# Port-Hamiltonian modeling of non-isothermal chemical reaction networks

- 717 Downloads

## Abstract

Motivated by recent progress on the port-Hamiltonian formulation of isothermal chemical reaction networks and of the continuous stirred tank reactor, the present paper aims to develop a port-Hamiltonian formulation of chemical reaction networks in the non-isothermal case, and to exploit this for equilibrium and stability analysis.

## Keywords

Chemical reaction networks Port-Hamiltonian systems Network dynamics Irreversible thermodynamic systems## 1 Introduction

Modeling of chemical reaction networks has attracted much attention in the last decades due to its wide application in systems biology and chemical engineering. Previous work, such as [8, 14, 15], provides the foundation of a structural theory of isothermal chemical reaction networks governed by *mass action kinetics*. From then on, a series of papers about the modeling and analysis of mass action kinetics chemical reaction networks appeared (see [1, 16, 29, 35]). In most of these papers, the chemical reaction networks are assumed to take place under *isothermal* condition. Consequently, the influence of in/outflow of heat can not be taken into account. Hence, non-isothermal chemical reaction networks still pose fundamental challenges.

In this paper, we aim to use the port-Hamiltonian framework for the modeling of non-isothermal mass action kinetics chemical reaction networks. Port-Hamiltonian systems theory (PHS) has been intensively employed in the modeling and passivity-based control of electrical, mechanical and electromechanical systems (see [20, 30, 31]).

In [32, 33], a port-Hamiltonian formulation of isothermal mass action kinetics chemical reaction networks was provided. A first step to non-isothermal chemical reaction networks was taken in [36]. Here, based on the previous works [5, 7, 24, 25], a quasi port-Hamiltonian formulation for non-isothermal chemical reaction networks was developed.

The main contributions of the present paper are as follows. First, based on mass and energy balance equations, a port-Hamiltonian formulation for non-isothermal mass action kinetics chemical reaction networks which are detailed-balanced is developed. This formulation directly extends the port-Hamiltonian formulation of isothermal chemical reaction networks of [32, 33], in contrast with the quasi port-Hamiltonian formulation in [36]. It exhibits the energy balance and the thermodynamic principles in an explicit way. Based on the obtained port-Hamiltonian formulation, we provide a thermodynamic analysis of the existence and characterization of thermodynamic equilibria and their asymptotic stability. Being directly related with the energy and entropy functions, this port-Hamiltonian formulation is easily applicable to chemical and biological systems. The second contribution of this paper is the extension of the port-Hamiltonian formulation and the thermodynamic analysis to non-isothermal chemical reaction networks with external ports.

The structure of the paper is as follows. In Sect. 2, some notation will be introduced which will be used in the remainder of the paper. Section 3 surveys the main elements of non-isothermal chemical reaction networks. Section 4 develops the port-Hamiltonian formulation of non-isothermal chemical reaction networks, and shows how this formulation is in line with the main laws of thermodynamics. In Sect. 5, a thermodynamic analysis will be carried out, including the characterization of equilibria and their asymptotic stability. In Sect. 6, an example—a genetic protein synthesis circuit with internal feedback and cell-to-cell communication—is discussed as an illustration of the developed theory. Section 7 extends the previous results to non-isothermal chemical reaction networks with external ports.

## 2 Notation

\(\mathbb {R}^m\) denotes the space of *m*-dimensional real vectors, and \(\mathbb {R}^m_+\) the space of *m*-dimensional real vectors whose entries are all strictly positive. The element-wise natural logarithm \(\mathrm {Ln}:\mathbb {R}^m_+\rightarrow \mathbb {R}^m\), \(x\mapsto \mathrm {Ln}(x)\), is defined as the mapping whose *i*th component is given as \((\mathrm {Ln}(x))_i := \ln (x_i)\). Similarly, \(\mathrm {Exp}:\mathbb {R}^m_+\rightarrow \mathbb {R}^m\), \(x\mapsto \mathrm {Exp}(x)\), is the mapping whose *i*th component is given as \((\mathrm {Exp}(x))_i := \exp (x_i)\). Note that \(\mathrm {Exp}(x+z)=\mathrm {Exp}(x)\mathrm {Exp}(z)\), \(\mathrm {Ln}(xz)=\mathrm {Ln}(x)+\mathrm {Ln}(z)\), and \(\mathrm {Ln}(\frac{x}{z})=\mathrm {Ln}(x)-\mathrm {Ln}(z)\), where \(xz\in \mathbb {R}^m\) is the element-wise product \((xz)_i:=x_iz_i\), \(i=1,\dots ,m\), and \(\frac{x}{z}\in \mathbb {R}^m\) is the element-wise quotient \((\frac{x}{z})_i=\frac{x_i}{z_i}\), \(i=1,\dots ,m\). Also, we define the mapping \(\mathrm {Diag}:\mathbb {R}^m\rightarrow \mathbb {R}^{m\times m}\), \(v\mapsto \mathrm {Diag}(v)\), where \(\mathrm {Diag}(v)\) is the diagonal matrix with \((\mathrm {Diag}(v))_{ii}=v_i\). Finally, the \(i\times i\) identity matrix is denoted as \(\mathrm {I}_i\), the \(i\times i\) zero matrix is denoted as \(\mathbf {0}_{i\times i}\), while the \(i\times 1\) zero vector is denoted as \(\mathbf {0}_i\). The notation \(z^{tr}\) is used to denote the transpose of the vector *z*.

## 3 The chemical reaction network structure

*r*reactions,

*m*species and

*c*complexes, given by the following reversible reaction scheme:

*stoichiometric coefficients*. The graph-theoretic formulation, according to [9, 11, 15], is to consider the chemical complexes defined by the left-hand and the right-hand sides of the chemical reactions, and to associate to each complex a vertex of a graph, while each reaction from left-hand to right-hand complex corresponds to a directed edge.

The concentrations of the species are denoted as \(x_i\), \(i=1,\dots ,m\), and the total vector of concentrations is denoted as \(x=[x_1,\dots ,x_m]^{tr}\). In order to capture the basic conservation laws of the chemical reactions, we define an \(m\times r\) matrix *C*, known as the *stoichiometric matrix*, whose (*i*, *j*)th element is the signed stoichiometric coefficient of the *i*th species in the *j*th reaction. Similarly, to define the connection between the complexes of each chemical reaction, we define an \(m\times c\) matrix *Z*, called the *complex stoichiometric matrix*, whose \(\rho \)th column captures the expression of the \(\rho \)th complex in the *i*th chemical species. Any directed graph is characterized by an \(c\times r\) matrix *B*, called the *incidence matrix*, whose (*i*, *j*)th element equals to \(-1\) if vertex *i* is the tail vertex of edge *j* and 1 if vertex *i* is the head vertex of edge *j* , while 0 otherwise. The relation between these three matrices for the graph of complexes is \(C=ZB\).

*v*is the r-dimensional vector of reaction rates, whose

*j*th element represents the

*j*th reaction rate of the chemical reaction network. Each reaction is considered to be a combination of a forward reaction with forward rate equation and a reverse reaction with reverse rate equation, both given by mass action kinetics. Thus, the reaction rate of the

*j*th reaction can be written as

*Z*. Here, \(k_{j}^{f}(T)\), \(k_{j}^{b}(T)\) are the forward/backward rate reaction coefficients of the

*j*th reaction defined by the

*Arrhenius equation*([4])

*activation energies*, \(k_j^f\) and \(k_j^b\) the forward and backward rate constants, and

*R*is the

*Boltzmann constant*. Using the element-wise natural logarithm \(\mathrm {Ln}(x)\) defined in Sect. 2, the

*j*th element of rate equation vector can be equivalently written as

*Z*corresponding to the substrate complex \(S_j\) and the product complex \(P_j\) of the

*j*th reaction.

## 4 The standard form of balanced mass action and balanced energy action for non-isothermal chemical reaction networks

In this paper, we will focus on *detailed balanced* mass action kinetics chemical reaction networks (see [10, 28, 34]). First, we assume that in the chemical reaction network, the equilibration following any reaction event is much faster that any reaction time scale. Thus, all intensive thermodynamic variables are well defined and equal everywhere in the system. Then, we assume that the chemical reaction network is closed and undergoes an adiabatic process. That means there is no heat or mass transfer between the system and external environment. Moreover, the chemical reaction network is isochoric so that the volume change can be neglected, i.e. \(dV=0\).

The definition of thermodynamic equilibrium for isothermal chemical reaction networks, see e.g. [32], is extended to non-isothermal networks as follows.

### Definition 4.1

A vector of concentrations \(x^*\) is called an *equilibrium* for the dynamics \(\dot{x}=Cv(x,T)\) for a certain temperature *T* if \(Cv(x^*,T)=0\), and a *thermodynamic equilibrium* if \(v(x^*,T)=0\). A chemical reaction network \(\dot{x}=Cv(x,T)\) is called detailed-balanced if it admits a thermodynamic equilibrium for every temperature *T*.

In order to stress the dependence on *T*, the thermodynamical equilibrium will be denoted by \(x^*(T)\). The conditions for existence of a thermodynamic equilibrium will be discussed in Sect. 5.1. Throughout this section we assume that there exists at least one thermodynamic equilibrium, like in the isothermal case, see e.g. [32, 33]. We will use the existence of this thermodynamic equilibrium to develop a port-Hamiltonian formulation.

### 4.1 Mass balance equations

*T*, i.e.,

*conductance*\(\kappa _j(T)\) of the

*j*th reaction as:

*j*th chemical reaction (6) can be rewritten as

*K*(

*T*) as

*B*defined in Sect. 3, the rate vector of a detailed balanced non-isothermal reaction network can be written as

*chemical potentials*and \(L:=BK(T)B^{tr}\) is the weighted Laplacian matrix for the reaction network graph, with weights given by the conductances \(\kappa _1(T),\dots ,\kappa _r(T)\).

*T*, but also on the choice of the thermodynamic equilibrium \(x^*(T)\). However, if the reaction network graph is connected, then for any other thermodynamical equilibrium \(x^{**}(T)\) for the same temperature

*T*, there exists a positive constant

*d*such that

*K*has been proved in [32]. It implies that the dependence on \(x^*(T)\) is minor; choosing another thermodynamical equilibrium only involves a uniform scaling of

*K*, and thus of

*L*. Another well-known property of

*L*is the fact that the matrix

*L*is independent of the orientation of the graph [2].

### 4.2 Energy balance equations

In this section we will express the energy conservation for a closed chemical reaction network in order to encompass the thermodynamic properties of the system.

*U*denotes the internal energy,

*S*the entropy, and the conjugated intensive variables are the chemical potential \(\frac{\partial U}{\partial x}=\mu \) and the temperature \(\frac{\partial U}{\partial S}=T\). This implies

### 4.3 Port-Hamiltonian formulation

*x*is the vector of concentrations and

*U*the internal energy. Then we define the Hamiltonian function \(H=-S\), where

*S*is the entropy. Note that the Gibbs’ relation (15) can also be written in the entropy formulation

*U*. This implies that the co-state vector corresponding to \(H=-S\) is

*T*can be expressed as function of the components of this co-state vector. Now define the skew-symmetric matrix

*R*is positive semi-definite. The formulation (23) is called ’quasi port-Hamiltonian’, since the structure matrices \(\mathcal {J}\) and \(\mathcal {R}\) depend on the co-state variables \(\frac{\partial H}{\partial z}=\frac{\partial (-S)}{\partial z}=\left[ \begin{array}{llll}\frac{\mu _{1}}{T}&\cdots&\frac{\mu _{m}}{T}&-\frac{1}{T}\end{array}\right] ^{tr}\), instead of on the state variables \(\left[ \begin{array}{llll}x_{1}&\cdots&x_{m}&U\end{array}\right] ^{tr}\) as in a standard port-Hamiltonian formulation. This formulation is comparable to the formulation of the mass balance and energy balance equations such as GENERIC, suggested in [18], or the port-Hamiltonian formulation with generating function being the availability function derived from the entropy function in [12].

### 4.4 Entropy balance equation

*S*

*L*) that for any \(\gamma \in \mathbb {R}^c\)

*S*is deduced from the port-Hamiltonian formulation (23) defined in Sect. 4.3. It is consistent with Eq. (19), which is deduced from the Gibbs’ relation.

*P*is set by the environment. It differs from the quasi port-Hamiltonian representation of the mass and entropy balance equations of chemical reaction networks in [26, 36], by the fact that it is based on the energy balance equation instead of on the entropy balance equation. Note that the description based on the energy balance equation is classical [6], and more easily derived than the description based on the entropy balance equation. Moreover, the quasi-port-Hamiltonian formulation given in (21), (22) and (23) fundamentally differs from the representation of chemical reaction networks as port Hamiltonian systems in [21] as well, by the fact that this quasi port-Hamiltonian representation is established on the whole space of concentration vectors instead of only locally around an equilibrium point, as in [21].

Finally the quasi port-Hamiltonian directly extends the port-Hamiltonian formulation of isothermal chemical reaction networks obtained in [32, 33] by including the energy balance equation.

## 5 Thermodynamic equilibria and asymptotic stability

The discussion in Sect. 4 is based on the assumption of existence of a thermodynamic equilibrium. Starting from the definition of a thermodynamic equilibrium of non-isothermal chemical reaction networks, we will derive in this section a full characterization of the set of equilibria, analogous to the case of isothermal chemical reaction networks in [32].

Subsequently, for stability analysis, we will use Lyapunov function as an availability function which is directly based on the quasi port-Hamiltonian representation given in (21), (22) and (23). Note that the use of availability functions for stability analysis is classical, see e.g. [12, 13, 19].

### 5.1 Thermodynamic equilibria

*T*, i.e., \(v(z^*)=0\). This implies that for any \(j=1,\dots ,r\),

*r*, and making use of the incidence matrix

*B*, we obtain the following condition for a thermodynamical equilibrium \(x^*(T)\)

*r*-dimensional vector with

*j*th element \(K_j^{eq}\), which is dependent on the temperature

*T*. Therefore, for a given temperature

*T*, there exists a thermodynamic equilibrium \(x^*(T)\in \mathbb {R}^m_+\) if and only if \(k_j^f>0\), \(k_j^b>0\) for all \(j=1,\dots ,r\), and

*T*. Then

*T*, once one thermodynamic equilibrium \(x^*(T)\) is given, the whole set of thermodynamic equilibria at the same temperature

*T*can be found. Furthermore, since \(dU=0\), we have \(U^*=U^{**}\). Denote \(z^{*}=(x^{*}(T), U^{*})\) and \(z^{**}(T)=(x^{**}(T), U^{**})\), then it follows that the set of thermodynamic equilibria at the same temperature

*T*can be written as

*T*, while the relation \(\mathrm {Exp}(C^{tr}\mathrm {Ln}(x^{**}))=\mathrm {Exp}(C^{tr}\mathrm {Ln}(x^*))\) is not dependent on temperature

*T*.

Since \(K^{eq}(T)=\mathrm {Exp}(C^{tr}\mathrm {Ln}(x^{*}))\) as a function of *T* is monotone and injective, it follows that the set of thermodynamic equilibria \({\varSigma }_{T_1}\) is disjoint from \({\varSigma }_{T_2}\), i.e., \({\varSigma }_{T_1}\cap {\varSigma }_{T_2}=\emptyset \) for any \(T_1\ne T_2\).

### 5.2 Asymptotic stability

For isothermal chemical reaction networks, it was shown in [15, 32, 33], that the Gibbs’ free energy can be used as a Lyapunov function for proving asymptotic stability towards a unique equilibrium depending on the initial condition. In this section we aim at proving a similar result for the non-isothermal case based on the port-Hamiltonian formulation obtained in the previous section, employing the availability function. Note that this is different from [24, 36], where an energy-based availability function was employed.

### Theorem 5.1

Consider a detailed balanced chemical reaction network given by (21), (22) and (23), with \(A:\mathbb {R}^{m+1}_+\rightarrow \mathbb {R}\) given by (31). Then *A* has a strict minimum at \(z^o\) with \(A(z^o)=0\), while the time-derivative \(\frac{dA}{dt}\) is less than or equal to zero with equality only at \(z^o\).

### Proof

*A*has a strict minimum at \(z^o\). Moreover, the time derivative of

*A*(

*z*) is given as

*L*is a balanced weighted Laplacian matrix, for any \(\gamma \in R^c\), we have \(\gamma ^{tr}L\mathrm {Exp}(\gamma )\ge 0\), while \(\gamma ^{tr}L\mathrm {Exp}(\gamma )=0\) if and only if \(B^{tr}\gamma =0\). Hence

*A*(

*z*) satisfies

### Theorem 5.2

Consider the detailed-balanced chemical reaction network (21), (22) and (23) with \(T\in \mathbb {R}_+\). Then for any \(x_1\in \mathbb {R}^{m+1}_+\), \(T_1\in \mathbb {R}_+\), there exists a unique \(x^*\in \mathbb {R}^{m+1}_+\) and \(T^*\in \mathbb {R}_+\), such that \(x^*-x_1\in \mathrm {im}\,C\), and \(z^*=(x^*,U^*(x^*,T^*))\in {\varSigma }_{T^*}\).

### Proof

Let \(W=\mathrm {im}\,C\). Then \(W^\perp =\mathrm {ker}\,C^{tr}\). Let \(z_1(x_1,T_1),z^{**}(x^{**},T^*)\in \mathbb {R}^{m+1}_+\), where \(z^{**}(x^{**},T^*)\in {\varSigma }_{T^*}\) is a thermodynamic equilibrium for temperature \(T^*\). As proved in [11, 32], there exists a unique \(\beta \in \mathrm {ker}\,C^{tr}\) such that \(x^{**}\mathrm {Exp}(\beta )-x_1\in \mathrm {im}\,C\). Define \(z^*(x^*,T^*)\in \mathbb {R}^{m+1}_+\) with \(x^*=x^{**}\mathrm {Exp}(\beta )\). Clearly, \(C^{tr}\beta =C^{tr}\mathrm {Ln}(\frac{x^*}{x^{**}})=0\), which is in line with (29) so that \(z^*(x^*,T^*)\in {\varSigma }_{T^*}\). Moreover, we have \(x^*-x_1=x^{**}\mathrm {Exp}\beta -x_1\in \mathrm {im}\,C\).

Combining with Theorem 4.1, this implies that the equilibrium \(z^*\) is asymptotically stable with respect to all initial conditions in near \(z^*\). Hence the asymptotic stability of the quasi port-Hamiltonian system defined by (21), (22) and (23) is proved. \(\square \)

## 6 Example: a genetic circuit with internal feedback and cell-to-cell communication

*E. Coli*in the large intestine of human beings [22]. When the cell of

*E.Coli*receives a ’message’ from the environment (a kind of transcription process from extracellular space into the

*E. Coli*cell), three chemical reactions will take place at the intercellular level:

*E. Coli*cell, with internal feedback and cell-to-cell communication.

### 6.1 Modeling

*K*becomes independent of

*T*and takes the form

*T*. Therefore, the port-Hamiltonian formulation (23) for the genetic protein synthesis circuit is

### 6.2 Equilibrium and Lyapunov function

*T*. We have \(A(z)=0\) at \(z=z^o\), and as discussed in Sect. 5.2, the time derivative of

*A*(

*z*) can be written as

*A*(

*z*) is a well-defined Lyapunov candidate. The port-Hamiltonian system (23) for the genetic protein synthesis circuit is asymptotically stable under temperature

*T*.

## 7 Non-isothermal chemical reaction networks with ports

In many application areas, the chemical reaction networks under consideration are not isolated. That is to say, there exist mass exchange or heat exchange between the chemical reaction network and its environment.

In this section we will extend the port-Hamiltonian formulation for non-isothermal chemical reaction networks to the case of mass and heat exchange. As in the previous work (see [12, 13, 24, 26]) when modeling and control of the Continuous Stirred Tank Reactor (CSTR), we define ‘external ports’ as the inflow/outflow of a mixture. Furthermore, we suppose that the output flow is such that the volume and pressure are constant [4].

*U*can be written as

*Q*is the heat flux from the environment, and \(F_{ei}\) and \(F_{si}\) are the

*i*th element of \(F_e\) and \(F_s\). Furthermore, \(h_{ei}\) and \(h_{si}\) are respectively the input and output specific enthalpies.

Note that in the port-Hamiltonian formulation for non-isothermal chemical reaction network with ports (38), we still use the thermodynamic equilibrium \(z^*(T)\) for the chemical reaction network without ports under given temperature *T*, as defined in Sect. 4.

*A*(

*z*) is a Lyapunov function. Obviously, we have \(A(z^{o})=0\).

*Z*, but with the different incidence matrix \(B'\) and different rate coefficients \(K'\).

*A*(

*z*) becomes

*T*. Let us illustrate this on the following example.

### 7.1 Example: a genetic circuit with internal feedback and cell-to-cell communication (continued)

*A*(

*z*) becomes

*A*(

*z*) is a well-defined Lyapunov candidate. We conclude that the port-Hamiltonian system for genetic protein synthesis circuit with a specific port \({\varDelta }F\) is asymptotically stable for temperature

*T*.

## 8 Conclusions and outlook

In this paper, a (quasi) port-Hamiltonian formulation has been developed for non-isothermal mass action kinetics chemical reaction networks. As an extensive result of the port-Hamiltonian formulation for isothermal chemical reaction network, and based on the mass balance and energy balance equations, this port-Hamiltonian formulation provides us a very explicit way to represent the chemical reaction networks and their thermodynamic properties, including the entropy balance and the conditions for the existence of thermodynamic equilibrium. As for the asymptotic stability, a comparable statement with the one in [27] has been found.

Moreover, this (quasi) port-Hamiltonian formulation and its thermodynamic analysis have been extended to non-isothermal chemical reaction networks with external ports. The results have been illustrated on a chemical reaction network in our body: the genetic circuit with internal feedback and cell-to-cell communication.

The focus of future work will be on the extension of current results to the modeling of interconnection of non-isothermal chemical reaction networks. Inspired by Rao and Esposito [27] and Qian and Beard [23], the interconnected port-Hamiltonian formulation will be considered as the combination of two *driven* (or *chemostatted*) chemical reaction networks with shared species.

## References

- 1.N. Balabanian, T.A. Bickart,
*Linear Network Theory: Analysis, Properties, Design and Synthesis*(Matrix Pub, York, 1981)Google Scholar - 2.B. Bollobas,
*Modern Graph Theory*, vol. 184 (Springer, New York, 1998)Google Scholar - 3.H. Callen,
*Thermodynamics*(Wiley, New York, 1960)Google Scholar - 4.F. Couenne, C. Jallut, B. Maschke, P. Breedveld, M. Tayakout, Bond graph modelling for chemical reactors. Math. Comput. Model. Dyn. Syst.
**12**(2), 159–174 (2006)CrossRefGoogle Scholar - 5.D. Eberard, B. Maschke, A.J. Van der Schaft, An extension of pseudo-Hamiltonian systems to the thermodynamic space: towards a geometry of non-equilibrium thermodynamics. Rep. Math. Phys.
**60**(2), 175–198 (2007)CrossRefGoogle Scholar - 6.A. Favache, D. Dochain, M. B, An entropy-based formulation of irreversible processes based on contact structures. Chem. Eng. Sci.
**65**, 5204–5216 (2010)CrossRefGoogle Scholar - 7.A. Favache, V. Dos Santos, B. Maschke, D. Dochain, Some properties of conservative control systems. IEEE Trans. Autom. Control
**54**(10), 2341–2351 (2009)CrossRefGoogle Scholar - 8.M. Feinberg, Complex balancing in general kinetic systems. Arch. Ration. Mech. Anal.
**49**(3), 187–194 (1972)CrossRefGoogle Scholar - 9.M. Feinberg, Chemical reaction network structure and the stability of complex isothermal reactors I. The deficiency zero and deficiency one theorems. Chem. Eng. Sci.
**42**(10), 2229–2268 (1987)CrossRefGoogle Scholar - 10.M. Feinberg, Necessary and sufficient conditions for detailed balancing in mass action systems of arbitrary complexity. Chem. Eng. Sci.
**44**(9), 1819–1827 (1989)CrossRefGoogle Scholar - 11.M. Feinberg, The existence and uniqueness of steady states for a class of chemical reaction networks. Arch. Ration. Mech. Anal.
**132**(4), 311–370 (1995)CrossRefGoogle Scholar - 12.H. Hoang, F. Couenne, C. Jallut, Y.L. Gorrec, The port Hamiltonian approach to modeling and control of continuous stirred tank reactors. J. Process Control
**21**(10), 1449–1458 (2011). https://doi.org/10.1016/j.jprocont.2011.06.014. (Special Issue: Selected Papers From Two Joint IFAC Conferences: 9th International Symposium on Dynamics and Control of Process Systems and the 11th International Symposium on Computer Applications in Biotechnology, Leuven, Belgium, July 5–9, 2010)CrossRefGoogle Scholar - 13.H. Hoang, F. Couenne, C. Jallut, Y.L. Gorrec, Lyapunov-based control of non isothermal continuous stirred tank reactors using irreversible thermodynamics. J. Process Control (2012). https://doi.org/10.1016/j.jprocont.2011.12.007 Google Scholar
- 14.F. Horn, Necessary and sufficient conditions for complex balancing in chemical kinetics. Arch. Ration. Mech. Anal.
**49**(3), 172–186 (1972)CrossRefGoogle Scholar - 15.F. Horn, R. Jackson, General mass action kinetics. Arch. Ration. Mech. Anal.
**47**(2), 81–116 (1972)CrossRefGoogle Scholar - 16.B. Jayawardhana, S. Rao, A.J. Van der Schaft, Balanced chemical reaction networks governed by general kinetics, in
*Proceedings of the 20th International Symposium on Mathematical Theory of Networks and Systems, Melbourne, Australia*(2012)Google Scholar - 17.K.R. Jillson, B.E. Ydstie, Process networks with decentralized inventory and flow control. J. Process Control
**17**(5), 399–413 (2007)CrossRefGoogle Scholar - 18.R. Jongschaap, H.C. Öttinger, The mathematical representation of driven thermodynamical systems. J. Non Newton. Fluid Mech.
**120**, 3–9 (2004)CrossRefGoogle Scholar - 19.J. Keenan, Availability and irreversibility in thermodynamics. Br. J. Appl. Phys.
**2**(7), 183 (1951)CrossRefGoogle Scholar - 20.B. Maschke, A.J. Van der Schaft, Port-controlled Hamiltonian systems: modelling origins and system theoretic properties, in
*Nonlinear Control Systems Design*, vol. 25 (1992), pp. 359–365Google Scholar - 21.I. Otero-Muras, G. Szederkényi, A. Alonso, K. Hangos, Local dissipative Hamiltonian description of reversible reaction networks. Syst. Control Lett.
**57**(7), 554–560 (2008)CrossRefGoogle Scholar - 22.E. Pico-Marco, Y. Boada, J. Pico, A. Vignoni, Contractivity of a genetic circuit with internal feedback and cell-to-cell communication. IFAC PapersOnLine
**49**(26), 213–218 (2016)CrossRefGoogle Scholar - 23.H. Qian, D.A. Beard, Thermodynamics of stoichiometric biochemical networks in living systems far from equilibrium. Biophys. Chem.
**114**(2), 213–220 (2005)CrossRefGoogle Scholar - 24.H. Ramırez, Y. Le Gorrec, B. Maschke, F. Couenne, Passivity based control of irreversible port Hamiltonian systems. IFAC Proc. Vol.
**46**(14), 84–89 (2013)CrossRefGoogle Scholar - 25.H. Ramirez, B. Maschke, D. Sbarbaro, Irreversible port-Hamiltonian systems: a general formulation of irreversible processes with application to the CSTR. Chem. Eng. Sci.
**89**, 223–234 (2013)CrossRefGoogle Scholar - 26.H. Ramırez, D. Sbárbaro, B. Maschke, Irreversible port-Hamiltonian formulation of chemical reaction networks, in
*21st International Symposium on Mathematical Theory of Networks and Systems*, Groningen, The Netherlands, 7–11 July 2014Google Scholar - 27.R. Rao, M. Esposito, Nonequilibrium thermodynamics of chemical reaction networks: wisdom from stochastic thermodynamics. Phys. Rev.
**6**(4), 041064 (2016)CrossRefGoogle Scholar - 28.S. Rao, A.J. Van der Schaft, B. Jayawardhana, A graph-theoretical approach for the analysis and model reduction of complex-balanced chemical reaction networks. J. Math. Chem.
**51**(9), 2401–2422 (2013)CrossRefGoogle Scholar - 29.S. Rao, A.J. Van der Schaft, K. Van Eunen, B.M. Bakker, B. Jayawardhana, A model reduction method for biochemical reaction networks. BMC Syst. Biol.
**8**(1), 1 (2014)CrossRefGoogle Scholar - 30.A.J. Van der Schaft, Port-Hamiltonian systems: an introductory survey, in
*Proceedings of the International Congress of Mathematicians*, vol. III (European Mathematical Society Publishing House, 2006), pp. 1339–1365Google Scholar - 31.A.J. Van der Schaft, B. Maschke, The Hamiltonian formulation of energy conserving physical systems with external ports. Arch. Elektron. Übertrag.
**49**(5–6), 362–371 (1995)Google Scholar - 32.A.J. Van der Schaft, S. Rao, B. Jayawardhana, On the mathematical structure of balanced chemical reaction networks governed by mass action kinetics. SIAM J. Appl. Math.
**73**(2), 953–973 (2013)CrossRefGoogle Scholar - 33.A.J. van der Schaft, S. Rao, B. Jayawardhana, On the network thermodynamics of mass action chemical reaction networks. IFAC Proc. Vol.
**46**(14), 24–29 (2013)CrossRefGoogle Scholar - 34.O.N. Temkin, A.V. Zeigarnik, D.G. Bonchev,
*Chemical Reaction Networks: A Graph-Theoretical Approach*(CRC Press, Boca Raton, 1996)Google Scholar - 35.A.B.O. Varma, Palsson, metabolic flux balancing: basic concepts, scientific and practical use. Biotechnology
**12**, 994–998 (1994)CrossRefGoogle Scholar - 36.L. Wang, B. Maschke, A.J. Van der Schaft, Irreversible port-Hamiltonian approach to modeling and analyzing of non-isothermal chemical reaction networks, in
*6th IFAC Conference on Foundations of Systems Biology in Engineering (FOSBE)*(2016)Google Scholar

## Copyright information

**Open Access**This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.