The model has both very long and very short pipes
A model with both long and short pipes requires a larger number of pipe sections in order to perform calculations. The more pipe sections in a model, the more computations the solver needs to perform, which results in a longer run time.
How are short and long pipe lengths contributing to long run times?
To run a transient simulation, AFT Impulse sections pipes to apply the Method of Characteristics solution to the waterhammer equations. In any system there will be a single controlling pipe. The controlling pipe is the pipe with the fewest sections - generally one. AFT Impulse uses the length of a section in the controlling pipe to divide all pipes in the model into sections and determine a uniform time step for the transient simulation.
When a model has both very long and very short pipes, the short pipes determine the controlling pipe length, and the long pipes must be divided into many sections to properly apply the Method of Characteristics. The run time is proportional to the square of the number of sections, meaning many sections in a model will often cause extremely long run times.
Consider a model that consists of two pipes: a very short pipe, such as 1 meter, and a very long pipe, such as 10,000 meters. The 1 meter pipe will be the controlling pipe, and the 10,000 meter pipe must be broken into 10,000 sections. The model thus requires 10,001 sections in order to perform the calculations, which results in a long run time. Increasing the short pipe’s length artificially from 1 to 10 meters decreases the number of sections required by a factor of 10, reducing run time by a factor of 100. While the length increase may also increase the error in the short pipe, experience shows that the error is usually negligible in a liquid transient context.
What can the user do?
Several different approaches are available to reduce run time. These approaches can work in parallel, and it is up to the engineer to make appropriate judgments as to which is best for the model.
Consider using the zero-length connector for very short pipes. Oftentimes two fittings in a system are connected with a very short pipe or are welded together directly. In AFT Impulse, it is common for users to attempt to model these connections using pipes with very short defined lengths. These short pipes offer little accuracy improvement and result in a large number of sections for the solver to calculate. This also causes significant time for the model to run. The zero-length connector option allows AFT Impulse to exclude the pipe from the sectioning process since it represents a lossless connection between two junctions. Zero-length connectors can be used in up to two sequential pipes, but intermediate branching flow is not allowed between two zero-length connector pipes.
Consider modeling fittings and losses in pipes and combining multiple short pipes. In the Pipe Properties window, use the Fittings & Losses tab to add in fittings to the pipe instead of using junctions on the workspace. This is useful to account for pressure losses without modeling large numbers of elbows or similar fittings. For example, a user can add valves and bends to a pipe, and the solver will still apply the losses directly to the pipe during the calculations. By removing these junctions and combining pipes in the workspace, many short pipes between these fittings are eliminated. Reducing and combing short pipes reduces the number of sections in the model and the overall run time.
Consider adding artificial length to short pipes. Adding length to short pipes increases the length of the controlling pipe, which results in fewer sections for the other pipes in the model. Reducing the overall number of sections in the model improves run time. Using pipe lengths that are common multiples, such as 5 feet, 10 feet, 15 feet, etc. provides improved sectioning since short pipes can be perfectly divided into longer pipes.
Do not include extremely long pipes unless you are directly analyzing the transient behavior in those pipes. Consider using the Infinite Pipe feature in the Assigned Pressure and Assigned Flow junctions. The infinite pipe feature prevents the junction from reflecting pressure waves, which can be useful in modeling the part of a pipe system which is so long that its communication time prevents wave reflection interactions with the rest of the system during the simulation.
Related Topics
Related Blogs
Analyzing Transients Faster - a Hacker's Guide to Pipe Sectioning