Elsevier

Journal of Process Control

Volume 106, October 2021, Pages 72-83
Journal of Process Control

A new multivariable control concept for the falling film evaporator process

https://doi.org/10.1016/j.jprocont.2021.08.015Get rights and content

Highlights

  • The main control challenges w.r.t. falling film evaporators (FFEs) are addressed.

  • Robustness w.r.t. parametric plant-model mismatch is shown in a simulation study.

  • Robustness w.r.t. structural plant-model mismatch is shown via a digital twin.

  • An automated ramp-up strategy for the FFE process is proposed and validated.

  • Simple implementation of the concept to standard digital control systems is enabled.

Abstract

The paper presents a new multivariable control concept for falling film evaporators (FFEs). Our concept solves the major challenges encountered in modern FFE control: large transport delays, additional control of the output mass flow, coupling of controlled variables, and disturbances due to time-varying input dry matter content. The challenges are addressed together, for the first time, by the following control design. Based on a dynamic nonlinear input–output model, we consider a linearizing output transformation to enable application of classical linear control methods composed of feedforward design, disturbance rejection, and a decoupling network. Due to these features, we are able to design robust PID and PI controllers that substantially compensate plant-model mismatches. Connecting our concept to a digital twin of the plant yields good performance, which encourages future application of the design in the real-world process.

Section snippets

Introduction to the falling film evaporator process

Falling film evaporators (FFEs) have a wide range of application in various industries. Especially, FFEs are very common in the food industry as part of a production line with downstream spray dryer to produce, e.g., milk or coffee powder. In this context, the FFE’s main task consists in increasing the dry matter content of the liquid. Due to hygienic constraints and to ensure high product quality, FFEs are often built up of multiple passes enabling slow and cautious evaporation of the liquid

State of the art on control of falling film evaporators

PI controllers are still standard in the industry [5]. However, it is commonly known that a PI controller is unable to adequately cope with dominant time delays since it typically leads to large-amplitude, long-lasting oscillations of wo. In particular, Winchester et al. [4] conclude that pure single-loop PI control is insufficient to reject disturbances due to wi.

Therefore, in the last decades, more advanced methods have been proposed and applied to enhance the wo-control loop. In [6], [7], a

Control-oriented model

Our control-oriented model was originally introduced in [3]. However, since this model is the basis for our control concept, we recall relevant details of the derivation in this section. Subsequently, we convert physical variables into control nomenclature, i.e., define manipulated variables u, states x, disturbances z, and outputs y. Thus, we specify the model’s state space representation, which reveals linear state equations and a nonlinear output equation. Finally, we transform the linear

New multivariable control concept

The design of our control concept is based on the Plant in Fig. 4 with transfer functions (21). In Section 4.1, a stepwise derivation of the concept is given. To avoid operation beyond the actuators’ limits, we derive a domain of feasible setpoints in Section 4.2.

Simulation

In this section, the control concept presented in Section 4 is tested w.r.t. parameter uncertainties, see Section 5.1, and validated via a digital plant twin in Section 5.2.

However, before the concept is simulated, we have to define a suitable test signal for the disturbance z=wi. To this end, the following scenario is considered: The liquid fed into the FFE comes from a feed tank, where the dry matter content at the bottom is larger than at the top due to imperfect mixing and sedimentation.

Conclusions

In this paper, a new multivariable control concept for the falling film evaporator process is developed. As single-loop control of only wo and manual ramp-up are common industrial practice, our concept additionally enables control of ṁo and an automated ramp-up process. Moreover, since our concept is represented in a block diagram structure, see Fig. 7, practical implementation into digital control systems is simplified. Consequently, it is readily accepted by plant operators and also cheaper

CRediT authorship contribution statement

Julian Hofmann: Methodology, Software, Validation, Investigation, Data curation, Writing – original draft. Anton Ponomarev: Methodology, Software, Formal analysis, Writing – review & editing. Veit Hagenmeyer: Funding acquisition, Project administration, Writing – review & editing. Lutz Gröll: Conceptualization, Supervision, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the Helmholtz Association under the program “Energy System Design”. Additionally, the authors thank GEA Wiegand GmbH for data and technical information regarding the falling film evaporator process.

References (43)

  • Bekiaris-LiberisN. et al.

    Stability of predictor-based feedback for nonlinear systems with distributed input delay

    Automatica

    (2016)
  • ClergetC.-H. et al.

    Dynamic optimization of processes with time varying hydraulic delays

    J. Process Control

    (2019)
  • AlbertC. et al.

    Direct numerical simulation of interfacial mass transfer into falling films

    Int. J. Heat Mass Transfer

    (2014)
  • PonomarevA. et al.

    Novel control-oriented models for liquid transport in falling film evaporator tubes

    Comput. Chem. Eng.

    (2021)
  • KharangateC.R. et al.

    Computational modeling of turbulent evaporating falling films

    Int. J. Heat Mass Transfer

    (2015)
  • HofmannJ. et al.

    Transport models for liquid films

    Automatisierungstechnik

    (2020)
  • HofmannJ. et al.

    Control loop pairing and interaction analyses of the falling film evaporator process

  • O’CallaghanD. et al.

    Modern process control techniques in the production of dried milk products – a review

    Le Lait

    (2005)
  • ParamalingamS.

    Modelling, Optimisation and Control of a Falling-Film Evaporator

    (2004)
  • KarimiM. et al.

    Nonlinear modeling and cascade control design for multi effect falling film evaporators

    Iran. J. Chem. Eng.

    (2006)
  • FarsiM. et al.

    A new control algorithm for concentration control in three effect falling film evaporators

    Iran. J. Sci. Technol. Trans. B Eng.

    (2009)
  • Cited by (2)

    View full text