Coal, Nuclear, Reactors

New Annealing Process Proves its Worth

Issue 6 and Volume 103.

The First-ever Indirect, Gas-fired, Radiantly Heated Annealing of a Nuclear Reactor Has Been completed successfully, at the never-completed Marble Hill nuclear plant in Indiana. An important component of the Marble Hill demonstration project was the use of a computer simulation that validated the process in advance.


The nuclear power industry is looking at thermal annealing as a remedy for neutron-induced embrittlement of welds in nuclear reactor pressure vessels. The pressure vessel contains the fuel and control mechanism for the nuclear reactor and is located within the containment building.

Click here to enlarge image

After years of neutron bombardment, nuclear reactor pressure vessel welds become brittle. In time, embrittlement can reach damaging proportions, as demonstrated in 1992 when Yankee Atomic Electric Co. was forced to decommission its Rowe power plant outside Boston, eight years before its license was due to expire.

Annealing is the only method available for recovering the material properties of a reactor pressure vessel. Metallurgical experiments have shown that annealing will restore the vessel to near its original metallurgical properties. By restoring the ductility and fracture toughness of the weld metal, many years of operating life can be added to the reactor vessel. The only alternative to annealing is to change the way the vessel is fueled and reduce the rate of embrittlement by reducing neutron bombardment.

Annealing, however, is not without its challenges. The internal components of the reactor must be disassembled (a standard operation during refueling) and a means must be provided to heat the 35-foot tall, 15-foot diameter vessel to 850 F for one week. The annealing approach pioneered by Westinghouse Electric Corporation and Cooperheat-MQS Inc. involves blowing superheated air from gas-fired burners into a heat exchanger that in turn heats the reactor vessel. This approach was validated, prior to the $6 million dollar demonstration project at Marble Hill, by building a functional scale mockup and by simulating heat transfer and air flow using computational fluid dynamics (CFD) software.


In developing the heat exchanger, the size of the reactor vessel was not a major obstacle; Cooperheat-MQS regularly heat treats many larger pressure vessels for the petroleum industry. The difficulty of the task relates more to the fact that the inside of nuclear reactors are contaminated with radiological materials that cannot be released to the environment. This eliminates the most common method used to anneal large pressure vessels, which is circulating superheated air from gas burners inside the vessel and insulating the outside of the vessel to minimize heat loss. The problem with this method, however, is that the hot combustion gases would pick up radioactive dust and other contaminants inside the vessel and spread them into the atmosphere.

Several nuclear reactors in Russia have been annealed using radiant electric heat. Cooperheat-MQS engineers decided not to use this approach for several technical and economic reasons. Ironically, despite the fact that the pressure vessel was located within an electric generating facility, obtaining power for the heaters would have been very costly. Another problem with this approach is that a failed electric heater element would have been virtually impossible to repair or replace inside the reactor. The use of redundant electric heaters would have nearly doubled the cost and the weight of the furnace.

The Westinghouse/Cooperheat-MQS approach uses gas-fired burners to superheat air and blow it into sealed ducts that are routed through existing openings in the containment building, such as the equipment hatch, and into the heat exchanger inside the reactor vessel. The superheated air is then discharged outside of containment through another duct to the atmosphere. Since the air never comes into contact with any contaminated surfaces, it does not become contaminated. The gas-fired heaters are located outside the containment building so they can be easily replaced in case of failure.

CFD Refinements

The indirect heating approach clearly had the potential to eliminate many of the problems with electric heating, but it also raised several potential difficulties. The main one was insuring that the heat exchanger would be able to maintain temperature uniformity throughout the entire reactor vessel annealing zone. If adequate temperature uniformity is not achieved, excessive thermal stresses can develop. To verify that the reactor vessel would not be over-stressed, the project team built and tested a scale model of the heat exchanger and reactor vessel to experimentally estimate heat transfer coefficients. Project engineers also elected to use the CFD method to simulate flow distribution and heat transfer within the heat exchanger and confirm the experimental results.

Click here to enlarge image

A CFD analysis provides fluid velocity, pressure and temperature values throughout the solution domain for problems with complex geometries and boundary conditions. As part of the analysis, a researcher may change the geometry of the system or the boundary conditions such as inlet velocity, flow rate, etc. and view the effect on fluid flow patterns or concentration distributions. CFD also can provide detailed parametric studies that can significantly reduce the amount of experimentation necessary to develop prototype equipment and thus reduce design cycle times and costs.

Cooperheat-MQS engineers selected FIDAP CFD software from Fluent Inc., because FIDAP uses a finite element method that is ideal for generating the complex and irregular geometries involved in the proposed heat exchanger design. The flexibility of the mesh generation tool provided with this software package makes it possible to handle very odd shapes.

The project team first developed an initial design for the heat exchanger (Figure 1) by modeling the flow within the heat exchanger under the assumption that the flow would be uniform. The initial design verified the experimental heat transfer coefficients, based on the assumptions that were made. Next, the team modeled the distribution system that provides hot gases to the heat exchanger to make sure it actually met the uniform assumptions of the first analysis. The initial design had assumed that a single injection point would provide uniform circulation within the heat exchanger. The analysis showed, however, that with only a single injection point, the hot gases impinged on the heat exchanger wall and created a recirculation zone. This could have caused a cold spot that would have prevented the heat exchanger from uniformly heating the reactor vessel. After changing the model several times and re-running the analysis, the project engineers finally settled on the use of four injection points, which provided the uniform flow required to achieve uniform heat transfer.

Marble Hill Demo

With the experimental results confirmed, the new indirectly heated annealing process was first tested at a demonstration at Public Service of Indiana`s never-completed Marble Hill nuclear plant near Paynesville, Indiana. The annealing demonstration at Marble Hill was carried out with the combined resources of the Department of Energy`s Sandia Laboratories and an industry consortium including the American Society of Mechanical Engineers, EPRI, Consumers Energy, Japan`s Central Research Institute of the Electric Power Industry, Westinghouse and Cooperheat-MQS.

The reactor vessel was heated for seven days and 10 hours, as the heat exchanger reached a temperature of about 1,100 F. The reactor pressure vessel was brought to its peak temperature at a rate of about 20 F per hour and cooled down at the same rate. Each of the five gas burners produced two to three million Btus of heat per hour. During the annealing process, the pressure vessel and surrounding components were monitored by over 500 thermocouples, strain gauges and displacement gauges.

An important concern during the testing was showing that the vessel maintains a fairly even temperature distribution during annealing in order to avoid the stresses associated with thermal variations (Figure 2). Another concern was avoiding damage from vessel expansion as it is heated, because the piping and other connections surrounding the vessel do not experience the same expansion.

The test was a complete success. Reactor wall temperatures in critical areas were closely monitored during the annealing process. Target temperatures for the embrittled regions of the vessel requiring annealing as well as lower targets for the other areas of the vessel were met and maintained during the process. The heating system demonstrated its ability to provide the required level of temperature uniformity in the annealing region and maintain desired gradients in the transition zones between embrittled and non-embrittled areas. Analytical results verified that the metal throughout the vessel walls, welds and attached piping expanded and contracted as predicted, without damage.

The Marble Hill test proved that aging reactor vessels can be rejuvenated and their operating life substantially extended, permitting old reactors to continue operations for many years.

Management of nuclear reactors throughout the world have expressed interest in this technology. In April 1998, for example, EPRI held a well-attended conference that focused on the topic of annealing, at which the Marble Hill project was discussed. p


Eric Bish is a Senior Engineer at Fluent, Inc., Lebanon, N.H. He provides engineering support to clients in the power generation industry. Bish received his Ph.D. in Aerospace Engineering from the University of Michigan in 1996.

Bob Nugent is Manager of Engineering for Cooperheat-MQS, Inc. in Piscataway, N.J., where he has worked for 10 years. His department provides engineering support for contract heat treating services and equipment sales, including the development of special heating processes and equipment such as that described in this article. Nugent received his BSME from Rutgers University College of Engineering in 1988.