Legendre Transformation Explained with Geometric Animation

A function is concave if for any two points x and y in its domain and any λ in [0,1], the function evaluated at the linear combination of x and y is greater than or equal to the linear combination of the function values: f(λx + (1-λ)y) ≥ λf(x) + (1-λ)f(y); conversely, a convex function satisfies the opposite inequality f(λx + (1-λ)y) ≤ λf(x) + (1-λ)f(y), making convex functions the direct opposite of concave functions in terms of their mathematical definition.

A function f(x₁, x₂, ..., xₙ) is convex if and only if its n×n Hessian matrix (composed of second-order partial derivatives) is positive semidefinite for all values of the variables, meaning zᵀHz ≥ 0 for every non-zero vector z; conversely, a function is concave if its Hessian is negative semidefinite (zᵀHz ≤ 0). For two-variable functions, convexity can be checked by verifying that f₁₁ ≥ 0, f₂₂ ≥ 0, and f₁₁f₂₂ - (f₁₂)² ≥ 0, while concavity requires f₁₁ ≤ 0, f₂₂ ≤ 0, and f₁₁f₂₂ - (f₁₂)² ≥ 0.

Common functions exhibit different convexity properties: the square function (x²) is convex on ℝ; the square root function (√x) is concave on [0, +∞); the exponential function (e^x) is convex on ℝ; the logarithm function (ln x) is concave on (0, +∞); and the inverse function (1/x) changes convexity, being concave on (-∞, 0) and convex on (0, +∞). Knowing convexity provides practical applications: it reveals growth rates (convex functions grow increasingly fast, concave functions grow increasingly slow) and enables deriving inequalities. For example, since ln x is concave, its curve lies below the tangent at x=1 (y=x-1), yielding ln x ≤ x-1 for all x > 0.

Convex functions have bowl-like shapes with single minima; concave functions have cave-like shapes with single maxima. Convex functions guarantee that any local optimum is global. Concave functions guarantee that any local maximum is global. Functions that are neither may have multiple local optima, making optimization more challenging.

Convex and concave functions describe curvature. A function is convex if its graph lies below chords connecting any two points. A function is concave if its graph lies above chords. These can be determined through: (1) the function itself, (2) its graph, or (3) the convex set property.

Physical systems follow trajectories that minimize a quantity called the action, which is the integral of the Lagrangian over time. The Lagrangian is typically kinetic energy minus potential energy. Euler and Lagrange discovered that this principle leads to the equations of motion for most physical systems. When a system has temporal translation symmetry (physics doesn't change with time), the Lagrangian changes by a term proportional to the time interval. Using the chain rule and Euler-Lagrange equations, this leads to the conservation of a specific quantity. For the simple Lagrangian (kinetic minus potential energy), this conserved quantity is the total energy.

Lagrangian mechanics, formulated by Joseph Louis Lagrange in the late 1700s, provides an alternative to Newtonian mechanics by describing physical systems through energy rather than forces. The Lagrangian L is defined as the difference between kinetic energy T and potential energy V (L = T - V), and the principle of stationary action states that the actual trajectory of a system corresponds to a stationary point (minimum, maximum, or inflection point) of the action S, which is the integral of the Lagrangian over time. This principle leads to the Euler-Lagrange equation, which can be applied to obtain the equations of motion for complex systems with multiple degrees of freedom more conveniently than Newton's force-based approach.

Lagrangian mechanics provides an alternative approach to classical mechanics by defining the Lagrangian as the difference between kinetic energy (T) and potential energy (U), and finding the path that minimizes the action (the integral of the Lagrangian over time) using the Euler-Lagrange equation; this method is particularly powerful for systems with constraints and generalized coordinates, as it eliminates the need to explicitly calculate constraint forces and allows the use of any convenient coordinate system to describe the system's configuration.

The Lagrangian (L) is defined as the difference between kinetic energy (T) and potential energy (V), so L = T - V. The action is the integral of the Lagrangian over time between two points. By applying the calculus of variations to minimize the action, one derives the equations of motion for a physical system. This approach, developed by Lagrange in the 18th century, provides an alternative to Newton's laws and is particularly powerful for complex systems with constraints. The Lagrangian formulation is fundamental to modern physics, including quantum mechanics and general relativity.

The Lagrangian approach offers major advantages over Newtonian mechanics. Simply write kinetic and potential energy, plug into the Euler-Lagrange equation, and obtain correct equations of motion. This works in multiple dimensions and with any coordinate system—including polar coordinates for rotational problems. For complex systems like the double pendulum, where forces become extremely difficult to write down, the Lagrangian approach is straightforward. This democratizes physics problem-solving, allowing mathematicians to tackle mechanics problems systematically.

Thermodynamics is the branch of physics that studies the transformation of heat into work, originating from the Industrial Revolution's steam machines. A thermodynamic system is a region enclosing working substance (typically ideal gas), classified as open (matter and energy transfer), closed (energy transfer only), or isolated (no transfer). An ideal gas has non-interacting molecules, negligible molecular size, and elastic collisions, described by five parameters: pressure, volume, temperature, number of moles, and internal energy. The ideal gas law PV = nRT (where R = 8.31 J/mol·K) defines the state equation, while processes are changes of state following P₁V₁/T₁ = P₂V₂/T₂. The First Law of Thermodynamics states that heat added equals work done plus change in internal energy (Q = W + ΔU), with specific sign conventions: Q positive when gaining heat, W positive when gas expands, and ΔU positive when temperature increases. Key processes include isobaric (constant pressure, Charles's Law), isochoric (constant volume, Gay-Lussac's Law), isothermal (constant temperature, Boyle's Law), and adiabatic (no heat transfer, PV^γ = constant).

Basic Thermodynamics is one of the most important subjects for SSC JE exam, carrying approximately 52-60 marks out of 100 annually. The subject is divided into four parts: Basic Thermodynamics, IC Engine, Power Plant Engineering, and Refrigeration and Air Conditioning. Mastering Basic Thermodynamics provides the foundation for understanding all other subjects in this domain. The syllabus follows a structured sequence: Basic Introduction, Zeroth Law, First Law, Second Law, Entropy, Availability and Irreversibility, and Pure Substances. This sequential approach ensures students build foundational knowledge before advancing to complex topics. The course emphasizes understanding concepts deeply rather than just memorizing notes.

Thermodynamics consists of two main parts: Basic Thermodynamics covering system definition, function, First and Second Laws, open/closed systems, and cycles; and Application Thermodynamics covering Power Plant Engineering (steam and gas turbines), Internal Combustion Engines, and Refrigeration and Air Conditioning. The application part carries significant weightage in GATE examinations, with Refrigeration and Air Conditioning alone accounting for approximately 425 marks out of 210 total marks in the application section.

Thermodynamics in Mechanical Engineering is divided into two main parts: Basic Thermodynamics and Applied Thermodynamics. Basic Thermodynamics covers fundamental concepts including thermodynamic systems, surroundings, universe, properties, cycles, enthalpy, entropy, internal energy, and the Laws of Thermodynamics (Zeroth, First, Second, and Third). Applied Thermodynamics applies these fundamentals to practical engineering applications including Heat Transfer, Power Plant Engineering (Thermal and Gas Power Plants), Internal Combustion Engines, and Refrigeration and Air Conditioning.

Thermodynamics combines 'thermo' (heat) and 'dynamics' (power), coined by Lord Kelvin in 1849. A system is matter or region under study, classified as control mass (fixed mass) or control volume (fixed region). Properties are measurable characteristics independent of past history, divided into intensive (pressure, temperature, density) and extensive (mass, volume, entropy). The zeroth law establishes thermal equilibrium through a third body, enabling temperature measurement. Temperature scales use arbitrary ice and steam points, with Kelvin scale offset by 273.15 from Celsius. Absolute zero (-273.15°C) represents complete cessation of molecular motion. Pressure arises from molecular collisions with container walls, depending on molecular number density, velocity, and mass. The ideal gas equation PV = nRT emerges from experiments showing PV/(nT) approaches the universal gas constant Ru = 8.314 kJ/kmol·K at low pressures. Two key assumptions enable ideal gas behavior: molecular volume is negligible and intermolecular forces are negligible at low pressures. The compressibility factor Z = PV_actual/(nRT) quantifies deviations from ideality, approaching 1 at low pressures and high temperatures. The first law is the law of conservation of energy applied to thermodynamic systems: energy in minus energy out equals change in system energy. For closed stationary systems, Q - W = ΔU, where Q is heat added (positive), W is work done by system (positive), and ΔU is internal energy change. Energy comprises macroscopic forms (kinetic from bulk motion, potential from position) and microscopic forms (molecular kinetic, intermolecular potential, lattice vibrations) included in internal energy. The Clausius inequality states that for any cyclic process, ∮(δQ/T) ≤ 0, with equality for reversible cycles and inequality for irreversible ones. This leads to two fundamental conclusions: entropy is a point function (state variable) independent of path, while entropy generation is a path function always ≥ 0 representing irreversibilities. The entropy change equation Δs = ∫(δQ_rev/T) + s_gen shows that entropy change depends on both reversible heat transfer and irreversibilities. Polytropic processes follow PV^n = constant, with n = 0 (isobaric), n = 1 (isothermal), and n = γ (isentropic) representing different behaviors. Isentropic processes (reversible adiabatic) have Δs = 0, representing idealized maximum efficiency processes used in cycle analysis.