Towards a general optimal model for minimizing nighttime road traffic accidents and road lighting power consumption

Full text

(1)LiU-ITN-TEK-A--10/021--SE. Towards a general optimal model for minimizing nighttime road traffic accidents and road lighting power consumption Ma Jun 2010-06-03. Department of Science and Technology Linköping University SE-601 74 Norrköping, Sweden. Institutionen för teknik och naturvetenskap Linköpings Universitet 601 74 Norrköping.

(2) LiU-ITN-TEK-A--10/021--SE. Towards a general optimal model for minimizing nighttime road traffic accidents and road lighting power consumption Examensarbete utfört i transportsystem vid Tekniska Högskolan vid Linköpings universitet. Ma Jun Handledare Ghazwan Al-Haji Examinator Kenneth Asp Norrköping 2010-06-03.

(3) Upphovsrätt Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare – under en längre tid från publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår. Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning. Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns det lösningar av teknisk och administrativ art. Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för upphovsmannens litterära eller konstnärliga anseende eller egenart. För ytterligare information om Linköping University Electronic Press se förlagets hemsida http://www.ep.liu.se/ Copyright The publishers will keep this document online on the Internet - or its possible replacement - for a considerable time from the date of publication barring exceptional circumstances. The online availability of the document implies a permanent permission for anyone to read, to download, to print out single copies for your own use and to use it unchanged for any non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional on the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility. According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement. For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its WWW home page: http://www.ep.liu.se/. © Ma Jun.

(4) Linköping University LITH- Department of Science and Technology. Linköpings universitet. Towards a general optimal model for minimizing nighttime road traffic accidents and road lighting power consumption. By. Jun Ma (Junma243@student.liu.se). Date: 2010-06-06. Examiners: Kenneth Asp, Professor, PhD Ghazwan Al-Haji, Assistant Professor, PhD. Department of Science and Technology Linköping universitet SE -581 83 Linköping, Sweden Linköping, 2010 I .

(5) II .

(6) Abstract Nowadays, NRTS (Nighttime Road Traffic Safety) and energy saving are very hot topics in transportation field. This thesis investigates a general optimal model for minimizing NRTAs (nighttime road traffic accidents) and power consumption of the road lighting. To establish this model, the relationship between N/D RTAs (Night to Day Road Traffic Accidents) ratio and road lighting condition and the relationship between power consumption and road lighting condition have been studied and explained. A media variable “economic cost” has been chosen which is used for making a connection between these two relationships. The evaluations of NRTAs and power consumption from cost point of view are introduced as well. The impacts of each internal factor defined by author are explained carefully. The result of the model based on these relationships and internal influencing factors is presented in the paper. Finally, the recommendations for reducing NRTAs and/or power consumption, as well as other interesting areas for further study are presented.. Keywords Nighttime Road Traffic Safety, Nighttime Road Traffic Accidents, Road (Street) Lighting, Power Consumption, Cost. III .

(7) Acknowledgement This master thesis in ITS (Intelligent Transport System) program was initiated by and performed at the Department of Science and Technology at Linköping University. First of all, I would like to thank my supervisor Assistant Prof. Dr. Ghazwan Al-Haji for his good advises, inspiration and motivation throughout the project. I think without his support and dedication it was not possible to do this work. I would also like to thank my examiner Prof. Dr. Kenneth Asp at Linköping University. Finally, I would like to thank VTI (Swedish Road and Transport Research Institute) Library in Linköping, Sweden. Because the literatures borrowed from them have helped and supported me a lot to finish this project.. IV .

(8) List of Abbreviations Abbreviations RTAs GNP RTS NRTS NRTAs DRTAs N/D C.I.E. NOK SI VL Uo TI LED DC AC HPS LPS CO SCO NCO VI ITS ACC ISA. Full version Road Traffic Accidents Gross National Product Road Traffic Safety Nighttime Road Traffic Safety Nighttime Road Traffic Accidents Daytime Road Traffic Accidents Night to Day Commission Internationale de L'Eclairage (International Commission on Illumination) Norwegian Kroner International System of Units (Systeme Internationale d’Unites) Visibility Level Luminance Uniformity Threshold Increment Light Emitting Diode Direct Current Alternative Current High-Pressure Sodium Low-Pressure Sodium Cut-Off Semi-Cut-Off Non-Cut-Off Visibility Index Intelligent Transportation System Adaptive Cruise Control Intelligent Speed Adaptation. V .

(9) List of Tables Number Table 1. Name. Page Accident rate (accidents per million vehicle-kilometers) in Sweden as a p7 function of geographical region, general lighting condition and road condition. Table 2 Table 3 Table 4. p8 Deaths per 10 million vehicle miles in the U.S.A. for 1965 p9 Relative personal injury accident rates for different light conditions, p9 Percentage of accidents occurring at night, in Victoria 1963. assuming no changes in weather or traffic. Table 5. Accident rates and severity on London – Birmingham motorway in 2 years, 1960-61. p10. Table 6. The darkness/daylight ratio of casualty rates per motor vehicle kilometer in the Metropolitan Police District in 1966. p11. Table 7. Classification of accidents (on unlit Federal German motorways by day and by night) by seriousness of results. p11. Table 8 Table 9. Accident frequencies before and after relighting. p12 Effects of improved road lighting on the number of accidents. Percentage p13 change in number of accidents. Table 10. The results of nighttime traffic situation before and after road lighting improvement. p14. Table 11. The effect of lighting conditions on the cost of accidents: Washington Area Vehicle Accident Study. p15. Table 12 Table 13 Table 14 Table 15 Table 16 Table 17 Table 18 Table 19 Table 20. Overall costs of street lighting. p16 p17 p23 p23 p35 p40 p47 P49 p55. Road lighting costs in Norway for the years 1986, 1990, 1991, and 1992 CIE recommendations for motor traffic based on luminance level CIE recommendations for motor traffic based on visibility level Source types comparison form Effect on accidents of increase in speed limits Summary of the impact of each internal influencing factor to the model A abbreviation list of internal influencing factors Summary of each methodology impact on NRTAs and power consumption. VI .

(10) List of Figures Number Figure 1 Figure 2. Name. Page Thesis structure diagram p6 Average rate of reported casualty accidents by hour and day in Melbourne p8 Metropolitan Area, 1963. Figure 3. The five primary factors influencing the visibility conditions in the night traffic. Figure 4 Figure 5. VL along a certain road section, depending on the driver’s age. p18. p22 ‘Best’-fitted relationship between the dark/day accident ratio and the p24 average road-surface luminance. Figure 6. The variation of night to day accidents with light level provided by street lighting. p25. Figure 7. The relationship between forward current and luminous intensity in LED 466-3582. p27. Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14. A general relationship between power consumption and luminous intensity. p28 General structure of the model p30 General curves of two functions p31 (a) Reflective clothes, (b) Reflective wrist or ankle ties p32 Age and fatality rates by time of day p33 Cross-sectional schematics of lanterns p35 Revealing power (RP) at the darkest location on the road as a function of p37 the average road-surface luminance (L av ) for three combinations of overall Uo and TI. Figure 15. Accident rate, including damage accidents of DM 1000 or over and/or personal injuries, by day and by night as a function of traffic volume on unlit 2 lane dual carriageway. P38. Figure 16. Recommended visibility index as a function of design velocity and probability of detection of a simple target. p41. Figure 17 Figure 18 Figure 19 Figure 20 Figure 21 Figure 22 Figure 23 Figure 24. (a) Reflective traffic sign, (b) Reflective road mark. p42 p44 p45 p46 p50 p52 p53 p54. A comparison of visibility distance between dry and wet road surface The relationship between speed and friction A photo of road lighting condition in the city of Norrkoping in Sweden A solar-powered street lamp in Beijing rural road An example of the actual different in road lighting Chinese drivers’ road traffic accidents causation proportion Night road traffic potential dangers. VII .

(11) Table of contents 1 . 2 . 3 . 4 . 5 . 6 . Introduction ............................................................................................................. 3 1.1 Background ............................................................................................... 3 1.2 Problem formulation ................................................................................. 4 1.3 Aim of the study........................................................................................ 4 1.4 Methodology ............................................................................................. 4 1.5 Thesis outline ............................................................................................ 4 1.6 Thesis structure diagram ........................................................................... 6 NRTAs and road lighting ........................................................................................ 7 2.1 NRTAs situation ........................................................................................ 7 2.1.1 NRTAs frequency ............................................................................... 7 2.1.2 NRTAs severity ................................................................................ 10 2.2 Road lighting improving NRTS .............................................................. 11 2.2.1 C.I.E. conclusion .............................................................................. 12 2.2.2 Norway ............................................................................................. 13 2.2.3 Hungary............................................................................................ 13 2.3 Cost evaluation........................................................................................ 14 2.3.1 Road traffic accidents cost evaluation ............................................. 14 2.3.2 Road lighting cost evaluation........................................................... 16 Road luminance in nighttime traffic ..................................................................... 18 3.1 Visibility determining factors in nighttime traffic .................................. 18 3.2 Road luminance indicators – Luminance & Visibility Level.................. 19 3.2.1 Lighting glossaries and units conversion ......................................... 19 3.2.2 Definition of VL .............................................................................. 20 3.2.3 Influencing factors of VL................................................................. 20 3.2.3.1 Contrast .................................................................................... 20 3.2.3.2 Target property ......................................................................... 21 3.2.3.3 Uniformity ................................................................................ 21 3.2.3.4 Glare effect ............................................................................... 21 3.2.3.5 Age of the observer .................................................................. 21 3.2.4 C.I.E. recommended standards ........................................................ 22 Relationship between N/D RTAs ratio and road lighting condition ..................... 24 4.1 Exponential function relationship ........................................................... 24 4.2 Logarithm function relationship ............................................................. 25 Relationship between power consumption and road lighting condition ............... 27 5.1 Forward current and luminous intensity relationship ............................. 27 5.2 Power and luminous intensity relationship ............................................. 28 General N/D RTAs ratio and power consumption optimal model based on luminous intensity variable .................................................................................. 30 6.1 General structure of the model ................................................................ 30 6.2 General curves and expressions of the functions .................................... 31 6.3 Influencing factors of the model ............................................................. 31 6.3.1 Human factors .................................................................................. 31 1.

(12) 7 . 8 9 . 6.3.1.1 Physiological response ............................................................. 31 6.3.1.2 Human behavior ....................................................................... 32 6.3.1.3 Driver age ................................................................................. 33 6.3.2 Lantern factors ................................................................................. 34 6.3.2.1 Lantern design .......................................................................... 34 6.3.2.2 Mounting height ....................................................................... 36 6.3.2.3 Uo TI influencing ..................................................................... 37 6.3.3 Traffic factors ................................................................................... 38 6.3.3.1 Traffic flow ............................................................................... 38 6.3.3.2 Speed limits .............................................................................. 39 6.3.3.3 Road environment .................................................................... 41 6.3.3.4 Road type .................................................................................. 42 6.3.4 Weather factors ................................................................................ 43 6.3.4.1 Rain .......................................................................................... 43 6.3.4.2 Fog ............................................................................................ 45 6.3.4.3 Snow ......................................................................................... 46 6.3.5 Summary .......................................................................................... 47 6.4 Mathematical expression of the model ................................................... 47 6.4.1 Objective function ............................................................................ 47 6.4.2 Constrains ........................................................................................ 49 Recommendations for reducing NRTAs and/or power consumption ................... 50 7.1 Engineering ............................................................................................. 50 7.1.1 Solar-powered road lighting ............................................................ 50 7.1.2 LED street lamp ............................................................................... 51 7.1.3 Dimming and intelligent control ...................................................... 51 7.1.4 ITS applications ............................................................................... 52 7.2 Enforcement – speed limits ..................................................................... 53 7.3 Education ................................................................................................ 54 7.4 Summary ................................................................................................. 55 Conclusion and further study ................................................................................ 56 Reference .............................................................................................................. 57 9.1 Book reference ........................................................................................ 57 9.2 Website reference .................................................................................... 59 . 2.

(13) 1 Introduction 1.1 Background With the increasing cars and limited road resource, RTAs (Road Traffic Accidents) are increasing every year. In the year 2020, RTAs will become the 3rd largest reason causing death. [1] The economic loss caused by RTAs in developing countries is estimated to cost 1%-4% of a country’s GNP (Gross National Product) annually. [2] Therefore, RTS (Road Traffic Safety) becomes a more and more serious and intractable problem all over the world. NRTS (Nighttime Road Traffic Safety) is one of the important branches in RTS, since it is easier to get accidents in nighttime than in daytime. In China, the statistic results show that NRTAs (Nighttime Road Traffic Accidents) are 2 – 2.5 times than DRTAs (Daytime Road Traffic Accidents) depend on different road lighting situations. [3] In America, according to PHILIP company statistic results, it shows that a half traffic accidents caused people death happen at night, if it is calculated by accidents per million vehicle miles, NRTAs are around 2.5 times than DRTAs. [4] In Europe, it also has data said that for motor vehicles, the risk of having an accident in the dark is about 1.5 – 2 times higher than in daylight. [5] Moreover, NRTAs also have more serious results comparing to DRTAs. Fatal injury has a higher proportion among all injuries in NRTAs. It will be illustrated and proved more in the later chapter. The main reason of the increasing accidents rate at night is because darkness reduces the visibility level of road consequently reduces human visual performance. [6] In the daytime, under bright sunlight, the illuminance on the road can be as high as 10 5 lux, but at night, the average illuminance on a lighted main road is about 10 lux, the driving task is much more onerous. [7] The recommended artificial lighting level in an office is about 500 lux, but to provide this order of lighting on a roadway would be prohibitively expensive. [8] Therefore, night drivers have to drive with around 0.01% of the light he is used to by daytime. In the night, road lighting has a big and important effect on traffic safety. According to the published report by C.I.E. (International Commission on Illumination), good road lighting system can reduce traffic accident rate 30% in urban road traffic, 30% in highway road traffic, and 45% in rural road traffic. [9] It is no doubt that road lighting can help on improving traffic situation at night and reducing NRTAs caused by darkness. However, on the other hand, it also costs lots of energy and money (installation cost and maintenance cost) to support these road lights working in the night. It has data shows that road lighting costs represent one of the largest components of a city government’s utility bill, often accounting for 10-38% of the total bill. (1) 3.

(14) Lighting represents 15-20% of the world yearly electricity, and 3% of the total electricity consumption goes to road lighting. In the United States, road lighting consumes 14 billion kWh annually. [10] Los Angeles’ streetlights use 168 gigawatt hours of electricity annually cost of $15 million. (1). 1.2 Problem formulation The problem need to be faced in this paper is to find out a solution that could calculate an optimal luminance of the road lighting in order to balance the problems between NRTAs reducing and energy saving.. 1.3 Aim of the study The purpose of this study is to establish a general model in order to find the optimal (balanced) point (luminance) for minimizing the total cost of NRTAs and road lighting power consumption which could be used in a specific case study in future. Meanwhile, point out how these factors inside model influence the result.. 1.4 Methodology 1) Study on some general concepts and standards concerning road lighting as a preparation. 2) Find the relationship between N/D (Night to Day) RTAs ratio and road lighting condition as well as the relationship between power consumption and the road lighting condition from the literature or datasheet. 3) Construct a general model for minimizing N/D RTAs ratio and road lighting power consumption, lay out constrains and analyze influences of the inner factors.. 1.5 Thesis outline Chapter 1 Introduction It introduces the background of NRTAs situation, road lighting condition as well as power consumption in real traffic. Problem formulation, aim, methodology and thesis outline are presented. Chapter 2 NRTAs and road lighting It describes how worse the situation of NRTAs is and how much road lighting can help on NRTS. Furthermore, it evaluates NRTAs and power consumption of the road lighting from cost point of view individually. Chapter 3 Road luminance in nighttime traffic It explains some basic and necessary knowledge towards road luminance in nighttime traffic which could help to understand later part of the paper.. 4.

(15) Chapter 4 Relationship between N/D RTAs ratio and road lighting condition It shows the relationship between N/D RTAs ratio and road lighting condition based on Exponential function and Logarithm function. Chapter 5 Relationship between power consumption and road lighting condition It shows the relationship between power consumption and road lighting condition based on quadratic function. Chapter 6 General N/D RTAs ratio and power consumption optimal model based on luminous intensity variable It introduces general structure of the model and general curve and expression of the function. Moreover, it explains how these internal factors defined by author affect to the model. Chapter 7 Recommendations for reducing NRTAs and/or power consumption It gives some recommendations for reducing NRTAs and/or power consumption of the road lighting based on ‘3E’ principle. Chapter 8 Conclusion and further study The final conclusion and decision are presented. Interesting areas of further study are also introduced.. 5.

(16) 1.6 Thesis structure diagram. Figure 1: Thesis structure diagram. The above figure shows the structure diagram of this thesis. It shows that the problem is to minimize (balance) the problems between NRTAs and power consumption. The central element of this structure is road lighting which can help to reduce NRTAs, but require dissipating more power. A general optimal model was developed and used to calculate (optimize) the luminous intensity of the road lighting. It mainly contains three parts which are relationship between accidents and luminance, relationship between power and luminance, and internal factors. In this diagram, it also marks relative chapter number for each block. Some basic and necessary knowledge towards road luminance in nighttime traffic presented in chapter 3 can be considered as a part of literature review which helps on model construction. The last part of this thesis is recommendations for improving NRTAs and/or power consumption.. 6.

(17) 2 NRTAs and road lighting 2.1 NRTAs situation Lots of published data shows that among these RTAs, NRTAs take up higher proportion comparing to DRTAs. The frequency and severity of NRTAs compared with DRTAs in different countries will be illustrated in figures or tables as followed.. 2.1.1 NRTAs frequency The most acceptable method of presenting accidents data is by means of an accidents rate, i.e. accidents per million vehicle miles. To give the percentage value of NRTAs, it requires knowing the quantity of travel occurring at night and the number of accidents occurring at night firstly. However, the total mileages of travel and total number of accidents are often available, the day-night split is rarely known. Various references strongly suggest that the percentage at night is about 25%. [6] This percentage is just an approximate value at here. It could be changed according to different areas of the earth and different seasons. For instance, in North Europe, the length of nighttime is even longer than daytime in winter. It will cause that the percentage of travel in darkness is higher in winter. Sweden The data from Rumar (1979) showed accidents rate in the daylight and in the darkness separately in Sweden which is shown as below. [11] Region North Central South. Daylight 0.37 0.46 0.55. Darkness 0.83 1.01 1.06. Table 1: Accident rate (accidents per million vehicle-kilometers) in Sweden as a function of geographical region, general lighting condition and road condition [11]. In Table 1, it can be found accidents rate in the darkness is much higher than the case in the daylight in all regions. Thus, high proportion on NRTAs is a serious problem in Sweden. Australia The data from the Traffic Commission of Victoria contrasted the night and day accidents situation for Victoria in 1963 which is shown in Table 2. [12]. 7.

(18) Accident Type Multiple vehicle Single vehicle Pedestrian All types. Fatal accidents 55% 52% 70% 60%. All injury accidents 38% 52% 39% 42%. Table 2: Percentage of accidents occurring at night, in Victoria 1963 [12]. From the data in Table 2, it can be seen that 42% of all casualty accidents occur at night and this value even rises higher to 60% for fatal accidents. The number of fatalities occurring at night exceeds those by day for each type of accident (multiple vehicle, single vehicle and pedestrian). This is in spite of the fact that only about 20% of vehicular traffic takes place at night. [13] Another example from Australia is in the Melbourne Metropolitan Area which was also applied by the Traffic Commission of Victoria. [12] It shows that the rate of casualty accidents occur between 6pm and 6am (nighttime period contains dusk and dawn) exceeds the rate of casualty accidents occur between 6am and 6pm (daytime period) about 2 times on weekdays and about 5 times at the worst periods like Figure 2 shows. The relative risk increase is even higher during holiday nights probably due to the influence of alcohol, fatigue and different driver populations.. Figure 2: Average rate of reported casualty accidents by hour and day in Melbourne Metropolitan Area, 1963 [12]. 8.

(19) U.S.A. The data from the National Safety Council which is shown in Table 3 showed that the overall ratio of N/D fatality rates in the U.S. was about 2.5 times (. Night Death rate ) Day Death rate. in different road types. [14] Road type National Urban Rural. Time period Day Night Day Night Day Night. Percentage of deaths 47% 53% 45% 55% 49% 51%. Death rate 3.8 9.8 2.3 6.5 5.1 12.8. Table 3: Deaths per 10 million vehicle miles in the U.S.A. for 1965 [14]. U.K. The Road Research Laboratory concluded that the personal injury night accidents rate was higher than the day rate, however, the actual N/D ratio depended on the standard of the road lighting as shown in Table 4. [17] Daylight 1.0. Good lighting 1.3. Dark Average lighting Poor lighting 1.6 1.8. No lighting 2.0. Table 4: Relative personal injury accident rates for different light conditions, assuming no changes in weather or traffic [17]. Like Table 4 shows N/D accidents ratio is decreasing with road lighting condition improved better and better. It implies that road lighting can help on NRTAs reduction. Christie (1968) showed up the time trend of the proportion of accidents occurring at night had been rising gradually but steadily: in 1955 the proportion was 27% but by 1965 the proportion had reached 35%. [18] The actual night casualty rate had increased by 6% within a decade. The recent data showed that in UK in 1980 there were 79432 accidents after dark, compared with 171515 during the day [19], but as vehicle mileage was less during the hours of darkness, the accident rate at night was about 1.8 times than the day accident rate.. 9.

(20) Germany Hartmann and Linde (1970) quoted studies which showed that 24% of all accidents in Germany took place at night at a time when the traffic volume accounted for no more than 10 to 15% of the total traffic in 1954. In one state, 31.5% of all personal injury accidents took place during the hours of darkness and accounted for 43% of all fatal injuries in 1962. [15] Belgium Lefevre (1962) described a study of accidents on 1,300 km of main roads situated outside built up areas. The ratio of N/D accidents rates was about 1.4 times for roads without lighting or with poor to average lighting and about 1.2 times for roads with good lighting. [16] Developing countries Since in developing countries, the research on NRTS towards road lighting was paid not enough attention, the material and literature concerning NRTAs are vey less. It is quite hard to find some data about NRTAs situation. However, the data from a developing country - China still has been found which can give us a consultation. Like it has been said in the background section, in China, NRTAs are 2 – 2.5 times than DRTAs depend on different road lighting situations. [3] It is a little higher than the value (1.5 – 2 times) in most of the European countries, due to higher traffic flow and worse traffic environment. Therefore, it is not hard to imagine that in developing countries, for instance the countries in Africa, NRTS should be worse than these developed countries, and NRTAs should have higher proportion, due to their poor road lighting system and bad driving behaviors.. 2.1.2 NRTAs severity NRTAs not only have higher occurring rate comparing to DRTAs, but also have more severe consequence than DRTAs. The following data and conclusions prove this point. U.K. Coburn (1965) showed that accidents at night on a UK motorway were more severe in their consequences than those by day in Table 5. [20] The ratio (dark/light) is increasing with class of accidents changed from slight to fatal. Class of accident Fatal Serious Slight All injury. Accident per million vehicle miles Light hours Dark hours 0.03 0.08 0.18 0.40 0.24 0.42 0.45 0.91. Ratio (Dark/Light) 2.67 2.22 1.75 2.02. Table 5: Accident rates and severity on London – Birmingham motorway in 2 years, 1960-61 [20] 10.

(21) One more example from U.K. shows a similar result. Christie (1968) gave the ratio of the darkness to daylight casualty rates for the London Police District in 1966. The ratio for each degree of injury is greater than 1, as shown in Table 6, the value is increasing with injury severity. [18] Degree of injury Fatal Serious Slight All severities. (Casualty rate per vehicle kilometer in dark hours) ÷ (Casualty rate per vehicle kilometer in daylight hours) 2.16 1.89 1.47 1.54. Table 6: The darkness/daylight ratio of casualty rates per motor vehicle kilometer in the Metropolitan Police District in 1966 [18]. Germany Hartmann and Linde (1970) gave data in Table 7 for unlit motorways in Germany which showed that accidents at night were more serious than those by day. [15] Heavy injuries and fatal injuries in total at night is 17% which is higher than the case in daylight 12%. Time. Day Night. Material damage 1,117 (55%) 540 (53%). Consequences of accident Light Heavy injuries injuries 681 235 (33%) (11%) 309 142 (30%) (14%). Total Fatal injuries 27 (1%) 25 (3%). 2060 1016. Table 7: Classification of accidents (on unlit Federal German motorways by day and by night) by seriousness of results [15]. New South Wales Fisher (1967) using accidents data for the 3-years period (1963-65) for New South Wales, showed that the severity ratio (number of persons killed divided by the total casualties) for the nighttime was markedly higher than for the daytime. The average number of fatalities per injury is about 1.6 times greater by night than by day. [12]. 2.2 Road lighting improving NRTS The main reason causing higher road accidents rate and more severe accidents consequence in the night for all classes of road is poor visual performance due to darkness. Therefore, to increase NRTS, the most important thing is to enhance the visibility at night.. 11.

(22) 2.2.1 C.I.E. conclusion At the 1960, the C.I.E. has concluded that road lighting definitely reduced both the incidence and severity of NRTAs. [21] The results are shown in Table 8. Pedestrian. Other. Total injury. Fatal. Serious. Slight. Total. Fatal. Serious. Slight. Total. Fatal. Serious. Slight. Total. Before B. 5. 84. 230. 319. 11. 140. 778. 929. 16. 224. 1,008. 1,248. After A. 11. 85. 238. 334. 6. 159. 926. 1,091. 17. 244. 1,164. 1,425. Before b. 15. 52. 92. 159. 13. 71. 262. 346. 28. 123. 354. 505. After a. 6. 31. 54. 91. 9. 59. 244. 312. 15. 90. 298. 403. Daylight. 2.20. 1.01. 1.04. 1.05. 0.54. 1.14. 1.19. 1.17. 1.06. 1.09. 1.16. 1.14. Darkness. 0.40. 0.60. 0.59. 0.57. 0.69. 0.83. 0.93. 0.90. 0.54. 0.73. 0.84. 0.80. r. 0.18. 0.59. 0.57. 0.55. 1.27. 0.73. 0.78. 0.77. 0.50. 0.67. 0.73. 0.70. 5%. -. 1%. 0.1%. -. -. 5%. 1%. -. 5%. 0.1%. 0.1%. Daylight. Darkness. After/before. Significance level. Table 8: Accident frequencies before and after relighting [21]. The combined data given in Table 8 are divided into three parts: accidents in which a pedestrian was injured, accidents in which no pedestrian was injured and all injury accidents together. Furthermore, in each part, the accidents are classified according to severity. In each part, it has been further classified according to severity of the accidents from slight to fatal. “Before” and “After” in this table mean before and after the improvement respectively. To estimate the effect of the improvements in road lighting, it is necessary to consider the fact that the increased traffic caused by this relighting will increase the number of accidents both by night and by day. Therefore, probably the best available estimated parameter of the effect of the lighting is the quantity r given in Table 8 which is defined as the ratio of the actual number of accidents in darkness before and after the improvement to the expected number. [22] If b and a denote the number of accidents in darkness before and after the change, and B and A are corresponding daylight accidents, then if the change had not been made and if the number of accidents in darkness had the same increased proportion as the number of accidents in daylight, bA the expected number of accidents in darkness in the after period would be , where, B A is the increased accidents ratio caused by increased flow. Thus: B r =a÷. bA a A = ( )÷( ) B B b 12. (2.1).

(23) If r = 1, the lighting has apparently no effect; if r = 0.7, the apparent effect is a reduction of 30 per cent in the night accidents. The “significance level” quoted in Table 8 is a measure of confidence that the observed change is real and not due to chance causes. [22] The 5% level means that there is only a probability of 1 in 20 that such a large change would have arisen from chance. The 5% level is usually accepted by the standard of statistical significance. The 0.1% level indicates a highly significant result. In Table 8, it can be seen that casualty accidents are reduced by 30% overall. In addition serious accidents are reduced 33% more than slight accidents 27%.. 2.2.2 Norway Studies in Norway of the effect on accidents of improving the existing level of lighting are included here which are presented in Table 9. [5] Accident severity. Percentage change in number of accidents Accident types affected Best estimate 95% confidence interval Increasing the level of lighting by up to double the previous lighting level Injury accidents Accidents in darkness -8 (-20; +6) Property-damage-only Accidents in darkness -1 (-4; +3) Increasing the level of lighting by up to 2 – 5 times the previous lighting level Injury accidents Accidents in darkness -13 (-17; -9) Property-damage-only Accidents in darkness -9 (-14; -4) Increasing the level of lighting by up to 5 times the previous lighting level or more Fatal accidents Accidents in darkness -50 (-79; +15) Injury accidents Accidents in darkness -32 (-39; -25) Property-damage-only Accidents in darkness -47 (-62; -25) Table 9: Effects of improved road lighting on the number of accidents. Percentage change in number of accidents [5]. From above table, with the level of road lighting is improving better and better, the number of various accidents in darkness is reduced more and more. The results clearly show that how big of the effect of improved lighting on accidents depends on how big of the improvement.. 2.2.3 Hungary In Hungary, the accidents occurred both within and outside built-up areas are more serious at night than in daytime. Within built-up areas, 10% of injury accidents 13.

(24) occurred at night is fatal, outside built-up areas are 20%. However, in daytime only about 5% of injury accidents are fatal. [23] In the year of 1985, an improvement of the road lighting had been done in cities. This up-to-date technology of road lighting was excepted to reduce the number and severity of accidents in nighttime. There were 11 sites chosen (mostly junctions and road section with some junctions) from different cities of the country, and accident statistics were compiled regarding those sites for the time before and after the improvement taken, concerning the road lighting. The results are shown in Table 10. [23] Items. Number of accidents (involving personal injury) Total economic losses (million Forint) Economic losses per accident (thousand Forint) Number of accidents during nighttime per number of accidents during daytime Number of accidents per 1km of road. Before (1983-84) 100% 42. After (1986-87) 35. Dropped to % (percent value) of the before 83%. 10.144 241.5. 4.112 117.6. 41% 49%. 0.22. 0.17. 77%. 1.8. 1.5. 83%. Table 10: The results of nighttime traffic situation before and after road lighting improvement [23]. From above table, it can be seen that after the improvement, the number of injury accidents dropped to 83% of the before situation. The estimated economic losses fell more than 50%, compared to the before situation due to the decrease of severity of accidents. Therefore, it can be concluded that road lighting has a big contribution on NRTAs reduction. The better road lighting system it has, the better NRTS it has.. 2.3 Cost evaluation 2.3.1 Road traffic accidents cost evaluation RTAs not only hurt or kill people’s life, but also bring a huge economic cost every year. Depend on different cities, road network environments, human behaviors, etc., the costs on RTAs could have a big difference. Furthermore, even in the same area with fixed conditions, the annual cost of RTAs is also changed yearly. Therefore, it is very hard to give an exactly value of how big the annual accidents cost in this evaluation. However, an example data from U.S.A. will be still given as followed which could give us a consultation in this field.. 14.

(25) Table 11: The effect of lighting conditions on the cost of accidents: Washington Area Vehicle Accident Study [24]. 15. 425,396 204,162 415,982 629,853 $1,675,393 817,905 314,625 1,197,744 1,547,961 $3,878,235 620,885 157,065 987,688 $1,765,638 3,127,187 1,377,216 2,471,306 5,024,585 $12,001,294. 17 7 12 7 43 17 3 17 16 53 10 2 10 22 86 28 60 81 255. Fatal injury accidents Involvements Total cost $ 42 1,263,001 18 858,429 29 701,515 48 1,859,083 137 $4,682,028. Collision between vehicles Daylight Dawn or dusk Dark with street lighting Dark without street lighting SUBTOTAL Collision with pedestrian Daylight Dawn or dusk Dark with street lighting Dark without street lighting SUBTOTAL Collision with fixed object Daylight Dawn or dusk Dark with street lighting Dark without street lighting SUBTOTAL Ran off roadway Daylight Dawn or dusk Dark with street lighting Dark without street lighting SUBTOTAL Total of four major types Daylight Dawn or dusk Dark with street lighting Dark without street lighting TOTAL *Four accident types selected as most prevalent. Light condition by type of accident*. 16,318 2,435 6,034 2,856 27,643. 194 40 75 310 619. 522 81 499 441 1,543. 1,353 227 281 55 1,916. 12,627,216 2,654,650 5,445,741 3,383,613 $24,111,220. 237,438 116,463 166,003 564,907 $1,084,811. 731,084 116,293 989,212 826,130 $2,662,719. 576,288 54,310 365,515 104,999 $1,101,112. Non-fatal injury accidents Involvements Total cost $ 14,249 11,082,406 2,087 2,367,584 5,179 3,925,011 2,050 1,887,577 23,565 $19,262,578. 41,266 5,346 13,992 4,198 64,802. 367 87 142 261 857. 616 148 412 321 1,497. -. 7,236,553 1,008,469 3,111,388 1,162,155 $12,518,565. 120,470 79,373 32,076 113,910 $345,829. 301,054 52,350 193,467 161,880 $708,751. -. 57,670 7,809 20,086 7,135 92,700. 571 127 219 581 1,498. 1,155 232 928 778 3,093. 1,370 234 293 62 1,959. 22,990,956 5,040,335 11,029,435 9,570,353 $48,631,079. 978,793 195,836 355,144 1,666,505 $3,196,278. 1,850,043 483,268 2,380,423 2,535,971 $7,249,705. 1,001,584 258,472 781,497 734,852 $2,776,505. Property damage only accidents All severity classes Involvements Total cost Involvements Total cost $ $ 40,283 6,815,029 54,574 19,160,436 45,111 876,746 7,216 4,102,759 13,438 2,885,845 18,646 7,512,371 3,616 886,365 5,714 4,633,025 62,448 $11,463,985 86,150 $35,408,591. 399 645 549 1,341 $525. 1,714 1,542 1,622 2,868 $2,134. 1,602 2,083 2,565 3,260 $2,344. 731 1,105 2,667 11,852 $1,417. $ 351 569 403 811 $411. Average cost.

(26) Table 11 in last page showed the result of accidents cost evaluation in Washington Area in 1965, which was researched by Berry (1968). [24] From Table 11, it can be seen that the average cost in dark without street lighting is much bigger than the case in dark with street lighting, especially in the type of collision with pedestrian. Furthermore, it also informs us street lighting helps on reducing the cost of NRTAs a lot. In principle, the annual total cost of street lighting should not be higher than the annual difference between NRTAs cost without lighting and with good lighting condition.. 2.3.2 Road lighting cost evaluation The total road lighting cost could be mainly divided into two parts according to Fisher (1971) which is shown as below. [6]. Capital. Maintenance. Overall costs of street lighting Lanterns Mountings Wiring Labour Amortisation Period Interest charges Energy consumption Lamp replacements Cleaning Table 12: Overall costs of street lighting [6]. In this paper, it assumes the street lights have already been installed on road. Meanwhile it only defined the energy consumption as our research target which is one of the maintenance cost. The difficulty with road lighting cost evaluation is the variation can occur depend on different technologies it has been used, electricity price, and city size, etc. Therefore, it is also very hard to give a general value on this cost. Furthermore, in a world-wide study of the costs of road lighting, a number of countries supplied their costs per kilometer for lighting and running two ‘standard’ roadways using their normal range of equipment and meeting their national road lighting standards. It was hoped that relationships could be derived between levels of lighting criteria and the cost of achieving them, but so great were the cost variations in road lighting around the world, that it was concluded that no useful generalizations could be made. [25] Two consulting statistics emerging from this study are: (a) The average annual costs of operating road-lighting installations are 7 dollars per meter for urban traffic routes and 9 dollars per meter for rural motorways 16.

(27) (b) Assuming road are lit to C.I.E. standards, the average energy demand is about 10 watts per linear meter of carriageway. An example concerning road lighting cost in a European country to show the differences is also illustrated at here. The following table for the costs of the road lighting installed along national and county highways in Norway was available for the years 1986, 1990, 1991, and 1992. [5] Year 1986 1990 1991 1992. Road type National highway National highway County highway National highway County highway National highway County highway. Km road 41 66 22 62 8 90 15. Cost per km road (NOK) 263,000 441,000 157,000 219,000 315,000 441,000 194,000. Table 13: Road lighting costs in Norway for the years 1986, 1990, 1991, and 1992 [5]. This overview shows relatively large variations in equipment costs for road lighting. A weighted average (with the length of road as weight) for national highways is NOK (in current Norwegian kroner) 360,000 per km of road (1986-1992 prices). Where 360,000 is calculated by the way: 263000 * 41 + 441000 * 66 + 219000 * 62 + 194000 * 90 = 360000 NOK 41 + 66 + 62 + 90. (2.2). It has used the total cost for these four years to be divided by the total length of roads in km. Furthermore, the annual running and maintenance costs for road lighting are NOK 10,000 – 20,000 per km road, depending on the standard of the equipment. [5] From above examples, it could be seen whatever which country it is and which standard it uses, road lighting always cost a lot of money and power.. 17.

(28) 3 Road luminance in nighttime traffic 3.1 Visibility determining factors in nighttime traffic According to Rumar (1975), seeing (detection) distance in the night traffic is determined mainly by five factors which are shown in Figure 3. [26] In this paper, it mainly focuses on road lighting under the light source. However, the internal influencing factors inside developed model are also related with the other factors such as: human eyes, visual targets, atmosphere and so on.. Figure 3 The five primary factors influencing the visibility conditions in the night traffic [26]. 1. Visual performance (human eyes) 2. Artificial illumination (light source) A. Vertical illumination (road lighting) B. Horizontal illumination (vehicle headlights) 3. Visual targets A. Vertical targets (other road users, signs, obstacles) B. Horizontal targets (road surface, road marking) 4. Signal lights A. Stationary signals (traffic signals, railway signals, road-work signals) B. Dynamic signals (vehicle signal lights, vehicle emergency lights) 5. Optical media A. Atmosphere (clear, haze, fog, snow, rain, smoke) B. Vehicle optical media (windows and mirrors: clean, tinted, dirty, scratched) 18.

(29) Short seeing distance could cause either a low safety speed or a high probability of collisions. To improve seeing distance in night traffic, one or more of these factors have to be changed. Among these influencing factors, road lighting is the one has the biggest impact on improving night traffic safety and also the one has the biggest energy consumption and economic cost.. 3.2 Road luminance indicators – Luminance & Visibility Level 3.2.1 Lighting glossaries and units conversion To explain the conversions between different lighting units, some glossaries concerning lighting are necessary to be introduced firstly which are summarized as followed. Illuminance – The density of the luminous flux incident on a surface. It is the quotient of the luminous flux by the area of the surface when the latter is uniformly illuminated. [27] Luminance – The luminous intensity of any surface in a given direction per unit of projected area of the surface as viewed from that direction. The unit is cd/m 2 . [27] Candela – The SI (International System of Units) unit of luminous intensity. Formerly the term candle was used. [27] Footcandle – The illuminance of a surface one square foot in area on which there is uniformly distributed light flux of one lumen or the illuminance produced on a surface all points of which are at a distance of one foot from a directionally uniform point source of one candela. 1 footcandle equals 10.76 lux. [27] Lumen (Flux) – A unit is used to measure the quantity of light. One lumen is the amount of light that falls on an area of one square foot every point of which is one foot from the source of one candela. A light source of 1 candela emits a total of 12.57 lumens (lm). [27] Lux – The SI unit of illuminance. It is defined as the amount of light on a surface of one square meter all points of which are one meter from a uniform source of one candela. 1 lux equals 0.0929 footcandle. [27] In the U.S.A. lighting levels are specified in terms of illumination which is the light flux incident per unit projected area. The American term is usually foot candles (f.c.). However, the preferred term which is used in Australia and U.K. is lumens per square foot (lm/ft 2 ). The metric equivalent is lux or lumen per square meter. Luminance is the photometric correlate of brightness and the imperial term is ft.l. The metric. 19.

(30) equivalent is cd/m 2 which is widely used to indicate the level of road lighting nowadays. [6] The conversions between these units are shown as below: [6] 10 lux = 1 f.c. = 1 lumen ft 2 3 cd/m 2 = 1 ft.l 1 lumen/ft 2 = 0.17 ft.1. 3.2.2 Definition of VL VL (Visibility Level) is used as an indicator to show the situation of the road luminance in nighttime traffic. According to Adrian’s visibility equations, VL of the critical object which is on the road surface can be expressed by the formula: [28]. VL =. ΔLactual L − Lb = t ΔLthreshold Lt 0 − Lb. (3.1). where, ΔLactual is the luminance difference between target and its back ground for the real road conditions, ΔLthreshold is the luminance difference needed for minimal (critical) visibility, between a target of certain angular size and its background. Lt is the target luminance, Lt 0 is the critical target luminance to make it visible, and Lb is the background luminance. The unit of VL is nothing which is not the same as luminance’s unit cd/m 2 .. 3.2.3 Influencing factors of VL The influencing factors of VL could be many and complex. Such as: contrast, target size and presenting time, uniformity, glare effect, as well as the age of the observer, etc. [29] [30] The below paragraphs will give a short description for each main factor to introduce how they influence VL individually. 3.2.3.1 Contrast According to the VL formula, it can be found the difference between background and target luminance which is named contrast is a directly influencing factor to VL. The bigger difference it has, the easier to be saw. Whereas, if target luminance is close to background luminance, whatever how big (bright) the road luminance it has, VL is still low and the situation is in danger.. 20.

(31) Experiments on threshold contrast perception indicate that contrast sensitivity by night is only about 10% of that by day. [31] It means that the brightness difference between an object and its background, expressed as a proportion of the background brightness, must be 10 times greater by night than by day for the object just to be seen. It is also one of the important reasons causing NRTAs. 3.2.3.2 Target property Target property could be the size of the target, presenting time and position on road, etc. Big target has a higher probability to be seen comparing with small one. Meanwhile, longer presenting time can be easier caught than the shorter one. The position of target should be in a suitable place. It should not be covered by some obstacles or at the place which blocks the way of the road users. 3.2.3.3 Uniformity Uniformity is the ratio between the minimum luminance and the maximum luminance in such road section and a certain condition. The bigger value it has, the closer luminance between the minimum and the maximum. In other words, big ratio means more uniform and safer condition. 3.2.3.4 Glare effect Glare can make drivers feel uncomfortable. In worst case, it even can lead drivers lose their vision. When glare comes into human eyes, it will generate a veiling luminance ( L v ) which can reduce the contrast between target and background. C.I.E. has introduced a variable TI (Threshold Increment) to describe road lighting glare effect on drivers’ vision function. It has also recommended to use percentage value to express the loss of contrast on the observing object caused by glare. The formula expression is: [3]. TI (%) =. 100 Lv % Lav + Lv. (3.2). where, Lv is the veiling luminance generated by glare in eyes, Lav is the average luminance of road surface. The stronger glare it has, the bigger percentage value it has. 3.2.3.5 Age of the observer With the age increasing, human vision ability will decrease. Therefore it will influence the visibility level. A relative research showed that the relationship between 21.

(32) age and visibility level. The selected road section was a road near Strasbourg in France. [32]. Figure 4: VL along a certain road section, depending on the driver’s age [32]. In Figure 4, the dashed line shows the threshold (VL = 7) between good condition (up) and bad condition (down). The VL is decreasing with the age increasing. Age factor has a bigger effect on good VL condition comparing to poor VL condition.. 3.2.4 C.I.E. recommended standards The recommendations based on road surface luminance level from C.I.E. publication No. 115 are shown in Table 14. [33] M1 to M5 indicates different lighting classes with different luminance levels for different road types. L avg is the average luminance level. U l is the recommended longitudinal luminance uniformity ratio based on visibility concept. U o is the overall luminance uniformity in the recommendations of C.I.E. based on visibility level. TI means Threshold Increment values that restrict glare.. 22.

(33) Table 14: CIE recommendations for motor traffic based on luminance level [33]. Moreover, there is another recommendation in the same C.I.E. Publication based on VL variable which is shown in Table 15. [33] L min and L max indicates minimum and maximum luminance level respectively.. Table 15: CIE recommendations for motor traffic based on visibility level [33]. From above tables, it can be seen the required luminance/visibility level for road lighting is increasing from M5 to M1. It means the level of road type becomes higher and higher, and NRTS becomes better and better. With the increased luminance/visibility level, NRTAs will be reduced, while the power consumption will be increased as well which is unexpected.. 23.

(34) 4 Relationship between N/D RTAs ratio and road lighting condition Since it is quite difficult to find the relationship between NRTAs rate and road lighting condition, N/D RTAs ratio is used to instead of NRTAs rate in this paper which has similar performance. However, with different areas and different road condition effects such as road shape, weather condition, local speed limits, etc., the relationship between N/D RTAs ratio and road lighting condition could be changed. In this paper, two most important and general functions to show this relationship will be discussed.. 4.1 Exponential function relationship The first research is the study of the effect of lighting on accidents involved measurements being made on 89 urban traffic routes throughout England and Scotland which has been considered as the most recent, and probably the most comprehensive one. A mobile laboratory was used to collect data on up to a dozen different parameters associated with lighting performance. Measurements were taken continuously over at least 1km of each road site where the lighting installation and road surface were reasonably constant. Reprehensive values of each lighting parameter were obtained for each site and related in turn to the corresponding accident figures accumulated over a period of 3 years. [34] The result of this relationship is shown in Figure 5 which is an exponential function can be seen from the expression.. Figure 5: ‘Best’-fitted relationship between the dark/day accident ratio and the average road-surface luminance L. [34]. Exponential function: dark / day = 0.66 exp( −0.42 L). 24. (4.1).

(35) In this function, the accidents statistic used was ‘dark/day ratio’ which means the ratio between NRTAs and DRTAs. The coefficients could be changed in different cases. The curve shows N/D RTAs ratio is decreasing with the lighting parameter L (average road surface luminance) increasing.. 4.2 Logarithm function relationship Another research is in Australian from Fisher (1967). [35] It has taken the accidents data of Turner which is shown in Figure 6. It has converted original light scale into the one with unit cd/m 2 .. Figure 6: The variation of night to day accidents with light level provided by street lighting, A: S.A.A. Code installations lie within this range on a L.H.L./100 ft road basis, B: S.A.A. Code 2. installations lie within this range on an lm/ft basis [35]. Logarithm function: N / D = 1.53 − 0.236 log( L). (4.2). In this figure, it still uses ‘night/day ratio’ to present accidents rate. The unit of street lighting is changed to other units. Thus, to make it comparable with Figure 5, a unit conversion has been done by using the formula has been introduced in the previous chapter. The results are showing in red color in Figure 6. An example of the calculation is shown as below which is used to calculate first element.. 25.

(36) 0.24 * 0.17 * 3 = 0.1224 cd/m 2. (4.3). Furthermore, according to the expression, it shows that it is a logarithm function relationship here. The coefficients could also be changed in different cases. N/D RTAs ratio is decreasing with the luminance increasing as well.. 26.

(37) 5 Relationship between power consumption and road lighting condition Depend on different lamps or LED (Light Emitting Diode) types and different technologies it has used, the relationship between power consumption and its luminous intensity is different. However, it has a general form of this function which is quadratic function. The type of the LED has been chosen to introduce in this paper is: 466-3582 from Kingbright. The reason of choosing this LED is because it is a high intensity, long life and low power consumption LED which can be use to simulate sunlight resource at night. Furthermore, LEDs technology has been supposed to replace HPS (High-Pressure Sodium) lamps which are common used now in the near future.. 5.1 Forward current and luminous intensity relationship The following figure which was taken from the datasheet of LED 466-3582 showed the relationship between forward current and luminous intensity. [36] It can be seen that luminous intensity is increasing with forward current increasing linearly. This forward current is a strongly relative indicator of the power.. Figure 7: The relationship between forward current and luminous intensity in LED 466-3582 [36]. However, in practical, it has some LEDs have a non-linear relationship between forward current and luminous intensity. The formulas used to express these relationships are quite complicated. Therefore, in this paper, it only considers the LEDs which have a linear relationship between forward current and luminous intensity. It means the model is only suit for the LEDs or the lamps which have a linear or more or less linear unit conversion between forward current and luminous intensity.. 27.

(38) 5.2 Power and luminous intensity relationship In direct current resistive circuits, electrical power is calculated by using Joule’s law: (2) P=V*I. (5.1). where P is the electric power, V is the potential difference, and I is the DC (direct current). If the using current is an AC (alternative current), then it needs to be converted to an equivalent DC by the formula: (2). I DC(eq) =. I AC ( peak ). (5.2). 2. where, I DC(eq) is the equivalent DC value of AC, I AC ( peak ) is the peak value of AC. According to Ohm’s law: (2) I=. V R. (5.3). where, I is the current, V is the potential difference, and R is the electrical resistance. In the case of resistive (linear) loads, Joule’s law can be combined with Ohm’s law in order to produce alternative expressions for the dissipated power: (2) P = I2 * R. (5.4). where, P is the power consumption, I is the current and R is the resistance which can be considered as a constant here. Since luminous intensity is increasing linearly with forward current increasing, luminous intensity has a quadratic relationship with power consumption which is like Figure 8 shows in the next page. It shows a general quadratic function curve in the figure. Because luminous intensity never gets a negative value, the curve only keeps the positive part (contains 0).. 28.

(39) Figure 8: A general relationship between power consumption and luminous intensity. Quadratic function: P = L2 * Z. (5.5). In this derivative quadratic function, P is still the power consumption, L is the luminous intensity, and Z is a coefficient determined by light source. Furthermore, depend on different types of LEDs or lamps, the slope of this curve could be sharper (with bigger Z value) or gentler (with smaller Z value). But the basic function is the same (which is quadratic function).. 29.

(40) 6 General N/D RTAs ratio and power consumption optimal model based on luminous intensity variable 6.1 General structure of the model A general structure of the model is presented by the figure shown as below. It mainly contains three parts which are: the relationship between N/D RTAs ratio and luminous intensity, the relationship between power consumption and luminous intensity, internal influencing factors to both functions. These factors will be analyzed individually by introducing how they affect N/D RTAs ratio or power consumption or both. Since the effect on N/D RTAs ratio has similar result as NRTAs rate, it will use NRTAs rate to evaluate which is more convenient in the following sections.. Figure 9: General structure of the model. In Figure 9, it can be seen that the internal influencing factors have been divided into four categories which are: Human factors, Lantern factors, Traffic factors, and Weather factors. In each category, it also contains several items which will be 30.

(41) described from their influences perspective individually later on.. 6.2 General curves and expressions of the functions Figure 10 shows general curves and expressions of two functions (exponential or logarithm function and quadratic function) which have been developed out by consulting literature data.. Figure 10: General curves of two functions. For these two graphs, they have same variable “L” and unit “cd/m 2 ” along the horizontal-axis, but different variables and units along the vertical-axis. Therefore, to make them comparable with each other, a media variable which has a connection with both N/D RTAs ratio and power consumption is needed in calculation. Economic cost has been chosen to use as a media variable in this paper. Afterwards, it is required to convert N/D RTAs ratio and power consumption to equivalent economic cost respectively. However, these conversions (or the scale in between these units) could be changed in different cities and different road lighting conditions according to the literature study. Thus, it will use a general form in the model instead of giving some exactly values in this paper.. 6.3 Influencing factors of the model 6.3.1 Human factors 6.3.1.1 Physiological response The human being has developed to be a daylight animal, they are not naturally nocturnal in their habits. Thus the human visual system has many drawbacks in night traffic. Such as: vision at night is slower than by day, i.e. the visual signal which stimulates the eye takes longer to reach the brain and lower luminance being viewed. This difference can be up to 0.15s for extremes of luminance level. [37] Visual acuity 31.

(42) or the resolving power of the eye also decreases at night. [38] Although various efforts to aid the visual system have been presented and tested so far (optical filters, glare shield glasses, special adaptation lamps, etc.), they haven’t been successful in improving vision in real night driving situation. [39] Furthermore, human’s physiological response to darkness is natural fatigue. It will reduce visual performance and communicating efficiency which could be one of the reasons causing more accidents occurring at night. Therefore, human’s physiological response has a naturally negative impact on NRTAs rate. 6.3.1.2 Human behavior On the other hand, the hours of darkness are also those of relaxation and pleasure seeking. It means the percentage of drivers who have drunk at nighttime is much more than at daytime. The effect of alcohol will also reduce human visual performance and commuting efficiency a lot. Therefore, it can be concluded that human behavior has a positive effect on NRTAs rate. If a person who has bad driving behavior (such as: drunk, fatigue, etc.), whatever how good road lighting condition it has, he still has a very high possibility to get an accident. Whereas, if a person could concentrate all his attention on driving and follow the traffic rules, then he will has a lower possibility to get an accident even if road lighting condition is not that good. Furthermore, for the vulnerable road users, such as: pedestrians, cyclists, road workers, etc., human behavior could also affect their safety in night traffic. For instance, like Figure 11 shows the examples of reflective clothes and some reflective wrist or ankle ties. They can be all used by these vulnerable people mentioned above. It can protect their safety by enhancing their visibility level at night. Therefore, it indirectly helps on improving the performance of road lighting, and has a positive impact on power consumption.. (a) (b) Figure 11: (a) Reflective clothes (3), (b) Reflective wrist or ankle ties (4) 32.

(43) 6.3.1.3 Driver age As Figure 4 shows in chapter 3, human vision ability will decrease with age increasing. Therefore, to provide these old people a road lighting environment as safe as young people, it needs paying more on cost and power consumption to increase the luminance level of road lighting. It means driver age has a negative effect on road lighting’s power consumption. Definitely, it is not possible to evaluate the proportion of the older drivers on a certain road section. However, the proportion of the older drivers in a city could be known in practical. Therefore, this data could be consulted when setting the luminance level of road lighting in a city. For instance, if a city has higher older drivers proportion, road lighting system may need providing higher luminance. Nevertheless, due to better driving experience and skill, old drivers could have a better way to deal with the emergency or unexpected situations when they are driving on the road. Furthermore, from the perspective of elder psychology, old drives may not drive as fast as young drivers in the night. Thus, it also reduces the possibility of NRTAs occurring on these old drives. Figure 12 shows a relationship between the involvement rate in fatal crashes and drivers’ age. [40] It can be seen that NRTAs rate is much higher than DRTAs rate except for very old drivers (above 70). Except for the age range above 60, NRTAs rate is decreasing with age increasing obviously. Especially for the age range between 16 and 44, the decrease speed is quite fast.. Figure 12: Age and fatality rates by time of day [40]. It can also be argued that the population of people travelling at night could be different from those using the roads by day. It may be that a disproportionate number of young drivers are out at night, because the older have family responsibilities and are less likely to be motivated to go out at night. [6] It has been well established that 33.

(44) young drivers have much higher accident rates than old drivers at night may due to lack of driving experience or poor driving skill, etc. This could be reflected in high NRTAs. From above analysis, it could get a small conclusion: driver age has a negative impact on road lighting’s power consumption, but has a positive impact on NRTAs rate. However, it is quite hard to determine which impact is bigger than the other one without implementing some experiments in real traffic.. 6.3.2 Lantern factors 6.3.2.1 Lantern design The chief determinants of lantern design are the type of lamp and the type of lighting distribution required. Lamp type: The type of lamp mainly contains: HPS (High-Pressure Sodium), LPS (Low-Pressure Sodium), metal halide and mercury vapor. [27] HPS lamps are the most commonly used light source for highway lighting. This is because they have a long operating life (an average is 24,000 hours), and high lamp efficacy. The disadvantage is an orange color rendition and a long restrike time in the event of momentary power interruption. LPS lamps also have a high lamp efficacy, however, they do not have as long an operating life as HPS or mercury vapor lamps. They have a relatively short restrike in the event of momentary power interruption. The cost of installation and maintenance can be a major consideration in product selection. These lamps have a very large arc tube, which makes it difficult to efficiently control light output from the lamp. LPS lighting produces a monochromatic yellow color. Metal Halide lamps have been found only limited use because of their relatively short lamp life. However, these lamps have a very good color rendition, efficacy, good light control because of the small size, and produce a light that is bluish white in color. Mercury Vapor lamps offer an exceptionally long operating life, but have a lumen efficiency only about half that of HPS lamps. Mercury vapor lamps produce a light that is bluish white in color.. 34.

(45) A comparison of the various lamps that are used in roadway lighting is given as below. [41]. Table 16: Source types comparison form [41]. From Table 16, it can be seen that LPS has the highest lumen efficiency, but it also has poor performance on both color rendition and optical control. Metal Halide has the best performance on that, but the average life is the shortest one and low lumen efficiency. The performance and average life of Mercury Vapor are both good, however, it has the worst efficiency and highest operational cost. It seems HPS is the most average one who has balance performance in all the evaluations. It is also the reason why HPS has been widely used in road lighting system. Lighting distribution: The purpose of controlling the distribution of light on the roadway is to limit the high-angle light (angle between light beam and vertical) towards the drivers from glare consideration. The categories of control include CO (Cut-Off), SCO (Semi-Cut-Off), and NCO (Non-Cut-Off). Actually only two general classes of light distribution are widely used in road lighting field which are SCO with a peak intensity at about 75° to the vertical, and CO where the peak is generally around 65° . An illustration of these two type controls is shown as followed. [8]. Figure 13: Cross-sectional schematics of lanterns, a) Cut-off lamp, b) Semi-cut-off lamp [8]. From Figure 13, it can be seen that the bigger angle (between light beam and lantern) it has, the wider area can be covered by light beam, but the heavier influence it will have on glare problem. SCO lanterns are usually used since they permit a reasonable 35.

(46) spacing/mounting height ratio. CO lanterns are generally thought to be more expensive. It has applications on some motorways and trunk roads where the lower level of glare is desirable. On the other hand, CO lanterns can also provide a better uniformity performance comparing to NCO design. [42] Furthermore, the use of CO lanterns is not necessarily more expensive than that of SCO or NCO luminaries. When the cost is greater, the difference is no more than 25%. [43] From above study, it can be seen that different lantern design technologies could impact on both operational cost and the quality of lighting. Better quality often needs paying more on operation. Therefore, lantern design has a positive impact on improving road lighting condition as well as on reducing NRTAs rate, but has a negative impact on cost and power consumption. 6.3.2.2 Mounting height The higher lantern mounting height, the greater beam spread along the road and the smaller will be the number of mounting points required within a certain area. Other pay offs from higher mounting height will enhance road surface luminance uniformity and lower glare. [6] It has a formula could show the relationship between luminance (L) of road surface at any point and mounting height (H) which is shown as below: [8] L = q * EH = q *. I cos 3 γ 2 H. (6.1). where, E H is the horizontal illuminance on the road surface, q is the luminance coefficient of road surface, I is the light intensity directed at the point, and γ is the angle of incidence of the light on the road surface. From Formula (6.1), it can be concluded that mounting height is inversely proportional function to the luminance of road surface. However, it does not mean the mounting height of street light should be as short as possible. Otherwise, the area that could be covered by each single light will be too small, and it will cause uniformity and glare problems. Therefore, a standard or an optimal mounting height is needed for road lighting design. For different countries, they may have different criteria. In the U.K. the standard mounting height for lanterns in trunk route lighting is 10m. [44] The U.S. Bureau of Public roads suggest that heights of 40 to 50 feet would be preferable in terms of economy, effectiveness, safety, aesthetics and flexibility for further modification. [45] To estimate mounting height influencing on power consumption, firstly it has to find a 36.

No results found