# ThermalLoad > Cold-side receiver of a physical inter-circuit thermal coupling. > Récepteur côté froid d'un couplage thermique physique inter-circuits. --- ## EN ### Physical model `ThermalLoad` models a hydronic load segment — e.g. the cooling-water side of a shared heat exchanger — that receives an **externally-determined heat rate Q [W]** from the solver's thermal-coupling layer. It follows the BOLT/Modelica boundary pattern (`BOLT.BoundaryNode.Coolant.Source → HX → Sink`): the loop's pressure and inlet temperature are fixed by **boundary components**, not by the load: ```text BrineSource(P_set, T_in) ──edge──▶ ThermalLoad ──edge──▶ BrineSink(P_back, T free) ``` The outlet temperature is **emergent**: `T_out = T_in + Q / (ṁ·cp)` (the sink temperature must be left free — do not set `t_set_c` on the `BrineSink`, or the loop becomes over-determined). ### Residual equations — `n_equations() = 2` ```text r0: ṁ − ṁ_design (imposed design flow) r1: ṁ_design·(h_out − h_in) − Q_ext (energy balance, Q_ext = state[q_idx]) ``` The energy balance uses the *design* flow (a constant): r0 already pins `ṁ = ṁ_design`, and the constant form keeps the block linear and structurally nonsingular even when the initializer starts at `ṁ = 0`. `Q_ext` is read from the per-coupling state unknown wired by `System::finalize()` via `Component::set_external_heat_index`. Unwired ⇒ `Q_ext = 0` (adiabatic pass-through). ### DoF balance (water loop) Unknowns: 1 ṁ (shared branch) + 2×(P,h) + 1 Q = 6. Equations: BrineSource 2 + ThermalLoad 2 + BrineSink 1 (T free) + coupling 1 = 6. ✓ ### Jacobian Exact and analytic (the whole block is linear): unit entry on the ṁ row, `±ṁ_design` on r1's enthalpy columns, `−1` on the coupling Q column. ### Operational states | State | r0 | r1 | |---|---|---| | `On` | `ṁ = ṁ_design` | `ṁ_design·Δh = Q` | | `Bypass` | `ṁ = ṁ_design` | `h_out = h_in` (adiabatic) | | `Off` | `ṁ = 0` | `h_out = h_in` | ### `measure_output` | Kind | Value | |---|---| | `Capacity` / `HeatTransferRate` | `abs(Q_ext)` [W] | | `MassFlowRate` | inlet ṁ [kg/s] | ### `energy_transfers` `(Q_ext, 0)` — heat added *to* the component is positive. The component is **excluded from cycle-performance aggregation** (`counts_in_cycle_performance() = false`): the absorbed Q is the primary cycle's rejected duty, not extra cooling capacity. It still participates in per-component First Law validation. ### JSON parameters (CLI) | Parameter | Unit | Default | Description | |---|---|---|---| | `mass_flow_kg_s` | kg/s | `0.5` | Imposed design mass flow (must be > 0) | ### Usage with `thermal_couplings` ```json "components": [ { "type": "BrineSource", "name": "cw_in", "fluid": "Water", "p_set_bar": 2.0, "t_set_c": 30.0 }, { "type": "ThermalLoad", "name": "cw_load", "mass_flow_kg_s": 0.9 }, { "type": "BrineSink", "name": "cw_out", "fluid": "Water", "p_back_bar": 2.0 } ], ... "thermal_couplings": [ { "hot_circuit": 0, "cold_circuit": 1, "ua": 5000.0, "efficiency": 1.0, "hot_component": "cond", "cold_component": "cw_load" } ] ``` The coupling owns one unknown Q closed against the hot component's measured duty (`Q = η·duty_hot` via `measure_output(Capacity)`); the `ThermalLoad` consumes Q in r1, so the heat genuinely crosses the circuit boundary and the First Law closes across circuits. Keep the water-loop conditions consistent with the hot component's secondary stream (same T_in, ṁ, cp). Full example: `crates/cli/examples/chiller_r410a_coupled_water_loop.json`. --- ## FR ### Modèle physique `ThermalLoad` modélise un segment de charge hydronique — par exemple le côté eau de refroidissement d'un échangeur partagé — qui reçoit une **puissance thermique Q [W] déterminée extérieurement** par la couche de couplage thermique du solveur. Il suit le pattern de frontières BOLT/Modelica (`BOLT.BoundaryNode.Coolant.Source → HX → Sink`) : la pression et la température d'entrée de la boucle sont fixées par des **composants frontières**, pas par la charge : ```text BrineSource(P_set, T_in) ──arête──▶ ThermalLoad ──arête──▶ BrineSink(P_back, T libre) ``` La température de sortie est **émergente** : `T_out = T_in + Q / (ṁ·cp)` (laisser la température du sink libre — ne pas mettre `t_set_c` sur le `BrineSink`, sinon la boucle est surdéterminée). ### Équations résiduelles — `n_equations() = 2` Débit imposé (`ṁ = ṁ_design`) + bilan d'énergie (`ṁ_design·(h_out − h_in) = Q_ext`). Le bilan utilise le débit de *conception* (constante) : r0 épingle déjà ṁ, et la forme constante garde le bloc linéaire et structurellement non singulier même si l'initialiseur part de ṁ = 0. `Q_ext` est lu depuis l'inconnu d'état du couplage, câblé par `System::finalize()` via `set_external_heat_index`. Non câblé ⇒ `Q_ext = 0` (passage adiabatique). ### Bilan DoF (boucle d'eau) Inconnues : 1 ṁ (branche partagée) + 2×(P,h) + 1 Q = 6. Équations : BrineSource 2 + ThermalLoad 2 + BrineSink 1 (T libre) + couplage 1 = 6. ✓ ### `energy_transfers` et performance `(Q_ext, 0)` — chaleur reçue positive. Le composant est **exclu de l'agrégation de performance du cycle** (`counts_in_cycle_performance() = false`) : le Q absorbé est la puissance rejetée du cycle primaire. Il participe néanmoins à la validation du 1er principe par composant. ### Paramètres JSON (CLI) `mass_flow_kg_s` (kg/s, défaut 0.5) — débit de conception imposé. P et T_in se règlent sur le `BrineSource` ; P_back sur le `BrineSink`. Exemple complet : `crates/cli/examples/chiller_r410a_coupled_water_loop.json`.