Economic Load Dispatch in Power Systems

Introduction

  • The goal is to minimize production cost while maintaining voltage magnitudes at each bus.
  • Load conditions vary with time (day/night, season), so constant power generation is unrealistic.
  • Power generation should follow load patterns, which change with seasons.
  • To optimize dispatch, we focus on two aspects:
    • Economic operation must adapt to load conditions.
    • Turbine-governor control must ensure optimal generation.

Economic Operation of Power Systems

  • Power was initially supplied by the most efficient plant at light load.
  • As load increased, the most efficient plant supplied until its max efficiency was reached.
  • The next most efficient plant then supplied power until its max efficiency.
  • This continued from the most to the least efficient plants to meet peak demand.
  • However, this method failed to minimize total cost of generation.
  • An alternative method is needed to account for total generation costs of all units.

Economic Distribution of Loads between the Units of a Plant

  • Variable operating costs of each unit are expressed in terms of power output.
  • The fuel cost is the dominant cost in thermal or nuclear units.
  • Fuel cost must be expressed as a function of power output.
  • Other costs, like operation and maintenance, can also be expressed in terms of power output.
  • Fixed costs, like capital and depreciation, are not included in fuel cost.
  • Fuel requirement is given in Rupees per hour.
  • For unit-\(i\), input cost \(f_i\) (Rs./h) is expressed as:
    \[f_i = \left[\dfrac{a_i}{2}P_i^2 + b_iP_i + c_i\right]~Rs/h\]
  • Operating cost is approximated by a quadratic function of power (MW) versus cost (Rs).
  • Incremental operating cost is calculated as:
    \[ \lambda_{i} = \dfrac{df_{i}}{dP_{i}} = \left(a_{i}P_{i} + b_{i}\right)~\text{Rs/MWh} \]
  • For two units with different incremental costs, load transfer from higher to lower incremental cost reduces the total cost.
  • This continues until incremental costs are equal for both units, which is the optimal point of operation.
  • This principle extends to \(N\) units in a plant.
  • Total fuel cost \(f_T\) is the sum of individual costs:
    \[f_{T}=f_{1}+f_{2}+\cdots+f_{N}={\displaystyle \sum_{k=1}^{N}f_{k}}\]
  • Total power \(P_T\) is the sum of powers supplied by each unit:
    \[P_{T}= P_{1}+P_{2}+\cdots+P_{N}={\displaystyle \sum_{k=1}^{N}P_{k}}\]
  • The objective is to minimize \(f_T\) for a given \(P_T\), achieved when:
    \[df_{T}=\dfrac{\partial f_{T}}{\partial P_{1}}dP_{1}+\dfrac{\partial f_{T}}{\partial P_{2}}dP_{2}+\cdots+\dfrac{\partial f_{T}}{\partial P_{N}}dP_{N}=0\]
  • Since total power is constant:
    \[dP_{T}= dP_{1}+dP_{2}+\cdots+dP_{N}=0\]
  • Multiplying by \(\lambda\) and subtracting:
    \[\left(\dfrac{\partial f_{T}}{\partial P_{1}}-\lambda\right)dP_{1}+\left(\dfrac{\partial f_{T}}{\partial P_{2}}-\lambda\right)dP_{2}+\cdots+\left(\dfrac{\partial f_{T}}{\partial P_{N}}-\lambda\right)dP_{N}=0\]
  • The equality holds when:
    \[\dfrac{\partial f_{T}}{\partial P_{i}}-\lambda=0,~i=1,\cdots,N\]
  • This leads to:
    \[\lambda=\dfrac{df_{1}}{dP_{1}}=\dfrac{df_{2}}{dP_{2}}=\cdots=\dfrac{df_{N}}{dP_{N}} \Leftarrow~\text{coordination equation}\]
    Economic Load Optimization Graph
    Figure 1: Economic Load Optimization Curve

Generating Limits

  • Not all units are always available to share the load due to scheduled maintenance.
  • It is not necessary to switch off less efficient units during off-peak hours.
  • Shutting down and starting up units incurs costs and may take over 8 hours to restore and synchronize.
  • To handle sudden power demand changes, extra units may be required, creating a spinning reserve.
  • The optimal load dispatch must include startup and shutdown costs while ensuring system security.
  • Power generation limits are constrained by:
    \[P_{min,i} \leq P_i \leq P_{max,i}, ~ i=1,\dots,N\]
  • \(P_{max}\) is the maximum power capacity of each unit.
  • \(P_{min}\) ensures thermal stability of the boiler in thermal or nuclear stations.
  • A unit must produce a minimum power to maintain the boiler's design operating temperature.

Economic Sharing of Loads between Different Plants

  • We previously considered the economic operation of a single plant, focusing on how load is shared among its units.
  • This analysis ignored transmission line losses, assuming they were part of the load.
  • When considering multiple plants connected by transmission lines, line losses must be included in the economic dispatch.
  • Including transmission losses in the economic dispatch:
    \[P_T = P_1 + P_2 + \cdots + P_N - P_{Loss}\]
    Since \(P_T\) is constant, we have:
    \[0 = dP_1 + dP_2 + \cdots + dP_N - dP_{Loss}\]
    where:
    \[dP_{Loss} = \dfrac{\partial P_{Loss}}{\partial P_1}dP_1 + \dfrac{\partial P_{Loss}}{\partial P_2}dP_2 + \cdots + \dfrac{\partial P_{Loss}}{\partial P_N}dP_N\]
  • Minimum generation cost implies \(df_T = 0\):
    \[0 = \left(\lambda \dfrac{\partial P_{Loss}}{\partial P_1} - \lambda \right) dP_1 + \cdots + \left(\lambda \dfrac{\partial P_{Loss}}{\partial P_N} - \lambda \right) dP_N\]
    Summing up:
    \[0 = \sum_{i=1}^{N} \left(\dfrac{\partial f_T}{\partial P_i} + \lambda \dfrac{\partial P_{Loss}}{\partial P_i} - \lambda \right) dP_i\]
  • The equation is satisfied when:
    \[\dfrac{\partial f_T}{\partial P_i} + \lambda \dfrac{\partial P_{Loss}}{\partial P_i} - \lambda = 0, \quad i = 1, \cdots, N\]
  • Thus:
    \[\lambda = \dfrac{df_1}{dP_1} L_1 = \dfrac{df_2}{dP_2} L_2 = \cdots = \dfrac{df_N}{dP_N} L_N\]
    where \(L_i\) is the penalty factor for load-\(i\), given by:
    \[L_i = \dfrac{1}{1 - \dfrac{\partial P_{Loss}}{\partial P_i}}, \quad i = 1, \cdots, N\]
  • For an area with \(N\) units, the power generated is represented by the vector:
    \[P = \left[\begin{array}{cccc} P_1 & P_2 & \cdots & P_N \end{array}\right]^T\]
  • Transmission losses are expressed as:
    \[P_{Loss} = P^T B P\]
    where \(B\) is a symmetric \(N \times N\) matrix, and \(B_{ij}\) are the loss coefficients.
    \[B=\left[\begin{array}{cccc} B_{11} & B_{12} & \cdots & B_{1 N} \\ B_{12} & B_{22} & \cdots & B_{2 N} \\ \vdots & \vdots & \ddots & \vdots \\ B_{1 N} & B_{2 N} & \cdots & B_{N W} \end{array}\right]\]
  • These coefficients vary with plant loading, but for simplicity, they are often assumed constant when calculating the penalty factor \(L_i\).

Economic Dispatch Problem-1

Consider two units of a power plant with fuel costs given by the following equations:

\[\begin{aligned} F_1 & = 0.2 P_1^2 + 40 P_1 + 120 ~ \text{Rs}/\text{h}\\ F_2 & = 0.25 P_2^2 + 30 P_2 + 150 ~ \text{Rs}/\text{h} \end{aligned}\]
  1. Determine the economic operating schedule and the corresponding cost of generation for a demand of 180 MW.
  2. If the load is equally shared by both units, determine the savings obtained by loading the units optimally.

Solution

Part 1 - Economic Dispatch

  • For economical dispatch, the fuel cost derivative relative to each unit's output power must be equal.
    \[\frac{dF_1}{dP_1} = \frac{dF_2}{dP_2}\]
  • Taking the derivatives of the fuel cost equations:
    \[0.4 P_1 + 40 = 0.5 P_2 + 30\]
  • The total demand must be met:
    \[P_1 + P_2 = 180\]
  • Solving this system of equations:
    \[P_1 = 88.89 \, \text{MW}, \quad P_2 = 91.11 \, \text{MW}\]
  • The total cost of generation is the sum of the individual costs:
    \[F_T = F_1 + F_2 = 10,214.43 \, \text{Rs}/\text{h}\]

Part 2: Equal Load Sharing

  • If the load is equally shared by both units:
    \[P_1 = 90 \, \text{MW}, \quad P_2 = 90 \, \text{MW}\]
  • The total cost of generation in this case is:
    \[F_T = 10,215 \, \text{Rs}/\text{h}\]
  • Thus, the savings obtained by optimally loading the units is:
    \[\text{Savings} = 10,215 - 10,214.43 = 0.57 \, \text{Rs}/\text{h}\]

Economic Dispatch Problem-2

The fuel cost functions for three thermal plants are:

\[\begin{array}{r} F_1 = 0.4 P_1^2 + 10 P_1 + 25 \, \text{Rs/h} \\ F_2 = 0.35 P_2^2 + 5 P_2 + 20 \, \text{Rs/h} \\ F_3 = 0.475 P_3^2 + 15 P_3 + 35 \, \text{Rs/h} \end{array}\]

The generation limits of the units are:

\[\begin{aligned} 30 \, \text{MW} &\leq P_1 \leq 500 \, \text{MW} \\ 30 \, \text{MW} &\leq P_2 \leq 500 \, \text{MW} \\ 30 \, \text{MW} &\leq P_3 \leq 250 \, \text{MW} \end{aligned}\]

Find the optimum schedule for a load of 1000 MW.

Solution

  • For optimum dispatch:
    \[\frac{d F_1}{d P_1} = \frac{d F_2}{d P_2} = \frac{d F_3}{d P_3}\]
  • This leads to the following equations:
    \[\begin{aligned} 0.8 P_1 + 10 & = 0.7 P_2 + 5 \\ 0.7 P_2 + 5 & = 0.9 P_3 + 15 \end{aligned}\]
  • The total power balance constraint:
    \[P_1 + P_2 + P_3 = 1000\]
  • By solving the above system of equations, we get:
    \[P_1 = 334.3829 \, \text{MW}, \quad P_2 = 389.2947 \, \text{MW}, \quad P_3 = 276.3224 \, \text{MW}\]
  • However, unit 3 violates its maximum limit, so we set:
    \[P_3 = 250 \, \text{MW}\]
  • The remaining load (750 MW) is distributed optimally between units 1 and 2.
  • The equations to solve now are:
    \[\begin{array}{r} 0.8 P_1 + 10 = 0.7 P_2 + 5 \\ P_1 + P_2 = 750 \end{array}\]
  • Solving these, we get the final load distribution as:
    \[P_1 = 346.6667 \, \text{MW}, \quad P_2 = 403.3333 \, \text{MW}, \quad P_3 = 250 \, \text{MW}\]

Economic Dispatch Problem-3

Consider a two bus system.

Two Bus System Diagram
Figure 2: Two Bus System Configuration

The incremental fuel cost characteristics of plant 1 and plant 2 are:

\[\begin{aligned} & \frac{d F_1}{d P_1}=0.025 P_1+14 \mathrm{Rs} / \mathrm{MWHr} \\ & \frac{d F_2}{d P_2}=0.05 P_2+16 \mathrm{Rs} / \mathrm{MWHr} \end{aligned}\]

If 200 MW of power is transmitted from plant 1 to the load, a transmission loss of 20 MW occurs. Determine the optimal generation schedule and the cost of the received power for a load demand of 204.41 MW.

Solution

\[P_L=\left[\begin{array}{ll} P_1 & P_2 \end{array}\right]\left[\begin{array}{ll} B_{11} & B_{12} \\ B_{21} & B_{22} \end{array}\right]\left[\begin{array}{l} P_1 \\ P_2 \end{array}\right]\]
  • Since the load is at bus \(2, P_2\) will not have any effect on \(P_L\).
    \[\begin{aligned} B_{12} & =B_{21}=0 ; \quad B_{22}=0 \\ \Rightarrow~ P_L & =B_{11} P_1^2 \end{aligned}\]
  • For 200 MW of \(P_1, P_L=20 \mathrm{MW}\).
    \[\begin{gathered} 20=B_{11} 200^2 \\ B_{11}=0.0005 \mathrm{MW}^{-1} \\ P_L=0.0005 P_1^2 \end{gathered}\]
  • For optimum dispatch,
    \[L_1 \frac{d F_1}{d P_1}=L_2 \frac{d F_2}{d P_2}=\lambda\]
  • Since \(P_L\) is a function of \(P_1\) alone,
    \[\begin{gathered} L_1=\frac{1}{1-\frac{\partial P_L}{\partial P_1}}=\frac{1}{1-0.001 P_1} \\ L_2=1 \\ \left(\frac{1}{1-0.001 P_1}\right) 0.025 P_1+14=0.05 P_2+16 \end{gathered}\]
  • On simplification,
    \[\begin{aligned} & 0.041 P_1-0.05 P_2+0.00005 P_1 P_2=2 \\ & P_1+P_2-0.0005 P_1^2=204.41 \end{aligned}\]
  • So,
    \[\begin{array}{r} f_1\left(P_1, P_2\right)=2 \\ f_2\left(P_1, P_2\right)=204.41 \end{array}\]
  • Let us solve them by \(\mathrm{N}-\mathrm{R}\) method.
    \[\begin{aligned} \Delta \mathrm{f} & =\mathrm{J} \Delta \mathrm{P} \\ \Delta f &=\left[\begin{array}{c} 2-f_1\left(P_1, P_2\right) \\ 204.41-f_2\left(P_1, P_2\right) \end{array}\right] \\ \mathrm{J} &=\left[\begin{array}{ll} \frac{\partial f_1}{\partial P_1} & \frac{\partial f_1}{\partial P_2} \\ \frac{\partial f_2}{\partial P_1} & \frac{\partial f_2}{\partial P_2} \end{array}\right] \end{aligned}\]
  • To find the initial estimate: Let us solve the problem without loss.
    \[\begin{gathered} 0.025 P_1+14=0.05 P_2+16 \\ P_1+P_2=204.41 \\ P_1^0=162.94 ; P_2^0=41.47 \end{gathered}\]
  • First Iteration:
    \[\begin{aligned} \Delta f^0 & =\left[\begin{array}{c} 2-f_1\left(P_1^0, P_2^0\right) \\ 204.41-f_2\left(P_1^0, P_2^0\right) \end{array}\right]=\left[\begin{array}{c} -2.9449 \\ 13.2747 \end{array}\right] \\ J^0 & =\left[\begin{array}{cc} 0.0431 & -0.0419 \\ 0.8371 & 0.9585 \end{array}\right] \\ \left[\begin{array}{l} \Delta P_1^0 \\ \Delta P_2^0 \end{array}\right] & =\left[\begin{array}{c} -29.7060 \\ 39.7906 \end{array}\right] \\ \left[\begin{array}{l} P_1^1 \\ P_2^1 \end{array}\right] & =\left[\begin{array}{l} P_1^0 \\ P_2^0 \end{array}\right]+\left[\begin{array}{c} \Delta P_1^0 \\ \Delta P_2^0 \end{array}\right]=\left[\begin{array}{c} 133.2340 \\ 81.2606 \end{array}\right] \end{aligned}\]
  • It took 6 iterations to converge.
    \[P_1=133.3153 \mathrm{MW} \quad P_2=79.9812 \mathrm{MW}\]
  • The cost of received power is
    \[\lambda=L_2 \frac{d F_2}{d P_2}=1 \times(0.05 \times 79.9812+16)=19.9991 \mathrm{Rs} . / \mathrm{MWh}\]