CHAPTER ONE

LOAD FORECASTING

1. INTRODUCTION
A Power system planning is a Process in which the aim is to decide on a new as well as upgrading
existing system elements to adequately satisfy the loads for a foreseen future. Elements can be:

 Generation facilities

 Substations

 Transmission lines and/ cables

 Capacitors Etc…..

 decisions can be:

 Where to allocate the element(for instance, the sending and receiving end of the line)

 When to install the element(for instance 2022)

 What to select, in terms of element specification (for instance, no of bundles and

conductor type)

 The load should be adequately satisfied

 Forecasting refers to the prediction of the load behavior for the future

 Words such as demand and consumption are also used instead of electric load

 Energy(MWh, KWh)and power(MW.KW) are the two basic parameters of the load:

 By load we mean the power

 Demand forecast

 To determine the capacity of generation, transmission and distribution required

 Energy forecast

 To determine the type of generation facilities required

 Load forecasting plays an important role in power system planning, operation and control.

 A good forecast reflecting current and future trends, tempered with good judgment, is the key to
all planning, and of financial success.

 Load forecasts can be divided into three categories:

 short-term forecasts which are usually from one hour to one week,
  medium forecasts which are usually from a week to a year, and

 Long-term forecasts which are longer than a year.

1. 1 Load forecasting factors

 Several factors affect Load Forecasting. Among them we mention:

1. Land use: (Residential, industrial, commercial, agricultural . . . etc.).

Different types of use affect the capacity of the substation, i.e. residential loads is different from
industrial loads.

2. Population growth:
As the population increases more loads are needed.

3. Historical Data:
Historical data plays an important role in forecasting since they can tell how the load will behave in the
future.

4. Load Densities (KVA/Km 2 ) :

 Load density must be put into account during load forecasting, we must consider the range of
enmities as it differs for different types of loads.

 Common figures for load densities are:

1000 KVA/Km2 for agricultural areas

3000 KVA/Km2 for residential areas

5000 KVA/Km2 for city center and

10,000 KVA/Km2 for industrial areas.

1.2 Load patterns

Types of Patterns :

• By the word pattern we mean how the load changes with time.

• Four basic types of patterns often exist in data series:

(1) Horizontal pattern: This exists when there is no trend in a data series. This can happen when there
is no more expected load or an increase in the area. Such pattern is generally referred to as stationary.

Example of this load pattern exists in parts of large cities where there is no space to build more
buildings on.

(2) Trend pattern:

This exists when there is an increase or decrease in the value of electric load consumption.

Such pattern is generally referred to as non-stationary.

(3) Seasonal pattern:

This exists when data series fluctuates according to some seasonal factors.

(4) Cyclical pattern:

This exists when data series fluctuates and doesn't repeat itself at constant time interval. One of
the factors that may cause the pattern to be so is an economic crisis in the state which may cause to stop
some industries in the state causing a dip in the load curve.

1.3 Load forecasting methods

 Forecasting is simply a systematic procedure for quantitatively defining future loads.

 Forecasting techniques may be divided into 3 broad classes:

 extrapolation

 correlation or

 Combination of both.

 Techniques may be further classified as deterministic, probabilistic, or stochastic.

(1) Extrapolation:

Extrapolation techniques involve fitting trend curves to basic historical data adjusted to reflect
the growth. It produces reasonable results in many cases.

• Such as Simple Linear Regression technique is to be classified as a deterministic extrapolation,

since no attempt is made to account for random errors in the data or in the analytical model.

• Some standard analytical functions are used in trend curves fitting, including:

1. Straight line

Y=a+bx

2. Parabola

Y = a + b x + cx2

• The most common curve - fitting technique for finding coefficients of function in a given
forecast is the method of least squares as will be discussed later.

(2) Correlation :

 Correlation techniques are used to relate system loads to various demographic and economic
factors.
  This approach has an advantage of forcing the forecast to understand clearly the
interrelationship between load growth patterns and other measurable factors.

 The most obvious disadvantage, however results from the need to forecast demographic and
economic factors, which can be more difficult than forecasting system load.

 Typically, these factors may be population, employment, building permits, weather data
and business indicators.

Calculation method

(1) Simple Regression :

Regression in general is a relationship between the variable we want to forecast (dependent) and
another variable (independent). Or, we can say

Y = f (x)

 If the independent variable is time, then we call it simple time-series regression, and simple
refers to a single independent variable.

In a simple regression the relationship is assumed linear, i.e

Ŷ=a+bt

 The principle of regression theory is that, any function

Ŷ = f (x) can be fitted to a set of data points so as to minimize the sum of errors squared at each data
point and this type of fitting is called least square fit in which the objective is to minimize:

Where n is the number of data points.

(2) Linear Regression :

Calculation method
Examples

Practical Example:

• The load forecasting of a new area is to be considered. This area consists of several zones.

(1)Agriculture.

(2) Residential.

(3) City center.

(4) Light industrial.

(5) Heavy industrial.

• The area under study is given below:

Example 1 linear model

Examples 2. quadratic model
• Example 3.
• From Least Squared Error Criteria use Quadratic model, and find the load for 2017