Tariffs, energy contracts, and optimal Distributed Energy Resources investment in campus microgrids

Introduction: Large institutional electricity consumers such as university campuses face a planning problem that cannot be reduced to photovoltaic sizing alone. Their annual cost may depend simultaneously on time-varying energy prices, peak-demand charges, third-party procurement contracts, battery operation, investment budgets, and rules governing surplus photovoltaic injection. Materials and methods: This paper develops a tariff-aware mixed-integer linear programming framework that jointly optimizes distributed-energy investment and hourly electricity procurement for a grid-connected university campus over a complete 8760 h horizon. The model includes spatially differentiated rooftop, parking, and ground-mounted photovoltaic (PV), independent battery energy and power sizing, technology-specific service lives and replacement costs, peak-demand charges, selectable energy contracts, surplus export, curtailment, battery cycling wear, and an optional grid-outage resilience case. A transparent synthetic benchmark representing a campus with 22 GWh/year of electricity demand and a 5.50 MW annual peak is used to evaluate the formulation. Results: PV-only planning reduces total annual cost by 4.52%, while the economically optimized PV–battery solution reduces it by 5.12% and lowers the billing peak from 5.447 MW to 5.038 MW with a 0.699 MWh/0.410 MW battery. Enforcing a 4.5 MW peak limit increases optimal storage to 2.692 MWh but raises annual cost by only 0.39% relative to the unconstrained PV–battery optimum. When third-party energy contracts are enabled, the minimum annual cost decreases to 2.470 MUSD/year, 24.43% below the existing grid-supplied case, while the optimal PV and battery capacities decrease substantially. A 45% minimum on-site PV-utilization target is shown to be infeasible under a 6 MUSD budget, and an analytical lower bound confirms that the minimum PV and inverter investment alone exceeds 6.43 MUSD. An eight-hour outage stress case further shows that explicit grid unavailability creates a planning value for dispatchable backup and storage, with the optimized system supplying the complete outage without non-served energy. Sensitivity analysis confirms that optimal Distributed Energy Resources (DER) sizing is strongly driven by demand-charge level, battery Capital Expenditure (CAPEX), and export remuneration. Conclusions: The results show that the economically efficient campus microgrid portfolio is an endogenous response to tariff, procurement, regulatory, and resilience conditions rather than a fixed renewable-sizing problem. The main contribution of the study is therefore an integrated and transparent planning framework for large institutional consumers, rather than the specific capacities obtained for the synthetic benchmark.