(A) novel method for fast partial defect detection for spent fuel safeguards

As the number of nuclear power plants increases worldwide, the importance of nuclear safeguards is increasing. International Atomic Energy Agency (IAEA) recommends all spent fuel assemblies has to be verified completely before spent fuel storage. The complete verification includes correctness verification and completeness verification. The correctness verifies whether the declared information of target material is correct or not. The completeness verifies whether the declared amount of nuclear material exists or not. The completeness verification includes gross and partial defect detection. However, methods for partial defect detection have some limitations to verify all spent fuel assemblies: long detection time and frequent maintenance. This dissertation designed a scintillator – photodiode based partial defect detector (SPDD) to overcome this limitations. Passive gamma photons from spent fuel is initially converted into visible photons via a scintillator. The scintillated photons are then converted into electric current using a photodiode. Since SPDD does not require signal amplifier and additional structure, it has fast screening capability and less maintenance compared to conventional methods. SPDD consists of a number of detector legs. Each detector leg consists of a CdWO4 scintillator, amorphous silicon photodiode, and structural material. Detector legs are inserted inside guide tubes of spent fuel assemblies. SPDD then measures the generated current at each guide tube location. Since dummy material do not emit gamma photons, generated current near dummy fuel rods decreases compared to the other locations. SPDD is operated based on the following two assumptions: 1) Given operator provided information and 2) No diversion in fresh spent fuel assemblies. SPDD performs initial verification for all spent fuel assemblies at the beginning of the storage. Initial verification compares the estimated and measured SPDD current. Estimated current can be calculated using operator provided information and lattice physics code. Since no dummy fuel rods exist in fresh fuel, initial verification generates current distribution database for all spent fuel assemblies. Once initial is finished, regular verification is performed. Regular verification compares the measured SPDD current distribution and current distribution in database. If the difference is larger than an SPDD detection criterion, target assembly becomes suspicious assemblies to have partial defects. The suspicious assemblies can be re-verified using conventional methods with high accuracy and long detection time. This dissertation developed a computational model to describe real measurement results. The model was validated by lab-scale experiments using gamma sources and mockup SPDD. The detection criterion was quantified based on uncertainty analysis. The performance of an SPDD was demonstrated using test case spent fuel assemblies with partial defects. The dissertation is organized as follows. Chapter 2 analyzes the characteristics of conventional detectors for spent fuel verification based on the literature and demonstrates the limitations of conventional partial defect detection methods. Chapter 3 describes the design of the proposed detector and operating principle. Chapter 4 develops a computational model to analyze the performance of the proposed detector since the direct use of spent fuel is almost impossible. Chapter 5 evaluates the performance of the proposed detector based on the results of Chapters 3 and 4. Chapter 6 consists of the overall summary, discussion, and conclusion of the dissertation