Search

Oil and gas pipelines transport millions of dollars of products every day making the optimization of their long-term maintenance strategies an essential target for practitioners. Thus, this paper aims to optimize the maintenance plan of such pipelines by maximizing their lifetime average condition with the minimum possible cost. This is achieved in three steps: (1) developing a life-cycle cost model for such pipelines to determine their net present value (NPV) based on the different rehabilitation actions applied throughout their lifetime; (2) establishing a strategy to determine the condition index of such pipelines before and after any rehabilitation action applied throughout their lifetime; and (3) formulating the optimization model and applying the elitist non-dominated sorting genetic algorithm (NSGA-II) to determine the optimum maintenance plans for such pipelines. The optimization model is applied to a real-life case study to demonstrate its applicability. [ABSTRACT FROM AUTHOR]

Construction companies need a knowledge management system to collate, share and ultimately apply this knowledge in various projects. One of the most important elements that determine the time estimates of any construction project is productivity. Such projects have a predilection towards uncertainty and therefore require new generation of prediction models that utilizes available historical data. The research presented in this paper develops, using fuzzy approach, a knowledge base to analyze, extract and infer any underlying patterns of the data sets to predict the duration and productivity of a construction process. A six-step protocol has been followed to create this model: ( i) determine which factors affect productivity; ( ii) select those factors that are critical; ( iii) build the fuzzy sets; ( iv) generate the fuzzy rules and models; ( v) develop the fuzzy knowledge base; and ( vi) validate the efficacy and function of these models in predicting the productivity construction process. The fuzzy knowledge base was validated and verified using a case study and the results were satisfactory with 92.00% mean validity. In conclusion, the developed models and system demonstrated the ability of a knowledge base management to predict the patterns and productivity of different construction operations. [ABSTRACT FROM AUTHOR]