[Bill-Becker]

This pull request introduces a new feature to constrain technology sizing based on a maximum simple payback period, along with supporting changes throughout the codebase. The main changes include extending the financial model to support a max_simple_payback_years constraint, updating technology selection logic, adding a lexicographic (maximize-size, then minimize-lifecycle-cost) solve, and updating results reporting and testing.

Financial model and tech selection enhancements:

• Added max_simple_payback_years and bau_year_one_operating_cost fields to the Financial struct, enabling users to specify a maximum allowed simple payback period for new technologies. bau_year_one_operating_cost comes from a prior BAU solve's results["Financial"]["year_one_operating_cost_before_tax_model"] and is required whenever max_simple_payback_years is set. (src/core/financial.jl)
• Extended the Techs struct and related logic to include a maximize_size field, which tracks technologies eligible for maximization under the payback constraint (all techs except ExistingBoiler/ExistingChiller). (src/core/types.jl, src/core/techs.jl)
• BAUScenario now clears both max_simple_payback_years and bau_year_one_operating_cost, so the BAU reference solve is never itself payback-constrained. (src/core/bau_scenario.jl)

Model building (build_reopt!):

• Builds an undiscounted Year1OperatingCost expression (elec bill, CHP standby charges, fixed/variable/per-hour O&M, GHP and storage O&M, and fuel costs) the same way regardless of whether max_simple_payback_years is set, so a BAU solve's value(m[:Year1OperatingCost]) is directly comparable to the optimal model's.
• When max_simple_payback_years is set, adds PaybackSavings (BAU year-one operating cost minus the model's year-one operating cost) and constrains InitialCapexNoIncentives <= max_simple_payback_years * PaybackSavings, with a small absolute tolerance so floating-point noise can't turn a satisfiable case (e.g. all new techs at size 0) into a false infeasibility.
• Builds a MaximizeSizeObjective expression over dvSize for techs.maximize_size plus dvStoragePower/dvStorageEnergy for all storage when the payback constraint is active. It uses sizes (not production) so it behaves correctly for both intermittent (PV, Wind) and dispatchable techs.
• build_reopt! no longer adds any size-maximizing term to the Costs objective expression — the previous -1e4 * size penalty-style incentive inside Costs was removed in favor of the lexicographic solve below. MaximizeProductionIncentive is retained as 0.0 so existing results/objective plumbing is unchanged.

Lexicographic solve (optimize_reopt!):

• New optimize_reopt!, used by run_reopt(m, p), replaces the single optimize! call:
  1. If max_simple_payback_years is set and there is anything to maximize, maximize MaximizeSizeObjective subject to the payback constraint, then pin it with MaximizeSizeObjective >= max_size - tol.
  2. Minimize the usual Costs + ObjectivePenalties objective while holding that maximum size.
• This guarantees the maximum size within the payback bound is chosen even when the lifecycle-cost-optimal size is smaller, without relying on a tuned penalty coefficient that could cause MIP numerical instability. Total solve time across both stages is reported as opt_time.
• Multinode: run_reopt(m, ps) now throws an explicit error if any node sets max_simple_payback_years, since the feature is not supported for multinode solves. (src/core/reopt_multinode.jl)

Results:

• Financial results report year_one_operating_cost_before_tax_model (the model's undiscounted year-one operating cost, used to seed bau_year_one_operating_cost) and retain maximize_tech_size_incentive_value (now always 0.0). (src/results/financial.jl)

Testing:

• Tests for the max_simple_payback_years threshold behavior covering unconstrained, loose, and strict payback scenarios, that BAUScenario clears the payback inputs, and a high-lifecycle-cost case (1-year analysis, expensive PV) where the LCC-optimal size is 0 but the lexicographic solve still sizes PV up to a ~9-year payback. (test/runtests.jl)

[Copilot]

Copilot review overview

🟡 Changes recommended

Critical BAU workflow and sizing-maximization issues remain, and multinode support does not enforce the requested constraint.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 3 · 1

Open (4)

• BAU solve retains and applies the payback constraint · New
• BAU workflow incorrectly applies new-build payback constraint · New
• Fixed size incentive does not guarantee maximum sizing · New
• Multinode solves silently ignore payback constraints · New

What changed in this PR

Adds maximum simple-payback constraints, technology sizing incentives, financial reporting, and related tests.

Changes:

• Adds payback configuration and eligible-technology tracking.
• Enforces payback limits and sizing incentives.
• Updates multinode handling, reporting, and tests.

File	Summary
test/​runtests.jl	Adds payback sizing tests.
src/​results/​financial.jl	Reports payback-related financial values.
src/​core/​types.jl	Adds maximization metadata.
src/​core/​techs.jl	Identifies eligible technologies.
src/​core/​reopt.jl	Implements payback constraints and incentives; BAU handling, capex selection, and maximization logic require changes.
src/​core/​reopt_multinode.jl	Multinode models currently ignore the payback constraint and incentive.
src/​core/​financial.jl	Adds payback inputs; BAU fields must be cleared and inputs documented.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Fixed in 93ff1e9.

[Copilot]

Fixed in 93ff1e9.

[Copilot]

Fixed in 93ff1e9 with a lexicographic maximize-then-minimize solve.

[Copilot]

Fixed in 93ff1e9 by rejecting financial.max_simple_payback_years in multinode solves.