Intelligent early-warning system of dam safety

For the effective control of dam life process, it is important to implement the real-time diagnosis and reasonable evaluation for dam safety based on prototype observations. The development of intelligent early-warning systems of dam safety (IEWSDS) is a significant approach to realize above aim. This paper regarded dam as a vital and intelligent system, thereby constructed a bionics model of dam safety, which consisted of sensing system (nerve), central processing unit (cerebrum) and decision-making implement (organism). With above model and systems engineering, this paper designed an IEWSDS, which was composed of integration control, intelligent inference machine, project database, model base, graphics base and input/output modules. In the system, metadata base based on object oriented data model was used to implement the integration of homogeneous or nonhomogeneous databases storing dam safety data from different information sources. The multilevel link style was adopted to manage model base. Intelligent inference machine is the central processing unit of IEWSDS implementing data analysis, abnormality diagnosis and evaluation for dam safety. Because of strong non-linear and dynamic characteristics, the system adopted a combination model based on wavelet networks to approximate and forecast the run characteristics of dam. The abnormality diagnosis is a process of discovering knowledge on the pattern between of disease and geneses. The methods of attributions reduction in rough sets theory were presented to diagnose adaptively the abnormal cases and find out the geneses. The safety status of dam changes dynamically, which emerges in the quantitative and qualitative change manners. Therefore, quantitative and qualitative changes need be considered comprehensively in the process of dam safety evaluation. This paper proposed an extension evaluation method. The quantitative change and qualitative change of dam safety property were integrated into a matter-element. Based on the matter-element extensibility, the contradictory problem for dam safety elevation can be solved. In practice, the proposed system has been used to monitor dam safety successfully. The applications show that the bionics model is feasible; the proposed key technologies are effective. The systems can supply technical support for improving the level of dam safety management, extending normal run time of dam and voiding dam failure.