Entropyk/apps/web/public/docs/components/flooded-evaporator.md

FloodedEvaporator

Config type: "FloodedEvaporator"
Source: crates/components/src/heat_exchanger/flooded_evaporator.rs
Example: crates/cli/examples/chiller_flooded_4port_watercooled.json (DoF 19=19, COP ≈ 6.45)

---

EN

Purpose & physical model

Shell-and-tube flooded evaporator. Refrigerant boils on the shell side; secondary (water/brine) flows in the tubes. Heat duty uses phase-change ε-NTU (C_min = C_sec, C_r → 0):

ε = 1 − exp(−UA / C_sec)
Q = ε · C_sec · (T_sec,in − T_evap(P))

• T_evap(P) = saturation temperature of the refrigerant at edge pressure
• C_sec = secondary heat-capacity rate [W/K]

There is no plate-geometry correlation inside this component (unlike BPHX). UA is a lumped parameter (possibly scaled by calibration z_ua via the inner HeatExchanger calib API).

Dual operating modes (both enter Newton residuals)

Mode	How secondary is defined	Secondary Newton unknowns	When to use
System (4-port)	Live edges secondary_inlet / secondary_outlet (e.g. BrineSource → HX → BrineSink)	Yes (n_secondary = 1 or 2)	Closed water loop, real machine
Rating	Scalars secondary_inlet_temp_* + C_sec (secondary_mass_flow_kg_s × cp or secondary_capacity_rate_w_per_k)	No (n_secondary = 0)	Qualification / open-loop duty; still coupled ε-NTU in residuals

coupled_ready = refrigerant indices ready and (live secondary edges or rating scalars).
Never falls through to generic four-port HeatExchanger::inner residuals for normal operation (seed path is local and finite).

Rating residual energy: uses full Q (not α(ṁ)·Q) so ṁ = 0 is not a trivial root when C_sec > 0.
System residual energy: uses effective_duty(Q, α_ref, α_sec) from flow_regularization for zero-flow safety.

Also exposed: rate(p_in) open-loop rating API for sweeps (same ε-NTU formulas).

Outlet closure (DoF-critical)

Setting	Residual r2	Typical use
Default (quality_control: false)	h_out − h_g(P) saturated vapor	Compressor suction after disengagement
quality_control: true	x_out − target_quality	Legacy recirculation / two-phase outlet

Both keep the same n_equations (quality replaces sat-vapor; it does not add an extra free residual by itself).
quality_control: true on a closed cycle often needs a free actuator elsewhere or the DoF gate rejects the graph.

Residuals & n_equations()

Row	Equation
r0	P_out − P_in (no refrigerant ΔP by default)
r1	ṁ_ref · (h_out − h_in) − Q_eff
r2	sat-vapor or quality target (see above)
r_sec mass	ṁ_sec,out − ṁ_sec,in only if live edges and not same-branch
r_sec energy	live secondary energy + duty (blended at low ṁ)

n_equations = 3 + n_secondary
n_secondary = 0          # rating (no live edges)
            | 1          # live edges, same-branch ṁ
            | 2          # live edges, independent ṁ in/out

Ports

Port	Index	Role
inlet	0	Refrigerant from EXV
outlet	1	Refrigerant to compressor suction
secondary_inlet	2	Water/brine in
secondary_outlet	3	Water/brine out

CLI aliases: water_in / brine_in → secondary_inlet, etc. (resolve_port_index).

Calibration

Factor	Effect	Default
z_ua (BOLT Z_UA)	UA_eff = z_ua · UA via inner calib	1.0
z_dp	pressure-drop scale if ΔP model used	1.0

Inverse calibration (CLI controls[] / UI Fixed checkboxes): impose a measure (e.g. SST = saturationTemperature) and free z_ua.

JSON parameters

Key	Meaning	Unit	Default
ua	UA	W/K	required
refrigerant	refrigerant id	–	primary fluid
secondary_fluid	secondary fluid	–	Water / MEG
quality_control	quality residual instead of sat-vapor	bool	false
target_quality	x target if quality_control	–	0.7
secondary_inlet_temp_c / _k	rating T_sec,in	°C / K	–
secondary_mass_flow_kg_s	rating ṁ_sec	kg/s	–
secondary_cp_j_per_kgk	rating cp	J/(kg·K)	4186
secondary_capacity_rate_w_per_k	rating C_sec direct	W/K	–
calib z_ua / z_dp	Z-factors	–	1.0

energy_transfers / mass

When coupled: cooling Q ≈ ṁ·(h_out − h_in) (positive heat absorbed by refrigerant).
port_mass_flows reports 4-port signs without calling generic inner four-port.

Zero flow

flow_regularization on system path: smooth |ṁ| for live C_sec, activity factors, secondary Δh hold. Seed residuals stay finite if P is non-physical.

---

FR

But & modèle

Évaporateur noyé tubes-calandre. Duty ε-NTU à changement de phase :

ε = 1 − exp(−UA / C_sec)
Q = ε · C_sec · (T_sec,in − T_evap(P))

Pas de corrélation géométrique type Longo/Shah (voir BPHX pour ça). UA est un paramètre global, modulable par z_ua (défaut 1).

Deux modes (tous deux dans le Newton)

Mode	Secondaire	Inconnues eau	Usage
Système	Ports live Source → HX → Sink	oui	Machine fermée
Rating	Scalaires T + ṁ·cp (ou C_sec)	non	Qualification ; Q ε-NTU dans les résidus

Clôture de sortie

• Défaut : vapeur saturée h_out = h_g(P) (aspiration compresseur).
• quality_control: true : x_out − x_cible (même nombre d’équations).

Résiduels

n_equations = 3 + n_secondary (0 / 1 / 2). Voir tableau EN.

Calibration

Imposer une mesure (SST) et libérer z_ua (UI case Fixed, ou controls[]).
z_ua = 1 = pas de correction.

Ports / JSON

Identiques aux tableaux EN.