Learning physicsbased reducedorder models for a singleinjector combustion process
Abstract
This paper presents a physicsbased datadriven method to learn predictive reducedorder models (ROMs) from highfidelity simulations, and illustrates it in the challenging context of a singleinjector combustion process. The method combines the perspectives of model reduction and machine learning. Model reduction brings in the physics of the problem, constraining the ROM predictions to lie on a subspace defined by the governing equations. This is achieved by defining the ROM in proper orthogonal decomposition (POD) coordinates, which embed the rich physics information contained in solution snapshots of a highfidelity computational fluid dynamics (CFD) model. The machine learning perspective brings the flexibility to use transformed physical variables to define the POD basis. This is in contrast to traditional model reduction approaches that are constrained to use the physical variables of the highfidelity code. Combining the two perspectives, the approach identifies a set of transformed physical variables that expose quadratic structure in the combustion governing equations and learns a quadratic ROM from transformed snapshot data. This learning does not require access to the highfidelity model implementation. Numerical experiments show that the ROM accurately predicts temperature, pressure, velocity, species concentrations, and the limitcycle amplitude, with speedups of more than five orders of magnitude over highfidelity models.
ROMpredicted pressure traces accurately match the phase of the pressure signal and yield good approximations of the limitcycle amplitude.
Nomenclature
=  System matrix for linear part 
=  Input matrix 
=  Matricized quadratic tensor 
=  State vector in finite dimensions 
=  External input vector 
=  Snapshot matrix 
=  Matrix of POD basis vectors 
=  Spatial coordinates 
=  Number of physical variables 
=  Spatial discretization dimension 
=  Reduced model dimension 
=  Time 
=  State vector in primitive, conservative, and learning variables 
=  Pressure, continuous and discretized 
=  Temperature, continuous and discretized 
=  Density, continuous and discretized 
=  Specific volume, continuous and discretized 
=  Velocity in direction, continuous and discretized 
=  Species mass fraction, 
=  Species molar concentrations, , also denoted as for chemical species 
=  Kronecker product 
=  Notation for ROM quantities 
Abbreviations  
CFD =  Computational Fluid Dynamics 
GEMS =  General equation and mesh solver; a CFD code 
PDE =  Partial differential equation 
POD =  Proper orthogonal decomposition 
ROM =  Reducedorder model 
1 Introduction
This paper presents an approach to learning lowdimensional surrogate models for a complex, nonlinear, multiphysics, multiscale problem in form of a multispecies combustion process. The need for repeated model evaluations in optimization, design, uncertainty quantification and control of aerospace systems has driven the development of reducedorder models (ROMs) for applications in aerodynamics [1, 2, 3, 4, 5, 6, 7], reacting flows [8, 9, 10] and combustion [11, 12, 13]. ROMs combine the rich information embedded in highfidelity simulations with the efficiency of lowdimensional surrogate models; yet, effective and robust ROM methods for nonlinear, multiscale applications such as combustion have remained an open challenge.
Most existing nonlinear model reduction methods are intrusive—that is, they derive the ROM by projecting the highfidelity model operators onto a lowdimensional subspace. In doing so, the physics of the problem is embedded in the reducedorder representation. The proper orthogonal decomposition (POD) [14, 15] is the most common way to define the lowdimensional subspace, using the singular value decomposition to identify lowdimensional structure based on training data. For some problems, the projection approach is amenable to rigorous error analysis and structurepreservation guarantees [16, 17, 18, 19], but these rigorous guarantees do not apply to nonlinear, multiphysics, multiscale models, for which projectionbased ROMs remain challenging to implement (due to the need for access to the highdimensional operators). Furthermore, ROMs for these problems typically require relatively high dimensionality (and thus high cost) to avoid problems with robustness and stability [12, 20]
There is increasing attention to nonintrusive model reduction methods (sometimes called blackbox or datadriven methods) that learn a model based on training data, without requiring explicit access to the highfidelity model operators. The nonintrusive philosophy aligns directly with the field of machine learning, where representations such as neural networks have been shown to induce nonlinear model forms that can approximate many physical processes [21]. However, neural networks require a large amount of training data, limiting their utility when the data comes from expensive largescale partial differential equation (PDE) simulations [22]. Moreover, the parametrization of the learned lowdimensional model is critical to the predictive accuracy and success of the learned model—in particular, it is critical to determining whether the ROM can issue reliable predictions in regimes outside of the training data.
For largescale PDE models, an important class of nonintrusive learning approaches tackle this challenge of model parametrization by embedding the structure of the problem into the learning formulation. Some approaches use sparse learning techniques to identify PDE model terms that explain the data [23, 24, 25]. When the model can be expressed in the form of a dynamical system with polynomial terms, then the learning problem can be formulated as a parameter estimation problem, as in the operator inference approach of [26]. An important advantage of nonintrusive learning approaches is that the user has the flexibility to choose the variables that drive the learning. This opens the way for variable transformations that expose system structure and, in doing so, transform the ROM learning task into a structured form. In some cases, the governing PDEs naturally admit variable transformations that reveal polynomial form, such as the specific volume representation of the Euler equations [27]. More generally, one can introduce new auxiliary variables to the problem—known as lifting—to produce a system that is polynomial in its expanded set of state variables [28, 29, 30]. This allows for a much broader class of nonlinear systems to be learned using the operator inference framework.
In this work, we build on the operator inference framework to learn structured, polynomial ROMs from simulated snapshot data of a singleinjector combustion model. While in our case the polynomial model parametrization is a model approximation, we show that the predictive capabilities of the learned ROM are excellent beyond the training data. Our proposed approach follows the steps below, which we develop in detail in the ensuing sections:

We identify a set of state variables in which many of the terms in the governing equations have quadratic form. We transform the snapshot data to these new state variables, as described in Section 3.3.
We present numerical results comparing our learned ROMs with GEMS test data in Section 4 and conclude the paper in Section 5.
2 Combustion model
Section 2.1 defines the computational domain under consideration, Section 2.2 presents the governing equations for the combustion model, and Section 2.3 briefly summarizes the numerical implementation. The combustion model follows the implementation of the General Equation and Mesh Solver (GEMS) CFD code [31] and more details can be found in Huang et al. [11].
2.1 Computational domain
A singleinjector combustor as in [32] is shown in Figure 0(a), with the computational domain outlined in red dashed lines. Our domain is a simplified twodimensional version of the computational domain, shown in Figure 0(b), which also shows the four locations where we monitor the state variables.
2.2 Governing equations
The dynamics of the combustor are governed by the conservation equations for mass, momentum, energy and species mass fractions. For this twodimensional problem, the conservation equations are
(1) 
and they describe the evolution of the conservative variables
where is the density (), and are the and velocity (), is the total energy (), and is the th species mass fraction with and is defined as the number of chemical species that are included in the model.
The total energy is defined as
(2) 
where pressure is given in (Pa), is the enthalpy corresponding to the th species and is a highly nonlinear function of temperature, , and is the stagnation enthalpy. The inviscid flux and viscous flux in Eq. (1) are
The twodimensional viscous shear tensor is defined as
(3) 
where is the mixture viscosity coefficient. The diffusive heat flux vector is defined as
(4) 
where quantifies thermal conductivity and is the diffusion coefficient for the th species into the mixture, which is an approximation used to model the multicomponent diffusion as the binary diffusion of each species into a mixture. The two terms in the definition of the heat flux (Eq. (4)) represent heat transfer due to conductivity and species diffusion. The diffusive mass flux vector of species is modeled as
(5) 
The source term, in Eq. (1) is
(6) 
and is defined by considering a 1step combustion reaction governed by
as presented in [33], with . The corresponding general stoichiometric equation is defined as where , HO and is the net stoichiometric coefficients of each species with and . The molar concentration of the th species is denoted by . In our case, , so and are the molar concentrations. Here, we use the standard bracket notation to indicate molar concentration of a species. The general relationship between a species molar concentration, , and a species mass fraction, , is
(7) 
where is the mass fraction of CH, is the mass fraction of O, is the mass fraction of CO and is the mass fraction of HO. The production rate of the th species in the source term in Eq. 6 is modeled as
(8) 
where are chemical reaction source terms whose dynamics are described below and is the reaction rate. The molar mass of CH is , the molar mass of O is , the molar mass of CO is , and the molar mass of HO is . The reaction rate is approximated by
where is the number of reactants, is the rate coefficient and is the reaction order of the th reactant. In our case and . The rate coefficient, , is described by the Arrhenius equation as
(9) 
where is the universal gas constant, is the preexponential constant and is the energy required to reach a chemical reaction, measured in Joules and referred to as the activation energy. In this work, we use the ideal gas state equation that relates density and pressure to temperature
(10) 
where and is the average molar mass of the mixtures.
At the downstream end of the combustor, we impose a nonreflecting boundary condition while maintaining the chamber pressure via
(11) 
where Pa, and Hz. The top and bottom wall boundary conditions are noslip conditions, and for the upstream boundary we impose constant mass flow at the inlets.
2.3 Numerical model
GEMS uses the finite volume method to discretize the governing equations, which are solved in primitive variables. For a spatial discretization with cells, this results in a dimensional system of nonlinear ordinary differential equations (ODEs)
(12) 
for , where is the number of unknowns in the PDE governing equations and here (four flow variables and four species concentrations). In Eq. (12), is the discretized state vector at time (for GEMS, it is the discretization of the primitive variables ), are the specified initial conditions, and is the time derivative of the state vector at time . The inputs arise from the timedependent boundary condition, defined in Eq. (11), applied at the combustor downstream end. The nonlinear function maps the discretized states and the input to the state time derivatives, representing the spatial discretization of the governing equations described in Section 2.2.
Solving these highdimensional nonlinear ODEs is expensive, motivating the derivation of a ROM that can yield approximate solutions at reduced cost. The nonlinear multiscale dynamics represented by these equations makes this a challenging task. To maintain computational efficiency in the ROM, stateoftheart nonlinear model reduction methods combine POD with a sparse interpolation method (often called hyperreduction) by evaluating the nonlinear functions only at a select number of points. For instance, POD together with the discrete empirical interpolation method (DEIM) has had some success, but also encountered problems in combustion applications [12] . Of particular challenge is the need to include a large number of interpolation points in the PODDEIM approximation, which means that the ROM loses its computational efficiency. Robustness and stability of the PODDEIM models is also a challenge [20]. In the next section, we present a different approach that uses nonintrusive ROM learning to enable variable transformations that expose system structure. This structure is then exploited in the derivation of the ROM and removes the need for the DEIM approximation.
3 Nonintrusive learning of a combustion reduced model
This section presents our approach to learn ROMs for the unsteady combustion dynamics simulation from GEMS. Section 3.1 writes a general nonlinear system in a form that exposes the underlying structure of the governing equations and shows how projection preserves that structure. Section 3.2 presents the operator inference approach from [26], which learns structured ROM operators from simulation data. Section 3.3 describes variable transformations that lead to the desired polynomial structure for the combustion governing equations presented in Section 2. These transformations yield the structure needed to apply the operator inference approach.
3.1 Projection preserves polynomial structure in the governing equations
Consider a largescale sytem of nonlinear ODEs written in polynomial form
(13) 
Relating this equation to the general nonlinear system in Eq. (12), we see that are the terms in that are linear in the state , with ; are the terms in that are quadratic in , with ; are the terms in that are cubic in , with ; are the terms in that are linear in the input , with ; and are constant terms in that do not depend on state or input. The abbreviation “HOT” in Eq. (13) denotes higherorder terms, and represents terms that are quartic and higher order, as well as any other nonlinear terms that cannot be represented in polynomial form.
We emphasize that we are not (yet) introducing approximations—rather, we are explicitly writing out the discretized equations in the form (13) to expose the system structure that arises from the form of the terms in the governing PDEs. For example, a term such as in Eq. (1) is linear in the state , while a term such as is quadratic in the states and . Also note that the term is quadratic in the states and , highlighting the important point that the structure of the nonlinear model depends on the particular choice of state variables.
A projectionbased ROM of Eq. (13) preserves the polynomial structure. Approximating the highdimensional state in a lowdimensional basis , with , we write . Using a Galerkin projection, this yields the ROM of Eq. (13) as
(14) 
where , , , and are the ROM operators corresponding respectively to , , , and , and is a constant vector. We note again that projection preserves polynomial structure, that is, (14) has the same polynomial form as (13), but in the reduced subspace defined by .
In what follows, we will work with a quadratic system in order to simplify notation. We note that the least squares learning approach described below applies directly to cubic, quartic and all higherorder polynomial terms (although it should be noted that the number of elements in the ROM operators scales with for the cubic operator, for the quartic operator, etc.). For terms in the governing equations that are not in polynomial form (such as terms involving , and the Arrhenius reaction terms) we discuss in Section 3.3 the introduction of variable transformations and auxiliary variables via the process of lifting [28, 29] to convert these terms to polynomial form.
3.2 Operator inference for learning reduced models
Here we summarize the steps of the operator inference approach from [26]. First, we collect snapshots of the state by solving the highfidelity model. We store the snapshots and the inputs used to generate them in the matrices:
(15) 
where and with . In general, , so the matrix is tall and skinny. Second, we identify the lowdimensional subspace in which we will learn the ROM. In this work, we use the POD to define the lowdimensional subspace, by computing the singular value decomposition of the snapshot matrix
where , and . The dimensional POD basis, , is given by the first columns of . Third, we project the state snapshot data onto the POD subspace spanned by the columns of and obtain the reduced snapshot matrices
(16) 
where the columns of are computed from using any time derivative approximation (see, e.g., [34, 35, 36]), or can be obtained—if available—by collecting and projecting snapshots of .
Operator inference solves a least squares problem to find the reduced operators that yield the ROM that best matches the projected snapshot data in a minimum residual sense. For the quadratic ROM
(17) 
operator inference solves the least squares problem
(18) 
where is the length column vector with all entries set to unity. Note that this least squares problem is linear in the coefficients of the unknown ROM operators , , and . Also note that the operator inference approach permits us to compute the ROM operators , , and without needing explicit access to the original highdimensional operators , , and .
We combine the unknown operators of Eq. (17) in the matrix
and the known lowdimensional data in the data matrix
(19) 
and then solve the minimization problem
(20) 
It was proven in [26] that Eq. (20) can be written as independent least squares problems of the form , for , where is a column of (row of ) and is a column of . This makes the operator inference approach efficient and scalable.
Regularization becomes necessary to avoid overfitting and to infer operators that produce a stable ROM. In this work, we use an L regularization penalty on the offdiagonal elements of the operator and on all elements of the remaining operators. With this regularization, our least squares problem becomes
(21) 
where is the regularization parameter and is the identity matrix with the th diagonal set to zero so that we avoid regularizing the diagonal elements of . It should be noted that the regularization parameter, , is problem specific and should be chosen accordingly. In Section 4.2, we discuss details of the operator inference implementation, a method for selecting , and the removal of redundant terms in the least squares problem in Eq. (21).
3.3 A structureexploiting ROM learning formulation for GEMS
A key contribution of this work is to recognize that the nonintrusive operator inference approach gives us complete flexibility in the set of physical variables we work with to define the ROM. We can identify choices of physical variables that expose the desired polynomial structure in the governing equations, and then extract snapshots for those variables by applying transformations to the snapshot data—we do not need to make any modifications to the highfidelity CFD simulation model itself. In theory, a classical intrusive ROM approach could work with transformed variables (e.g., in the work of [37]); however, this would involve rewriting the highfidelity simulator, a task that would be not only timeconsuming but also fraught with mathematical pitfalls, especially for unusual choices of variables. This is where the datadriven perspective of machine learning becomes extremely valuable.
The Euler equations admit a quadratic representation in the specific volume variables; in that case, a transformation of the snapshots from conservative (or primitive) variables to specific volume variables can be exploited to create quadratic ROMs [27]. Other PDEs may not admit polynomial structure via such straightforward transformations, but the process of lifting the equations via the introduction of new auxiliary variables can produce a set of coordinates in which the governing equations become polynomial in the lifted state [28, 29, 30]. For example, the tubular reactor example of [29] includes Arrheniustype reaction terms similar to those in Eq. (9). The introduction of auxiliary variables permits the governing equations to be written equivalently with quartic nonlinearity in the lifted variables.
Lifting to polynomial form for the GEMS equations described in Section 2 is made difficult by several of the terms, in particular through some of the gas thermal properties such as the nonlinear dependence of enthalpy on temperature. A complete lifting that converts all equations to a polynomial form is possible, but would require the introduction of a large number of auxiliary variables and would also result in the introduction of some algebraic equations. However, as the analysis below shows, the GEMS governing equations admit a transformation for which many terms in the governing equations take polynomial form when we use the variables
(22) 
Here is the specific volume, and recall that and are the molar concentrations with .
Below, we derive the governing PDEs for specific volume and velocities . These three governing PDEs all turn out to be quadratic in the learning variables . In Appendix A we present the lifting transformations for the source term dynamics in the vector in Eq. (6). In Appendix B we derive the equations governing the pressure and the species molar concentrations . These equations have some terms that are not polynomial in the chosen learning variables .
To keep notation clean in application of the chain rule, let the conservative variables be denoted as . Throughout, we frequently use the relationship
(23) 
and similarly for . Note also that we are assuming the existence of these partial derivatives, that is, we do not consider the case of problems with discontinuities.
Specific Volume . We use the constitutive relationship for the density in Eq. (1) in the derivation:
(24) 
Inserting Eq. (23) into the above, we obtain
(25) 
which is quadratic in the learning variables .
Velocities : We have
(26) 
and from Eq. (1) we have that as well as . Thus, we obtain
(27)  
(28)  
(29)  
(30) 
and we get a similar expression for . Both dynamics are
quadratic in the learning variables .
As noted above, Appendix A and Appendix B present the derivations for the chemical source terms, pressure, and chemical species.
4 Numerical Results
We now apply the variable transformations and operator inference framework to learn a predictive ROM from GEMS highfidelity combustion simulation data.
4.1 GEMS Dataset
The computational domain shown in Figure 0(b) is discretized with spatial discretization points. Each CFD state solution thus has dimension . The problem considered here has fuel and oxidizer input streams with constant mass flow rates of and , respectively. The fuel is composed of gaseous methane and the oxidizer is 42% gaseous O and 58% gaseous HO, as described in [11]. The forcing input Eq. (11) is applied at the right side of the domain.
To generate training data, GEMS is simulated for a time duration of 1ms with a time step size of s resulting in snapshots. The GEMS output is transformed to the variables given in Eq. (22). The recorded snapshot matrix is thus
Our numerical experiments were parallelized on a cluster with two computing nodes. Each node has two 10core Intel XeonE5 processors (20 cores per node) and 128 GB RAM. The training data generation took approximately 200h in CPU time for the 1ms, 10000 snapshots of highfidelity CFD data.
The range of variable values for the training data is shown in Table 2. Note that the data covers a wide range of scales. Pressure is of the order while species concentrations can be as low as 10. This large scaling difference presents a challenge when learning models from data. To deal with the numerical issues related to large differences in scaling and small species concentrations and velocities, we scale each variable to the interval . Variables are scaled before computing the POD basis and projecting the data.
State variable  Minimum  Mean  Maximum 

Pressure in Pa  
Velocity in  222.930  69.637  307.147 
Velocity in  206.990  1.304  186.548 
Specific volume in  0.0533  0.220  0.674 
Molar concentration  0.0  0.063  1.169 
Molar concentration  0.0  0.056  0.097 
Molar concentration  0.0  0.002  0.012 
Molar concentration  0.0  0.154  0.232 
To obtain snapshots of the projected state time derivative, we approximate the derivative with a fivepoint approximation . This approximation is fourth order accurate. The first two and last two time derivatives are computed using firstorder forward and backward Euler approximations, respectively.
4.2 Learning a quadratic reducedorder model
To learn the operators of the quadratic ROM, we solve the regularized least squares problem shown in Eq. (21). In what follows next, we describe several important implementation details.
Singular value decomposition implementation and POD basis selection
Due to the large size of this dataset, we implement the randomized singular value decomposition algorithm, introduced in [39], to compute the leading 500 singular values and vectors of the snapshot matrix. The randomized singular value decomposition algorithm can be implemented in a scalable way for very large datasets. The POD basis is chosen as the leading left singular vectors. The dimension is typically chosen so that the cumulative energy contained in the subspace is greater than a user specified tolerance , i.e.,
where are the squared singular values of the data matrix . To guide the choice of we also use the relative projection error
(31) 
Removing redundant terms in least squares problem
Due to the redundant terms computed in in equation (19), the least squares problem may become illposed. Thus, the Kronecker product is replaced with the term
where . Each vector is defined, according to [26], as
and is the th element of the vector . Now, instead of learning the operator , the operator is learned, which satisfies the equivalent least squares problem
(32) 
The least squares problem is again of the form (20) but we use the data matrix and solve for the operators . Once we have solved for the operator we can easily transform it to obtain .
Regularization
We use an Lregularization (also known as Tikhonov regularization or ridge regression) to solve the operator inference problem, as shown in Eq. (32). The regularization term introduces a tradeoff between operators that fit the data well and operators with small values. This regularization is used to avoid overfitting to the data, which in this setting causes our learned ROMs to be unstable. The regularization parameter affects the performance of this algorithm—we require enough regularization to avoid overfitting, but if is too large, the data will be poorly fit.
To help determine appropriate values of , we consider the “Lcurve” criterion discussed in [40]. The Lcurve is a way of visualizing the effects of different values of on the norm of the residual (data fit) against the norm of the solution. The Lcurve criterion recommends choosing a value for that lies in the corner of the curve, nearest the origin. In our numerical experiments, we compute the Lcurve to help determine appropriate values for .
Regularization also helps to reduce the condition number of the regularized least squares data matrix, . In Figure 2 we show the condition number of the data matrix from Eq. (32) (with the data already scaled to [1,1]) when we include and snapshots in the training set. The condition number is quite large for this application, yet it decreases as we add more training data. This effect is due to the fact that as we add more training data, the first POD basis vectors become richer. This in turn means that the projected data in are richer for a given dimension of the POD basis. Our numerical experiments reinforced this finding by confirming that training snapshots were required to achieve a sufficiently rich POD basis to obtain accurate ROMs, an indication of the complexity of the combustor dynamics.
Error measures
To evaluate the ROM performance, we define appropriate error metrics for each variable. The error is computed after the ROM solutions have been reconstructed back into the full CFD model dimension and scaled back to the original variable ranges. Recall Table 2, which showed the range of values for each variable. Below we provide details on how error is computed for each variable:

For pressure and temperature, the values are always positive and well above zero, so we use a standard relative error, defined as
(33) where and denotes the th entry of a vector, and is the CFD solution variable and is the solution variable obtained from the ROM simulation.

Due to the small values of species concentrations (on the order of ), dividing by the true value can skew a small error. Similarly, velocities range from positive to negative, including zero. Thus, for species concentrations and velocities, we use a normalized absolute error, defined as
(34) where and denotes the maximum entry of , i.e., the maximum absolute value over the discretized spatial domain.
4.3 Learned reduced model performance
The given snapshots representing 1ms of GEMS simulation data are used to learn the ROM. We are also given another 1ms of testing data at the monitor locations shown in Figure 0(b), which we use to assess the predictive capabilities of our learned ROMs beyond the range of training data.
The cumulative energy of the singular values of the snapshot data matrix is shown in Figure 3. The singular values that correspond to a cumulative energy of 97.5% and 99% are indicated with the red triangle and green square, respectively. We use basis sizes of , capturing 97.5% of the total energy, and , capturing 99% of the total energy.
In Figures 3(a), 3(b) we show the Lcurve for each basis size. The Lcurve for a basis of is somewhat uninformative in this case. The regularization parameters chosen were E+04 and E+05, which lie in the middle of the Lcurve and are therefore a recommended choice by the Lcurve criterion discussed above. For a basis of size of , the Lcurve indicates a regularization parameter around E+04. Stable systems are produced for E+04 and E+04.
We simulate the learned ROM for the two model sizes and with the same initial value and time step size (s) as those used to generate the training set. Figures 5 and 6 compare the time trace of pressure computed by GEMS (our “truth” data) with the ROM predictions for all 20000 time steps at the cell located at in the domain (denoted as monitor location 1 in Figure 0(b)). The performance of the ROM on the training data (first 1ms of data) is good in both cases, although the ROM (Figure 6) is more accurate. For test data predictions beyond the training data, Figure 6 shows that a basis of size of yields accurate phase predictions and pressure oscillation amplitudes that are good approximations of the truth. We notice an improved amplitude prediction when using a smaller regularization value.
We also compute the average error of each variable over the entire domain at the last time step of the training set (the 10000th time step). The normalized absolute error, defined in Eq. (34), is shown for each species and for and velocity in Figure 7. This figure also shows the relative error, defined in Eq. (33), for pressure and temperature. These plots show that overall, the error is decreasing with an increasing basis size. At a basis size of 18, 22, and 24 the systems were unstable and so these basis sizes are excluded from the figure. The cause of this may be due to the fact that the same regularization parameter was used for each of these, E+04, and ideally one would pick a parameter specific for the basis size.
In Figure 8, we show the integrated species concentrations over time. To compute these, at each time step in our simulation, we sum all elements of a species vector. This measure monitors whether our ROM conserves species mass, an important feature of a physically meaningful simulation. As the discretization of the highfidelity model becomes finer, pointwise error may become large and misleading if the mass is shifted slightly into the neighboring cells. The integrated species concentration complements the evaluation of pointwise errors and provides a global view of the error in the domain.
We also compare the state variables over the entire domain predicted by the learned ROM with the given GEMS data at the last time step (which corresponds to s). We provide the true field, the ROMpredicted field, and an error field for each variable in Figures 9–16. Again, for pressure and temperature, we use a relative error from Eq. (33). For and velocity and for species molar concentrations we use a normalized absolute error from Eq. (34). The plots show that the ROM predictions are, as expected, not perfect, but indeed they capture well the overall structure and many details of the solution fields.
Table 3 shows timing results for the ROM generation and simulation, as performed using python 3.6.4 on a dualcore Intel i5 processor with 2.3 GHz and 16GB RAM. We report the following CPU times: solving the operator inference least squares problem (32); ROM runtime for the two different basis sizes for 2ms of real time prediction; and reconstruction of the highdimensional, unscaled combustion variables. The latter is required since the ROM is evolving the dynamics in the scaled variables, and thus a postprocessing step is required to obtain the true magnitudes of the variables. In comparison with the approximately 200h of CPU time needed on a 40core architecture (more details in Section 4.1) to compute the first 10000 snapshots of data, the ROMs provide five to six orders of magnitude in computational speedup.
ROM order  LS solve of (21)  ROM simulation  Reconstructing highdim. field 

0.97s  2.88s  0.03s  
6.05s  3.73s  0.05s 
5 Conclusion
Operator inference is a datadriven method for learning reducedorder models (ROMs) of dynamical systems with polynomial structure. This paper demonstrates how variable transformations can expose quadratic structure in the nonlinear system of partial differential equations describing a complex combustion model. This quadratic structure is preserved under projection, providing the mathematical justification for learning a quadratic ROM using operator inference. An important feature of the approach is that the learning of the ROM is nonintrusive—it requires state solutions generated by running the highfidelity combustion model, but it does not require access to the discretized operators of the governing equations. This is important because it means that the variable transformations can be applied as a postprocessing step to the simulation data set, but no intrusive modifications are needed to the highfidelity code. While the quadratic model form is an approximation for this particular application problem, the numerical results show that the learned quadratic ROM can predict relevant quantities of interest and can also conserve species accurately. Many nonlinear equations in scientific and engineering applications admit variable transformations that expose polynomial structure. This combined with the nonintrusive nature of the approach make it a viable option for deriving ROMs for complex nonlinear applications where traditional intrusive model reduction is impractical and/or unreliable.