The point of SEU is that the figure on the board slide is the figure the engineer can defend. Nothing is a fitted curve or a magic constant — each engine rests on a published standard, runs deterministically, and is calibrated to your meters. Here is exactly how.
Thermodynamics and psychrometrics, not regression on someone else’s plant.
The same plant always returns the same result — seeded pseudo-randomness only, never an unrepeatable draw.
Reconciled against your interval meter data, so the model answers to reality.
Saturation temperatures, enthalpies and the full steam tables come from the international standard formulation for water and steam — driving every boiler, header, PRV, flash vessel, deaerator and condensate calculation.
The heat-loss method: stack, radiation and blowdown losses subtracted from fuel input. Efficiency reflects the real fired plant and its losses, not a nameplate figure.
Chiller energy uses the integrated part-load value across the load curve, and the best-practice target is bounded by the Carnot limit at the actual refrigeration lift — so the tool never proposes upgrading a −40 °C process chiller to a comfort chiller’s COP.
Cooling, heating, dehumidification and humidification loads from moist-air enthalpy, the sensible/latent split and the apparatus dew point — the physics that makes classified cleanroom HVAC so demanding.
Steady-state duties are spread across all 8,760 hours using a synthetic typical-meteorological-year and equipment schedules. The annual total is preserved exactly; what changes is when demand falls — so cost and carbon reflect demand and price actually coinciding.
Measures are ranked on net present value with time-of-use electricity prices and hourly grid carbon, with a Monte-Carlo confidence band — not on simple payback, which flatters short-lived measures.
Upload an interval electricity or gas series and SEU reconciles the model against measured consumption — a whole-facility baseline and the model-versus-measured variance, in the spirit of IPMVP Option C. The gap is shown honestly: a small variance is confidence, a large one is a finding worth chasing.
Trust comes from candour about the edges as much as confidence in the core.
A generated plant is a researched segment template. For an audit-grade result, calibrate it to your nameplate data and reconcile it against your meters.
The typical-meteorological-year is a smooth analytical signal — accurate for load shape and free-cooling hours, not for a specific weather station. A site EPW file can be substituted.
A comfort chiller’s duty follows the air handlers it serves; process and low-temperature refrigeration still use nameplate duty until explicitly wired into the schematic.
SEU produces the analytical core of an audit. Formal certification, regulatory sign-off and on-site verification remain with your qualified auditor.
Open a sample audit and trace any figure back to the engine and the standard behind it.