Automated construction of compressor performance maps in pSeven Enterprise

August 24, 2023

Introduction

Compressor performance map, also known as a compressor characteristic map or just compressor map, is a graphical representation of the performance of a compressor. It is typically presented as a two-dimensional graph with the mass flow on the horizontal axis and the pressure ratio (ratio of the outlet pressure to the inlet pressure of the compressor) on the vertical axis.

Compressor performance map shows the compressor's efficiency, pressure ratio and surge limit across its operating range. The efficiency is shown as contour lines, with higher efficiency represented by lines that are closer together. The pressure ratio is usually shown as a series of curves (branches), each corresponding to a different speed (example shown in Fig.1).

When the stages form a gas turbine, the range of possible operating conditions depends on the equilibrium of the gas turbine as a whole. The equilibrium points can also be plotted on the compressor map to form the operating line. Sometimes it may be necessary to calculate intersection points of the selected branches with the operating line.

The surge limit is the point at which the compressor is unable to maintain stable flow and begins to surge or stall. It is shown as a line on the compressor map, and represents the upper limit of the compressor's operating range. Surge points can only be determined approximately, therefore, depending on the required tolerance, further refinement may be required after the map is constructed.

A compressor map is an important tool for designers and engineers who work with compressors, as it allows them to analyse and optimize the compressor's performance under different operating conditions.

One way to build a compressor map is to run a series of CFD simulations. In this case, each point of the map corresponds to a single simulation and high quality of surge line is ensured by manual iteration over multiple values of the boundary conditions. Obviously, such approach requires a lot of manual work of a simulation engineer. Moreover, multiple parameters may change during the design phase: geometry of the blades, control program (dependency between the orientation of the blades and the rotational velocity), CFD simulation model, etc. Thus, it may be necessary to reconstruct the map several times as the design parameters change.

In this article, we will show how the process of a compressor map construction can be automated using pSeven Enterprise and what are the benefits of this approach. The CAD and CAE software used is only an example and can be replaced by any other software suitable for these tasks.

Process of compressor performance map construction

Steps of compressor map construction are shown in Fig. 2.

Construction begins by building geometric models of the blades of each stage in a CAD software. These models are then imported into an existing project of mesh generator to create the new mesh. The process repeats for each new branch as the blade orientation angle depends on rotational velocity of the branch.

After that, a series of CFD simulations is carried out using the newly generated mesh. Each branch is constructed by incrementing the value of the boundary conditions until the solution becomes numerically unstable (i.e., the surge limit is reached). Once this cycle is complete, one may further refine the surge boundary or compute intersections with the operating line.

The construction process requires manual data transfer between several applications. The process therefore requires constant attention from the simulation engineer. You cannot just run a simulation and pick up the results when it is finished, at least without a proper automation tool.

Automating the process in pSeven Enterprise

In pSeven Enterprise the process of the compressor map construction can be captured in the form of a workflow (Fig. 3).

The workflow consists of three main blocks:

• Block “Main driver” controls the calculation cycle of the branches and intersections.
• Composite block “Single branch” calculates one branch of the compressor map (for single rotation speed).
• Composite block “Single point” calculates the intersection of the operating line with one branch of the compressor map.

The nested structure of the “Single branch” block is shown in Fig. 4.

Block “File parsing” in Fig.4 reads the control program and the information about the number of blades in each stage, after that “geomTurbo generator” block runs a cycle of creating input files for the mesh generator based on the text files of the blade profile in NX.

Block “cgns generator” is responsible for the generation of the mesh of the flow part of the compressor in Numeca. Block “Condition” checks if it is necessary to calculate the branch, it either triggers the execution of the “Exit” block to exit the iteration or launches the branch construction cycle with the surge line refinement.

The nested structure of the “Single point” block is shown in Fig. 5.

Block “File management” in Fig.5 copies files for Ansys CFX study at a given point, searching for the desired initial approximation file, then “CFX Pre 1” and “CCL modification” blocks read information about the structure of Ansys CFX template project and update it with the new boundary conditions and rotation speed.

Block “CFX Pre 2” generates an input file for the solver with the updated mesh, boundary condition and rotation speed. Block “CFX Solver” triggers the simulation. If the run was successful, the found values are added to the branch by blocks “CFX Mondata” and “Parse CSV”. If numerical instability is encountered, then “Return NaN” block sends a message about reaching the surge line.

Workflow construction has been facilitated by the use of special blocks (Fig. 6). These blocks provide a seamless connection to engineering applications (NX, Numeca, Ansys CFX). Such blocks can be developed for any type of software that supports batch execution.

At the start of the workflow all the files needed for performance map construction must be placed in the «Files» folder on the Explorer panel according to the pre-agreed rules. After the workflow is completed successfully, all the produced files can be found on the same Explorer panel (shown in Fig.7).

Conclusions

The main benefit that pSeven Enterprise brings is increased efficiency and time savings. This is achieved by automating repetitive manual operations such as building geometry models, mesh generation, post-processing of the results and data exchange between different software. This approach can be applied to a wide range of tasks that can be standardised and need mass reuse.

The next level of automation can be achieved by developing a special user interface (UI). Example of UI is shown in Fig. 8. With such UI the workflow could be used as a Web App from AppsHub, a dedicated library for Apps in pSeven Enterprise. This approach would give the ability to conduct a study to an ordinary user who is not familiar with the case by simplifying the process of providing input data and configuring the task to construct the performance map. The user needs only to prepare the input data. However, it should be noted that developing such an interface would of course require some frontend programming skills.