Engineering Validation
How We Prove the Numbers
Every CogenS™ engine is cross-validated against the energy industry’s most trusted public-domain references — the same engines and standards your independent engineer and lender already accept. Below is what we test, against what, and where the results land.
Modules Complete
10
Test Scenarios
3527
Pass Rate
99.3%
The validation process, end to end
For each equipment module, we build a test matrix of realistic operating scenarios, run the same scenarios through both CogenS and the reference engine, and score every output against published tolerance bands.
Step 01
Build matrix
Vary equipment size, part-load fraction, climate, control mode across hundreds of cases.
Step 02
Run CogenS
Execute the matrix through the production CogenS engine just as a project simulation would.
Step 03
Run reference
Build matched EnergyPlus IDF or analytical solver inputs for each case; execute side by side.
Step 04
Score per case
Compare hourly profiles using ASHRAE Guideline 14 metrics (annual %, NMBE, CV(RMSE), peak).
Step 05
Publish
Report every gate, every threshold, every documented exclusion — nothing hidden.
Pass criteria
A case passes when every gate falls inside its published tolerance.
We use ASHRAE Guideline 14-2023 tolerance bands as the default gate. Where two correct engines diverge on a documented architectural difference (control law, sizing convention), we apply the federal FEMP M&V relaxed band, named explicitly in the report. We never silently widen a gate to make a case “pass.”
The Standard We Score Against
ASHRAE Guideline 14-2023
ASHRAE Guideline 14, “Measurement of Energy, Demand, and Water Savings,” is the industry-recognized standard for comparing energy model output against a reference. It is what independent engineers, lenders, and federal M&V auditors look for when they ask “how do you know your model is right?”
Note on edition: we use the current 2023 publication. The calibrated-simulation tolerance bands shown below (NMBE ±10% / CV(RMSE) ±30% on hourly data, ±5% / ±15% on monthly) are carried forward unchanged from the 2014 publication; the 2023 update refines instrumentation guidance and adds measurement-approach clarifications without altering the calibration thresholds CogenS scores against.
CogenS adopts the Guideline 14 framework module by module. For each stage, we report the three core metrics the guideline defines, with the tolerance band applied to each:
Annual percent deviation
±5% (annual), ±10% (peak)
Formula
( ΣMᵢ − ΣDᵢ ) / ΣDᵢ × 100%
How close the model's annual energy total is to the reference. Lender-grade energy estimates fall here.
NMBE (Normalized Mean Bias Error)
±10% hourly, ±5% monthly
Formula
[ Σ(Mᵢ − Dᵢ) / ( n × D̄ ) ] × 100
Whether the model is systematically over- or under-predicting across the year. Catches calibration drift.
CV(RMSE) (Coefficient of Variation, Root Mean Square Error)
≤30% hourly, ≤15% monthly
Formula
[ √( Σ(Mᵢ − Dᵢ)² / n ) / D̄ ] × 100
How well the hourly shape of energy use matches the reference. Important for time-of-use rate analysis.
Variables (per ASHRAE Guideline 14-2023 §5)
- Mᵢ
- CogenS model value at interval i
- Dᵢ
- Reference engine (EnergyPlus / PySAM) value at interval i
- n
- number of intervals (8760 hourly, 12 monthly, 1 annual)
- D̄
- mean of reference values across the period
Guideline 14 applies to every CogenS module that has a side-by-side comparison against a reference engine — today: Boiler, Chiller, Cooling Tower, CHP, BESS, and Utility Tariff and Metering (the last scored against NREL System Advisor Model PySAM Utilityrate5 rather than EnergyPlus, with the same ASHRAE Guideline 14 calibrated- simulation tolerance bands). Financial modules (NPV, IRR, MACRS) use the published NIST and IRS conventions directly because there is no uncertainty band on a tax schedule.
Read the free ASHRAE Journal overview of Guideline 14 (Haberl, ASHRAE 2023) →· purchase the full standard via ANSI
What We Compare Against
Reference engines and standards
CogenS is only as credible as the references it’s compared against. We use only public-domain engines and government-published standards — the same ones your independent engineer already trusts.
EnergyPlus
U.S. Department of Energy / NREL
What it is: Whole-building energy simulation engine used by ASHRAE 90.1, LEED Energy Modeling, and Title 24 compliance paths.
Why we use it: It is the most rigorously tested public-domain engine for HVAC equipment performance. When CogenS results match EnergyPlus to within ASHRAE Guideline 14 tolerance, an independent engineer can defend the numbers.
Applies to: Boiler, Chiller, Cooling Tower, CHP, BESS modules.
NREL System Advisor Model (SAM)
U.S. National Renewable Energy Laboratory
What it is: Free techno-economic analysis software covering distributed generation, storage, microgrids, and project finance. The industry reference for solar PV, battery storage, microgrid composition, and financial pro forma.
Why we use it: SAM is the canonical tool finance and engineering teams use for the storage / microgrid / project-finance layers above the equipment-physics layer. We document SAM Generic System, Battery, Microgrid, and Single Owner / Commercial Owner as the industry reference engines for the CHP and BESS modules' degradation, storage, microgrid, and financial stages.
Applies to: CHP module Stages 3, 9, 10, 11, 12; BESS module Stages 1, 3, 8, 10, 11; Utility Tariff and Metering module Stages 1-11 (PySAM Utilityrate5 as the bill-engine reference across every tariff structure and metering policy) and Stage 12 (REAL PySAM Singleowner 7.1.1 as the Layer B 20-year financial-layer reference engine, with mixed-asset MACRS depr_custom_schedule and 30% PV ITC).
NIST Handbook 135
U.S. National Institute of Standards and Technology
What it is: The federal life-cycle costing manual that defines net present value, payback, and discount-factor conventions used by U.S. Federal Energy Management Program (FEMP) project evaluation.
Why we use it: Lifecycle financial KPIs must follow a published, defensible convention. NIST HB 135 is the standard our LifeCycleCost engine targets for direct comparison.
Applies to: All financial validation stages (boiler/chiller/CT/CHP Stage 6; BESS Stage 11; Utility Tariff and Metering Stage 6 multi-year escalation and Stage 12 20-year NPV).
ASHRAE Guideline 14-2023
American Society of Heating, Refrigerating and Air-Conditioning Engineers
What it is: Industry-standard 'Measurement of Energy, Demand, and Water Savings.' Defines acceptable error bands for energy models compared to measured or reference data.
Why we use it: We adopt its three core metrics (annual percent deviation, NMBE, CV(RMSE)) as the pass/fail gate language so every result reports against a recognized tolerance.
Applies to: Every side-by-side reference-engine comparison stage on the platform: Boiler, Chiller, Cooling Tower, CHP, BESS (vs EnergyPlus); Utility Tariff and Metering Stages 1-11 (vs PySAM Utilityrate5, scored against the same calibrated-simulation tolerance bands).
FEMP M&V Guidelines 4.0 (PDF)
U.S. Department of Energy
What it is: Federal M&V guidance (Guidelines 4.0) that provides a 'relaxed' tier of tolerance bands for cases where two physically valid models legitimately diverge.
Why we use it: Used as a Tier 2 fallback when two correct engines model the same equipment under different simplifying assumptions (e.g. control law differences).
Applies to: Regime-specific scoring where strict ASHRAE 14 over-penalizes documented architecture differences.
IRS Publication 946
U.S. Internal Revenue Service
What it is: How to depreciate property. Defines the MACRS GDS schedules used for U.S. energy-asset tax treatment.
Why we use it: Project after-tax NPV and payback depend on getting MACRS right. We validate every depreciable life class CogenS uses against the IRS schedules directly.
Applies to: All Stage 6 financial validation modules (MACRS depreciation deviation); BESS Stage 11 (5-year MACRS GDS for energy storage assets per IRC Section 168); Utility Tariff and Metering Stage 12 (5-year MACRS GDS for PV plus BESS, 7-year for CHP per IRC Section 168 and IRS Pub 946 half-year convention).
Per-Module Detail
Module reports
Thermal Systems
Boiler
Eight validation stages across 280 in-scope scenarios; every stage cross-validated against EnergyPlus 26.1 (Boiler:HotWater, WaterHeater:Stratified, Schedule:File, LifeCycleCost). 278 / 280 PASS strict ASHRAE Guideline 14-2023 on annual fuel energy, hourly NMBE / CV(RMSE), peak fuel rate, and the financial-layer KPIs (0.0000%). The 2 documented FAILs are sub-minimum-cycling sinusoidal scenarios at PLR <= 0.10 on the 100 kW non-condensing boiler.
Scenarios
280
Pass
99%
Reference-engine validated
8 of 8 stages
Reference engines and standards (9)
Thermal Systems
Chiller
Ten validation stages totalling 289 in-scope scenarios. 287 / 289 PASS strict ASHRAE Guideline 14-2023 cross-validated head-to-head against EnergyPlus 26.1 Chiller:Electric:EIR, Chiller:Absorption:Indirect (single + double effect), PlantEquipmentOperation:CoolingLoad SequentialUniformPLR, and LifeCycleCost. Five chiller technology classes (Air-Cooled Screw, Air-Cooled Scroll, Water-Cooled Centrifugal, Indirect-Fired Absorption Hot Water / Steam, Direct-Fired Absorption Natural Gas) run for the full 8,760-hour DOE Reference Building Large Hospital cooling demand at Chicago O'Hare TMY3 weather. Stage 5 chiller + chilled-TES coupled dispatch passes 12 / 12 under both setpoint-buffer and dynamic deadband coupling modes, scored on an analytical energy-balance harness — the EnergyPlus side-by-side for the storage tank itself is published separately in the Thermal Energy Storage report. Stage 6 financial layer at 0.0000 percent on operating PV, NPV, IRR, payback, MACRS vs EnergyPlus LifeCycleCost + analytical references.
Scenarios
289
Pass
99%
Reference-engine validated
8 of 10 stages (2 analytical)
Reference engines and standards (9)
Thermal Systems
Cooling Tower
Six validation stages across 321 in-scope scenarios, every stage cross-validated against EnergyPlus 26.1. Stages 1-5 cover the evaporative wet tower (CoolingTowerPerformance:CoolTools / :YorkCalc, multi-cell staging, free cooling, chiller-coupled plant, and the financial layer). Stage 6 adds the air-cooled DRY fluid cooler as a separate equipment class, scored against FluidCooler:SingleSpeed and :TwoSpeed over a 90-case parameter sweep plus two 8760-hour annual canonicals — 92 / 92 PASS at strict ASHRAE Guideline 14-2023 bands, with makeup water exactly zero on every case.
Scenarios
321
Pass
100%
Reference-engine validated
6 of 6 stages
Reference engines and standards (5)
Thermal Systems
Thermal Energy Storage (Hot + Chilled)
Five validation stages across 84 in-scope scenarios, every one cross-validated head-to-head against EnergyPlus 26.1 WaterHeater:Stratified. 84 / 84 PASS. The stratified-tank engine is scored on the COMPLETE duty cycle — charge and discharge, hot-water and chilled-water — at a 15-minute timestep, on three independent metrics per scenario: bulk tank-mean temperature, thermocline stratification, and the delivered supply temperature the plant actually serves the building with. Chilled-water storage is cross-validated against EnergyPlus here for the first time. The same 72-scenario charge-and-discharge matrix run against the previous engine passes 19 / 72, and both sets of numbers are published rather than the older one quietly superseded.
Scenarios
84
Pass
100%
Reference-engine validated
5 of 5 stages
Reference engines and standards (2)
CHP Suite
CHP
Twelve validation stages across 743 in-scope scenarios. 743 / 743 PASS at ASHRAE Guideline 14-2023 / FEMP M&V Tier 2. Six stages cross-validated against EnergyPlus Generator:* per equipment class (RICE / Combustion Turbine / MicroTurbine / Fuel Cell) and ElectricLoadCenter:Distribution; six stages validated against NREL SAM Generic System / Battery / Microgrid / Singleowner. Five CHP technology families at OEM-grounded part-load curves.
Scenarios
743
Pass
100%
Reference-engine validated
9 of 12 stages (3 analytical)
Reference engines and standards (19)
Microgrid
Utility Tariff + Metering
Thirteen validation stages across 174 in-scope scenarios. Flat rate, time-of-use, tiered block rate, demand charges, combined utility bundles, multi-year escalation, all five implemented net-metering policies, three on-site DER architectures (Solar Only, CHP plus BESS, full Solar plus CHP plus BESS microgrid) under realistic combined-bundle tariffs with live linear-programming dispatch, and a twenty-year financial layer with NPV / IRR / payback / IRS Publication 946 MACRS depreciation - every bill stage matches the NREL SAM PySAM Utilityrate5 reference engine to floating-point precision, the 20-year financial layer matches both an independent closed-form analytical reference and a Layer B cross-check against REAL PySAM Singleowner 7.1.1.
Scenarios
174
Pass
100%
Reference-engine validated
12 of 13 stages (1 analytical)
Reference engines and standards (7)
Microgrid
BESS
Twelve validation stages across 984 in-scope scenarios including a full-year 8760-hour realistic-load annual extension on Stage 1 and DOE Commercial Reference Building cross-occupancy coverage (Hospital, Large Office, Standalone Retail) on the annual extension and on every coupled-plant stage. Core BESS dispatch physics cross-validated directly against EnergyPlus ElectricLoadCenter:Storage:Simple within ASHRAE Guideline 14-2023 strict. Dynamic response, lifetime degradation, dispatch strategies, multi-unit scaling, coupled plant integration with PV and CHP, full microgrid, and the building-owner and rate-based Energy-as-a-Service financial layers validated against analytical engineering gates.
Scenarios
984
Pass
100%
Reference-engine validated
5 of 13 stages (8 analytical)
Reference engines and standards (13)
Thermal Systems
Domestic Hot Water
Every domestic-hot-water equipment type validated against EnergyPlus 26.1 in one body of evidence — direct-fired storage and tankless on electric, gas and oil under both cycling and modulating control; hydronic indirect tanks heated through a boiler-fed coil; and heat-pump water heaters. 27 of 37 in-scope scenarios inside ASHRAE Guideline 14-2023 band on annual and monthly energy. The hydronic indirect water heater is cross-validated against EnergyPlus here for the first time, on a source-side plant-loop reference built for this study. The direct-fired storage model previously zeroed both of its standby-loss terms and understated annual fuel by 6 to 32 percent depending on jacket insulation; it now runs the same node-resolved stratified tank as the heat-pump type, and both the defect and its correction are published.
Scenarios
37
Pass
73%
Reference-engine validated
3 of 3 stages
Reference engines and standards (7)
Thermal Systems
DHW Heat Pump Water Heater
Two validation stages across 28 in-scope scenarios, cross-validated against EnergyPlus 26.1 (WaterHeater:HeatPump:WrappedCondenser / :PumpedCondenser + WaterHeater:Stratified). Scored on ANNUAL and MONTHLY energy — the basis a techno-economic analysis rests on. Stage 2, the annual real-building test across four climates on the DOE Large Hotel prototype, passes 15 / 16 with monthly NMBE and monthly CV(RMSE) both at 16 / 16. Stage 1, a 30-day constant-draw steady-state shakedown, passes 4 / 12. All four exception classes are named, mechanism-characterised, and published rather than excluded from the count.
Scenarios
28
Pass
68%
Reference-engine validated
2 of 2 stages
Reference engines and standards (2)
Thermal Systems
Heat-Pump Boiler
Six validation stages across 587 scenarios and 5,407 individually scored gates for the commercial air-to-water and water-to-water heat-pump boiler. 587 of 587 PASS, cross-validated head-to-head against EnergyPlus 26.1 HeatPump:PlantLoop:EIR:Heating on a closed hydronic plant loop, plus published manufacturer operating envelopes from eight vendors. Core capacity and coefficient of performance agree to 0.12 percent; defrost energy to 0.10 percent; full-year 8,760-hour runs across four climates from 218 MWh to 17.9 MWh of annual heating.
Scenarios
587
Pass
100%
Reference-engine validated
5 of 6 stages (1 analytical)
Reference engines and standards (7)
Manifest last updated: 2026-07-29
Run the validated platform
The CogenS™ engines you see validated here power every project simulation. Eight-thousand-seven-hundred-and-sixty hourly results, plus financial KPIs, in minutes.