Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Algorithmic Method for Scraper Load-Time Optimization

Marina Marinelli 1 and Sergios Lambropoulos 2

Abstract: Scrapers have established an important position in the earthmoving field as they are independently capable of accomplishing an
Downloaded from ascelibrary.org by University Of Central Florida Libraries on 10/05/13. Copyright ASCE. For personal use only; all rights reserved.

earthmoving operation. Given that loading a scraper to its capacity does not entail its maximum production, optimizing the scraper’s loading
time is an essential prerequisite for successful operations management. The relevant literature addresses the loading time optimization through
a graphical method that is founded on the invalid assumption that the hauling time is independent of the load time. To correct this, a new
algorithmic optimization method that incorporates the golden section search and the bisection algorithm is proposed. Comparison of the
results derived from the proposed and the existing method demonstrates that the latter entails the systematic needless prolongation of the
loading stage thus resulting in reduced hourly production and increased cost. Therefore, the proposed method achieves an improved modeling
of scraper earthmoving operations and contributes toward a more efficient cost management. DOI: 10.1061/(ASCE)CO.1943-7862
.0000624. © 2013 American Society of Civil Engineers.
CE Database subject headings: Construction management; Construction equipment; Earthmoving; Productivity; Optimization;
Algorithms.
Author keywords: Construction management; Construction equipment; Earthmoving; Productivity; Optimization; Algorithms.

Introduction obtained has to physically lift or force its way through all of the
material already loaded. Hence, allowing a scraper to obtain the
Earthmoving operations constitute an inextricable part of most con- maximum load implies neither maximum production nor minimum
struction projects. Scrapers have established an important position cost (Carmichael 1987). Therefore, the optimal loading time needs
in the earthmoving field as they are independently capable of to be determined.
excavating, loading, hauling, and placing materials under a wide In the following sections, the existing method for determining
variety of conditions. Although neither as effective as hoes and the optimal scraper load time is evaluated and a new one is pro-
shovels in excavating nor as efficient as trucks in hauling and posed. Moreover, the results of the two approaches are compared
placing, the fact that this one machine performs all these tasks in proof of the systematic error that the existing method entails.
makes it the equipment of choice when moving large quantities
of soil (Eldin and Mayfield 2005).
The literature attributes increased significance to the stage of Presentation and Evaluation of the Existing
loading. Gransberg (1996) and Schexnayder et al. (1999) consider Load-Time Optimization Method
it determinant for the development of the whole earthmoving op-
eration. Kannan et al. (1999) note that the interrelationship between In each scraper earthmoving operation, the scraper’s cycle time
payload and load time makes loading more complicated than the (tc;sc ) consists of the load time (tl ), transport time (ttr ), spot and
other activities. In scraper earthmoving operations, the evolution dumping time (ts ), and return time (tr ) [Eq. (1)]:
of the loading stage is precisely described by the load growth curve,
that is, the graphic representation of the increase in payload during tc;sc ¼ t1 þ ttr þ ts þ tr ð1Þ
loading time. Independently of the specific equipment and material
characteristics, a scraper’s load growth curve is expected to The literature addresses the determination of the scraper’s opti-
conform to the shape of Fig. 1. The changing slope represents mal load time through the respective load growth curve, according
the rate at which the material enters the bowl. At the beginning, to the graphical process that follows (Day and Benjamin 1991;
the rate is high; however, as the loading time increases the rate Singh 1993; Nunnally 2007). First, the scraper’s cycle time after
decreases. This is due to the fact that the additional material loading, that is, the sum of ttr , ts , and tr , is plotted (point A) on
the left horizontal semiaxis of the load growth curve graph
1 (Fig. 2). Then, the tangent to the load growth curve from point
Ph.D. Candidate, Dept. of Construction Engineering and Management,
School of Civil Engineering, National Technical Univ. of Athens, 9 Iroon A is drawn and the point of tangency is specified (point B). The
Polytechniou St., 157 73 Athens, Greece (corresponding author). E-mail: optimal load time (point C) is finally indicated by the projection
marina01022@yahoo.gr of point B.
2
Associate Professor, Dept. of Construction Engineering and Manage- Evidently, the above approach lies on the assumption that ttr , ts ,
ment, School of Civil Engineering, National Technical Univ. of Athens, and tr are independent of the weight loaded and, as such, they can
9 Iroon Polytechniou St., 157 73 Athens, Greece.
be used to determine tl . This assumption could be considered rea-
Note. This manuscript was submitted on December 27, 2010; approved
on June 11, 2012; published online on July 25, 2012. Discussion period sonable for tr, given that during the return, the scraper is empty.
open until October 1, 2013; separate discussions must be submitted for in- Likewise, ts is not expected to be substantially affected by the load.
dividual papers. This paper is part of the Journal of Construction Engi- However, this is certainly not the case for ttr, because according to
neering and Management, Vol. 139, No. 5, May 1, 2013. © ASCE, ISSN Eq. (2) the maximum speed of the vehicle during hauling U max
0733-9364/2013/5-459-465/$25.00. (km=h) depends on the weight of the load transported, Btr :

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MAY 2013 / 459

J. Constr. Eng. Manage. 2013.139:459-465.


Table 1. Example Data
Parameter Value
N e (kW) 335
Bo (kN) 464.75
Bmax (kN) 372.85
L (m) 450
wr (N=kN) 50
ws (N=kN) 10
me 0.95

Fig. 1. Typical shape of a scraper load growth curve


Downloaded from ascelibrary.org by University Of Central Florida Libraries on 10/05/13. Copyright ASCE. For personal use only; all rights reserved.

New Algorithmic Method for Scraper Load-Time


Optimization

This section addresses the determination of the scraper’s optimal


load time through an algorithmic process that takes into account
the actual effect of the weight hauled on the scraper’s cycle time
and hourly production.

Loading Optimization for a Single Scraper


Fig. 2. Graphical determination of the scraper’s optimal load time The objective is to detect, for given operation conditions, the load-
ing time for which the scraper’s hourly production is maximized.
This requires generation of the scraper’s production function
N e × me Qsc ðt1 Þ by means of the respective load growth curve. For this pur-
U max ¼ 3600 × ð2Þ pose, several load cases (i.e., pairs of loading time and respective
ðBo þ Btr Þ × ðwr þ ws Þ
load) are obtained by the load growth curve and the respective
This is also the case for the mean speed U mean [Eq. (3)], given hourly production Qsc is derived for the specific operation accord-
that the derivation of the relevant speed coefficient cs (Fig. 3) ing to Eq. (4). Evidently, the calculation of tc;sc from Eq. (1) re-
depends on the value of the ratio N e =ðBo þ Btr Þ (Kuhn 1974): quires the employment of Eqs. (2) and (3) to yield the transport
and return mean speed that determine ttr and tr . The plotting of
U mean ¼ U max × cs ð3Þ the load time and the corresponding hourly production values visu-
alize Qsc ðtl Þ. Its typical shape is presented in Fig. 4. Evidently,
A numerical example may clarify the above. Assuming the
Qsc ðtl Þ is a continuous and unimodal curve with a unique ex-
data of Table 1, if the load corresponds to the 90% of the scraper’s
tremum, which corresponds to the optimum load time tsc;opt :
max payload, that is, Btr ¼ 0.9 × 372.85 ¼ 335.56 kN, the
maximum travel speed calculated using Eq. (2) is equal to V tr × nexp
U max;90 ¼ 23.9 km=h. Instead, if the load corresponds to the Qsc ¼ ð4Þ
70% of the scraper’s capacity, that is, Btr ¼ 261.00 kN, the respec- tc;sc
tive speed is U max;70 ¼ 26.3 km=h. Apparently, the maximum
speed has been increased by approximately 10%. Furthermore, The detection of tsc;opt constitutes a one-dimensional optimiza-
the speed coefficients emerging from Fig. 3 are also different, that tion problem, which could be addressed by the algorithmic process
is, cs;90 ¼ 0.54 and cs;70 ¼ 0.59. As a result, according to Eq. (3), of the golden section search (GSS). The GSS brackets the maxi-
the respective mean speed for the first case is U mean;90 ¼ mum of the unimodal function Qsc ðtl Þ by successively narrowing
12.9 km=h, while for the second it is U mean;70 ¼ 15.5 km=h, which the range of the interval where it is being sought. Assuming that the
is an increase of 20%. loading time ranges between t1 and t4 , each iteration of the GSS
Given the dependence between mean speed and respective time, employs the golden section equation [Eq. (5)] to yield the inter-
the above differentiations disprove the fundamental consideration mediate values t2 and t3 . Fig. 5 illustrates the geometrical relation-
of the existing graphical method that there is no dependence ship that the golden section entails for the line segments defined by
between loading and transport time. Therefore, a new method t1 , t2 , t 3 , t4 :
detecting the optimal scraper loading time is required.

Fig. 3. Nomograph for speed coefficient determination (N e in PS; Bo ,


Btr in t) Fig. 4. Scraper load time–hourly production typical curve

460 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MAY 2013

J. Constr. Eng. Manage. 2013.139:459-465.


maximum value is expected to occur in the subinterval [t1 , t3 ]; thus,
the boundary value t4 is rejected and replaced by t3, and the next
iteration of the search is applied in the new subinterval [t1 , t3 ].
Likewise, if Qsc ðt3 Þ > Qsc ðt2 Þ, then the maximum is expected
to occur in the subinterval [t2 , t4 ]; thus, the boundary value t1
is rejected and replaced by t2, and the next iteration concerns
the new subinterval [t2 , t4 ]. Therefore, based on the evaluation
Fig. 5. Golden section geometry of the function Qsc ðtl Þ at t2 and t3 , each iteration leads to the
one of the two boundary values t1 or t4 being rejected, as well
as to a new intermediate value t3 or t2 being created. Overall,
the new interval [t10 , t40 ] emerging at the end of the iteration is re-
duced by the factor b of Eq. (5), compared to the original interval
Downloaded from ascelibrary.org by University Of Central Florida Libraries on 10/05/13. Copyright ASCE. For personal use only; all rights reserved.

[t1 , t4 ]. An illustration of the process is presented in Fig. 6.


It is noted that the precise determination of the material volume
corresponding to the new load time arising requires the prior em-
ployment of the least squares methodology to determine the
polynomial that best fits the observations constituting the load
growth curve.
After successive iterations of the GSS, the final determination of
the maximum hourly production is achieved and thus the optimal
loading time tsc;opt;GSS is detected. Given that the load time for full
capacity does not usually exceed 1.5 min, 10 iterations suffice for
the detection of a reasonably small time range (e.g., 0.02 min)
Fig. 6. Golden section search where the optimal load time lies.
The algorithmic procedure achieving the loading time optimi-
zation for a single scraper is presented in the flow chart of Fig. 7.
2
t3 − t1 ¼ t4 − t2 ¼ b × ðt4 − t1 Þ where b ¼ pffiffiffi ¼ 0.618
1þ 5
Loading Optimization for a Fleet of Scrapers and a
ð5Þ
Pusher
Given that Qsc ðtl Þ is unimodal and t1 < t2 < t3 < t4 , the scrap- The previous section addressed the load-time optimization for a
er’s maximum production Qsc;max lies either to the subinterval single scraper. However, the external assistance of a pusher is often
[t1 , t3 ] or to the [t2 , t4 ] one. Therefore, following the comparison necessary. The involvement of a pusher requires the GSS optimi-
of Qsc ðt2 Þ and Qsc ðt3 Þ the search could be limited to one of the zation process to yield not only the scraper’s load time tsc;opt;GSS but
above subintervals. To be specific, if Qsc ðt2 Þ > Qsc ðt3 Þ, then the also the pusher’s load time tpush;opt;GSS . For this additional purpose,

Fig. 7. Optimal load-time determination by means of the GSS (single scraper)

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MAY 2013 / 461

J. Constr. Eng. Manage. 2013.139:459-465.


the pusher’s hourly production function Qpush ðtl Þ is required, as
deriving from Eqs. (6) (Peurifoy et al. 2006) and (7):
tc;push ¼ 0.25 þ 1.4 × t1 ð6Þ

V tr × nexp
Qpush ¼ ð7Þ
tc;push

Its typical shape is presented in Fig. 8. Thereafter, and because


Fig. 8. Pusher load time–hourly production typical curve Qpush ðtl Þ is also unimodal, the GSS is applied to detect the maxi-
mum production of the pusher Qpush;max ¼ Qpush ðtpush;opt;GSS Þ.
Downloaded from ascelibrary.org by University Of Central Florida Libraries on 10/05/13. Copyright ASCE. For personal use only; all rights reserved.

However, the pusher is usually combined with more than one


scraper in order to avoid significant idleness. The scrapers maxi-
mum number is reansonably restricted to the minimum number
required to render the pusher critical, because no economic advan-
tage is gained by having more scrapers. In any case, assuming a
fleet of one pusher and n scrapers, the hourly production of the
fleet is equal to the minimum between the pusher’s production
and the scraper(s) production (Fig. 9).
Therefore, it is necessary to examine whether the pusher’s
maximum production Qpush;max ¼ Qpush ðtpush;opt;GSS Þ is smaller
than the respective production of the n scrapers Qsc;n ¼
n × Qsc ðtpush;opt;GSS Þ, as well as whether the scrapers’ maximum
Fig. 9. Example of pusher and scraper(s) hourly production curves
production Qsc;n;max ¼ n × Qsc ðtsc;opt;GSS Þ is smaller than the

Fig. 10. Optimal load-time determination by means of the GSS (pusher-scrapers fleet)

462 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MAY 2013

J. Constr. Eng. Manage. 2013.139:459-465.


respective production of the pusher Qpush ðtsc;opt;GSS Þ. If the first Results and Discussions
event occurs, the optimal load time for the fleet is tfl;opt ¼
tpush;opt;GSS ; if the second one occurs, the optimal load time is The proposed method’s possibilities and gains can be assessed by
tfl;opt ¼ tsc;opt;GSS . In a theoretical case that both of these events implementing both the proposed and the existing method under the
occur, the optimal time is the one corresponding to the greater same conditions, as this allows to compare the loading time and the
between Qpush;max and Qsc;n;max . The sequence of comparisons respective hourly production that they entail. Addressing this com-
required to yield the optimal loading time for a pusher–scrapers parison on-site would probably lead to misleading conclusions,
fleet tfl;opt is presented in the flow chart of Fig. 10. given that the observations are prone to the impact of randomly
However, there is still the possibility that none of these events occurring/varying factors, such as the competence and the actions
occur, in case that both tsc;opt;GSS and tpush;opt;GSS are outside the of the operator, the soil composition, and the obstacles on the route.
load-time range where the corresponding unit is critical. In such All these affect the practical verification of both the proposed and
Downloaded from ascelibrary.org by University Of Central Florida Libraries on 10/05/13. Copyright ASCE. For personal use only; all rights reserved.

a case, the goal shifts to the elimination of the fleet’s idle time the existing optimization method; deviations between planning and
and thus the optimization process aims at determining the load time execution are inherent in earthmoving operations. The comparative
that concurrently renders the scrapers and the pusher critical. evaluation of the two approaches can soundly be addressed through
Evidently, this corresponds to the point of intersection of the scrap- the numerical investigation of scraper operations with typical
ers’ and the pusher’s production curves. Therefore, the optimization characteristics. As such were considered the hauling distances
process aims at detecting the root of the function Fðtl Þ ¼ (L) from 100 to 1,000 m and the grades at hauling direction be-
jQsc;n ðtl Þ − Qpush ðtl Þj. For the attainment of this goal, the well- tween 0% and 6% (ws ¼ 0–60 N=KN). Furthermore, the load
known bisection algorithm is applied at [t1 , t4 ], given that Fðtl Þ growth curve of Fig. 1, the scraper of Table 2, and the operation
is continuous and has a root within this interval. The bisection al- conditions of Table 3 were assumed.
gorithm (Fig. 11) divides the initial interval into two halves [t1 , t ] Given its extent, the computational process had to be automated.
and [t , t4 ], where t ¼ ðt1 þ t4 Þ=2, and checks the sign of the Therefore, the graphical determination of the point of tangency
product Fðt1 Þ × Fðt Þ. If Fðt1 Þ × Fðt Þ < 0, then the root lies in (point B, Fig. 2), which anyway involves an inadequate level of
[t1 , t ]; otherwise, it lies in [t , t4 ]. The half-interval that contains accuracy, had to be substituted by an analytical one. The algorith-
the root is preserved and bisected again in the next iteration. After mic process of the GSS was selected to address this objective too,
several repetitions in which the interval becomes smaller and given that the point of tangency corresponds to the maximum angle
smaller, its boundaries converge to the root, which is the optimal formed by the horizontal axis and the line segment commencing
loading time of the fleet tfl;opt . from point A.
Indicative results of the numerical investigation are presented in
Tables 4–7. Comparing the values of tsc;opt;exist and tsc;opt;GSS , that
is, the load time that the existing and the proposed method yield as
optimal, it is evident that the existing method consistently overes-
timates the loading time. The overestimation roughly ranges
between 10% and 40%, depending on hauling distance and grade.
Fig. 12 indicatively compares the results of the two approaches for
hauling routes of grade 3%. There is no doubt that the significant
unnecessary prolongation of the loading stage entails additional
cost resulting from the increased fuel consumption and wear of
the machine’s tires, ground engaging parts, and power train com-
ponents. Furthermore, the maximum hourly production that can be
achieved by means of the existing method Qsc;max;exist is lower
by 0.2–1.5% than the respective hourly production under the
GSS method, Qsc;max;GSS .

Table 2. Numerical Investigation Assumptions


Parameter Value
N e (kW) 335
Bo (kN) 464.75
Bmax (kN) 372.85
V sc ðm3 Þ 26
Ulim (km=h) 53
Note: Data in the table are from 631G scraper characteristics (Caterpillar
2010).

Table 3. Numerical Investigation Assumptions of Operation Conditions


Parameter Value
wr (N=kN) 50
g (t=m3 ) 1.4
ts (min) 0.9
Fig. 11. Optimal load-time determination by means of the bisection me 0.95
algorithm (pusher-scrapers fleet) nexp 0.83

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MAY 2013 / 463

J. Constr. Eng. Manage. 2013.139:459-465.


Table 4. Comparison of Existing and Proposed Method Results for
Grade 0%
tsc;opt;exist tsc;opt;GSS Qsc;max;exist Qsc;max;GSS
L (m) (min) (min) (m3 =h) (m3 =h)
100 0.67 0.59 510.70 512.84
200 0.74 0.64 415.82 418.16
300 0.79 0.68 356.93 358.73
400 0.85 0.72 315.68 317.43
500 0.91 0.76 282.88 284.39
600 0.98 0.81 257.13 258.42
700 1.05 0.83 236.40 238.04
Downloaded from ascelibrary.org by University Of Central Florida Libraries on 10/05/13. Copyright ASCE. For personal use only; all rights reserved.

800 1.12 0.87 215.27 216.75


900 1.16 0.91 197.95 199.06
1,000 1.19 0.96 183.34 184.14
Fig. 12. Optimal load time according to the existing and the proposed
method for grade 3%

Table 5. Comparison of Existing and Proposed Method Results for


Grade 2% Following the comparative evaluation of the two approaches it is
clear that the proposed method not only achieves an improved
tsc;opt;exist tsc;opt;GSS Qsc;max;exist Qsc;max;GSS
modeling of scraper earthmoving operations but also enhances
L (m) (min) (min) (m3 =h) (m3 =h)
the cost control effectiveness by averting the reduction of the
100 0.71 0.60 461.20 464.41 hourly production as well as the increased cost that the extended
200 0.79 0.65 366.77 369.92 loading duration entails.
300 0.86 0.70 310.78 313.19 Given that the proposed method lends itself quite well to a
400 0.93 0.74 272.59 274.96
computer-aided spreadsheet application, future work could include
500 1.03 0.79 242.56 244.78
600 1.11 0.85 219.98 221.60 addressing the more complex algorithmic procedure of pusher-
700 1.16 0.87 202.02 203.60 scrapers fleet through a user-friendly tool which would examine
800 1.19 0.92 183.34 184.43 for given operation conditions the cost implications of all the pos-
900 1.22 0.98 167.94 168.66 sible scraper fleet sizes, thus enabling the simultaneous optimiza-
1,000 1.23 1.04 154.98 155.46 tion of the loading time and the number of scrapers.

Conclusions
Table 6. Comparison of Existing and Proposed Method Results for
Detecting the optimal load time in scraper earthmoving operations
Grade 4%
is of critical importance, given that filling the scraper bowl to maxi-
tsc;opt;exist tsc;opt;GSS Qsc;max;exist Qsc;max;GSS mum capacity implies neither maximum production nor minimum
L (m) (min) (min) (m3 =h) (m3 =h) cost. The existing literature addresses the optimization of the scrap-
100 0.74 0.61 418.59 422.18 er’s load time through a graphical method, founded on the consid-
200 0.84 0.66 325.86 329.43 eration that the hauling time is independent of the load time. This
300 0.93 0.73 272.92 275.80 research proves the invalidity of the aforementioned consideration
400 1.05 0.77 237.44 240.48 and presents a new approach for establishing the optimal load time
500 1.14 0.83 210.70 212.95 by means of the GSS and the bisection algorithm. The comparative
600 1.18 0.91 190.96 192.18 implementation of the proposed and existing method demonstrates
700 1.21 0.93 175.06 176.11
that, depending on the route characteristics, the latter entails the
800 1.23 1.00 158.25 158.90
900 1.25 1.07 144.42 144.86 systematic overestimation of the loading time, thus resulting in re-
1,000 1.26 1.13 132.87 133.17 duced hourly production and additional cost implications. There-
fore, the contribution of this research lies in the improvement of
both the modeling and the cost management of scraper earthmoving
operations. Future work on this issue could include the integration
of the proposed time-optimization procedure into a computer-aided
Table 7. Comparison of Existing and Proposed Method Results for
Grade 6%
spreadsheet application, additionally enabling the user to detect the
optimal scrapers fleet size for given operation conditions.
tsc;opt;exist tsc;opt;GSS Qsc;max;exist Qsc;max;GSS
L (m) (min) (min) (m3 =h) (m3 =h)
100 0.77 0.62 383.05 387.04 Notation
200 0.89 0.67 293.13 296.99
300 1.03 0.75 242.98 246.46 The following symbols are used in this paper:
400 1.14 0.79 210.80 213.76 Bmax = scraper max payload;
500 1.19 0.87 186.82 188.54 Bo = scraper tare weight;
600 1.22 0.97 169.02 169.79 Btr = weight of the transported load; p
700 1.24 1.01 154.59 155.28 b = golden section factor = 2=ð1 þ 5Þ ¼ 0.618;
800 1.26 1.08 139.28 139.70 cs = speed coefficient;
900 1.27 1.13 126.76 127.04 Fðtl Þ = difference between scraper and pusher hourly
1,000 1.28 1.16 116.33 116.52
production as a function of the load time;

464 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MAY 2013

J. Constr. Eng. Manage. 2013.139:459-465.


g = material specific weight; U mean = scraper mean transport speed;
L = hauling distance; V sc = scraper volumetric capacity;
me = scraper mechanical efficiency factor; V tr = volume of the transported material;
Ne = scraper flywheel power; wr = rolling resistance coefficient; and
n = number of scrapers; ws = grade resistance coefficient.
nexp = scraper/pusher exploitation coefficient;
Qpush = pusher hourly production;
Qpush ðtl Þ = pusher hourly production as a function of the load
time; References
Qsc = scraper hourly production;
Carmichael, D. (1987). “Optimal pusher–scraper loading policies.” Eng.
Qsc;n = hourly production of n scrapers;
Optim., (Bellingham, WA, U.S.) 12(4), 255–267.
Qsc ðtl Þ =
Downloaded from ascelibrary.org by University Of Central Florida Libraries on 10/05/13. Copyright ASCE. For personal use only; all rights reserved.

scraper hourly production as a function of the load Caterpillar. (2010). Caterpillar performance handbook, 40th Ed.,
time; Caterpillar, Peoria, IL.
Qsc;max;exist = scraper hourly production corresponding to the Day, D. A., and Benjamin, N. (1991). “Scrapers for earthmoving.”
optimal load time that the existing method Construction equipment guide, 2nd Ed., Wiley, New York.
yields; Eldin, N., and Mayfield, J. (2005). “Determination of most economical
Qsc;max;GSS = scraper hourly production corresponding to the scrapers fleet.” J. Constr. Eng. Manage., 131(10), 1109–1114.
optimal load time that the GSS yields; Gransberg, D. D. (1996). “Optimizing haul unit size and number based
t1 = lower bound of the load-time range; on loading facility characteristics.” J. Constr. Eng. Manage., 122(3),
t2 = intermediate time value determined by the GSS; 248–253.
Kannan, G., Vorster, M. C., and Martinez, J. C. (1999). “Developing the
t3 = intermediate time value determined by the GSS;
statistical parameters for simultaneous variation in final payload and
t4 = upper bound of the load-time range; total load time.” Winter Simulation Conf. Proc., P. A. Farrington, H. B.
tc;push = pusher cycle time; Nembhard, D. Sturrock, and G. W. Evans, eds., Association for
tc;sc = scraper cycle time; Computing Machinery (ACM), New York, 1016–1022.
tfl;opt = load time determined by the GSS/the bisection Kuhn, G. (1974). “Bagger-Lkw-Betrieb.” Die mechanik des baubetriebes,
algorithm as optimal for the pusher-scrapers fleet; teil 1: Transport mechanik, Bauverlag, Wiesbaden, Berlin.
tl = scraper load time; Nunnally, S. W. (2007). “Loading and hauling.” Construction methods
tpush;opt;GSS = pusher optimal load time according to the GSS; and management, 7th Ed., Pearson Education Inc., Upper Saddle
tr = scraper return time; River, NJ.
ts = scraper spot and dumping time; Peurifoy, R. L., Schexnayder, C. J., and Shapira, A. (2006). “Scrapers.”
Construction planning, equipment and methods, 7th Ed., McGraw-Hill,
tsc;opt;exist = scraper optimal load time according to the existing
New York.
method; Schexnayder, C., Weber, S. L., and Brooks, B. T. (1999). “Effect of
tsc;opt;GSS = scraper optimal load time according to the GSS; truck payload weight on production.” J. Constr. Eng. Manage.,
ttr = scraper transport time; 125(1), 1–7.
U lim = scraper speed limit; Singh, J. (1993). “Earthmoving machines.” Heavy construction planning,
U max = scraper maximum transport speed; equipment and methods, A/A Balkema, Rotterdam, Amsterdam.

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / MAY 2013 / 465

J. Constr. Eng. Manage. 2013.139:459-465.

You might also like