Free evolution and outdoors weather implicate dynamic testing conditions, as it is clearly seen in the recorded data in Figure 3 to Figure 6, so dynamic testing is suitable for data analysis in this case.

Several references report the use of dynamical tools to estimate the main characteristics of building components from outdoors testing in real buildings (Ralb, 1988) and also using tests cells (SIC 1, 1994), (SIC 2, 1996). Usually these tests impose an optimum testing strategy as previously mentioned.

In this case analysis is complicated as no test strategy was implemented to optimise test sequence regarding analysis, and also because of the low range of some of the available signals due to the high degree of insulation of each wall.

Dynamic ARX multi-output models have been used for data analysis. General mathematical details about the analysis tools are described in (Ljung, 1999). More particular details about the estimation of the physical parameters are reported in (Jimenez et al, 2003). Indoors temperature and heat flux have been considered as output and outdoors temperature has been considered as input.

Taking into account that all walls where very thick, opaque and highly insulating solar radiation was not included into the used model.

Once obtained a suitable model for the dynamical system, U value was obtained taking into account that the obtained dynamical model particularised to the case that all inputs and outputs were constants must coincide with the steady state equation.

Results

The following results have been obtained:

All results were according to their expectable range, but all walls were found more insulating than expected.

Conclusions

Satisfactory results have been obtained, using non-intrusive measurements, besides testing sequence was not optimum.

Realistic estimations for U value of each wall of the building have been made available for other simulations works to estimate actual heating and cooling loads and also to evaluate the effect of the implemented saving strategies.

Discrepancies regarding values estimated according to the NBE-CT-79 could be due to the faulty sealing in the air chamber into the walls that could lead to produce ventilation phenomena that where not taken into account.