The Effect of Delayed Neutrons and Detector Dead-time in Feynman Distribution Analysis
Open Access
- Author:
- Brener, Mathieu W
- Graduate Program:
- Nuclear Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- June 24, 2016
- Committee Members:
- Gary Lee Catchen, Dissertation Advisor/Co-Advisor
- Keywords:
- delayed neutrons
Feynman analysis
detector dead time
neutron analysis
neutron coincidence counting
neutron chains
neutron multiplicity
monte carlo simulation
hansen-dowdy
feynman distribution - Abstract:
- I have explored the effect of delayed neutrons on neutron chains in a sub-critical nuclear system. Every fission has an approximately 0.7% chance of creating a delayed neutron that will go on to cause another fission. My hypothesis was that in systems with a very high sub-critical multiplication, there will be enough delayed neutrons created that they will in effect become a stronger starting neutron source than the original source. I used a point Monte-Carlo style simulation of sub-critical neutron chains to explore the physics of neutron chains. The simulation was designed to keep the same assumptions behind the Hage-Cifarelli derivation and Prasad-Snyderman formulation, to preserve the statistical variations in the list-mode data found in a physical experiment, and to be as simple as possible while maintaining accuracy. I present this method and benchmark it against both measured data and a theoretical model. I also use the model to simulate a Feynman Beta curve and 1/M approach-to-criticality experiments. I modified the simulation to take into account delayed neutrons and found that delayed neutrons have no measurable effect on a sub-critical systems for subcritical multiplication (M) less then 200. I wrote an algorithm to take into account detector dead-time with multiple detectors and found that increasing the number of detectors is the only effective method of compensating for dead-time effects. I investigated the minimum number of events needed for the analysis methods to produce accurate results and found that, for low values of M, there are statistical variations of the calculated multiplicity that will produce inaccurate results. For high values of M, the statistical variation is low, but an insufficient number of events will produce M-values that are inaccurately low.