Chapter 3

HYDRO THERMAL SCHEDULING

Contents
Problems definition Mathematical model of long and short term problems Dynamic and incremental dynamic programming Methods of local variation Hydro thermal system with pumped hydro units Solution by local variation Treating Pumped Hydro unit for load management and spinning reserve

Ethiopian power system

Generation potential
Type of energy source Hydropower Wind Geothermal Potential Above 4,000MW Above 1,350,000MW Above 5000MW

Supply system
ICS supplied from hydro power plants SCS mini hydro plants and diesel generators

Generation trend
Generation trend is increasing

Electrification status
According to EEPCO report 48.3%

Installed capacity

Installed capacity

Percentage contributions
Largest portion is supplied from hydro 89.8%

hydro diesel

geothermal
wind

Transmission system
Total length 11,796.32Km with 5 voltage levels

There are 143 substations with 127 supplying distribution systems

Distribution system
Distribution system consists of 148745.5km and 18,888 transformers

Customers

Energy sales
Total number of customers by tariff group is as shown below

Historic energy sales

Total energy soled

Problem definition
Generation scheduling
Determining optimal operation strategy for the next scheduling period subject to a variety of constraints
Limited energy storage capability of water reservoirs Stochastic nature of availability of water

Can be broken down into sub problems

Hydrothermal coordination (yearly, monthly and weekly ) Unit commitment (weekly or daily) Economic dispatch (hourly)

Problem definition
Generation planning problem is
Nonlinear optimization problem With continuous and discrete variable s Non explicit objective function With equality and inequality constraints

Problem formulation
Hydro thermal coordination (HCP)
Determining optimal amount of hydro and thermal generation to be used in a scheduling period Can be classified into short, mid and long term scheduling Constraints
Irrigation Recreation Effect on downstream users

Problem formulation
Hydrothermal scheduling problem
Long range
This involves
Long range forecasting of water availability Scheduling of water releases

Affected by
assumption used to calculate water replacement (Use statistical mean or Worst case ) Load, hydraulic inflows, unit availabilities

Medium range Short range

Mathematical model of long and short term problems

Mid term hydro scheduling (1week to 1 day)
This involves
Long range forecasting of water availability Scheduling of water releases

Affected by
assumption used to calculate water replacement
Use statistical mean Worst case

Load, hydraulic inflows, unit availabilities

Problem formulation
Short range ( 1 day to 1 week)
Hour by hour scheduling of all generators to achieve minimum production cost Load, hydraulic inflow and unit availability is known

Operational factors affecting hydrothermal scheduling

Type of turbine used
Francis turbine may operate at 65 to 125% of rated net head

Loading
At light unit loading, efficiency may drop to 70% and at full load may rise to 87%

Hence
unit loading must be at best efficiency gate position Water releases schedule must coordinate with inflow

Note: better one full than two halves

Types of hydrothermal scheduling

All hydro
Difficult to meet the demands and operate economically Simulate the water system Find a schedule which leaves maximum stored energy in reservoir

Hydro dominated
Schedule the system for minimum cost of thermal

Balanced or with most thermal

Minimize thermal production costs

Scheduling of mostly hydro case

Problem formulation
One hydro and one thermal Rated hydro power is enough to cover load but available hydro energy is not enough Use thermal energy to fill the gap

Problem formulation
Symbols used
EH- total available hydro energy Time period to use all energy is TMAX

Hence

Where PHj is hydro generation limit at time j

Problem formulation
ES=total thermal energy used

PSj- thermal generation level at time j TS- total thermal run time

Problem formulation
Total energy used by the load

Where PLj is load at time j The thermal energy used is then

Problem formulation
Then the problem statement is to minimize the cost of generation of Es

Solution using Lagrangian method

Form the Lagrangian

Apply necessary condition

Solution using Lagrangian Method

Solving the two equations, from the first one

Which means the thermal must run at constant generation level for all time If the generation level is assumed to be PS*

Solution by Lagangian Method

Solution by Lagrangian Method

Substituting into the second equation

(*)

If the cost function of the thermal unit is given bye

Solution by Lagrangian Method

The total cost of running the thermal unit is then Substituting from equation (*) above

Now we have nonlinear unconstrained optimization with single variable

Solution by Lagrangian Method

Solving for Ps

This means the thermal unit must be operated at the maximum efficiency point for as long as it produced the demanded energy Es

Example 7A page 217

Given the following characteristics for the hydro and thermal plants, they have to supply a load of 90MW for a week (15120MWhr)

Case I- if the hydro is limited to 10, 000MWh energy, solve the run time of the thermal Required energy from thermal is 5120MWhr

Example 7A
The steam plants maximum efficiency point is at 50MW, hence it has to run for 102.4hr Hence
Steam run at 50MW for the first 102.4 hr of the week Hydro runs at 40MW for the first 102.4hr and then for the rest 65.6hr, it runs at 90MW

Example 7A
Case II- if the limit is on the amount of water drawn from the reservoir 250,000 acre-fit The water drawn from the reservoir is determined by the amount of power generated

Giving Ts=36.27h

Short term hydrothermal scheduling

Problem definition
Given a hydro and thermal unit supplying a load and constraints of total water discharge, starting volume, ending volume etc, find a schedule which minimizes the total cost

Short term hydrothermal scheduling

Problem formulation

Short term hydrothermal scheduling

Forming the Lagangian

Applying the necessary conditions

Short term hydrothermal scheduling

Example 7B
Solve example 7B

Assignment
Explain the idea of pumped hydro systems List some of world famous pumped hydro systems and explain how they are used, i.e
Load management Spinning reserve

Discuss the hydrothermal scheduling of generation units which are in series cascad like that of Gibe dam units

Dynamic programming
Is a systematic way of determining optimal solution in a problem having multiple stages and with many possible decisions at each stage Is developed by Bellman in 1957 Has nothing to do with computer programming

Dynamic programming
Consider the following graph, numbers correspond to time to travel between nodes
Find a path from A to B, travelling on to the right, such that the total sum of numbers is minimum

DP algorithm
Forward DP- start with A and find the cost of moving forward until B Backward DP start with B and find backward the optimal path

In both cases, we have to decisions at each stages

DP algorithm
There are 20 possible routes The optimal path can be computed using DP as follows
Start at B and move one step back Find the minimum path to arrive at B from one step back Repeat this for each stage Stop when arriving at A

DP algorithm
Solution

DP used for hydrothermal scheduling

Consider the following situation where a single hydro and single thermal units supply power to a load
Total scheduling time Tmax is divided into time periods j Given: loads at time periods, starting and ending volume Vo and Vmax

DP for hydrothermal scheduling

Terms used are

DP for hydrothermal scheduling

Assumptions
The thermal unit is on for the whole time period

Water use rate of hydro is given by

Maximum flow rate qmax is limited corresponding to the maximum power output of the hydro

DP for hydrothermal scheduling

The volume in storage changes as

For two consecutive intervals Vj and Vj-1

Problem is find the volume at each stage so that the total cost is minimum

DP for hydrothermal scheduling

DP algorithm

DP hydrothermal scheduling Example

A thermal unit and hydro unit have the following characteristics curves
F 700 4.8Ps Ps2 2000 and 200 Ps 1200

DP hydrothermal scheduling Example

Initial and final volumes are 10,000m3 and storage volume limits are 6000 and 18000m3 The scheduling period is 24h divided by 4h intervals

DP for hydrothermal scheduling Example

If the storage volume steps are 2000m3
Discharge at any time interval j is given by

DP for hydrothermal scheduling Example

The first and second stage computations becomes

DP example

Linear programming
Is a mathematical method used to allocate scarce resources to competing alternatives in an optimal manner Linear objective function and Constraints are linear inequality

Example Linear programming

Consider the following profit function Subject to the constraints

Example LP
Solution methods
Graphical method for two dimensional case
Draw the graph of constraints The closed boundary formed from constraints is the feasible solution region Corner points of the boundary are max/min points Check for these points

Simplex method for multi variable case

Hydrothermal scheduling with LP

To apply LP
Objective function has to be linear
Divide the cost function of the steam turbine into sections and approximate it by linear function with known slope Hydro turbine output versus the discharge has to be approximated by linear

Constraints have to be linear

Approximate water

Hydrothermal scheduling using LP

For thermal characteristics, if 3 steps are used

For the hydro

Hydrothermal scheduling using LP

Thermal and hydro linear approximation

Hydrothermal scheduling using LP

Spill characteristics approximated by linear
Water spills out if volume is above a limit

Hydrothermal scheduling using LP

Problem formulation
Minimize the linear cost function

Subject to