FYP Slides

8/9/2019 FYP Slides

1/31

Presented by LE THU HIEN

Environmental Engineering

November 11, 2009

Nanyang Technological University

School of Civil and Environmental Engineering

FINAL YEAR PR!ECT

Prediction of

Total Bed-Material Load

in Alluvial Streams

8/9/2019 FYP Slides

2/31

8/9/2019 FYP Slides

3/31

INTRODUCTION

BACK GROUND

Sediment transport is known as one o many problems to

rivers and natural streams

Monitoring the eect o sediment discharge on alluvial

stream is a challenge to many scientists and engineers

Modeling and orecasting the sediment rate in rivers is animportant stage

Le Thu Hien

8/9/2019 FYP Slides

4/31

INTRODUCTION

OBJECTIVE !ormulate an e"uation to predict the rate o

sediment discharge in alluvial streams

Minimi#e the inaccuracy and amend any

unreliable assumptions o other previous studies

Le Thu Hien

8/9/2019 FYP Slides

5/31

PASTSTUD

SEDIMENT DISCHARGE RATE

Indirect definition

$he sum o two separated elements% bed load andsuspended load

"irect definition

Consider $otal &ed Material 'oad ($&M') as the

most signiicant part *ash load is neglected in natural rivers

Le Thu Hien

8/9/2019 FYP Slides

6/31

PASTSTUD

Sediment discharge vs#

Flo$ discharge Discontinuous

Double valued

Sediment discharge vs#

%ean &elocity Continuous

Single valued

Le Thu Hien

8/9/2019 FYP Slides

7/31

PASTSTUD

'ennedy ( 'arim)s *rediction model

Pro+lems

Unrelia+le *arameters

Ins,fficient data so,rce

Ina**ro*riate ass,m*tions

Le Thu Hien

8/9/2019 FYP Slides

8/31

M!T"ODOLO#

Dimensional Analysis

Regression Analysis

Evaluation and Discussion

Le Thu Hien

8/9/2019 FYP Slides

9/31

DATASOURC!S

Collected by *illiam R+ &rownlie rom ield

investigation

,ncluding -. lows and /01. records

Sediment discharge rate was given in orm o

Cppm 23s4(3s 5 3) and can be obtained by

Le Thu Hien

8/9/2019 FYP Slides

10/31

DIM!NSIONALANALSIS

Le Thu Hien

8/9/2019 FYP Slides

11/31

DIM!NSIONALANALSIS

Selection of parameter

Physical consideration

Mean velocity should be used instead o *ater discharge

*ater 6 sediments inter6relationship should be taken into

consideration

Stream power includes the incipient motion7 which is not

a reliable actor should be discarded

Le Thu Hien

8/9/2019 FYP Slides

12/31

DIM!NSIONALANALSIS

Le Thu Hien

8/9/2019 FYP Slides

13/31

DIM!NSIONALANALSIS


Regression conditions

8igh correlation between dependent variable and

independent variables

'ow correlation between independent variables

Le Thu Hien

8/9/2019 FYP Slides

14/31

DIM!NSIONALANALSIS

Correlation +et$een varia+les

Le Thu Hien

8/9/2019 FYP Slides

15/31

DIM!NSIONALANALSIS


Chosen parameters

Dependent% P-9 ,ndependent% (P-7 P.7 P: )

(P-7 P; 7 P: )

(P07 P.7 P: )

(P07 P; 7 P: )

Correlation was ound in each data set at

speciic location

Le Thu Hien

8/9/2019 FYP Slides

16/31

DIM!NSIONALANALSIS


Correlation was computed in each data set at

speciic location

Le Thu Hien

8/9/2019 FYP Slides

17/31

R!#R!SSIONANALSIS

8/9/2019 FYP Slides

18/31

R!#R!SSIONANALSIS

$hree approaches

Regression on all data

Regression on data at speciic location

Regression on all data with the neglection o

abnormal and unreliable value

Le Thu Hien

8/9/2019 FYP Slides

19/31

R!#R!SSIONANALSIS

Approac! "#

Regression on all data

%odel -% ln(P-9) 2 >95 >/ln(P.) 5 >-ln(P/) 5 >?ln(P0)

%odel .% ln(P-9) 2 >95 >/ln(P.) 5 >-ln(P/) 5 >?ln(P0) 5 >.ln(P.) @ ln(P/)

%odel 1% ln(P-9) 2 >95 >/ln(P.) 5 >-ln(P/) 5 >?ln(P0) 5 >;ln(P/) @ ln(P0)

%odel /% ln(P-9) 2 >95 >/ln(P.) 5 >-ln(P/) 5 >?ln(P0) 5 >1ln(P0) @ ln(P.)

%odel 2% ln(P-9) 2 >95 >/ln(P.) 5 >-ln(P/) 5 >?ln(P0) 5 >0ln(P.) @ ln(P/) @ ln(P0)

%odel 3% ln(P-9) 2 >95 >/ln(P.) 5 >-ln(P/) 5 >?ln(P0)

5 >.ln(P.) @ ln(P/) 5 >;ln(P/) @ ln(P0) 5 >1ln(P0) @ ln(P.)

5 >0ln(P.) @ ln(P/) @ ln(P0) Le Thu Hien

8/9/2019 FYP Slides

20/31

R!#R!SSIONANALSIS

Le Thu Hien

8/9/2019 FYP Slides

21/31

R!#R!SSIONANALSIS

Le Thu Hien

8/9/2019 FYP Slides

22/31

R!#R!SSIONANALSIS

Approac! $#

Regression on data at s*ecific location

8/9/2019 FYP Slides

23/31

R!#R!SSIONANALSIS

Le Thu Hien

8/9/2019 FYP Slides

24/31

R!#R!SSIONANALSIS

Approac! %#

Regression on all data e4ce*t for a+normal and

,nrelia+le records

Data set . and -? are discarded

Some o abnormal points picked rom the previous

graph are also ignored

Le Thu Hien

8/9/2019 FYP Slides

25/31

R!#R!SSIONANALSIS

Le Thu Hien

8/9/2019 FYP Slides

26/31

!$ALUATION % COMPARISON

RESU&TS Coeicients are almost consistent in magnitude and signs

8igh reliability and itness% very high R6s"uare value

Le Thu Hien

8/9/2019 FYP Slides

27/31


DISCUSSION All models could be used to predict sediment discharge

Approach - is less accurate compared to other two

Approach ? can discard abnormal records and limit thesmaller range

A**roach - A**roach . Le Thu HienA**roach 1

8/9/2019 FYP Slides

28/31


RECOMMENDATIONS

Model ? is most suitable and reliable to use in normal

cases

8igh accurate range applied to the model% P-9 rom

(6?+; ;+9)

Errors o measurement made by human and e"uipments

should be avoided as much as possible to increase thereliability o prediction

Le Thu Hien

8/9/2019 FYP Slides

29/31


comparion

'ennedy ( 'arim )s *rediction

Same physical consideration% Mean velocity7 Shear

velocity B sediment si#e

Critical shear velocity eceed the mean velocity7 make the

regression terms not applicable to use

Le Thu Hien

8/9/2019 FYP Slides

30/31

CONCLUSION

A model of sediment load has +een com*,ted $o help civil and hydraulic engineers to predict the sediment

discharge in natural systems

$o be a tool in urther design and controlling o streams problems

Advantages More accurate than most o the previous studies7 even not yet

/99 but still it is acceptable Applicable in very wide range o low7 ecept or laboratory range

Le Thu Hien

8/9/2019 FYP Slides

31/31

THANK

'OU L Th Hi

Nanyang Technological University

School of Civil and Environmental Engineering

Supervisor% Assoc Pro Cheng =ian Sheng

Moderator% Assoc Pro 'im Siow ong

FYP Slides

Documents