node-red-contrib-predictive-heating

Predictive outdoor temperature node for heating and cooling curve control. Uses weather forecast and solar UV correction to calculate a more realistic outdoor temperature — saving energy without sacrificing comfort.

How it works

Instead of using only the current outdoor temperature for heating/cooling curve control, this node calculates a weighted predictive temperature that takes upcoming weather into account:

• Cold front coming → AT_pred drops → heating reacts earlier
• Warm front coming → AT_pred rises → cooling reacts earlier
• High UV / solar gain → AT_pred rises → passive solar heating is accounted for
• No sun → no correction → heating/cooling runs normally

Works automatically for both heating and cooling — no mode switching needed.

Inputs

| Property | Type | Description | |----------|------|-------------| | msg.at_current | number | Required - Current outdoor temperature (°C) | | msg.avg_day0 | number | (optional) Forecast avg temperature today (°C) | | msg.avg_day1 | number | (optional) Forecast avg temperature tomorrow (°C) | | msg.avg_day2 | number | (optional) Forecast avg temperature day after tomorrow (°C) | | msg.uv_day0 | number | (optional) UV index today (0–12) | | msg.uv_day1 | number | (optional) UV index tomorrow (0–12) |

Forecast values are stored in node context — only send them when they change (e.g. once per hour from your weather API). at_current can be sent as often as needed. If a forecast value is missing and no stored value exists, it falls back to at_current.

Output

{
  "at_predictive": 3.5,
  "correction": -1.5,
  "solar_offset": 0.5,
  "mode": "cooling"
}

| Field | Description | |-------|-------------| | at_predictive | Calculated predictive outdoor temperature → use as AT input for heat/cool curve | | correction | Delta between predictive and current AT (forecast influence only) | | solar_offset | Solar UV correction applied (°C) | | mode | "warming" / "cooling" / "stable" |

Algorithm

AT_pred = (AT_current × w0 + avg_day0 × w1 + avg_day1 × w2 + avg_day2 × w3) / total_weight

uv_avg = (uv_day0 + uv_day1) / 2
solar_offset = (uv_avg / 12) × uvMaxOffset   // positive: high UV = warmer

AT_final = AT_pred + solar_offset

Configuration

| Parameter | Default | Description | |-----------|---------|-------------| | Weight Current | 4 | Weight of current outdoor temperature | | Weight Day 0 | 2 | Weight of today's forecast | | Weight Day 1 | 1 | Weight of tomorrow's forecast | | Weight Day 2 | 0.5 | Weight of day after tomorrow's forecast | | Max UV Offset | 2 | Maximum solar correction in °C (at UV index 12) |

Use Cases

| Application | Description | |-------------|-------------| | Heating curve | Feed at_predictive as outdoor temp → pre-heat before cold front | | Cooling curve | Feed at_predictive as outdoor temp → pre-cool before heat wave | | HVAC control | Reduce energy by accounting for solar gains |

Tips

Tip: Instead of using the current momentary outdoor temperature as at_current, use a 24h rolling average. This eliminates short-term fluctuations and represents the actual thermal load of the building — exactly how professional DDC/BMS systems work. The weighted forecast then builds on a stable base, making the prediction much more accurate.

Weather API Integration

Example function node to map WeatherAPI response:

var w = msg.payload;
msg.at_current = global.get('gWeatherAvg24hTemp') ?? w.current.temp_c;
msg.avg_day0   = w.forecast.forecastday[0].day.avgtemp_c;
msg.avg_day1   = w.forecast.forecastday[1].day.avgtemp_c;
msg.avg_day2   = w.forecast.forecastday[2].day.avgtemp_c;
msg.uv_day0    = w.forecast.forecastday[0].day.uv;
msg.uv_day1    = w.forecast.forecastday[1].day.uv;
return msg;

Node Status

• AT 5 → 3.5°C (cooling) - Cold front detected (blue)
• AT 5 → 7.2°C (warming) - Warm front detected (yellow)
• AT 5 → 5.4°C (stable) - Stable weather with solar gain (green)

Author

sr.rpo

License

MIT