Multi-field coupling dynamic modeling and simulation of turbine test rig gas system

https://doi.org/10.1016/j.simpat.2014.03.004Get rights and content

Highlights

  • Multi-field coupling numerical system of flow, heat transfer, combustion, rotation.

  • Turbine is connected to main test system by partially integrated zooming method.

  • A low-cost and convenient virtual test platform is established for engineering test.

  • 42-component system makes improvements against many weaknesses of previous research.

  • Key factors which affect simulation accuracy and stability are revealed.

Abstract

On the basis of early research on the 38-component system, this paper extends the established finite volume model to the multi-field coupling numerical system which can describe flow, heat transfer, combustion and rotation, and conducts, based on the experiment, dynamic modeling and simulation research on the turbine test rig gas system which includes turbine and load. The comparison among the simulation results, test data and early simulation results indicates the 42-component system established by this paper has made improvements against many weaknesses of the Previous simulation in an all-round way. Accordingly, it can be concluded that the case setting and algorithm improvement are effective. It is also found that the modeling of module with chemical reaction, the spool throttling modeling of various regulator valves, the modeling of the turbine characteristics and the modeling of wall heat transfer are four key factors which affect simulation accuracy, and that the coupling modeling of flow and combustion is the key factor which affects simulation stability.

Introduction

The idea of modularization [1] is gradually formed to meet the demand of versatile system modeling. Its basic idea is firstly viewing the system as an assembly of some typical components which are called modules, secondly establishing numerical model of every module and encapsulating it as an independent function module, finally establishing numerical model of the whole system by combination of relevant modules according to special rules. As all components of the same type are described by one module, the modeling and simulation problem of all kinds of systems with different structures can be solved conveniently by computer. In the liquid propulsion system (LPS) simulation field, this idea is deemed to be put forward by A.P. Tishi and L.P. Gurova [2]. Since late 1980s, relevant researches have been evolving toward maturity and put into application in the form of simulation software [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25]. Ref. [26] summarizes the relevant researches and points out that future improving direction in the aspect of algorithm will focus on more detailed and accurate modular modeling of typical component, and integration of simulation and optimization function for optimization application.

At the level of component research, the high-fidelity full-scale three-dimensional (3D) or two-dimensional (2D) modeling and simulation can be conducted by employing the leading CFD (Computational Fluid Dynamics) software tools (e.g. CFX, FLUENT) for many components of LPS. However, at the level of system research, that has not become the mainstream practice mainly because: for a transient complex system consisting of tens of or even hundreds of components, the CFD softwares like CFX and FLUENT, with current technological conditions, are not able to undertake the system-level modeling and simulation in terms of complexity and task volume because it involves multi-component and multi-disciplinary combined transient modeling of components with different structures and functions, flow of different fluids and various non-pure flow phenomena; even if the modeling was successfully made with the CFD tools, the computing cycles would not be acceptable due to the huge computation.

Therefore, a reasonable balance must be made between the accuracy of component models and the complexity of the whole system for the system-level research. The general approach is to use 1D or even 0D simplified models to conduct modeling and simulation, with focus on the overall performance of the system and the specific role of a single component in the dynamic change process of the system. Currently, the relatively mature system-level simulation software tools, such as ROCETS [3] and GFSSP [6], [7] (USA), AMESim [10] (France) and Flowmaster [12] (UK), are developed in this way and put into commercial use. Moreover, according to the researches [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25] of recent years, the modeling method in the system-level simulation field, represented by such software tools as ROCETS [13], [14], GFSSP [15], [16], AMESim [17], CARINS [18], Flowmaster [19], Matlab/Simulink [20], [21] and so on [22], [23], [24], [25], does not change intrinsically. These software tools just have more functions and more extensive application scope by constantly adding new modules and improving algorithm, and are applied to modeling, simulation and optimization of some specific complex systems.

It is worth noting that NASA (National Aeronautics and Space Administration) has realized multidisciplinary analysis on the whole engine system at a variety of resolution levels (from 1D steady state to 3D transient state) in the NPSS (Numerical Propulsion System Simulation) project [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47] which has been carried out for more than a decade, but the 1D framework of the flow field simulation, NCP [27], [28] (National Cycle Program, released by the Lewis Research Center, the predecessor of NASA Glenn Research Center, during 1997–1998), was still developed by the foregoing method. Furthermore, the high-dimensional frameworks [29], [30], [31] are coupling systems consisting of many single-discipline software tools such as SINDA (Heat transfer), ENG10 and ADPAC (Fluids), TETRA and NASTRAN (Structures). Particularly emphasizing on 2D or 3D quasi-steady simulation, the high-dimensional frameworks do not give a solution to high-dimensional multi-disciplinary transient combined simulation of flow, heat transfer and combustion.

The NPSS project was proposed by NASA Glenn Research Center in late 1990s [32] with the primary purpose of achieving high-fidelity system analysis ability by the integrated computer simulation of many disciplines including fluid mechanics, heat transfer, combustion, structural mechanics, materials, controls, manufacturing and so on in order to improve the confidence level of design and decrease the cost of test and related hardware facilities. The ultimate goal of NPSS is modeling the entire propulsion system at the highest level of fidelity (3D transient and multidisciplinary) for revealing the physical mechanisms that affect system characteristics, while two problems prevent this from being a viable option in most cases [33], [34]: first, the level of detailed information needed as boundary and initial conditions to get a converged, validated solution would be extremely difficult to collect; second, the computational time and cost would be prohibitively high for effective use in a design environment. Therefore, a kind of important method called zooming [29], [30], [31], [32], [33], [34], [35], [36], [37] is presented for multi-components coupling simulation in NPSS framework. Zooming is defined as the “automated, seamless integration of one or more high resolution analyses with a full system simulation” [35], [36]. Its basic idea is integrating the high-dimensional models of some key components (such as fan, compressor, combustor and turbo) in an aerospace system with the low-dimensional full engine model by the directly linking approach (fully integrated zooming) or the characteristic-map approach (partially integrated zooming) [34], [38], [39] to conduct variable-fidelity simulation of the entire system. The zooming methods have been extensively used in the simulation and optimization design of aerospace propulsion systems and components [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45]. However, there is a lack of explicit application report on full-system high-dimensional transient simulation.

As a continuation of the numerical study on the turbine test rig system, this paper connects the turbine component to main test system [26] by the characteristic-map approach in zooming methods, and carries out the modeling and simulation of full gas system of the test rig by expanding the algorithm system and improving the regulating time sequence.

Section snippets

Physical model of the system

Fig. 1 shows the schematic diagram of the turbine test rig system which can be approximately divided into three parts: main test system, fuel–oil supply and ignition system, turbine and load system. The function of main test system is to provide driving gas for rotation of the test turbine. The function of fuel–oil supply and ignition system is to provide fuel, which then mixes with air to generate high-temperature and high-pressure gas after combusting in the BK-1 type heater as the driving

Modeling approach

Based on the finite volume model system [48], [49] developed by us which attributes to the improvement and evolution of the finite element state-variable model system [1], Ref. [26] has given an introduction to the finite volume model used for calculating the transient flow field of quasi-one-dimensional compressible variable-section pipe flow, the axisymmetric two-dimensional finite volume model used for calculating pipe-wall transient heat transfer, the valve spool and orifice throttling

System model

The turbine test rig is a complex system, where there are numerous physical and chemical phenomena such as atomization, ignition, combustion, mixing, rotation, flow and heat transfer, as well as flows of various media like fuel oil, water, normal-temperature air, high-temperature combustion gas and mixed gas. Furthermore, each flow is controlled by many regulator valves of corresponding types. Considering the great difficulty in the entire system modeling, Ref. [26] only conducts modeling and

Results and discussion

Fig. 5 shows the experimental curves of static pressures, total pressures, static temperatures, total temperatures, wall temperatures, fuel–oil flow rate and turbine rotate speed in the turbine test rig system which are measured by sensors in this rig regulating test. The data acquisition frequency stands at 1 Hz, i.e. a data point per second. The experimental curves of ∼2500 s have recorded the variations of state parameters at various measuring points in the speed-up process of the turbine from

Conclusions

Compared with the Previous research [26], the research in this paper adds two brand-new modules – gas turbine and rotor, improves the algorithm of module with chemical reaction to solve the problem in the stability of simulation, adds the radiation heat transfer algorithm based on the algorithm of natural convection heat transfer between wall and environment to enhance wall heat transfer, adds the regulator valve F7 and improves the spool throttling models, modifies the model size of the heater

Acknowledgments

This work is financially supported by the National Natural Science Foundation of China (No. 11101023) and the China Scholarship Council (No. 201203070237). The authors would like to thank Dr. Peng Xu from the research team for his guidance in software interface, Professor Haixing Wang from BUAA and the anonymous referees for their revision advice, Dr. Huasheng Wang from Queen Mary, University of London and Ms. Jihong Zhao for their language support.

References (51)

  • M. Sozen, A.K. Majumdar, A novel approach for modeling chemical reactions in generalized fluid system simulation...
  • A. Tarafder, S. Sarangi, CRESP-LP: a dynamic simulator for liquid-propellant rocket engines, in: AIAA 2000-3768,...
  • N. Yamanishi, T. Kimura, M. Takahashi, et al., Transient analysis of the LE-7A rocket engine using the rocket engine...
  • IMAGINE S.A., AMESim4.2 User Manual,...
  • V. Leudiere, G. Albano, G. Ordonneau, et al., CARINS: a versatile and flexible tool for engine transient prediction...
  • Flowmaster International Ltd., Flowmaster version 6.2 user guide, Flowmaster International Ltd., Towcester, 2006....
  • J.E. Crowley, J.R. Olds, Co-OPT: a constrained optimizer for propulsion tools, in: AIAA 2005-4126,...
  • K.W. Nelson, S.P. Simpson, Engine system model development for nuclear thermal propulsion, in: AIAA 2006-5087,...
  • R.D. Salvo, S. Deaconu, A.K. Majumdar, Development and implementation of non-Newtonian rheology into the generalized...
  • A.C. LeClair, A.K. Majumdar, Computational model of the chilldown and propellant loading of the space shuttle external...
  • X.Y. Ren et al.

    Simulation of turbofan engine main fuel control system based on AMESim

    J Aerosp Power

    (2010)
  • V. Leudiere, P. Supie, M. Villa, KVD1 engine in LOX/CH4, in: AIAA 2007-5446,...
  • H.W. Ng et al.

    Simulation of fuel behaviour during aircraft in-flight refueling

    Aircraft Eng. Aerosp. Technol.

    (2009)
  • B.T. Burchett, Simulink model of the Ares I upper stage main propulsion system, in: AIAA 2008-6544,...
  • D.K. Frederick, A new method for constructing fast models of jet engines in Simulink, in: AIAA 2009-5419,...
  • Cited by (0)

    View full text