Microsoft Word - 211.docx


 CHEMICAL ENGINEERING TRANSACTIONS  
 

VOL. 61, 2017 

A publication of 

 

The Italian Association 
of Chemical Engineering 
Online at www.aidic.it/cet 

Guest Editors: Petar S Varbanov, Rongxin Su, Hon Loong Lam, Xia Liu, Jiří J Klemeš 
Copyright © 2017, AIDIC Servizi S.r.l. 

ISBN 978-88-95608-51-8; ISSN 2283-9216 

Integrated Process Design and Control: A Review of the 

Current State-of-the-Art and Research Challenges 

Alexios-Spyridon Kyriakides, Theodoros Zarogiannis, Panos Seferlis* 

Department of Mechanical Engineering, Aristotle University of Thessaloniki,  P.O. Box 484, 54124 Thessaloniki, Greece  

seferlis@auth.gr 

In chemical and energy conversion processes, the achievable dynamic performance cannot be improved unless 

design limitations are waived, despite the employment of sophisticated control strategies. It is therefore, quite 

important to incorporate process operability as an additional criterion during process design in order to ensure 

the achievable of the desirable dynamic performance. The two main purposes of this survey is to concisely 

present, classify, and eventually assess the current state-of-art technologies towards the integration of process 

design and control and further outline the outstanding research challenges in the field. Ideally, the simultaneous 

calculation of the process and the control system design variables within a holistic framework appears as the 

most efficient approach. Decomposition strategies facilitate the solution of the highly complex problem. The 

available technologies can be classified based on their features in the decision, process and uncertainty 

description, and solution technique levels. Several challenges remain in utilizing engineering knowledge and 

hierarchical design methods in defining a reasonable design space and the smart decomposition of the resulting 

complex and difficult to solve optimisation problems. Heat and mass integration as well as intensified processes 

add to the complexity of the process system by increasing interactions and interdependencies among various 

process subsystems, which eventually make the integrated design and control absolutely necessary. 

1. Introduction 

Nowadays, a wide variety of fine chemicals are produced under stringent sustainability requirements utilizing 

clean and renewable energy sources. A significant research effort has been invested in order to further improve 

process efficiency and limit exergy losses through a process design procedure that aims to optimize economic 

criteria (e.g., total annualized cost). Process design procedure involves several steps. A first step would be 

process synthesis, where alternative flowsheet structure configurations and unit interconnections are 

determined. Subsequently, the calculation of the structural parameters (e.g., equipment type, geometry, and 

dimensions) must be performed along with the determination of the nominal operating conditions. Usually, the 

evaluation of alternative process flowsheets is performed within an optimisation framework that utilizes 

engineering knowledge and experience in order to define a meaningful design space. Within such framework 

performance specifications, safety regulations and environmental constrains are explicitly taken into account, 

whereas a representative objective function encompasses the economic impact of the process system. 

However, process systems operate in a continuously changing environment due to exogenous disturbances, 

intrinsic process uncertainties, switching between different operating conditions and so forth. In order to satisfy 

product and safety specifications and environmental regulations under such variable conditions, the design of 

an efficient automatic control system is absolutely essential to achieve an acceptable dynamic performance. 

The selection of the control objectives that best serve the process specifications, the definition of the input-

output structure, the selection of the appropriate control algorithm and its tuning are key features of the control 

system design. In a sequential approach, where the major process design decisions for the process flowsheet 

have been determined before the design of the control system the achievable dynamic performance cannot 

overcome the restrictions imposed by the process structure. It is therefore important to simultaneously design 

process and control systems so that such burdens are lifted and superior performance is achieved under both 

nominal and variable operating conditions (Seferlis and Georgiadis, 2004). A schematic of the integrated 

process and control system design that incorporates the decisions that need to be made is shown in Figure 1. 

                               
 
 

 

 
   

                                                  
DOI: 10.3303/CET1761234

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Please cite this article as: Kyriakides A.-S., Zarogiannis T., Seferlis P., 2017, Integrated process design and control: a review of the current 
state-of-the-art and research challenges, Chemical Engineering Transactions, 61, 1417-1422  DOI:10.3303/CET1761234  

1417



 

Figure 1: Integrated process design and control procedure. 

2. Motivating process example 

Heat and mass integration of processes as well as process intensification generally improve greatly the 

efficiency and economic performance but usually results in highly complex processes where strong interactions 

among individual units make the control system design even more challenging and demanding, leading to plants 

that are difficult to control. Such an illustrative example is presented by Kyriakides et al. (2017), in which the 

integration of a catalytic reactor along with a membrane based separation unit are combined into an integrated 

membrane reactor (IMR) as shown in Figure 2a, for H2 production via low temperature methane steam reforming 

(MSR). The IMR is a process where the step of production (through MSR reaction) and the step of product 

purification (utilizing a Pd-based membrane) are conducted simultaneously. Reactants (CH4 and steam) are fed 

into the IMR, where the reforming reactions take place and at the same time H2 is separated through the 

membrane that is placed on the surface of the inner tube separating the reaction from the permeation zone. 

Since the equilibrium is shifted towards H2 production due to its removal from the reaction zone, a very high 

CH4 conversion at relatively low temperature levels can be achieved in such a reactor. Interaction in such a 

process system is extremely high as a disturbance in the balance of the reforming reactions may greatly affect 

both the reacting and the separation tasks. Disturbances in this system are frequent as the catalytic reactions 

need to be thermally balanced with the outer heat source and the removed H2 through the selective membrane. 

A simple alternative design consists of a cascaded arrangement between a reactor (Figure 2b) and a membrane 

based separator (Figure 2c) module in series (CRM), where H2 purification takes place only after the reaction 

step is completed. The reactive mixture is fed into the reactor and only after exiting the reactor module (at 

equilibrium composition) is fed to the membrane separator, where H2 is separated through the membrane (inner 

tube). Such configuration is simpler to construct and probably involves lower investment costs than the 

integrated membrane reactor module. However, the manipulation of the H2 purification conditions can be 

performed only after the manipulation of the reaction conditions. For example, when biogas (mainly consisted 

of CH4 and CO2) is utilized instead of pure CH4, where its composition may deviate greatly, concentration 

disturbances will affect H2 separation with a time delay at least equal to the mean residence time of the reactor. 

Such time delays may result in longer response periods by the control system with significant violation of the 

product specifications. On the other hand, the direct contact between the reactive mixture and the membrane, 

although it increases the complexity of the system, provides the opportunity for the simultaneous manipulation 

of the reaction and permeation zone operating conditions, making the system more prone to disturbances but 

simultaneously enabling a quicker response. 

A process system such as the one described above is difficult to be designed to operate optimally under variable 

operating conditions unless an integrated process and control system design approach is employed. Truly, a 

highly intensified process may possess several advantages in terms of controllability effectiveness even though 

higher investment costs are associated due to the uniqueness of the process equipment and the more 

sophisticated control algorithms that are required to be installed. In addition to the key flowsheet configuration 

decisions, the equipment sizing and connectivity of the streams and in particular recycle streams offer key 

opportunities to assist the ability of the overall process system to cope with disturbances and uncertainty. 

1418



 
a) 

b) 

c) 

Figure 2: a) Integrated membrane reactor module geometry (IMR flowsheet) and Cascaded b) reactor and c) 

membrane module geometry (CRM flowsheet). 

3. Integration of process design and control technologies 

A large variety of methods are proposed in the literature that address the simultaneous process and control 

system design utilizing various sets of comprehensive criteria, that aim to take into account the underlying 

interactions among the design decisions. Such criteria, that would consider the ability of the process to achieve 

a desirable dynamic performance, usually incorporate properties like feasibility, flexibility, controllability and 

operability. Feasibility is the ability of ensuring feasible steady-state operation, meaning satisfying the 

operational inequality constraints (e.g., operational limits, product specifications, safety regulations) for a chosen 

control action when design variables and uncertain parameters are given (Swaney and Grossmann, 1985). 

Flexibility, as described by Swaney and Grossmann (1985), is the ability of ensuring feasible steady-state and 

dynamic operation over a variety of operating conditions. Controllability, among different definitions that can be 

found in literature, most commonly is considered as an inherited characteristic of the system indicating the 

existence of the system’s (process and control system) ability to achieve the desired performance despite the 

presence of disturbances and uncertainties using available inputs and measurements (Skogestad and 

Postlethwaite 1996). Operability, which includes all the aforementioned features, describes the capability of the 

process (given the available set of external inputs) on satisfying the desired steady-state and dynamic conditions 

under the presence of uncertainties and anticipated disturbances, without violating any performance process 

constraints. 

3.1 Problem formulation 
The conceptual formulation of the problem of integration of design control in its most general form can be 

expressed as a mixed-integer dynamic optimisation (MIDO) problem. As shown in Eq(1), the optimisation 

problem that involves the maximization of a performance index is subject to the differential and algebraic 

equations of the process, the inequality constraints that define the feasible region and possibly the flexibility 

space and the controller differential and algebraic equations. 

Max (Performance Index) 

(1) 

s.t.: 
Process Physical Model (Differential and Algebraic Equations, Inequality Constraints) 

Controller Relations (Differential and Algebraic Equations) 

The decision variables of the problem in Eq(1) are the integer variables associated with flowsheet configuration, 

equipment and material selection and control system input-output structure as well as the continuous process 

and controller variables associated with the sizing of the selected equipment, the operating conditions and the 

tuning of the selected control algorithm. 

Mohideen et al. (1996) introduced a MIDO framework to simultaneously design and control dynamic systems in 

the presence of process uncertainty and disturbances. The optimisation was performed through an iterative 

decomposition technique that separated the problem into the solution of a multi-period design and control 

problem followed by the solution of a dynamic feasibility problem. Matrix norms were also used to account for 

process stability in the presence of uncertainty and disturbances. Luyben and Floudas (1994) proposed a similar 

problem formulation which combined process synthesis, design and control. The proposed procedure was 

consisted of four steps, where process synthesis was addressed through the search within a set of flowsheet 

superstructures that included several possible design alternatives. Then the superstructures were translated 

into mathematical models and the multi objective optimisation problem was solved using a generalized Benders 

decomposition (GBD) algorithm and finally the best compromise solution was obtained. Narraway and Perkins 

(1993) have proposed a method based on the minimization of the economic penalties associated with the back-

1419



off from active constraints. The importance of this contribution was the incorporation of economic goals into the 

problem definition. In most cases economic objectives are fulfilled on the intersection of the constraints but 

uncertainties and disturbances tend to violate these constraints. Therefore, the operating point should be moved 

within the feasible region away from the constraints intersections to enable feasible operation at the presence 

of disturbances. The back-off method was also extended to time domain controller design by considering 

decentralized (Heath et al.,2000) and centralized (Kookos and Perkins, 2003) PI controllers. The performance 

index is usually formulated in such way that takes into account the annualized process capital cost (based on 

steady state economics) and the operational cost. The effect of the dynamic performance was expressed by the 

variability cost (Ricardez-Sandoval et. al, 2008) which reflects to the possible losses during dynamic transitions 

by applying a penalty for economic losses due to non-optimum or not-on-specs performance usually expressed 

in terms of aggregate or average deviations. Sharifzadeh (2013) reported that the incorporation of the 

controllability measures into the process design as a multi-objective optimisation was helpful in obtaining design 

of superior control performance. 

3.2 Decision Level 
The first key aspect of differentiation of the proposed approaches is related to the extent of decisions that are 

included in the integrated design procedure of Figure 1. For example, the process synthesis step that also 

involves equipment and material selection is not usually incorporated due to the high combinatorial complexity 

it introduces. Similarly, the control algorithm is selected a priori based on guidelines that are derived from the 

nature of the process system and the facilitation of the calculations. However, the controller tuning parameters 

have been part of the optimisation problem (Mohideen et al., 1996). Recent challenges involve the incorporation 

of heat integration within the process and control system design framework, taking into account the flexibility 

and structural controllability in a heat exchanger network as presented by Hafizab et al. (2016). Patraşcu et al. 

(2017) dealt with the process dynamics utilizing an effective control structure of a proposed heat integrated 

design of a heat pump assisted extractive distillation process. An attempt in the field of the simultaneous material 

selection and process design has been described by Papadopoulos et al. (2017). They have developed a 

systematic methodology addressing the integrated computer-aided molecular and process design (CAMPD) 

under variability by considering a static operability analysis without however considering explicitly the process 

dynamics or the control system. 

3.3 Description Level 
The selection of the appropriate process model and the characterization of the process uncertainties and 

disturbances are two features that distinguish the proposed integrated design techniques. While the most 

accurate approach would be to utilize a non-linear dynamic process model, a comparative increase in 

computational effort is inevitable. Most methodologies utilize linear dynamic or non-linear steady-state models 

if the degree of nonlinearity in the actual process is moderate. Hamid et al. (2010) presented a systematic model-

based methodology for the integration of process design and controller design. The primal problem is 

decomposed into four sub-problems, namely: pre-analysis, design analysis, controller design analysis and final 

selection and verification. An extended version of this methodology on binary element reactive distillation 

processes is presented by Mansouri et al. (2016). Similarly, the selection of the type of the control algorithm to 

be implemented and the technique used for the pairing of manipulated and controlled variables in the case of 

decentralized controllers, can significantly affect the complexity of the mathematical model and its solution. 

Conventional PID controllers in multi-loop schemes (Ricardez Sandoval et. al, 2008) or perfect control, where 

controlled variables are kept constant with manipulated variables adjusted accordingly (Kookos and Perkins, 

2003) are usually implemented. In some cases, advanced model-based control schemes, like model predictive 

control (Sanchez-Sanchez and Ricardez-Sandoval, 2013) improved the achieved process economics and 

controllability. A comparison between the utilization of PI and MPC control scheme is performed under an 

integrated procedure by Sanchez-Sanchez and Ricardez-Sandoval (2013) and Gutierrez et al. (2014). The main 

drawback when advanced control schemes are used is that solving a dynamic optimisation problem requires an 

increased computational effort. 

Another feature related to the decomposition of the proposed framework is the way that these frameworks deal 

with the uncertainty of several model parameters and the disturbances that are taken into account. Ricardez-

Sandoval et al. (2008) attempted to implement uncertainty in the process model parameters, in which linearized 

dynamic process models are utilized. Even though, the process configuration as well as the control algorithm 

was specified a priori, the optimisation framework was formulated in such way that different process models 

(e.g., finite impulse response) and alternative control algorithms could be used. In addition, stochastic 

approaches, such as worst-case variability (Ricardez-Sandoval et al., 2012), were implemented in order to study 

the behaviour of the system to the time dependent disturbance profile that produced the largest variability in the 

system. The flowsheet synthesis and control structure selection was then performed using Monte Carlo 

sampling techniques to characterize the worst-case variability scenario. An approach of analysing the sensitivity 

1420



of the system dynamics through eigenvalue and transmission zeros tracking with respect to variations in 

operating conditions, design changes and disturbances was presented by Seferlis (2010). 

3.4 Solution Level 
A third key feature of the proposed approaches is related to the chosen solution procedure of the decomposed 

procedure. The main categories can be identified towards this direction, with the first category assess 

controllability analysis (e.g. RGA and condition number) during the design of the process and utilizing it as a 

decision criterion for the identification of a more meaningful and tractable design space. Furthermore, the 

formulation of the objective function can distinguish one approach from another, where indices that measure 

controllability (like interaction and error metrics) are used. Such methods, although consider the dynamic 

performance of the system, use arbitrarily selected weights in the multi-objective function to scale controllability 

and capital or/and operational cost that are usually not equivalent to the effect on actual economics of the 

process. A back-off methodology was presented by Rafiei-Shishavan et al. (2017) where power series 

expansions is employed in order to represent the cost function and the constraints of the optimisation problem. 

Malcom et al. (2007) presented a procedure that finds the optimal trade-offs between design and control 

decisions, based on a two-stage decomposed problem that is reduced on terms of size and complexity by 

utilizing a state-space model. Fuentes-Cortes et al. (2016) presented a multi-objective mixed-integer nonlinear 

optimisation approach for the optimal design and control of combined cooling, heat and power systems. Nordin 

et al. (2015) presented a methodology where sustainability criteria are taken into account leading to cost efficient 

and controllable designs, in a constrained optimisation problem, decomposed in six sequential hierarchical sub-

problems. 

The second category is consisted of methods that incorporate simultaneous process design and control system 

design aspects into a single holistic model-based optimisation problem and employ a successive execution of 

steps for the determination of the optimal solution. To deal with the large computational effort, Bansal et al. 

(2000) applied dual information and GBD algorithm, whereas full discretization of the dynamic system based on 

orthogonal collocation was also applied. Sakizlis et al. (2003) and Khajuria and Pistikopoulos (2011) 

incorporated the concept of parametric programming and extended this method by including multi-parametric 

model predictive controllers. 

4. Conclusions - Challenges 

Generally, a truly generalized approach that incorporates every aspect of the integrated process and control 

system design would require the solution of an extremely complex problem where the computational burden 

would become cumbersome. In order to obtain a meaningful and realizable solution the integrated problem is 

required to be decomposed into successive and interacting stages where the decision criteria are applied 

gradually using models and uncertainty characterization techniques of various resolution in order to simplify the 

calculations but without compromising the validity of the results. Based on the resulting sub-problem an 

appropriate solution procedure is employed. 

Obviously, significant effort should be invested in defining a reasonable design space for the process system. 

General engineering knowledge, process design heuristics, and hierarchical decomposition of the design 

decision procedure should be utilized to ensure that the design space encompasses the best performing 

solutions but efficiently eliminates options of definite inferior performance. Similarly, in the design of the control 

system, even though controllability is basically considered as an inherent property of the process, the selection 

of the control algorithm and its employment in the process system plays a significant role. Basically, the control 

system would enable a more compact process system by transferring process variability from the economically 

important process variables (e.g., product quality) to the less important variables (e.g., auxiliary utilities). The 

interaction of the control algorithm with the achieved process design is still quite unexplored. The identification 

of the optimal model reduction that would allow an accurate process and control system representation but with 

the least computational effort is a task that requires to become more systematic. 

The explicit inclusion of the control system performance in the design of heat and mass integrated process 

system has started to attract the attention of the research community. Moreover, the integration of materials 

selection in process and control system design is a new challenge with very few contributions that attempt to 

investigate the effect of catalysts, reactants, solvents and working media in the overall closed loop system 

performance, appears as one of the new fields with great challenges to be addressed. 

Acknowledgments  

This research has received funding from the State Scholarship Foundation (IKY), and especially under the 

framework of the program: “Fellowships of Excellence for Research Programs IKY/SIEMENS». 

1421



References 

Bansal V., Perkins J.D., Pistikopoulos E.N. Ross R., Van Schijndel J.M.G., 2000, Simultaneous design and 

control optimisation under uncertainty, Comput. & Chem. Eng., 24(2) 261-266. 

Fuentes-Cortes L.F., Dowling A.W., Rubio-Maya C., Zavala V.M., Ponce-Ortega J.M., 2016, Integrated design 

and control of miltigeneration systems for building complexes, Energy, 116, 1403-1416. 

Gutierrez G., Ricardez-Sandoval L.A., Budman H.M., Prada C., 2014, An MPC-based control structure selection 

approach for simultaneous process and control design, Comput. & Chem. Eng., 70, 11-21. 

Hafizan A.M., Alwi S.R.W., Manan Z.A., Klemeš J.J., 2016, Optimal Heat Exchanger Network Synthesis with 

Operability and Safety Considerations, Clean Technologies and Environmental Policy, 18(8), 2381–2400 

Hamid M.K.A., Sin G., Gani R., 2010, Integration of process design and controller design for chemical processes 

using model-based methodology, Comput. & Chem. Eng., 34(5), 683-699. 

Heath L.A., Kookos I.K., Perkins J.D., 2000, Process control structure selection based on economics, AIChE J., 

46(10), 1998-2016. 

Khajuria H., Pistikopoulos E.N., 2011, Dynamic modeling and explicit/multi-parametric MPC control of pressure 

swing adsorption systems, J. of Process Control, 21, 151-163. 

Kookos I.K., Perkins J. D., 2003, Regulatory control structure selection of linear systems , Comput. & Chem. 

Eng., 26, 875-887. 

Kyriakides A.-S., Seferlis P., Voutetakis S., Papadopoulou S., 2017, Investigating the simultaneous process 

design and control of membrane reactor for hydrogen production via methane steam reforming, Chemical 

Engineering Transactions, 61,  

Luyben M.L., Floudas C.A., 1994. Analysing the Interaction of Design and Control – 1. A Multiobjective 

Framework and Applixation to Binary Distillation Synthesis, Comput. and Chem. Eng., 18-10, 933–969. 

Malcom A., Polan J., Zhang L., Ogunnaike B.A., Linninger A.A., 2007, Integrating Systems Design and Control 

using Dynamic Flexibility Analysis, AIChE J., 53, 2048-2061. 

Mansouri S.S., Huusom J.K., Gani R., 2016, Systematic Integrated Process Design and Control of Binary 

Element Reactive Distillation Processes, AIChE J., 62, 3137-3154. 

Mohideen M.J., Perkins J.D., Pistikopoulos E.N, 1996, Optimal design of dynamic systems under uncertainty, 

AIChE J., 42(8), 2251-2272. 

Narraway L.T., Perkins J.D., 1993, Selection of process control structure based on linear dynamic economics, 

Industrial & Engineering Chemistry Research, 32(11), 2681-2692. 

Nordin M.Z., Othman M.R., Hamid M.K.A., 2015, Application of sustainable integrated process design and 

control for four distillation column systems, Chemical Engineering Transactions, 45, 211-216. 

Papadopoulos A.I., Seferlis P., Linke P., 2017, A framework for the integration of solvent and process design 

with controllability assessment, Chemical Engineering Science, 159, 154-176. 

Patraşcu I., Bildea C.S., Kiss A., 2017, Dynamics and control of a heat pump assisted extractive dividing-wall 

column for bioethanol dehydration, Chemical Engineering Research and Design, 119, 66–74. 

Rafiei-Shashavan M., Mehta S., Ricardez-Sandoval L.A., 2017, Simultaneous design and control under 

uncertainty: A back-off approach using power series expansions, Comput. & Chem. Eng., 99, 66-81. 

Ricardez-Sandoval L.A., 2012, Optimal design and control of dynamic systems under uncertainty: A probabilistic 

approach, Computers and Chemical Engineering, 43, 91-107. 

Ricardez-Sandoval L.A., Budman H.M., Douglas P.L., 2008, Simultaneous design and control of process under 

uncertainty: A Robust model approach, J. of Process Control, 18, 735-752. 

Sakizlis V., Perkins J.D., Pistikopoulos E.N., 2003, Parametric controllers in simultaneous process and control 

design optimisation, Industrial & Engineering Chemistry Research, 42(20), 4545-4563. 

Sanchez-Sanchez K.B., Ricardez-Sandoval L.A., 2013, Simultaneous Design and Control under Uncertainty 

Using Model Predictive Control, Industrial & Engineering Chemistry Research, 52 (13), 4815–4833. 

Seferlis P., 2010, System poles and zeros sensitivity for dynamic process controllability, American Control 

Conference, 1868–1873. 

Seferlis P., Georgiadis M.C., 2004, The Integration of Process Design and Control, Comput. Aided Chem. Eng., 

17, 1-10. 

Sharifzadeh M., 2013, Integration of process design and control: A Review, Chemical Engineering Research 

and Design, 91, 2515-2549. 

Skogestad S., Postlethwaite I., 1996, Multivariable Feedback Control Analysis and Design, U.K., J. Wiley. 

Swaney R.E., Grossmann I.E., 1985. An Index for Operational Flexibility in Chemical Process Design, Part I: 

Formulation and Theory, AIChE J., 31-4, 621–630. 

1422