Research On Offshore Platform Project Cost Estimation Technology

The competition between offshore platform construct companies becomes fierce due to the increasing demand for offshore platform.Cost estimation can exert a significant influence on the tender price and the more accurate the tender price is,and the easier estimation will be.Full-scale project cost breakdown structure and effective cost prediction methods are the foundation of improving the accuracy of project cost estimation.However,offshore platform project cost breakdown method and cost prediction method have not yet come into being currently,and in order to solve these issues,research on offshore platform project cost estimation technology and method were finished in this paper.Firstly,the research of domestic and international situations about work breakdown structure,project cost estimation method and neural network optimization algorithm were analyzed.Secondly,offshore platform project cost breakdown strategy was studied based on the analysis of offshore platform project cost element characteristics,and then,cost tool bag conception and offshore platform project cost element technical route was put forward;the cost tool bag coding strategy was designed and cost tool bag collection algorithm was designed to solve the cost tool bag collection problem.Considering the influence of time-cost value,capital cost,taxes and unpredictable cost,offshore platform project cost estimation correction model was established;cost tool bag prediction problem was decomposed into inner and outer problem.Mathematical model was established to solve the outer problem.Neural network which activation function and training algorithm were improved was designed;comparative analysis of designed neural network and SPSS software about prediction accuracy and convenience based on example was given.Finally,prototype system structure of offshore platform project cost estimation was designed and prototype system was developed based on the Java language,the feasibility and effectiveness of the prototype system was verified.