Вы находитесь на странице: 1из 42

CONVENTIONAL MACROSCOPIC

TRAFFIC FLOW MODEL

An overview class for Master in Transportation Engineering, IoE


Pulchowk Campus

Hemant Tiwari, MSc (Transportation)


Freelancer, Traffic & Transportation Engineer
Project Coordinator, National College of Engineering
Chairperson, Safe and Sustainable Travel Nepal
Email: hemu.ioe@gmail.com
COMMITTED TOWARDS IMPROVEMENT!!!!
Background:
Transportation is regarded as Derived Demand

Classical four step modeling is the basic fundament of


traffic flow

Appropriate indicator is must for the representation of


the complex nature of Transport Network, either
macroscopic or microscopic model.
6/14/2017 Hemant Tiwari 2
Mathematical Traffic Flow Model

Macroscopic Traffic Flow Models: Analyze the traffic flow by


modeling relationships among system level traffic characteristics
like density, flow, speed of a traffic stream.

Microscopic Traffic Flow Model: Analyze the traffic flow by


modeling driver-driver and driver-road interactions within a
traffic stream. Parameter like: spacing, headway are considered.

6/14/2017 Hemant Tiwari 3


Important Macroscopic traffic parameter:

Speed(V)
Distance per unit time.

Flow or Volume (Q)


Vehicles per unit time

Density or Concentration (K)


Vehicle per unit length.

Fundamental relationship: Q = K*V


6/14/2017 Hemant Tiwari 4
Greenshield Model
Greenshield (1935) assume that speed & density are linearly
related.

Model: U=AB*K

Alternatively,

6/14/2017 Hemant Tiwari 5


Speed Density Graph

6/14/2017 Hemant Tiwari 6


Flow-Density Graph

6/14/2017 Hemant Tiwari 7


Speed-Flow Graph

6/14/2017 Hemant Tiwari 8


Greenberg Model
Greenberg (1959) gave logarithmic V-K models

Shows good agreement with for congested flow condition data.


Mathematically,
V Vc *ln(K j )
K
Kj= Jam Density flow)
Vc= Speed corresponding to maximum flow or capacity

6/14/2017 Hemant Tiwari 9


Greenberg Speed-density model

6/14/2017 Hemant Tiwari 10


Underwoods Exponential Model:
Underwood (1961) proposed exponential V-K relationship.
( K
V V f *e k0 )

Give better fit for the uncongested conditions

Unreliable for predicting speeds at high densities.

6/14/2017 Hemant Tiwari 11


Underwood Speed-Density Model

6/14/2017 Hemant Tiwari 12


The Drake Model (Bell-Shaped Curve
Model)

Drake (1961) gave bell shaped Speed-Density relationship.

Best fit for uncongested conditions, but fails to for congested


conditions.

6/14/2017 Hemant Tiwari 13


Drake Speed-Flow Model:

6/14/2017 Hemant Tiwari 14


Pipes' generalized model:
A new parameter (n) was introduced for more generalized
modeling approach.

When n =1, Pipe's model resembles Greenshields' model.

Family of models can be developed for different n

6/14/2017 Hemant Tiwari 15


Modified Greenberg Model
Ardekani and Ghandehari (2008) modify Greenberg model by
introduction of a non-zero average minimum density(K0)

K K
U U *ln( ) j 0
KK
c
0
When density approaches zero, it yields finite free flow speed of
Uf=Uc* ln(1 + kj/ko)

6/14/2017 Hemant Tiwari 16


Modified Greenberg Speed-Density Model

6/14/2017 Hemant Tiwari 17


Underwood Model with Taylor Series Expansion:
For numerical approximation of Jam density,
exponential function is expanded in Taylor series:

Taking up the expansion to the term containing


third power of k yields

6/14/2017 Hemant Tiwari 18


Drake Model with Taylor Series Expansion:

Using Taylor series expansion for numerical approximation for


the jam density, as follows:

6/14/2017 Hemant Tiwari 19


A THESIS ON CALIBRATION OF CONVENTIONAL MACROSCOPIC
TRAFFIC FLOW MODELS

FOR NEPELESE ROADS:

(A case study of Jadibuti - Suryabinayak section)

Identification of conventional macroscopic traffic flow


models and calibration in context of Nepal

Selection of best fit model among various models.

Validation of recommended model.


6/14/2017 Hemant Tiwari 20
Site Selection:
Stretch of 200m towards Thimi from Laxmi steel
building, located at Radhe Radhe Chowk was
considered.

6/14/2017 Hemant Tiwari 21


Method of Data Collection:

Video-graphic method was


used.

Installation of camera on 4th


storey of Laxmi steel building

Six day of data, covering peak,


off peak and holiday duration
was taken.

6/14/2017 Hemant Tiwari 22


Volume data Extraction:

Vehicle Type Adopted PCU


15 minute Classified volume
was extracted and expressed Car, utility vehicle, 1
as PCU per 15 minute per
total width of road (7m) per Jeep Van
each direction
Microbus 1.5
Mini bus and Bus 3
NMT and Vehicle entering via
shoulder was not considered. Two wheeler (Bike) 0.5
Tipper 1.5
Final report NRS 2070 was
considered for PCU values: Heavy Truck 3
6/14/2017 Hemant Tiwari 23
Average Speed Data Extraction:
Proportion based Stratified random sampling (10% of
each type) was conducted.

Vehicle arriving at every 30th second irrespective of mode


and lane was considered.

For extracted sample less than the required, the vehicle


arriving at every 15th and 45th second was considered.

n* L
Space mean speed (Vs ) was calculated. Vs
t
n
i
i1
6/14/2017 Hemant Tiwari 24
Density Data Extraction

Vehicles occupying the section at each 30th second was noted and
converted to PCU

Then averaged to obtain the density corresponding to that 15


minute period.

6/14/2017 Hemant Tiwari 25


Calibration & Validation of Speed Density model

Calibration of various model is done using 12 hour of data (48 points)


along Koteshwor bound direction.

Models with strong fit ( better R2 value) and realistic value of free-
flow speed and jam density were recommended for validation.

The models were validated using 6 hour data of Shallaghari bound


direction data

Finally the most suitable model is recommended as the Best Fit


Model.
6/14/2017 Hemant Tiwari 26
Data Analysis and Result
Plotting of Speed Flow density relationship

Koteshwor bound direction data was used for calibration.

Suitable model based on R2 value, jam density and free flow speed
are refer for validation.

Shallaghari bound direction was used in validation process.

Most realistic model is recommended as the BEST FIT MODEL

6/14/2017 Hemant Tiwari 28


Greenshield Model

70
Observed data
60 V = 61.97 - 2.01 * K
Greenshield curve
R2 = 0.797
50
Speed (kmph)

40

30

20

10

0
0 4 8 12 16 20 24 28 32
Density (PCU/200m)

6/14/2017 Hemant Tiwari 29


Greenberg model

80

70
V = 14.925 ln (171.14/K)
60 Greenberg curve
R2 = 0.845
50 Observed data
Speed (Kmph)

40

30

20

10

0
0 20 40 60 80 100 120 140 160 180
Density (PCU/200km)
6/14/2017 Hemant Tiwari 30
Underwood model:
70

60 Observed data

50 Underwood curve

40 V = 66.365 e-K/21.516
Speed (kmph)

R2 = 0.815
30

20

10

0
0 10 20 30 40 50 60 70 80 90
Density (PCU/200m)
6/14/2017 Hemant Tiwari 31
Underwood Model With Taylor Series Expansion:
70
Underwood with Taylor Expansion
60
Observed points
50
Speed (kmph)

40
V = 66.16*(1-K/21.8+K2/948.6-K3/61976.96)
30
R2 value = 0.814
20

10

0
0 5 10 15 20 25 30 35 40
Density (PCU/200m)
6/14/2017 Hemant Tiwari 32
Conventional Drake Model:
60
Drake Model Curve
50 Observed data

40
Speed (kmph)

30
V = 56.729 Exp (- K2/ 305.737)
R2 = 0.75
20

10

0
0 5 10 15 20 25 30 35 40
Density (PCU/200m)
6/14/2017 Hemant Tiwari 33
Drake with Taylor Series Expansion:
60
Drake with Taylor expansion Observed data
50

40
Speed (kmph)

30 V = 56.71*(1-K2/306.84+K4/188304.85-K6/173339891.3)

20 R2 Value = 0.75

10

0
0 2 4 6 8 10 12 14 16 18 20 22 24
Density (PCU/200km)

6/14/2017 Hemant Tiwari 34


Pipes-Munjal Model:
60 Pipes-Munjal model (N=2)

50 Observed Data

40
Speed (Kmph)

30

V = 55.834*[1-(K/19.29)2]
20
R2 Value = 0.72
10

0
0 2 4 6 8 10 12 14 16 18 20 22
Density (PCU/200m)

6/14/2017 Hemant Tiwari 35


Polynomial Fit
90

80 Two degree polynomial fit Observed Data

70
Speed (kmph)

60

50

40 V = 0.4018*K2 - 7.7448*K + 81.696


30 R = 0.8778
20

10

0
0 2 4 6 8 10 12 14
Density (PCU/200m)

6/14/2017 Hemant Tiwari 36


Comparison of Various models:
Model Calibrated Equation Vf Vc kmph Kj Kc R2
kmph PCU/200m PCU/200m
Greenshields V = 61.97 (1 K/30.83) 61.97 30.98 30.83 15.42 0.80
Greenberg V = 14.99 Ln (168.61/K) _ 14.99 168.61 62.03 0.84
Conventional Underwood V = 66.36Exp ( -K/21.52) 66.36 24.42 _ 21.52 0.82
Underwood with Taylor V = 66.16 * (1- K/21.8+K2/948.6-
66.16 22.05 34.81 21.8 0.82
Expansion K3/61976.96)

Conventional Drake V = 56.73Exp (- K2 / 305.737) 56.73 34.41 _ 12.36 0.75

V = 56.71*(1-
Drake with Taylor Series
K2/306.84+K4/188304.85 - 56.71 34.26 22.13 12.39 0.75
Expansion
K6/173339891.3)

Pipes-Munjal (N = 2) U = 55.83 * {1 (K/19.29)2} 55.83 37.22 19.29 11.14 0.72

Polynomial
6/14/2017(Degree two) V = 0.4018*K2 - 7.7448*K +Hemant
81.696Tiwari81.69 _ _ _ 0.88
37
Underwood model:
70

60 Observed data

50 Underwood curve

40 V = 66.365 e-K/21.516
Speed (kmph)

R2 = 0.815
30

20

10

0
0 10 20 30 40 50 60 70 80 90
Density (PCU/200m)
6/14/2017 Hemant Tiwari 38
Validation Process:
Validation of Recommended Calibrated Model:
Model Calibrated Equation R2 Value Standard Error

Greenshields V = 61.97 (1 K/30.83) 0.868 0.97827


Greenberg V = 14.99 Ln (168.61/K) 0.9 0.841
Conventional Underwood V = 66.36 Exp ( -K/21.52) 0.878 0.942

Recalibration of Models based on validation data


Model Re-Calibrated Equation Vf Vc kmph Kj PCU/200m Kc PCU/200m R2
kmph

Greenshields V=62.93 (1 K/30.11) 62.93 30.11 31.46 15.06 0.87

Greenberg V=12.94Ln(290.33/K) _ 12.94 290.33 106.81 0.89


Conventional V = 64.94 Exp ( - K
64.94 23.89 _ 23.59 0.88
Underwood /23.59)
6/14/2017 Hemant Tiwari 39
Recommendation of Best fit model: (Conventional
Underwood model)

70
Recommended Model fit Observed data
60

50
V = 66.22 * Exp(-K/21.75)
Speed (Kmph)

40
R2 Value = 0.83
30 Standard Error = 1.364
20

10

0
0 5 10 15 20 25 30 35 40 45 50
Density (PCU/200m/7mwidth
6/14/2017 Hemant Tiwari 40
Future Scope:

Lane wise separate calibration of macroscopic traffic flow model

Assessment of present and future level of service offered by the road


based on present and anticipated future traffic.

Development of traffic flow model need to be replicated for urban


roads within the valley.

6/14/2017 Hemant Tiwari 41


THANK YOU!!!!!
6/14/2017 Hemant Tiwari 42

Вам также может понравиться