- Email: [email protected]
Application of fuzzy control method in a tunnel lighting system
Mathematical and Computer Modelling 54 (2011) 931–937
Contents lists available at ScienceDirect
Mathematical and Computer Modelling journal homepage: www.elsevier.com/locate/mcm
Application of fuzzy control method in a tunnel lighting system Chao Yang ∗ , Shijuan Fan, Zhiwei Wang, Wei Li School of Mechatronic Engineering, East China Jiaotong University, Nanchang, 330013, PR China
article
info
Article history: Received 12 August 2010 Accepted 4 November 2010 Keywords: Fuzzy control method Tunnel lighting system Energy-saving Simulation experiment Environment luminance
abstract In order to meet the requirements of driving safety and energy-saving of lighting in tunnel, a fuzzy control method is adopted to design the tunnel lighting control system. A fuzzy control model for the tunnel lighting control system is established with tunnel exterior environment luminance, traffic volume and vehicle speed information as inputs and tunnel interior light luminance as output. Membership functions of environmental parameters and fuzzy control rules are designed based on ‘‘Specifications for the Design of Ventilation and Lighting of Highway Tunnels (China)’’ and experts’ relevant experiences. The tunnel exterior environment luminance fit-curve and the tunnel interior luminance fitcurve are constructed by experimental data with the fuzzy control model on sunny days. Errors between theoretical luminance data and simulation luminance data with the fuzzy control model are less than 5 %. Comparison between luminance data from fit-curves and luminance data from original curves (no light-tuning or class light-tuning) shows that the fuzzy control system has a notable energy-saving effect (saving more than 50 % energy to no light-tuning and more than 20 % to four-steps light-tuning), and nice adaptability. © 2011 Published by Elsevier Ltd
1. Introduction Tunnels are important parts of highway transport, and the operating cost of tunnel traffic is huge. How to reduce operating costs of the tunnels with good safety performance has become a focus issue that the transport department is be concerned with. Tunnel lighting is essential for the tunnel system, its quality directly affects driving safety, it is also the largest energy consumption unit in the tunnel’s engineering [1]. In the mountainous regions of China, most tunnel lighting systems have been at a low level and high energy consumption stage for a long period of time because of the restriction of geographical locations and limit of economic conditions. China has advocated vigorously to build an economic society and it is imperative to improve the quality of tunnel lighting systems [2]. In this paper, a tunnel lighting control system is designed using a fuzzy control method to adjust tunnel interior luminance constantly in accordance with the changes of tunnel exterior environmental luminance, traffic volume and driving speed to meet the demands of tunnel lighting, which ensures driving security in the tunnel and saves the energy of tunnel lighting greatly. 2. Vision problems of tunnel lighting Unlike road lighting, lighting is also needed in tunnels during the daytime, and the lighting problems are more complex than night lighting. Like road lighting, a certain luminance level for pavement, traffic volume, driving speed and other factors needed to be considered in tunnel lighting and the luminance level should be synthetically determined by the driver’s safety and comfort, especially in the tunnel entrance and corresponding sections of the tunnel where human visual adaptation must be considered. According to the CIE (Commission International d’Eclairage) technical report, ‘‘when driver’s approach, enter
∗
Corresponding author. E-mail address: [email protected] (C. Yang).
0895-7177/$ – see front matter © 2011 Published by Elsevier Ltd doi:10.1016/j.mcm.2010.11.018
932
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
and exit the tunnel from a bright environment, those will bring about variety vision problems’’ [3–5]: (1) Vision problems before entering the tunnel (daytime): the car drivers arrive from the external environment with a natural lighting level, which can be very high. Consequently, they look at a tunnel like a black hole in which they are supposed to perceive the presence of dangerous situations, like obstacles, queues or traffic stops. (2) Vision problems after entering the tunnel (daytime): when drivers even enter a not too dark tunnel from a bright environment, it needs an interval before drivers can discern the situation of the tunnel interior, which is called the ‘‘adaptation lagging phenomenon’’, this reason is because the abrupt change of luminance, and the fact that people’s vision adapts slowly. (3) Vision problems of tunnel interior: unlike general roads, the road in the tunnel interior mainly lies in auto exhaust releases which cannot disperse quickly and form smoke. Lighting that car headlights and road lighting emit is absorbed and scattered by the smoke, and then reduces visibility. (4) Vision problems of tunnel exit: during the day, when a vehicle passes through a long tunnel and approaches tunnel exit, because of the high luminance of the tunnel exterior, the tunnel exit appears as a bright hole, and the strong glare will cause great discomfort to drivers. On the contrary, the tunnel exit is not bright hole but a black hole at night, so that alignment and obstructions in the road cannot be discerned. In the special circumstances of tunnels, lighting facilities must be installed in order to ensure traffic safety. However, high energy consumption becomes another problem, lighting energy consumption takes up much of the total energy consumption in the tunnel in accordance with theoretical design, so various effective methods should be adopted to ensure safety and energy-saving during the course of tunnel lighting design. 3. Control methods of tunnel lighting Logic switch style is adopted in most current tunnel lighting control systems in which different permutations and combinations of lighting lamps are used to achieve required light intensity. The strong points of logic switch style are simple programming and circuitry design, flexible lamp selection and easy maintenance, but its continuity and uniformity of luminance is not good enough, and the maximum luminance level is often considered in tunnel lighting design, which causes considerable energy consumption [6]. Another control method of tunnel lighting is stepless control that uses the electronic controller based on an SCR (siliconcontrolled rectifier). In this method, location information of lamps, tunnel exterior environmental luminance, traffic volume and vehicle speeds are given values, tunnel interior luminance is a controlled object, the difference between the given luminance value and tunnel interior luminance value is a control variable to change the conduction angle of SCR. With the tunnel exterior environmental luminance changes, the entire lighting control system is in a dynamic equilibrium state and feasible tunnel interior luminance is obtained [7,8]. The stepless control method can realize continuous adjustment of tunnel lighting and is an undoubted ideal control mode of tunnel lighting with the consideration of energy saving. In order to realize stepless control of tunnel lighting, a mathematical model of tunnel interior luminance must be established based on changes of tunnel exterior environmental luminance, traffic volume and vehicle speeds. Because of the randomness of environmental parameters and the different degrees of light efficiency attenuation along with time for different kinds of lamps, it is very difficult to establish a precise mathematical model. Fuzzy control is an intelligent method to control effectively systems which are not easy to obtain accurate mathematical models and complex nonlinear systems for. The reliability, accuracy and adaptability of tunnel lighting are improved by using fuzzy control theory. Stepless control of tunnel lighting can be realized by a reasonable control program and corresponding hardware device. 4. Fuzzy control scheme of tunnel lighting 4.1. Structure of tunnel lighting control system Fig. 1 is a structure diagram of tunnel lighting control system. The control system is mainly consisted of vehicle detectors, luminance detectors, lighting control computer, data converters, DALI (Digital Addressable Lighting Interface) local controllers, electronic ballasts and electrodeless discharge lamps. Luminance detectors disposed in each section of the tunnel are used to collect luminance information; vehicle detectors and luminance detectors disposed in the tunnel entrance are used to collect traffic volume, vehicle speeds and environmental luminance information. The data collected from the detectors are passed to the lighting control computer, and then the lighting control computer sends control commands to electronic ballasts through the DALI bus after operation of a given fuzzy control logic. Here the electronic ballast and DALI sub-controller are manufactured into one single-entity, each ballast is addressed separately by the DALI master controller, and stepless control of tunnel luminance is achieved based on tunnel exterior luminance, traffic volume and vehicle speeds. 4.2. Fuzzy control model for tunnel lighting The fuzzy control model for tunnel lighting is divided into four layers shown in Fig. 2: the first layer is the input layer of fuzzy control, accepting language variables information including environmental luminance L, traffic volume N, and vehicle speed V ; the second layer is the membership function generation layer, completing the fuzzy operation; the third layer
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
933
Fig. 1. Control structure diagram of tunnel lighting.
Fig. 2. Fuzzy control model of tunnel lighting.
is the fuzzy inference layer, calculating the fitness of each rule; the fourth layer is the output layer, realizing anti-fuzzy ¯ The electronic ballasts are adjusted through the DALI bus based on operations, the output is tunnel interior luminance L. the luminance difference between required tunnel interior luminance from the fuzzy control model and the actual tunnel interior luminance collected in the field. In this way, the required tunnel interior luminance is gained. 4.2.1. Transformation of input/output domains Actual inputs should be converted into the required domain range by means of scale transformation. Supposing x∗0 is the actual input, its transformation range is [x∗min , x∗max ], if the required domain range is [xmin , xmax ], the input x0 is calculated with the formula (1) by a linear transformation. x0 =
xmin + xmax 2
xmax − xmin k= ∗ xmax − x∗min
∗
+ k x0 −
x∗max + x∗min 2
(1) (2)
where x∗min , x∗max is the minimum and maximum of the transformation range respectively; xmin , xmax is the minimum and maximum of required domain range respectively. Tunnel exterior environmental luminance L, as an input of the fuzzy controller, is translated into L∗ by domain transformation and fuzziness. Supposing the actual domain range is [−eL , +eL ], and then its quantificational domain is {−4, −3, −2, −1, 0, 1, 2, 3, 4}. Traffic volume N, as an input of fuzzy controller, is translated into N ∗ by domain transformation and fuzziness, supposing the actual domain range is [−eN , +eN ], then its quantificational domain is {−4, −3, −2, −1, 0, 1, 2, 3, 4}.
934
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
Fig. 3. Membership function of environmental luminance.
Fig. 4. Membership function of traffic volume.
Fig. 5. Membership function of vehicle speed.
Fig. 6. Membership function of interior luminance.
Vehicle speed V , as an input of fuzzy controller, is translated into V ∗ by domain transformation and fuzziness, supposing the actual domain range is [−eV , +eV ], then its quantificational domain is {−4, −3, −2, −1, 0, 1, 2, 3, 4}. ¯ as an output of fuzzy controller, is translated into L¯ ∗ by domain transformation and fuzziness, Tunnel interior luminance L, supposing the actual domain range is [−eL¯ , +eL¯ ], then its quantificational domain is {−4, −3, −2, −1, 0, 1, 2, 3, 4}. 4.2.2. Fuzzy language and membership functions of input/output Vague language of tunnel exterior environmental luminance L is defined as (dark, mild dark, medium, mild light, light), that is {LNB , LNS , LZR , LPS , LPB }, its membership function is shown in Fig. 3. Vague language of traffic volume N is defined as (very small, small, medium, large, larger), that is {NNB , NNS , NZR , NPS , NPB }, its membership function is shown in Fig. 4. Vague language of vehicle speed V is defined as (slow, medium, fast), that is {VNB , VZR , VPB }, its membership function is shown in Fig. 5. Vague language of tunnel interior luminance L¯ is defined as (dark, mild dark, medium, mild light, light), that is ¯ {LNB , L¯ NS , L¯ ZR , L¯ PS , L¯ PB }, its membership function is shown in Fig. 6.
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
935
Table 1 Fuzzy control rules. VNB
VZR
N
L
N
LNB
LNS
LZR
LPS
LPB
NNB NNS NZR NPS NPB
L¯ NB L¯ NB L¯ NS L¯ NS L¯ ZR
L¯ NB L¯ NS L¯ NS L¯ ZR L¯ ZR
L¯ NS L¯ NS L¯ ZR L¯ ZR L¯ PS
L¯ NS L¯ ZR L¯ ZR L¯ PS L¯ PS
L¯ ZR L¯ ZR L¯ PS L¯ PS L¯ PB
If L is LNB If L is LNB If L is LNB If L is LNB ....
and N and N and N and N
NNB NNS NZR NPS NPB
VPB L
N
LNB
LNS
LZR
LPS
LPB
L¯ NB L¯ NS L¯ NS L¯ ZR L¯ ZR
L¯ NS L¯ NS L¯ ZR L¯ ZR L¯ PS
L¯ NS L¯ ZR L¯ ZR L¯ PS L¯ PS
L¯ ZR L¯ ZR L¯ PS L¯ PS L¯ PB
L¯ ZR L¯ PS L¯ PS L¯ PB L¯ PB
NNB NNS NZR NPS NPB
L LNB
LNS
LZR
LPS
LPB
L¯ NS L¯ NS L¯ ZR L¯ ZR L¯ PS
L¯ NS L¯ ZR L¯ ZR L¯ PS L¯ PS
L¯ ZR L¯ ZR L¯ PS L¯ PS L¯ PB
L¯ ZR L¯ PS L¯ PS L¯ PB L¯ PB
L¯ PS L¯ PS L¯ PB L¯ PB L¯ PB
is NNB and V is VNB then L¯ is L¯ NB . is NNS and V is VNB then L¯ is L¯ NB . is NZR and V is VNB then L¯ is L¯ NS . is NPB and V is VNB then L¯ is L¯ ZR .
4.2.3. Fuzzy control rules and fuzzy inference The integrity of control rules must be guaranteed while establishing the fuzzy control rules, that is to guarantee that all input states are covered, and to make corresponding control rules take effect in each input state. In addition, contradictory control rules must be avoided while designing rules [9]. A fuzzy control rule table is established based on ‘‘Specifications for the Design of Ventilation and Lighting of Highway Tunnels’’ [10] and expert’s relevant experience. The fuzzy control rule table is shown in Table 1. The fuzzy controller of the control system is composed of three input variables and one output variable, and the input variables have five linguistic values, five linguistic values and three linguistic values respectively, therefore the control rule number is 75 (5 × 5 × 3). Every vague statement corresponds to one input/output fuzzy relation Ri . Ri = [Lj × NK × Vm ]T1 × L¯ l
(3)
where Lj is the language value of tunnel exterior environmental luminance; NK is the language value of traffic volume; Vm is the language value of vehicle speed; L¯ l is the language value of internal luminance; T1 is the dimensions of matrix [Lj × NK × Vm ]; i = 0, 1, . . . , 74; j = k = l = 0, 1, . . . , 4; m = 0, 1, 2. Fuzzy relationship matrix R is composed of 75 fuzzy relations by amalgamative computation. R=
74
Ri .
(4)
i =0
The tunnel interior luminance L¯ is calculated in formula (5). L¯ = [L × N × V ]T2 × R
(5)
where T2 is the dimensions of matrix [L × N × V ]. 4.2.4. Certainty of output and anti-fuzzy operation The results are judged by the gravity center method.
µL¯ (L¯ i )L¯ i EL¯ = ∑ µL¯ (L¯ i ) ∑ i
(6)
i
where EL¯ is the certainty value of internal luminance; µL¯ is the membership function of internal luminance; L¯ i is the language value of internal luminance. In order to get the actual tunnel interior luminance, an inverse transform is carried out for the domains. Supposing the domain range of EL¯ is [ELmin , ELmax ], the domain range of L¯ is [L¯ min , L¯ max ], it is calculated by a linear transformation: ¯ ¯ L¯ =
k=
L¯ min + L¯ max 2
EL¯ + k EL¯ −
L¯ max − L¯ max ELmax − ELmin ¯ ¯
ELmax + ELmin ¯ ¯ 2
(7)
(8)
where L¯ min , L¯ max is the minimum and maximum of language values of internal luminance respectively; ELmin , ELmax is the ¯ ¯ minimum and maximum of certainty values of internal luminance respectively. When the speed is constant, the fuzzy control surface of tunnel lighting is shown in Fig. 7.
936
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
Fig. 7. Fuzzy control surface of tunnel lighting. Table 2 Theoretical calculation data. L (cd/m2 )
N (vehicles/h)
V (km/h)
Lth
Ltr1
Ltr2
Ltr3
Lin
Lout
4000 3000 2000 4000 3000 2000
700 1000 1500 2000 1500 1000
60 90 80 90 80 70
62 94 63 167 95 45
18.6 28.2 18.9 50.1 28.5 13.5
6.2 9.4 6.3 16.7 9.5 4.5
2.2 3.3 2.2 5.8 3.3 1.6
1.5 3.7 3.2 5.4 3.2 1.9
7.5 18.5 16.0 27.0 16.0 9.5
Lth is the entrance zone luminance; Ltr1 , Ltr2 , Ltr3 are the transition zone luminances respectively; Lin is the internal zone luminance; Lout is the exit zone luminance. Table 3 Simulation data. L (cd/m2 )
N (vehicles/h)
V (km/h)
Lth
Ltr1
Ltr2
Ltr3
Lin
Lout
4000 3000 2000 4000 3000 2000
700 1000 1500 2000 1500 1000
60 70 80 90 80 70
69 102 72 172 104 51
20.7 30.6 21.6 51.6 31.2 15.3
6.9 10.2 7.2 17.2 10.4 5.1
2.4 3.6 2.5 6.0 3.6 1.8
1.9 4.1 3.9 5.7 3.5 2.2
9.5 20.5 19.5 28.5 17.5 11.0
Lth is the entrance zone luminance; Ltr1 , Ltr2 , Ltr3 are the transition zone luminances respectively; Lin is the internal zone luminance; Lout is the exit zone luminance.
5. Simulation In order to estimate the control effect, simulation experiments of tunnel lighting are made with the constructed fuzzy control model in MATLAB. Assuming that the range of tunnel exterior environmental luminance is 0–4000 cd/m2 , the range of traffic volume is 0–2400 vehicles/h, the range of vehicle speed is 0–100 km/h, the range of tunnel interior luminance is 0–200 cd/m2 . With different combinations of tunnel exterior environmental luminance, traffic volume and vehicle speed are inputs of the fuzzy control model. The theoretical data calculated based on ‘‘Specifications for the Design of Ventilation and Lighting of Highway Tunnels’’ are shown in Table 2. The output data of simulation experiments are shown in Table 3. Comparing the date of Tables 2 and 3, the errors of two sets of data are small and the fuzzy control method can reflect the changes of environmental luminance, traffic volume and vehicle speed very well. The energy-saving effect of the fuzzy control system can be evaluated simply by the following method: only considering the change of environmental luminance and neglecting the influence of traffic volume change and vehicle speeds change, supposing traffic volume as 1000 vehicles/h, vehicle speed as 80 km/h, and setting 24 luminance values to construct, the change curve of environmental luminance on sunny days is shown in Fig. 8. With the established fuzzy control model, the change curve of tunnel interior luminance is obtained and shown in Fig. 9. In Figs. 8 and 9, the abscissa axis indicates the time of one day (from 6 AM to 6 PM). For comparison, the tunnel interior luminance curves of the four-step hierarchical control model and the normal control model only considering maximum luminance of tunnel lighting are also shown in Fig. 9. The areas surrounded by the curves and abscissa axis represent relative power consumption of different control methods. It can be seen from Fig. 9 that the energy-saving effect of the fuzzy control system is very obvious: the fuzzy control system can
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
937
Fig. 8. Given environmental luminance of the fuzzy control system.
Fig. 9. Interior luminance of different control models in the same condition.
save more than 50% energy than the system only considering maximum luminance, and can save more than 20% energy than four-step control system. The energy-saving effect will be more considerable if the influences of traffic volume and vehicle speed are taken into account. 6. Conclusions Safety and smoothness of a tunnel’s operation must be ensured, energy-saving is also an important problem which cannot be ignored. In this paper, a fuzzy control model is designed and the fuzzy control method is adopted in the tunnel lighting control system to realize energy-saving by adjusting tunnel interior luminance in real time along with the changes of environmental luminance, traffic volume and vehicle speed. Comparing with traditional control methods of tunnel lighting, the fuzzy control method has a great energy-saving effect in addition to meeting the demands of luminance and security of ‘‘Specifications for the Design of Ventilation and Lighting of Highway Tunnels (China)’’. Acknowledgements This work is supported by Key Laboratory of Ministry of Education for Conveyance and Equipment (East China Jiaotong University). References [1] V.B. Wout, Tunnel lighting practice world-wide, Journal of Lighting Research and Technology 2 (13) (1981) 80–86. [2] Z.Y. Zhang, X.L. Ding, Study on energy-saving of highway tunnel lighting in mountain, in: The 9th Seminar of China Highway Information Management and Technologies, Hefei, Anhui, 2007, pp. 327–330. [3] T. Ishimura, Road tunnel lighting, Journal of the Japan Society of Mechanical Engineers 107 (1030) (2004) 687–689. [4] M.C. Beka, A study on tunnel lighting, Journal of Lighting Design & Application (6) (2005) 10–16. [5] Z.L. Chen, Study on highway tunnel lighting energy-saving, Journal of Light & Lighting 32 (3) (2008) 6–16. [6] L.Y. Guo, B. Lang, Study on lighting control system of the tunnel on expressway, Journal of Microcomputer Information 25 (3–1) (2009) 38–40. [7] X.F. Lv, The application of stepless intelligent control system to LED illuminating brightness in tunnel, Journal of China Lighting 10 (2008) 96–98. [8] F.W. Huang, Y. Chen, Research and realization tunnel illumination energy-saving control system, Journal of Highway Engineering 33 (6) (2008) 111–114. [9] G.Y. Li, Theory and Application of Neural-Fuzzy Control, Publishing House of Electronics Industry, Beijing, 2009, pp. 235–236. [10] JTJ 026.101999, Specifications for Design of Ventilation and Lighting of Highway Tunnel, China Communications Press, Beijing, 2000.
Contents lists available at ScienceDirect
Mathematical and Computer Modelling journal homepage: www.elsevier.com/locate/mcm
Application of fuzzy control method in a tunnel lighting system Chao Yang ∗ , Shijuan Fan, Zhiwei Wang, Wei Li School of Mechatronic Engineering, East China Jiaotong University, Nanchang, 330013, PR China
article
info
Article history: Received 12 August 2010 Accepted 4 November 2010 Keywords: Fuzzy control method Tunnel lighting system Energy-saving Simulation experiment Environment luminance
abstract In order to meet the requirements of driving safety and energy-saving of lighting in tunnel, a fuzzy control method is adopted to design the tunnel lighting control system. A fuzzy control model for the tunnel lighting control system is established with tunnel exterior environment luminance, traffic volume and vehicle speed information as inputs and tunnel interior light luminance as output. Membership functions of environmental parameters and fuzzy control rules are designed based on ‘‘Specifications for the Design of Ventilation and Lighting of Highway Tunnels (China)’’ and experts’ relevant experiences. The tunnel exterior environment luminance fit-curve and the tunnel interior luminance fitcurve are constructed by experimental data with the fuzzy control model on sunny days. Errors between theoretical luminance data and simulation luminance data with the fuzzy control model are less than 5 %. Comparison between luminance data from fit-curves and luminance data from original curves (no light-tuning or class light-tuning) shows that the fuzzy control system has a notable energy-saving effect (saving more than 50 % energy to no light-tuning and more than 20 % to four-steps light-tuning), and nice adaptability. © 2011 Published by Elsevier Ltd
1. Introduction Tunnels are important parts of highway transport, and the operating cost of tunnel traffic is huge. How to reduce operating costs of the tunnels with good safety performance has become a focus issue that the transport department is be concerned with. Tunnel lighting is essential for the tunnel system, its quality directly affects driving safety, it is also the largest energy consumption unit in the tunnel’s engineering [1]. In the mountainous regions of China, most tunnel lighting systems have been at a low level and high energy consumption stage for a long period of time because of the restriction of geographical locations and limit of economic conditions. China has advocated vigorously to build an economic society and it is imperative to improve the quality of tunnel lighting systems [2]. In this paper, a tunnel lighting control system is designed using a fuzzy control method to adjust tunnel interior luminance constantly in accordance with the changes of tunnel exterior environmental luminance, traffic volume and driving speed to meet the demands of tunnel lighting, which ensures driving security in the tunnel and saves the energy of tunnel lighting greatly. 2. Vision problems of tunnel lighting Unlike road lighting, lighting is also needed in tunnels during the daytime, and the lighting problems are more complex than night lighting. Like road lighting, a certain luminance level for pavement, traffic volume, driving speed and other factors needed to be considered in tunnel lighting and the luminance level should be synthetically determined by the driver’s safety and comfort, especially in the tunnel entrance and corresponding sections of the tunnel where human visual adaptation must be considered. According to the CIE (Commission International d’Eclairage) technical report, ‘‘when driver’s approach, enter
∗
Corresponding author. E-mail address: [email protected] (C. Yang).
0895-7177/$ – see front matter © 2011 Published by Elsevier Ltd doi:10.1016/j.mcm.2010.11.018
932
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
and exit the tunnel from a bright environment, those will bring about variety vision problems’’ [3–5]: (1) Vision problems before entering the tunnel (daytime): the car drivers arrive from the external environment with a natural lighting level, which can be very high. Consequently, they look at a tunnel like a black hole in which they are supposed to perceive the presence of dangerous situations, like obstacles, queues or traffic stops. (2) Vision problems after entering the tunnel (daytime): when drivers even enter a not too dark tunnel from a bright environment, it needs an interval before drivers can discern the situation of the tunnel interior, which is called the ‘‘adaptation lagging phenomenon’’, this reason is because the abrupt change of luminance, and the fact that people’s vision adapts slowly. (3) Vision problems of tunnel interior: unlike general roads, the road in the tunnel interior mainly lies in auto exhaust releases which cannot disperse quickly and form smoke. Lighting that car headlights and road lighting emit is absorbed and scattered by the smoke, and then reduces visibility. (4) Vision problems of tunnel exit: during the day, when a vehicle passes through a long tunnel and approaches tunnel exit, because of the high luminance of the tunnel exterior, the tunnel exit appears as a bright hole, and the strong glare will cause great discomfort to drivers. On the contrary, the tunnel exit is not bright hole but a black hole at night, so that alignment and obstructions in the road cannot be discerned. In the special circumstances of tunnels, lighting facilities must be installed in order to ensure traffic safety. However, high energy consumption becomes another problem, lighting energy consumption takes up much of the total energy consumption in the tunnel in accordance with theoretical design, so various effective methods should be adopted to ensure safety and energy-saving during the course of tunnel lighting design. 3. Control methods of tunnel lighting Logic switch style is adopted in most current tunnel lighting control systems in which different permutations and combinations of lighting lamps are used to achieve required light intensity. The strong points of logic switch style are simple programming and circuitry design, flexible lamp selection and easy maintenance, but its continuity and uniformity of luminance is not good enough, and the maximum luminance level is often considered in tunnel lighting design, which causes considerable energy consumption [6]. Another control method of tunnel lighting is stepless control that uses the electronic controller based on an SCR (siliconcontrolled rectifier). In this method, location information of lamps, tunnel exterior environmental luminance, traffic volume and vehicle speeds are given values, tunnel interior luminance is a controlled object, the difference between the given luminance value and tunnel interior luminance value is a control variable to change the conduction angle of SCR. With the tunnel exterior environmental luminance changes, the entire lighting control system is in a dynamic equilibrium state and feasible tunnel interior luminance is obtained [7,8]. The stepless control method can realize continuous adjustment of tunnel lighting and is an undoubted ideal control mode of tunnel lighting with the consideration of energy saving. In order to realize stepless control of tunnel lighting, a mathematical model of tunnel interior luminance must be established based on changes of tunnel exterior environmental luminance, traffic volume and vehicle speeds. Because of the randomness of environmental parameters and the different degrees of light efficiency attenuation along with time for different kinds of lamps, it is very difficult to establish a precise mathematical model. Fuzzy control is an intelligent method to control effectively systems which are not easy to obtain accurate mathematical models and complex nonlinear systems for. The reliability, accuracy and adaptability of tunnel lighting are improved by using fuzzy control theory. Stepless control of tunnel lighting can be realized by a reasonable control program and corresponding hardware device. 4. Fuzzy control scheme of tunnel lighting 4.1. Structure of tunnel lighting control system Fig. 1 is a structure diagram of tunnel lighting control system. The control system is mainly consisted of vehicle detectors, luminance detectors, lighting control computer, data converters, DALI (Digital Addressable Lighting Interface) local controllers, electronic ballasts and electrodeless discharge lamps. Luminance detectors disposed in each section of the tunnel are used to collect luminance information; vehicle detectors and luminance detectors disposed in the tunnel entrance are used to collect traffic volume, vehicle speeds and environmental luminance information. The data collected from the detectors are passed to the lighting control computer, and then the lighting control computer sends control commands to electronic ballasts through the DALI bus after operation of a given fuzzy control logic. Here the electronic ballast and DALI sub-controller are manufactured into one single-entity, each ballast is addressed separately by the DALI master controller, and stepless control of tunnel luminance is achieved based on tunnel exterior luminance, traffic volume and vehicle speeds. 4.2. Fuzzy control model for tunnel lighting The fuzzy control model for tunnel lighting is divided into four layers shown in Fig. 2: the first layer is the input layer of fuzzy control, accepting language variables information including environmental luminance L, traffic volume N, and vehicle speed V ; the second layer is the membership function generation layer, completing the fuzzy operation; the third layer
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
933
Fig. 1. Control structure diagram of tunnel lighting.
Fig. 2. Fuzzy control model of tunnel lighting.
is the fuzzy inference layer, calculating the fitness of each rule; the fourth layer is the output layer, realizing anti-fuzzy ¯ The electronic ballasts are adjusted through the DALI bus based on operations, the output is tunnel interior luminance L. the luminance difference between required tunnel interior luminance from the fuzzy control model and the actual tunnel interior luminance collected in the field. In this way, the required tunnel interior luminance is gained. 4.2.1. Transformation of input/output domains Actual inputs should be converted into the required domain range by means of scale transformation. Supposing x∗0 is the actual input, its transformation range is [x∗min , x∗max ], if the required domain range is [xmin , xmax ], the input x0 is calculated with the formula (1) by a linear transformation. x0 =
xmin + xmax 2
xmax − xmin k= ∗ xmax − x∗min
∗
+ k x0 −
x∗max + x∗min 2
(1) (2)
where x∗min , x∗max is the minimum and maximum of the transformation range respectively; xmin , xmax is the minimum and maximum of required domain range respectively. Tunnel exterior environmental luminance L, as an input of the fuzzy controller, is translated into L∗ by domain transformation and fuzziness. Supposing the actual domain range is [−eL , +eL ], and then its quantificational domain is {−4, −3, −2, −1, 0, 1, 2, 3, 4}. Traffic volume N, as an input of fuzzy controller, is translated into N ∗ by domain transformation and fuzziness, supposing the actual domain range is [−eN , +eN ], then its quantificational domain is {−4, −3, −2, −1, 0, 1, 2, 3, 4}.
934
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
Fig. 3. Membership function of environmental luminance.
Fig. 4. Membership function of traffic volume.
Fig. 5. Membership function of vehicle speed.
Fig. 6. Membership function of interior luminance.
Vehicle speed V , as an input of fuzzy controller, is translated into V ∗ by domain transformation and fuzziness, supposing the actual domain range is [−eV , +eV ], then its quantificational domain is {−4, −3, −2, −1, 0, 1, 2, 3, 4}. ¯ as an output of fuzzy controller, is translated into L¯ ∗ by domain transformation and fuzziness, Tunnel interior luminance L, supposing the actual domain range is [−eL¯ , +eL¯ ], then its quantificational domain is {−4, −3, −2, −1, 0, 1, 2, 3, 4}. 4.2.2. Fuzzy language and membership functions of input/output Vague language of tunnel exterior environmental luminance L is defined as (dark, mild dark, medium, mild light, light), that is {LNB , LNS , LZR , LPS , LPB }, its membership function is shown in Fig. 3. Vague language of traffic volume N is defined as (very small, small, medium, large, larger), that is {NNB , NNS , NZR , NPS , NPB }, its membership function is shown in Fig. 4. Vague language of vehicle speed V is defined as (slow, medium, fast), that is {VNB , VZR , VPB }, its membership function is shown in Fig. 5. Vague language of tunnel interior luminance L¯ is defined as (dark, mild dark, medium, mild light, light), that is ¯ {LNB , L¯ NS , L¯ ZR , L¯ PS , L¯ PB }, its membership function is shown in Fig. 6.
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
935
Table 1 Fuzzy control rules. VNB
VZR
N
L
N
LNB
LNS
LZR
LPS
LPB
NNB NNS NZR NPS NPB
L¯ NB L¯ NB L¯ NS L¯ NS L¯ ZR
L¯ NB L¯ NS L¯ NS L¯ ZR L¯ ZR
L¯ NS L¯ NS L¯ ZR L¯ ZR L¯ PS
L¯ NS L¯ ZR L¯ ZR L¯ PS L¯ PS
L¯ ZR L¯ ZR L¯ PS L¯ PS L¯ PB
If L is LNB If L is LNB If L is LNB If L is LNB ....
and N and N and N and N
NNB NNS NZR NPS NPB
VPB L
N
LNB
LNS
LZR
LPS
LPB
L¯ NB L¯ NS L¯ NS L¯ ZR L¯ ZR
L¯ NS L¯ NS L¯ ZR L¯ ZR L¯ PS
L¯ NS L¯ ZR L¯ ZR L¯ PS L¯ PS
L¯ ZR L¯ ZR L¯ PS L¯ PS L¯ PB
L¯ ZR L¯ PS L¯ PS L¯ PB L¯ PB
NNB NNS NZR NPS NPB
L LNB
LNS
LZR
LPS
LPB
L¯ NS L¯ NS L¯ ZR L¯ ZR L¯ PS
L¯ NS L¯ ZR L¯ ZR L¯ PS L¯ PS
L¯ ZR L¯ ZR L¯ PS L¯ PS L¯ PB
L¯ ZR L¯ PS L¯ PS L¯ PB L¯ PB
L¯ PS L¯ PS L¯ PB L¯ PB L¯ PB
is NNB and V is VNB then L¯ is L¯ NB . is NNS and V is VNB then L¯ is L¯ NB . is NZR and V is VNB then L¯ is L¯ NS . is NPB and V is VNB then L¯ is L¯ ZR .
4.2.3. Fuzzy control rules and fuzzy inference The integrity of control rules must be guaranteed while establishing the fuzzy control rules, that is to guarantee that all input states are covered, and to make corresponding control rules take effect in each input state. In addition, contradictory control rules must be avoided while designing rules [9]. A fuzzy control rule table is established based on ‘‘Specifications for the Design of Ventilation and Lighting of Highway Tunnels’’ [10] and expert’s relevant experience. The fuzzy control rule table is shown in Table 1. The fuzzy controller of the control system is composed of three input variables and one output variable, and the input variables have five linguistic values, five linguistic values and three linguistic values respectively, therefore the control rule number is 75 (5 × 5 × 3). Every vague statement corresponds to one input/output fuzzy relation Ri . Ri = [Lj × NK × Vm ]T1 × L¯ l
(3)
where Lj is the language value of tunnel exterior environmental luminance; NK is the language value of traffic volume; Vm is the language value of vehicle speed; L¯ l is the language value of internal luminance; T1 is the dimensions of matrix [Lj × NK × Vm ]; i = 0, 1, . . . , 74; j = k = l = 0, 1, . . . , 4; m = 0, 1, 2. Fuzzy relationship matrix R is composed of 75 fuzzy relations by amalgamative computation. R=
74
Ri .
(4)
i =0
The tunnel interior luminance L¯ is calculated in formula (5). L¯ = [L × N × V ]T2 × R
(5)
where T2 is the dimensions of matrix [L × N × V ]. 4.2.4. Certainty of output and anti-fuzzy operation The results are judged by the gravity center method.
µL¯ (L¯ i )L¯ i EL¯ = ∑ µL¯ (L¯ i ) ∑ i
(6)
i
where EL¯ is the certainty value of internal luminance; µL¯ is the membership function of internal luminance; L¯ i is the language value of internal luminance. In order to get the actual tunnel interior luminance, an inverse transform is carried out for the domains. Supposing the domain range of EL¯ is [ELmin , ELmax ], the domain range of L¯ is [L¯ min , L¯ max ], it is calculated by a linear transformation: ¯ ¯ L¯ =
k=
L¯ min + L¯ max 2
EL¯ + k EL¯ −
L¯ max − L¯ max ELmax − ELmin ¯ ¯
ELmax + ELmin ¯ ¯ 2
(7)
(8)
where L¯ min , L¯ max is the minimum and maximum of language values of internal luminance respectively; ELmin , ELmax is the ¯ ¯ minimum and maximum of certainty values of internal luminance respectively. When the speed is constant, the fuzzy control surface of tunnel lighting is shown in Fig. 7.
936
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
Fig. 7. Fuzzy control surface of tunnel lighting. Table 2 Theoretical calculation data. L (cd/m2 )
N (vehicles/h)
V (km/h)
Lth
Ltr1
Ltr2
Ltr3
Lin
Lout
4000 3000 2000 4000 3000 2000
700 1000 1500 2000 1500 1000
60 90 80 90 80 70
62 94 63 167 95 45
18.6 28.2 18.9 50.1 28.5 13.5
6.2 9.4 6.3 16.7 9.5 4.5
2.2 3.3 2.2 5.8 3.3 1.6
1.5 3.7 3.2 5.4 3.2 1.9
7.5 18.5 16.0 27.0 16.0 9.5
Lth is the entrance zone luminance; Ltr1 , Ltr2 , Ltr3 are the transition zone luminances respectively; Lin is the internal zone luminance; Lout is the exit zone luminance. Table 3 Simulation data. L (cd/m2 )
N (vehicles/h)
V (km/h)
Lth
Ltr1
Ltr2
Ltr3
Lin
Lout
4000 3000 2000 4000 3000 2000
700 1000 1500 2000 1500 1000
60 70 80 90 80 70
69 102 72 172 104 51
20.7 30.6 21.6 51.6 31.2 15.3
6.9 10.2 7.2 17.2 10.4 5.1
2.4 3.6 2.5 6.0 3.6 1.8
1.9 4.1 3.9 5.7 3.5 2.2
9.5 20.5 19.5 28.5 17.5 11.0
Lth is the entrance zone luminance; Ltr1 , Ltr2 , Ltr3 are the transition zone luminances respectively; Lin is the internal zone luminance; Lout is the exit zone luminance.
5. Simulation In order to estimate the control effect, simulation experiments of tunnel lighting are made with the constructed fuzzy control model in MATLAB. Assuming that the range of tunnel exterior environmental luminance is 0–4000 cd/m2 , the range of traffic volume is 0–2400 vehicles/h, the range of vehicle speed is 0–100 km/h, the range of tunnel interior luminance is 0–200 cd/m2 . With different combinations of tunnel exterior environmental luminance, traffic volume and vehicle speed are inputs of the fuzzy control model. The theoretical data calculated based on ‘‘Specifications for the Design of Ventilation and Lighting of Highway Tunnels’’ are shown in Table 2. The output data of simulation experiments are shown in Table 3. Comparing the date of Tables 2 and 3, the errors of two sets of data are small and the fuzzy control method can reflect the changes of environmental luminance, traffic volume and vehicle speed very well. The energy-saving effect of the fuzzy control system can be evaluated simply by the following method: only considering the change of environmental luminance and neglecting the influence of traffic volume change and vehicle speeds change, supposing traffic volume as 1000 vehicles/h, vehicle speed as 80 km/h, and setting 24 luminance values to construct, the change curve of environmental luminance on sunny days is shown in Fig. 8. With the established fuzzy control model, the change curve of tunnel interior luminance is obtained and shown in Fig. 9. In Figs. 8 and 9, the abscissa axis indicates the time of one day (from 6 AM to 6 PM). For comparison, the tunnel interior luminance curves of the four-step hierarchical control model and the normal control model only considering maximum luminance of tunnel lighting are also shown in Fig. 9. The areas surrounded by the curves and abscissa axis represent relative power consumption of different control methods. It can be seen from Fig. 9 that the energy-saving effect of the fuzzy control system is very obvious: the fuzzy control system can
C. Yang et al. / Mathematical and Computer Modelling 54 (2011) 931–937
937
Fig. 8. Given environmental luminance of the fuzzy control system.
Fig. 9. Interior luminance of different control models in the same condition.
save more than 50% energy than the system only considering maximum luminance, and can save more than 20% energy than four-step control system. The energy-saving effect will be more considerable if the influences of traffic volume and vehicle speed are taken into account. 6. Conclusions Safety and smoothness of a tunnel’s operation must be ensured, energy-saving is also an important problem which cannot be ignored. In this paper, a fuzzy control model is designed and the fuzzy control method is adopted in the tunnel lighting control system to realize energy-saving by adjusting tunnel interior luminance in real time along with the changes of environmental luminance, traffic volume and vehicle speed. Comparing with traditional control methods of tunnel lighting, the fuzzy control method has a great energy-saving effect in addition to meeting the demands of luminance and security of ‘‘Specifications for the Design of Ventilation and Lighting of Highway Tunnels (China)’’. Acknowledgements This work is supported by Key Laboratory of Ministry of Education for Conveyance and Equipment (East China Jiaotong University). References [1] V.B. Wout, Tunnel lighting practice world-wide, Journal of Lighting Research and Technology 2 (13) (1981) 80–86. [2] Z.Y. Zhang, X.L. Ding, Study on energy-saving of highway tunnel lighting in mountain, in: The 9th Seminar of China Highway Information Management and Technologies, Hefei, Anhui, 2007, pp. 327–330. [3] T. Ishimura, Road tunnel lighting, Journal of the Japan Society of Mechanical Engineers 107 (1030) (2004) 687–689. [4] M.C. Beka, A study on tunnel lighting, Journal of Lighting Design & Application (6) (2005) 10–16. [5] Z.L. Chen, Study on highway tunnel lighting energy-saving, Journal of Light & Lighting 32 (3) (2008) 6–16. [6] L.Y. Guo, B. Lang, Study on lighting control system of the tunnel on expressway, Journal of Microcomputer Information 25 (3–1) (2009) 38–40. [7] X.F. Lv, The application of stepless intelligent control system to LED illuminating brightness in tunnel, Journal of China Lighting 10 (2008) 96–98. [8] F.W. Huang, Y. Chen, Research and realization tunnel illumination energy-saving control system, Journal of Highway Engineering 33 (6) (2008) 111–114. [9] G.Y. Li, Theory and Application of Neural-Fuzzy Control, Publishing House of Electronics Industry, Beijing, 2009, pp. 235–236. [10] JTJ 026.101999, Specifications for Design of Ventilation and Lighting of Highway Tunnel, China Communications Press, Beijing, 2000.