Peak Load Prediction Using Fuzzy Logic For The 150 kV Sulselrabar System

. Prediction of electrical load on 150 kV Sulselrabar electrical system, analyzed using approach at night peak load using Fuzzy Logic based intelligent method. The peak load characteristics are certainly different from the load in normal time, therefore a special approach is needed to predict the peak night load. As input data will be used data of night peak load in 2010 until 2015, on the same day and date, each 4 days before day-H or day date which will be predicted load. For the data processing stage is divided into several stages, namely pre-processing, processing, and post-processing. The load data processing follows several procedures, ie computing WDmax, LDmax, TLDmax and VLDmax each year. Data processing is processed using excel software and then using Matlab software to run Fuzzy Logic. From the analysis results obtained, Error Prediction The peak evening load is very small that is equal to - 0.039035754%. As comparison data used actual day-H data is April 2016. The graph of analysis result also shown in this paper.


Introduction
Sulselrabar electrical system studies previously been widely done, in order to review the electrical system sulselrabar [1][2][3][4][5][6][7][8][9][10][11][12].Forecasting burden is one of the studies that need to be done in sulselrabar system, because it is growing.In this research will be conducted a study of electrical load prediction on sulselrabar system using intelligent fuzzy logic method.Previously smart methods have been widely applied to power system optimization, one of them forecasting the electrical load.Some research related techniques for forecasting the load conventionally to use smart methods such as, Regression [13], Support Machine [14], Fuzzy Logic [15,16], Genetic Algorithm [17,18], Neural Network [19,20], Firefly [21].
Load forecasting is a study conducted in order to anticipate the development of load and operating system strategy in the future and also in evaluating the current system.Some of the above research shows that the application of load forecasting technique in sulselrabar electrical system still has not been done, and only using conventional method so that the resulting error is still relatively big.In this research we will use load data several years ago as input data for fuzzy logic.Fuzzy Logic is one of the smart methods that have been widely used in various fields.Therefore the authors use this method as a media forecasting techniques on the electrical load system Sulselrabar.p-ISSN: 2540-9433; e-ISSN: 2540-9824

Load Forecasting
Forecasting is a phenomenon of the calculation or estimation of measurements in the period of time to come.In the operation of power system, load forecasting problem is a very important problem in the company.Both in terms of management and in operations, so the load forecasting has special attention.The time load forecasting is divided into several groups: 1.Long Term Load Forecasting is forecasting the load for a period of over five years.For long term load forecasting is commonly used for planning and development of a system (planning), the highest peak load on a power system is often used as a reference in system development.In addition, external factors such as macroeconomic factors also determine in forecasting long-term expenses.2. Medium load forecasting (medium term load forecasting) is forecasting the load for a period of one month to five years.Forecasting medium-term expenses can not be separated from long-term forecasting of loads, so that long-term forecasting costs will not be much distorted from long-term forecasting expenses.In mid-term forecasting is used for the operational aspects of power systems such as the capacity of the Circuit Breaker (CB) or Transformator panel capacity, expanding the distribution network so that not much is done in medium-term forecasting loads.3. Forecasting the short term load (for example loading forecasting) is forecasting the load for a period of several hours to a week.In short-term forecasting for forecasts there is a maximum load limit (Pmax) and a minimum load limit (Pmin) determined by medium-term forecasting loads.Short-term load forecasting is most widely used for the operation of a power system 4. Very short term forecasting forecasting (for example load forecasting for less than one hour (hour, minute, second).Very short term load forecasting is used for some special cases.

Fuzzy Logic
Fuzzy logic was first introduced by professor Zadeh (California University) in 1965 by describing the mathematical calculations based on set theory to describe vagueness in the form of linguistic variables, in other words fuzzy logic theory developed the theory of boolean set (0 and 1) into set which has an obscure membership value (between 0 and 1) so that fuzzy logic is also called fuzzy logic.Fuzzy inference is doing reasoning using fuzzy input and fuzzy rules that have been determined so as to produce fuzzy output.The main structure of type-1 fuzzy logic system is as follows:

Fuzzy Logic Implementation for Sulselrabar Load Forecasting
Forecasting the electrical load on the electrical system of 150 kV Sulselrabar analyzed using 24 hours short-term forecasting approach.As data input used data load of electricity year 2010 until 2015, the data obtained from UPB PT.PLN Region Sulselrabar.For the data processing stage is divided into several stages, namely pre-processing, processing, and post-processing.as described in the previous chapter.Software used in this research using Microsoft Excel to process load data and Matlab Software to perform load forecasting optimization using Fuzzy Logic. 4

Calculation of Input
Calculating TLDMax Value (1 Year Before Forecast)

Calculating VLDMax Value (1 Year Before Forecast)
VLD value max 1 year before forecasting year become input variable X, and VLD value max year forecasting become input variable Y and Z.After obtained value of input variable X, Y, Z Fuzzy, next make Fuzzy Logic design (Membership Function & Fuzzy Rule) using M-File and Matlab toolbox.Here's a picture of Fuzzy Logic input-output design.p-ISSN: 2540-9433; e-ISSN: 2540-9824

Fuzzy Rules Design
The set of Fuzzy input variables (X, Y) of VLDmax national holidays for each linguistic degree membership membership fuzzy input variable (X, Y) is mathematically described as follows: p-ISSN: 2540-9433; e-ISSN: 2540-9824

Result and Analysis
For example the result of calculation of input variable Fuzzy Logic, can be seen in table 2 below, by using sample of calculation of Input X variable on load forecasting for load on april 2016.

Discussion
Prediction of electrical load on 150 kV Sulselrabar electrical system, analyzed using approach at night peak load using Fuzzy Logic based intelligent method.The peak load characteristics are certainly different from the load in normal time, therefore a special approach is needed to predict the peak night load.As input data will be used data of night peak load in 2010 until 2015, on the same day and date, each 4 days before day-H or day date which will be predicted load.For the data processing stage is divided into several stages, namely pre-processing, processing, and post-processing.The load data processing follows several procedures, ie computing WDmax, LDmax, TLDmax and VLDmax each year.Data processing is processed using excel software and then using Matlab software to run Fuzzy Logic.From the analysis results obtained, Error Prediction The peak evening load is very small that is equal to -0.039035754%.As comparison data used actual day-H data is April 2016.p-ISSN: 2540-9433; e-ISSN: 2540-9824 Figure 6-9.Shows the prediction results of power load on april 2016.From the results of the analysis of the above electrical load prediction, obtained a very small load forecasting error of -0.039035754%.Based on Fig. 6-9 it can be seen that, the biggest error occurs in forecasting the 5th, 8th, 9th, and 12th load date.
Electric load prediction using Fuzzy Logic's intelligent method of optimization is highly accurate and is recommended for use in long-term forecasting studies.In addition, other input variables for Fuzzy Logic can be added to optimize for more complex electrical load forecasts.As a method development, it also proposed another combination of intelligent algortms with fuzzy logic.

Conclusion
Prediction of load sulselrabar system on april 2016 by using the data burden of 2010-2015 for Fuzzy Logic input, obtained Error forecasting the load is very small that is equal to -0.039035754%.The biggest error occurs in forecasting the 5th, 8th, 9th, and 12th load date.Electric load prediction using Fuzzy Logic's intelligent method of optimization is highly accurate and is recommended for use in long-term load forecasting studies.

Fig. 8 .
Fig. 8. Error Prediction of Peak Night Burden in April 2016

Table 1 .
Processing Results of Load Data

Table 2 .
Results Calculation of Input Variables X, Y, Z April 2016 After calculating input variables X, Y, Z for Fuzzy Logic, optimizing the prediction of electrical load can be done, and the analysis results are shown in the following figure.