LABORATOIRE CENTRAL DES PONTS ET CHAUSSEES

DIVISION EXPLOITATION, SIGNALISATION, ECLAIRAGE (ESE)

58, boulevard Lefebvre
75732 PARIS cedex 15, FRANCE

OPTIMISED DESIGN OF WEIGH-IN-MOTION MULTIPLE-SENSORS ARRAY

BY AN ENERGY APPROACH
Mohamed BOUTELDJA Bernard JACOB Victor DOLCEMASCOLO
[email protected] [email protected] [email protected]

INTRODUCTION / OBJECTIVES
Piezoceramic Sampling of the
TRUCKS OVERLOADING CONTROL sensors of the impact force
„ Fair competition between transport companies or transport type, MS-WIM array
„ Infrastructure preservation,
„ Road safety.
PRINCIPE OF WEIGH-IN-MOTION MULTIPLE-SENSORS
„ Vehicle/road dynamics interaction: impact forces variation (in function of: speed, acceleration, road profile, …),
„ Sampling impact force and static weight estimation by various algorithms.
OBJECTIVES: Accurate in motion measurements of axles weights for enforcement purposes
„ Multiple Sensor - WIM systems Æ automatic enforcement,
Figure 1: Principle of weight-in-motion multiple-sensors
„ Accuracy improvement: Sensor’s quality, Array design, Static weight estimation algorithms.

PRINCIPLES OF A MS-WIM ARRAY DESIGN METHOD

∂J
9 Optimization the sensors location and spacing by considering.
∂t
- The vehicle vertical dynamics in interaction with the infrastructure (road profile and slope)
X3 X2 X1 times
- The driver behaviour (vehicle acceleration or deceleration) θ
x
Road (u)
This method takes into account :
k3 C3k2 C2
k1 C1 Sensor
Driver data Reel System Energy criterion of
- the infrastructure characteristics, the selected zone
- the traffic characteristics, Traffic and road data
Zr Yr Xr
Traffic data Fd is

- the interaction vehicle/road, Model 1

- the trucks model. Road profile

The sensor location is so that the total Model n Array location; Figure 2: Bump effect (localised road profile default) on the truck vertical dynamics
variation of the vehicle energy due to the Deflection distance spacing
Models of the literature
pulses is minimal (Figure 3 represents the
algorithm’s scheme).
Figure 3: MS-WIM array optimization algorithm

SIMULATION RESULTS OF MS-WIM ARRAY OPTIMIZATION

Evaluation along the road profile of the RN4 WIM site of the total energy of the vehicle, infrastructure and Mass (m) x

driver system by simulation:

F = Fs + Fd
- First step, the optimal array location with respect to the road profile was studied. C
k Sensor’s
- Second step, various sensor spacings were calculated for an array of six sensors.
STEP 1: Definition of the optimal section ⎛ ∂J ⎞
∂⎜ ⎟
Evaluation of the vehicle dynamic along the road section ⎝ ∂t ⎠ = 0, Energetic criteria u
∂u
J : total energy of each vehicle type, u : the road profile. F Δ Δ1 Δ2 Δ3
9 Determination the sampling rate to estimate the energy derivative by choosing a window of Lf in length
(to view the road profile),
Figure 4: Quarter vehicle model on the MS-WIM array
9 Lf length = maximum array length (24 m) + the influence length (200 m) # 230 m,
9 The influence length is the distance on which the road profile characteristics influence the vehicle impact
forces.
STEP 2: Optimal sensors spacing (CP+ k )P
Determination of inter-distance Fd = m&x& = k(− x + u) + C(− x& + u& ) ⇒ Fd =
(P + CP+ k m)u&(P)
2

u& is the first derivative of the road profile.

Using Fd function and minimizing the dynamic variation of the impact force leads to
- the optimal location of each sensor,
- and thus the sensors spacing in the optimal road section to get a minimal energy.
Table 1: Optimized sensors spacing

N of sensor’s 1 2 3 4 5 6 …. n
Position of the 523 526,1 528,4 532 534,2 537,6 Moyen (3,1m)
sensor / zone Figure 5: Road profile windows for the RN4 site

CONCLUSIONS & PERSPECTIVES

MS-WIM arrays are promising solutions to develop an efficient overload control tool.
The MS-WIM array design method proposed is original and based on an energy criterion optimization.
This method takes into account the infrastructure characteristics (slope and road profile), the traffic characteristics (speed, silhouette) and the interaction
infrastructure/vehicule (by use of trucks model).
The advantage compared to the current methods which does not take into account the infrastructure characteristics and the traffic silhouette distribution.
However, this method must still be validated on site with real measurements. That is planned to be done with the data of the multiple-sensor WIM system installed on the RN4.

• No Significant influence of sensor’s location along the road profile of the RN4 weight-in-motion site (MAULAN, Dep: 55) because the road profile was excellent.
→ Validation planned with real measurements on RN4 test site.
→ Approach to be correlated with theoretical considerations.

ACKNOWLEDGEMENTS
Daniel Stanczyk (CETE Est)
Yves Marchadour, Charles Thioun (DGMT/MEDADE)