Every mechanical design starts with a load calculation. The number that comes out determines what equipment gets specified, how big it is, how much it costs, and how the building will feel a year from now. In our experience, most owners assume the load calculation is a rigorous, single-answer engineering exercise, like calculating the weight a beam has to carry. It is not. Two competent engineers with the same drawings can produce cooling loads that differ by 30 percent, and both can be defensible. This note is about why that happens, what the calculation actually is, and how to make sure yours is honest.
What a load calculation actually is
A load calculation is a computer model of a building at design conditions. The engineer enters the geometry, the wall and roof assemblies, the glass properties, the internal heat sources (people, lights, equipment), the ventilation air, and the weather. The software walks through every hour of a design day and computes how much heat has to be added or removed to hold the setpoint. The peak hour becomes the design load. Equipment gets sized to handle it.
That model is a simplification of a real building, and every simplification is a choice the engineer made. The choices are where the disagreement lives.
Where the assumptions hide
Weather data. The design temperature depends on which weather file the engineer picked and which percentile they used (0.4 percent, 1 percent, or 2 percent are all in ASHRAE). Peak dry bulb is set. Peak wet bulb, which drives humidity load, is often a separate design day the engineer may or may not have run. If only dry-bulb is checked, the equipment can be undersized on humid shoulder-season mornings.
Envelope assumptions. See the R-value note. Nominal cavity R gets used instead of assembly U. Windows are modeled with generic properties instead of the actual submittal. Thermal bridging is skipped.
Internal loads. How many people are in the space? At what time of day? What are the lights? What is the equipment (computers, servers, appliances)? Small differences in these inputs cascade because the peak often occurs at the hour they overlap.
Ventilation and infiltration. How much outdoor air was assumed? Was it code minimum (which is often too little for the actual use), or the real value? Was infiltration modeled at 0.00 air changes per hour (unrealistic) or a defensible rate for the envelope tightness?
Setpoints and dead bands. Was 72°F used, or 75°F? Was a nighttime setback modeled, or was the design load calculated for constant occupancy? These change the peak hour and the peak magnitude.
Safety factors. Some engineers add a 10 or 20 percent multiplier at the end. Some do not. Some pad individual inputs (people count, equipment) and skip the final multiplier. Two engineers with the same defensible inputs can end up with different sized equipment because of where they hid the cushion.
Why the numbers still disagree
Say two engineers calculate the cooling load for the same 20,000 square foot office renovation. Both use ASHRAE 0.4 percent weather data and apply a 15 percent safety factor at the end. Engineer A models the walls at assembly U-value including thermal bridging, uses the actual occupancy schedule the tenant provided, and evaluates a separate wet-bulb design day so the equipment can handle humid shoulder-season mornings. Engineer B uses nominal R-value ignoring studs, assumes peak occupancy for the whole design day, and only checks the dry-bulb design condition.
Engineer A calculates 65 tons of cooling with proper dehumidification capacity. Engineer B calculates 78 tons of sensible cooling and no dedicated latent equipment. Neither is wrong on paper. But Engineer B's building will run a 78-ton chiller at 40 percent load most of the year, cycling on and off and dehumidifying poorly on every mild humid morning because the wet-bulb load was never sized for. Engineer A's building will run its 65-ton system closer to design most days, hold humidity through shoulder seasons, and use less energy per ton delivered. In both cases, the calculation was "correct" for the inputs used.
| Assumption | Engineer A | Engineer B |
|---|---|---|
| Weather percentile | 0.4 percent | 0.4 percent |
| Wall U-value | Assembly U-value | Nominal cavity R |
| Occupancy | Actual schedule | Peak, all day |
| Wet-bulb design day | Yes | No |
| Safety factor | 15% | 15% |
| Result | 65 tons, sized for latent | 78 tons, dry-bulb only |
The two failure modes
Undersized. Peak conditions occur and the equipment cannot keep up. The space runs warm or humid on the worst days. Complaints, expensive retrofits, and a hard conversation with the owner. Uncommon on new construction because engineers naturally pad, common on renovation projects where existing conditions were misread.
Oversized. Equipment is too big for typical operating conditions. Cycles on and off instead of running steady. Dehumidifies poorly. Uses more energy per ton delivered. Costs more to buy up front. This is the far more common failure mode in North American commercial construction, and it is the primary reason so many buildings run cool and clammy.
Both failures come from the same root cause: assumptions that were never questioned.
What a good load calculation looks like
- The weather source and percentile are documented in the report
- Wall, roof, and window U-values come from the actual assemblies, not shortcuts
- Internal loads reference real occupancy, lighting, and equipment schedules
- Both peak dry-bulb (cooling) and peak wet-bulb (dehumidification) design days are checked
- Ventilation matches ASHRAE 62.1 for the actual occupancy category
- Infiltration is included at a defensible rate for the envelope
- Safety factors are called out separately, not buried inside every input
- Equipment is selected against the calculation, not against a rule of thumb like 400 square feet per ton
What to ask your engineer
- Can I see the load calculation report with all inputs listed?
- What weather percentile was used, and was a wet-bulb design day also evaluated?
- What wall U-values were used, and where did they come from?
- What safety factor was applied and where in the calculation?
- How does the resulting equipment tonnage compare to rules of thumb for this building type, and if it is significantly different, why?
- If the equipment turns out to be oversized in operation, what is the plan (variable-speed compressors, staged capacity, dedicated dehumidification)?
The one-line version
A load calculation is a story built out of assumptions. Ask what the assumptions are, ask where the safety factors live, and ask for the numbers on both a dry-bulb and a wet-bulb design day.