Having established the requirement for a new building evaluation tool, an inspection is required of the ways in which previous research has analysed tools. Previously, this has been done with the aid of computer simulation to show proof of concept, the use of fully climate controlled laboratory facilities or semi-controlled test cell facilities to show real world applications, and comparisons with other established tests to prove validity.
Test Facilities
When analysing the application of a building evaluation tool, tests are either carried out on a residential home, or using a test facility, either in a laboratory or using a test cell.
Researchers use both approaches extensively. Each has different advantages. In short, the advantage of using a real home is complete demonstration of the tool as it can be applied in the field, while the test facility allows for a much higher level of control over test conditions (Baker & Van Dijk 2008; Cattarin et al. 2016; Dewsbury 2011).
2.10.1.1 Occupied Buildings
Analysis of a method as it applies to a real building is a direct evaluation of the test’s overall application, beyond analysis of the theory alone. It provides insight into the difficulties that will arise when testing in other homes, such as site access and disrupting occupant routines.
It can be difficult, however, to differentiate all the factors that influence results due to the complex nature of the building system (Cattarin et al. 2016). Subbarao et al. (1988) demonstrated the application of the PSTAR method on a residential home, though this study did not compare the PSTAR results with other benchmarks. Butler & Dengle (2013) used two residential homes to demonstrate different influences on the co-heating test, and its comparison with the design parameters. This showed the application of the co-heating test in the real world, and was applied to two different homes, one control and one with various alterations. It showed the test’s repeatability, and the influences of real world conditions (such as sun and wind) on test reliability. Window shading use enabled analysis of how high solar gains would influence the test. Stamp (2011) also applied the co-heating test to two high efficiency buildings to evaluate test uncertainty.
These studies are valuable as they show the application of the tests in field, which both provides information regarding their feasibility, and evaluates performance of the tests.
58 However, they are not the initial application; the examples provided are of tests that have undergone much theoretical development and experimentation in controlled environments.
2.10.1.2 Test Cells and Laboratories
Applying test methods in field is the culmination of much work in designing and adjusting methods on the basis of controlled testing, either in laboratory settings or outdoor test cell buildings. These test facilities provide a means to simulate real world performance on a real building under fully known or controlled stimulus (Achenbach 1981; Baker & Van Dijk 2008;
Cattarin et al. 2016; Dewsbury, Nolan & Fay 2007b; Lomas et al. 1997). Using test cell buildings is exceptionally popular since they give the researcher full access and control over all elements. The preference is for the test cell to be specifically designed for the study, such as those used by Dewsbury, Nolan & Fay (2007b), Hitchin, Delaforce & Martin (1993), Lundin, Andersson & Östin (2004) and Sugo et al. (2004): this enables incorporation of specific design requirements. For example, Lundin, Andersson & Östin (2004) included ways to alter airflow into the test cell. Dewsbury, Nolan & Fay (2007b) designed the timber frame to ,among other things, allow windows to be retrofitted for future studies. However, it is not always feasible to build a new test facility from scratch, and cells are re-used when they are available and still fit the requirements, such as the PASLINK buildings used by Baker & Van Dijk (2008) and Cesaratto, De Carli & Marinetti (2011). These studies all investigated elements which require significant knowledge of the building as part of the validation process, or they would be negatively impacted by increased complexity. The experiments discussed by Baker & Van Dijk (2008) require a number of days, and the test cell allows this without disrupting building residents. These studies all investigated new building evaluation procedures, or quantified different influences. Other examples of using test cells to investigate building evaluation tools include Cucumo et al. (2006), Letherman & Palin (1982), and Loutzenhiser et al. (2009).
Validation
The approaches to validating testing procedures are logical and follow general principles of a controlled experiment. Results of the new test procedure are compared to results of a previously known method. The control method can range from theoretical values based on an assessment calculation method or a simulated model, results of a long term monitoring project, or a different post-construction test (Mangematin, Pandraud & Roux 2012).
59 2.10.2.1 Simulation tools for Validation
Simulated models are often used as the first step in showing an analysis technique or approach works on a theoretical level and warrants additional research. It can show that conclusions can be drawn when certain data is available.
In some cases, comparison with the real world is unnecessary. Wang & Chen (2002) used computer simulations to demonstrate a different procedure for calculating conduction transfer functions and thermal response factors of multilayer walls. There was no need for external, real world validation as the study’s aim was to provide the simulation mechanic with a faster and more reliable method of arriving at the same answer. Ballarini & Corrado (2012) implemented a similar approach, using EnergyPlus to show how the principle of superposition can be used to determine the level of influence of the forces driving the cooling load of a building. This type of analysis is impossible to do in the real world, as the simulation allows alteration of physics and application of adiabatic building elements which are effectively impossible to apply to an in situ case study of a real building. Karlsson, with the simulation program (Hitchin, Delaforce & Martin 1993). A clear example is that of the ‘perfect mixing’ assumption that Hitchin, Delaforce & Martin (1993) showed as incorrect when compared with experimental data from a well-mixed room.
2.10.2.2 Experimental Validation
Comparisons with other post-construction tools generally show how the tool improves upon either the application or the accuracy of previous tools. The benchmark tool to be compared against is generally well understood, if not the standard procedure for the particular test.
Cucumo et al. (2006) utilised the average method of heat flux testing as the benchmark for proposing a different analysis of heat flux data. The aim was to increase the application of heat flux meters for in situ evaluation of R-values. However, Ficco et al. (2015) used alternative tests to evaluate parameters affecting heat flux tests. Judkoff et al. (2000) used co-heating tests as their benchmark in evaluating performance of the STEM test for
60 evaluating HTC. Cesaratto, De Carli & Marinetti (2011) provided an alternative, using a series of trials to show what effects the presence of the heat flux meter itself might have on the results. Repeating heat flux tests under different conditions, and noting the variation, indicated that differences in emissivity might have a small effect, and that using the heat flux meter surface temperature would also improve accuracy.
These are not tests designed to evaluate the building’s energy use. They are all designed to estimate, in some way or format, the overall R-value of the building. With previous work linking how R-values influence building performance, there is no need to conduct long term monitoring tests to validate each new tool evaluating the building R-value. It is far easier to use previously developed tools as benchmarks for comparison.