CHAPTER III CASE COMPLETION This chapter contains a description of the problem about the adaptation of a job training report of PT Indah Kiat Pulp and Paper Tbk consists of background
CHAPTER III
CASE COMPLETION
This chapter contains a description of the problem about the adaptation of a job training report of PT Indah Kiat Pulp and Paper Tbk consists of background, the cornerstone of the theory of the collection, data processing and analysis of the results of the data processing and the conclusion.
Introduction
Introduction this menejelaskan about early stages in the research work of this practice, which consists of the background , problem formulation, objective, and research scope.
Background
The production process in manufacturing companies is the core of all activities of the company. Production activities shall be performed in accordance with proper planning so that those activities run more effectively and efficiently so that it can increase the profits of the company. The quantity of the product that is produced shall also comply with the planning that has been done by the company. This is due to a lack of product produced, will could result in consumer beralihnya to other companies. This is certainly not very desirable by any company.
Forecasting is one of the tools that can be used in the conduct of company decision-making for one or several periods in the future. The forecast is based on the history of several periods ago by looking at the pattern of demand from consumers. Forecasting itself is essentially an allegation or estimates about the occurrence of an event or events in the future. In this case the note, namely the quantity of the product to be produced. The accuracy of forecasting the needs of consumers largely determine the quality of the planning undertaken by the company.
PT Indah Kiat Pulp and Paper Tbk is one of the companies engaged in the production of paper. On its own paper production requires the Recovery Boiler to residual waste composting the pulp from the pulp making laundering be BL, where the results of the processing of waste from paper mills cannot be discarded carelessly, cause contain toxic chemicals can contaminate the environment when not prepared in advance. The recovery Boiler is itself one of the units that use waste treatment methods most used in the pulp and paper industry. Black liquor (BL) sent to Vacum Evaporator (VE) subsequently sent to the Recovery Boiler units are processed to produce green liquor (GL) and steam. Green liquor (GL) further processed to produce white liquor (WL) are in process in the RC that was then reused as a cooking liquor on pulp making (PM). Meanwhile, steam can be used to drive a steam turbine power plant so that the generated electricity. It is therefore necessary to determine how consumption forecasting utility that is used to support the process of the production of steam in Recovery Boilers 13 so that for the production of future periods can be foreseen.
Problem Formulation
Based on the background that have been described above, then the formulation of research problem is how to predict the production of steam on RB 13.
Objective Research
Based on the existing problems, then the objective research is how to determine the steam production forecasting in RB 13 to 12 the next period.
Research Scope
Scope of this research on the issue is as follows:
Data processed by data taken from the Recovery Boiler 13
The data about Steam Production 2016 and 2017
Literature Rivew
This section contains about the theories that support the research conducted in PT Indah Kiat Pulp and Paper Tbk Perawang.
Forcasting
Forecasting is the beginning of a process of taking a decision to memperkiraan the expected level of demand for a product or multiple products in a certain period of time in the future. Basically forecasting is simply an estimate (guess), but using the approaches and specific techniques, forecasting can serve as a good guideline in helping the decision making (Ginting, 2007).
Time of forecasting can be classified into three parts namely (Ginting, 2007):
Short-Term Forecasting
The forecasting results of one year or less, usually used for forecasting quantity scheduling product to be produced, the production time scheduling in one time period.
Medium-term forecasting.
To predict the form of one to five years to come. At this period forecasting-oriented forecasting capacity, labor recruitment, adding machines, etc.
Long term forecasting
Used to take decisions on product planning, market planning and expenses of the company, the plant feasibility study, budget, etc.
Common procedures used in the quantitative forecasting is as follows (Ginting, 2007) :
Define the objective of forcasting.
Create a scatter diagram.
Select at least two methods that fulfill the purpose of forecasting and in accordance with the plot data.
Countdown-forecasting function parameters.
Calculate the error forecasting is happening.
Select the best method.
Verify the forecasting.
An important step after forecasting forecasting verification is done in such a way to reflect the past data and the underlying causal system request. All representations of these reliable forecasting, forecasting results will continue to be used. If during the verification process found doubts the validity of the forecasting methods used, other methods must be used that is more suitable. Validity must be determined by the appropriate statistical test. After a forecast is made, always doubts arise when to be created a new forecasting method. Forecasting should always be compared with demand on a regular basis. At one point the action to be taken when forecasting revision found evidence of a change in the pattern of demand that is convincing. In addition, it causes changes in the pattern of demand must be known. There are some tools that can be used to verify the forecasting and detecting changes in the causal system aspects influenced changes in patterns of demand. One of them is map control forecasting (Kusuma, 1999).
3.2.1.1 Forecasting Methods
The methods used in forecasting there are two methods of qualitative and quantitative forecasting forecasting. Each method has its own advantages. Quantitative forecasting is better used for forecasting short term, whereas qualitative forecasting is better used for long-term forecasting. Explanation of the two methods is as follows:
Qualitative Methods
Qualitative methods are typically used when there is no or little data of the past is available. In this method, the expert opinions and predictions they relied upon to set the request to come. Qualitative methods are widely known is the method Delphi method and nominal Group (nominal group technique). (Baroto, 2002).
Qualitative forecasting methods, among others, are:
The method Delphi is forecasting based on the process of the convergence of opinion of some person or expert interactively without mentioning his identity.
Market research is forecasting method based on the results of market surveys carried out by staffs marketers product or who represent him.
The semi-historical Analogy is forecasting techniques based on patterns of past data of products which can be likened in analogy.
The panel’s Consensus forecasting is solely based on the consideration of the management, generally by senior management.
Quantitative Method
Quantitative methods are forecasting that is the analysis of past data to gain wisdom in the days to come. Kuantatif forecasting is better used for short and medium-term forecasting. Quantitative methods consist of two techniques are (Makridakis, dkk, 1999) :
Time Series
This method treats the system as a black box with no effort to find other factors that affect the behavior of the system. This method is suitable for forecasting short term or medium. The method is often used in regular sequence dinataranya techniques are:
Smoothing Methods
The methods included smoothing methods are:
Simple Average
Average method simply calculates rataan from available data (some T). Method of average equations are:
…(3.1)
Single Moving Average
The term moving average illustrates the procedure if there is new data, the new average can be calculated and the data is then deleted. Align the new flat – will be used to predict. The equation is used to predict. The equation of a single moving average is:
…(3.2)
Double Moving Average
Double Moving Average forecasting covers 3 aspects, namely:
Use a Single Moving Average at time t.
Going adjustments between Single Moving Average
Double Moving Average (S’t – S “t) at time t.
Going adjustment trend t – N + 1.
Eksponensial Smoothing Methods
This methods consist of :
Single Eksponensial Smoothing
Double Eksponensial Smoothing: Brown’s One Parameter Linear
This technique is done if possible three data and one ?. values of the equation is as follows:
Double Eksponensial Smoothing : Holt’s Two Parameter
Triple Eksponensial Smoothing : Brown’s One Parameter Quadratic
Triple Eksponensial smoothing : Winter’s Three Parameter Trend and Seasonality
The decomposition Method
The decomposition has the assumption that data is arranged as follows:
Data = pattern + errors
= f (trend, cycle, seasonal) + errors
The writing of mathematical approach to decomposition is (Makridakis, dkk, 1999) :
Xt = f(It, Tt, Ct, Et) …(3.3)
Where:
Xt = The value of the periodic sequence (actual data) in period t
It = Components (Index) seasonal period t
Tt = The components of the trend in the period t,
Ct = Cycle at period t, and
Et = Component error at period t
Methods of Causal or Eksplanatoris
The methods included in this eksplanatoris method is a method of regression. Regression techniques generally discuss the approach to causation (causal) or coercive clarify (explanatory) for forecasting. These techniques try to estimate future circumstances come up with discover and measure the factors (independent) are important, along with the influence of the variables is not free which will be foreseen.
The regression method consists of 2 parts, namely simple and multiple regression regression (Makridakis, et al, 1999) described as follows:
Simple Regression.
Common model used in forecasting is essentially the form of the polynomial are
…(3.4)
With a number of poliomial n-order, where Y (t) is the value estimation against actual data Y (t), t and a, b, and c, namely adjustment coefficients of the polynomials. A simple regression method is also divided into several methods, namely:
Constant Method
The value of forecasting for each period t is obtained by using the equation:
…(3.5)
If we deduct the value of forecasting with actual historical value, then it can be determined the magnitude of error in period t. Equation as follows:
…(3.6)
Linear Method
The following forecasting methods used in case of fluctuations in historical data in the form of a straight line either towards the top or bottom of the direction of the X and Y fields all the time.
An error occurred can be minimize with the equation:
…(3.7)
The process of adjustment by using the linear method starts by computing the value of , with the equation:
…(3.8)
The value obtained on the substitution equation 3.7 for obtaining parameter values of .
…(3.9)
Quadratis Method
Next, do the calculation for the value of forecasting the future in case of fluctuations of random data with quadratic curves. Sum of the squared error is used to manage the value of the method compared to:
…(3.10)
Stimulation of the equation are given on the values a ?, b ? and c ? are obtained by making zero equation for each parameter. The equations used are:
…(3.11)
After the value b ? obtained, then it can do calculations for values with equation (Makridakis, dkk, 1999):
…(3.12)
Kemudian, dapat diperoleh nilai dengan persamaan :
…(3.13)
Exponential Method
This Exponential method using the equation:
…(3.14)
With transformation algorithms for the above equation are calculated from the following equation:
…(3.15)
Siklis Method
This method is a method that follows the cycle of the actual demand. As for the equation for this method are (Makridakis, dkk, 1999):
y’ = a + b sin (2?t/n) + c cos (2?t/n) …(3.16)
Multiple Regression
This method considers multiple variables is not free will be foreseen. Multiple regression is a common form of
Y=b_0+b_1+b_2+?+b_k X_k+? …(3.17)
3.2.1.2 Error of forcasting result
Implementation of forecasting in the American production of course requires parameter acceptance. These parameters are described in the form of measurements errors of forecasting results. The magnitude of the error in the i (ei) is expressed as:
…(3.18)
Dengan : ei = Error for i period
Xi = Actual Data for i period
Fi = Forcasting Value for i
Some statistical forecasting error measurements that can be used include (Tersine, 1994) :
Mean Absolute Deviation (MAD) = …(3.19)
Mean Squared Error (MSE) = …(3.20)
Standar Deviation of Regretion (Sr) = …(3.21)
Mean Absolute Percent Error (MAPE) = …(3.22)
Mean Error (ME) = …(3.23)
Mean Percent Error (MPE) = …(3.24)
Statistics on the size of the others are forecasting errors by using the tracking signal (TS), with the following formula:
Tracking Signal (TS) = …(3.25)
The value of the tracking signal is a good measure of a forecasted by estimating the actual values. If the value of the tracking signal is positive, meaning the actual value request smaller than forecast, while if is negative, meaning that the actual value is less than the demand forecast. The value of a good is said to be in TS the sum of the difference between the actual data with forecasting approaches zero, or in other words the number of positive error negative error in balance with. If the value of the tracking signal has been calculated can be built, then the next map control to see the distribution and movement of data from the value of the tracking signal. According to George Plossl and Oliver Wight, the value of the tracking signal is preferably a maximum ± 4 as his limit. If the value obtained the maximum threshold, which means that forecasting model needs to be reviewed because of the accuracy of forecasting is not acceptable (Gasperz, 2001).
3.2.1.3 Verification Results Forecasting
Forecasting results verification is needed to see whether the forecasting method is obtained against representative data (Ginting, 2007). The verification process is done using pete moving range. Map moving range is used to compare the values of the actual observation with a value of forecasting of a request.
The price of moving range retrieved from:
…(3.26)
Where :
MRt = ?(y’t – yt) – (y’t-1 – yt-1)? …(3.27)
The upper and lower control limits on the map moving range is:
BKA = +2,66 …(2.28)
BKB = -2,66 …(2.29)
Based on this map, it will be seen whether the spread of the data still in control or not. If the distribution is outside the control limits, then the function or method of forecasting are not appropriate or not representative.
Research Methodology
Research methodology describes the steps that are performed in this research as follows :
Identification of Problems
Based on the observations that have been made on PT Indah Kiat Pulp and Paper Tbk then can be seen there is a problem on the steam production according to the level of demand for several periods ago.
Problem Formulation
Based on the background that have been described above, then the formulation of research problem is how to predict the production of steam for the next period of RB 13.
Data Collection
The data collected in this study i.e. the steam production data on RB 13 Indah Kiat Pulp and Paper Tbk period in 2016-2017.
Data Processing
The next step after data collection is doing data processing. Data that is processed is the steam production data in PT Indah Kiat Pulp and Paper Tbk period in 2016-2017 for RB 13. Forecasting the production done for 12 years in the fore period 2018 using the method of forecasting i.e. methods quadratis, exponential, and linear. The third method is based on the chosen method that has the smallest error value.
Analysis
After doing the calculation, then conducted an analysis of the results of forecasting production steam PT Indah Kiat Pulp and Paper Tbk for the period 12 to the fore in the year 2018
Closing
This section presented the conclusions of the calculations and analysis that has been done and suggestions towards the problem of forecasting steam PT Indah Kiat Pulp and Paper Tbk for the period 12 to the fore in the year 2018
The following flowchart of research methodology conducted:
Picture 3.1 Research Metodology Flowchart
3.4 Case Resolution
This section aims to resolve the problems that have been formulated in the previous section, here will be explained the steps in problem solving.
3.4.1 Collection Data
The data collected in the research that is in the form of general steam on RB 13 PT Indah Kiat Pulp and Paper Tbk period in 2016-2017. To do the actual production data needed forecasting from 2 years earlier. The production data can be seen as follows:
Tabel 3.1 Actual Data From Two Years Ago
3.4.2 Data Processing
The data processing is done doing forecasting the production of steam for 12 priode fore by using exponential method method, quadratic method and linear methods. The forecasting method is chosen based on the most minor error value. Error calculation is done using the method of the Mean Absolute Percent Error (MAPE). Processing of data is useful for predicting the production of steam on RB 13 PT Indah Tips Pulp and Paper Tbk Perawang for 12 periods in the future.
3.4.2.1 Calculation of the Forecast production of Steam in PT Indah Tips Pulp and Paper Tbk Perawang Fore Period for 12
Calculation of the forecast was the foundation for the company to plan the budget of the company to be next period. Forecasting is done for the next period with 12 methods compared to exponential method, quadratic method and the method of linear. The following calculation is forecasting production of general steam.
3.4.2.1.1 Exponential Method
Forecasting is done by using the method of exponential can be seen in Table 3.2 as follows:
Table 3.2 Parameters used in the exponential Method
Forecasting function with exponential method belows here:
Based on the calculations made, the obtained forecasting function values above as follows:
a = 153900,6485 b = 0,007328954
ln a = 11,94406 y’ = a x e b(t)
Example:
Period 1
y’ = 153900,6485 x e 0,00733(t)
= 153900,6485 x e 0,00733(1)
= 161574,52
Period 2
y’ = 153900,6485 x e 0,00733(t)
= 153900,6485 x e 0,00733(2)
= 162334,22
Picture 3.2 Graph Results Forecasting Exponential Method
3.4.2.1.2 Quadratic Method
Forecasting is done by using the quadratic method can be seen in Table 3.3:
Table 3.3 Parameters used in the quadratic Method
Forecasting function with this method is as follows:
Based on the calculations made, the obtained forecasting function values above as follows:
dt= -18302480
d = -18347964
?= -323659204
? = -690000
?= -27600
b = 3872,44
c = -128,31
a = 149111,33
y'(dt’) = a + (b (t)) + (c)(t2)
Example :
Period 1
y’ = 149111,33 + (3872,44 (t)) + (-128,31)(t2)
= 149111,33+ (3872,44 (t)) + (-128,31)(12)
= 152984
Period 2
y’ = 149111,33+ (3872,44 (t)) + (-128,31)(t2)
= 149111,33+ (3872,44 (t)) + (-128,31)(22)
= 156856
Picture 3.3 Forcasting Graph Result with Quadratic Method
3.4.2.1.3 Linier Method
Forecasting is done by using the linear method can be seen in Table 3.4 below:
Tabel 3.4 Parameters used in the linear Method
Forecasting function with this method are:
y^’=a+bt+ct^2
Example :
Period 1
y’ = 163011,19 + 66478 t + -323659229,06t^2
= 163011,19 + 66478 t + -323659229,06(1)^2
= 163675,97
Period 2
y’ = 163011,19 + 66478 t + -323659229,06t^2
= 163011,19 + 66478 t + -323659229,06?(2)?^2
= 164340,75
Gambar 3.4 Forcasting Graph Result with Linier Method
3.4.2.2 Calculation The Error of Forecasting
Trial calculation error forecasting is done to look for the best method of forecast results. Forecasting error method used was the Mean Absolute Percentage Error (MAPE). The results of a calculation error value from the recap of the third method of divination used can be seen in Table 3.4
Table 3.4 Recap of The Calculation The Error of Forecasting Method
Based on the results of the calculation of the value of the error with the third method, then the method with a value of error obtained the least linear method. This method was chosen to be implemented in forecasting the production of general steam for 12 periods in future.
3.4.2.3 Verify The Selected Forecasting Method
Verification is done to assess the data requests that are in the upper limit and lower limit. Data verification is done so that there is no extreme data on the data to be processed.
Tabel 3.5 Verification Method Of Linear Forecasting
Example :
1. Moving Range of period 2 = absolut (z 2 – z 1) = 159047,03 – (110639)
= 4807,72
2. Moving Range of period 3 = absolut (z 3 – z 2) = 175781 – (159047,03)
= 16734,72
Average of MR = 26029,15 /(24-1)
= 11317,09
UCL = 2,66 x Average of MR = 2,66 x 11317,09
= 30103,47
LCL = -2,66 x Average of MR = -2,66 x 11317,09
= -30103,47
Based on the value of the UCL and LCL, can be seen there is data out UCL and LCL as shown in the following figure:
Gambar 3.5 The Moving Range of Forecasting with Linear Method
3.4.2.4 Forecasting the production of Steam for PT Indah Kiat Pulp and Paper Tbk Perawang Fore Period for 12
Based on the method of forecasting the Linear method, it can be predicted that steam production on RB 13 PT Indah Kiat Pulp and Paper Tbk Perawang to 12 period ahead as shown in Table 3.6:
Table 3.6 Forecasting The Production Of Steam For 12 Periods In The Future by Using The Selected Forecasting Methods (Linear Method)
Based on the method of forecasting the Linear method, it can be foreseen production Steam on RB 13 PT Indah Kiat Pulp and Paper Tbk Perawang to 12 period ahead as shown in Table 3.7:
Table 3.7 Forecasting the production of Steam for 12 Periods in the future by using the Peralaman Method (Linear Method)
Based on the results of the forecasting, then obtained a recapitulation number of Steam production at RB 13 PT Indah Kiat Pulp and Paper Tbk Perawang for the period January – December 2018.
3.5 Analysis
Based on the data processing has been done, then it can be analysed things as follows:
3.5.1 Data Analysis of Production General Steam and The Use of Utilities.
The actual data of the production of Steam from the Vitim RB 13 PT Indah Kiat Perawang Pulp and Paper. The data provided in the form of actual data in 2016-2017.
3.5.2 The Analysis Of Forecasting Production Of Steam
Forecasting is done for 12 period ahead between January-December 2018. Forecasting is done by using the time series forecasting method consists of a method of exponential, quadratic and linear. The selection of methods of forecasting is done because this method can find and measure of several factors (independent) are important, along with the influence of the variables is not free which will be foreseen in the form of the equation of the line inclination for an equation.
After forecasting with that method, then do the calculation error from the third method. The method of calculation used is MAPE (Mean Absolute Percentage Error). The third method of the retrieved value is the smallest MAPE i.e. Linear methods, which have a value of MAPE 6% this result shows that a linear method can be chosen as the best method for predicting the production of steam for 12 periods ahead.
Based on the results of forecasting for 12 period ahead we can see that the amount of steam production in the next year has increased from period to period.
3.6. Closing
This section contains conclusions from data processing and analysis that has been done as well as suggestions for the company as well as for further research.
3.6.1 Conclution
Based on the results obtained from the calculations that have been done, it can be concluded that forecasting is done against the production of steam and utility usage on RB 13 has increased to 12 the next period in the year 2018. The calculation result obtained is used as a consideration for determining a target production of steam according to the needs.
3.6.2 Advise
Based on the calculations and analysis that has been done then advice on reporting the work practices is a product should be the object of forecasting not only products but all 13 RB RB that is in PT Indah Kiat Pulp and Paper Tbk Perawang.
