Multi-Agent Technique Based Substation's Operating-Sequences Intelligent System

Automatic generating system of operating sequences is an important guarantee for the transformer substation of the power system．Nowadays with the development of integrated automation of substation, more AI theory and computer techniques are introduced into this system．A general study on its composed parts——the drawing system and generating system is carried out and the following problem is found still in existence: The separation between the graph and the DB in the drawing system made the communication with the generating system not smoothly enough; the generating system simply based ES is mainly coded under the DOS operating system and for a given substation's topology structure, its commonality and maintainability is greatly limited．This paper presents the distributed structured expression of the substation and establish a Multi-Agent system composed of equipment Agent,knowledge Agent,explanation Agent,corresponding and controlling Agent．The equipment Agents use the NN modeling technique to realize the connotative generation of the operating sequences; the corresponding and controlling Agent controls different Agent to work under parallel or corresponding state; the explanation Agent transfers the content of knowledge Agent to interpret the sequences and forms a formatted operating ticket．In the drawing system this paper compiles a VB program with a friendly interface and high visualization．Based on ActiveX technique of VB6.0 this paper builds the object-oriented electric equipment controls．Being packaged they have ones own attributes and events, can response to the users' different actions．The drawing system also has a topology automatic forming function and is connected with Access database by Microsoft jet of VB6.0．In the generating system this paper summarizes the representative tickets of some substations and confirms four equipment Agents．In each Agent this paper forms a RBFNN model and corresponding sample for training．The trained RBFNN is connected with generating system, so according to the NN's output it can generate the sequences automatically．The whole MAS accomplishes the question's solution．In the language forming aspect, the explanation Agent accomplishes this task through the interaction with the distributed knowledge Agent which has packaged the equipment knowledge,operation knowledge,protection knowledge and automatic device knowledge together．This function makes the system has a high response speed．Implantation of the Agent technique and RBFNN theory into the VB based automatic generating system of operating sequences makes the system has high commonality,rapidity,maintainability and highly intelligence．...