The Application of Dynamic Programing to Fuel Management Optimization

Pattern optimization of PWR reactor using hybrid parallel Artificial Bee Colony

“…Several algorithms have been developed and successfully applied to optimize reactor core loading problem such as Dynamic Programming (Wall and Fenech, 1965), direct search (Stout, 1973), Variational Techniques (Terney and Williamson, 1982), Backward Diffusion Calculation (Chao et al, 1986), Reverse Depletion (Downar and Kim, 1986;Kim et al, 1987), Linear Programming (Stillman et al, 1989), Simulated Annealing (Stevens, 1995), Ant Colony algorithm (Schirru et al, 2006), Safarzadeh et al (2011) applied ABC algorithm to power flattening of PWR reactor, continuous Genetic Algorithm (GA) introduced for flatting power distribution (Zolfaghari et al, 2009;Norouzi et al, 2011), discrete PSO (Babazadeh et al, 2009), continuous PSO (Khoshahval et al, 2010), Mohseni et al used GA in multi-objective optimization of lowering power peaking factor, maximization of the effective multiplication factor (Mohseni et al, 2008), Cellular Automata for maximizing initial excess reactivity and minimizing power peaking factor , Perturbation Theory (Stacey, 1974;Hosseini and Vosoughi, 2012), ArtificialIntelligence techniques like Artificial Neural Networks (ANNs) (Sadighi et al, 2002), and combination of fuzzy logic and ANN (Kim et al, 1993) are the ones most commonly used in core fuel management. A further study based on hybrid algorithms was performed (Stevens, 1995;Erdog and Geçkinli, 2003;).…”

Application of differential dynamic programming to nuclear fuel management optimization

“…Another approach to fuel management optimization which has promised some degree of success and usefulness is dynamic programming. Wall and Fenech [16] considered this approach in optimizing the loading strategy of a three zone pressurized water reactor over the lifetime of the plant. The method of dynamic programming hinges on Bellman's Principle of Optimality [17,18] which states that at any stage of operation, only the current state of the system and future control actions deter mine optimality, not the previous history of the system.…”

“…The constraints used in the search were approximated and must be confirmed by detailed calculations at this point. Other constraints imposed upon the solution but not considered during the search include maximum burnup and gross-power peaking.Two common assumptions used in previous studies[6,12,13,16] and used for certain problems considered in this studyare 1. Fuel (lattice) burnup calculations can be separated from the spatial burnup problem.…”