The durability model presented is only one, of the needed durability indi-cators in the service life framework. The simulation tool is, therefore, not complete as a tool for calculating the end of service life for any given con-crete. Some issues in the model needs to be improved in future versions in order to facilitate simulations used in the design of concrete structures.
A general issue, for which similar reactive mass transport models may suer from as well, is related to the FE boundary node and the chemical equilibrium calculated at this node. The ion concentrations at the boundary node are prescribed in each FE time step, which eliminates the time depen-dency for this node. The implicit dissolution rates of the solid phases at the boundary node are only dependent on the number of time steps. The ulti-mate future goal should be a complete kinetic description of the dissolution reactions involved which would solve the above discussed problem. Other dierent solution methods may be relevant in this context, e.g., identica-tion of the primary solid phases and their dissoluidentica-tion rate descripidentica-tion or the assignment of multiple nodes in the FE system for the description of the boundary solution itself.
The current version of the durability model is sensitive to the chemical reactions used as input. The model requires that a single stable chemical model is dened for the whole spatial and transient domains. Prior knowl-edge of which chemical reactions occurs in specic environments is needed in order to get a valid result. This is not optimal in the sense of using the service life framework as an engineering tool. The durability model would be strengthened if stable validated chemical models are predened. On the other hand, the model should continue the development in an open format where it also is possible to dene custom chemical models for research purposes.
It is important that the model is continuously updated with the recent thermodynamic databases which also includes descriptions of supplementary cementitious materials. Supplementary cementitious materials are essential in modern concretes and especially in the the development of new more environmentally friendly concretes.
The durability model includes a description of water and vapor diusion in order to simulate non-saturated concrete. Diusion of ions in the gas phase is not accounted for in the current version of the model. The gaseous species establish an equilibrium condition primarily with the pore solution and thereby aects the whole chemical equilibrium. The diusion of gaseous carbonates are of interest in this context due to the eects of carbonation of the concrete. A set of multi-species gaseous diusion equations should be added to the existing dierential equations in a future version. The chemical model needs to be extended along with the gaseous diusion extension to account for the gas-liquid equilibrium.
The mass transport rate of ions is sensitive to the dierent reduction factors on the diusion coecient as shown in Paper IV. The tortuosity is assumed constant over time in the current version of the durability model even though the amount of solid phases are changed by ingress and leach-ing and thereby change the geometrical conguration of the pore structure.
The model should account for the pore structure evolvement in terms of a tortuosity factor which is dependent on the actual solid phase composition at a given time and spatial position. The porosity could be calculated by the solid phase composition and then relate the porosity and tortuosity. The changed geometrical conguration may also create stresses and strains in the solid structure due to volume expansions. The mixture theory includes the theoretical description of mechanical impacts, but it may be extremely com-plicated to solve numerically, i.a. due to the strong non-linearities introduced and fracture mechanical considerations that need to be taken into account as well.
The service environment temperature and the variation of this is of great importance for the physical and chemical properties. The mass transport is therefore aected by temperature in the diusion and convection process.
The sorption isotherms, are changed with temperature and thereby also the sorption hysteresis described by the phenomenological model. The chemical reactions are individually temperature dependent and the overall chemical equilibrium state is aected. The temperature should be included as an additional state variable into the established durability model. Coupling of the temperature to the diusion and chemical equilibrium calculations is pretty much straight forward whereas the coupling to the sorption hysteresis model may be more complex.
Computational speed is important in the further development of dura-bility models but also for the complete service life framework calculation.
Parallelization should be investigated even though a simple parallel version in some parts failed to succeed in this work. Computational speed is also an issue for extending the current one-dimensional FE model to two- or three-dimensions. Two- and three-dimensions are of interest for the durabil-ity modeling when for instances cracks are taken into account or in systems where dierent types of materials are combined, like in a brick wall or foun-dations with concrete combined with light weight blocks.
How should a coupled non-linear model be validated? The mass transport models, coupled in this work are validated individually and the reactions in the chemical models are validated by dierent separated experiments. As-sembling these models and comparing them with assembled experimental tests may not always give compatible results and a complex problem arises due to this. Validation of coupled models are dicult as it is dicult to experimentally separate the individual physical and chemical processes in order to detect the reasons for inconsistency between simulation results and experimental results. It is therefore dicult to know where to improve the numerical model.