Award Date


Degree Type


Degree Name

Master of Architecture (MArch)



First Committee Member

Alfredo Fernández-González

Second Committee Member

Daniel Ortega

Third Committee Member

Firas Al-Douri

Fourth Committee Member

Robert Boehm

Number of Pages



Roofponds mimic the ways in which nature tempers and controls the global climate; they utilize higher heat capacity of water to passively control the temperature of the interior space. From a thermal standpoint, roofponds are strong performers, providing high solar savings fractions, interior temperature stability, enhanced thermal comfort and very low operational power requirements. Moreover, due to convective heat transfer within the water bags, heat gains or losses are quickly distributed throughout the roofpond to create a very homogeneous distribution of heat throughout the floor area covered by the system.

Research by Harold Hay and John Yellott (Hay & Yellott, 1968) studied the feasibility of the roofpond system and tried to develop a heat transfer model for roofponds with insulation. During the late 1960s, several of their publications discussed the heating and cooling potential produced by various roofpond strategies. Throughout the 70's and 80's, a number of heat transfer models were developed to simulate different roofpond systems. Researchers at Trinity University tried to simplify the complexity of the simulation model without compromising the accuracy of its prediction. However, only a couple of them considered the whole building's heat transfer mechanisms.

This research intends to develop a model to predict the hourly indoor air temperatures in a single-zone building featuring a roofpond. Like most of the passive design strategies, roofponds are difficult to model as they have; too many independent variables, mostly climatic parameters that influence the performance of the roofpond. However, the indoor air temperature of such a passive building (without mechanical conditioning systems) is highly influenced by the change in the daily outdoor air temperature profile as well as the incoming solar radiation. A transfer-function unsteady state model can predict the indoor air temperature of a roofpond building quite accurately. Such model can be greatly handy to design professionals for quick evaluation of such system during the early schematic design phase.

The study herewith presented uses data collected from a roofpond test cell located at the NEAT Laboratory of the University of Nevada, Las Vegas, and implements unsteady-state thermal heat transfer principles to predict average interior temperatures. The three distinct phases of the project are: first, to predict indoor air temperatures using transfer-function heat transfer equations; second, to statistically fine-tune the model by finding the correlation between the predicted and the measured temperature; and third, to validate the model using a different data set.

A thermal network model of the roofpond using the transfer-function method with a time step of one hour is used to calculate the indoor air temperature. Measured data of 14 days is used to develop the unsteady state heat-transfer model that can predict the average indoor air temperature. The predicted temperature then is regressed against the measured temperature to find the correlation. The cyclic patterns observed in residuals indicate the daily change in the outdoor temperature profile and imply that time-series model with Fourier series is apt for de-trending the pattern.

The model is then empirically modified to increase accuracy. Auto Correlation Factor (ACF) and Partial Auto Correlation Factor (PACF) tests suggested that either Auto Regressive (AR) or Auto Regressive Integrated Moving Average (ARIMA) model would neutralize residuals. The empirically developed AR / ARIMA model is then added with the physical model to predict the interior air temperature. The AR(2) model which yielded the best fit model, was tested against data from another summer month for validation. The proposed validated hybrid model is capable of addressing the change in configuration of the roofpond building and can accurately predict the indoor air temperature.


Heat – Transmission; Heat transfer model; Passive heating and cooling; Predictive model; Roofpond buildings; Solar heating; Solar ponds; Solar heating – Passive systems; Solar air conditioning – Passive systems


Architecture | Environmental Design | Oil, Gas, and Energy