High Temperature Bolt Load Measurement

The pre-requisite for high temperature measurement of bolt loads in service is measurement and control on installation tightening under ambient conditions.

Bolted joint reliability depends on assuring the joint design is correct, that all joint component quality is assured and then the design bolt tension/joint compression/gasket seating stress is assured on installation tightening. If all 3 reliability factors are measured and assured, bolted joint reliability is also assured. This technology driven, lowest operational cost, lowest cost of assurance can be delivered by a fit and forget strategy.

Out of necessity this strategy has to be applied for High temperature bolted joints especially where production runs are lengthy and even more so where expected (or unexpected) thermal swing takes place during the production process. This assured installation strategy has been extremely successful. What added value would there be to this fully assured strategy if bolt tension measurement could also be achieved at high service temperatures? Could current bolt tension measuring systems provide accurate reliable bolt tension indication at high service temperatures?

At the moment long term reliable, accurate, bolt tension measurement at high service temperatures is not possible. There are a number of factors that make such monitoring problematic. Bolting alloys are subject to metallurgical creep and oxidisation at elevated temperatures; both mechanisms are time temperature related. The effect on the bolt, from either mechanism, is its length grows – the bolt length extends in addition to any applied or residual elastic extension. Creep is a mechanism that induces elastic and plastic elongation of the fastener. Oxidation growth also increases the bolt length.

All calibration of current technology systems is made from the ambient datum length of the bolt whether direct length or gauge pin type technology is used. If a change in bolt length could be practically measured at high temperatures, the datum change renders any in service elastic reading inaccurate. The more relevant measurement systems use gauge pins. Short term, reliable measurement is possible where the bolt and gauge pin are metallurgically compatible and have good creep and oxidisation resistance. This invariably means bolts made from higher nickel alloys and ideally super alloys e.g. Inconel 718, Nimonics, alloy 625, Hastelloys etc. The measurement gauge pin has to be manufactured from the same material as the parent bolt.

Some experimental/ development measurement systems have used high temperature alloy gauge pins with standard petrochemical medium carbon low alloy steel bolts (B7, B16) used at elevated operating temperatures. The purpose was to make relative datum faces resist oxidisation and maintain/sustain their ambient datum tightening characteristics. By using a higher nickel alloy gauge pin in a B7/B16 and then additionally protecting the low alloy steel bolt datum surface with a similar super alloy barrier, original surfaces are maintained.

Unfortunately the relative creep rates between low alloy steels and super alloys are not accounted for. The ambient measurement datum is lost, irrespective of oxidisation resistance. Subsequently accurate bolt load measurement at service temperatures is not possible thru’ the normal elastic extension relationship. Petrochemical flange bolt staples, B7 and B16 steel, at bolt temperatures 350C and above, start to creep significantly. Relative creep is even higher at service temperatures of 400-550C. The bolt will grow plastically so it’s ambient datum changes. Despite the super alloy gauge pin being relatively stable from a creep and oxidisation viewpoint (no growth) a significant change in the original respective datum face calibration has taken place. The bolt stress suggested by the ambient relationship is likely to be substantially different and misleading.

Theoretical work evaluating super alloy bolting with compatible gauge pins e.g. Inco 718 at 800C, suggests that reasonable accuracy /reusability of +, – 5%, is possible for up to 3-4 months subject to installed bolt loads. Gauge pins are not subjected to stress compared to the in service bolt so their relative creep rates are different. Subject to the accuracy tolerance required, acceptable accuracy is possible for a period of time.  Beyond this time the measurement becomes inaccurate.

The high temperature ‘fit and forget’ assured installation strategy has been extremely successful but remarkably is still in its infancy and petrochemical process industries have still not fully taken advantage of the lowest cost of ownership that bolt tension measurement provides. Massive maintenance and delayed start-up costs savings are there for the taking. To accrue additional value from full in service bolt monitoring, the operator has to establish bolt measurement and control on ambient installation first.