PV or wind profile must therefore be supplied to MEME separately, as a precomputed series or through TentaCron's resolvents.

MEME target compatibility

The produced job passes MEME's decoding and model validation. The per-target gates then apply:

Target	Result
calliope	accepted; warnings that highs is replaced by CBC and that the unmet-demand penalty is approximated by bigM
pypsa	rejected: PyPSA does not support allow_unmet_demand; remove that constant rule for this target
adopt-net0	rejected: AdOpT-NET0 has no transmission mapping in MEME yet

The reverse direction

t1k -reverse rebuilds the calculation payload from a job. Everything the tables above mark as a copy or conversion comes back exactly (numbers to their digits, kWh rounded to whole numbers, node ids with their prefix stripped); the reverse constants restore the payload's required keys with the placeholders listed. Three limits:

• A building without a connection has a node and a demand but no transmission arc, so the reverse has nowhere to place it in topology; it is not restored.
• user_id, session_id, callback_url, country and lkr are not in a MEME job; the reverse writes empty strings, and the caller must set them (the simulation engine's schema requires a non-empty country).
• A transformer whose rating is 0 gets no grid trade, and the reverse, which tells transformers from buildings by that trade, restores it as a building.

examples/ holds the payload, the job and the reverse output; pkg/t1k/testdata/ pins both directions, and the tests check that converting the reverse output forward again reproduces the job apart from the standalone building.

Adapting the mapping

t1k -print-config > my-mapping.json writes the embedded mapping out. Typical edits: change the tariff constants, drop the unmet-demand rule for PyPSA, replace the constant demand by a series (demand_profile naming a model.timeseries entry), or add rules for further building properties. To change the node prefix, edit it in the three places that agree on it: the key template model.nodes{n$k} of the two node rules, the n$k templates of the node references, and the affix prefix of the arc references. t1k -config my-mapping.json runs the edited copy; the configuration reference describes every construct.