• No results found

Parameter Estimation in Biofilm Models

N/A
N/A
Protected

Academic year: 2021

Share "Parameter Estimation in Biofilm Models"

Copied!
2
0
0

Loading.... (view fulltext now)

Full text

(1)

Parameter estimation in biofilm models

Alma Masic, Niels Chr. Overgaard, Anders Heyden

Applied Mathematics Group, Malmo University, 20506 Malmo, Sweden; alma.masic@ts.mah.se, nco@ts.mah.se, heyden@ts.mah.se

Keywords: biofilm models, parameter estimation, moving bed reactor, detachment

Mathematical modeling of biofilm development has been an active research topic during the latest years. The main purpose of these models is to understand and predict biofilm growth and devel-opment. One of the first mathematical model was the so called Wanner-Gujer model, presented in [4], where a one-dimensional model capable of describing multiple species development was proposed. Later on, more advanced models, incorporating more dimensions, eps-production, cell-to-cell-signalling and other properties have been developed. cf. [2]. The major obstacle when trying to predict biofilm development using these mathematical models is that a huge number of parameters are needed, such as parameters in the bacterial metabolism, growth and detachment rates. Some of these parameters are possible to measure in experimental conditions, but others are more difficult, such as the detachment rate. In this paper, we will propose a methodology for estimating parameters in mathematical models of biofilm development from comparison of model prediction and experimental data. The proposed method is based on parameter identification meth-ods used in automatic control theory, see [3]. We will especially focus on the determination of the detachment rate in the Wanner-Gujer model.

The model proposed in [4] considers a one-dimensional biofilm growth in the direction z >0. The biofilm consists of NB bacterial species and NS different substrates, where fi denote the volume

fractions of the different species. The model is based on a conservation law for both the substrates

and the biomass, which results in ∂fi

∂t + u ∂fi

∂z = (µOi− ¯µO)fi, i= 1, . . . , NB , (1)

µOidenote the specific growth rate,µ¯Odenotes the average growth rate and u denotes the velocity

field, defined by u= Z z 0 ¯ µO(z′)dz′ , (2)

and the corresponding equations for the substrates ∂Sk ∂t + ∂ ∂z(Dk ∂Sk ∂z ) = rk, k= 1, . . . , NS , (3) where Dkdenote the diffusivities and rkthe consumption rates. The thickness of the biofilm, L(t)

evolves according to

d

dtL= u(t, L) − σL

2

, (4)

where σ denotes the detachment rate.

We propose to estimate parameters by observing substrate concentrations, Sk, from real

experi-ments and compare them with substrate concentrations predicted from the mathematical model, ˆ

Sk(p1, . . . , pn), where pi denote the unknown parameters, obtained from (1)–(4) under NC

dif-ferent running conditions. The optimal estimate of the parameters is obtained by minimizing the functional min p1,...pn V(Sk, ˆSk) = min p1,...pn NXC,NS j,k (Skj −Sˆ j k) 2 , (5) 1

(2)

Figure 1: Comparison of estimated and measured global reaction rates.

where the index j denotes the different running conditions. In continuous time the functional in (5) can be replaced by the corresponding integral. The actual optimum is found by calculating the first order variation, dV of the functional in (5) and applying a gradient descent method resulting in the following differential equation

dpi

dτ = −dV (pi(τ )) , (6)

where τ denotes artificial time.

The proposed method will be used to estimate the detachment rate in a moving bed biofilm reac-tor for waste water treatment presented in [1], see Fig. 1 for an illustration of the predicted and measured nitrification rates. Based on the Wanner-Gujer model in (1)–(4) and the gradient descent scheme for solving (5), we will estimate the detachment rate.

References

[1] E. Alpkvist, J. Bengtsson, N.-C. Overgaard, M. Christensson, and A. Heyden. Simulation of nitrification of municipal wastewater in moving-bed biofilm reactor: a bottom-up approach based on a 2d continuum model for growth and detachment. In Proc. International Conference

on Biofilm Systems, 2006.

[2] E. Alpkvist and I. Klapper. A multidemensional multispecies continuum model for heteroge-neous biofilm development. Biotechnology and Bioengineering, 69:765–789, 2007.

[3] L. Ljung. System Identification - Theory for hte User. Prentice-Hall, Upper Saddle River, N.J., 2nd edition, 1999.

[4] O. Wanner and W. Gujer. A multispecies biofilm model. Biotechnology and Bioengineering, 28:314–328, 1986.

Figure

Figure 1: Comparison of estimated and measured global reaction rates.

References

Related documents

The aims of the present study were to (i) characterize and compare saliva and tooth biofilm microbiota via deep sequencing using the Illumina MiSeq and the PacBio SMRT plat- forms

The structural form estimator is for its use dependent on a priori spe­ cifications on model and parameter residual covariances, the initial parameter vector and the transition

However, the injection barrier can be tuned with the assistance of the stray field of the ferroelectric polarization charges.[7, 9] The ascending curve in Figure

Focusing on a specific crowd dynamics situation, including real life experiments and measurements, our paper targets a twofold aim: (1) we present a Bayesian probabilistic method

Since the study’s focus is on finding parameters that should be considered in SDP’s  to improve SDP’s success, theories that support fast product delivery, software  development

Plotting the dOFV distributions obtained from the M proposal samples and from the m SIR resamples against a Chi square distribution with degrees of freedom equal to the number

As mentioned in section 3.1, the present model uniformly predicted initial void formation centrally in the microcolonies at the substratum; in the simulations

Coherent laser radar for vibrometry: Robust design and adaptive signal processing Ingmar Renhorn, Christer Karlsson, and Dietmar Letalick National Defence Research Establishment