Public lighting: the calculation comes before the decision, not after

Specifications and supply

The order of the documents decides the outcome of a retrofit. If the calculation arrives after publication, every bidder proposes its own configuration and nothing is comparable any more.

The document that makes bids comparable

In a street lighting retrofit the order in which the documents appear matters more than almost anything else. If the lighting design arrives after the tender is published, every bidder proposes its own configuration, and the authority ends up comparing proposals that meet different requirements.

With the calculation done first, for each homogeneous stretch, the specification can ask for a precise performance instead of a generic description. At that point the bids are compared on price, which is the only comparison a tender does well.

Not a formality, but the unit of measure

The lighting calculation is often treated as an annex to be produced because a standard asks for it. In fact it establishes what is being bought: without it the authority chooses between documents that do not speak the same language, and whoever wrote the most reassuring report wins.

A specification saying «good quality LED fittings» has asked for nothing verifiable. One saying maintained illuminance, uniformity and glare for each stretch has asked for something measurable on site, and contestable if it is missing.

The homogeneous stretch, not the municipality

An average across a whole territory says nothing and serves nobody. The calculation is done stretch by stretch: with its geometry, its width, its real spacing between columns, its classification and its lighting class.

Two roads that look alike on paper may need different classes because of the traffic they carry or the crossings on them. Averaging everything means over-lighting where it is not needed and under-lighting where it is, and paying for both.

What has to be in hand before starting

A calculation does not come from a wish but from a survey. The four things always needed are few, and they almost always exist somewhere already, scattered across different offices.

• position and height of the light points, with the real spacing
• carriageway width and the presence of footways or cycle tracks
• supply cabinets, with the circuits each of them serves
• road classification and the lighting class required

Cutting consumption has a limit, and the limit is written down

Dropping below the required lighting class is not a saving: it is a defect. Optimisation works at constant performance, meaning it looks for the configuration that costs least while meeting the same requirement, not the one that simply costs least.

It is a distinction that looks obvious written this way and often gets lost in a tender, because energy saving is easy to show in a table while an unmet requirement only appears when somebody measures it at night.

The survey is worth as much as the design

A calculation run on old data produces a design that cannot be built. It happens more often than one would think: columns moved and never recorded, circuits shared with other services, cabinets feeding stretches other than those on the map.

A site visit is not there to take good photographs: it is there to find those differences before they become variations during the works. A variation in a public tender costs administrative time, which is the dearest currency an authority has.

Shared circuits and shared cabinets

In historic centres and outlying hamlets it often happens that the street lighting circuit also feeds something else, or that two neighbouring municipalities share a cabinet without anyone ever having written it down. These are normal situations, and they only become a problem when they surface after the works have begun.

They have to be surveyed and declared in the specification, because they change what the contractor can do and how much time is needed. An installation that can be switched off for two hours and one that can never be switched off are two different jobs with the same drawing.

The maintenance factor, the assumption nobody reads

Inside every calculation there is one number that decides the result more than almost any other: the maintenance factor. It states how much the installation is assumed to soil and decay over time, and changing it changes the result without touching a single fitting.

That is why it has to be declared, not left to the software. Two reports starting from different factors reach different conclusions about the same installation, and both are formally correct.

What to ask bidders, in a verifiable way

A well written specification asks for quantities, not adjectives. The items that make two bids comparable are these, and there are few of them.

• maintained average illuminance and uniformity, for each stretch
• disability glare, declared rather than implied
• the maintenance factor adopted, with its justification
• photometric files for the fittings proposed, in a readable format

Glare is the part the public actually feels

People walking or driving do not notice average illuminance. They notice a source going into their eyes, and they notice it at once. An installation that meets the average values and gets glare wrong produces more complaints than one running slightly below that dazzles nobody.

It is also the quantity most easily ruined during installation: a fitting tilted a few degrees more changes the result. It belongs in the specification and it belongs in the acceptance test, not only on paper.

One-for-one replacement is an expensive shortcut

The temptation in a retrofit is to swap every old fitting for a new one in the same position. It is quick and looks prudent, but it carries over every mistake of the previous installation, including columns placed wherever was convenient.

The calculation serves this purpose too: saying where a light point can be removed and where one more is needed. On many stretches the number of fittings falls and the result improves, and the saving is not only energy but future maintenance.

Dimming control is decided early, not added later

Reducing output in the middle hours of the night is nearly always worthwhile, but it is decided at design stage because it changes which fittings to choose and which circuits to prepare. Adding it afterwards means going back over what has just been installed.

It also has to be stated clearly what it must do: which time bands, which minimum level, and what happens when the system does not answer. An installation that switches off because the command did not arrive is worse than one that was never dimmed at all.

The acceptance test is written with the specification

If the verification measurements are defined at the end, what gets measured is whatever is convenient to measure. Written alongside the specification, they say in advance at which points and under which conditions the result will be checked, and that disciplines the bidders too.

It is the part that protects the authority most and costs least: a few lines, and they change how the tender is read by the people reading it.

The quantities, and when each becomes known

Lined up, the quantities that decide a retrofit do not arrive at the same moment. Some are known before the tender, others only after the survey, and one — administrative time — depends on neither.

What changes for the technical office

For a technical office, having the calculation first means being able to answer bidders without improvising, and being able to reject a proposal that misses a requirement without opening a dispute. It is a more comfortable and more defensible position.

It also means being able to explain the choice to the administration with numbers instead of opinions, which in a public body is often the difference between a decision that passes and one that comes back.

Who signs the calculation, and what signing entails

A lighting design with no name on it binds nobody. If the value is not reached on site, somebody has to be accountable, and the answer cannot be the software that ran the numbers. Responsibility for the calculation is an administrative choice as much as a technical one.

When the document comes from the bidder, that is not a flaw in itself: the supplier is the one holding the photometric data. It becomes a flaw when it serves to justify a decision already made rather than to verify it, and the sign is always the same — the calculation states the average and stays silent on everything else.

The territorial data that often already exists

Many authorities have a light point survey carried out years ago for other reasons, and do not know they can use it. Even incomplete, it is a better starting point than a survey from scratch, because it holds the most laborious part: where the columns are and which cabinet feeds them.

The practical rule is to start from that data and verify it by sampling, not to rebuild it. A full survey costs weeks; a targeted check on the doubtful stretches costs days and produces an equally reliable result.

What happens when the order is reversed

It is worth describing the opposite case, because it is the one most often met. The tender goes out with a generic description, five bids arrive with five different configurations, and the committee realises it has no criterion for comparing them.

From there every road is worse than the last: asking for clarifications stretches the timetable, awarding on lowest price puts an unverified installation on site, and cancelling means starting again. None of those three exits costs less than the calculation that was not done.

The three lines to settle before publishing

If only one thing survives from all this, let it be three points to close before the tender goes out. They are short, and they close most of the problems that otherwise surface later.

• the calculation exists, and it is per homogeneous stretch
• the specification asks for measurable quantities, not adjectives
• the acceptance test is described, with its points and conditions

• A bespoke piece
• The buildings we have lit
• Free lighting calculation
• Lighting knowledge