As to question concerning "cushions between steel and concrete", Mr. Russell Barnard of ARMCO Steel delivered a paper at the 1946 AWWA Annual Convention on the development of steel water pipe standards (that was reprinted as a stand alone pamphlet in 1948). This paper/reprint referred to "suitably padded concrete saddles" and that the padding consist of "several layers of felt with graphite in between", or instead pipe can be supported with ring girders or flange rings welded to the pipe. In 1970, the "Welded Steel Water Pipe Manual" of the Steel Plate Fabricators Association (SPFA) provided the most elaboration on shape and dimensional information on supports to date, and in one location where axial movement was to be lubricated referred to "two layers" of "sheet packing with graphite grease". It also mentioned "it is good practice to construct the saddles with a radius about 1% larger than that of the outside of the pipe". While the reason(s) for this suggestion was not provided, I have a hunch unstated candidates could well include minimization of the situation mentioned by another on this thread, where a flexible pipe (that by nature often has at least slightly imperfect circularity) might rest predominantly on the upper corners instead of the arc of the cushion (increasing at least some localized stress there) and/or that some designers might not realize the geometric reality that any thickness of "cushion" effectively makes the radius of the cushion contact arc smaller than any at least the underlying pre-formed structural pier or saddle form, and/or also that it may really not hurt even a perfectly shaped flexible pipe when it deflects very slightly under weight reaction load to bear well on a slightly larger radius arc. AWWA manuals M41 Ductile-Iron Pipe and Fittings as well as M11 Steel Water Pipe a Guide to Design and Installation provide similar calculation procedures for determining maximum localized stresses at saddle supported piping locations, and with regard to the former well-instrumented and published research at American Cast Iron Pipe Company found that for pipes that fit the test saddles well that level of maximum localized stress might occur a few degrees above the saddle tips, or instead a short distance axially off the edge of the saddles at the bottom of the pipes. As to cushioning or lubrication means, the verbiage in the fourth edition of AWWA M11 for steel pipe is remarkably similar to that of Barnard nearly six decades earlier, reading "several layers of felt with graphite in between or joint filler material" (with really only the last four words, it appears allowing some alternative, added since). As to rubber sheet (as one responder on this thread has mentioned "neoprene"), it should probably be mentioned (and while it is probably mostly academic) while a layer of graphite is indeed a "lubricant", at least unlubricated rubber is on the other hand a very high coefficient of friction material. In a water pump stations and plants with generally rather short reaches of pipes, legacy materials with very low coefficients of expansion, flanged or other restrained joints, and often also not huge swings in temperature (even well strapped down), this difference may be of little or no consequence as rubber sheets and other methods not mentioned in the manuals are commonly used. [I only mention this in that much greater thermal axial loads might be encountered on piers or supports at long spans and reaches of pipes strapped tightly down on rubber pads than when lubricated saddles or rollers/bracketed rollers etc. are employed.] Other than explaining the effects on maximum localized stress due to varying saddle angles and axial width, and mentioning other considerations and support mechanisms in e.g. bridge crossings etc., AWWA M41 on ductile iron pipe has not really gotten into specific support design and construction, perhaps due to the fact the support field has so greatly expanded, with a huge variety/numbers of support schemes and specialized support schemes/vendors used for insulated and non insulated pipelines it appears with generally good success.
