Street Lighting: A Metric for Specifying White Light

This project will investigate lighting in subsidiary (residential) streets. In recent decades this has been provided mainly using low pressure sodium (LPS) and high pressure sodium (HPS) lamps. These lamps have high luminous efficacy (light output) and a long working life, and these characteristics were considered sufficient to outweigh their poor colour properties / the orange appearance and poor colour rendering. Brightness is an important consideration when lighting pedestrian environments because it is linked to the perception of safety / brighter streets are considered to be safer. Street lighting engineers suggested that using whiter light sources, such as metal halide (MH) and compact fluorescent (CFL) lamps, would create a clearer and brighter environment, and furthermore that street lighting using these whiter light sources could adopt a reduced illuminance (light level) whilst maintaining the same level of brightness and clarity. This illuminance reduction would offset the efficacy of MH and CFL lamps, which tends to be lower than HPS and LPS lamps, and may even lead to a reduction in energy consumption. This anecdotal evidence has been supported through a series of experiments carried out by the applicant (EPSRC grant ref. GR/25074/01) demonstrating savings of approximately 30% in capital and energy costs. A primary reason for this interaction between lamp type and illuminance is that the human eye responds differently at the low levels of light found at night-time than it does in the higher levels typical of daytime, but illuminance specification is calibrated only to the daytime response. The BSI road lighting committee (CPL/34/8) took account of the white light effect in the 2003 revision of BS5489-1. This document interprets the S-series of lighting classes for subsidiary streets specified in BS EN 13201-2:2003 according to the rates of crime and traffic flow. The metric by which BS5489-1:2003 identifies whiter lighting is the General Colour Rendering Index (CRI) and this is because it provides an obvious distinction between the white lamps and the standard HPS and LPS lamps. Thus, BS5489-1:2003 permits a reduction of one class of the S-series, a reduced illuminance, when using lamps having CRI >60. Unfortunately, CRI, by definition and in practice, cannot reliably be used to predict anything except perhaps the naturalness of colour appearance. This is exacerbated at night-time because CRI is based on the visual response under daytime light levels. Even at daytime levels CRI fails to predict the visual appearance of an environment under different types of lamp, and colour appearance under LED lamps, for example, is not well predicted. Although the CRI specification may work with lamps now commonly used for street lighting there will be occasions when it gives inappropriate guidance, for example when extrapolated to distinguish between some other types of lamp.A better method of specification is needed to avoid misapplication. Of equal importance is the need to provide research evidence that can be translated into guidance that is readily accessible to practising lighting engineers. Otherwise it may be ignored, wasting the opportunity for improvements in the visual environment and energy consumption.This project will carry out further tests, supplementing the evidence from recent studies, to generate a dataset against which to test tools for predicting whether an illuminance reduction is appropriate. These tests include brightness matching and visual acuity, using an extended range of lamps. Consultations with lighting engineers will be used to determine how the findings should be presented to ensure an appropriate balance between accuracy and ease of application.