Bolting and the Space Shuttle Challenger Disaster

The decision was made to cut the handle with a hacksaw. A maintenance worker ran over from the service building while live TV showed President Reagan watching the drama from the Oval office. Eventually, after being stopped by security, the worker took the elevator up the side of the 184-foot shuttle. When he reached the door, the team used the tool until the handle broke off and fell between the gangway and the shuttle.

By now, a weather front had come into the cape and the winds started to climb past the threshold of the shuttle’s “return to launch site emergency constraints.” NASA decided to abort the launch and try again the next day.

The story of the hatch bolt seizure that happened the day before the Challenger disaster is not often discussed. Instead, the sealing of the O-ring in the booster rockets is named as the root cause for the horrendous accident that delayed NASA’s space advancement several years.

However, bolting is important and did contribute to this horrible accident. And just as it did on the space shuttle, bolting can cause headaches for plant operations and create safety hazards.

Tightening Valve Glands

A main issue of debate in industry has been how to accurately apply bolt load to valve packing. When tightened in excess, valve packing can cause major damage to the valve and interfere with general operations. But if the load is too low, the valve packing will not be able to maintain an adequate seal, resulting in costly leakage.

Many years ago, the term that would show up in installation instructions for valve packing was to use “the skill of the craft.” This relied on tribal knowledge to tighten valve glands but was not a particularly accurate measurement.

Older valve packing designs had a loose core of fibers that in some cases could cause a blow-out failure mode when load was lost. This caused many injuries to maintenance workers and resulted in people applying extra load to valve packing in attempt to prevent another accident. This extra load would usually take the form of using a pipe (or cheater bar) to add to the length of a wrench tightening the gland follower bolts. This would create extreme gland forces that would then create havoc with the valve’s operation.

The three main issues that can arise from excessive gland load on valve packing bolts are:

Bending of glands. The stress on the bolts can cause the gland to bend and distort so they are not perpendicular to the stuffing box. This distortion also causes the gland follower not to move into the stuffing box freely resulting in loss of load and failure.

Bent stems. With the excess gland forces, the packing increases in packing friction, which makes it impossible to use the handwheel to operate the valve. One way around this issue is using cheater bars on the handwheel to increase the force and allow the stem to move. But the loads applied by increasing the fulcrum distance can be so strong that they exceed the yield strength of the stem and cause it to bend. A bent stem will create a serious valve operability issue and serious leakage, since the packing needs to compensate the change in diameter.

Seat leakage. When using a cheater bar on the handwheel, the amount of force that strikes the seat to close the valve is often more than the design engineers had calculated. This added force can exceed the limits of the seat material, creating cracks resulting in internal leakage.

In the 1980s the Electric Power Research Institute (EPRI) was tasked with lowering the frequency of unscheduled maintenance in nuclear power plants. One issue they flagged was the high percentage of lost power generation because of leaking valves. Through its research, EPRI showed the importance of correct loading for longer packing service life. They concluded that using the “skill of the craft” method resulted in a large percent of scatter on loading accuracy.

Torque

One specific change after the landmark EPRI report was to apply gland load by using a torque wrench. Torqued valves started to become the norm and were written into packing procedures, training manuals and textbooks. Torque changed the industry, but it also has its drawbacks.

Studies have shown that the accuracy of applying load using a torque wrench is only about 30%. The main problem with these studies is they use bolts that are in good condition and have been maintained with good bolting practices (flat washer, thread condition, anti-seize, etc.). This is often not the practice in real world situations.

Without these practices, a false positive will often be seen in the torque wrench and will result in an inaccurate load being applied. Bolt condition is also one of those issues that could result in a false positive and one of the only ways to ensure proper loading of the packing is to replace studs when repacking in the field. Something as simple as a bent thread on a new or old bolt can give the illusion of a tight bolt without any load being applied to the gland.

Valve packing is a wearing part in a valve. Even with precise loading of the packing, over time as the valve is opened and closed, the packing will lose volume and consolidate. This reduction in shape will move the gland into the stuffing box to take up this loss.

The bolts that are stretched a small distance will loosen, resulting in lowering the gland force on the packing. This effect happens every time a valve is actuated. This loss of load is hidden.

One technology that started around the EPRI report was using Belleville springs sets that can be added to the bolting on a valve, resulting in more travel in the joint. This is called live loading and results in reducing the loss of load over time, resulting in valves being kept in service longer.

But one thing that live loading can’t overcome is imprecise gland loads. The inaccuracy of torque still results in premature failure of packing, even with live loading.

Another issue that is seldom talked about, but is well known in bolting circles, is that re-applying a torque value in the field adds drastic errors. Dried anti-seize does not have the same K factor as when it was applied while it was in a wet state. The scatter is wide and hard to predict. People think they are re-energizing a valve when they are re-applying the same torque value while the valve is in service but are actually applying 50% less load.

Cartridge Live Loading

A modern technology that has changed the entire relationship between correct loading of packing is cartridge live loading. This also uses Belleville springs, but in an outer cup design that is cut to a precise height. This is a way to ensure a very accurate load when the flat washer on the top of the springs come even with the cup.

This simple design is the next generation of accurate valve loading and does not rely on a torque wrench to apply gland load. Instead it uses the spring compression. In addition to getting the spring travel and extending the time of bolt load loss, it also gives a visual aid of the actual valve loading while in service.

With this technology, valves can be checked to see if the assembly is still loaded correctly before it starts to leak and by using a simple box wrench re-loaded to the correct load (not torque). These spring assemblies, like all live loading, must be sized by valve dimensions (stem diameter, stuffing box OD, stud diameter, clearances) and the valve’s operating conditions like system pressure and temperature.

With cartridge live loading, plant maintenance teams now have a consistent and reliable way to ensure accurate load packing that will result in fewer packing failures and longer reliability.

Ron Frisard is field product manager for global marketing at A.W. Chesterton. Frisard is also chairman of the Packing Division of the Fluid Sealing Association.