Research On Refining Furnace Power Supply Curve Methods Based On CBR

Refining technology can improve the production capacity of steel-making devices to make the quality of steel and alloys better.Also it can reduce the energy and consumption of ferroalloys and refractories.And the utilization of time can be improved.Refining furnace consumes a lot of energy,so the research on the refining furnace power-supply characteristic and the establishment of a reasonable power-supply strategy has a theoretical and practical significance in the lower cost and high efficiency production.The paper describes the development process and the refining furnace research status after consulting a great deal of document.Currently,the establishment and optimization of the refining furnace power supply curve are mostly based on the mechanism modeling.This method need make an in-depth analysis of refining furnace electrical characteristics,so it is complex and its accuracy is not high.There are many problems when it is applied to practice.However,this problem easy to describe can find the solution of the problem simply and quickly using case-based reasoning.And case-based reasoning algorithm has self-learning ability.Case-based reasoning algorithm can work well even if industrial devices become aging and updated.Firstly,in this paper,refining furnace power system is divided into three stages on the base of the refining furnace steelmaking equipment,technology and electrical characteristics.And it develops a refining furnace supply curve evaluation index for case-based reasoning solution and case-based maintenance.Secondly,we propose the case representation of the refining furnace power supply curve and establish the database of cases.We develop an optimized power supply curve of the furnace using the case-based reuse based on support vector machine after retrieving the database by the nearest neighbor case retrieval method based on improved grey relation grade.Finally,update,optimize and maintain this case and casebase by evaluating the curve after smelting.So it can enhance the system's ability to solve problems,improve the retrieval rate and optimize case database.Meanwhile,the computer system has been implemented based on the algorithms.