Grid Connection and Availability¶

MicroGridsPy can represent a weakly grid-connected mini-grid in addition to the fully isolated off-grid configuration. When grid connection is enabled, the system can import electricity and, optionally, export surplus. Grid interaction is constrained by the physical exchange capacity of the interconnection line and a time-dependent grid-availability matrix capturing outages and upstream reliability.

Point-of-common-coupling convention

Grid import and export variables represent raw energy exchanged at the point of common coupling (PCC) — the electricity crossing the grid interface before internal transmission/conversion efficiency is applied. Grid efficiency \(\eta^{\text{grid}}\) is therefore accounted for in the energy balance and in grid-related emissions, but not in the line-capacity constraints.

At hourly resolution \(\Delta t = 1\,\text{h}\), line capacity (kW) and per-step exchange (kWh) coincide. Let \(E^{\text{imp}}\) and \(E^{\text{exp}}\) be imported/exported electricity at the PCC. Grid exchange is governed by: the binary availability \(A^{\text{grid}}\in\{0,1\}\); the line capacity \(\overline{P}^{\text{grid}}\); the transmission/conversion efficiency \(\eta^{\text{grid}}\); the allow_export switch; and, in multi-year, first_year_connection. The relevant dimensions are \((t,\omega)\) for typical-year and \((t,y,\omega)\) for multi-year.

Grid import and export capacity¶

Import (and, if enabled, export) is bounded by line capacity and availability. Typical-year:

\[ 0 \le E^{\text{imp}}_{t,\omega} \le A^{\text{grid}}_{t,\omega}\,\overline{P}^{\text{grid}}, \qquad 0 \le E^{\text{exp}}_{t,\omega} \le A^{\text{grid}}_{t,\omega}\,\overline{P}^{\text{grid}} \qquad \forall t,\omega \]

Multi-year (with explicit year indexing):

\[ 0 \le E^{\text{imp}}_{t,y,\omega} \le A^{\text{grid}}_{t,y,\omega}\,\overline{P}^{\text{grid}}, \qquad 0 \le E^{\text{exp}}_{t,y,\omega} \le A^{\text{grid}}_{t,y,\omega}\,\overline{P}^{\text{grid}} \qquad \forall t,y,\omega \]

Exchange is thus possible only when the upstream grid is available and within the rated interconnection capacity. Note that \(\eta^{\text{grid}}\) does not appear here — capacity limits apply to raw PCC flows.

Grid efficiency in the energy balance¶

Because grid variables are defined at the PCC, the electricity effectively received by (or delivered from) the mini-grid is adjusted by \(\eta^{\text{grid}}\). The relevant terms in the energy balance are \(+\eta^{\text{grid}} E^{\text{imp}}\) and \(-\eta^{\text{grid}} E^{\text{exp}}\): imported electricity is reduced by internal losses before serving demand, and exported electricity is similarly adjusted before leaving the system.

Grid cost and emissions accounting¶

Grid electricity is costed on the raw PCC variables. Import cost and export revenue (typical-year; multi-year adds a year index):

\[ C^{\text{grid,imp}}_{\omega} = \sum_t E^{\text{imp}}_{t,\omega}\,\pi^{\text{imp}}_{t,\omega}, \qquad R^{\text{grid,exp}}_{\omega} = \sum_t E^{\text{exp}}_{t,\omega}\,\pi^{\text{exp}}_{t,\omega} \]

entering the scenario-weighted annual cost as \(+\,C^{\text{grid,imp}}_{\omega} - R^{\text{grid,exp}}_{\omega}\) (discounted with the other cash flows in multi-year). Grid-related scope-2 emissions are instead based on the delivered imported electricity — the efficiency-adjusted quantity:

\[ \text{CO}_2^{\text{grid}}_{\omega} = \left( \sum_t \eta^{\text{grid}} E^{\text{imp}}_{t,\omega} \right) \phi^{\text{grid}}_{\omega} \]

where \(\phi^{\text{grid}}\) is the emissions factor of imported electricity.

Grid transmission efficiency

The model distinguishes what crosses the grid connection point from what is delivered inside the mini-grid. If 10 kWh are imported and \(\eta^{\text{grid}}=0.95\): the import cost is charged on the full 10 kWh at the PCC; only 9.5 kWh reach the internal balance; and scope-2 emissions are computed on those delivered 9.5 kWh. This convention is internally consistent — transmission/conversion losses occur between the grid interface and the mini-grid balance.

Grid-availability simulation¶

For realistic weak-grid modelling, MicroGridsPy generates an exogenous availability matrix \(A^{\text{grid}}\) from user-defined outage statistics in grid.yaml. The result is stored as grid_availability.csv, a derived backend artifact, not a primary user input. The inputs are average_outages_per_year, average_outage_duration_minutes, and (multi-year) first_year_connection.

In the typical-year case the matrix has dimensions \(|\mathcal{T}|\times|\Omega|\) with \(A^{\text{grid}}_{t,\omega}\in\{0,1\}\). In the multi-year case it has dimensions \(|\mathcal{T}|\times|\Omega|\times|Y|\) with \(A^{\text{grid}}_{t,y,\omega}\in\{0,1\}\); if first_year_connection is specified, all years before connection are forced to zero availability.

The stochastic outage process follows a simple alternating scheme:

Time between outages (TBO) is sampled from a Weibull distribution;
Outage duration (OD) is sampled from a second Weibull distribution;
the two are alternated to generate a binary availability sequence over the horizon.

The Weibull samples are rescaled so the final sequence matches, in expectation, the target outage frequency and duration. Special cases: if the expected frequency or duration is zero, the grid is perfectly available after connection; if the connection year lies beyond the horizon, the grid remains unavailable throughout.

Stochastic grid-availability generation: sampled time-between-outages and outage-duration distributions with fitted Weibull curves, and the resulting binary hourly availability trajectory over a year

The stochastic grid-availability generation process. Top left: the sampled distribution of time between outages (TBO) versus the fitted Weibull distribution. Top right: the sampled outage-duration (OD) distribution. Bottom: the resulting binary hourly availability trajectory (1 = available, 0 = outage) over a representative year. This lets the model represent fully reliable, fully unavailable, and weakly connected grids with realistic interruption patterns.