An Optimization Procedure for the Aerodynamic Model Tuning of Centrifugal Compressor Stages

2011 ◽  
Author(s):  
Susanne Svensdotter ◽  
Omar El Shamy ◽  
Nidal Ghizawi ◽  
Vittorio Michelassi ◽  
Sivasubramaniyan Sankaran
Keyword(s):  
Experimental Data ◽  
Centrifugal Compressor ◽  
Optimization Procedure ◽  
Optimization Method ◽  
Design Flow ◽  
Flow Coefficient ◽  
Prediction Tool ◽  
Tuning Parameters ◽  
Model Tuning ◽  
And Performance

This paper presents an automated optimization procedure for tuning and optimizing the performance parameters of centrifugal compressor stages in order to improve the accuracy of a 1D performance prediction tool and performance database. An in-house, well-validated 1D tool is used to predict the performance of centrifugal compressor stages. The stages are usually tested under similitude conditions in order to verify the predicted performance with the experimental data. Continuous improvements have been done on the tool to improve its accuracy, but the tuning to test data is still done manually and separately for each tested design flow coefficient. As a further leap in this activity, an in-house developed optimization code (PEZ) is interfaced with the 1D prediction tool to provide the best possible solution within the given tuning limits. This provides the possibility to use an extended number of tuning parameters and to tune the entire design family simultaneously, thereby ensuring a smooth evolution of the tuning parameters within the database. The optimization plan consists of a Differential Evolution (DE) genetic algorithm followed by a simplex-based optimization method (AMOEBA) with an objective of reducing the Root Mean Square (RMS) value of the error with the specified constraints. The procedure was successfully challenged with several families of similar stages but with various design corrected mass flows, by setting different objective/constraints combinations. The Optimizer was able to reduce the total RMS value of the error by approximately 80% with respect to the baseline for one of the recently tuned families. The result is a minimal deviation between predicted and experimental data for entire families, as well as a significant time reduction compared to the previous tuning methodology.

Redesign and Performance Analysis of Medium Flow Coefficient Centrifugal Compressor

2019 ◽  
Author(s):  
Khaled Yousef ◽  
Abraham Engeda ◽  
Ahmed Hegazy
Keyword(s):  
Experimental Data ◽  
Centrifugal Compressor ◽  
Program Analysis ◽  
Cfd Simulation ◽  
Analysis Data ◽  
Performance Comparison ◽  
Flow Coefficient ◽  
And Performance ◽  
Performance Discrepancy ◽  
House Design

Abstract In-house design/analysis program is used in this study to redesign and predict the entire performance of an existing centrifugal compressor. The centrifugal compressor experimental data from a cooperative company is validated first with the 3D CFD simulation for a single passage flow. The validation shows an acceptable performance discrepancy between experimental data and CFD for the existing impeller. Then, a redesign procedure by the in-house program is carried out to get more efficient impeller based on the experimental data. This redesigned impeller is re-implemented in the CFD simulation to be compared with in-house program analysis data. The validation shows a good performance comparison between CFD and in-house program for the redesigned impeller. Moreover, this procedure succeeds to improve the redesigned stage efficiency and work coefficient with 6.79%, and 10.55% compared to the experimental impeller, respectively.

Steady and Transient Computations of Interaction Effects in a Centrifugal Compressor With Different Types of Diffusers

2010 ◽  
Author(s):  
S. Anish ◽  
N. Sitaram
Keyword(s):  
Centrifugal Compressor ◽  
Static Pressure ◽  
Leading Edge ◽  
Design Flow ◽  
Flow Coefficient ◽  
Pressure Recovery ◽  
Maximum Efficiency ◽  
Vaned Diffuser ◽  
Lower Flow ◽  
Transient Simulations

A computational study has been conducted to analyze the performance of a centrifugal compressor under various levels of impeller-diffuser interactions. The study has been conducted using a low solidity vaned diffuser (LSVD), a conventional vaned diffuser (VD) and a vaneless diffuser (VLD). The study is carried out using Reynolds-Averaged Navier-Stokes simulations. A commercial software ANSYS CFX is used for this purpose. The intensity of interaction is varied by keeping the diffuser vane leading edge at three different radial locations. Frozen rotor and transient simulations are carried out at four different flow coefficients. At design flow coefficient maximum efficiency occurs when the leading edge is at R3 (ratio of radius of the diffuser leading edge to the impeller tip radius) = 1.10. At lower flow coefficient higher stage efficiency occurs when the diffuser vanes are kept at R3 = 1.15 and at higher flow coefficient R3 = 1.05 gives better efficiency. It is observed that at lower flow coefficients positive incidence causes separation of flow at the suction side of the diffuser vane. When the flow rate is above design point there is a negative incidence at the leading edge of the diffuser vane which causes separation of flow from the pressure side of the diffuser vane. Compressor stage performance as well as performance of individual components is calculated at different time steps. Large variations in the stage performances at off-design flow coefficients are observed. The static pressure recovery coefficient (Cp) value is found to be varying with the relative position of impeller and diffuser. It is observed that maximum Cp value occurred at time step where Ψloss value is lowest. From the transient simulations it has been found that the strength and location of impeller exit wake affect the diffuser vane loading which in turn influences the diffuser static pressure recovery.

Affinity Law Modified to Predict the Pump Head Performance for Different Viscosities Using the Morrison Number

2018 ◽  
Vol 141 (2) ◽  
Author(s):  
Abhay Patil ◽  
Gerald Morrison
Keyword(s):  
Experimental Data ◽  
Reynolds Number ◽  
Flow Coefficient ◽  
Design Parameters ◽  
Sharp Change ◽  
Fluid Viscosity ◽  
Pump Head ◽  
Laminar And Turbulent Flow ◽  
Computational Fluid Dynamics Cfd ◽  
And Performance

The goal of this study is to provide pump users a simple means to predict a pump's performance change due to changing fluid viscosity. During the initial investigation, it has been demonstrated that pump performance can be represented in terms of the head coefficient, flow coefficient, and rotational Reynolds number with the head coefficient data for all viscosities falling on the same curve when presented as a function of ф*Rew−a. Further evaluation of the pump using computational fluid dynamics (CFD) simulations for wider range of viscosities demonstrated that the value of a (Morrison number) changes as the rotational Reynolds number increases. There is a sharp change in Morrison number in the range of 104<Rew<3*104 indicating a possible flow regime change between laminar and turbulent flow. The experimental data from previously published literature were utilized to determine the variation in the Morrison number as the function of rotational Reynolds number and specific speed. The Morrison number obtained from the CFD study was utilized to predict the head performance for the pump with known design parameters and performance from published literature. The results agree well with experimental data. The method presented in this paper can be used to establish a procedure to predict any pump's performance for different viscosities; however, more data are required to completely build the Morrison number plot.

Flow Studies on a Centrifugal Compressor Stage With Low Solidity Diffuser Vanes

2002 ◽  
Author(s):  
Prasad Mukkavilli ◽  
G. Rama Raju ◽  
A. Dasgupta ◽  
G. V. Ramana Murty ◽  
K. V. Jagadeshwar Chary
Keyword(s):  
Centrifugal Compressor ◽  
Design Flow ◽  
Flow Coefficient ◽  
Compressor Stage ◽  
Vaneless Diffuser ◽  
Centrifugal Compressor Stage ◽  
Setting Angle ◽  
Set Up ◽  
Flow Improvement ◽  
Stagger Angle

Diffusers are found to play a significant role in the performance of centrifugal compressors. Extensive studies have been in progress in various research laboratories for improvement of performance with various types of diffusers. One such effort for study of performance of a centrifugal compressor stage with Low Solidity Diffuser (LSD) vanes is presented in this paper. The study was conducted at a tip mach number of 0.35. An exclusive test rig was set up for carrying out these flow studies. The LSD vane is formed using standard NACA profile with marginal modification at the trailing edge region. The study encompasses the variation of setting angle of the LSD vane and the vane solidity. The effect of solidity and the setting angle on overall stage performance is evaluated in terms of flow coefficient, head coefficient and efficiency normalised with respect to these parameters for the case of vaneless diffuser at design flow. Improvement in performance as well as static pressure recovery was observed with LSD as compared to vaneless diffuser configuration. It is concluded from these studies that there is an optimum solidity and stagger angle for the given stage with LSD vanes for the chosen configuration.

Aerodynamic Design and Rig Test Validation of a High Flow Coefficient Process Centrifugal Compressor Stage

2021 ◽  
Author(s):  
Louis Larosiliere ◽  
Vishal Jariwala ◽  
Kapil Panchal
Keyword(s):  
Centrifugal Compressor ◽  
High Flow ◽  
Design Flow ◽  
Flow Coefficient ◽  
Test Validation ◽  
Lapse Rate ◽  
Compressor Stage ◽  
Aerodynamic Design ◽  
Centrifugal Compressor Stage ◽  
Very High

Abstract Efficient and diametrically compact very high flow coefficient stages with wide operability are desirable for economic reasons in many process multistage centrifugal compressor applications. Such stages present special aerodynamic and mechanical design challenges. There is often a sizeable efficiency lapse rate as well as substantial reduction in useable operating range for traditional stages having design flow coefficients greater than 0.15 and moderate to high machine Mach numbers. This paper describes aerodynamic design and rig test validation of a very high flow coefficient (φ0 = 0.237) process centrifugal compressor stage. Some useful experience of the detailed design work required to navigate certain technical challenges and its rig test validation are reflected in the manuscript. A relatively high machine Mach number (MU ∼ 0.878) mixed-flow shrouded impeller matched with a curved radial vaneless diffuser and return channel was developed. Test results confirmed that the principal aerodynamic design intents were met or exceeded. A sensible design strategy guided by a well-anchored design method is shown to successfully extend an existing stage portfolio to very high-flow coefficients for multistage process centrifugal compressor applications.

An Aerodynamic Optimization Procedure for the Preliminary Design of Centrifugal Compressor Stages

2008 ◽  
Author(s):  
Omar Elshamy ◽  
Nidal Ghizawi ◽  
Ce´line Yon ◽  
Simone Pazzi ◽  
Denis Guenard
Keyword(s):  
Centrifugal Compressor ◽  
Optimization Procedure ◽  
Optimization Method ◽  
Computational Time ◽  
Preliminary Design ◽  
Trial And Error ◽  
Aerodynamic Optimization ◽  
Peak Efficiency ◽  
Stage Design ◽  
Design Variables

This paper presents an automated aerodynamic optimization procedure for the preliminary design of centrifugal compressors. The proposed procedure interfaces a well-validated prediction tool with a GE in-house developed optimization code (PEZ). In GE Oil & Gas this tool is used to predict the performance of a single centrifugal compressor stage the outline of which requires more than thirty geometric parameters to be set. In the early phase of a new stage design, the designer manually varies all related parameters in the framework of a trial-and-error approach. The optimization procedure eliminates the inconvenience of a vast amount of manually launched simulations required by variations of the large number of design variables. Additionally, this procedure can perform trade-off studies and sensitivity analysis. In this case the optimization plan consists of a differential evolution (DE) genetic algorithm followed by a simplex-based optimization method (AMOEBA). The procedure was challenged with several existing designs by setting different objective/constraints combinations. The optimizer was often able to improve the predicted performance, as for an old 2D design where it was possible to increase the peak efficiency of approximately 2.6%. Also, the algorithm proved able to maximize the polytropic head (+12% with respect to baseline), while keeping unaltered both surge and choke limits. The computational time was about 40 hours per case on a Windows workstation (3.20 GHz, 3.5 GB RAM).

Spanning Tree Based Topological Optimization of Compliant Mechanisms

2005 ◽  
Author(s):  
Hong Zhou ◽  
Kwun-Lon Ting
Keyword(s):  
Spanning Tree ◽  
Spanning Trees ◽  
Compliant Mechanisms ◽  
Optimization Procedure ◽  
Optimization Method ◽  
Topological Optimization ◽  
Deformation Analysis ◽  
Discrete Optimization Problem ◽  
Minimum Number ◽  
And Performance

In graph theory, spanning trees connect all the vertices together using minimum number of edges. A topological optimization method of compliant mechanisms is presented based on spanning tree theory. A valid topology is regarded as a network connecting input, output, support and intermediate nodes, which contains at least one spanning tree among the introduced nodes. Invalid disconnected topologies can be weeded out if no spanning tree is included. No further deformation analysis and performance evaluation is needed for invalid disconnected topologies. Problem-dependent objectives are optimized for topological optimization of compliant mechanisms. Constraints about maximum input displacement and input force, maximum stress and overlapping connections are directly imposed during the optimization process. The discrete optimization problem is solved by genetic algorithm with penalty function handling constraints. An example is presented to verify the effectiveness of the proposed optimization procedure.

An Optimization Method for Centrifugal Compressor Design Using the Surrogate Management Framework

2011 ◽  
Author(s):  
Kyoung Ku Ha ◽  
Shin Hyoung Kang
Keyword(s):  
Optimal Design ◽  
Centrifugal Compressor ◽  
Design Space ◽  
Optimization Method ◽  
Pattern Search ◽  
Flow Coefficient ◽  
Design Parameters ◽  
Centrifugal Compressors ◽  
Management Framework ◽  
Design Point

A variety of centrifugal compressors are used in various fields of industry these days. The design requirements are more complicated, and it is difficult to determine the optimal design point of a centrifugal compressor. The aim of this study was to propose an efficient optimization method for centrifugal compressors considering the impeller, the vaneless diffuser, and the overhung type volute. The optimization was performed using the surrogate management framework (SMF). The design parameters were the impeller exit radius, the exit blade angle, and the flow coefficient. Sample points in the design space were selected according to the Design of Experiments (DoE) theory. The CFD simulations were executed on the impeller and the diffuser at every sampled point. The volutes were described using a one-dimensional but reliable theory to reduce the simulation time. An approximation model based on the Kriging method was constructed using this dataset. Then, an optimal design point that minimized the objective function was determined in a substitute design space using the pattern search method because of its efficiency and rigorous convergence. The optimization process, underlying methods, and results are described in this paper.

Effects of Inlet Pressure Distortion on the Performance and Flow Field of a Centrifugal Compressor at Off-Design Conditions

2012 ◽  
Author(s):  
N. Sitaram ◽  
M. Govardhan ◽  
K. V. Murali
Keyword(s):  
Flow Field ◽  
Centrifugal Compressor ◽  
Total Pressure ◽  
Static Pressure ◽  
Design Flow ◽  
Flow Coefficient ◽  
Low Flow ◽  
Performance Measurements ◽  
Perforated Sheet ◽  
Sector Angle

The present paper presents experimental results on the effects of inlet total pressure distortion on the performance and flow field of a centrifugal compressor. The total pressure at inlet is artificially distorted by means of a perforated sheet, which is supported by a support mesh. A total of eleven configurations, including clean inlet configuration, are tested. Performance measurements and impeller inlet and exit flow studies at three flow coefficients, one near design flow coefficient, one below design flow coefficient and one above design flow coefficient, are carried out. The present paper presents and discusses results at off-design flow coefficients and the effects of stage loading on the distortion effects are presented. A new parameter, Distortion Index (DI) is introduced. As DI increases, the mass averaged total pressure at exit stations decreases. Distortion sector angle of 60° having the lowest total pressure is found to be the critical sector for circumferential distortion configurations. As the Distortion Correlation parameter, DC(60) increases, the mass averaged total pressure for circumferential distortion configuration decreases, except in the case of low flow coefficient where DC(60) is nearly constant. DC(60) also increases with sector angle. The static pressure normalized with static pressure for clean inlet decreases as the distortion sector angle is increased. Distortion attenuates the static pressure as the flow passes through the vaneless diffuser. The attenuation increases with the distortion sector angle.

Analysis and Redesign of the Stationary Components for a Low Flow Coefficient Centrifugal Compressor Stage

2009 ◽  
Author(s):  
Zhiheng Wang ◽  
Guang Xi
Keyword(s):  
Centrifugal Compressor ◽  
Design Flow ◽  
Flow Coefficient ◽  
Low Flow ◽  
Meridional Plane ◽  
Compressor Stage ◽  
Centrifugal Compressor Stage ◽  
Lower Stage ◽  
Wide Range ◽  
Low Flow Coefficient

A low flow coefficient centrifugal compressor stage is characterized by the small relative outlet width, and is often one of the latter stages in the multistage compressor. The low flow coefficient stage is known to give lower stage efficiency in comparison with the conventional stage, which still leaves much more space to be improved with modern tools such as CFD techniques. In the paper the flow in a CO2 centrifugal compressor stage with a low design flow coefficient of 0.008 is simulated based on the 3D viscous CFD codes. The analysis shows the impeller gives a favorable performance over a wide range of low flow coefficient, but the high losses exist in the stationary components and this incurs the poor performance of the whole stage. In this case, the diffuser, the return channel and the meridional plane are redesigned. The redesigned stage has distinct improvements on the performance and the flow structure.

Sign in / Sign up

Export Citation Format

Share Document