On-board artificial intelligence

Control that learns your home, night after night

Every installation is different: insulation, heat emitters, habits, climate. Rather than applying generic settings, synapsus continuously records its own operation and, every night, derives models suited to that particular home.

On the tabletThe models are computed on site. No connection is needed and the data stays with you.

Learned every nightThe day's history enriches the models. The longer the installation runs, the more accurate they become.

CautiousThe AI only acts once its model is judged reliable. Until then, conventional control stays in charge.

Under controlEach AI function can be enabled or disabled on screen, and its learning status is displayed.

Help assistant

You type your question in your own words, "the house is cold", and the assistant opens the right help page. It works without Internet, in all 5 languages.

Learned heating curves

For each zone, the tablet compares the temperature achieved with the one requested and learns the curve that really maintains comfort, without overheating.

Hot water to suit your habits

Draw-off times are learned: hot water is ready when it is needed, and not heated unnecessarily the rest of the time.

Self-adapting expansion valve

The response of the refrigerant circuit is learned for each operating condition, for stable superheat and better efficiency.

Compressor speed

Three models learn the right starting speed, the maximum permissible speed and the right rate of temperature rise.

Anticipating demand

Prediction of heating and cooling demand, fan speed, taking solar thermal and photovoltaic production into account.

AI control screen — Each AI function is enabled separately

Help assistant: question in everyday language — The assistant understands questions in everyday language

AI functions and model location — On-board model today, shared model tomorrow

Example

The heating curve that really fits

A heating curve is usually set at commissioning, then adjusted in use, winter after winter. synapsus observes the temperature achieved in each zone every day and suggests the curve that maintains comfort without heating more than necessary.

Displayed next to the set curve: the difference can be seen at a glance.
Applied only if you choose: the AI function is enabled zone by zone.
Always within limits: the minimum and maximum limits set by the installer are respected.

Heating curve and curve suggested by the AI — Set curve and curve suggested by the AI

Example

An expansion valve that learns its machine

Superheat is controlled by the opening of the electronic expansion valve. Each operating condition, depending on compressor speed and outdoor temperature, responds differently: synapsus learns these responses and, at every start-up, resumes the position and gains that worked the previous time.

Fast approach, fine holding: one gain to reach the setpoint, another to hold it.
Confirmed start-up: control waits until the pressure difference proves that the compressor is really running.
Refrigerant synoptic: pressures, temperatures, superheat and current phase, with the time remaining.

Refrigerant circuit synoptic diagram — The refrigerant circuit, live

AI working for the installer

Fewer adjustments after commissioning, more comfort and efficiency for the occupant.