Page 193 - AJWEP-22-6

Municipal street lighting retrofit in Bulgaria

min fx = P (II) (i) Use a global optimizer over (P, H, θ).
( )
x
(ii) For each candidate:
2.2.2.4. Constraints (EN 13201 class M ) (a) Set P via luminous flux.
k
Based on the photometric simulation, the design (b) Set H and θ.
must meet the quantitative and qualitative criteria of (c) Run the simulation → extract L , U , U, TI, and EIR.
l
o
ave
street lighting according to EN 13201 for the relevant (d) Select the best feasible point with the lowest P.
class Mk, including average luminance, overall and Practical bounds and starting points used in the
longitudinal uniformity, glare limitation (TI, accounting DIALux evo simulation are:
for veiling luminance Lv), edge illuminance ratio, and (i) H: 4–12 m (road dependent), step 0.2 m.
the geometric limits of the lighting system layout. (ii) θ: 0–20° (per luminaire specification), step 1–2°.
x
(i) Average road luminance: L () ≥ L req (III) (iii) P: From the minimum dimming level to nominal
ave
power (driver-specific steps from Table 4).
L x () The process of modeling and optimization in DIALux
(ii) Overall uniformity: Ux() min U o,req (IV)
o
x
L () evo for street lighting on a given street was carried out
ave
(iii) Longitudinal uniformity: in the following sequence:
(i) Fix geometry, R-table, and pole spacing S.
L x ()
Ux() min,along line U (V) (ii) Choose luminaire optics (fixed photometric file).
l
L x () l,req (iii) Run a coarse grid over (H,θ); for each pair,
max,along line
(iv) Glare (threshold increment [TI]): determine the minimum P that meets L ≥ L .
ave
req
(iv) From the candidate solutions, check uniformities
Lx () and TI; remove non-compliant cases.
v
TI x() 65 TI max (VI)
x
L () . 08 (v) Fine-tune around the best points (local search) to
ad
where L is veiling luminance from glare sources and obtain (P, H, θ) that satisfy all EN 13201 criteria.
v
L ≈L for typical road viewing conditions. In this study, and following the above-described
ad
ave
(v) Edge illuminance ratio (if required): methodology using DIALux evo (Figure 9), a multi-
variant lighting modeling and optimization process
E ()x
EIR () x = outside edge lane ≥ EIR was conducted for all streets with diverse geometrical
E inside edge lane () x req (VII) characteristics in Pavlikeni and Byala Cherkva,
(vi) Geometric limits: involving the optimization of the power rating, mounting
height, and tilt angle of the street LED luminaires.
H H H , , S const (VIII) For park lighting, the project included an LED
min max min max
luminaire with a rated power of 14.4 W, a luminous flux
2.2.2.5. Solving approach of 1,730 lm, and a luminous efficacy of 120.1 lm/W, with
Such lighting optimization tasks can be solved using its light distribution shown in Figure 10. This luminaire
specialized software tools, such as DIALux evo (DIAL is intended to replace the existing park fixtures mounted
GmbH, Germany). This software enables the solution on 4.5-m-high poles in pedestrian zones in Pavlikeni,
of a constrained non-linear optimization problem, classified under lighting class P6.
because the relationship between the decision variables A maintenance factor of MF = 0.81 was used in
(e.g., luminaire power, mounting height, tilt) and the lighting calculations, reflecting luminous flux
the objectives/constraints (e.g., L , U , U , TI, and depreciation to 90% of the initial value over the
ave
l
o
EIR) is non-linear. This problem can be classified as luminaire’s lifetime (L90F10). If a luminaire with greater
black-box optimization, as no analytical formulas are luminous flux depreciation is supplied, its electrical
available for the objectives/constraints (e.g., L , U )— power must be proportionally increased. Conversely, for
o
ave
they are obtained from the simulation software. In the luminaires equipped with drivers that maintain constant
simulation-based solution, the variables are restricted to luminous output over time (constant light output [CLO]),
discrete steps (e.g., height in 0.2 m increments, tilt in the required power may be reduced accordingly.
1° increments), thereby forming a mixed-integer non- The LED street luminaires were mounted on existing
linear programming problem. poles made of tubular steel and reinforced concrete.
The constrained search approach includes the The project included the replacement of existing park
following steps: luminaires with new street-type LED luminaires on

Volume 22 Issue 6 (2025) 187 doi: 10.36922/AJWEP025310242