Retrofitting MJTs onto combustion gas turbine. Industrial combustion gas turbines are well known for being especially hard to bolt and unbolt. No only are the bolts and nuts very large, they are not forgiving and will leak when not preloaded properly since the flanges are metal-to-metal with no gaskets. The problem can be exacerbated when the bolts are in a hard-to-access area or lack heating holes for thermal tightening.

One particular unit in a refinery suffered from all of these problems with a large number of its bolts. Some bolts literally took hours to bolt and unbolt, especially when they had not been opened in a long time. Sometimes bolts could not be freed and would have to be cut with a torch to be removed. In addition, the unit was notorious for developing hot air leaks after being overhauled and, consequently, would have to be shut down for bolt retightening. Even then the leak would not completely disappear, since the casing faces were distorted due to previous leaks.

All of these problems made this unit a very good candidate to replace the original turbine bolting with MJTs. The first step was to decide which bolts to replace with MJTs during the next unit overhaul. The unit had many bolts and it would have been uneconomical to replace them all. Instead, efforts were concentrated on selecting a few large-diameter bolts in inaccessible and critical areas to be replaced. Fig. 4 shows a cross-section of the gas turbine casing, with arrows pointing to the bolts that were selected to be changed. As can be seen from Fig. 5, these bolts are situated at the corners of the casings and are among the largest in diameter (2 in.). Furthermore, the original bolts had a large hexagonal socket cap, which meant that a special socket had to be used to torque them. It was felt that replacing these bolts would not only take care of the most problematic bolts in the area but would also have the most positive effect on reducing leaks.

Once it was decided which bolts would be retrofitted with MJTs, a custom design was developed with help from the manufacturer. The bolt would have to fit in the recessed area intended for the original bolt. It would also have to allow enough clearance above it to access the jacking bolts. Since detailed drawings for the bolts and the casing were not available, detailed measurements of some spare bolts along with the casing had to be taken. Once these details were collected, they were sent to the manufacturer. It came up with a design similar to what is shown in Fig. 6.

The next step was to come up with an appropriate material for this application, taking into account both the designed preload of the bolting and the temperature that the bolts will have to withstand. After some background search as well as consultation with different turbine manufacturers, a preload close to the original design was selected that would help in minimizing the leak problem. Actual casing operating temperature was measured by taking a thermal image of the unit when it was running (Fig. 7). Consequently, an estimate of the maximum temperatures the bolting will be exposed to over its life was determined. Samples of the original bolting material were analyzed to try to guess its composition and the design constraints of the original bolting designer. A final selection of bolting materials was reached with help from the MJT manufacturer as well as in-house company engineering expertise. Finally, the designed bolts were manufactured and installed in a major unit overhaul.

MJT Assessment. We found that they were more expensive than the original OEM bolts (about 40% more in this particular application). In addition, most machinists do not know how to properly install MJTs, so some time has to be spent explaining the proper installation sequence. Having said this, we feel that MJTs have performed very well in our particular application. time required to install bolts has shrunk from about 30-60 min. to less than 5 min. per bolt. Moreover, all the problems that we were facing with the original bolts have now disappeared. No heavy hammering or heating with torches is required to set these bolts. In our application, the required torque has decreased from the original 3,313 ft lb to 65 ft lb (an advantage of more than 50:1). This new torque can be easily and accurately applied by hand. This is achieved with confidence that the bolts were tightened to the target preload.

As has been mentioned earlier, this turbine had a long history of developing leaks in its horizontal and vertical flanges. In the past, the leaks were so severe tha tthey have caused unevenness in the turbine casing (Fig. 8). After MJTs were installed and the unit was run, some clear leaks did develop but only in the areas where the new bolts were not installed on the casing. This shows that these bolts are as good ormost likely better than the original bolts in applying the proper preload and preventing hot air leaks on the casing.

The decision to use MJTs should not be automatic; each bolting problem should be evaluated separately to decide if they can be used and are worth the engineering effort to retrofit them to solve the problem. Other bolting methods are very useful and effective in many applications. In our opinion, however, MJTs should at least be considered for large or especially critical bolting applications to improve safety and reliability. They should be considered for new installations that require large bolts or high torquing values and preloads.

ACKNOWLEDGMENTS

Many thanks go to a number of engineers working in various departments of Saudi Aramco for suggesting and supporting this project, particularly A. R. Al-Khowaiter from Aramco and the engineering staff at Superbolt Inc.

BIBLIOGRAPHY

Steinbock, R., "Multi-Bolt Mechanical Tensioners," ASME PVP - Vol. 158, Advances in Bolted Joint Technology, 1989.

Dekker, M., An introduction to the design and behavior of bolted joints, New York, 1981.