|
Abstract
|
In recent years，Life Cycle Cost（LCC）and Economic Life（EL）have been widely used to determine the decommissioned life of equipment for energy companies. LCC records all cost in-curred for equipment from the start of operation to decommissioning; EL is the year when the average annual cost of LCC is lowest, decommissioning equipment at EL can maximize economic benefits. But researches on LCC and EL in the last two decades were still simple, they mainly selected a part of equipment as a sample to record all cost incurred and recorded it as LCC ,by directly calculated the average annual cost to obtain EL. In this paper, we will use 75 220KV transformers put into service in 1986 for the study (this type of equipment was first strictly recorded as LCC step in 1986 at a representative energy company).We first utilize Isolated Forest (IF) and Weibull distribution to screen the sample for outliers and correct the probability distribution, then get the average annual cost of equipment during the first 37 years by Monte Carlo stochastic simulation (MC) and use various algorithms to predict it in 38~41 years , compare MAE, MAPE, RMES to select the most suitable Artificial Hummingbird Algorithm (AHA) and Bi-directional long and short-term memory algorithm (Bi-LSTM) to get EL, finally EL is compared with the design life (manufacturer's recommended service life) to determine the optimal retirement life of equipment. In this paper the improved model is named MC-AHABi-LCC, it in the result is closer to the actual equipment decom-missioning life than that obtained by the traditional method, and the improved model is more accurate and scientific.
|