Research On Wind Power Project Unit Cost Estimation Index System And Model Construction

Engineering cost information plays an adjustive and indicative role in the whole process of engineering cost management. All cooperated-building parties need to make relevant decisions by engineering cost information and to make related decisions and methods about cost in the process of project construction. Cost estimation index system is a way of engineering cost information application, but now the cost estimation index system in our country cannot make good use of the engineering cost information. The main reason is that the cost index extraction is not completed or flawed. Therefore, this paper puts forward to rebuild a new cost estimation index system in wind power projects through the study of wind power project cost structure and extracts completed and handy cost estimation index. Establishing the model on basis of the above objective, the subject possesses superior practical application value.This paper expounds the concept of data mining and data mining methods and analyzes the characteristics of all kinds of data mining methods. Through researching the model of data mining process named CRISP-DM (Cross-Industry Standard Process for Data Mining), the paper presents the process of engineering cost information data mining in order to guide the later research work. The work uses the WBS project decomposition structure method, combined with the cost structure of wind power engineering and related cost documents, to divide different kinds of project and establish the unit cost estimation index system. This system embraces two parts:sub project cost estimation index system, equipment and materials cost estimation index system. In the established index system of cost estimation index system of the wind power project, the paper collects the estimates information of practical engineering project and other applicable. In order to estimate unit cost of the wind power project more accurately, the original index data have been collected within data preprocessing, including the processing of data cleaning, data transformation, and data reduction. Based on the index data after data preprocessing, the cost estimation model of wind power engineering unit is properly established. The model is included the neural network estimation model and the support vector machine (SVM) regression estimation model. On the fundamental of the calculation program, the two estimation models are evaluated by means of an example to assess the results of these two methods. The analysis results show that the support vector machine regression estimation model has the better generalization ability, faster convergence rate and higher estimation accuracy.