Artificial Neural Network - Water Quality Management: Assessment and Interpretation - page 75

Chapter 5
Artificial Neural Network

Abstract Often in water quality management, understanding the relationship
between input and output data might be a complicated process. In this situation Data
Driven Models using information and collected data (input data)

find out the rela-

tionship between inputs and outputs. In this regard, Arti cial Neural Network (ANN)

is one of the Data Driven Models which has recently been applied as a tool for
modeling complicated processes. In this chapter, after reviewing the developing
process of ANN in water quality management, the theory of the ANN is mentioned in
detail for both static and dynamic methods. Data preparation, learning rate and model
ef ciency including selection of number of neurons in hiding layer which has a

minimum error in learning rate and network ef ciency is described in detail. At the

end step, as a case study water quality of Zaribar Lake located in the Northwestern part
of Iran, using Multilayer Perceptron (MLP) neural network method are described.

5.1 Introduction

Lack of water resources and optimum management has been two recent challenges
of water resources engineering. Population growth, decline of useable water
resources, improvements in lifestyle, growing rate of consumption, climate change
and several other parameters have caused useable water to be a noteworthy problem
for the future. Economic and ef

cient use of water resources and its management
have an increasingly important role.
Other challenges which water quality managements and environmental engi-
neers are facing are controlling the nutrients released to surface waters. Despite all
efforts, eutrophication is also other major problems with water quality management.
Eutrophication is de

ned as a cultural or accelerated enrichment of nutrient in lakes,
rivers, estuaries and marine waters in which the natural eutrophication process has
gone forward by hundreds or more years of human activities that add nutrients
(Burkholder 2000 ). Two important parameters which cause eutrophication are
phosphor and nitrogen. Analyzing sample data of Phosphor and Nitrogen param-
eters, in surface water is necessary for understanding the eutrophication situation