Frames, Grids & the Structural Bay

For thousands of years the structure of a building was its walls. In a load-bearing masonry building the outer and cross walls carry the floors and roof straight down to the ground, so the walls must be thick, continuous and pierced only sparingly for openings. This is honest and durable, but it fixes the plan: you cannot move a structural wall without rebuilding what it holds up, and you cannot make the walls mostly glass because the glass would carry nothing.

The skeleton frame separates these two jobs. A cage of columns (vertical members carrying compression) and beams (horizontal members carrying the floor loads in bending) takes all the structural weight and delivers it to the foundations. The walls are then demoted to mere enclosure - a skin hung on or filling in between the frame, carrying nothing but itself and the wind on its own face. This is why the modern office can have a fully glazed facade, an open floor plate and internal partitions that can be moved at will: none of them are holding the building up.

The consequences for design are enormous and worth naming. The plan becomes free - Le Corbusier called the open plan and the free facade two of his five points of architecture, and both depend on the frame. Floors become repeatable, so a tower is essentially one framed floor stacked many times. And the building's logic shifts from where are the walls to where are the columns - the grid. Reading a building as a frame rather than as a set of walls is the first mental move of this whole module.

Not all frames are equal, and the difference lies in the joints. In a post-and-beam frame the beams simply rest on the columns and the connections are effectively pinned - they can transfer vertical load, but they allow the beam and column to rotate freely relative to each other. This is the oldest frame there is, from Greek temples to a timber verandah: elegant, simple, and perfectly good at carrying gravity straight down. But a pinned frame has a fatal weakness on its own - push it sideways and it folds up like a cardboard box, because nothing at the joints resists the beam and column changing their angle. A pure post-and-beam frame cannot stand against wind or earthquake without help.

A rigid frame, also called a moment frame, fixes this by making the beam-to-column joints stiff - strong enough to resist rotation so that when the frame is pushed sideways, the joints fight back by developing bending moments. Beam and column now act as one continuous structure. A moment frame can therefore stand up to lateral load by itself, with no diagonal bracing and no walls, which keeps the bays open and the facade clear - a huge architectural advantage. The price is that the connections become large and demanding (a moment connection in steel is a serious, carefully detailed thing) and the members must be sized for that bending, so moment frames are heavier and costlier than the alternatives for the same height.

The practical world mixes the two. Many buildings carry gravity on simple pinned connections - cheap and fast - and resist lateral load separately through bracing or shear walls, reserving expensive moment frames for where their openness is truly needed. Knowing which kind of frame you are drawing tells you immediately how the building resists a push, and what it will cost to keep the bays clear.

Every building has to resist two very different demands: gravity pulling straight down, and lateral loads - wind and earthquake - shoving sideways. Carrying gravity is the easy part; a simple frame does it happily. The hard part, and the thing that separates a stable building from a house of cards, is resisting the sideways push without swaying too far or falling over. The system that does this is called the lateral load resisting system, and there are three great families of it.

The first is bracing: diagonal members added into the frame so that a rectangular bay, which racks (deforms into a parallelogram) under sideways load, is turned into stable triangles. A braced frame is extremely efficient - the diagonals carry the lateral load in simple tension and compression - which is why steel frames so often show X, K or chevron bracing, especially around lift cores and gable ends. The catch is architectural: a diagonal in a bay blocks that opening, so bracing is usually pushed into the walls of service cores or onto blank facades. The second family is the shear wall: a solid wall of concrete (or reinforced masonry) placed to act as a deep vertical cantilever that stiffens the building against sway, most often forming the lift-and-stair core at the centre of a plan. The third is the moment frame we just met, which needs no diagonals or walls but pays for it in connection cost.

An unbraced frame is one that relies purely on rigid joints (a moment frame) for its stability; a braced frame relies on those diagonals or walls, and can then use cheap pinned connections everywhere else. Most real buildings combine them - a braced core doing the heavy lateral work while simple gravity frames fill the floors around it. The architectural stakes are high, because where you put the bracing or the shear walls decides where the plan is permanently blocked and where it can stay open. Deciding the lateral system early, with the engineer, is one of the most design-critical conversations in the whole project.

Lay the columns out and you have drawn a grid - usually a rectangular pattern of gridlines, lettered one way and numbered the other, so any column can be named (A-1, C-4) and any point located. The rectangle of structure bounded by four columns is a bay, and the bay is the true module of a framed building. Its two dimensions are the column spacing (how far apart the columns march along a line) and the span (how far the beams reach across between the lines). Almost every quantity that matters flows from these two numbers.

The span, above all, sets the depth of the floor structure: beams get deeper as they span further, roughly in proportion, so a long-span bay means a deep floor zone, which either raises the building or eats into the ceiling height and the space for ducts. A short span gives a shallow, economical floor but plants columns close together, cluttering the plan. Somewhere in between sits the sweet spot for each building type - a car park wants columns spaced to suit parking bays, an office wants a column-free depth that suits desks and a lettable floor plate, a warehouse wants the longest clear span it can afford. The bay is where structural economy and spatial usefulness are negotiated.

The grid does far more than locate columns. It becomes the planning module the partitions align to, the facade module the mullions and cladding panels repeat on, the ceiling module the tiles and lights follow, and the coordination grid every other discipline references. When the structural grid, the facade module and the services grid are chosen together and share a common rhythm, the building is calm and buildable. When they are chosen in isolation - a structural grid that ignores the parking layout, or a facade module that does not divide into the column spacing - the mismatches show up as awkward junctions, wasted material and endless site queries. Choosing the bay is therefore not a narrow engineering act; it is one of the earliest and most powerful architectural decisions in the project, usually best made hand-in-hand with the structural engineer.

Frames reward regularity. A grid that repeats cleanly - columns stacking vertically floor over floor, bays roughly equal, beams running in consistent directions - gives a building that is cheaper to build, easier to coordinate and, crucially, better behaved in an earthquake, because regular structures have smooth, predictable load paths and no sudden weak spots. Irregular frames - columns that stop partway up, bays that jump in size, big offsets in plan - concentrate forces where the geometry breaks and are a well-known source of seismic failure. IS 1893 explicitly penalises plan and vertical irregularity for exactly this reason. So the first rule of framing is: be as regular as the architecture honestly allows.

But architecture rarely wants a perfectly uniform grid all the way down, and this is where the transfer structure earns its keep. A common real situation is a tower of small residential bays sitting over a ground floor that needs a big column-free lobby or a clear retail span or a basement car park on a different grid. The columns above cannot all continue down, so their loads are gathered by a deep transfer beam or transfer slab and redistributed to a sparser set of columns below. Transfers are powerful and everyday, but they are expensive, deep and demanding, and they interrupt the clean load path - so the design instinct should be to minimise them, not to scatter them thoughtlessly because the grid was never coordinated.

The mature attitude to the frame is therefore twofold. First, treat the grid as a shared design decision made early and holding the plan, the structure, the facade and the services in one rhythm - because retrofitting a grid onto a finished plan is painful and wasteful. Second, know that every irregularity, cantilever, transfer and change of grid is a deliberate cost you are choosing to spend for an architectural reason, not an accident to be discovered late by the engineer. Command of the frame is command of that trade: knowing where to be disciplined and regular, and where a considered exception is worth its price.