Published Resources Details
Journal Article
- Title
- The analysis of networks of pipes and pumps
- In
- Journal of the Institution of Engineers, Australia
- Imprint
- vol. 37, no. 4-5, Apr-May 1965, pp. 111-117
- Description
Paper No.1975, submitted by the author on 13th November, 1964.
The author, C. J. Bellamy, BE PhD AMIEAust, is Officer-in-charge, Computer Centre, Monash University, Melbourne.- Abstract
The problem of determining the flows in a network of pipes and pumps occurs in many situations; in water reticulation systems, in chemical plants and in oil refineries to name but a few. In many instances where the system is essentially a tree structure the problem is relatively trivial, however, where there exists a high degree of interconnection the problem gives rise to a set of nonlinear equations which may be prohibitively time-consuming to solve by hand using the accepted Hardy-Cross methods.
To facilitate the solution analogue network analysers have been developed and more recently digital computers have been used. This report describes a general method of solution that has been programmed for a Ferranti Sirius digital computer and the way in which the program is used.
In practical situations the most difficult problem is to establish the characteristics of pipes and pumps so that any analysis will provide significant results. Corrosion, and deposition in pipes, may substantially alter their flow characteristics from those obtained by assuming the internal diameter and roughness characteristics for new pipes. Ideally, before a network analysis is undertaken, the pipe and pump characteristics should be measured experimentally over the range of flows which occur in the system. Failing this an attempt may be made to estimate the characteristics by repeatedly analysing the system and adjusting the pipe characteristics until the system conforms with the available information on total heads and flows. If accurate data on the pipes and pumps in a system are not available then it is imperative that measurements be made to establish whether the mathematical model of the system is adequate.