Should you internet-of-things your lighting?

In brief: not yet.

Lighting networks are becoming popular for the modern lighting system - modern being defined as doing everything right given the technology available. From our perspective as a lighting equipment supplier and long-term operator, we will try to evaluate here the real value of a networked lighting system and whether there are positive returns associated with implementing one. The short answer is that when you strip out the benefits from things that do not require networking (ie upgrading to LEDs and installing vacancy sensors), network lighting systems have a payback period of over 20 years and are, in our view, economically unjustified.

What does a lighting network involve?

Currently, most architects and lighting designers we see in the marketplace specify lighting networks from the conventional lighting control players like Crestron and Lutron. To a lesser extent we see newer, networking-dedicated companies like Enlighted.

A lighting control system, at its core, will turn your lights down or off if the system believes they are not being used.

Lighting controls systems typically involve

  • Hardware at the fixture level, with a sensor either in every fixture on by zone (it varies - generally more sensors means more customization and more cost). Sensors are designed to detect motion, distinguish between people and other objects, and occasionally do other things like detect temperature levels or measure fixture energy consumption. The control for light levels sometimes exists within the fixture, either via a dimmable ballast or directly with an LED, and sometimes the control exists remotely through a wall dimmer.
  • Hardware to hub everything together. They tend to live in a server room or closet, and communicate to all the fixture hardware using any number of communication standards, but usually something called Zigbee (others options include radio-frequency mesh, WiFi, or Bluetooth - all with pros and cons). The hub also interfaces with the software, which is varying degrees of user-friendly, but ultimately allows the facility management team to determine their usage pattern.
  • Software to control and communicate. The software allows users to set up profiles that tell the system what do with the sensor data - when to turn lights down or off based on time of day, day of the week, etc. This also allows managers to adjust lighting behaviors for different parts of a facility based on occupancy.

An example from Lutron:

Simplicity.

Simplicity.

The complexity of these systems is the first strike against them. Complexity can easily lead to misuse, or more commonly underuse. The buyer of the lighting network is usually not the operator (think design team vs facilities team), so the probability that the system sits in default settings is very high, which eliminates the value of the adjustability of the network in the first place.

When the system is complex, implementing it correctly is hard, too. A buyer of a system from one of the conventional players mentioned to us said that it took more than a year to get everything up and running. Removing all of first-year savings from your project value model is not good, to say nothing of the soft costs associated with countless visits from the tech over the course of a year. Challenges persist during operation as troubleshooting requires input from the manufacturer: the networks operate on proprietary software which is otherwise unfamiliar to the client.

How much does a lighting network cost?

$0.85-$1.00 per square foot for the typical lighting controls players, and more ($2.00 per square foot in some cases) from some of the new network-focuses entrants. For frame of reference, an LED implementation (top end lighting, like what Wavelength works with) costs $0.80 per foot, and that includes installation and a multi-year service commitment.

How much does a lighting network save?

We believe networked lighting can reduce lighting energy consumption by 5-15%.

In assessing lighting network savings, we assume that basic non-network lighting control technology is already in place. That primarily includes vacancy sensors and timers.

Sensors: Infrared vacancy sensors in private offices or meeting rooms should be installed.

  • Vacancy sensors cost $15 each and cover about 100 square feet each. For an commercial space where 50% of the lighting is in common areas and 50% in rooms, the per square foot cost of vacancy sensors is 7.5 cents per square foot.
  • Sensors can cut your lighting consumption in affected areas by half - a hugely worthwhile investment, and one that does not require a network. 

Timers: You want to make sure the lights will not be on all night by accident, and the solution we see there is lighting that automatically switches off at 8pm, but can be turned back easily if someone is still working. This is even less expensive than sensors, and again can be valuable.

We think the above lighting control measures make financial sense and should be incorporated in most commercial facilities. Looking then at the functions served specifically by networking to determine the value of "smart" lighting, we focus on things like the lowering of light levels in common areas and daylight harvesting.

  • Dynamic response to occupancy in common areas - It is much easier for a lighting network to cut back light output dynamically than a simple sensor. Perhaps you want the common area lights to dim slightly when not in use during the weekdays, but to have them turn off completely when not in use on weekends or holidays. This would be difficult to accomplish without being able to program settings specific to different times of week.
  • Daylight harvesting - Harvesting daylight, which means shutting off fixtures near windows when the sun is shining can be accomplished by simple sensors, but those simple sensors tend to struggle against a few factors. Pillars or drawn shades can confuse sensors into leaving the lights on during the day, and typical daylight sensors need to be within 20 feet of a window to be effective. A lighting network could allow a user to override these issues and program dimming or cutoff of relevant lights during certain times of day/week.

In aggregate, these networking-specific benefits affect 20-30% of the square footage of a facility and can have a 20-50% savings impact on those areas. We arrive at a 5-15% total savings, which feels in our experience like an appropriate range of expectations.

Do networked lighting systems make economic sense?

Let's run some quick numbers on the return on a lighting network. We will use a typical 200,000 square foot New York City commercial space as the baseline, assume the mid-point of the network cost range (which was $0.85-1.00 per sq ft), and assume the high end of network savings (15% savings on total lighting electricity). We will use payback (rather than ROI) for it's simplicity and popularity:

Screen Shot 2017-03-09 at 3.27.37 PM.png

Other than remarking that the payback does not look particularly good, we see that LEDs are in a way the enemy of lighting networks, which makes sense - the more efficient your lighting is to begin with the less there is to save.

Again, given that a full LED upgrade is only slightly more expensive than the lighting network but delivers something like $65,000 per year in savings in the above example, you would see a payback of under 2 years. This is why, even though we'd be happy to roll networking into our service now, our LEDs make networks very difficult to justify.

Dimming doesn’t count!

There are lighting network owners out there, however, who will testify to the 60% and 70% savings as promised by the lighting network companies, and many of them already had the vacancy sensors and timers we already mentioned above. So what is Wavelength missing?

The answer is dimming. Initially, dimming sounds sensible, because the idea that you would dim down lights not in use would be a great way to save energy without disrupting the overall interior lighting scheme. Additionally, dynamic dimming is definitely the kind of thing that can only be easily accomplished with a networked system.

However that kind of dimming wouldn’t accomplish anywhere near 70% savings on your total lighting cost. The way some of these systems work, with Enlighted as an example, is by dimming an entire set of lights all the time. You are reading that correctly - the smart lighting innovation is to turn your lights down. A facility might upgrade to smart lighting, and use that system to reduce hallway or open office light output by 50%, 60%, sometimes 70%. Again, these are not dynamic actions based on usage patterns, but are simple changes to how bright your space is.

So, the savings promised by networks can be real, but to us accomplishing this by having less light does not create value. People may not really notice less light and so much work is done with screens now anyway that less overhead light might be good. This savings method is therefore reasonable in its way, but it is not a technology-enabled efficiency measure - it is simply having less and paying less in proportion. As an alternative, you could go around your facility removing half of your light bulbs, have the same effect, and save hundreds of thousands of dollars on networking (sort of kidding).

How Wavelength will incorporate lighting networks in the future

Networked lighting will be integral to Wavelength’s work in the coming years, but we feel that right now it is not a justifiable investment in the context of the LED transition. You could pay $1 per square foot on controls that might save you 15% under the right conditions or pay about the same amount and do an LED implementation that saves 60% under all conditions. Meanwhile, implementing a controls system provides a disincentive to go to LEDs because you need expensive lighting to justify the controls system. That’s backward.

Part of what concerns us about existing lighting network platform is similar to why we advise caution around integrated LED fixtures - they are not future-proof. Existing software and network systems will improve dramatically in the next five years, and hardwiring today’s system into your facility will make it difficult to adapt in the future. When lighting systems companies claim future-proofing, they do so in the context of being the future re-supply partner.

The future, to us, is in dramatic cost reduction and in lamp-based networking. As always, we advocate the use of conventional fixtures, which have a multi-decade useful life. Then we will lay into those fixtures LED lamps that have networking capability on board. There will be no hardwiring of networks into your structure, granular lighting management down to individual lamps will be possible, and this accomplishes true future-proofing. After the 5 or 7 year useful life of the lamp, replacement and upgrade to a more modern/effective networking system is as easy as changing a light bulb.

A number of companies are in the early stages of lamp-based networks, we are watching the space closely for when a product displays the right functionality at the right cost - we'll tell you when it's ready.