The Improved BP Neural Network In The Application Of Hydropower Engineering Investment Prediction

Hydropower project investment is a huge system and a whole comprehensive control process. Its investment should be effectively controlled to obtain the biggest benefit using the limited approval of investment quota. Hydropower project is characterized by long cycle, big investment and cooperative work of many departments, and not only influenced greatly by natural resources, topography, geology, hydrology meteorological conditions, but also limited by local economic development level, transportation and other resources restrictions. The uniqueness and complexity of the hydropower project add the difficulty for its cost estimate and investment control. Hydropower projects are the national important economic and social development infrastructure, and the status and function of water conservancy industry in the national economy should be taken into account for the project investment. The project investment should strengthen government macro-control to realize equilibrium of the rapid development social relation, the paid service and production operational industry, to guarantee the normal movement of water conservancy, to realize capital investment and water conservancy assets value. Since the reform and opening up, with the development of economy and construction industry's rapid rise, there are some runaway investments and investment failure problems, such as final budget and the budget exceeds the budget estimates, super estimate phenomenon. As a result, it was difficult to obtain the investment efficiency. The main reason for this phenomenon is the lack of accuracy and comprehensive investment decision and a reliable investment analysis that has no accurate cost estimate standards. Therefore, it is necessary to improve the investment forecast analysis, make reasonable determination and effective control of the project cost; human, material and financial resources should be utilized more reasonably; fixed asset investment benefits should be controlled effectively.Investment estimation is the preliminary estimate of amount of investment on decisions-making process. Estimation of work was very important to the investment plan that directively affected the project for investment decisions. Especially,large hydroelectric engineering as the important domestric economic and social development infrastructure in government investment,is an accurate investment forecasting which appears very important on the phase of bidding to avoid the blindness of investment and reduce unnecessary economic loss.The analysis of benefit of investment on project cost management is also an important part of investment basic work that make clear the investment structure and the results.The result of scientific method for determining reasonable construction project investment provides comprehensive, systematic, reliable basis for decision-making. Investment analysis full or not will give the impact and bias for project-decision. Therefore, in order to strengthen the investment forecast analysis, pay attention to the early stage investment analysis research, avoid the wrong decision.With the market competition intensifying and global economic integration, the demand of project cost estimation is more scientific and reliable.Although,Our water resources and hydropower engineering started early in engineering fields, consultancy institutions needs strengthened, social foundations is not stable,the system of investment prediction is still not perfect. So, it is very necessary to improve and perfect the current investment method that researchs and develops project cost estimate. With the artificial neural network theory itself and related theory, related technology unceasing developing, the application of neural network has been widely into the neural physiological science, computer science, cognitive science, psychology, and mathematical science, information science, biology, electronics, microelectronics, optical etc, also be well applied to engineering cost estimate fields. Make full use of the fuzzy mathematics method and neural network method to build a cost estimate model for water electrician engineering investment estimation. Investment analysis includs technical, economic, environmental, social, and other important complicated factors,its analysis and decision process is a multi-stage, multi-state of dynamic points decision process. The BP neural network features and performance is very suitable for solving investment appraisal and analysis and decision this kind of multi-index, multi-factor analysis, judgement to resolve problems. So, the BP neural network can analysis aspects of construction projects and decision analysis.In this paper, access and in-depth study of domestic and foreign construction engineering cost estimation method, based on the characteristics of large hydroelectric engineering cost, put forward the fuzzy mathematics and the BP neural network combined forecasting model of investment.First, use the fuzzy mathematics method to screening historical data sample classification, in order to improve the investment estimate model accuracy;Through the comprehensive analysis of the main features of the hydropower project, connect these features and its engineering cost, construct based on BP neural network hydropower project investment estimation model.On the basis of the model, use MATLAB language to train, simulation and test the model, and use engineering example to validate the model, the results show that the model has better generalization ability, able to accurately estimate of the project cost.Then, puts forward the BP neural network in the application of hydropower project investment analysis,make investment forecast more authoritative.