Errors arising from the digital implementation of integration and lead/lag networks

摘要

In the simulation of aircraft and other physical systems, the solution of differential equations is required and lead and/or lag networks have to be simulated. Many techniques have been devised for numerical integration when digital computers are used to solve ordinary differential equations. Obviously, since the digital computer uses discrete samples rather than continuous variables, such integration schemes are an approximation to the solution of a differential equation. The question naturally arises to how good the approximation offered by the use of a particular integration algorithm or lead network implementation is. This paper explores the sources and magnitudes of the errors which accrue from the use of different integration algorithms and compares such errors quantitatively.