Economic Load Dispatch Basics: Complete Beginner's Guide

Economic load dispatch answers a daily question for every power utility: given tomorrow's demand, how much power should each generator produce so the total fuel cost is minimum? A system with ten generators has an infinite number of ways to meet 5,000 MW of demand — dispatch finds the cheapest combination that still respects every unit's limits.

The key insight is beautifully simple: at the cheapest operating point, every generator that is free to move operates at the same incremental (marginal) cost. If one unit can produce its next megawatt for ₹8 while another needs ₹14, shifting one megawatt from the expensive unit to the cheap one saves ₹6 — so the dispatch was not optimal. Costs keep shifting until all incremental costs are equal, a common value called lambda (λ).

Key Concepts

• Incremental cost curve: dC/dP, the cost of the next MW from a unit; typically rising linearly as output increases.
• Lambda (λ): the system incremental cost in ₹/MWh; all dispatched units operate where their incremental cost equals λ.
• Generator limits: each unit has Pmin and Pmax; if the λ rule pushes a unit outside its limits, it is clamped and the rest re-optimized.
• Penalty factor: when transmission losses are included, each unit's effective cost becomes λ × (1 − ∂Ploss/∂Pi)⁻¹, favouring units electrically closer to load.
• Unit commitment: the bigger day-ahead decision of which units to start up or shut down; economic dispatch then fine-tunes the running units.
• Cost curves: usually modelled as quadratic: C = a + bP + cP², where a is no-load cost, b the linear coefficient and c the curvature.

Formulas

• Optimality condition (no losses): dC1/dP1 = dC2/dP2 = … = λ, with Σ Pi = demand.
• Quadratic cost: Ci = ai + biPi + ciPi², so incremental cost dCi/dPi = bi + 2ciPi.
• Unit output from λ: Pi = (λ − bi) / (2ci).
• With losses: (dCi/dPi) × Li = λ, where Li = 1/(1 − ∂Ploss/∂Pi) is the penalty factor.

Worked Example

Two units supply 500 MW. Unit 1: C1 = 200 + 8P1 + 0.01P1²; Unit 2: C2 = 180 + 7P2 + 0.02P2² (₹/h). Incremental costs: IC1 = 8 + 0.02P1, IC2 = 7 + 0.04P2. Setting both equal to λ with P1 + P2 = 500: P1 = (λ − 8)/0.02 = 50λ − 400, P2 = (λ − 7)/0.04 = 25λ − 175. Adding: 75λ − 575 = 500, so λ = 14.33 ₹/MWh. Then P1 = 50 × 14.33 − 400 ≈ 316.7 MW and P2 = 25 × 14.33 − 175 ≈ 183.3 MW. Check: 316.7 + 183.3 = 500 MW ✓, and IC1 = 8 + 0.02 × 316.7 = 14.33 = IC2 ✓ — the equal-lambda condition holds, so this is the cheapest split.

Common Misconceptions

• "The cheapest unit should take all the load." Average cost misleads; dispatch follows incremental cost, and every unit also has minimum and maximum limits that bind first.
• "Dispatch ignores transmission losses." Real dispatch includes them through penalty factors — a cheap far-away unit can lose to a pricier nearby one once losses are counted.
• "Lambda is the electricity price." Lambda is the system's marginal production cost; the market price consumers pay adds transmission, distribution, taxes and margins on top.

Real-World Applications

• Load dispatch centres run economic dispatch every 15 minutes (and security-constrained dispatch every 5) to schedule India's 400+ GW fleet.
• Power exchanges use the same marginal-cost logic to clear day-ahead markets — the most expensive accepted bid sets the price.
• Utilities evaluate fuel contracts and plant upgrades by how they shift the system's lambda-duration curve.
• Renewables with near-zero incremental cost dispatch first (merit order), which is why solar depresses daytime wholesale prices.

Economic dispatch is one rule with enormous consequences: equalize every unit's incremental cost and the total fuel bill is minimized. From that single principle flow lambda, penalty factors, merit-order markets and the daily scheduling of entire national grids.