TL;DR: Sizing behind-the-meter energy systems with spreadsheet heuristics consistently over-specifies battery capacity. In a 30-truck Fresno depot benchmark analysis, moving from heuristic rules of thumb to Mixed-Integer Linear Programming (MILP) via X4G Architect reduced initial capital expenditure and prevented a $148,000 waste in unnecessary hardware capacity.
EPC engineers and energy consultants frequently design behind-the-meter energy systems using simplified Excel models. While a flat spreadsheet calculation can verify directional feasibility, it fails to solve the circular, multi-dimensional interdependencies of asset sizing, where battery size dictates solar capacity, which dictates grid limits, which alters tariff optimization.
To quantify the financial delta between heuristic estimation and mathematical optimization, we analyzed a benchmark 30-truck commercial depot located in Fresno, California.
Baseline Parameters: The Fresno Depot
The scenario evaluates a 30-truck fleet transitioning from diesel to electric. The facility operates with an existing non-fleet electrical building load alongside the new charging infrastructure.
Financially, the model applies the local commercial time-of-use (TOU) utility rate structure, including peak demand charges and a 11-month demand ratchet mechanism. The physical environment is mapped using site-specific typical meteorological year (TMY) weather data to capture temperature impacts on vehicle HVAC consumption, charger efficiencies, panel performance, and battery degradation.
Heuristic Screen vs. Hardware Optimization
Using standard industry heuristics, an initial screen suggests an infrastructure deployment utilizing a 500 kWh Battery Energy Storage System (BESS). At this stage, the project appears fundamentally feasible, but it lacks dynamic dispatch logic to back up the capital request.
To achieve procurement-grade precision, the scenario was promoted to X4G Architect. The solver rejected the 500 kWh guess and located the mathematical global optimum: 352 kWh. Sizing the asset to this exact requirement trimmed 148 kWh of unnecessary capacity, saving the client $148,000 in upfront capital expenditure.
Where the Optimization Alpha is Found
X4G Architect proved that a 352 kWh battery, discharged precisely during high-tariff windows, achieved the same peak shaving and demand charge reductions as the unoptimized 500 kWh estimate.
Furthermore, by identifying flexible onsite building loads as deferrable scheduling opportunities, the MILP solver automatically shifts consumption profiles to align with low-tariff periods or peak solar generation. This minimizes grid cost and increases project net present value (NPV) with zero additional infrastructure asset investment.
Architect also replaces flat depreciation assumptions with a multi-mechanism battery degradation model using Arrhenius kinetics and Rainflow cycle counting. It independently tracks state of health (SOH) across the 20-year proforma, explicitly mapping the year and cost of cell-replacement events so finance teams can structure capital reserves accurately.
Engineering Specifications for Project Finance
Upgrading from a heuristic screening to an optimized engineering design directly de-risks the capital proposal. The resulting proforma relies on a physically executable dispatch schedule rather than static spreadsheet assumptions.
For projects requiring significant capital commitments, mathematical precision separates an approved investment from a rejected bid. X4G is a self-service platform that gives engineers the speed to run these simulations directly, providing full data lineage from raw weather files to the final proforma.
Frequently Asked Questions
What is the difference between heuristic models and MILP optimization? Heuristic models rely on generalized rules of thumb to estimate energy requirements, which routinely over-specifies asset capacity. MILP optimization uses mathematical algorithms to co-optimize hardware sizing and dispatch schedules simultaneously, minimizing total lifecycle costs.
Why do heuristic models over-specify BESS capacity? Spreadsheet-based heuristics cannot calculate dynamic hourly dispatch, nor can they handle the cascading impacts of complex TOU tariffs and demand ratchets. To compensate for this lack of precision, traditional workflows build a 15% to 25% safety buffer into the hardware specs.
How does X4G Architect handle battery degradation? Instead of assuming a flat annual wear rate, X4G Architect models calendar and cycle aging independently using Arrhenius kinetics, Rainflow counting, and local temperature profiles. This delivers an explicit timeline for capacity fade and necessary replacement capital injections.
X4G opens in June. Stop modeling one option and start seeing the full picture.
Follow us on X4G for weekly technical insights.