Capacity value estimation of random solar power through associative searching in a k‐d tree

Abstract

Summary The estimation of capacity values of unconventional sources like solar and wind is quite difficult due to their intermittent nature. Mainly reliability‐based methods are generally followed to calculate the capacity values. Intermittent solar power cannot be clubbed directly in the process of reliability calculation. Solar power is generally taken as negative load in adequacy assessment and subtracted from the load on hourly basis. Random solar radiation creates redundant solar power generation. Bypassing multiple estimations of capacity values for the redundant solar power, an approach is made to construct a database considering the incremental added capacities with their resulting loss of load expectation. The database is portrayed in a multidimensional k‐d tree to locate queried expected loss of load of solar power effectively using nearest neighbor search algorithm. The algorithm is providing the best match record as well as the associated information like the capacity values stored along with that record. A case study of 75 MW of conventional with 15 MW solar powers is considered for the validation of the proposed methodology. Capacity values corresponding to each reliability‐based methods resulted from the database seem to be satisfactory in evaluating the capacity values of solar power. The applicability of the proposed method is extended through implementing on the IEEE RTS system.