Modified Glaser ' s Condensation Model

It is well-known that Glaser's model []-[7], a combined graphical and numerical method to assess the condensation of water vapor in building stnrctures t8l-t91, suffers from some drawbacks which make the model rather debatable. For example, the model includes neither hygroscopic nor liquid transports and also omits the transition from the liquid into the solid phase. In addition, the numerical part of the model is based on isothermal diffusion. However, real building envelopes, especially in the winter season, are considerably non-isothermal. These inconsistencies raise the question: What will happen with the model if fully non-isothermal conditions are incorporated into its scheme. This paper aims to implement non-isothermal diffusion into Glaser's model and discusses some features of the modified scheme.


I Introduction
It is well-known that Glaser's model []- [7], a combined graphical and numerical method to assess the condensation of water vapor in building stnrctures t8l-t91, suffers from some drawbacks which make the model rather debatable.
For example, the model includes neither hygroscopic nor liquid transports and also omits the transition from the liq- uid into the solid phase.In addition, the numerical part of the model is based on isothermal diffusion.However, real building envelopes, especially in the winter season, are considerably non-isothermal.These inconsistencies raise the question: What will happen with the model if fully non-isothermal conditions are incorporated into its scheme.This paper aims to implement non-isothermal diffusion into Glaser's model and discusses some features of the modified scheme.
2 Non-isothermal steady-state diffusion In our previous paper I I 0] we developed two basic models for describing non-isothermal steady-state diffirsion of water vapor through porous building materials, namely the DIAL and DRAL models.
The DIAL model (Diffusion through an Immobilized Air I-ayer) led to the following basic relations (ll(4) that hold for a non-isothermal structure (wall) of thickness d, which is embedded into an air environmenr Dry air is supposed to have a constant concentration co on both sides of the wall.The wall shows, however, different temperatures ?,, ?", and concentrations c,*, c2w of water vapour on its two sides (internal and external).The concentration profile y(r) and diffi-rsion flux qi, expressed within the DIAL non-isothermal model, read *r r C,,,( X\ The symbols p and fijn" are the diffusion resistance factor and diffusion resistance of the structure, respectively, while p., stands for the pressure of dry air.More details concerning the derivation of (l)-(a) can be found in [0].
Let us have, for example, a building envelope realized by a plain brick wall (without plaster, g = 9) of thickness d--44 cm.-fhe wall seParates a heated room of a usual environment (surface temperature and relative humidity: Tr=293.15 K, 9F60Vo RH) from an outdoor space (Tr=293.15K,qr=60 7o RH).Table I and Frgs' l, 2, 3 show the results of Glaser's schemes based on isothermal and non-isothermal diffusions and applied to the structure given above.Since the structure is only 'weakly' non-isothermal (7,-Tr< 40K), Iarge differences in results cannot be ex- pected (see the discussion in [10]).The amount of condensate L,Qtrg extracted fiom an area of 1 m2 per second is calculated as the difference between the diffusion flux g, entering the condensation region at pointl (see ligs.l, 2, 3) and leaving the region (qu) at point 'B Fig. 3: Glaser's condensation scheme with DRAL approximatron (Glaser's snnd.ardmodel: 4.4'10-8 kg'm-2's-r; DIAL:   4.2 1 . 1 0-s kg' m-2' s-t ; DRAL: 3'5' l 0-8 kg' m-2' s-r) show certain differences.The largest amount of condensate is yielded by Glaser's standard model, while the smallest amount of condensate is forecasted the DRAL model, As was mentioned (e) The corresponding amounts of condensate Lqg (lab- le l) determined by means of the three investigated models earlier [10], the applicability of the DRAL model is restricted especially to materials possessing closed pores, which is not our case, so this result does not represent a relevant condensation prognosis.On the other hand, the DIAL model has been developed [l0] especially for building materials having mutually interconnected open pores.Thus only the comparison between the standard and DIAL values is meaningful.This comparison shows that the standard value (4.4.10-8 kg.mr.s-]) is larger than that of the DIAL model (4.21.l0-8 kg.--t.s-t) by about four percent, which is a negligibly small difference in this field.Nevertheless, in extreme climatic regions (Tr-Tr> 40 K) the difference may be significantly larger [10], and in such cases the DIAL model will yield a more realistic prognosis as compared with the standard model.which overestimates the amount of condensate occurring in building stmctures.However, thanks to the ability to overestimate the condensate, Glaser's standard model yields prognoses that are on the 'safe side', though not always on the 'economical side'. 4

Conclusion
It has been illustrated that Glaser's standard condensat- ion model can be modified to include fully non-isothermal calculations ofwater condensate appearing in building structures.Such calculations do not lead to very diflerent results in normal Central European climatic regions (AT < 40 K).
Howeve! in extreme climatic regions (LT > 40 K) essential differences can be expected and in such events the DIAL model offers a good tool for a more realistic assessment of condensation problems.In conclusion, it should be stressed that Glaser's standard condensation scheme provides an as- sessment that is on the safe side, but especially in extreme non-isothermal cases it will not lead to an economicallv opti- mal design of building stmctures.