Modeling of a double effect evaporator: Bond graph approach

https://doi.org/10.1016/j.cherd.2018.07.007Get rights and content

Highlights

  • Chemical engineering process is modeled using bond graph approach.

  • Interactions of different thermal and hydraulic phenomena are considered.

  • Energetic parameters are determinate by simulation study.

  • Experimental results are confronted to simulation results.

  • Analytical redundancy relations are developed for diagnosis and isolation of default.

Abstract

This paper deals with the modeling of a chemical engineering system, which is the double effect evaporator (DEE). The main physical phenomenon encountered in this system is the heat transfer during distillation and concentration process. The modeling of DEE is carried out using bond graph approach. It allows description of interdisciplinary systems involving energetic, thermodynamic, hydraulic and thermal phenomena. In this study a dynamical models for the different parts of the DEE system are established. Simulation data are presented and confronted to the experimental results. To supervise the DEE system, the bond graph model is used to obtain analytical redundancy relations (ARR). They allow detection and isolation of faults.

Introduction

Modeling and analysis of dynamic behavior of industrial processes needs acknowledgment of analytical expressions governing process physical laws. In literature, different modeling methods and tools are used such as software modeling tools for process engineering based on a stationary behaviors (Westerberg and Benjamin, 1985, Raman, 1986). Mathlab/Simulink libraries and Modelica modelling language are widely used in dynamic systems (Mathworks, 2004, Elmqvist et al., 1993, Wöllhaf et al., 1996). Object-oriented Dynamic Simulation Software Environment is used to model chemical processes for the simulation of their dynamic behavior, using fixed input/output causalities (Moyse et al., 1999). Aspen HYSYS software also allows modeling of chemical processes (Sunny et al., 2016, Lewis et al., 2010). Theses modeling softwares are based generally on a predefined model used just for simulation which means that the physical phenomena are not explicitly displayed to the user and derive symbolic state equations for non-linear systems (Ould Bouamama and Bouamama, 2003).

The bond graph approach has many advantages. The bond graph model is written with a unified representation language and obtained with a structure procedure. It conserves the physical meaning of parameters. Although with different physical domains, the model shows explicitly power flows and cause to effect relationships between all the elementary subsystems. The model leads to a graphical diagram allowing fast analyzing. Thus, to take into account a further phenomenon in an existing model, the latter is only extending without needing to rebuild the entire whole model. The model can also be rewritten to a Symantec language. A state space representation can easily be obtained and a reduced model can be deduced. The condition of the monitorability, observability and controllability can be graphically established and the configuration of sensors and actors can be discussed. Based on a state observer, a control low can designed even with non-measurable physical phenomena. The model can easily simulate a physical fault. The supervision and fault diagnosis and localization can be carried throw simple residues analyses. The bond graph approach was introduced for the first time in 1959 by Henry Paynter (Paynter, 1959, Paynter, 1958). It was further more developed by Rosenberg, Thoma, Margolis and Karnopp (Karnopp, 1979, Margolis, 1977, Rosenberg and Karnopp, 1972, Thoma, 1975). Some other works deal with the definition of state variables, the representation, and equation formulation in bond graph language (Jr and Rosenberg, 1980, Breedveld, 1988, Karnopp, 1975, Karnopp, 1979). This approach was used to study linear systems in terms of energy variables and was expanded later to cover systems containing nonlinearities interconnection of different energetic fields (Ould Bouamama and Bouamama, 2003, Karnopp, 1981, Martens, 1973, Samanta and Mukherjee, 1985, Tagina et al., 1995). The complex interactions of many types of energy in thermodynamics, hydraulic, mechanic, electric and chemical systems can be represented by bond graph language (Balino et al., 2006, Ould-Bouamama et al., 2012, Couenne et al., 2006, Ould Bouamama et al., 2009, Mezghanni et al., 2007, Youping et al., 2011, Thoma and Bouamama, 2013, Cauffriez et al., 2016, Shekhar et al., 2017).

The purpose of this study is to model a double effect evaporator (DEE) using the bond graph approach. Then, physical analytical redundancy relations are established for fault detection and isolation.

Section snippets

Process description

Evaporation is used to obtain a product with high added value such as distilled water or concentrated solution. The use of multiple effect evaporation is recommended to reduce energy consumption (Zhang et al., 2014). Only the energy used for the first effect is of external origin and each of the other effects using the steam produced in the previous one. The choice of the number of effect is determined according treated product (Al-Mutaz and Wazeer, 2014). DEE process is described by Fig. 1.

Mass balance

  • Preheating exchanger:m˙ee=m˙se

    where m˙ee and m˙se are the water inlet and outlet flow rate, respectively.

  • First effect evaporator:m˙se=m˙se1+m˙vp1

    m˙vp1 is the flow rate of the produced vapor at the first effect evaporator.

  • Second effect evaporator:m˙se1=m˙se2+m˙vp2

    m˙vp2 is the flow rate of the produced vapor at the second effect evaporator.

  • Condenser:m˙vp2=m˙c

  • Cooler:m˙sv=m˙er

    m˙ is a flow rate and the indices e, s, vp1, vp2, c and er designed inlet, outlet, vapor in first evaporator, vapor in

Modeling of a DEE system

Applying the concept of bond graph approach to the modeling of the evaporator system requires an identification of the possible subsystems which can be represented using words bond graph approach. It consists on a decomposition of the system into sub systems with two conjugate variables: effort and flow for example: in hydraulic respectively (pressure, mass flow) in thermal transfer respectively (temperature, flow enthalpy). The subsystems are all modeled.

Conclusion

To reduce energy consumption multiple effect evaporators are widely used in chemical industries (Kumar et al., 2013, Simpson et al., 2008, Higa et al., 2009, Sogut et al., 2010). In this process, vapor produced in the first effect is used to evaporate solution in the others one. Only in the first evaporator an external steam is used as a heating medium. Nevertheless it needs regularly servicing which the modeling and diagnosis of such system can preview it. This work provides a bond graph model

Acknowledgements

This work was supported by the Ministry of the Higher Education and Scientific Research in Tunisia.

References (48)

  • B. Samanta et al.

    Dynamics of a class of repeated system with nonidentical elastic and visco-elastic interconnections, a bond graph approach

    J. Franklin Inst.

    (1985)
  • R. Simpson et al.

    Optimum design and operating conditions of multiple effect evaporators: tomato paste

    J. Food Eng.

    (2008)
  • Z. Sogut et al.

    Energetic and exergetic performance evaluation of the quadruple-effect evaporator unit in tomato paste production

    Energy

    (2010)
  • A. Sunny et al.

    Syngas production from regasified liquefied natural gas and its simulation using Aspen HYSYS

    J. Nat. Gas Sci. Eng.

    (2016)
  • Y. Touati et al.

    Robust diagnosis to measurement uncertainties using bond graph approach: application to intelligent autonomous vehicle

    Mechatronics

    (2012)
  • A.W. Westerberg et al.

    Thoughts on a future equation-oriented flowsheeting system

    Comput. Chem. Eng.

    (1985)
  • J. Yang et al.

    A comparison of four numerical modeling approaches for enhanced shell-and-tube heat exchangers with experimental validation

    Appl. Therm. Eng.

    (2014)
  • G. Youping et al.

    Multi-functional loader steering hydraulic system model construction and simulation based on power bond graphs

    Proc. Eng.

    (2011)
  • X. Zhang et al.

    Performance analysis on a new multi-effect distillation combined with an open absorption heat transformer driven by waste heat

    Appl. Therm. Eng.

    (2014)
  • I.S. Al-Mutaz et al.

    Comparative performance evaluation of conventional multi-effect evaporation desalination processes

    Appl. Therm. Eng.

    (2014)
  • W. Borutzky

    Bond graph model-based fault detection using residual sinks

    Proc. Inst. Mech. Eng. I: J. Syst. Control Eng.

    (2009)
  • P.C. Breedveld

    Fundamentals of bond graphs

    Proc. 12th World Congr. Sci. Comput.

    (1988)
  • F. Couenne et al.

    Bond graph modelling for chemical reactors

    Math. Comput. Model. Dyn. Syst.

    (2006)
  • H. Elmqvist et al.

    Object-oriented modeling of hybrid systems

    Proc. Eur. Simul. Symp.

    (1993)
  • Cited by (7)

    View all citing articles on Scopus
    View full text