J. Cent. South Univ. (2018) 25: 2040-2048

DOI: https://doi.org/10.1007/s11771-018-3893-6

Study of road tunnel threshold zone lighting reduction coefficient

DU Feng(杜峰)^{1, 2}, WENG Ji(翁季)¹, HU Ying-kui(胡英奎)³, CAI Xian-yun(蔡贤云)¹

1. Faculty of Architecture and Urban Planning, Chongqing University, Chongqing 400045, China;

2. College of Architecture and Urban Planning, Fujian University of Technology, Fuzhou 350118, China;

3. Periodic Press, Chongqing University, Chongqing 400045, China

Central South University Press and Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract: The luminance in the road tunnel threshold zone attracts broad attention due to its enormous energy consumption and direct influence on tunnel transportation security. Current lighting design methods in threshold zones mostly adopt the reduction coefficient method. However, the determination of reduction coefficient k simply considers tunnel design speed and flow rate, while excluding outside tunnel luminance and threshold zone color temperature and luminance, which have a major impact on driver visual adaptation. Existing problems in the determination of k value are analyzed; a visual performance experiment is utilized; and the reaction time of drivers in changeable outside tunnel luminance and threshold zone color temperature and luminance conditions is obtained; thus, the equations concerning reduction coefficient variation law are derived. In the end, a comparative analysis is made of the k values of the reduction coefficient stipulated by various norms under different color temperature conditions.

Key words: tunnel lighting; visual performance; threshold zone; reduction coefficient; reaction time

Cite this article as: DU Feng, WENG Ji, HU Ying-kui, CAI Xian-yun. Study of road tunnel threshold zone lighting reduction coefficient [J]. Journal of Central South University, 2018, 25(9): 2040–2048. DOI: https://doi.org/10.1007/ s11771-018-3893-6.

1 Introduction

Drivers will encounter obvious dark adaptation visual hysteresis phenomena while entering a tunnel, particularly in the threshold zone with the largest luminance changes. As a result, severe traffic accidents can easily occur in road tunnel threshold zones [1, 2]. Intelligent lighting [3, 4] is a development direction of road tunnel lighting that gradually applies new-type and stepless adjustable lighting sources. In this context, the study of dynamic change laws on tunnel threshold zone luminance (L_th) and adaptation luminance (L₂₀) could effectively overcome “invalid lighting” and “excessive lighting” phenomena rapidly and accurately, while dynamically solving people’s lighting requirements and reaching high uniformity of security, comfort, maximal energy conservation, and personalized lighting in road tunnels.

L_th is the benchmark of tunnel lighting design. According to commission internationale de l'eclairage (CIE) [5, 6] and specifications in each country [7–12], there are three main methods that can be used to determine L_th, namely the reduction coefficient (k value) method, subjective rating number (SRN) subjective assessment method, and perceived correlation method. Among others, the reduction coefficient method is the most popular method in practical application. However, there remain numerous disputes concerning the value of this method. BLASER et al [13] found a new method to gain L₂₀ based on daylight and mean sunshine time distribution. ADRIAN [14] found that there is poor correlation between L₂₀ and equivalent veiling luminance (L_seq) and that the required luminance level in the entrance zone is influenced by the size of the object and contrast [15]. The research of BLASER et al and ADRIAN provided the base for the development of this study, but they lacked an analysis of the relationship between L₂₀ and L_th. YIN et al [16] compared the norms to determine the calculation method of L_th. HU et al [17, 18] determined multiple methods concerning the determination of L_th based on equivalent veiling luminance and visual adaptation. NARISADA et al [19] pointed out that the difference with Schreuder in k value is caused by the difference of observation point distance. In addition, the determination of lighting parameter L_th in the threshold zone refers to a static research thought. After briefly considering L₂₀, traffic flow and average vehicle speed, the current specifications design the fixed value of luminance in the threshold zone. In fact, L₂₀ will constantly change together with season and time. Moreover, the luminance and color temperature of the light source in the threshold zone is different. In other words, it deserves in-depth exploration to discover whether there is a traditional linear relationship between L_th and L_20.

In terms of research methods, NAKAMICHI et al [20], and KABAYAMA [21] put forward the visual performance method and utilize reaction time data to evaluate visual performance based on adjustable and energy-efficient LED light sources in previous studies. ZHANG et al [22] and LI [23]. study the impact of reaction time on visual performance in road lighting. HU et al [24] and DU et al [25] take the pupil change of the driver as an indicator to observe road tunnel lighting. Previous studies have proved the feasibility of the visual performance method in road tunnel lighting research. The authors use the visual performance method and take the reaction time as an assessment indicator to study the relationship of L_th and L₂₀ on the premise of guaranteeing vehicle security. At the same time, the dynamic reduction coefficient of different color temperature LED light sources in the threshold zone in case of any changes of external environmental conditions is also determined.

2 Visual performance experiment

2.1 Experimental principle

This thesis simulates the L_th and L₂₀ road tunnel lighting experimental devices, represents road tunnel lighting scenes, changes light environment parameters, simulates the visual adaptation process of drivers in the access zone and threshold zone, and obtains the reaction time of different lighting parameters to build a correlation model between L_th and L₂₀ and derive the change rule of reduction coefficient k in different lighting environments. The devices in the road tunnel lighting experiments contain an L₂₀ simulation light box, visual performance test apparatus, reaction time test apparatus, BM-5A luminance meter, and CS-2000 spectral radiation luminance meter as shown in Figures 1 and 2.

In accordance with the principle that shorter reaction time will result in higher and more secure visual performance, the relationship between the L_th and L₂₀, equations and relevant rules can be discovered. The study selected 13 research respondents, with seven males (54%) and six females (46%), aged between 20 and 30 years. The naked vision of research respondents was above 0.8 without color blindness or color weakness. There were 50 groups of data altogether in the experiment. See the data in Tables 1–3.

2.2 Experimental procedure

1) The research respondents did the experiment orderly. The respondent was required to focus on an L₂₀ simulation light box (the luminance of the light box should be set as a random value between 2500–5500 cd/m² in advance, such as 3200, 4500 or 5340 cd/m²).

Figure 1 Experimental devices:

Figure 2 Plan and elevation of road tunnel experimental device

Table 1 Data of L₂₀ and L_th with light source of 2829 K LED

2) After opening the road tunnel visual performance test devices for a certain adaptation time (20–30 s), the light source of the L₂₀ simulation light box was cut off. In the meantime, the light source of the simulation illuminance in the threshold zone was opened. The L_th in the threshold zone was successively set as the constant value of 5, 10, 20, 40, 60, 80, 100, 120, 140, 160, 180 and 200 cd/m². There were three kinds of light source color temperature, including 2829, 3814 and 5257 K which belonged respectively to warm (low), medium, and cool (high) color temperatures commonly used in current road tunnel lighting. Simultaneously, the sighting mark placed in the visual performance test apparatus popped up the sighting target (small target). To ensure the objectivity of the targets, the sighting target (size 0.01 m×0.01 m) randomly appeared in three places with the angles (–10°, 0°, 10°). Upon realizing the connection between the sighting mark and the reaction time test apparatus, the device began to recon by time.

Table 2 Data of L₂₀ and L_th with light source of 3814 K LED

Table 3 Data between L₂₀ and L_th with light source of 5257 K LED

3) The research respondent was supposed to press the reaction time button after seeing the sighting mark (small target) the first time. The time difference between releasing the sighting target to pressing the button was the reaction time.

2.3 Data processing

The individual differences among research respondents (such as age, gender, emotions, and physical condition) has a great influence on reaction time data (for instance, the reaction time of some research respondents was longer than 500 ms, while some was shorter than 400ms). This implies that there is no comparability in data among research respondents. Accordingly, this thesis simply simulates the minimum value of the reaction time when L_th changes and other experimental parameters remain unchanged. As shown in Figure 3, if the simulated outside tunnel color temperature was 2829 K and L₂₀ was 5092 cd/m², the reaction time data changing conditions of the research respondent could be discovered when L_th turned from 5 to 200 cd/m².

Figure 3 Reaction time of a research respondent with L₂₀ of 5092 cd/m² (color temperature: 2829 K)

As shown in Figure 3, the relationship between reaction time and L_th conforms well to the form of a parabola. The minimum value of the reaction time was 377 ms and the corresponding L_th was 140 cd/m². This implies that on the premise that L₂₀ is 5092 cd/m² and the LED light source color temperature in the threshold zone is 2829 K, then the most secure L_th should be 140 cd/m². The corresponding most secure L_th value in different L₂₀ and color temperature LED light source conditions could also be measured with this analytical method.

3 Analysis of experimental results

3.1 Experimental data with color temperature of 2829 K LED

When the simulated LED light source in the tunnel threshold zone had the color temperature of 2829 K LED, there were 17 L_th data groups corresponding to different L₂₀ values among the 13 research respondents. The resulting basic data are shown in Table 1 and Figure 4.

Equation (1) is derived according to fitting results.

Figure 4 Relationship graph between L₂₀ and L_th with light source of 2829 K LED

         (1)

Data distribution conforms to the parabola relationship. In this condition, the R² value, namely the multiple correlation coefficient, is 0.7788. This indicates favorable fitting conditions.

The analysis of the fitting graph is as follows:

First, when the color temperature LED of the light source in the threshold zone is 2829 K, if L₂₀ ranges between 2500 and 4800 cd/m², then L_th will present increment tendency together with an increase of L_20. If L₂₀. ranges between 4800 and 5500 cd/m², L_th will present decrement tendency. The relationship between the two is a parabolic relationship.

The second concern is deviation analysis. Through the analysis of points with relatively large deviation, such as the analysis of representative points (L₂₀: 3238 cd/m², L_th: 80 cd/m², L₂₀:4260 cd/m²; L_th: 180 cd/m²), the conclusions could be reached that the mental state (impatience) of research respondents had certain impacts on reaction time data. Considering the high precision of the experimental equipment, there were only a few deviations. If the deviation was within the allowed range, this study did not remove the data.

3.2 Experimental data with color temperature of 3814 K

When the simulated color temperature of the LED light source in the tunnel threshold zone was 3814 K, there were 16 groups of data as shown in Table 2 and Figure 5.

Equation (2) is derived according to fitting results.

Figure 5 Relationship graph between L₂₀ and L_th with LED light source of 3814 K

                    (2)

The data distribution conforms to a linear relationship. In this condition, the R² value, namely the multiple correlation coefficient value, is 0.754. This indicates favorable fitting conditions.

The analysis of the fitting graph is as follows:

First, when the color temperature LED of the light source in the threshold zone is 3814 K, if there is a positive correlation between L₂₀ and L_th, then L_th will present increment tendency together with an increase of L_20. The relationship between the two obeys a linear relationship.

The second concern is deviation analysis. Through the analysis of points with relatively large deviation, such as the analysis of representative points (L₂₀: 3600 cd/m², L_th: 100 cd/m², L₂₀:4260 cd/m², L_th: 120 cd/m²), the conclusions could be reached that the mental state (nervousness) of research respondents had certain impacts on reaction time data. Considering the high precision of the experimental equipment, there were only a few deviations and such deviations were within the allowed range.

3.3 Experimental data with color temperature of 5257 K

When the simulated color temperature of the LED light source in the tunnel threshold zone was 5257 K, there were 17 groups of data as shown in Table 3 and Figure 6.

Equation (3) is derived according to fitting results.

          (3)

Figure 6 Relationship between L₂₀ and L_th in LED light source (5257 K)

The data distribution conforms to a parabolic relationship. In this condition, the R² value, namely the multiple correlation coefficient value, is 0.7529. This also indicates favorable fitting conditions.

The analysis of the fitting graph is as follows:

First, when the color temperature LED of the light source in the threshold zone is 5257 K, if there is a positive correlation between L₂₀ and L_th, then L_th will present increment tendency together with an increase of L_20. The relationship between the two obeys a parabolic relationship.

The second concern is deviation analysis. Through the analysis of points with relatively large deviation, such as the analysis of representative points (L₂₀: 4950 cd/m², L_th: 120 cd/m² and L₂₀: 5180 cd/m², L_th: 200 cd/m²), the conclusions could be reached that the mental state (anxiety) of research respondents had certain impacts on reaction time data. Considering the high precision of the experimental equipment, there were only a few deviations and such deviations were within the allowed range.

4 Discussion

4.1 Comparison of reduction coefficient in different color temperatures

In order to have a comparative analysis, this thesis compares the reduction coefficient graphs in three color temperature conditions, as shown in Figure 7. There are five conclusions: First, among the three types of reduction coefficient relationships between L₂₀ and L_th, they are not all linear; there are also parabolic relationships.

Figure 7 Reduction coefficient comparison in different color temperature LED conditions

Second, when the LED color temperature is 2829 K, then the corresponding reduction coefficient curvity is maximal. If L₂₀ ranges between 3400 and 4700 cd/m², L_th value will be determined in the case of medium and high color temperature conditions. The result is the opposite in the case of low and high lighting conditions. For better accuracy, the relationship among the curves should be based on Eqs. (1)–(3) (similarly hereafter).

Third, when LED color temperature is 3814 K, the reduction coefficient will encounter linear variation. This is different from the rule of 2829 K LED. When L₂₀ ranges between 3600 and 4800 cd/m², the L_th will be smaller than that of L_th in the case of low and high color temperature conditions. It will be the opposite in the case of low and high lighting conditions.

Fourth, when reduction coefficient curvity is medium and LED color temperature is 5257 K, then L_th value will be relatively smaller in any lighting conditions. However, when L₂₀ ranges between 3600 and 4600 cd/m², the value of L₂₀ will be higher than that in the case of 3814 K LED color temperature. When L₂₀ is above 4600 cd/m², the value of L_th will be relatively higher than that in the case of 5257 K LED color temperature.

Fifth, pursuant to the comparison of the reduction coefficient in different color temperature conditions, the minimum value is 2829 K (low color temperature) in general, from the perspective of energy efficiency, while the reduction coefficients in 3814 K (medium color temperature) and 5257 K (high color temperature) conditions have unique advantages. Above all, 5257 K (high color temperature) has optimal energy efficiency.

4.2 Data contrast

In order to further compare with the current k value method, the thesis presents CIE and k value norms in China in Table 4.

Table 4 Road tunnel threshold zone luminance reduction coefficient (k)

Supposing a 80 km/h travel speed, 100 m parking visual range, and symmetrical lighting conditions, k value will be set as 0.060 according to the CIE88-1990 guidebook. [5] Under the same conditions, k value is 0.035 in the Road Tunnel Lighting Design Specification in China 2014 [12].

The relationship between L₂₀ and L_th could be made according to the current k value method in Table 5. Among others, the value of L₂₀ is determined with predetermined norms.

Table 5 Calculation results made with current k value method (k=0.060 and 0.035)

After comparing the reduction coefficient graph reflected by LED in different color temperature conditions and the graph calculated by the traditional k value method, Figure 8 could be derived.

As indicated in Figure 8, there are two points to be mentioned. First, as for the LED light source in any norm, the curve is always subject to the k value method straight line prescribed in CIE1990 norms; moreover, the difference is rather prominent. This implies that in terms of the CIE 1990 guidebook specifications, the L_th value is much lower in identical L₂₀ conditions. In other words,CIE sets higher values on parameters. It is also related with the preference of the high-voltage sodium light in the guidebook. Together with the development of the tunnel light source, in particular an LED light source, lighting efficiency has been greatly elevated and energy resources have been greatly saved. Second, in comparison with the road tunnel lighting design rules in China, the conclusions could be reached that there exists a crossing relationship between the relation curve of the light source in different color temperature LED conditions and the direct line made by the current k value method. In the case of 2829 K color temperature, the LED light source will meet the straight line. In the case of 3814 K color temperature, the LED light source is higher than the current k value method straight line in low lighting conditions but lower than the current k value method straight line in high lighting conditions. In the case of 5257 K color temperature, the LED light source is basically lower than the current k value method straight line in all conditions. This fully reveals the high illuminance and energy efficient properties of medium and high color temperature LED light sources.

Figure 8 Comparison graph of reduction coefficient and current k value in different color temperature LED

4.3 Mechanism analysis

There are three main causes that lead to the variation rule of three curves.

First, k value prescribed by the k value method in current norms is linear. It is similar to the rule of the reduction coefficient in the case of medium color temperature such as 3814 K. In the past, the main road tunnel light source has usually adopted high-voltage sodium light; moreover, the color temperature of such light often ranges between 2500 and 4000 K.

Second, people tend to feel uncertain and uneasy in the case of low color temperature and high lighting conditions. In such conditions, luminance (for example, L_th) is also rather low. This conforms well to the fact that L_th is above 160 cd/m² in the case of 2829 K color temperature.

Third, people tend to generate depressive feelings and uneasiness in high color temperature and low luminance. In such conditions, luminance (for example, L_th) is also rather low. This conforms well to the fact that L_th is below 80 cd/m² in the case of 5257 K color temperature.

5 Conclusions

Aimed at the problems in lighting specifications from the road tunnel access zone to the threshold zone, this thesis adopts a simulated road tunnel lighting experimental device in a laboratory experiment. By using the visual performance method with reaction time as the instrument, a correlation test between L₂₀ and L_th was subsequently conducted to derive the dynamic reduction coefficient value in various color temperature conditions. The main research conclusions are presented as follows:

1) The relation equation between L₂₀ and L_th in different color temperature conditions, namely the reduction coefficient rule, is not generally linear. This means that there still exists a slight difference in various color temperature conditions. The relevant equations contain (1)–(3).

2) According to the comparison of the reduction coefficient rule graphs in different color temperature LED light source conditions, the overall luminance performance in the case of 2829 K color temperature was relatively low. In the case of 3814 K color temperature, the change rule was linear, but the corresponding slope was below the standards of road tunnel lighting design specifications in China. In the case of 5257 K color temperature, LED lighting was maximal and most energy efficient.

3) Throughout the comparative analysis of the reduction coefficient curve in different color temperature LED conditions and straight line in CIE1990 road tunnel lighting norms of China, it is obvious that in LED of various color temperature conditions, the L₂₀ value is lower than L_th in the CIE 1990 norms but mutually intersected in the road tunnel lighting norms of China.

4) In light of the 80 km/h speed limit conditions in road tunnels, this thesis will further observe the conditions in 40, 60, 100 and 120 km/h speed limit conditions and continue to experiment with various driving scenes. The research conclusions in this thesis require the support of on-site experimentation.

References

[1] JIAN Sheng-yeung, YIIK Diew-wong. The effect of road tunnel environment on car following behavior [J]. Accident Analysis and Prevention, 2014, 70: 100–109.

[2] KATJA K, CHRISTER A. The impact of tunnel design and lighting on the performance of attentive and visually distracted drivers [J]. Accident Analysis and Prevention, 2012, 47: 153– 161.

[3] YANG Cui, WANG Shao-fei, HU Guo-hui, ZHAO Yi, YU Bo, SHI Hang. Intelligent lighting system of highway tunnel based on IOT [J]. Highway, 2015 (5): 153–157. (in Chinese)

[4] WENG Ji, CHEN Zhong-lin. Research of highway tunnel intelligent lighting [J]. Light and Lighting, 2015, 39(4): 1–4. (in Chinese)

[5] CIE 88-1990. Guide for the lighting of road tunnels and underpasses [S].

[6] CIE 88-2004. Guide for the lighting of road tunnels and underpasses [S].

[7] CR14380:2003. Lighting applications—Tunnel lighting [S].

[8] ANSI/IESNA RP-22-05. American national standard practice for tunnel lighting [S].

[9] BS 5489-2:2003+A1:2008. Code of practice for the design of road lighting—Part 2: Lighting of tunnels [S].

[10] JIS Z 9116. Lighting of tunnels for motorized traffic [S].

[11] JTJ 026.1-1999. Specification for design of ventilation and lighting of highway tunnel [S].

[12] JTG/T D70/2-01-2014. Guidelines for design of highway tunnels [S].

[13] BLASER P, DUDLI H. Tunnel lighting: Method of calculating luminance of access zone L₂₀ [J]. Lighting Research and Technology, 1993, 25(1): 25–30.

[14] ADRIAN W K. Adaptation luminance when approaching a tunnel in daytime [J].Lighting Research & Technology, 1987, 19(3): 73–79.

[15] ADRIAN W. Investigations on the required luminance in tunnel entrances [J]. Lighting Research & Technology, 1980, 14(3): 151–159.

[16] YIN Ying, CHEN Zhong-lin, CUI Lu-lu. Comparative study on the computing method of luminance of tunnel threshold zone [J]. Light and Lighting, 2008, 32(2): 20–24. (in Chinese)

[17] HU Ying-kui, CHEN Zhong-lin, SUN Chun-hong. Calculation method of road tunnel threshold zone luminance based on equivalent veiling luminance [J]. Journal of Highway and Transportation Research and Development, 2011, 28(5): 98–101, 120. (in Chinese)

[18] HU Ying-kui, CHEN Zhong-lin, ZHANG Qing-wen, WENG Ji, HUANG Ke, LIN Yong. Tunnel threshold zone luminance determined method based on drivers’ visual adaptation [J]. Journal of Civil, Architectural & Environmental Engineering, 2016, 38(4): 20–26. (in Chinese)

[19] NARISADA K, YOSHIKAWA K. Tunnel entrance lighting-effect of fixation point and other factors on the determination of requirements [J]. Lighting Research and Technology, 1974, 6(1): 9–18.

[20] NAKAMICHI F, NARISADA K. Experiment on the Visibility of the tunnel lighting [J]. Journal of the Illuminating Engineering Institute of Japan, 1967, 47(9): 4–12.

[21] KABAYAMA H. Study on adaptive illumination for sudden change of brightness [J]. Journal of the Illuminating Engineering Institute of Japan, 1967, 47(10): 9–12.

[22] ZHANG Qing-wen, LI Yi, WENG Ji, HU Ying-kui. The applicability of different color temperature LED light sources in road lighting [J]. China Illuminating Engineering Journal，2013, 24(5): 70–77. (in Chinese)

[23] LI Yi. Study on the influence of road lighting to traffic safety based on photobiological effect [D]. Chongqing: Chongqing University, 2009. (in Chinese)

[24] HU Ying-kui WENG Ji, ZHANG Qing-wen CHEN Zhong-lin. Calculation method of road tunnel threshold zone luminance based on drivers’ pupil size change rules [J]. Light and Lighting, 2011, 35(4): 10–13. (in Chinese)

[25] DU Zhi-gang, PAN Xiao-dong, GUO Xue-bin. Experimental studies of visual adaptation on driving through freeway tunnel’s entrance and exit [J]. Journal of Harbin Institute of Technology, 2007, 39(12): 1998–2001. (in Chinese)

(Edited by FANG Jing-hua)

中文导读

公路隧道入口段亮度折减系数研究

摘要：公路隧道入口段照明因能耗巨大且直接影响隧道交通安全而备受关注。现行入口段亮度设计方法多利用折减系数法，在确定折减系数k时仅考虑了隧道设计车速和车流量，对驾驶员视觉适应有较大影响的隧道洞外亮度及入口段的色温和亮度未被考虑。文章分析了现行k值确定过程中存在的问题，利用视觉功效实验获得了驾驶员在隧道洞外亮度不断变化且隧道入口段色温和亮度不同条件下的反应时间，进而得到折减系数变化规律的关系式，最后将不同色温下的折减系数与各规范规定的k值进行了对比分析。

关键词：隧道照明；视觉功效；入口段；折减系数；反应时间

Foundation item: Project(51278507) supported by the National Natural Science Foundation of China; Project(cstc2017jcyjAX0346) supported by Chongqing Association for Science and Technology, China

Received date: 2017-03-28; Accepted date: 2017-11-07

Corresponding author: WENG Ji, PhD, Professor; Tel: +86–13608302496; E-mail: wengji0403@163.com; ORCID: 0000-0001-7807- 8965