Font Size: a A A

Sequential quadratic programming algorithm and an optimum mechanism design

Posted on:1992-01-01Degree:Ph.DType:Dissertation
University:North Dakota State UniversityCandidate:Dandu, Raju SivaramaFull Text:PDF
GTID:1470390014997999Subject:Engineering
Abstract/Summary:
Pshenichny's linearization method, which belongs to sequential quadratic programming (QP) methods, has been modified to improve its robustness and numerical efficiency. The application of optimization in engineering design has been demonstrated by improving the design of the spring-loaded four-bar mechanism.; The Pshenichny's has been modified by implementing the solution of a QP subproblem by Lemke's method, a feasible starting point method, an active set strategy, a step size calculation, the selection of penalty parameter, and a Hessian update. Implementations of Pshenichny's baseline algorithm and modified algorithm were tested against 54 standard test problems. The results indicate the modified algorithm was more efficient than the baseline algorithm.; The sequential quadratic programming technique was used to synthesize optimal performance of a spring-loaded four-bar mechanism that, when actuated, lifts a farm tillage implement out of the soil. Maximum vertical travel of 12 inches of the sweep and maintenance of a near constant actuation force of 650 lbs were significantly improved over a current design. Computer-aided design was used to confirm that interference design constraints were not violated during optimization. Infield performance of the mechanism when encountering an obstacle was verified using a dynamic model to realistically simulate the event.
Keywords/Search Tags:Sequential quadratic programming, Mechanism, Algorithm, Modified
Related items