The flow temperature of brazing alloys approaches or exceeds the lower phase transition temperature limit of steel alloys, and it is nearly impossible to manually control the amount of heat applied to the work to prevent exceeding the LPTTL, and it is nearly impossible to achieve, let alone maintain, even heating of the work while staying below the LPTTL and preventing the creation of multiple heat affected zones and/or thermal stress transition regions. Any amount of heating/cooling will cause dimensional changes in steel and such changes can very well become permanent once the LPTTL is reached and depending on the ambient and applied conditions, such permanent changes can be realized even when temperatures remain below the LPTTL.
If you were to place one end of a barrel in a solid fixture and allow the other end to hang free, then place dial indicators on the free-end at the 12, 2, 4 & 6 o'clock positions, as you applied even a soft LPG/ambient air flame, the movement of those dial indicators would scare the $#@* out you. It's only been in the last 15 years or so that thermal transition and stress concerns have really come to light and only because there have been so many engineering disasters for which the "real" causes are finally coming to light. The name and location of the structure escape me at the moment but for years the collapse was blamed on the contractor & workers, only to protect the "famous architect" ... but decades later it finally comes out that the sole cause of the failure was failure of the "famous architect" to take the applied thermal stresses of the fabrication process and need for post-fabrication heat-treatment into account ... of course, had the "famous architect" bothered to use proper safety factoring to account for non-typical loading, the post-fabrication heat-treatment of the structural members would not have resulted in total catastrophic collapse of the structure. Just more proof as to how ignorance and assumptions results in disaster.
Barrel liners do NOT eliminate stresses on the existing barrel - even if the liner is thick enough to safely contain the working pressure of the load, it is by no means thick enough to prevent being subject to the radial displacement caused by the annular pressure waves nor the linear displacement caused by the linear pressure forces not to mention the multidimensional changes resulting from thermal variations. Most liners are of insufficient structural strength to contain the normal operating pressures of the load without having the additional structural support of the existing barrel. Utilizing a non-metal like epoxy to affix a liner is not without issue, there's been a rash of liner affixation and performance failures in recent years which primarily boil down to being caused by the use of inferior materials and/or improper preparation which in itself boils down to ignorance and assumptions like "there's no stress transferred from the liner to the existing barrel". Starting with the materials selection, if the liner material is not of similar alloy to the existing barrel, failure should be expected irregardless of the additional issues. Design an fitting of the liner to barrel joint is also a critical factor as the use of adhesive/filler material that has a lower proportional limit than the liner and existing barrel alloy will result in failure. Inferior materials primarily relates to the type of epoxy as there are thousands of specific blends, each designed for a specific application yet it's nearly impossible to explain the fact that generic and hardware store epoxy bends are best left sitting on the seller's shelf. Two of the most common liner failures being seen are slippage where, as the name implies, the liner slips out of place; the second is loss of performance which is most often related to two causes, use of inferior adhesive and improper joint preparation. If there joint is not at or near an interference fit, one is relying upon the adhesive/filler material to take up the slack where if the adhesive is of the non-metal type, the result is structural failure most often seen as the epoxy becoming pulverized into powder. In the case of a metal adhesive/filler such as solder, a loose fit will allow for displacement of the material with the low proportional limit resulting in asymmetrical response of the liner to the operating conditions of the load. That's just an overview as liners are a topic unto themselves. In short, anyone who tells you that an assembly formed with a typical barrel liner is not highly dependent upon the joint preparation, adhesive material and the existing barrel is full of $#@* - it's an "assembly", not a stand-alone unit and every "assembly" is dependent upon its individual components and anything that affects one component also affects the entire assembly.
