Director, Consortium for Nonproliferation Enabling Capabilities
North Carolina State University
Neutron Interaction in a Nuclear Reactor (0:30)
(Yousry Azmy recalls his path to academia in the nuclear space and explains how neutrons interact in a nuclear reactor)
Q: Where did it all begin for you? How did you know about nuclear?
A: Back in Egypt in the late 1960’s and early 1970’s, Yousry Azmy was finishing high school and decided that engineering was where his heart was. In Egypt, the education system is a little bit different. In a five year system, the first year is preparatory where everyone takes the same classes. During this year, Yousry stayed in Cairo to study mechanical engineering, but later decided he didn’t want to do mechanical and moved mid-semester to transfer to the nuclear engineering program in Alexandria. Yousry was drawn to nuclear because of the rigor and good mix between technology and basic science. The mathematically and computationally-oriented studies appealed to Yousry from the early stages. Nuclear was a bit abstract, which also appealed to his type of thinking and attitude.
Yousry Azmy came to the U.S. in 1980 to pursue his graduate program at the University of Illinois-Urbana Champaign, receiving his Master’s in 1982 and his PhD in 1985. The focus of his career has been transport theory. Yousry’s three-piece dissertation was on basic math of numerical methods, CFD (computational fluid dynamics), and transport, primarily diffusion acceleration, which is a way of solving the transport equation faster under certain conditions. Radiation has three features that characterize it: space, the energy of the particles, and the direction of their flight. There are two major classes of methods for angles: spherical harmonics and discrete ordinates, which is Yousry’s focus. Sit at any point in the reactor and there will be neutrons flying in every possible direction. Beam tubes create a path of least resistance that neutrons are not going to interact with much in a containment that is either non-absorbing or non-permitting. It does not prevent the waste from the neutrons that don’t make it in. They are just reducing the number that are not going to make it by reflecting them into the path.
Neutrons are born in fuel pins at very high energy because fission produces very high energy neutrons, a majority in the meV range. In light water reactor technology, water is used because hydrogen happens to slow them down quite a bit. The probability of the neutrons having an interaction with fuel to generate even more neutrons and sustain the chain reactor increases with decreasing energy in the thermal range. Neutrons are born at fission energy and, if they aren’t slowed down, there is a very little likelihood they will produce more fission and sustain the chain reaction. The slower the neutrons get, the higher the probability that they would interact. There are certain minute differences in the way a neutron interacts with the nucleus. Fission occurs in the first place because, once the neutron is absorbed, the compound nucleus is excited and needs to stabilize by ejecting energy.
Transport Theory Calculations on Unstructured Grids
(Yousry Azmy discusses his research in solving the transport equation on unstructured grids and explains the levels of a reactor transport problem)
Q: Does your lab focus on things other than transport?
A: Yousry Azmy’s lab does transport and applications of transport if it happens to be a heavy lift problem numerically or computationally. One of the big projects he has in his group is a new code for solving the transport equation on unstructured grids. People in computational fluid dynamics (CFD) have done this for ages, for example, in the design of aircraft wings. In neutronics, there are a lot of curved surfaces, such as a fuel pin and control rod. The method of long characteristics does not require discretizing into a spatial mesh. The pins and cladding are kept as curved surfaces and the problem is solved using the method of long characteristics and homogenization to represent a fuel assembly as a block without the inside of it. Unstructured grids are not needed in reactors at this stage, but in 5-10 years from now when quantum computing becomes a reality, people may want to do that. In applications like medicine, unstructured grids may be needed to capture the detail of a bone, tissue, or the surface of a tumor.
There are three levels of doing a reactor transport problem. The level of the actual reactor is done with diffusion theory, which is a simplified version of transport theory. The previous level is the assembly level, which is done with transport, on a single assembly at a time. The input to the intermediate level comes from another transport solver prior to that, which deals frequently with individual fuel pins or clusters of pins to reduce the number of energy groups. Yousry’s codes are mostly for research, but there is a code system at Oak Ridge National Lab that is available for engineers. Universities are not designed to do this type of production. They look at the next big thing, not the current big thing.
Currently, the major thrust in research is in nuclear security and non-proliferation. The transport solver is an engine that could be attached to anything you want to drive. If the obstacle in the next design is one portion of the transport calculation, Yousry will then go out to surmount this obstacle by thinking of something novel and put his students on it for their PhD’s. The nuclear space is very difficult to predict because of the speed at which computational speed is evolving. In the next few years, the GPUu technology is going to see continuous growth. The push towards exascale computing is going to make a big splash when it happens. The question for applications people is where they can carve a niche and exist in it as nuclear computational scientists. The unpredictable wild card is quantum computing which will change everything.