Overview

(Dietmar Fischer)

In a certain sense, piping technology has always been the pioneer for the use of computer programs. This is true for the area of isometry creation via CAD programs as well as for the area of material management with material extracts and pipe class management. This early appearance of EDP was particularly pronounced in the field of hydraulics and stress analysis.

The need to computationally record the expansion of spatial piping systems and thus to achieve a certain operational reliability was recognized early on. An ASCE article on this problem appeared as early as 1910. However, while a spatial bar system, e.g. for bridge structures, had long been calculable, an application of these calculation methods to a corresponding pipeline bar system failed due to the special features of pipe bends. The movements that a loaded piping system makes in space were not understood for a long time. Only the pioneering work of Karman in 1911 and Wahl and Hovgaard in 1928 paved the way for a calculation of three-dimensional piping systems with pipe bends.

With the knowledge gained, it was now possible to calculate simple frame structures such as expansion loops, L-legs and spatial Z-propagations. The expansion problem of complex line structures was reduced to the consideration of these known elementary frame structures. This approach is still used today for a rough estimate of flexibility or for initial support design. Nevertheless, the growing hazard potential in the chemical and petrochemical industry, and especially the problems in power plant engineering, demanded a calculation of the entire pipeline system, including the consideration of singular supports, spring hangers and fixed point displacements. Out of this need, in the 1930s the Kellogg Company, in cooperation with various American universities, developed a general analytical method for calculating spatial piping systems. These methods were refined over the years and officially published in 1941.

It was also the Kellogg Company that considered the use of digital computers for calculating spatial piping systems and had the first programs on punched cards available as early as 1954. The economic efficiency of these programs quickly became apparent. While the time required with these punched card programs had already shrunk to one tenth of the time required for manual calculations, the use of magnetic data carriers from 1956 onward reduced the calculation time even further. The expansion of these programs was further accelerated. The demands made on the programmers of the Kellogg Company at the beginning of the 1960s seemed almost modern.

"Every conceivable pipeline configuration must be calculable, regardless of the number of branches, fixed points, supports and internal ring connections. Above all, the program must get by with a minimum of input while providing a maximum of automation."

Why were such demands made on a piping program already in the early days of computer science? The calculator was freed from the pure mathematical activity with the help of the program, but now he had to describe the piping system to the program in an understandable way. Furthermore, the calculator cannot decide whether it exceeds the possible complexity of a system that can just be calculated. The piping program must therefore be able to handle any possible configuration and it must communicate with the user in a suitable way.

It is obvious that a piping system with all its variety and variation cannot be described as a whole. The practicable solution can only be to reduce the whole problem to small, simple subproblems. The fineness of the reduction is determined by the fixed piping elements. Consequently, only the type, number, position and sequence of these finite elements must be described to the program. The calculation method of the finite elements (FEM) was mathematically explored, the problem solvable via EDP!

The various piping programs that were also developed in Germany in the following years always follow this principle. The user first defines a Cartesian coordinate system to be able to describe points in space with a corresponding location vector. Then the given piping system is divided into individual elements. These individual elements are distinguished from each other by defining connection nodes. Consequently, a line element is always located between two nodes. Conversely, this means that between two nodes all properties such as diameter, wall thickness, material, but also the line course are constant. Only at a node can a line characteristic change, and only at defined nodes can forces be introduced or removed and can supports be located. The position of all nodes in space is defined by a vector, which also defines the position of all elements. Also the occurring load of a pipeline in operation is not calculated as a whole, but it is at least divided into the two basic load cases dead weight and temperature.

The programs themselves now process these finite elements according to the known mathematical methods. As a calculation result, the displacements, the occurring forces and moments and the existing stresses are determined for each element and for each load case and documented accordingly.

Now, how is the semantic description of a line system done? The programmer defines a set of keywords with which the user can determine the individual elements by type, or with which the special properties of the elements and nodes can be defined. Furthermore, the programmer defines a grammar, which determines the position and order of the keywords in a data set.

These keywords as well as the corresponding grammar have to be learned by the user. This learned common language enables the user to communicate with the program, i.e. to describe his problem to the program. The program itself now understands the problem, can process it and generate corresponding results.

LV-PIPE II is a stress calculation program for spatial piping systems. Due to the integration in Excel, it is characterized by an exceptionally simple operation and flexibility. It enables fast and comprehensive analysis of piping systems of any geometric configuration and complexity.

The program was created in 1978 on mainframes, first ported to DOS, then to Windows 3.x /9x and finally to Windows NT, 2000, XP, Vista, Windows 7. We decided to port the program because it has been successfully in use for all these years, a rebuild can always be buggy and we are convinced of the modernity of our program.

The LV-PIPE II program determines, among others, the following quantities:

Data are determined for the pure weight load case (assembly), the pure temperature load case (startup/shutdown) and for the combination load case (operation). For easier evaluation, the most important variables can be displayed graphically.

Due to the integration with Excel, LV-PIPE II has extensive possibilities for communication with other programs. The fastest to integrate are the modules of LV, which also have an Excel option. Determined piping data can thus be processed e.g. directly into a flange calculation according to EN 1591, ASME etc. or embedded into a pressure loss calculation.