To solve for the given generating station, let's calculate each item step by step based on the data

provided:

Data Table Recap:

Time Interval (Hours) Load (MW) Duration (Hours)

0–6 20 6
6–10 25 4
10–12 30 2
12–16 25 4
16–20 35 4
20–24 20 4

1. Load Curve

The Load Curve is a graph showing the variation of load (in MW) with time (hours of the day).
This requires plotting the data points:

 X-axis: Time (hours)

 Y-axis: Load (MW) The curve consists of horizontal segments connecting the load values
over the specified time intervals.

2. Load Duration Curve

The Load Duration Curve is obtained by arranging the load values in descending order and
plotting them against the cumulative hours of operation. Sorting the loads:

 35 MW for 4 hours
 30 MW for 2 hours
 25 MW for 8 hours (4 + 4)
 20 MW for 10 hours (6 + 4)

The cumulative hours and load values form a stepped graph:

 X-axis: Cumulative hours

 Y-axis: Load (MW)

3. Maximum Demand
The Maximum Demand is the highest load during the day: Maximum Demand=35 MW\
text{Maximum Demand} = 35 \, \text{MW}

4. Units Generated per Day

The total energy generated is the sum of energy for each time interval:

Units Generated=∑(Load×Duration)\text{Units Generated} = \sum (\text{Load} \times \

text{Duration}) Units Generated=(20×6)+(25×4)+(30×2)+(25×4)+(35×4)+(20×4)\text{Units
Generated} = (20 \times 6) + (25 \times 4) + (30 \times 2) + (25 \times 4) + (35 \times 4) + (20 \
times 4) Units Generated=120+100+60+100+140+80=600 MWh/day\text{Units Generated} =
120 + 100 + 60 + 100 + 140 + 80 = 600 \, \text{MWh/day}

5. Average Load

The Average Load is the total energy generated divided by the total time:

Average Load=Units GeneratedTime Period\text{Average Load} = \frac{\text{Units

Generated}}{\text{Time Period}} Average Load=60024=25 MW\text{Average Load} = \
frac{600}{24} = 25 \, \text{MW}

6. Load Factor

The Load Factor is the ratio of the average load to the maximum demand:

Load Factor=Average LoadMaximum Demand\text{Load Factor} = \frac{\text{Average Load}}

{\text{Maximum Demand}} Load Factor=2535=0.714 or 71.4%\text{Load Factor} = \frac{25}
{35} = 0.714 \, \text{or} \, 71.4\%

Final Results:

1. Load Curve: A plot of load vs. time (to be drawn).

2. Load Duration Curve: A descending step graph of load vs. cumulative hours (to be
drawn).
3. Maximum Demand: 35 MW35 \, \text{MW}
4. Units Generated per Day: 600 MWh600 \, \text{MWh}
5. Average Load: 25 MW25 \, \text{MW}
6. Load Factor: 71.4%71.4\%

Let's calculate each item step by step based on the given data:
Data Table:

Group Demand (kW) Time Interval Duration (Hours)

Group A 200 8 AM to 6 PM 10
Group B 100 6 AM to 10 AM 4
Group C 50 6 AM to 10 AM 4
Group D 100 10 AM to 6 PM, 6 PM to 6 AM 16

1. Load Curve

The Load Curve represents the variation of total demand over time. For this, we calculate the
total load in each time interval:

 6 AM to 8 AM: Group B (100 kW) + Group C (50 kW) + Group D (100 kW) = 250 kW
 8 AM to 10 AM: Group A (200 kW) + Group B (100 kW) + Group C (50 kW) + Group
D (100 kW) = 450 kW
 10 AM to 6 PM: Group A (200 kW) + Group D (100 kW) = 300 kW
 6 PM to 6 AM: Group D (100 kW) = 100 kW

2. Load Duration Curve

The Load Duration Curve is obtained by arranging the loads in descending order:

 Load of 450 kW for 2 hours

 Load of 300 kW for 8 hours
 Load of 250 kW for 2 hours
 Load of 100 kW for 12 hours

3. Maximum Demand

The Maximum Demand is the highest load during the day:

Maximum Demand=450 kW\text{Maximum Demand} = 450 \, \text{kW}

4. Units Generated per Day

The total energy generated is the sum of energy consumed in each interval:
Units Generated=∑(Load×Duration)\text{Units Generated} = \sum (\text{Load} \times \
text{Duration}) Units Generated=(250×2)+(450×2)+(300×8)+(100×12)\text{Units Generated} =
(250 \times 2) + (450 \times 2) + (300 \times 8) + (100 \times 12)
Units Generated=500+900+2400+1200=5000 kWh/day\text{Units Generated} = 500 + 900 +
2400 + 1200 = 5000 \, \text{kWh/day}

5. Average Load

The Average Load is the total energy generated divided by the total time:

Average Load=Units GeneratedTime Period\text{Average Load} = \frac{\text{Units

Generated}}{\text{Time Period}} Average Load=500024≈208.33 kW\text{Average Load} = \
frac{5000}{24} \approx 208.33 \, \text{kW}

6. Load Factor

The Load Factor is the ratio of the average load to the maximum demand:

Load Factor=Average LoadMaximum Demand\text{Load Factor} = \frac{\text{Average Load}}

{\text{Maximum Demand}} Load Factor=208.33450≈0.463 or 46.3%\text{Load Factor} = \
frac{208.33}{450} \approx 0.463 \, \text{or} \, 46.3\%

7. Diversity Factor

The Diversity Factor is the ratio of the sum of individual maximum demands to the maximum
demand of the station:

Diversity Factor=Sum of Individual Maximum DemandsMaximum Demand of the Station\

text{Diversity Factor} = \frac{\text{Sum of Individual Maximum Demands}}{\text{Maximum
Demand of the Station}}

Sum of individual maximum demands:

 Group A: 200 kW200 \, \text{kW}

 Group B: 100 kW100 \, \text{kW}
 Group C: 50 kW50 \, \text{kW}
 Group D: 100 kW100 \, \text{kW}

Sum of Individual Maximum Demands=200+100+50+100=450 kW\text{Sum of Individual

Maximum Demands} = 200 + 100 + 50 + 100 = 450 \, \text{kW} Diversity Factor=450450=1.0\
text{Diversity Factor} = \frac{450}{450} = 1.0
Final Results:

1. Load Curve: Variation of load vs. time (to be plotted).

2. Load Duration Curve: Stepped descending plot (to be plotted).
3. Maximum Demand: 450 kW450 \, \text{kW}
4. Units Generated per Day: 5000 kWh5000 \, \text{kWh}
5. Average Load: 208.33 kW208.33 \, \text{kW}
6. Load Factor: 46.3%46.3\%
7. Diversity Factor: 1.01.0

Here are the explanations of the mentioned terms related to power system reserves:

i. Installed Reserves

 Definition:
The installed reserves refer to the total capacity of all installed generating units in a
power system, minus the peak load demand of the system.
It represents the excess generating capacity available beyond the maximum demand.
 Purpose:
Installed reserves ensure that the system can handle unexpected increases in demand or
the unavailability of some generating units due to maintenance or outages.
 Formula:

Installed Reserves=Total Installed Capacity−Maximum Demand\text{Installed Reserves} = \

text{Total Installed Capacity} - \text{Maximum Demand}

ii. Spinning Reserves

 Definition:
The spinning reserves are the portion of generating capacity that is online and
synchronized to the grid but not fully loaded. These reserves can respond immediately to
fluctuations in power demand or the loss of a generating unit.
 Purpose:
Spinning reserves provide instantaneous support to stabilize the grid in case of sudden
changes, such as a generator tripping or a sharp increase in load.
 Characteristics:
o These reserves are already connected to the grid (hence "spinning").
o They can ramp up their output within seconds to minutes.

iii. Cold Reserves

  Definition:
The cold reserves refer to generating units that are not currently operational but are
maintained in a state of readiness to be brought online when needed. These reserves are
offline and may require hours or even days to start generating power.
 Purpose:
Cold reserves act as a backup for long-term emergencies, such as sustained high demand
or extended outages of multiple generating units.
 Characteristics:
o Not immediately available.
o Typically consist of units kept offline for economic or operational reasons (e.g.,
old or inefficient plants).

iv. Hot Reserves

 Definition:
The hot reserves are generating units that are not actively generating power but are kept
in a state where they can quickly be brought online. These reserves are closer to
operational status compared to cold reserves and can begin producing power within a few
minutes to hours.
 Purpose:
Hot reserves provide faster backup than cold reserves and are used to supplement
spinning reserves during medium-term demand spikes or outages.
 Characteristics:
o Typically consist of plants operating at idle or standby mode.
o Faster response time than cold reserves but slower than spinning reserves.

Comparison of Reserves:

Type Operational Status Response Time Purpose

Installed To meet peak demand +
Total system capacity Long-term planning
Reserves outages
Spinning Online and Instantaneous Grid stability and sudden
Reserves synchronized (seconds) demand
Backup for long-term
Cold Reserves Offline Hours to days
emergencies
Backup for medium-term
Hot Reserves Near-operational Minutes to hours
demand

Load Forecasting:
Load forecasting refers to the process of predicting future electricity demand (load) based on
historical data, weather patterns, economic indicators, and other influencing factors. It is
essential for efficient operation, planning, and management of power systems, as it helps utilities
to ensure reliable supply, optimize generation, reduce costs, and maintain system stability.

Types of Load Forecasting:

Load forecasting can be classified into three main categories based on the time horizon:

1. Short-Term Load Forecasting (STLF):

 Definition:
Short-term load forecasting predicts the electrical load for a period ranging from a few minutes
to a few days (typically up to a week).
 Purpose:
Used for daily or hourly operational decisions such as:
o Scheduling generation units.
o Managing spinning reserves.
o Optimizing energy dispatch.
 Factors Considered:
o Weather conditions (e.g., temperature, humidity, wind speed).
o Time of day and day of the week.
o Seasonal variations.
 Techniques:
o Machine learning models.
o Time-series analysis (e.g., ARIMA).
o Regression models.

2. Medium-Term Load Forecasting (MTLF):

 Definition:
Medium-term load forecasting predicts the electrical load for a period ranging from a few weeks
to a few months (up to a year).
 Purpose:
Used for mid-level planning such as:
o Maintenance scheduling of generation and transmission systems.
o Fuel procurement planning.
o Determining energy trading strategies.
 Factors Considered:
o Historical load trends.
o Economic and demographic changes.
o Seasonal load variations.
 Techniques:
 o Statistical models.
o Econometric models.
o Neural networks.

3. Long-Term Load Forecasting (LTLF):

 Definition:
Long-term load forecasting predicts the electrical load for a period ranging from a year to
several decades.
 Purpose:
Used for strategic planning such as:
o Infrastructure expansion (new generation plants, transmission lines).
o Investment decisions.
o Policy formulation for energy supply and demand.
 Factors Considered:
o Population growth.
o Economic development.
o Technological advancements.
o Policy changes (e.g., renewable energy targets).
 Techniques:
o Scenario-based models.
o Econometric models.
o System dynamics modeling.

Comparison of Load Forecasting Types:

Aspect Short-Term Medium-Term Long-Term

Time Horizon Minutes to a week Weeks to a year Years to decades

Purpose Operational decisions Maintenance and trading Infrastructure planning

Factors Considered Weather, time Historical trends Economic and policy

Accuracy Requirement High Moderate Lower (scenario-based)

Applications of Load Forecasting:

1. Utility Operations: Helps in load scheduling, demand-side management, and peak load shaving.
2. Cost Optimization: Reduces generation costs by optimizing the mix of energy resources.
3. Grid Reliability: Prevents overloads and ensures system stability.
 4. Energy Market: Enables effective energy trading in competitive markets.

Automatic Load Frequency Control (ALFC)

Automatic Load Frequency Control (ALFC) is a critical mechanism in power system operation
that maintains system frequency and tie-line power flows at their desired values. It achieves this
by automatically adjusting the output of the generators in response to changes in load or
disturbances.

Derivation of Equations for ALFC

1. Governor Equation

The governor controls the turbine input power PmP_m based on frequency deviations (Δf\Delta
f):

ΔPm=−1RΔf\Delta P_m = -\frac{1}{R} \Delta f

Where:

 ΔPm\Delta P_m: Change in mechanical power.

 RR: Speed regulation or droop characteristic.
 Δf\Delta f: Frequency deviation.

2. Turbine Dynamics

The turbine converts mechanical power into electrical power. Its dynamics can be approximated
using a first-order model:

ΔPe(s)=11+TtsΔPm(s)\Delta P_e(s) = \frac{1}{1 + T_t s} \Delta P_m(s)

Where:

 TtT_t: Turbine time constant.

 ss: Laplace transform variable.

3. Generator and Load Model

The generator and load are modeled using the swing equation. The system frequency deviation is
related to the mismatch between generated power (PeP_e) and load demand (PdP_d):

MdΔfdt+DΔf=ΔPm−ΔPdM \frac{d\Delta f}{dt} + D \Delta f = \Delta P_m - \Delta P_d

Where:

 MM: Inertia constant of the system.

 DD: Damping coefficient.
 ΔPd\Delta P_d: Load change.

In the Laplace domain:

Δf(s)=1Ms+D(ΔPm(s)−ΔPd(s))\Delta f(s) = \frac{1}{Ms + D} \left( \Delta P_m(s) - \Delta P_d(s) \right)

4. Closed-Loop Transfer Function

Combining the governor, turbine, and generator models, the closed-loop transfer function for
frequency deviation can be derived as:

Δf(s)=−ΔPd(s)(Ms+D+1R11+Tts)\Delta f(s) = \frac{-\Delta P_d(s)}{\left( M s + D + \frac{1}{R} \frac{1}{1 +

T_t s} \right)}

5. Integral Controller

To eliminate steady-state error in frequency, an integral controller is added. The control signal is:

ΔPint(s)=Ki1sΔf(s)\Delta P_{int}(s) = K_i \frac{1}{s} \Delta f(s)

Where:

 KiK_i: Integral gain.

The integral controller modifies the governor input to ensure that frequency deviation becomes
zero in steady state.

Block Diagram of Automatic Load Frequency Control

Below is the description of the components in the block diagram:

1. Governor: Adjusts the mechanical power input based on frequency deviation (Δf\Delta f).
2. Turbine: Converts mechanical power to electrical power with a first-order dynamic response.
3. Generator and Load: Represents the dynamic relationship between load changes, frequency,
and generated power.
4. Integral Controller: Ensures zero steady-state error in frequency.
To solve the given problem, we need to find the load sharing between the two generators and the
system frequency at the load of 1200 MW. Here's the step-by-step solution:

Step 1: General Equation for Governor Droop

The droop characteristic relates the change in frequency (ff) to the change in load (PP) as:

Δf=R⋅ΔP\Delta f = R \cdot \Delta P

Where:

 Δf=f0−f\Delta f = f_0 - f (frequency deviation),

 ΔP\Delta P is the load change from no load,
 RR is the droop (in Hz per MW or percentage droop).

The droop constant can be expressed as:

R=ΔfPratedR = \frac{\Delta f}{P_{\text{rated}}}

Where PratedP_{\text{rated}} is the rated load of the generator.

Step 2: Droop for Each Generator

For Generator 1:

 Rated capacity (PratedP_{\text{rated}}) = 500 MW,

 Droop (R1R_1) = 3% = 0.03,
 Full-load frequency deviation (Δf1\Delta f_1) = 0.03×50=1.5 Hz0.03 \times 50 = 1.5 \, \
text{Hz},
 Droop constant (R1R_1) = Δf1Prated=1.5500=0.003 Hz/MW\frac{\Delta f_1}{P_{\
text{rated}}} = \frac{1.5}{500} = 0.003 \, \text{Hz/MW}.

For Generator 2:

 Rated capacity (PratedP_{\text{rated}}) = 800 MW,

 Droop (R2R_2) = 4% = 0.04,
 Full-load frequency deviation (Δf2\Delta f_2) = 0.04×50=2 Hz0.04 \times 50 = 2 \, \
text{Hz},
 Droop constant (R2R_2) = Δf2Prated=2800=0.0025 Hz/MW\frac{\Delta f_2}{P_{\
text{rated}}} = \frac{2}{800} = 0.0025 \, \text{Hz/MW}.
Step 3: Load Sharing Between the Generators

The generators operate in parallel, so they will share the load Ptotal=1200 MWP_{\text{total}} =
1200 \, \text{MW} based on their droop characteristics. The relationship between load and
frequency deviation is:

Δf=R1⋅P1=R2⋅P2\Delta f = R_1 \cdot P_1 = R_2 \cdot P_2

And the total load is:

P1+P2=1200 MWP_1 + P_2 = 1200 \, \text{MW}

From Δf=R1⋅P1=R2⋅P2\Delta f = R_1 \cdot P_1 = R_2 \cdot P_2, we can write:

P1=R2R1⋅P2P_1 = \frac{R_2}{R_1} \cdot P_2

Substitute R1=0.003 Hz/MWR_1 = 0.003 \, \text{Hz/MW} and R2=0.0025 Hz/MWR_2 =

0.0025 \, \text{Hz/MW}:

P1=0.00250.003⋅P2=0.8333⋅P2P_1 = \frac{0.0025}{0.003} \cdot P_2 = 0.8333 \cdot P_2

Now substitute into the total load equation P1+P2=1200P_1 + P_2 = 1200:

0.8333⋅P2+P2=12000.8333 \cdot P_2 + P_2 = 1200 1.8333⋅P2=12001.8333 \cdot P_2 = 1200

P2=12001.8333=654.55 MWP_2 = \frac{1200}{1.8333} = 654.55 \, \text{MW}

Thus:

P2=654.55 MW,P1=1200−654.55=545.45 MWP_2 = 654.55 \, \text{MW}, \quad P_1 = 1200 -

654.55 = 545.45 \, \text{MW}

Step 4: System Frequency

The system frequency deviation is the same for both generators, given by:

Δf=R1⋅P1=R2⋅P2\Delta f = R_1 \cdot P_1 = R_2 \cdot P_2

Using R1=0.003 Hz/MWR_1 = 0.003 \, \text{Hz/MW} and P1=545.45 MWP_1 = 545.45 \, \

text{MW}:

Δf=0.003⋅545.45=1.6364 Hz\Delta f = 0.003 \cdot 545.45 = 1.6364 \, \text{Hz}

The system frequency is:

f=f0−Δf=50−1.6364=48.3636 Hzf = f_0 - \Delta f = 50 - 1.6364 = 48.3636 \, \text{Hz}

Final Results

1. Load Sharing:
o Generator 1: P1=545.45 MWP_1 = 545.45 \, \text{MW},
o Generator 2: P2=654.55 MWP_2 = 654.55 \, \text{MW}.
2. System Frequency:
f=48.36 Hzf = 48.36 \, \text{Hz}.

Pumped-Storage Scheduling by a Gradient Method

Pumped-storage scheduling refers to the optimization of energy production and storage in a

pumped-storage hydroelectric system. In these systems, water is pumped from a lower reservoir
to a higher reservoir during low electricity demand (or off-peak hours) and released to generate
electricity during high demand (or peak hours).

Gradient methods are a family of optimization algorithms that use the gradient (or partial
derivatives) of the objective function to iteratively find the optimal solution. When applied to
pumped-storage scheduling, gradient methods optimize the operation of the system to maximize
profit, minimize operational cost, or meet system reliability and energy balance constraints.

Key Elements of Pumped-Storage Scheduling with a Gradient Method

1. Objective Function
The scheduling problem is typically formulated as an optimization problem with the
following goals:
o Maximize revenue by utilizing price differences between peak and off-peak
electricity.
o Minimize overall operational costs, including energy losses during pumping and
generation.
o Achieve system balancing or support grid stability.

The objective function is usually expressed in terms of energy, cost, or profit:

Maximize/Minimize: f(x)=∑t(Pg(t)⋅p(t)−Pp(t)⋅cp(t)),\text{Maximize/Minimize: } f(x) = \

sum_t \left(P_g(t) \cdot p(t) - P_p(t) \cdot c_p(t)\right),

where:

oPg(t)P_g(t): Power generated during time tt,

oPp(t)P_p(t): Power used for pumping,
op(t)p(t): Electricity price at time tt,
ocp(t)c_p(t): Cost of pumping at time tt.
2. Constraints The optimization problem includes constraints such as:
o Energy Balance: Ensuring that the energy stored in the upper reservoir accounts
for inflows, outflows, and pumping losses:
 Et+1=Et+ηp⋅Pp(t)⋅Δt−Pg(t)⋅Δt,E_{t+1} = E_t + \eta_p \cdot P_p(t) \cdot \Delta t -
P_g(t) \cdot \Delta t, where:
 EtE_t: Energy stored at time tt,
 ηp\eta_p: Pumping efficiency,
 Δt\Delta t: Time interval.
o Reservoir Limits: Upper and lower limits for reservoir storage:
Emin≤Et≤Emax.E_{\text{min}} \leq E_t \leq E_{\text{max}}.
o Power Limits: Constraints on the maximum and minimum power during
generation and pumping: Pg(t)∈[0,Pg,max],Pp(t)∈[0,Pp,max].P_g(t) \in [0,
P_{g,\text{max}}], \quad P_p(t) \in [0, P_{p,\text{max}}].
3. Gradient Method
o The gradient method iteratively adjusts decision variables (e.g., Pg(t)P_g(t) and
Pp(t)P_p(t)) in the direction of the steepest descent (or ascent) of the objective
function. The update rule can be expressed as: x(k+1)=x(k)−α∇f(x(k)),x^{(k+1)}
= x^{(k)} - \alpha \nabla f(x^{(k)}), where:

 ∇f(x(k))\nabla f(x^{(k)}): Gradient of the objective function at

 x(k)x^{(k)}: Decision variables at iteration kk,

x(k)x^{(k)},
 α\alpha: Step size or learning rate.
4. Implementation Steps
o Initialization: Start with an initial guess for the decision variables (e.g., initial
pumping/generation schedule).
o Gradient Calculation: Compute the gradient of the objective function with
respect to the decision variables. This involves partial derivatives of f(x)f(x) with
respect to Pg(t)P_g(t), Pp(t)P_p(t), etc.
o Update: Adjust the decision variables using the gradient method formula.
o Constraints Handling: Ensure the solution satisfies the operational constraints
(e.g., by projection or penalty methods).
o Convergence Check: Stop the iterations when changes in the objective function
or decision variables are below a predefined threshold.
5. Advantages of Gradient Methods
o Simple and computationally efficient for large-scale scheduling problems.
o Suitable for real-time or near real-time optimization, especially when the
objective function is smooth and differentiable.
6. Challenges
o Nonlinearity: Pumped-storage systems often exhibit nonlinearities (e.g., due to
water flow dynamics or efficiency curves), which can make gradient calculation
more complex.
o Local Minima: Gradient methods may converge to a local minimum if the
objective function is not convex.
o Computational Efficiency: For large systems or complex constraints, ensuring
rapid convergence may require advanced techniques like adaptive step sizes or
hybrid methods.

Applications
Gradient methods are widely used in:

 Day-ahead market bidding for pumped-storage plants.

 Real-time grid balancing and frequency regulation.
 Long-term reservoir operation planning under varying electricity prices.

The shut-down algorithm in a priority list scheme is commonly used for unit commitment
problems in power systems, where decisions are made about which generating units should be
online or offline during a specific time period. The goal of the algorithm is to shut down
generating units based on their priorities while ensuring system constraints, such as load demand
and reserve requirements, are still met.

Here’s a step-by-step development of a simple shut-down algorithm:

Simple Shut-Down Algorithm for a Priority List Scheme

1. Inputs:
o List of generating units sorted by priority (e.g., based on cost, efficiency, or other
criteria).
o Current load demand (DD).
o Spinning reserve requirement (RR).
o Maximum and minimum power outputs of each unit (PmaxP_{\text{max}},
PminP_{\text{min}}).
o Status of each unit (on/off).
o Current power generation of each unit.
2. Initialization:
o Start with all necessary generating units online to meet the current demand and
reserve requirements.
o Maintain the priority list of units (e.g., units with higher operating costs or lower
efficiencies are given lower priority).
3. Iterative Shut-Down Process: a. Step 1: Calculate Total Online Capacity
Compute the total power output capacity of online units:

Ponline=∑i∈onlinePmax,i.P_{\text{online}} = \sum_{i \in \text{online}} P_{\

text{max}, i}.

Ensure this is sufficient to meet:

Ponline≥D+R.P_{\text{online}} \geq D + R.

b. Step 2: Select the Unit to Shut Down

From the priority list, identify the unit with the lowest priority (e.g., highest operating
cost or least efficiency) that is currently online.
 c. Step 3: Check Feasibility of Shut-Down
Temporarily remove the selected unit and recalculate:

o Remaining Online Capacity: Premaining=Ponline−Pmax, selected.P_{\

text{remaining}} = P_{\text{online}} - P_{\text{max, selected}}.
o Ensure that shutting down the unit does not violate constraints:
 Premaining≥D+RP_{\text{remaining}} \geq D + R (load demand and
reserve are met).
 The remaining units can handle the load without exceeding their
maximum capacities.

d. Step 4: Shut Down the Unit

If the constraints are satisfied, shut down the selected unit and update its status.
Otherwise, move to the next lowest-priority unit in the list and repeat Steps 3 and 4.

4. Termination Condition:
o Stop the shut-down process when no more units can be shut down without
violating load or reserve constraints.
o The remaining online units form the final committed set of units.
5. Output:
o The final list of online units after shut-down.
o Updated statuses and power outputs for each unit.

Key Features of the Algorithm

1. Priority-Based: The units with the lowest priority (e.g., highest cost or lowest efficiency)
are considered for shut-down first.
2. Constraint-Sensitive: Ensures that load demand and reserve requirements are always
satisfied.
3. Simple and Iterative: The algorithm iterates through the priority list and updates the
system until no further shut-downs are feasible.

Applications

 Unit commitment in power system operations.

 Optimization of thermal or hydroelectric power plants to reduce operational costs.
 Planning during periods of reduced electricity demand.
To solve the economic operating schedule and calculate the savings obtained by loading the units
as per Equal Incremental Production Cost (EIPC), let us proceed step by step:

Step 1: Problem Setup

Given:

 Fuel cost functions for the two plants:

F1=0.2P12+40P1+120(Rs/hr)F_1 = 0.2 P_1^2 + 40P_1 + 120 \quad (\text{Rs/hr})

F2=0.25P22+30P2+150(Rs/hr)F_2 = 0.25 P_2^2 + 30P_2 + 150 \quad (\text{Rs/hr})

 Maximum and minimum loading limits:

25≤P1≤100MW,25≤P2≤100MW25 \leq P_1 \leq 100 \quad \text{MW}, \quad 25 \leq

P_2 \leq 100 \quad \text{MW}

 Total demand:

PD=P1+P2=180MWP_D = P_1 + P_2 = 180 \quad \text{MW}

 Transmission losses are neglected.

Step 2: Equal Load Sharing

If the load is equally shared by both units:

P1=P2=PD2=1802=90MWP_1 = P_2 = \frac{P_D}{2} = \frac{180}{2} = 90 \quad \text{MW}

 Fuel costs:

F1=0.2(90)2+40(90)+120=1620+3600+120=5340Rs/hrF_1 = 0.2(90)^2 + 40(90) + 120 =

1620 + 3600 + 120 = 5340 \quad \text{Rs/hr}
F2=0.25(90)2+30(90)+150=2025+2700+150=4875Rs/hrF_2 = 0.25(90)^2 + 30(90) +
150 = 2025 + 2700 + 150 = 4875 \quad \text{Rs/hr}

 Total cost for equal load sharing:

Ftotal=F1+F2=5340+4875=10,215Rs/hrF_{\text{total}} = F_1 + F_2 = 5340 + 4875 =

10,215 \quad \text{Rs/hr}
Step 3: Economic Load Dispatch (Equal Incremental Cost)

To minimize total cost, we apply the condition of Equal Incremental Production Cost:

dF1dP1=dF2dP2\frac{dF_1}{dP_1} = \frac{dF_2}{dP_2}

1. Compute the incremental cost functions:

dF1dP1=0.4P1+40\frac{dF_1}{dP_1} = 0.4P_1 + 40 dF2dP2=0.5P2+30\frac{dF_2}

{dP_2} = 0.5P_2 + 30

2. Set the incremental costs equal:

0.4P1+40=0.5P2+300.4P_1 + 40 = 0.5P_2 + 30 0.4P1−0.5P2=−100.4P_1 - 0.5P_2 = -10

3. Use the power balance equation:

P1+P2=180orP2=180−P1P_1 + P_2 = 180 \quad \text{or} \quad P_2 = 180 - P_1

4. Substitute P2=180−P1P_2 = 180 - P_1 into the incremental cost equation:

0.4P1−0.5(180−P1)=−100.4P_1 - 0.5(180 - P_1) = -10 0.4P1−90+0.5P1=−100.4P_1 - 90

+ 0.5P_1 = -10 0.9P1=800.9P_1 = 80 P1=800.9≈88.89MWP_1 = \frac{80}{0.9} \approx
88.89 \quad \text{MW}

5. Compute P2P_2:

P2=180−P1=180−88.89=91.11MWP_2 = 180 - P_1 = 180 - 88.89 = 91.11 \quad \

text{MW}

Step 4: Total Cost with EIPC

 Compute the fuel costs for the economic dispatch:

F1=0.2(88.89)2+40(88.89)+120≈1577.78+3555.56+120=5253.34Rs/hrF_1 =
0.2(88.89)^2 + 40(88.89) + 120 \approx 1577.78 + 3555.56 + 120 = 5253.34 \quad \
text{Rs/hr} F2=0.25(91.11)2+30(91.11)+150≈2076.85+2733.33+150=4960.18Rs/hrF_2
= 0.25(91.11)^2 + 30(91.11) + 150 \approx 2076.85 + 2733.33 + 150 = 4960.18 \quad \
text{Rs/hr}

 Total cost with economic dispatch:

Ftotal, EIPC=F1+F2=5253.34+4960.18=10,213.52Rs/hrF_{\text{total, EIPC}} = F_1 +

F_2 = 5253.34 + 4960.18 = 10,213.52 \quad \text{Rs/hr}
Step 5: Savings

 Cost with equal load sharing:

Ftotal, equal=10,215Rs/hrF_{\text{total, equal}} = 10,215 \quad \text{Rs/hr}

 Cost with EIPC:

Ftotal, EIPC=10,213.52Rs/hrF_{\text{total, EIPC}} = 10,213.52 \quad \text{Rs/hr}

 Savings:

Savings=Ftotal, equal−Ftotal, EIPC=10,215−10,213.52=1.48Rs/hr\text{Savings} = F_{\

text{total, equal}} - F_{\text{total, EIPC}} = 10,215 - 10,213.52 = 1.48 \quad
\text{Rs/hr}

Final Results

1. Economic Operating Schedule:

o P1=88.89 MW, P2=91.11 MWP_1 = 88.89 \, \text{MW}, \, P_2 = 91.11 \, \
text{MW}
2. Savings:
o Savings by using the EIPC method: 1.48 Rs/hr.

State estimation in power systems involves determining the most accurate values of the system's
state variables, such as bus voltages and phase angles, based on available measurements (e.g.,
power flows, injections, and voltages). Accurate state estimation is critical for real-time
monitoring, control, and operation of power systems.

There are several techniques for state estimation, which can be categorized based on their
mathematical approach, measurement types, and system characteristics. Below is a detailed
explanation of the most commonly used state estimation techniques:

1. Weighted Least Squares (WLS) State Estimation

The Weighted Least Squares (WLS) method is the most commonly used state estimation
technique. It minimizes the weighted sum of squared differences between the measured values
and the calculated values based on the system model.
Objective Function:
J(x)=∑i=1mwi(zi−hi(x))2J(x) = \sum_{i=1}^{m} w_i (z_i - h_i(x))^2

Where:

 ziz_i: Measured value of the ii-th measurement.

 hi(x)h_i(x): Calculated value of the ii-th measurement from the system model (nonlinear
function of state variables xx).
 wi=1σi2w_i = \frac{1}{\sigma_i^2}: Weight assigned to the ii-th measurement, inversely
proportional to its variance (σi2\sigma_i^2).
 mm: Total number of measurements.

Steps:

1. Formulate Measurement Equations: Express measurements (ziz_i) as nonlinear functions

(hi(x)h_i(x)) of the state variables (e.g., bus voltages, phase angles).
2. Linearize the Equations: Use the Taylor series expansion around an initial estimate to linearize
the nonlinear equations.
3. Iteratively Solve: Apply an iterative numerical method (e.g., Newton-Raphson) to minimize the
objective function and estimate the state variables.

Advantages:

 Robust to measurement noise.

 Handles large power systems effectively.
 Provides statistical insight into the confidence of estimates.

Disadvantages:

 Sensitive to bad data (outliers in measurements).

 Requires a good initial guess for fast convergence.

2. Kalman Filter (KF) State Estimation

The Kalman Filter is a recursive method for state estimation in dynamic systems where the
states evolve over time. It is particularly useful for real-time state estimation in power systems
with rapidly changing conditions.

Key Equations:

1. Prediction Step:

x^k∣k−1=Akx^k−1∣k−1+Bkuk\hat{x}_{k|k-1} = A_k \hat{x}_{k-1|k-1} + B_k u_k

Pk∣k−1=AkPk−1∣k−1AkT+QkP_{k|k-1} = A_k P_{k-1|k-1} A_k^T + Q_k
 Where:

o x^k∣k−1\hat{x}_{k|k-1}: Predicted state vector at time kk.

o Pk∣k−1P_{k|k-1}: Predicted error covariance.
o Ak,BkA_k, B_k: System matrices.
o QkQ_k: Process noise covariance.

2. Correction Step:

Kk=Pk∣k−1HkT(HkPk∣k−1HkT+Rk)−1K_k = P_{k|k-1} H_k^T (H_k P_{k|k-1} H_k^T + R_k)^{-1}

x^k∣k=x^k∣k−1+Kk(zk−Hkx^k∣k−1)\hat{x}_{k|k} = \hat{x}_{k|k-1} + K_k (z_k - H_k \hat{x}_{k|k-1})
Pk∣k=(I−KkHk)Pk∣k−1P_{k|k} = (I - K_k H_k) P_{k|k-1}

Where:

o KkK_k: Kalman gain.

o RkR_k: Measurement noise covariance.
o zkz_k: Measurement vector at time kk.

Advantages:

 Suitable for real-time dynamic state estimation.

 Efficient for systems with time-varying states.
 Handles process and measurement noise effectively.

Disadvantages:

 Requires accurate system models and noise covariances.

 Limited robustness to gross measurement errors.

3. Extended Kalman Filter (EKF)

The Extended Kalman Filter is a variant of the Kalman Filter designed for nonlinear systems. It
linearizes the system dynamics and measurement equations around the current estimate using a
first-order Taylor series expansion.

Steps:

1. Linearize the nonlinear equations.

2. Apply the standard Kalman filter steps (prediction and correction) using the linearized model.

Advantages:

 Handles nonlinear systems.

  Maintains the real-time capability of the standard Kalman filter.

Disadvantages:

 Accuracy depends on the quality of linearization.

 Computationally intensive for large systems.

4. Unscented Kalman Filter (UKF)

The Unscented Kalman Filter is an improvement over the EKF that does not require explicit
linearization. Instead, it uses a set of sigma points to approximate the nonlinear transformations.

Key Features:

 Generates a deterministic set of sigma points around the current state estimate.
 Propagates these points through the nonlinear equations.
 Uses the propagated points to compute the mean and covariance of the state.

Advantages:

 More accurate than EKF for highly nonlinear systems.

 Avoids the need for Jacobian computations.

Disadvantages:

 Higher computational complexity compared to EKF.

5. Linear State Estimation (LSE)

In Linear State Estimation, the measurement equations are assumed to be linear functions of the
state variables. This technique is suitable for systems with PMUs (Phasor Measurement Units),
which provide direct voltage and current phasor measurements.

Key Features:

 Direct application of least-squares or weighted least-squares methods.

 Eliminates the need for iterative numerical methods.

Advantages:

 Fast and computationally efficient.

 High accuracy with PMU measurements.
Disadvantages:

 Requires sufficient PMU deployment to achieve observability.

6. Bayesian State Estimation

Bayesian state estimation techniques use probabilistic models to estimate the state variables.
They rely on the prior probability distribution of states and update it using measurement data.

Key Features:

 Combines prior knowledge (e.g., system dynamics) with measurement data.

 Provides a posterior probability distribution for the states.

Techniques:

 Particle Filters: Use a set of particles to represent the probability distribution.

 Gaussian Mixture Models: Approximate distributions as a mixture of Gaussians.

Advantages:

 Robust to uncertainties and noise.

 Handles nonlinear and non-Gaussian systems.

Disadvantages:

 Computationally expensive for large systems.

7. Robust State Estimation

Robust state estimation techniques aim to minimize the impact of bad data (gross errors) on the
estimation results. Common approaches include:

 Iterative Reweighted Least Squares (IRLS): Assigns lower weights to suspicious measurements.
 Median-Based Filters: Use median values to reject outliers.
 Least Absolute Value (LAV): Minimizes the sum of absolute errors instead of squared errors.

Advantages:

 Improves reliability in the presence of bad data.

 Increases robustness to measurement outliers.
Disadvantages:

 May require more computational effort.

Summary of Techniques

Technique Key Feature Application Complexity

WLS Weighted least squares optimization Static systems Moderate

Real-time estimation for linear

Kalman Filter (KF) Dynamic systems Moderate
systems

Extended Kalman Filter Linearized dynamic systems Nonlinear dynamic systems High

Unscented Kalman Filter Sigma points for nonlinear systems Highly nonlinear systems High

Linear State Estimation PMU-based linear equations PMU-enabled systems Low

Bayesian State Estimation Probabilistic approach Uncertain systems Very High

Robust Estimation Bad data handling Systems with outliers Moderate