I was walking through the street and saw that old steel bridges have lots of rivets in them. The newer ones have welds and bolts. How did field welding evolve when steel became plentiful for construction? How long did it take to displace rivets for connections, if at all?
The first type of welding is generally referred to as "forge welding". It is a process where two surfaces iron or steel are treated with a flux, such as borax, and then heated in a forge until almost burning (oxidizing), then they are hammered together, fusing the two surfaces into one. This process has been around for centuries. You can see it described in texts like Joseph Moxen's Mechanic Exercises. The process is rather difficult to perform on larger pieces, so it is mostly restricted to the blacksmith at the anvil.
Technically, the next welding process worked out was metal arc welding in the late 19th century. This eventually evolved into what is now called Stick welding or Shielded Metal Arc Welding (SMAW). But in its early incarnation, the welding electrode did not have the jacket of shielding chemicals. As a result, metal arc welds came out heavily oxidized, and as a result, far too weak to be used in structural projects. Another issue was that the process required a high amp power supply; and for structural processes would almost certainly require a generator in order to be portable.
The next welding process to occur can be generically referred to as "gas welding". Some fuel must be burned at a temperature decently higher than the melting point of iron/steel. The first reasonable source to satisfy this came from the electrolysis of water, which was first worked out in 1800. By using a source of electricity (original, the voltaic pile, the first feasible wet-cell battery) you can convert water into it's constituents, oxygen and hydrogen. This can then be burned. 2*H2 + O2 -> 2*H2O. Oxyhydrogen welding is still done today, but only in niche applications.
Oxyhydrogen (occasionally called HHO) welding was first marketed in 1901 out of Belgium. Chemically, it has a two problems. First, the H2O (in the form of steam) is not a terribly effective shield from the atmospheric oxygen; so the hot weld would immediately oxidize and weaken. Second, and usually considered a bigger problem, a small portion of hydrogen can become dissolved into the liquid metal before it gets a chance to react with the oxygen. This causes hydrogen porosity and hydrogen embrittlement, both of which seriously weakens welds.
As far as I've been able to find, the cause for the century wait between the discovery of electrolysis and the marketing of HHO welding was again, the great amount of electricity needed to produce enough gas to allow welding of metal of any reasonable thickness. Starting around the 1880s, the technology was developed to compress the isolated gasses into transportable steel cylinders, meaning the electrolysis did not need to take place on site.
For other fuels to burn hot enough, they must be mixed with pure oxygen before igniting; creating a far more efficient & hotter flame than burning with regular atmosphere. The first major source of commercial oxygen was from Brin's Oxygen Company, formed in 1886. The next fuel that burned hot enough to weld to be discovered was Acetylene, made by mixing lime (calcium oxide) with coke (purified coal) at high temperature to produce Calcium carbide, and then dripping water on that. When oxygen and acetylene are mixed and burned, it creates CO2, which is an excellent shield gas, that allows the weld to cool without oxidizing.
Unlike hydrogen and oxygen, acetylene cannot simply be compressed into an empty tank. When acetylene reaches a high enough pressure, it spontaneously explodes. In 1896, Albert Hess and Georges Claude worked out that the gas can be safely dissolved at high pressures into a cylinder filled with acetone and some sort of sediment. This is still used today, using diatomacious earth. Welding torches were patented by French Edmond Fouché and Charles Picard in 1902.
1903 saw the development of thermite welding. Thermite is a high temperature reaction between iron-oxide and aluminum powders, resulting in aluminum-oxide and iron. By containing the material within a matrix, it can fuse to steel. The primary application for this is in laying railroad track. The thermite weld bonds all the way through the very thick steel, and can be ground completely smooth, allowing for comfortable high-speed travel.
Oxyacetylene was fine for thinner plates; but was not appropriate for welding the thicker stock required for things like bridge construction. In 1909, patents were secured for the improved metal arc welding that used electrodes coated in chemicals that produce slag and shielding gasses that protect the hot welds from oxidation and nitrogen porosity. The components of this shield coatings continued to improve over the next few decades.
In 1920, Britain launched the HMS Fullagar, the first ship to have an all welded hull. Also that year, the first steel frame welded building was constructed.
In 1928, the Maurzyce Bridge in Poland became the first all-welded bridge. Redesigning the bridge to be welded instead of the original riveted plan reduced the weight by 20%.
In the 1930s, submerged arc welding was developed that allowed thick plate to be welded cleanly and quickly by hiding the electric arc beneath a shielding powder. The process is rarely used in the field, but it is used to weld plates in factory settings prior to shipping them to construction sites.
Prior to WWII, the Germans used welding techniques on battleships in order to save weight compared to rivets. This allowed them to build very fast, heavily armored, heavily armed ships while still remaining in compliance with maximum weight allowances for ships that were worked out in the treaties at the end of WWI.
During WWII, the US produced hundreds of liberty ships, which were all-welded construction. By the end of the war, the US was able to launch a new ship every 11 hours.
Riveted construction pretty much reached an end after WWII. By that point, welds were stronger, faster, and lighter weight than rivets and welding standards were in place to provide quality assurance for those welds.
Sources: