Font Size: a A A

Research On Analog Circuit Simulation Technology Based On Evolution Method

Posted on:2008-09-06Degree:MasterType:Thesis
Country:ChinaCandidate:C H CenFull Text:PDF
GTID:2178360272968899Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
Circuit simulation is key approach to design, test, diagnosis and maintenance of electronic product, the optimization of solving circuit equations is the emphases and difficulty of analog circuit simulation. Aimed on the problems of convergence instability, repeated derivative operations and one-solution-one-time in DC analyses of analogue circuit simulation, the principles and development of analogue circuit simulation are discussed, and the circuit simulation technique based on evolution method is proposed, which is applied to the analogue circuit simulation system.First, the development and actuality of analogue circuit simulation is introduced. The problem in analogue fault-free simulation and fault simulation is discussed, also the necessity of improve simulation efficiency by optimizing the solving of circuit equations is analyzed.Second, the principle and algorithm of analogue circuit simulation are discussed by SPICE for instance. Two key steps including equations setup and algorithm solving in circuit simulation are introduced, also the convergence restriction issues in DC analyses and Transient analyses are discussed.Third, the optimization of solving circuit equations based on evolution method is studied. With the direct evolution method, the configuration of fitness functions, design of transferring rules and update mechanism in evolution method are discussed in details. Based on the direct method, a new hybrid method-ACA-NR, which integrates ACA and NR method, is proposed.Finally, the implementation of SPICE program based on ACA-NR and analogue circuit simulation software based on the program is introduced. The performances of NR, ACA, and ACA-NR are compared by the simulation results of benchmark circuits; also the relationship between MIF (Mean Individuals Fitness) and number of generation or evolution is described by the figure. Experiments results show that ACA has the ability of stable convergence and finding multiple solutions, ACA-NR is the best choice to the optimization of solving circuit equations in DC analyses.The research in this paper is valuable for the improvement of analog circuit simulation.
Keywords/Search Tags:Ant colony algorithm, Newton-Raphson method, DC analyses, Analog circuit simulation
PDF Full Text Request
Related items