The principle of virtual work provides an alternative to the direct force-equilibrium approach for finding unknown forces in systems of connected rigid bodies. It states that if a system is in equilibrium, the total virtual work done by all external forces through any compatible virtual displacement is zero: delta_U = ΣF . delta_r + ΣM . delta_theta = 0. A virtual displacement is an imaginary, infinitesimally small displacement consistent with the system's geometric constraints. The power of this method is that constraint forces (pin reactions, normal forces at smooth contacts) do no virtual work because their points of application move perpendicular to the forces or not at all, so they drop out entirely. This reduces a multi-body equilibrium problem with many internal reactions to a single scalar equation involving only the active (applied) forces and the unknown of interest. For conservative systems, virtual work can be reformulated using potential energy: equilibrium occurs where dV/dq = 0, and the stability of that equilibrium depends on the sign of d^2V/dq^2.
Start with single-DOF mechanisms (toggle clamps, scissors lifts, linkages) where one coordinate q defines the configuration. Express every active force's displacement in terms of delta_q, apply delta_U = 0, and solve for the unknown. Then verify the result with a conventional FBD approach to build confidence. Practice the potential energy method on spring-gravity systems to classify equilibrium as stable (d^2V/dq^2 > 0), unstable (< 0), or neutral (= 0).
In equilibrium analysis using free-body diagrams, you cut a body apart, expose all reaction forces, and write ΣF = 0 and ΣM = 0. This works well for a single body. But for a mechanism with multiple connected links — a scissors jack, a toggle clamp, a robotic arm — every pin connecting one link to another introduces unknown reaction components, and the system of equations grows quickly. The principle of virtual work offers a completely different strategy: instead of exposing the internal reactions and solving for them, you make the system move an infinitesimal imaginary amount and ask how much work would be done. If the system is in equilibrium, the answer must be zero.
A virtual displacement δr is a hypothetical, infinitesimally small motion consistent with the geometric constraints — the links and pins are still connected, the wheels still roll on the ground, and so on. It is not an actual motion that occurs in time; it is a geometric probe. The crucial insight is that constraint forces do no virtual work: a pin exerts equal and opposite forces on the two bodies it connects, and those bodies move the same amount at the pin location, so the works cancel. The ground normal force under a wheel does no virtual work because the contact point cannot move vertically. These forces, which would appear as unknowns in a free-body diagram approach, simply drop out of the virtual work equation. What remains is a single equation involving only the active forces — the applied loads, springs, gravity — and any one unknown you choose to leave in.
The procedure for a single-degree-of-freedom mechanism is: (1) choose a generalized coordinate q that describes the configuration (say, the angle of a crank or the extension of a slider); (2) express the position of every active force's point of application in terms of q; (3) differentiate to find the virtual displacements δr_i in terms of δq; (4) write ΣF_i · δr_i = 0, which factors as (some expression) · δq = 0; since δq is arbitrary and nonzero, the expression in parentheses must equal zero. This gives you one equation for one unknown — no need to find any reactions at pins or smooth contacts.
For conservative systems (springs and gravity only, no friction), the virtual work principle takes its most elegant form: potential energy V is a function of q, and equilibrium requires dV/dq = 0. This is simply the condition that V is stationary with respect to configuration. The sign of the second derivative tells you the stability: d²V/dq² > 0 means the equilibrium is stable (a valley — perturbations return the system), d²V/dq² < 0 means unstable (a hill — perturbations grow), and d²V/dq² = 0 means neutral (a flat — perturbations neither grow nor decay). The potential energy method is the most compact tool available for equilibrium and stability analysis of conservative mechanisms.