newsletter link
mondo arc

Lighting Sustainability: Beyond Energy Use

Bonus web material

Kevan Shaw, Director for Sustainability for the PLDA, gave the following lecture on working towards a PLDA Statement on Sustainability at the Buliding Green Expo in Athens in December 2010.

This lecture has evolved over the last year and a half as a result of discussions and new information becoming available. Lighting Design is being challenged by many new regulations from the European Commission including the Energy Using products legislation that has so far effectively banned the incandescent lamp and is going on to develop further legislation controlling energy use at a lighting system level and potentially labelling light fittings as to their efficacy. We have local legislation, for example UK building regulations mandating the energy use of lighting equipment and we have an increasing number of voluntary schemes that aim to reduce the energy use of buildings such as BREEAM, LEED and others. The universal focus of these is the energy in use of lighting equipment and power density of a lighting installation. I do not believe that these are effective in the aim of producing holistically sustainable lighting.

Much of this stems from a publication by the International Energy agency in 2006, “Light’s Labour’s Lost” apart from the title being a dreadful pun on a Shakespeare play, it is based on a single dogma:
“If you increase the efficiency of a lamp you will create a consequent, demonstrable, equivalent saving in energy in use for every application”

This is a massive and wholly incorrect simplification, however this document is a framework for legislation complete with quotable “sound bites” in the text. It is this document that has driven legislation for the past 5 years.

To understand the problems with current approaches we need to look at how lighting efficiency is measured. The value used for the assessment of lamps and luminaries is the Lumen per Watt (Lm/W) typically this is considered to represent the amount of light per watt of electricity however to understand the effective meaning of these units we need to look more closely at what this unit is a measure of.

We have to remember that the Lumen represents the energy useful to the human eye in proportion to the total energy within the visible electromagnetic spectrum. This is described by the V lamda curve. If we had a perfect light source that matched the V lamda curve this would only achieve 245 Lumens per watt of electromagnetic energy not accounting for any losses in changing energy from one form, say electricity to another, light. Even if we channeled all the energy into a single spectral frequency at the peak of human light sensitivity , 550nm, the efficiency would only be 648 Lm/W however I doubt that a single wavelength green light would be acceptable. When you measure direct sunlight in this manner you find that it achieves only 93 Lm/W. if we accept that sunlight is the best light it becomes obvious that anything more efficient is bound to be of a lesser quality.

We hear a lot about the efficiencies achieved by LEDs. Currently claims are being made for LEDs achieving efficiencies up to 200 Lumens per watt in the lab (see Fig 1). We are also used to seeing “roadmap” efficiency curves reaching ever upward in a straight line. We know that this is impossible therefore the curves have to tail off. The following diagram was published by the USA Department of the Environment earlier this year. This shows a more realistic tail off of efficiency towards the absolute limit. It also shows a significant tail off of gains. Another important thing to note is the expanding gap between the performance of laboratory and commercial offerings. There is an explanation or several for this. Primarily the laboratory results are optimsed to achieve headline grabbing numbers. This is done by working with very small devices with low absolute outputs. Testing is done instantaneously with chips at 25°C, a standard temperature that has been used for quoting performance of electronic devices for many years. In an attempt to understand the difference between laboratory tests and our practical experience of LED fittings I have had a number of discussions with chip manufacturers to understand the composition of this gap.

Firstly the reason for small chips. One of the problems with LEDs is that light is created throughout the junction so much light is buried in the chip and can never escape. The ratio between the surface of the chip and volume of the chip diminishes as the chip gets bigger. This accounts for approximately 10% of a chip output for typical lighting devices around 1 or 2 watts. There is a limit to the efficiency of the actual conversion of electricity to light within the chip - this is called the Quantum efficiency. This is already very high at 95% and is unlikely to get much better. We are also doing well with phosphor efficiency converting 95% of the light from the chip from blue to yellow. This is an area where we are likely to see falls in efficiency as we demand higher color rendering needing both a greater proportion of the light converted from blue to other colours and through the necessity to use less efficient phosphors to add red and green into the total spectrum. There is one possible development in light frequency conversion that may maintain or improve on this - it is called Quantum dot filtering. This is a nanotechnology providing a coating that physically changes the wavelength of light with a theoretical 100% efficiency, however it needs to be coated onto a substrate that will inevitably absorb some light energy. Once we have something resembling white light we need to gather it and direct it. Even at chip level this requires optics, at least a mirror to gather light from behind the chip and a basic collimating lens.

As with any optics there are losses and light directed or reflected back into the chip. Typically this is a 10% loss at present. As previously mentioned there is a marked difference between the efficiency of a chip operating at 25°C and one operating at, say, 70° or 80° in a LED fitting.

This is an area of improvement however currently we are seeing between 5% and 15% loss across this temperature difference. Coming back to the issue of spectral incompatibility the “twin peaks” of blue and yellow ensure a loss of 15% of efficiency. Finally we come to the electrical efficiency. The LED is a difficult and fussy device when it comes to being powered up. It requires DC rather than AC that comes from our electricity supply. It needs a relatively low voltage, typically around 3.5V however each needs a slightly different voltage to operate at peak output so rather than provide a voltage controlled supply a current controlled supply is required. This increases the electronic complexity of a power supply and results in more devices and correspondingly less efficiency in converting the mains supply to a suitable supply for the LED. Typically the best LED converters are 85% efficient with many more being much less efficient. If we start with our perfect LED theoretically capable of achieving 245 Lm/W and apply all these losses sequentially we find we can only get to a practical plug top efficiency of around 145Lm/W, a loss of 47%. This does tally with the difference in performance between the stated performance of the best LEDs in the market currently round 120Lm/W and the best LED fittings achieving between 60 Lm/W and 70 Lm/W. When we look for conversion to warm white rather than cool white the performance is worse as indicated by the yellow line on the graph. This is predominantly an issue of the necessity to convert more of the LED blue light to yellow so greater losses in the phosphor conversion.

We currently have fluorescent lamps achieving 113Lm/W available in the market with control gear achieving 98% efficiency. It seems that we may have a very long wait for LED systems that honestly challenge this level of efficiency. Even the extended life of the LED is challenged by long life fluorescent lamps that rate in excess of 40,000 hours life to 80% of initial lumens compared with LEDs that are rated at 50,000 to only 70% of initial Lumens. I am not saying that LEDs do not have a place in the mix of useful light sources. A fluorescent lamp will never provide the tight beam control that is possible with an LED based system.

The next measure for lighting system efficiency is the installed load. This is measured in Watts per square meter of floor area. This measure is much beloved of the voluntary building assessment schemes such as LEED and BREEAM. From a lighting designer’s perspective this is very troubling. We generally do not spend our entire lives looking at our feet!

As lighting designers we all know that the apparent brightness of a space is dependent on the vertical lit surfaces as much if not more than the horizontal and that if we want to create a cave all we need is a dark unlit ceiling! Despite this knowledge the vast majority if not all lighting standards ad recommendations are based on target illuminances in the horizontal plane. Where ceilings and walls are mentioned at all they are referred to in percentages of this horizontal illuminance. When electric light was expensive and difficult to create the focus on the horizontal task area did have some logic and sense. The gradual increase in recommended levels also had logic when visual tasks revolved around paper based communications created on typewriters and often consisted of indistinct carbon copies. Are these lighting levels relevant today where the majority of work is done on self illuminating screens and that which is committed to paper is output in crisp and stylish text direct from very high resolution laser or inkjet printers and copies come from photocopiers? We are at a stage where we seriously need to reconsider the whole basis for lighting standards.

“The metrics currently used to specify, measure, and calculate lighting levels are inappropriate for this purpose. The concept of Mean Room Surface Exitance is proposed as a basis for lighting standards.”  Kit Cuttle

Kit Cuttle that lighting iconoclast from New Zealand has come up with a possible solution. What we see is not the light directly coming from a lamp or light fitting, it is that which is reflected from the surfaces around us. Kit’s approach is to consider this as the measure for lighting. Mean Room Surface Exitance is a measure for all the light that reflected and therefore is seen. It takes account of the reflectivity of surfaces as well as the quantity of light falling on them and in the work he has done to date shows a remarkable correlation with data gathered in other experiments on apparent brightness. What is the significance of this from the point of view of sustainability? It totally changes what is measured as effective light as it doesn’t over emphasise the horizontal plane. It takes account of the reflectivity of surfaces and gives a sense of how brightly lit a space is. If you imagine an identical room lit by downlights wall washers or up lights, under current principles the downlit space will be most “efficient” as most of the light if falling exclusively on the horizontal plane, however it will appear the least lit room. With wall washing little direct light reaches the horizontal plane however you would achieve a space that appeared more lit. With an up lighting solution the apparent efficiency will be very poor however the space will be very well lit, the bright ceiling filling the vertical and horizontal planes with a broad sweep of soft reflected light. Really you should read Kit’s article to get a full appreciation of the potential gains of such a fundamental change in how we measure and quantify lighting.

Returning to  the issue of power density as a measure for lighting efficiency I would like to discuss a project that we completed some many years ago. This was an entrance for a corporate headquarters (Fig 2). It was a tricky space as it had to rationalise the grids of three separate buildings that it connected to. It was also a space in use 24/7 and the aim of the project was to change the feel over the course of the day. Our solution was to create two separate schemes in effect, one based on wall washing to open the space up and another using down lights to create paths of light pools along the axes of the buildings. In terms of power density there was in effect twice as much lighting equipment than was required to light the space however the control set up limited the lighting to that which was appropriate for the particular time of day and scene. This scheme is absolutely not compliant to current expectations for power density of lighting but actually uses power appropriate to the lighting tasks in each scene.

Basically what we need is a measure that allows us to design schemes that:

Provide the right light at the right place at the right time.
Such a measure does exist, it is the Lighting Energy Numeric Indicator usually known as LENI. This is set out in BS/EN 15193. This measure determines the amount of lighting power used per square meter of floor per year and is described in kWH / M2 / Annum. As is usual with standards reading the document makes this look like a very complicated measure to calculate, however it is very logical and with little study it becomes clear very quickly that this is a real answer to future energy legislation. As we are heading towards much more detailed metering of commercial electricity supplies the information will be there to determine actual energy use attributable to lighting. From a lighting Designer’s perspective it is possible, with little work, to determine the likely lighting energy use providing the building owner provides reliable data for occupancy and space use. If measures require that the calculated LENI for a building is met or there will be significant financial penalties such as a significantly more expensive unit cost for electricity consumed in excess of that calculated then building owners will do the work to provide good occupancy data and Facilities Managers will take lighting control settings determined by the Lighting Designer more seriously.

Such legislation will hopefully put a stop to new buildings leaving lights on unnecessarily but will not impact existing buildings where the failure to properly manage lighting wastes incredible amount of energy every year. Given the amount of complaints directed at China for pollution it is remarkable to see that they have taken a distinct lead in dealing with this particular issue. In Hong Kong a voluntary code has been implemented to ensure that Office Lighting is switched off in the evenings. They have also implemented a scheme to reduce the street lighting at night after the public transport stops. Both these measures result in a very visible reduction in waste light (Figs 3 and 4).

If we are to reduce lighting energy use there is much more to be gained by switching lights off or running  them at very reduced power when they are not required. It has to be remembered that we are animals that have evolved under natural lighting of open skys. We have evolved with very high levels of light in the daytime and with darkness or low light being a threat with our ability to hunt and gather lost and the threat from predatory animals increased. Our reaction to this was the harnessing of fire to provide light and warmth at night. So if we are programmed that bright is good it is not surprising that we naturally try to increase light levels whenever we can. Equally if darkness is a threat we are not programmed to make areas dark by switching off lights. This fundamental aspect of psychology requires strongly learned responses. It is interesting to note that during and after the Second World War, in the UK particularly, light at night became strongly associated with a greater threat from bombing and a whole generation learned that switching off lights increased safety (Fig 5). This lasted for that generation and a few of the generation that followed immediately after the war. However the raw natural drive for more light has quickly overcome this learned response.

If we cannot rely on a learned response to get people to switch unnecessary lighting off then we have to come up with ways of doing this for them, however this has to be done in a way that does not make people inconvenienced or feel they are lacking in control of their environment. Typically occupancy sensors are based on Passive Infra Red (PIR) technology. This was not invented for lighting control, it was originally, and still is, used in security systems to identify movement of people into and across spaces. It is certainly not optimised to identify someone sitting at a desk concentrating on a task such as reading a document or updating their daily activity on Facebook! The result is that many if not most PIR based systems incorrectly switch lighting off. As previously mentioned we have a natural desire for more light so if light is taken away from us we are immediately concerned. The sudden switching on or off of light in our field of peripheral vision is also major distraction. Just as we are programmed to respond to light or dark our peripheral vision is programmed to detect change and movement that may be a threat so we immediately focus on that area and our attention is directed to it. If this happens frequently then we have real difficulty maintaining concentration. If lights are switched on and off according to changes in daylight availability the same distractions will occur.

“ What we have to do is make it easier for people to do the right thing and when we are doing it for them, do it so they don’t notice, otherwise they will feel things are being imposed on them or they are lacking in control.”

There is little excuse for lighting control companies not to provide better methods of detection for presence. Machine vision systems are now well tried and widespread technology in use in many areas from production lines where cameras check product quality and robots that “see” the tasks they perform to the humble office printer that can tell if paper has been put in upside down or whether it is matt or glossy! Philips demonstrated a machine vision based system to control an LED ceiling as long ago as 2005, why has this technology not been brought to market yet?

If we are frustratingly to be stuck using PIR technology we can still go someway to get the best out of it by thoughtful application. We were faced with this issue on a recent project, Tanfield House, Edinburgh an office building that was being converted from single occupancy with two large open floors to multiple occupancy (Fig 6). Naturally we did not have any idea of the nature of the future tenants business requirements so we had to start with a simple “one size fits all” solution. We had a good floor to ceiling height and worked with the architects to retain the exposed waffle slab as this gives a distinct advantage from available thermal mass to reduce heating and cooling plant sizes. We chose a direct/indirect lighting system using High Efficiency T5 fluorescent lamps and controlled these with combined daylight availability and presence detectors. The system is set up so that it will respond to absence or increased daylight availability by dimming the fittings up and down over 10 seconds. The fittings never switch off completely but ramp down to around 10 % of energy used. The result is very successful, so successful in fact that the first tenants, a building services consultancy, were not even aware that the lighting was controlled during the day until it was pointed out to them!

Another simple technique was employed to deal with toilets. Again we wanted occupancy control however we wanted to avoid forcing people to walk into a dark space before the lighting came on. This was easily accomplished by locating the sensor in the lobby outside the toilets and leaving a single lamp in each toilet for anyone who extended their stay unreasonably beyond the delay time set.

The ideas behind this are not restricted to internal office lighting. As more controllable light sources such as LEDs are used for outdoor public area lighting such as streets, they can be similarly applied. There is a scheme that was commissioned this year in Toulose, France where WEEF modular LED streetlights are controlled by PIR. They normally run at 30 Watts but if a pedestrian or cyclist is detected they switch up to 50W. This is well accepted by the populace and provides an annual saving in energy cost sfor the scheme of €4000. As street and roadway lighting is one of the major light polluters as well as uncontrolled energy users this kind of system is bound to reap many benefits.

Daylight is a much underused resource for building lighting. Over the past century development has grown outward and outward in an effort to maximise the lettable or saleable floor area. Equally floor to ceiling heights are reduced to minimise building envelope costs. The progress of lighting technology and somewhat suspect energy pricing policies have made electric lighting cheaper and brighter. The idea that tall buildings and deep floor plans represent economic development we have seen the glass tower pop up in entirely unsuitable climatic regions where land costs are not such an issue. Daylight is then compromised by the application of heavy solar filtration and blinds to reduce glare devaluing the quality and quantity of daylight that penetrates the building.

In more temperate regions daylight availability is being constrained by energy legislation aimed at reducing heat loss and gain through glazing. In the UK the glazed area for building sis restricted to 40% of the façade, if this is used by the architect for big atriums or entrance statements in curtain wall this results in mean windows to occupied areas barely capable of providing a view of the outside world let alone provide meaningful levels of useful daylight.

As I have been involved with Lighting Sustainability I have become increasingly concerned that the data and statistics used to base potential savings of lighting energy on. For example in various places in the EU documentation supporting the banning of incandescent lamps the proportion of energy attributed to domestic lighting to energy use varies between 3% and 15%. The truth is that the statistics and analysis used to identify lighting power use among total power use are based on inadequate data. All that one can illustrate is that for most buildings lighting power is a relatively small proportion of power required (Fig 8).

This graph obtained from a UK government website illustrates the relatively small proportion of power actually used when this is actually metered rather than estimated. Admittedly a certain proportion of cooling energy may be attributable to lighting but this still remains a relatively small proportion of total load. Lighting has become a specific target for energy saving as it is so visible.

I came across the following graphic on he USA Department of the Environment website (Fig 9). It shows how inefficient electricity generation and distribution is. I have not found sufficient information to reconstruct this for Europe, however the general principles remain.

Approximately two-thirds of energy used for electricity generation is wasted in generating and transmission losses, no matter what we do in lighting we are not going to touch that level of energy wastage. While we can do a little it is very obvious that, to meet the current targets to reduce CO2 emissions significant changes to the generating mix and distribution systems need to be addressed. So far in the most of the developed world, we are barely scratching the surface of the possibilities of renewable generation.

Keeping with Light as a theme we are beginning to look at solar electricity generation as a serious contributor. Photovoltaic systems are not yet economically viable without significant benefit from special feed in tariffs and really only work on quite small scale. From a lighting perspective they generally work when we don’t or shouldn’t need electric light! Large scale Solar / Thermal systems are more interesting. There is considerable development of the technology with plants in use and under construction in the USA, Spain and the Middle East totalling some 3 GigaWatts. In conjunction with this is the Desertec scheme that proposes a network of high power interconnects linking solar generation in the equatorial regions with wind and hydro schemes in the North of Europe reaching as far as Iceland’s seemingly limitless resources of geothermal energy. Such ambitions schemes will be necessary to really address the provision of renewable electricity generation.

If CO2 is the real enemy as we are currently told we also have to address the provision of new Nuclear power generation. As the main downside of this technology is the problems of disposal and clean up the next generation of nuclear plants should be designed to last much longer than the 20 or 30 years typical of the first and second generation, this way the disposal and clean up can be dealt with over a longer time frame so there is much less problematic waste per megawatt hour of power generated.

If we accept that some serious change to power generation is necessary then the issues of CO2 output will recede over the forthcoming years. This is far from the only consideration we should have for lighting sustainability however these other concerns are very much out of people’s thinking compared to the issue of energy in use. Manufacture of new lighting equipment consumes considerable natural resources, not the least of which is Aluminium.

Bauxite, the mineral from which aluminium is extracted is one of the most plentiful minerals in the world however winning it does have environmental impacts. To get at it you have to strip all the topsoil and destroy any habitat, You then basically dig a very big deep hole that will not fill itself in, neither is it likely to be in a convenient location to use for waste landfill! Much Bauxite is won in Northern Australia for example where big holes are not very useful as the majority of the population live thousands of miles away on the Southern and Eastern coast. Incidentally problems are caused by the lighting on the vast piers used to load bauxite into ships. These lights attract wildlife who mistake the light for moonlight, particularly turtle hatchlings who should be attracted to the sea during hours of darkness, instead they move along the beach towards the lights until daylight reveals them to predators or the sunlight dehydrates and kills them.

To extract Aluminium from Bauxite massive amounts of electricity are required in big electric furnaces. Iceland, whoa re lucky enough to have plentiful geothermal energy are now a major site for aluminium smelting to the extent that several times more power is used in this industry than all other energy uses put together. This is nice clean non CO2 producing energy however the carbon electrodes used for the process of smelting are sacrificial so Iceland’s major CO2 production comes from this source effectively producing as much CO2 as if Iceland wholly depended on coal fired electricity for its domestic energy needs! Iceland is not noted as an industrial centre so the aluminium is shipped in semi finished form to the USA and China to be turned into products including lighting equipment and is then shipped back around the world as finished lighting equipment.

Surprising as it may seem all his world travel does not contribute as much CO2 as the relatively short road trip from wherever the fittings are landed to site, however the question does need to be asked if sufficient use has been made of recycled materials in the lighting equipment you specify? This is a question we must ask of manufacturers. Aluminium is easily recycled and the energy required to turn recycled aluminium into light fittings is a fraction of that required to create new aluminium from Bauxite mined in Australia or Africa. Glass is also heavily recycled, again the raw material for new glass is sand that is very abundant but the energy required to turn sand into glass is much more than that required to recycle glass. Finally plastics can be recycled however many that are used in lighting equipment use brominated fire retardants making them non re-useable. If we are going to replace lighting equipment we expect the old stuff to be re-cycled, however if there is no demand for recycled material in new equipment where do you think it goes?

Of course re-use is better than recycling. This is a very big challenge as the lighting industry would far prefer that you bought new stuff with the appropriate profit margin than even support the repair of existing equipment. In the project at Tanfield House we could not re-use any of the interior equipment. The original scheme used floor mounted metal halide uplights that were life expired and we could not continue with a floor-mounted solution. On the exterior however we wanted to re-use the Bega bollards that were tired, used old technology lamps and gear but the main shell of the fitting was still in good condition or was until the contractor, unused to removing fittings carefully for re-use damaged some of them in removal with force! We did manage to recover almost all of them, replacing lamps, gear and reflectors repaired the contractor’s damage and ended up with a quality fitting at the cost of a very cheap new one. The company who did the work were not so happy and are unlikely to do this for us again. We need to get a market developed for some specialist lighting refurbishment companies to get started in. It is clear that the cost of new will not always be less than the cost of quality refurbishment however the environmental benefit is substantial. It is a real issue that no credits are given by LEED or BREAM for re-use of equipment in refurbishment schemes.

Our lighting schemes also need to consider the introduction of materials none to create environmental problems. Foremost among these is Mercury. This is essential for the majority of common “energy efficient” light sources including metal halide and fluorescent lamps. Currently mercury is being banned from almost every use except lighting. In conjunction with this the EU has banned the export of Mercury as much was finding its way to artisinal gold mining and uncontrolled release into the environment. The result of this has been a one way traffic between newly reopened Cinnabar mines in China, Chinese lamp factories, domestic lamp users in Europe and contained landfill in German salt mines for mercury contaminated lamps debris that is not economic to remove the mercury from. While much effort is being put into the development of LED and other non-mercury containing lamps fluorescent and other mercury containing lamps will be with us for much longer.

In conclusion, to maintain quality of lighting we need to look very carefully at the direction lighting legislation and consequent practice are moving. We need to become a stronger influence in these processes. While addressing current concerns on energy in use we must be mindful of the potential damage to other aspects of the environment and ensure that we are being sensible with resource use. We need to be sure when we specify new equipment that there is a proportion of recycled materials used preferably enough to account for any materials disposed of from previous installations. Finally we must avoid a blinkered view and look at lighting in isolation. The energy use is a building or scheme wide phenomenon extending before the project is started to the end of life disposal of the building and its fixtures and fittings.

Lighting Res. Technol. 2009; 00: 1–21
Towards the third stage of the lighting profession


Post a Comment
Security Code* Get another image
Kevan Shaw

  • Figure 1

  • Figure 2: Corporate headquarters foyer

  • Figure 3. Hong Kong at 8PM before the code. Pic: David Chan

  • Figure 4. Hong Kong at 8PM now. Pic: David Chan

  • Figure 5

  • Figure 6

  • Figure 7

  • Figure 8
Related Articles


Follow us on…

Follow Mondo Arc Magazine on Twitter Follow Mondo Arc Magazine on Facebook Follow Mondo Arc Magazine on Linked In

mondo arc india

darc awards DWLF IALD PLDC LRO