1. INTRODUCTION
The
load and weight process in a cement industry is ones of logistic process that
generate the biggest time of occurrence, increasing the queues and increasing
the lead time. Another concern is related to costs of production and productive
capacity. According to (WIENDAHL, 1995), shorts lead times facilitate
compliance with the due dates. The loading full and probable increases the
occupation of productive resources and reduces the cost, given same goal
factory.
In
the cement industry, the transportation of raw materials and finished products
is done by cargo vehicles, with capacity for large batches of cement either
bagged or bulk cement, reflecting thus in occupation of active load (and
equipment to load in bulks forklifts) for long periods of time, increasing the
queues. One solution to the above problem would be to acquire more charging
equipment; moreover, the costs to purchase, install and put into use this new
equipment will increase.
Thus,
the company faces a trade-off between investing in new resources or not, with
no knowledge if the modification may bring significant benefits to the process
and what these benefits would be. A useful technique in solving problems of
this nature, where there is variation and choices involved in the process, is
the use of computer simulation. Through simulation, you can test and analyze
different settings for resources and their potential impacts in productive
system. Because this technique is done virtually, no need physical intervention
is required in the real system, thus demanding low investment.
The
objective of this work is to simulate alternatives to solve the problem of
queues in the activities of weighing and loading cargo vehicles in a cement
factory in the interior of the Brazilian state Rio Grande do Norte. The study
considers the loads carried by trucks and bulk, looking for an alternative that
reduces the average waiting time of customers per charge and the consequences
of alternative resource use and costs for the company.
2. SIMULATION
There
are a lot of definitions of simulation. Some are general and include both the
computational models as the physical models, the definition of Gordon (1978)
that says that systems simulation is the technique of solving problems by
observing the performance of a dynamic system´s model over time. Shannon (1975)
also provides a global view to say that simulation is the process of developing
a real system´s model that conducts experiments in this model, in order to
understand the behavior of the system and / or evaluating various strategies
(with is limited by a criterion or a set of criteria) for system operation.
Naylor
et al. (1971) offers a definition based on computers for experimentations in
which simulation is a numerical technique for conducting experiments on a
digital computer, involving some types of logic models that describe the
behavior of an economic system or business (or a partial aspect of one) over
long time intervals.
According
Doukidis (1987), the primary function of a simulation model is to examine how
the system behaves over a period of time. To achieve this goal, the model
should provide facilities to represent the current state of the system, and
various pre-conditions which, if met, will result in a future state.
3. METHODOLOGY
The
methodology adopted for the development of this study follows the stages of the
simulation process shown in Figure 1.
This
paper is a case study of a descriptive nature, also involves the use of standardized
techniques of data collection, in which we used systematic observation.
The
study was started with the company's choice. It was later searched a better
understanding of the production process, in order to define, clearly and
objectively, the problem to be solved, and the study plan. At this stage, each
relevant information was analyzed in order to identify the problem at level of
goals, as well as the constraints and complexities, since it is not possible
solve a problem without knowing it deeply. The first step is to clearly define
the related goals to the problem to be solved. At this stage the process was
modeled with the aid of two tools - Flowchart and IDEF-SIM.
The
second stage of the study is to perform preliminary planning and observation of
data collection in order to obtain a faithful representation of the problem.
These data relate to the arrival intervals of clients, here called 'customers -
entities', and the length of stay of the system, i.e. the duration of
attendance.
Using
these data entry was performed followed by a statistical analysis (with input
analyzer ®), whose main objective is to measure the theoretical distribution
curve of probabilities that best reproduces the behavior of each data set.
After statistical treatment done this, the next step is to develop a computer
program for the proposed scenario using the software Simpy. Thus, it appears to
ensure that all specifications were met and implemented, and finally, the model
must be validated and executed.
Figure 1: Stages of Simulation Process
4. PROBLEM DEFINITION
The
cargo vehicles arrive at the factory and wait in a queue to be served. After
the balance becomes available, then the cargo vehicle starts to be served in
order to identify, collect license plate number, to collect the name of the
driver, to collect the name of the company, to collect the name of the
application, the day of tare of the truck, and then receive authorization for
loading. This done, it follows to the supply, which can be bagged or in bulk.
The
bagged cement is stacked on pallets, ready to be loaded. The operator carries
out routine procedures such as the preparation and cleaning of the vehicle
body. The forklift then starts placing the pallets on the car, in which
depending on the capacity, accommodates up to 7 pallets with 40 bags of 50kg
each (14000kg). In normal operating conditions, this procedure takes around 20
to 30 minutes. It is worth mentioning that there are two forklifts to carry the
load of vehicles therethus waiting for customers at certain hours of the day. The figure below illustrates what has been described previously:
Figure 2: Cargo
vehicles waiting loading
Figure 3: Loading complete
The
other form for the loading of vehicles is the in bulk supply, where the empty
vehicles are performing the same procedure weight (tare) defined above and are
directed to one of two equipment to load in bulks loading. As the loading of
bagged cement, the duration of this procedure is around 30 minutes per truck.
After loading, the vehicle returns to the entrance where your new weight is
determined, now with the cargo.
Figure 4: Supply in
bulk
Figure 5: Queue of vehicles waiting supply
5. MODELING OF THE SERVICE PROCESS
The
procedures described in the previous section consume around 30 minutes. This
study is considering the full-time load as the sum of the times: 1) The vehicle's access to the equipment to
load in bulk / industry; 2) Preparation of the equipment to load in bulk and
loading; 3) Losing time because malfunction during fueling; 4) Set manual
canvas - for expedition bagged; 5) Access to the weighing pan, final weighing
and calculation of tax. Then, using modeling tools (Flowchart and IDEF - SIM),
it will demonstrate the behavior of vehicle from the moment of arrival in
industrial plant until its release.
Figure 6: Modelling the service process - Flowchart
Figure 7: Modeling the processo of attended –
IDEF-SIM
5.1. The problem
The
arrival rate of vehicles in the industrial plant is more intense between 10h00
and 17h00. In the factory there are two weighing – One for vehicles that will
supply and another for vehicles that will make loading. It was observed that
the intervals between arrivals of vehicles are greater than the processing
time, characterizing the emergence of queues in both loaging (bagged or in
bulk). A service process for cargo vehicles consists in a loading order that
comes to an operation waits its turn in the queue, it is processed and finally
dispatched. The order of service is subject to priorities exchange and
interruptions for maintenance or lack of materials. (SELITTO, BOCHARDT,
PEREIRA, 2009). It is noticeable that vehicles face more than one queue, with
the highest waiting time is the first, in weighing scale. Á priori, this
happens because of the reduced number of attendants (3 employees) to handle
various demands and still receiving raw material. With background, there is the
limitation of resources (2 equipment to load in bulk and 2 forklifts) that
contributes to increasing the waiting time.
In
general, the specific objective of this study is to identify alternatives,
through computer simulation, to decrease the waiting time. The parameters used
for decision making were based on the traditional characteristics of the queue
according to Table 1, with real data.
Table 1: Queue
characteristics
|
Arrival interval |
11,31 minutes |
|
Rate of arrival |
0,09 vehicles per minute |
|
Average service time
weighing scale – loading – weighing scale |
45,62 minutes |
|
Service rate |
0,02 vehicles per minute |
|
Average waiting time weighing scale –
loading – weighing scale |
47,88 minutes |
5.2. Analysis of output data
The
simulation was performed with 3 different scenarios to achieve the research
objective, which is to find the best alternative to minimize the time customers
wait queue in loading operation.
In
the first scenario it is considered the acquisition of a weighing scale to work
in parallel with that already exists in the input stream of vehicles. This
would divide the existing single queue at this stage, also requiring the hiring
of at least one attendant to supply the deficiency. Therefore, it was projected
the proposed scenario 1, which it was obtained the results shown in Table 2.
Table 2: Scenario 1
simulated
|
Scenario 1 |
|||
|
Characteristics |
Average time in queue |
Simulated dverage
time in queue |
Reduction of waiting time (%) |
|
weighing scale (entrance) + Hiring an attendant |
26,45 minutes |
13,69 minutes |
48,24% |
In
a second and third scenario, the change occurred only in loading operations for
optimization. It is observed that a major average waiting time occurs in the
operations of loading of bags (forklift) and in bulk (Equipment to load in
bulk). It was then searched in the second scenario, the reduction for the time
in queue by adding a forklift making three resources available for loading
bags. This new context requires hiring a new forklift operator. Likewise in the
third scenario, a equipment to load in bulk was increased to accelerate the
cement supply in bulk, with hiring a new worker handling the equipment to load
in bulk. The following tables gather the data obtained from these situations:
Table 3: Scenario 2
simulated
|
Scenario 2 |
|||
|
Characteristics |
Average time in queue |
Simulated dverage
time in queue |
Reduction of waiting time (%) |
|
Forklift + Hiring an attendant |
33,77 minutes |
23,16
minutes |
31,41% |
Table 4: Scenario 3
simulated
|
Scenario 3 |
|||
|
Characteristics |
Average time in queue |
Simulated average
time in queue |
Reduction of waiting time (%) |
|
Equipment to load in bulk + Hiring an attendant |
29,41 minutes |
23,18
minutes |
21,18% |
The
fourth scenario is similar to first on, this one proposed to buy another
weighing scale. The new resource would work in parallel with the existing one,
to minimize the size and the average queue time. The new acquisition, as
already shown in the first scenario, requires hiring a new clerk. Considering
all other variables unchanged simulation, the results are shown on Table 5.
Table 5: Scenario 4
simulated
|
Scenario 4 |
|||
|
Characteristics |
Average time in queue |
Simulated average
time in queue |
Reduction of waiting time (%) |
|
weighing scale + Hiring an attendant |
31,23 minutes |
19,34
minutes |
38,07% |
The
fifth, and the last, proposed scenario is also called optimal scenario in
which includes all the changes suggested above. The average time in queue is
considered real is the sum of the times in queue for each step (+ Load
Balance-entry - Trunk / + Forklift Scale-out). Obviously, this scenario
requires a high initial investment, but in terms of simulation time in queue,
it is important get a more holistic view. The
table 6 shows the fifth scenario analyzed.
Table 6: Scenario 5
simulated
|
Scenario 5 |
|||
|
Characteristics |
Average time in queue |
Simulated average
time in queue |
Reduction of waiting time (%) |
|
2 weighing scale + 2 attendant + 1 Equipment to load in bulk + Forklift + Operators |
47,88 minutes |
8,31
minutes |
82,64% |
6. CONCLUSION
This
paper presented a suggestion for improving the cargo vehicles service at cement
industry reducing the time spent of vehicles in logistics operations and cargo
dispatch. For this purpose, the program Simpy was used for simulates scenarios
more optimized, reducing the waiting time and therefore the queue of vehicles.
As
can be observed, the highest times of vehicles queued are the operations supply
and load. Therefore, it was proposed the acquisition of new physical resources
and staffing (scales, trunks, forklifts and operators and attendants) in order
to speed up the processing of entities.
Thus,
it was shown that the inclusion of new features is reflected by a decrease of
at least 21.18% in waiting time.
7. REFERENCES
DOUKIDIS,
G. (1987) Ananthology on the homology os simulation with artificial
intelligence. Journal of operation
Research Society, v. 38, n. 8. Grã-Bretanha.
GORDON,
G. (1978) System Simulation, 2nd ed.
Englewood Cliffs, New Jersey, Pratice Hall, p. 423.
NAYLOR,
T. H.; JOSEPH L.; BALINTFY, D. S.; BURDICK, K. C. (1971) Técnicas de Simulação em Computadores.
São Paulo: Editora Vozes em colaboração com a Editora da Universidade de São
Paulo.
SELITTO, M. A.; BOCHARDT, M.; PEREIRA, G.
M. (2009) Análise de uma operação logística de carregamento e expedição de
cimento por simulação computacional. Revista Gestão industrial –
UTFPR.
SHANNON,
R. E. (1975) Systems simulation: the
art and science. Englewood Cliffs: Prentice Hall, Inc. 387p.
WIENDAHL,
H. P. (1995) Load-oriented manufacturing
control. Berlin: Springer-Verlag.