Structural identifiability analysis of the Anaerobic Digestion Model No. 1 using a local algebraic observability approach

摘要

Structural identifiability is a prerequisite for a successful parameter estimation of a dynamical model. ‘Structural’ relates to the theoretical possibility of estimating the parameters under the idealized conditions that all measurements are perfect and continuous, and that the model is a perfect representation of the real process. In this research, the Anaerobic Digestion Model No. 1 (ADM1) is analysed. The model is of great importance in anaerobic digestion modelling. Due to its numerous parameters, however, its structural identifiability is an open question. The used method investigates the local algebraic observability, which is a sufficient condition for structural local identifiability. Different measurement scenarios were tested, as well as the influence of the choice of inputs. The obtained results are optimistic, since in most conditions, nearly all of the initial conditions of the states and the kinetic, stoichiometric and composition parameters are found to be structurally locally identifiable if almost any one of the states is measured. Unidentifiability occurs for the parameters pertaining to inerts and the cations and anions. For these parameters, nevertheless, structurally identifiable parameter combinations are found.