Cooling Tower FAQ (Deep Dive)

Short answer

Because the DEM split uses approximations, so its “sensible + latent” doesn’t exactly equal the water-side total. The ECM model is the one that enforces energy closure to the water side by construction.

How DEM is defined

We first solve the air exit state from psychrometrics (at your chosen RH_out) so that the air enthalpy rise matches the water load as closely as possible. Then we split that air-side change into:

• Sensible: Q_s,DEM = ṁ_da · c_p,moist,avg · ΔT, with c_p,moist,avg ≈ 1.006 + 1.86 · w̄.
• Latent: Q_ℓ,DEM = ṁ_da · Δw · h_fg(T_film).

Those two pieces are not a mathematically exact partition of ṁ_da(h_out − h_in); they’re a convenient engineering split.

Cross terms that don’t cancel

The exact enthalpy change is: When we approximate with c_p,moist,avg ΔT for sensible and Δw·h_fg(T_film) for latent, you introduce two gaps:

• 1.86 (w_outT_out − w_inT_in) ≠ 1.86 w̄ ΔT (the cross term),
• h_fg(T_film) ≠ 2501 and also doesn’t carry the 1.86T “sensible part” attached to vapor exactly the same way.

Those small differences accumulate into the residual you see.

Temperature dependence & averaging

Using a single h_fg at a film temperature and a single averaged humidity for c_p is an approximation. In reality, c_p and h_fg vary with T and w along the fill. So the split is model-dependent, not identity-true.

Exit-RH feasibility / capping

If the chosen RH_out cannot exactly meet the water-side enthalpy target, we cap the exit state and show a residual (DEM total vs water total). Even with a feasible solution, iteration tolerances add a tiny mismatch.

Bottom line

ECM ensures Q_s + Q_ℓ = Q_water by definition (closure enforced). DEM shows a physically intuitive split on the air side but, due to the above approximations, Q_s,DEM + Q_ℓ,DEM can differ slightly from the water total. That’s why we display the “Residual vs water” pill—so you can see the non-closure magnitude.

Short answer

Sensible here is defined from the air-side temperature change across the tower: Q_s ∝ (T_out − T_in), not by comparing T_db,in to T_w,in. If the air leaves warmer than it entered, the sensible term is positive.

Why that happens

• Direction is set by the interface, not bulk inlets. Heat moves based on the local film temperature along the packing. The water film is hottest near the hot-water inlet and cools as it falls; the air warms and humidifies as it rises. Comparing bulk T_db,in to T_w,in doesn’t determine the net sensible sign.
• Evaporation dominates. Vapor pickup needs energy, mostly drawn from the water. The air’s total enthalpy must rise to match the water-side load. With near-saturated exit, the solution usually yields T_out ≥ T_in, keeping sensible ≥ 0.

Model notes (ECM vs DEM)

• ECM sensible: Q_s,ECM = ṁ_da·c_p,dry·(T_out − T_in).
• ECM latent: Q_ℓ,ECM = Q_water − Q_s,ECM (forced energy closure).
• DEM sensible: Q_s,DEM = ṁ_da·c_p,moist(avg)·(T_out − T_in); DEM latent: Q_ℓ,DEM = ṁ_da·Δw·h_fg(T).

When it flips

Sensible turns negative only if T_out < T_in (air cools). That requires unusual conditions: very strong evaporative effect vs small duty, very dry/hot inlet air with different exit constraints, etc.

Bottom line

Raising T_db,in (same RH) increases inlet air enthalpy; to still absorb the water-side load, the model typically solves with T_out ≥ T_in, leaving a nonzero positive sensible term.

Short answer

ECM latent is defined by energy closure: Q_ℓ,ECM = Q_water − Q_s,ECM. It does not depend on which h_fg correlation you pick.

Why that happens

ECM anchors to the water-side energy balance. Once total heat rejection is fixed by the water side, splitting into sensible (dry-air c_p×ΔT) and latent (remainder) does not need h_fg.

Model notes

Changing h_fg only affects calculations that translate energy into evaporated mass (DEM), not ECM’s energy-conserving split.

When it differs

ECM latent vs DEM latent will generally differ when DEM’s exit RH/property assumptions keep DEM from hitting the water-side total; the difference shows up as the DEM residual.

Bottom line

ECM latent won’t change when you toggle h_fg. That toggle is a DEM diagnostic, not an ECM driver.

Short answer

It affects the DEM latent (since Q_ℓ,DEM = ṁ_da·Δw·h_fg) and the inferred evaporation rate. It does not affect ECM’s split.

Why that happens

DEM derives latent directly from humidity change and h_fg(T). Different correlations shift the latent value and thus the DEM total.

Model notes

Choices: temperature-dependent ASHRAE-like h_fg(T) vs fixed (~2450 kJ/kg). The T-dependent one better tracks physical behavior over range.

When it matters

At higher film temperatures or wide ΔT, the h_fg(T) choice can move DEM latent by several percent, and the inferred evaporation accordingly.

Bottom line

Prefer temperature-dependent h_fg(T) for DEM diagnostics; ECM remains unchanged.

Short answer

DEM uses psychrometric approximations and a fixed exit RH. Those constraints can prevent an exact hit on the required air-side enthalpy rise, so DEM total may differ from the water-side total.

Why that happens

• DEM sensible uses c_p,moist(avg) over the pass; DEM latent uses Δw·h_fg(T).
• A fixed exit RH may cap the feasible state. Then the derived h_out won’t equal h_in + Q_water/ṁ_da.

Model notes

The tool reports Residual = Q_water − Q_air,DEM so you can see the mismatch explicitly.

When it flips

If you relax the exit RH (e.g., allow saturation) or iterate DEM to match enthalpy, residual → 0—but then DEM becomes ECM-like.

Bottom line

Use ECM for reporting; use DEM for diagnostics. Expect a residual when constraints bind.

Short answer

Residual = Q_water − Q_air,DEM. It measures how far the DEM air-side result is from the water-side duty.

Why that happens

Property approximations and fixed exit RH can make the DEM solution infeasible at the exact enthalpy target; the residual flags that gap.

Model notes

Residual ≈ 0 → DEM aligns with water side. Large residual → assumptions/constraints limit feasibility; recheck exit RH, flows, or rely on ECM.

Bottom line

Treat residual as a diagnostic indicator, not an error. For energy accounting, use ECM.

Short answer

When the solved state gives T_out < T_in — the air cools across the tower.

Why that happens

Needs unusual combinations: very strong evaporative effect relative to total duty, very dry/hot inlet air with different outlet limits, small water load, or constraints allowing the air to cool.

Model notes

With near-saturated outlet (typical), evaporation drives T_out ≥ T_in, keeping sensible ≥ 0 in both ECM and DEM.

Bottom line

Negative sensible is possible but rare; if seen, review inputs/constraints.

Short answer

Approach tells how close the cold water gets to the entering wet-bulb limit. Lower approach ⇒ stronger transfer (more KaV/packing/fan) but generally higher cost/energy.

Why that happens

Wet-bulb is the thermodynamic limit for evaporative cooling. Reducing approach demands more driving force/contact area, implying bigger towers or more fan power.

Model notes

The tool estimates Approach using provided T_wb,in (or infers it from inlet enthalpy if not given).

Bottom line

Track Approach with range, airflow, and water flow to characterize performance.

Short answer

Simplified indicators inspired by the Merkel method, relating duty to an average enthalpy driving force; good for comparing operating points, not for certification.

Why that happens

The full Merkel integral varies through the packing. Our simplified expressions use average quantities to give a quick diagnostic trend.

Model notes

We estimate saturated enthalpy at a film temperature (h*) and compare to average air enthalpy, then scale with range/flows to form KaV/L and KaV/G.

Bottom line

Use KaV/L & KaV/G to compare runs; for rigorous selection, use vendor tools with full Merkel integration.

Short answer

ECM pie shows the balanced split that always sums to the water-side total. DEM pie shows the psychrometric split at the assumed exit RH with chosen properties; it may not sum to the water-side total.

Why that happens

DEM uses c_p,moist, Δw, and h_fg(T); any mismatch appears as the residual pill.

Model notes

ECM: Q_s=ṁ_da·c_p,dry·ΔT, Q_ℓ = Q_water − Q_s. DEM: Q_s=ṁ_da·c_p,moist·ΔT, Q_ℓ=ṁ_da·Δw·h_fg.

Bottom line

Expect differences; use ECM for reporting, DEM for psychrometric insight.

Short answer

ECM. It matches the water-side energy balance and is appropriate for plant reporting and budgeting.

Why that happens

Water-side measurements are usually the most reliable basis for tower duty; enforcing closure avoids psychrometric approximation errors.

Model notes

DEM is excellent for understanding evaporation rate, property sensitivity, and RH effects, but may show a residual by design.

Bottom line

Report with ECM; analyze behavior with DEM.

Short answer

Because a fixed exit RH plus a high enthalpy target can be infeasible at that temperature. The solver caps at the RH boundary (or moves to saturation) and reports a residual.

Why that happens

A physical tower follows the Merkel path with varying driving force; a single-point exit constraint can’t always hit the same end state without adjusting other assumptions.

Model notes

If you allow exit to saturate or iterate DEM to match enthalpy, the “beyond saturation” impression disappears—at the cost of changing assumptions.

Bottom line

Treat this as a flag to revisit constraints; ECM remains valid for energy accounting regardless.

Short answer

1 TR = 3.517 kW. We convert each kW (total, sensible, latent) to TR by dividing by 3.517.

Why that happens

TR is a conventional capacity unit (historically tied to ice melting) used for cooling comparisons.

Model notes

The TR shown is a simple conversion of instantaneous kW; it doesn’t imply chiller rating conditions or compressor power.

Bottom line

Use TR for communication/comparisons; use kW for energy and calculations.