There are several uncertainties associated with power injections inadistribution system network because of intermittent renewable power generation, forecasting errors, measurement errors, etc. One simpler way of incorporating uncertainty into power system analysis is by using Monte Carlo (MC) simulations, where a large number of samples are extracted from within the specified uncertainty range, and repeated analysis is carried out with each of the extracted samples. However, it requires very large computational time due to the repeated number of simulations which may not be suitable for online studies.

In the present work, improvedaffine arithmetic (AA) based Backward Forward Sweep (BFS) based power flow analysis, improved AA based optimal DG allocation and improved AA based optimal network reconfiguration are proposed for a distribution network with uncertainty in power injections. Existing AA operations generate additional noise terms as a result of nonaffine operations like multiplication and division. The proposedimproved AA based multiplication operation does not produce any extra noise terms as compared to existing AA multiplication and improved AA based division operation produces one less noise term compared to existing AA based division, thereby improving accuracy in solution bounds.The proposed improved AA based methods are implemented on IEEE 33, 85 and 119 bus balanced radial distribution systems and an unbalanced IEEE 123 bus radial test feeder. Thedetailed simulation results will be presented.