Most bridges are fixed bridges, meaning they have no moving parts and stay in one place until they fail or are demolished. Temporary bridges, such as Bailey bridges, are designed to be assembled, and taken apart, transported to a different site, and re-used. They are important in military engineering, and are also used to carry traffic while an old bridge is being rebuilt. Movable bridges are designed to move out of the way of boats or other kinds of traffic, which would otherwise be too tall to fit. These are generally electrically powered.
Double-decked (or double-decker) bridges have two levels, such as the George Washington Bridge, connecting New York City to Bergen County, New Jersey, US, as the world's busiest bridge, carrying 102 million vehicles annually; truss work between the roadway levels provided stiffness to the roadways and reduced movement of the upper level when the lower level was installed three decades after the upper level. The Tsing Ma Bridge and Kap Shui Mun Bridge in Hong Kong have six lanes on their upper decks, and on their lower decks there are two lanes and a pair of tracks for MTR metro trains. Some double-decked bridges only use one level for street traffic; the Washington Avenue Bridge in Minneapolis reserves its lower level for automobile and light rail traffic and its upper level for pedestrian and bicycle traffic (predominantly students at the University of Minnesota). Likewise, in Toronto, the Prince Edward Viaduct has five lanes of motor traffic, bicycle lanes, and sidewalks on its upper deck; and a pair of tracks for the Bloor–Danforth subway line on its lower deck. The western span of the San Francisco–Oakland Bay Bridge also has two levels.
Robert Stephenson's High Level Bridge across the River Tyne in Newcastle upon Tyne, completed in 1849, is an early example of a double-decked bridge. The upper level carries a railway, and the lower level is used for road traffic. Other examples include Britannia Bridge over the Menai Strait and Craigavon Bridge in Derry, Northern Ireland. The Oresund Bridge between Copenhagen and Malmö consists of a four-lane highway on the upper level and a pair of railway tracks at the lower level. Tower Bridge in London is different example of a double-decked bridge, with the central section consisting of a low-level bascule span and a high-level footbridge.
A viaduct is made up of multiple bridges connected into one longer structure. The longest and some of the highest bridges are viaducts, such as the Lake Pontchartrain Causeway and Millau Viaduct.
A multi-way bridge has three or more separate spans which meet near the center of the bridge. Multi-way bridges with only three spans appear as a "T" or "Y" when viewed from above. Multi-way bridges are extremely rare. The Tridge, Margaret Bridge, and Zanesville Y-Bridge are examples.
A bridge can be categorized by what it is designed to carry, such as trains, pedestrian or road traffic (road bridge), a pipeline or waterway for water transport or barge traffic. An aqueduct is a bridge that carries water, resembling a viaduct, which is a bridge that connects points of equal height. A road-rail bridge carries both road and rail traffic. Overway is a term for a bridge that separates incompatible intersecting traffic, especially road and rail. A bridge can carry overhead power lines as does the Storstrøm Bridge.
Some bridges accommodate other purposes, such as the tower of Nový Most Bridge in Bratislava, which features a restaurant, or a bridge-restaurant which is a bridge built to serve as a restaurant. Other suspension bridge towers carry transmission antennas.
Conservationists use wildlife overpasses to reduce habitat fragmentation and animal-vehicle collisions. The first animal bridges sprung up in France in the 1950s, and these types of bridges are now used worldwide to protect both large and small wildlife.
Bridges are subject to unplanned uses as well. The areas underneath some bridges have become makeshift shelters and homes to homeless people, and the undertimbers of bridges all around the world are spots of prevalent graffiti. Some bridges attract people attempting suicide, and become known as suicide bridges.
The materials used to build the structure are also used to categorize bridges. Until the end of the 18th century, bridges were made out of timber, stone and masonry. Modern bridges are currently built in concrete, steel, fiber reinforced polymers (FRP), stainless steel or combinations of those materials. Living bridges have been constructed of live plants such as Ficus elastica tree roots in India and wisteria vines in Japan.
The Tank bridge transporter (TBT) has the same cross-country performance as a tank even when fully loaded. It can deploy, drop off and load bridges independently, but it cannot recover them.
Unlike buildings whose design is led by architects, bridges are usually designed by engineers. This follows from the importance of the engineering requirements; namely spanning the obstacle and having the durability to survive, with minimal maintenance, in an aggressive outdoor environment. Bridges are first analysed; the bending moment and shear force distributions are calculated due to the applied loads. For this, the finite element method is the most popular. The analysis can be one, two or three-dimensional. For the majority of bridges, a two-dimensional plate model (often with stiffening beams) is sufficient or an upstand finite element model. On completion of the analysis, the bridge is designed to resist the applied bending moments and shear forces, section sizes are selected with sufficient capacity to resist the stresses. Many bridges are made of prestressed concrete which has good durability properties, either by pre-tensioning of beams prior to installation or post-tensioning on site.
In most countries, bridges, like other structures, are designed according to Load and Resistance Factor Design (LRFD) principles. In simple terms, this means that the load is factored up by a factor greater than unity, while the resistance or capacity of the structure is factored down, by a factor less than unity. The effect of the factored load (stress, bending moment) should be less than the factored resistance to that effect. Both of these factors allow for uncertainty and are greater when the uncertainty is greater.
Most bridges are utilitarian in appearance, but in some cases, the appearance of the bridge can have great importance. Often, this is the case with a large bridge that serves as an entrance to a city, or crosses over a main harbor entrance. These are sometimes known as signature bridges. Designers of bridges in parks and along parkways often place more importance to aesthetics, as well. Examples include the stone-faced bridges along the Taconic State Parkway in New York.
To create a beautiful image, some bridges are built much taller than necessary. This type, often found in east-Asian style gardens, is called a Moon bridge, evoking a rising full moon. Other garden bridges may cross only a dry bed of stream washed pebbles, intended only to convey an impression of a stream. Often in palaces a bridge will be built over an artificial waterway as symbolic of a passage to an important place or state of mind. A set of five bridges cross a sinuous waterway in an important courtyard of the Forbidden City in Beijing, China. The central bridge was reserved exclusively for the use of the Emperor and Empress, with their attendants.
The estimated life of bridges vary between 25 to 80 years depending on location and material. However, bridges may age hundred years with proper maintenance and rehabilitation. Bridge maintenance consisting of a combination of structural health monitoring and testing. This is regulated in country-specific engineer standards and includes an ongoing monitoring every three to six months, a simple test or inspection every two to three years and a major inspection every six to ten years. In Europe, the cost of maintenance is considerable and is higher in some countries than spending on new bridges. The lifetime of welded steel bridges can be significantly extended by aftertreatment of the weld transitions. This results in a potential high benefit, using existing bridges far beyond the planned lifetime.
While the response of a bridge to the applied loading is well understood, the applied traffic loading itself is still the subject of research. This is a statistical problem as loading is highly variable, particularly for road bridges. Load Effects in bridges (stresses, bending moments) are designed for using the principles of Load and Resistance Factor Design. Before factoring to allow for uncertainty, the load effect is generally considered to be the maximum characteristic value in a specified return period. Notably, in Europe, it is the maximum value expected in 1000 years.
Bridge standards generally include a load model, deemed to represent the characteristic maximum load to be expected in the return period. In the past, these load models were agreed by standard drafting committees of experts but today, this situation is changing. It is now possible to measure the components of bridge traffic load, to weigh trucks, using weigh-in-motion (WIM) technologies. With extensive WIM databases, it is possible to calculate the maximum expected load effect in the specified return period. This is an active area of research, addressing issues of opposing direction lanes, side-by-side (same direction) lanes, traffic growth, permit/non-permit vehicles and long-span bridges (see below). Rather than repeat this complex process every time a bridge is to be designed, standards authorities specify simplified notional load models, notably HL-93, intended to give the same load effects as the characteristic maximum values. The Eurocode is an example of a standard for bridge traffic loading that was developed in this way.
Most bridge standards are only applicable for short and medium spans - for example, the Eurocode is only applicable for loaded lengths up to 200 m. Longer spans are dealt with on a case-by-case basis. It is generally accepted that the intensity of load reduces as span increases because the probability of many trucks being closely spaced and extremely heavy reduces as the number of trucks involved increases. It is also generally assumed that short spans are governed by a small number of trucks traveling at high speed, with an allowance for dynamics. Longer spans on the other hand, are governed by congested traffic and no allowance for dynamics is needed. Calculating the loading due to congested traffic remains a challenge as there is a paucity of data on inter-vehicle gaps, both within-lane and inter-lane, in congested conditions. Weigh-in-Motion (WIM) systems provide data on inter-vehicle gaps but only operate well in free flowing traffic conditions. Some authors have used cameras to measure gaps and vehicle lengths in jammed situations and have inferred weights from lengths using WIM data. Others have used microsimulation to generate typical clusters of vehicles on the bridge.
Bridges vibrate under load and this contributes, to a greater or lesser extent, to the stresses. Vibration and dynamics are generally more significant for slender structures such as pedestrian bridges and long-span road or rail bridges. One of the most famous examples is the Tacoma Narrows Bridge that collapsed shortly after being constructed due to excessive vibration. More recently, the Millennium Bridge in London vibrated excessively under pedestrian loading and was closed and retrofitted with a system of dampers. For smaller bridges, dynamics is not catastrophic but can contribute an added amplification to the stresses due to static effects. For example, the Eurocode for bridge loading specifies amplifications of between 10% and 70%, depending on the span, the number of traffic lanes and the type of stress (bending moment or shear force).
There have been many studies of the dynamic interaction between vehicles and bridges during vehicle crossing events. Fryba did pioneering work on the interaction of a moving load and an Euler-Bernoulli beam. With increased computing power, vehicle-bridge interaction (VBI) models have become ever more sophisticated. The concern is that one of the many natural frequencies associated with the vehicle will resonate with the bridge first natural frequency. The vehicle-related frequencies include body bounce and axle hop but there are also pseudo-frequencies associated with the vehicle's speed of crossing and there are many frequencies associated with the surface profile. Given the wide variety of heavy vehicles on road bridges, a statistical approach has been suggested, with VBI analyses carried out for many statically extreme loading events.
The failure of bridges is of special concern for structural engineers in trying to learn lessons vital to bridge design, construction and maintenance. The failure of bridges first assumed national interest during the Victorian era when many new designs were being built, often using new materials.
In the United States, the National Bridge Inventory tracks the structural evaluations of all bridges, including designations such as "structurally deficient" and "functionally obsolete".
There are several methods used to monitor the condition of large structures like bridges. Many long-span bridges are now routinely monitored with a range of sensors. Many types of sensors are used, including strain transducers, accelerometers, tiltmeters, and GPS. Accelerometers have the advantage that they are inertial, i.e., they do not require a reference point to measure from. This is often a problem for distance or deflection measurement, especially if the bridge is over water.
An option for structural-integrity monitoring is "non-contact monitoring", which uses the Doppler effect (Doppler shift). A laser beam from a Laser Doppler Vibrometer is directed at the point of interest, and the vibration amplitude and frequency are extracted from the Doppler shift of the laser beam frequency due to the motion of the surface. The advantage of this method is that the setup time for the equipment is faster and, unlike an accelerometer, this makes measurements possible on multiple structures in as short a time as possible. Additionally, this method can measure specific points on a bridge that might be difficult to access. However, vibrometers are relatively expensive and have the disadvantage that a reference point is needed to measure from.
Snapshots in time of the external condition of a bridge can be recorded using Lidar to aid bridge inspection. This can provide measurement of the bridge geometry (to facilitate the building of a computer model) but the accuracy is generally insufficient to measure bridge deflections under load.
While larger modern bridges are routinely monitored electronically, smaller bridges are generally inspected visually by trained inspectors. There is considerable research interest in the challenge of smaller bridges as they are often remote and do not have electrical power on site. Possible solutions are the installation of sensors on a specialist inspection vehicle and the use of its measurements as it drives over the bridge to infer information about the bridge condition. These vehicles can be equipped with accelerometers, gyrometers, Laser Doppler Vibrometers and some even have the capability to apply a resonant force to the road surface in order to dynamically excite the bridge at its resonant frequency.