To make buildings smarter, one could place a tiny processor at every location and network them together using common software.
From an engineering point of view, we can easily make buildings smarter since we already have the required processors and software. Engineers can place a tiny $3 processor at every location and network them together. These locations include light switches, light sockets, motors that move thermal covers over windows, and pumps that move water from thermal storage tanks to radiator valves.
Then there are large appliances, thermostats, temperature sensors, occupancy sensors, and fire detectors. These devices can control central air that flows to each room, move air from one room to another with a central HVAC system, and heat or cool a tank of thermal storage water via solar for use when the sun is not shining. Moreover, they can move underground 60°F water into heat pumps, control motorized thermal window covers embedded in the wall that slide out as needed, and adjust illumination at each light bulb.
So what’s the problem?
There are several reasons why the above is not happening:
How to ensure reliability
When one turns on a physical wall light switch, the communication between the switch and the ceiling bulb is operational ≥99.999% of the time. It’s a subtle point that gets little attention, yet is important. Occupants and builders don’t accept less reliability from common building infrastructure.
It’s worth noting that wireless and power-line communications are significantly less reliable with failure rates on the order of 1% to 10%. This is due to dead zone, crowded spectrum, low signal-to-noise ratio, too small antennas, and blocked signals. Power-line communication involves placing a data signal on a power wire, yet the signal must be routed into the fuse box and then out; it mixes with massive dynamic voltage drops along the power cable, and that leads to frequent errors.
Here, if engineers want to use low-cost MCUs to network a building, they need a wire that supports CAN bus, the networking system used by automobiles to interconnect sensors and actuators. It will protect data wire from damage in the event it’s accidentally connected to the power wire.
There is a type of wiring topology called “tree,” which means one cable connects to multiple devices and has off-shoot branches. One would need a data wire system that supports this, since power cables and building geometry are configured like branches in a tree. It’s different from Ethernet, which has a single wire between two devices and from daisy-chain that has multiple devices along one wire with no branches.
Light and heavy applications
One can divide consumers in a smart building into two categories: light and heavy. Light category consumes less than 20 W, whereas heavy users consume more. Light category includes LED bulbs, light switches, thermostats, temperature sensors, occupancy sensors, fire detectors, motors for window thermal covers, motors for curtains and blinds, motors for dampers in ducts/vents, and radiator valves. Heavy category encompasses 110/220 VAC power outlets, HVAC, large appliances, and fans.
For example, a 10-W LED bulb consumes 0.1 A at 110 VAC, and it’s 1/200th of a 20-A fuse. Most devices in a building fall under the light category. To save money, engineers can connect light devices with a lower power voltage and less bulky power cable. For instance, light may route 48 VDC power on 18-awg wire while heavy applications can use traditional 110/220 VAC power on 14-awg wire.
The 48 VDC power involves lower cost electronics and data wire protection. Furthermore, 48 VDC entails building codes with fewer wiring restrictions. So, if the majority of devices are powered with the less costly 48 VDC, engineers can potentially redirect saved money into incorporating networked smart devices at every location.
An automation network connects from a central location throughout the building. Source: Manhattan 2
Common software on all devices
If engineers want smart devices at low cost and high reliability, there is only one way to do this: place the same software onto all devices. It’s also the only way to get the world to agree to make it free and open, which means anyone can use and change at no cost. There is one more requirement: quality. The system will not be well received if it’s buggy and not well-documented.
There are existing networking protocols that define how devices interact, yet they do not include software that facilitates a complete smart system. So, engineering students are working on a free and open smart device operating system at UMass Amherst and other schools called BuildingBus.
Any device can send a message to any other device; can read or write any port within any other device; can receive a library that contains information about other devices; can monitor sensors from any other device in pseudo real-time; and can send a command to any other device. Since each device knows what software is running on every other device, it can easily coordinate activities while featuring fault tolerance, high reliability, and plug-and-play.
The same operating system on all devices and a reliable communication system could make buildings smarter and more energy efficient at low cost. Researchers are already working on this. However, it’s not clear which of the various initiatives will produce the best solution, and in the coming years, we will probably see several solutions emerging to make buildings smarter.
This article was originally published on EDN.
Glenn Weinreb is co-founder and chief technology officer of Manhattan 2, an R&D initiative to solve global warming and depletion of fossil fuel problems.