Non-linear grey-box modelling for heat dynamics of buildings

Time-dependent solar aperture estimation of a building: Comparing grey-box and white-box approaches

“…So, the real gA curve of this particular building will obviously not be constant. Therefore, some researchers have reported that assuming gA constantly might yield 'unqualified' grey-box models [43], while other suggest that a qualified grey-box model for model predictive control (MPC) requires on-site measured [38] or pre-simulated [44] dynamic solar gain data, as extra input, rather than estimating solar gains by Eq. ( 5).…”

Indoor Air Quality Assessment via Experimentally Calibrated Dynamic Simulation: A Case Study in an Office Building in Panama

“…According to this, the major differences was recorded during the last day from 5:00 a.m. until 3:00 p.m., about 1.9 °C to 3.4°C. Nevertheless, there were three other days with errors above 1.7°C, ranging between 1:00 p.m. to 6:00 p.m. those periods reflected a rise in temperature and then a slowly declining, this phenomenon happened in [24] but no more than 1°C. To know the error of the implemented model, the average error within the ten days was calculated (Fig.…”

“…The outcomes presented, showed that there was a correlation between opening windows with indoor and outdoor environmental condit io ns; als o, s enso r recorded data can be employed to take to do thermal an d environmental predictions. This is the case of C. Thilker et al [19] who developed an indoor temperature prediction modeling through air condition monitoring and S. Ryu and Y. Chen [20] who performed an occupancy prediction with the data obtained via motion sensors.…”

Estimating dynamic solar gains from on-site measured data: An ARX modelling approach

“…a valid model for predictive control, pre-simulated (BES-model based) solar gain dynamic details [16] or accurate measurements of dynamic solar gain [27] are strongly required. The invariant solar aperture (gA) assumption might lead to unqualified statistical models since it estimates the gA parameter with a large uncertainty and without statistical significance [34]. Therefore, it is argued that a precise 3/29 dynamic 𝑔𝐴 𝑡 curve (i.e.…”