ALGOR Helps Badger Meter Design a Cooling Valve for Nuclear Reactor
For over a century, Badger Meter, Inc., in Tulsa, Oklahoma has been a leading manufacturer and marketer of products using
flow measurement and control technologies.
Its products are used to measure and control the flow of liquids in a variety of applications.
With a reputation for producing quality, precision-control, metered valves, Badger was selected by Atomic Energy of Canada, Ltd.,
a subcontractor of Societatea Nationala Nucleoelectrica S.A. (SNN) of Romania, to model, test and produce a set of precision valves
for Cernavoda Unit 2, the second nuclear power plant in Cernavoda, Romania.
For the nuclear power plant in Cernavoda, Romania (top), Badger Meter, Inc. of Tulsa, Oklahoma, USA,
modeled, tested and produced a set of process control valves (bottom left).
One variation of the valve assembly, consisting of valve and switching units, was modeled in Solid Edge (bottom right).
|
The main concern for the valve’s construction and operation was its survivability and continued functioning after enduring an earthquake.
In nuclear power plants, such valves control the cooling of the nuclear reactors where
continued flow of water around the nuclear core is essential for safety.
After the earthquake that precipitated the eruption of Mt. Saint Helen’s in 1980,
testing criteria for valves have routinely included their capability to ensure the safe functioning of the reactor after seismic events,
at least in terms of cooling capacity.
Using Solid Edge to model the complex valve for controlling water temperature and flow and ALGOR to analyze its response to earthquake forces,
Badger Meter, Inc. was able to provide a valve with survivability estimates well within the acceptable range,
in accordance with the criteria set by the Canadian contractor and the Romanian government.
Modeling the Process Control Valve
The original process control valve assembly, consisting of valve and switching units, was modeled in Solid Edge by
Badger Mechanical Designer, Wayne Hall.
The assembly was designed to be part of a temperature control system.
In all, the complete system consists of 16 separate units, including 11 major assemblies and 4 to 5 minor variations of one part.
The assembled 3-D model interfaces with the piping structures (lower horizontal pipes), also modeled in Solid Edge’s piping program.
Hall defined materials for the various parts using Solid Edge’s built-in library.
The parts consist of 300 series stainless steel, brass, higher-end nickel alloys and high-pressure alloys of brass and stainless steel,
depending on the stresses and temperatures to which the various parts would be exposed.
Simplifying and Defining the Model in ALGOR
After receiving the Solid Edge model from Hall,
Senior Valve Engineer Wayne Hays reduced the model from its original configuration to a simplified model of 5 components.
The removed components were simplified into a series of lumped nodal masses and later re-attached to the ALGOR FEA model with mass-less,
virtual beams at appropriate locations.
The mass, centroid and moment of inertia data for the nodal masses was formulated in Solid Edge and then applied in ALGOR.
Simplifying the model, Hays used only the necessary properties for analysis and thus reduced the solving time.
For finite element analysis, Badger Meter developed a simplified valve model of five components.
|
Hays defined the important material properties using 316 stainless steel alloys from the Solid Edge, ALGOR and
MatWeb material libraries.
He then performed a mesh convergence analysis to increase mesh density as needed while eliminating unnecessary detail.
In some parts, the mesh refinement was as great as 1/1000th inch, while in others it was as coarse as 1/8th inch.
In all, the mesh yielded 29,448 elements.
Constraints were then defined at the rigid connection of the bonnet to the valve body.
Analyzing the Model in ALGOR
The goal of the analysis was to simulate the valve assembly’s condition and functionality after an earthquake.
That is, failure was defined not in terms of status at the moment of seismic impact but rather after the event’s impact was initially felt.
As part of an effort to improve the accuracy of the experiment, Hays was not given access to the maximum allowable stress target for the entire assembly.
That is, Badger was in a blind data scenario with regards to the overall target stresses.
The impact event—an earthquake—was defined as inputs of 1 to 4 Gs applied over a period of 1/10 second.
The event was simulated by using a
natural frequency analysis,
incorporating the results into the model and then conducting a
response spectrum analysis,
which is available as one of ALGOR’s linear dynamics tools.
Results
The results of the analysis showed that the maximum stress of 1811 psi was found in the yoke,
the section connecting the valve’s mount to the piping.
This is less than the maximum allowable stress of 23100 psi for material of this kind (ASTM A-27 Cast Steel), according to the ASTM standards.
The results of the ALGOR analysis showed the stress distribution in the valve model.
The impact event—an earthquake—was defined as inputs of 1 to 4 Gs applied over a period of 1/10 second.
The event was simulated by using a natural frequency analysis, incorporating the results into the model
and then conducting a response spectrum analysis, which is available as one of ALGOR’s linear dynamics tools.
The maximum stress of 1811 psi was found in the yoke (as shown in the inset).
This is less than the maximum allowable stress of 23100 psi for material of this kind (ASTM A-27 Cast Steel), according to the ASTM standards.
|
In its use of ALGOR analysis, Badger Meter, Inc. realized enormous savings in terms of the time, materials, testing funds and parts
that would have been required to prototype the valve assembly and perform physical tests.
Given a choice of either physical prototyping and testing or using simulation software to conduct virtual testing, Badger chose ALGOR.
Using ALGOR, Badger learned that its valve assembly met the criteria for use in the Cernavoda nuclear facility,
whose construction is now nearly 40% complete.
Badger’s valve assembly will be part of the new, functioning nuclear facility when it opens this year.
"FEA seismic events were new to me," said Hays.
"ALGOR is the easiest and most accessible tool available for handling such events without having to overcome an enormous learning curve.
We have several projects in the pipeline, for which we plan to use ALGOR."
Senior Valve Engineer Wayne Hays used ALGOR FEA software to analyze a process control valve for the Cernavoda nuclear power plant.
|
|