PCM Inputs Explanation

The input files for the HOPE model could be one big .XLSX file or multiple .csv files. If you use the XLSX file, each spreadsheet in the file needs to be prepared based on the input instructions below and the spreadsheet names should be carefully checked. If you use the csv files, each csv file will represent one spreadsheet from the XLSX file. If both XLSX file and csv files are provided, the XLSX files will be used.

network_model setting

In HOPE_model_settings.yml, set:

• network_model: 0 for no network constraints (copper plate).
• network_model: 1 for zonal transport.
• network_model: 2 for nodal DCOPF (angle-based).
• network_model: 3 for nodal DCOPF (PTDF-based).

When using nodal modes (2/3), provide nodal mapping via busdata and branchdata (or ensure linedata includes equivalent bus columns). Native nodal load must come from load_timeseries_nodal, and optional bus-level NI can come from ni_timeseries_nodal.

Related run-control settings:

• unit_commitment: 1 makes PCM a MILP.
• write_shadow_prices: 1 triggers MILP -> fixed-LP re-solve for dual/LMP recovery when unit_commitment: 1.
• summary_table: 1 writes summary analytics to output/Analysis/Summary_*.csv.
• transmission_loss: 1 enables a piecewise-linear transmission loss approximation based on absolute line flow. This is currently supported for network_model = 1 and 2, but not 3.

Price-output note:

• HOPE exports power_price*.csv as the marginal objective cost of +1 MW load.
• Because raw equality-constraint duals can flip sign under equivalent reformulations, HOPE applies a formulation-specific sign mapping internally.
• For network_model = 0/1/2, exported price follows the raw dual of the load-balance row.
• For network_model = 3, exported price uses the negative of the raw dual of PTDFInjDef_con.
• This mapping is regression-tested on both the ISO-NE nodal angle case and the RTS24 nodal PTDF case.

Simple workflow for line losses:

1. Keep transmission_loss: 0 if you want the default lossless PCM run.
2. Set transmission_loss: 1 only when your active line table includes a Loss (%) column.
3. Use network_model = 1 or 2 with transmission losses.
4. Keep network_model = 3 with transmission_loss: 0; PTDF mode remains lossless in the current HOPE release.
5. In nodal angle-based PCM (network_model = 2), HOPE now reports a nonzero Loss price component when transmission losses are active.

internal network helper utilities

These functions are implemented in src/network_utils.jl and used by src/PCM.jl:

• first_existing_col: finds the first available column name among aliases (for robust input parsing).
• resolve_reference_index: resolves reference bus/zone from either integer index or ID value.
• compute_ptdf_from_incidence: computes PTDF matrix from branch incidence and reactance.
• compute_zone_ptdf_from_linedata: convenience wrapper to build zonal PTDF from linedata (reserved for a future zonal-PTDF mode; not used in current PCM flow).

They are automatically called in PCM network preprocessing; users do not need to call them directly.

zonedata

This is the input dataset for zone-relevant information (e.g., demand, mapping with state, etc.).

gendata

This is the input dataset for existing generators.

genavailabilitytimeseries (optional)

This optional dataset provides hourly generator-level availability factors for PCM. It can be used for plant-based VRE profiles and for thermal derating.

Accepted names:

• CSV: gen_availability_timeseries.csv
• XLSX sheet: gen_availability_timeseries

Expected generator columns:

• G1, G2, ..., G(N)
• ordering follows the row order of gendata.csv

Required time columns:

• Time Period
• Hours

Optional time columns:

• Month
• Day

Behavior in current PCM:

• If any generator column is provided here, PCM uses that hourly AF for that generator.
• For wind/solar generators with no hourly G# column, PCM falls back to zonal wind_timeseries_regional or solar_timeseries_regional.
• For non-VRE generators with no hourly G# column, PCM falls back to the static AF column in gendata.
• For thermal generators, the hourly AF only derates the generation upper bound. PCM keeps reserve, ramp, and UC availability logic on the original 1-FOR basis.

So this file is now the main generator-level availability override for PCM, with zonal VRE profiles and static generator AF used as fallbacks when a generator-specific hourly column is not supplied.

linedata

This is the input dataset for existing transmission lines (e.g., transmission capacity limit for each inter-zonal transmission line).

busdata (optional for zonal modes, recommended for nodal modes)

This dataset defines nodal buses and bus-to-zone mapping.

Required columns for nodal modes:

• Bus_id
• Zone_id
• one nodal peak-load basis column:
  • Demand (MW), or
  • Load (MW), or
  • Pd, or
  • PD, or
  • Load_share

Optional columns:

• State for nodal state-policy accounting. When network_model is 2 or 3, PCM now uses bus-to-state mapping for carbon/RPS accounting if this column is provided. If omitted, PCM falls back to the state of the bus's assigned zone.

CSV name: busdata.csv XLSX sheet name: busdata

branchdata (optional; used preferentially in nodal modes)

This dataset defines nodal transmission branches.

Recommended columns:

• from_bus/to_bus (or MATPOWER-style F_BUS/T_BUS)
• Capacity (MW) (line thermal limit)
• X or Reactance (for DC angle/PTDF physics)
• Loss (%) (optional; supported when network_model = 2 and transmission_loss = 1)
• delta_theta_max (optional): per-line max angle difference (radians). If omitted/<=0, disabled.

If branchdata is provided and network_model is nodal, HOPE uses it as network branch input.

PTDF note:

• network_model = 3 uses a lossless PTDF-based DCOPF. Keep transmission_loss: 0 in that mode.

CSV name: branchdata.csv XLSX sheet name: branchdata

ptdfmatrixnodal / ptdf_matrix (optional)

This optional dataset is used only when network_model = 3 (nodal PTDF-based DCOPF).

If this file/sheet is provided, HOPE reads the PTDF matrix directly. If not provided, HOPE computes nodal PTDF from branch endpoints/reactance and reference_bus.

Required format:

• One row per line, in the same row order as the active branch table (branchdata if provided in nodal mode, otherwise linedata).
• One column per bus, with column names exactly matching busdata.Bus_id (or inferred bus labels).
• Cell value = PTDF coefficient for that line and bus (reference bus convention).

Accepted names:

• CSV: ptdf_matrix_nodal.csv (preferred) or ptdf_matrix.csv
• XLSX sheet: ptdf_matrix_nodal (preferred) or ptdf_matrix

storagedata

This is the input dataset for existing energy storage units (e.g., battery storage and pumped storage hydropower).

solartimeseriesregional

This is the input dataset for the annual hourly solar PV generation profile in each zone. Each zone has 8760 data points and the values are per unit.

Even if gen_availability_timeseries is provided, this zonal file is still used as the fallback profile for any solar generator that does not have its own G# column in the generator-level AF input.

windtimeseriesregional

This is the input dataset for the annual hourly wind generation profile in each zone. Each zone has 8760 data points and the values are per unit.

Even if gen_availability_timeseries is provided, this zonal file is still used as the fallback profile for any wind generator that does not have its own G# column in the generator-level AF input.

loadtimeseriesregional

This is the input dataset for the annual hourly load profile in each zone. Each zone has 8760 data points and the values are per unit.

In current HOPE PCM:

• network_model = 0 or 1: this file is used directly for load modeling
• network_model = 2 or 3: this file is still required as the chronology anchor for aligned time-series inputs, but native nodal load must come from load_timeseries_nodal
• NI in this file remains the system/zonal NI input and the fallback NI source for nodal PCM when ni_timeseries_nodal is not provided

loadtimeseriesnodal (required for nodal PCM)

This dataset provides the annual hourly nodal load profile for nodal PCM. It is required when:

• network_model = 2
• network_model = 3

HOPE interprets each bus column as a unitless hourly multiplier and builds native bus load as:

• Bus peak load * nodal hourly multiplier

The bus peak-load basis comes from busdata, then HOPE rescales it within each zone so nodal peak loads sum to the zone peak load in zonedata.

Accepted names:

• CSV: load_timeseries_nodal.csv
• XLSX sheet: load_timeseries_nodal

Required time columns:

• Time Period
• Hours

Accepted legacy aliases:

• Period -> Hours
• Hour -> Hours

Optional time columns:

• Month
• Day

Required load columns:

• one column per bus
• column names must match busdata.Bus_id exactly

Example:

Important behavior:

• HOPE no longer falls back to zonal hourly load * static bus share in nodal modes.
• If load_timeseries_nodal is missing in nodal PCM, HOPE will stop with an input error.

nitimeseriesnodal (optional for nodal PCM)

This optional dataset provides the annual hourly bus-level NI injections for nodal PCM. It is used only when:

• network_model = 2
• network_model = 3

Each bus column is interpreted directly in MW:

• positive value = net import injection into the modeled system at that bus
• negative value = net export withdrawal from the modeled system at that bus

Accepted names:

• CSV: ni_timeseries_nodal.csv
• XLSX sheet: ni_timeseries_nodal

Required time columns:

• Time Period
• Hours

Accepted legacy aliases:

• Period -> Hours
• Hour -> Hours

Optional time columns:

• Month
• Day

Required NI columns:

• one column per bus
• column names must match busdata.Bus_id exactly

Example:

Behavior:

• If ni_timeseries_nodal is provided, HOPE uses it directly in nodal power balance instead of allocating load_timeseries_regional.NI by bus load share.
• HOPE still reads load_timeseries_regional.NI; when both inputs are present, HOPE compares the hourly row sums and uses the nodal NI input as authoritative for nodal PCM.
• When ni_timeseries_nodal is omitted, nodal PCM falls back to the legacy load-share allocation of system NI.
• In the ISO-NE 250-bus example, ni_timeseries_nodal is generated in preprocessing from official ISO-NE interface chronology, then scaled to the case-level NI magnitude and distributed with interface-centered nodal weights plus a small load-share balancing tail.

nitimeseriesnodaltarget / nitimeseriesnodalcap (optional pair for flexible nodal PCM)

These two optional datasets activate flexible nodal NI in nodal PCM:

• network_model = 2
• network_model = 3

Accepted names:

• CSV: ni_timeseries_nodal_target.csv, ni_timeseries_nodal_cap.csv
• XLSX sheets: ni_timeseries_nodal_target, ni_timeseries_nodal_cap

Interpretation:

• ni_timeseries_nodal_target is the preferred hourly bus-level NI profile.
• ni_timeseries_nodal_cap is the hourly bus-level NI bound the solver may move toward when internal generation and network limits make the target infeasible or too expensive.

Both datasets use the same time-column rules and bus-column rules as ni_timeseries_nodal.

Required behavior:

• both files/sheets must be present together
• bus columns must match busdata.Bus_id
• abs(target) must not exceed abs(cap) at any bus-hour
• nonzero target/cap entries must have consistent signs at a bus-hour

Model behavior:

• when both target/cap inputs are present, HOPE ignores fixed ni_timeseries_nodal for nodal balance and instead optimizes actual nodal NI between the target and cap bounds
• deviation from the target is penalized in the objective using PT_NI_DEV
• outputs include:
  • power_ni_nodal.csv
  • power_ni_zonal.csv
  • power_ni_deviation_nodal.csv

The ISO-NE 250-bus example uses this workflow to treat official ISO-NE interchange as the lower NI target while preserving a higher synthetic NI cap needed to keep the case workable.

repperiodweights (required when external_rep_day = 1)

This optional-to-required dataset is used when users provide their own clustered representative periods for PCM.

Accepted names:

• CSV: rep_period_weights.csv
• XLSX sheet: rep_period_weights

Required columns:

• Time Period
• Weight

Behavior:

• If external_rep_day = 1, PCM requires this file/sheet.
• Time Period values must match the Time Period values used in load_timeseries_regional and the other aligned time-series inputs.
• Weight is the annual scaling/count for each representative period.

This matches the same external representative-period workflow used by GTEP.

flexddata (required when flexible_demand = 1)

This dataset defines DR resources for PCM backlog load shifting (dr_DF, dr_PB, b_DR over resource set R).

drtimeseriesregional (required when flexible_demand = 1)

This is the annual hourly DR availability profile in each zone (per unit).

carbonpolicies

This is the input dataset for carbon policies.

rpspolicies

This is the input dataset for renewable portfolio standard (RPS) policies. It defines renewable credits trading relationship between different states (i.e., the states must be neighboring states) and the renewable credit requirement for each state.

single parameters

This is the input dataset for some parameters that can be directly defined based on users' need. If not changed, they remain with default values.

Implementation note: PCM reads these fields through an internal helper (get_singlepar) that returns the file value when present, otherwise a built-in default. This keeps backward compatibility when older cases do not include newly added single-parameter columns.

Notes:

• planning_reserve_margin and Inv_bugt_* are currently not used by create_PCM_model (they are relevant to GTEP workflows).