A Comparative Analysis of Forecasting Reservoir Inflow using ARMA Model & Holt Winters Exponential Smoothening Technique

Hits: 807
Select Volume / Issue:
Year:
2016
Type of Publication:
Article
Keywords:
Auto Regression Moving Average, Exponential Smoothening, Krishnagiri Reservoir, Reservoir Inflow
Authors:
Naresh Vijayakumar; Soorya Vennila
Journal:
IJISM
Volume:
4
Number:
2
Pages:
85-90
Month:
March
ISSN:
2347-9051
Abstract:
Statistical analysis in hydrology includes estimating the future frequency or probability of hydrologic events based on the information contained in the past hydrologic records. The present study aims at applying time series analysis using Auto Regression Moving Average (ARMA) model and Holt Winters exponential smoothing technique for reservoir inflow forecast enabling better management of reservoir inflow. The Krishnagiri Reservoir in the state of Tamilnadu, India is selected as the focus area for this study. Using Box Jenkins approach, the model identification for ARMA model was done by Auto Correlation function plot (ACF) and Partial Auto Correlation Function Plot (PACF); then the parameters for the model are estimated by Maximum likelihood estimation method followed by diagnostic checking of estimated value of the model. ARMA (2, 4) shows better statistical parameters of generated annual reservoir inflow. One of the deficiencies of ARMA model is that it is not able to reproduce in a satisfactory manner the trend, seasonality, and periodicity component of a time series. In the present study, that deficiency was overcome by using Holt Winters Exponential Smoothening technique. The statistical parameters of generated reservoir inflow using exponential smoothening replicates the historical observed reservoir inflow with R2= 0.9659 and the percentage of mean error to 19.3%. Although ARMA model is used as the best fit model for any time series analysis in hydrology, exponential smoothening technique gives a better matching between observed and computed values.

About Us

Our Journals

IJECCE
International Journal of Electronics Communication and Computer Engineering
ISSN(Online): 2249 - 071X
ISSN (Print) : 2278 – 4209
www.ijecce.org
Submissions open
IJAIR
International Journal of Agriculture Innovations and Research (ISSN(Online) : 2319 – 1473)
www.ijair.org
Submissions open
IJISM
International Journal of Innovation in Science and Mathematics
ISSN : 2347 – 9051
www.ijism.org
Submissions open
IJEIR
International Journal of Engineering Innovations and Research
ISSN(Online) : 2277 – 5668
www.ijeir.org
Submissions are open.

IJAIM
International Journal of Artificial Intelligence and Mechatronics (ISSN(Online) : 2320 – 5121)
www.ijaim.org
Submissions open
IJRAS
International Journal of Research in Agricultural Sciences
www.ijras.org
Submissions open

Indexed By:

 width= width= width= width= width=