Kalman Filter based Newton Extremum Seeking Control for Maximum Gases Production Rates of Anaerobic Digestion Process

Nicolai Christov,Haoping Wang,Ivan Simeonov,Yang Tian,Maobo Hu,Lyudmila Kabaivanova

doi:10.5755/j01.itc.49.4.25367

Nicolai Christov, Haoping Wang + Show 4 more

Open Access

https://doi.org/10.5755/j01.itc.49.4.25367

Copy DOI

Abstract

This paper proposes a Kalman filter (KF) based Newton extremum seeking control (NESC) to maximize production rates of hydrogen and methane in anaerobic digestion process. The Kalman filtering algorithm is used to obtain more accurate gradient and Hessian estimates which makes possible to increase the convergence speed to the extremum and to eliminate input and output steady-state oscillations. The simulation examples demonstrate the performances of the proposed approach.

Highlights

Anaerobic digestion (AD) gets more and more attention nowadays because it is a relatively cheap, sustainable and efficient method to treat organic wastes
Most mathematical models of AD processes are based on mass-balance equations or on a zonal description of the bioprocess variables 8
The first one consists of complex models, such as ADM1 (Anaerobic Digestion Model 1) 6 using 35 state equations to describe the AD process in detail

Summary

Introduction

Anaerobic digestion (AD) gets more and more attention nowadays because it is a relatively cheap, sustainable and efficient method to treat organic wastes. In order to obtain good performances and high quality production, various control and observer algorithms have been developed and used for AD process control: sliding mode observer based optimal control 3, non-parametric adaptive control 25, and robust nonlinear observer 17. A KF based NESC is proposed to optimize an AD system for hydrogen and methane production in order to obtain maximum gas production rates. The AD system consists of a cascade of two continuously stirred anaerobic bioreactors which can produce hydrogen and methane separately. The proposed KF based NESC algorithm do not require any process model information and uses only input and output measurements – dilution rate and the biogas outflow rate. There are no steady-state oscillations because the amplitude of the perturbation signal converges to zero during steady state regime This ensures the smoothness of the dilution rate and the gas production rate.

Model of Anaerobic Digestion Process

K D2 where ratio o bioreact parame

Simulation Results

Maximizing Total Gas Production Rate

Conclusion