Heatlytics – Condensing Boiler Gas & Heat Calculator (Ideal Logic Heat H series)

What it is

An interactive calculator and set of ready‑to‑use Home Assistant Jinja templates that estimate fuel input (gas power) and water‑side heat output for Ideal Logic Heat H boilers (H12/H15/H18/H24/H30). It models how efficiency rises sharply in condensing mode (low return temperatures) and flattens in non‑condensing mode, and it respects the manufacturer’s min/max input and output data.

What it’s for

1. Visualising how modulation, flow/return temperatures and gas basis (Gross / Net CV) change fuel use and delivered heat.
2. Generating validated Home Assistant sensors for live monitoring of gas consumption (W) and heat output (W).

---

Key ideas (one page)

• Anchors from the datasheet. The app and templates are calibrated to Table 2 – Performance Data (Central Heating) for the Logic Heat H range. For example on the H24:

  • Net input (kW): MIN 4.9, MAX 24.3; Gross input (kW): MIN 5.4, MAX 27.0.
  • Water‑side output at 40 °C mean water temperature (MWT): MIN 5.1, MAX 25.6.
  • Water‑side output at 70 °C MWT: MIN 4.8, MAX 24.2.
  • Manual CVs and conversion guidance: 38.7 MJ/m³ (GCV), 34.9 MJ/m³ (NCV); m³/h = kW × 3.6 / CV.

• Condensing behaviour from research. Efficiency is modelled with the Baldi et al. return‑temperature & load polynomial. It reproduces the steep efficiency rise below the dew point (~56 °C) and the gentler decline above it. We map that shape so it passes through the manufacturer’s anchors.

• Modulation matters. We use the boiler’s relative modulation as the gas‑side load $L∈[0,1]$. This affects both efficiency and where we are between the datasheet’s min/max lines.

• Gross vs Net CV. The app lets you choose either basis. The Jinja sensors below are given on a Net CV basis (common for energy analysis). If you prefer Gross, swap to the Gross anchors and basis in the same framework.

• CH / DHW coefficient In the calculator/templates this is a simple scalar you apply to the modelled fuel–to–heat relationship to reflect real-world losses (or gains) that the pure efficiency curve doesn’t capture. You keep two because space-heating (CH) and hot-water (DHW) typically run at different set-points and hydraulics: DHW usually drives higher flow/return temperatures (often near or above the flue-gas dew point ~56 °C), so the boiler spends more time out of condensing mode and suffers extra start/stop, purge and cylinder/pipe losses; CH often runs lower returns (more condensing) but still sees distribution and cycling penalties. Practically, the coefficients let you nudge the estimate so it matches metered gas: e.g., CH ≈ 1.00 around low-temp weather-compensation, while DHW may be > 1 to account for hotter, non-condensing operation and coil/pipe losses. They’re grounded in the manufacturer’s 40 °C vs 70 °C MWT anchors (showing higher efficiency at low MWT) and the research curve that is bimodal in return temperature (sharp drop above the dew point); the coefficients simply layer on site-specific inefficiencies so your live sensor aligns with reality.

---

What the calculator shows

• Inputs: model (H12–H30), flow & return temperatures (with an optional ΔT linkage), modulation (%), and CV with Gross/Net toggle.
• Outputs: mean water temperature, whether you’re in condensing (<~56 °C return) or non‑condensing mode, min/max heat output lines, dynamic heat output (kW), efficiency (% on chosen basis), gas factor (input ÷ output), estimated gas input (kW) and m³/h using your selected CV. (See the m³/h formula in the manual.)

---

How the numbers are calculated

1) Heat‑output envelope from the datasheet

For each model we form two linear output lines vs MWT using the two anchor points (MWT ≈ 40 °C and 70 °C). Example for H24:

• min_out_kW(MWT) = -0.01·MWT + 5.5 (5.1 kW @ 40 °C → 4.8 kW @ 70 °C)
• max_out_kW(MWT) = -0.0467·MWT + 27.467 (25.6 kW @ 40 °C → 24.2 kW @ 70 °C) These numbers come directly from Table 2.

With modulation $r$ (0→1), the instantaneous heat output is:

dyn_out_kW = min_out_kW + r · (max_out_kW − min_out_kW)

2) Efficiency surface with a condensing knee

We take Baldi’s form η_poly(Tr, L) (return temperature & load), which embeds a knee near the dew point. We evaluate it at 30 °C and 60 °C return and then apply a simple affine map so the curve passes exactly through the datasheet’s Net‑CV efficiencies at those two points, blended by load:

• At ≈ 60 °C (≈ 80/60 system): η = output / input ≈ 4.8/4.9 at MIN, 24.2/24.3 at MAX.
• At ≈ 30 °C (≈ 50/30 system): η ≈ 5.1/4.9 at MIN, 25.6/24.3 at MAX.

This retains the physics‑driven bend at the dew point, but ties the scale to the boiler you own.

Why Net‑CV efficiency can exceed 100 % On a Net CV basis, latent heat is not counted in the fuel’s heating value; condensing recovers part of that heat. Hence the 104–106 % region at low returns is expected—and seen in the manual’s 40 °C data.

3) Fuel input and gas volume

• Net basis: gas_in_kW = dyn_out_kW / η_ncv.
• Gross basis: we compute η_gcv using the Gross anchors and then gas_in_kW_gross = dyn_out_kW / η_gcv.
• m³/h: gas_m3h = gas_in_kW × 3.6 / CV (MJ/m³) using your selected Gross/Net CV.

---

Home Assistant – example sensors

These are the reference templates you can paste into HA. They are aligned with the app’s maths and calibrated to the H24 model. Replace entity IDs to match your installation.

1) Gas input (Net CV, W) – condensing‑aware, modulation‑aware

{{- "" -}}
{# ===== Dynamic boiler gas consumption (Net CV, W) with condensing efficiency & explicit load =====
   Model: Ideal Logic Heat H24
   Output: Estimated gas input power on a Net CV basis (watts)

   References:
   - Efficiency surface η(Tr, L) with coefficients a1..a8 and load L on gas side (Baldi et al., Eq. 42–43). 
   - Two calibration anchors from H24 Table 2 (Net CV): 
       ~60 °C return (~80/60):  4.8/4.9 (min), 24.2/24.3 (max)
       ~30 °C return (~50/30):  5.1/4.9 (min), 25.6/24.3 (max)
   - Dew-point transition around 56 °C return; condensing efficiency improves rapidly below this point.
   Method:
   1) Compute dynamic HEAT OUTPUT from modulation between min/max output lines (vs mean water temp).
   2) Set gas-side load L directly from relative modulation (0→1).
   3) Evaluate Baldi efficiency shape η_poly(Tr, L); map it affinely so it passes through the H24 Net‑CV anchors at ~60 °C and ~30 °C
      for the current L.
   4) Net gas input (W) = HEAT_OUTPUT_W / η_ncv.
#}

{# --- Live inputs --- #}
{% set modulation = states('sensor.opentherm_gateway_otgw2_otgw2_relative_modulation_level') | float(0) %}
{% set t_flow    = states('sensor.opentherm_gateway_otgw2_otgw2_boiler_flow_water_temperature') | float(0) %}
{% set t_return  = states('sensor.opentherm_gateway_otgw2_otgw2_return_water_temperature') | float(0) %}
{% set mean_temp = (t_flow + t_return) / 2 %}

{# --- Gas-side load (use OpenTherm relative modulation as load L \in [0,1]) --- #}
{% set L = (modulation / 100) %}
{% set L = [0, [L, 1] | min] | max %}

{# --- Min/Max HEAT OUTPUT lines vs mean water temperature (kW), from H24 Table 2 --- #}
{#     5.1 kW @ 40 °C → 4.8 kW @ 70 °C (MIN) ; 25.6 → 24.2 (MAX) #}
{% set min_out_kW = (-0.01   * mean_temp + 5.5) %}
{% set max_out_kW = (-0.0467 * mean_temp + 27.467) %}
{% set min_out_W = min_out_kW * 1000 %}
{% set max_out_W = max_out_kW * 1000 %}

{# --- Dynamic HEAT OUTPUT from modulation (W) --- #}
{% set FINAL_OUTPUT_W = (max_out_W - min_out_W) * L + min_out_W %}

{# --- Baldi efficiency surface (fraction) η_poly(Tr, L); Tr = return temperature --- #}
{#     a1..a8 are from the identified polynomial; dew-point kink ~56 °C is inherent in the dataset they fit. #}
{% set a1 = 8.841   %}
{% set a2 = -0.2564 %}
{% set a3 = 0.003945 %}
{% set a4 = -0.000017983 %}
{% set a5 = 12.024  %}
{% set a6 = 0.2710  %}
{% set a7 = 0.001230 %}
{% set a8 = 7.147   %}

{% set Tr = t_return %}
{% set Tr2 = Tr * Tr %}
{% set Tr3 = Tr2 * Tr %}
{% set Tr4 = Tr3 * Tr %}
{% set eta_poly = (a1*Tr + a2*Tr2 + a3*Tr3 + a4*Tr4 + a5*L + a6*Tr*L + a7*Tr2*L + a8) / 100 %}

{# --- Evaluate the same shape at two calibration returns: ~60 °C (non‑condensing) and ~30 °C (deep condensing) --- #}
{% set t60 = 60 %}{% set t30 = 30 %}
{% set t60_2 = t60 * t60 %}{% set t60_3 = t60_2 * t60 %}{% set t60_4 = t60_3 * t60 %}
{% set t30_2 = t30 * t30 %}{% set t30_3 = t30_2 * t30 %}{% set t30_4 = t30_3 * t30 %}
{% set eta60_poly = (a1*t60 + a2*t60_2 + a3*t60_3 + a4*t60_4 + a5*L + a6*t60*L + a7*t60_2*L + a8) / 100 %}
{% set eta30_poly = (a1*t30 + a2*t30_2 + a3*t30_3 + a4*t30_4 + a5*L + a6*t30*L + a7*t30_2*L + a8) / 100 %}

{# --- Manufacturer NET‑CV anchor efficiencies η = output / net input, blended by gas‑load L --- #}
{% set eta60_min = 4.8 / 4.9 %}
{% set eta60_max = 24.2 / 24.3 %}
{% set eta30_min = 5.1 / 4.9 %}
{% set eta30_max = 25.6 / 24.3 %}
{% set eta60_target = eta60_min + (eta60_max - eta60_min) * L %}
{% set eta30_target = eta30_min + (eta30_max - eta30_min) * L %}

{# --- Affine calibration: map η_poly so it passes through η60_target and η30_target for current L --- #}
{% set den = (eta30_poly - eta60_poly) %}
{% if den | abs < 1e-9 %}
  {% set slope = 0 %}
{% else %}
  {% set slope = (eta30_target - eta60_target) / den %}
{% endif %}
{% set offset = eta60_target - slope * eta60_poly %}
{% set eta_ncv_raw = slope * eta_poly + offset %}

{# --- Sanity clamps for NET CV efficiency: 90% ≤ η ≤ 111% (theoretical upper bound for NCV basis) --- #}
{% set eta_ncv = [ [ eta_ncv_raw, 0.90 ] | max, 1.11 ] | min %}

{# --- Final GAS INPUT (Net CV, W) = HEAT_OUTPUT_W / η_ncv --- #}
{% if states('binary_sensor.opentherm_gateway_otgw2_otgw2_flame') != 'on' %}
  0
{% else %}
  {{ (FINAL_OUTPUT_W / eta_ncv) | round(0) }}
{% endif %}

2) Heat output (W) – symmetric to the gas‑input sensor

{{- "" -}}
{# ===== Dynamic HEAT OUTPUT (W, Net CV basis), symmetry with gas-input template =====
   Model: Ideal Logic Heat H24
   Output: Estimated heat delivered to water circuit (watts).

   Method:
   • Gas-side load L from OpenTherm relative modulation (0→1).
   • Net-CV efficiency η_ncv(Tr, L) from Baldi et al. polynomial (return-led + load),
     affinely mapped to pass the manufacturer H24 Net-CV anchors for ~60 °C and ~30 °C return
     at the current L (MIN/MAX blended by L).  [See Baldi 2017, Fig. 2 (dew-point ~56 °C),
     eq. (42) and coefficients (43).]
   • Gas input (kW, NCV) from H24 Table 2: MIN 4.9, MAX 24.3; interpolate by L.
   • Heat output (W) = gas_in_kW * 1000 * η_ncv.

   Notes:
   • This makes the HEAT OUTPUT sensor the exact “mirror” of your Net-CV gas-input sensor.
   • If flame is off, return 0.
#}

{# --- Live inputs --- #}
{% set modulation = states('sensor.opentherm_gateway_otgw2_otgw2_relative_modulation_level') | float(0) %}
{% set t_return  = states('sensor.opentherm_gateway_otgw2_otgw2_return_water_temperature') | float(0) %}

{# --- Gas-side load from modulation --- #}
{% set L = (modulation / 100) %}
{% set L = [0, [L, 1] | min] | max %}

{# --- Baldi polynomial (fraction) η_poly(Tr, L) --- #}
{% set a1 = 8.841   %}
{% set a2 = -0.2564 %}
{% set a3 =  0.003945 %}
{% set a4 = -0.000017983 %}
{% set a5 = 12.024  %}
{% set a6 =  0.2710  %}
{% set a7 =  0.001230 %}
{% set a8 =  7.147   %}

{% set Tr  = t_return %}
{% set Tr2 = Tr * Tr %}
{% set Tr3 = Tr2 * Tr %}
{% set Tr4 = Tr3 * Tr %}
{% set eta_poly = (a1*Tr + a2*Tr2 + a3*Tr3 + a4*Tr4 + a5*L + a6*Tr*L + a7*Tr2*L + a8) / 100 %}

{# Evaluate at ~60 °C and ~30 °C for mapping #}
{% set t60 = 60 %}{% set t30 = 30 %}
{% set t60_2 = t60 * t60 %}{% set t60_3 = t60_2 * t60 %}{% set t60_4 = t60_3 * t60 %}
{% set t30_2 = t30 * t30 %}{% set t30_3 = t30_2 * t30 %}{% set t30_4 = t30_3 * t30 %}
{% set eta60_poly = (a1*t60 + a2*t60_2 + a3*t60_3 + a4*t60_4 + a5*L + a6*t60*L + a7*t60_2*L + a8) / 100 %}
{% set eta30_poly = (a1*t30 + a2*t30_2 + a3*t30_3 + a4*t30_4 + a5*L + a6*t30*L + a7*t30_2*L + a8) / 100 %}

{# --- H24 Net‑CV anchor efficiencies (output / net input), blended by load L --- #}
{% set eta60_min = 4.8  / 4.9  %}  {# ~80/60, MIN #}
{% set eta60_max = 24.2 / 24.3 %}  {# ~80/60, MAX #}
{% set eta30_min = 5.1  / 4.9  %}  {# ~50/30, MIN #}
{% set eta30_max = 25.6 / 24.3 %}  {# ~50/30, MAX #}
{% set eta60_target = eta60_min + (eta60_max - eta60_min) * L %}
{% set eta30_target = eta30_min + (eta30_max - eta30_min) * L %}

{# --- Affine calibration of η_poly to pass through anchors for current L --- #}
{% set den = (eta30_poly - eta60_poly) %}
{% if den | abs < 1e-9 %}
  {% set slope = 0 %}
{% else %}
  {% set slope = (eta30_target - eta60_target) / den %}
{% endif %}
{% set offset = eta60_target - slope * eta60_poly %}
{% set eta_ncv_raw = slope * eta_poly + offset %}

{# --- Clamp within physically sensible Net‑CV bounds (≈90–111%) --- #}
{% set eta_ncv = [ [ eta_ncv_raw, 0.90 ] | max, 1.11 ] | min %}

{# --- Gas input (kW, Net‑CV) from modulation between 4.9→24.3 --- #}
{% set net_in_min = 4.9 %}
{% set net_in_max = 24.3 %}
{% set gas_in_kW = net_in_min + (net_in_max - net_in_min) * L %}

{# --- Final: HEAT OUTPUT (W) = gas_in_kW * 1000 * η_ncv --- #}
{% if states('binary_sensor.opentherm_gateway_otgw2_otgw2_flame') != 'on' %}
  0
{% else %}
  {{ (gas_in_kW * 1000 * eta_ncv) | round(0) }}
{% endif %}

Tip: If you want a Gross version of the gas sensor, use the same structure but replace the Net anchors with the Gross input anchors from Table 2 and clamp to a sensible Gross range (≈ 86–99.5 %).

---

Running the web app

# Python ≥3.10
pip install -r requirements.txt
flask --app app.py run  # or however your entrypoint is named

• Use Gross/Net toggle to change basis for efficiency anchoring and m³/h calculation.
• Use the ΔT slider to link flow and return: moving one will adjust the other to maintain ΔT (you can break the link by moving both independently).

---

Validating against the datasheet

• Set return ≈ 60 °C, MWT ≈ 70 °C (e.g. flow 80 °C / return 60 °C).

  • At low modulation, water‑side output and (Net) efficiency should be near 4.8 kW and ~98 %; at high modulation, near 24.2 kW and ~99.6 %.

• Set return ≈ 30 °C, MWT ≈ 40 °C (e.g. flow 50 °C / return 30 °C).

  • Expect outputs ~5.1 kW (min) and 25.6 kW (max) with Net‑CV efficiencies ≈ 104–106 %.

• Observe the knee near 56 °C return as per Baldi’s figure and discussion.

---

Known assumptions / limitations

• Anchors are static (from Table 2); real installations vary with gas quality, hydraulics, air‑fuel ratio, flue conditions, etc.
• Baldi model captures the condensing knee but is still a simplified regression; we calibrate scale to two points (30 °C & 60 °C) so intermediate returns are interpolated rather than lab‑tested.
• Electricity used by pumps/fans is not included in heat output; standby is ignored.
• Ensure your OpenTherm entities (flow/return/modulation) are stable; sensor noise will move the instantaneous estimates.

---

References

• Ideal Logic Heat H – Installation & Servicing Manual (2018): Table 2 performance data (min/max input and output) and CV guidance for gas m³/h conversion.
• Baldi, S. et al. (2017) Energy Conversion and Management 136:329–339 – condensing boiler efficiency dependence on return temperature with a dew‑point knee and a polynomial/load model used here for shape.

---

Why these results look “right”

• At low returns (e.g. ~30 °C), Net‑CV efficiency >100 % is expected because recovered latent heat isn’t counted in NCV. The manual’s 40 °C data show exactly that trend for the H24.
• At returns above ~56 °C, condensate stops forming and the curve flattens; the model reproduces this with the dew‑point knee seen in the literature.

---

Licence & contributions

MIT. Issues and PRs welcome. Please include your boiler model, CV basis (Gross/Net), and a short description of your system (ΔT, set‑points, weather, etc.) when reporting results.

---