We are not a fan of the current crop of smart home thermostats. Most are gadgets designed to sell you needlessly expensive hardware and many have flaws, such as not working without an Internet connection. Many of them allow anyone to play with your heating system, providing little in the way of access control.
When you break the smart thermostat down into its constituent parts, it is basically a simple user interface, a temperature sensor, some control software and an actuator (a simple switch or boiler control link).
For optimum control, a good quality temperature sensor is a must. They must have a fast response time, good accuracy and very low latency.
For optimum HVAC control a good quality humidity sensor is also required.
The user interface is often the weakest link, in that anyone with physical access can use it. In reality though, many possible user interfaces could be provided, all working in parallel. All would enforce permissions and you would choose the best one that works for the task at hand and your current situation. Rarely does this involve using unauthenticated voice control but that isn't to say that voice control does not have a part to play.
Most people just want the thermostat to just work in the background and just to know what temperature a room or zone should be, but many companies have made a business selling expensive gadgets and toys that are not really needed.
In our contextual smart home all the "thermostats" can work with full context and this includes things like knowing the current status, knowing if doors are open, how the zones are physically linked together, etc. The heating system may also factor in multiple heat sources and types of localised heating in each room (e.g. towel rails and baths/showers). This can include physical characteristics of the space such as passive solar heating through windows and how well each room is insulated.
The useful context can also include external factors like local weather conditions and weather forecasts.
The key driver behind our smart heating approach is the best possible user experience, that is simple, intuitive and protects our privacy. It must also be completely reliable, enable powerful and intelligent control, without adding complexity. We also want to enforce permissions and access control, so not just anyone with physical access to our home can alter the way the heating works.
In guest rooms we do want to allow guests to have some (limited) control over the heating and cooling in their rooms though.
For a great user experience, all of the user interfaces in our smart home can be used to configure and set any number of thermostats in our contextual smart home. Each thermostat is essentially just an instance of some software that uses any of the temperature sensors in our smart home. Because our home is a contextual smart home it can make better decisions and more informed decisions, using all the information and context available to it. This means is just knows what to do and when.
Our approach to heating control makes even more sense in a multi-zone environment and results in huge cost savings over deploying multiple smart thermostats in their current form.
Some homes have multiple heat sources within a room, such as under-floor heating, central heating connected radiators, electric towel rails, gas fires, etc. Very few of the current smart thermostats on the market have any awareness of these other heat source and none that we are aware of allow multiple heat sources to be used intelligently.
By intelligently, we mean using the heat source that gives the best user experience or comfort level (e.g. using the highest output heat source to reach the set temperature more quickly or using the cheapest heat source to maintain the set point).
Intelligent heating requires intelligent boiler control. Especially when the boiler may be one of several heat inputs to each zone. We plan to use the OpenTherm standard to optimise the use of a boiler and maximise efficiency. OpenTherm provides much more precise control mechanism that can continuously adjust the boiler's water temperature set point (the 'Control Setpoint'). Our can calculate the amount of heat required from the boiler based on all the available context and knowledge of each zone's requirements, combined with control of the distribution manifold. By setting the boiler's water temperature to the correct temperature set point as it leaves the boiler, OpenTherm improves the boiler's efficiency by running at lower flow temperatures for longer periods.
As well as intelligent control of the boiler, intelligent control of pumps is also desirable. Modern boilers have a pump over-run to remove excess heat from the heat exchanger and the pump should not be run against a closed circuit. You can fit a "bypass valve" but with the manifold under intelligent control this is not necessary. The system can always ensure some flow is possible whilst the pump is running.
TPI is a feature in a thermostat that ensures that the boiler fires for the shortest possible time in order to reach and maintain chosen set point. It does this by using software to learn how the room temperatures are achieved and maintained, using this data and the current temperature to predict how long the boiler needs to be fired for.
With thermostats that don't support TPI, the boiler will fire until it gets past the set point. When the temperature then drops to a specified amount below the set point the boiler is fired again. This creates temperature fluctuations and a less comfortable environment.
contextual smart home has external temperature sensors and this is part of its wider context.
A big part of this project is an on-going study and assessment of its operation and the effect these external factors have on my control algorithms. I aim to optimise and improve them in a generic way, such that they can be applied to any home or zone within it.
OpenTherm is covered on this separate page.
