Bruce J. Taylor

Analytical investigation of the steady-state behaviour of dynamic and diffusive building envelopes

Abstract A one-dimensional model describing the steady-state dynamic and diffusive behaviour of a three-layer building envelope element is developed with the objective of elucidating the physics of simultaneous heat and vapour transport through dynamically insulated building envelopes. The equations are simple to program on a spreadsheet, enabling architects to build tools which will allow them to design “breathing” envelope constructions. The variables at the designers disposal are air and vapour permeability, thermal conductivity and thickness of the layers comprising the envelope. Users wishing to consider more complicated constructions will find that they can readily extend, by inspection, the equations from the three layers presented here to any number of layers. Whilst mass transfer has been discussed in terms of water vapour transport, the equations are very general and can be applied to the transport of any gas through a permeable wall.

The building envelope as an air filter

Recent research suggests that fine-particulate air pollution increases the incidence of lung disease and premature death. In this paper, single fibre filter theory was used to predict the theoretical particulate collection efficiency of air permeable walls (dynamic insulation). The relationship between particle diameter and filtration efficiency for dynamic insulation, as a function of flow rate, is examined and compared to that for a conventional filter. Factors such as filter penetration as a function of flow rate, filter thickness, and packing density for a range of particle diameters are also presented. The findings suggest that, in addition to reducing heat loss through the building fabric, dynamic insulation can act as a high performance air filter in naturally ventilated buildings, thus providing a viable and attractive alternative to mechanical air-conditioning in congested urban environments.

Case-based reasoning for matching SmartHouse technology to people's needs

SmartHouse technology offers devices that help the elderly and people with disabilities to live independently in their homes. This paper presents our experiences from a pilot project applying case-based reasoning techniques to match the needs of the elderly and those with disabilities to SmartHouse technology. The SmartHouse problem is decomposed into sub-tasks, and generalised concepts added for each sub-task. This decomposition and generalisation enables multiple case reuse employing a standard decision tree index based iterative retrieval strategy. Documented real situations are used to create a small case base. An initial prototype is evaluated empirically using leave-one-out testing, and separately with the domain expert on newly created test cases. Results show that system generated solutions to be comparable to those of a domain expert. Importantly, the iterative retrieval strategy employing multiple indices generated solutions that were significantly better compared to a one-shot retrieval without indices.

Case Authoring: From Textual Reports to Knowledge-Rich Cases

SmartCAT is a Case Authoring Tool that creates knowledge-rich cases from textual reports. Knowledge is extracted from the reports and used to learn a concept hierarchy. The reports are mapped onto domain-specific concepts and the resulting cases are used to create a hierarchically organised case-based system. Indexing knowledge is acquired automatically unlike most textual case-based reasoning systems. Components of a solution are attached to nodes and relevant parts of a solution are retrieved and reused at different levels of abstraction. We evaluate SmartCAT on the SmartHouse domain looking at the usefulness of the cases, the structure of the case-base and the retrieval strategy in problem-solving. The system generated solutions compare well with those of a domain expert.

Automatically Acquiring Structured Case Representations: The SMART Way

Acquiring case representations from textual sources remains an interesting challenge for CBR research. Approaches based on methods in information retrieval require large amounts of data and typically result in knowledge-poor representations. The costs become prohibitive if an expert is engaged to manually craft cases or hand tag documents for learning. Thus there is a need for tools that automatically create knowledge-rich case representations from textual sources without the need to access large volumes of tagged data. Hierarchically structured case representations allow for comparison at different levels of specificity thus resulting in more effective retrieval than can be achieved with a fiat structure. In this paper, we present a novel method for automatically creating, hierarchically structured, knowledge-rich cases from textual reports in the Smart-House domain. Our system, SMART, uses a set of anchors to highlight key phrases in the reports. The key phrases are then used to learn a hierarchically structured case representation onto which reports are mapped to create the corresponding structured cases. SMART does not require large sets of tagged data for learning, and the concepts in the case representation are interpretable, allowing for expert refinement of knowledge.

Dynamic insulation in multistorey buildings

Dynamic insulation permits the movement of air and moisture through the external walls of a building to reduce heat loss and achieve high indoor air quality. The present paper details a pilot study carried out to examine the influence of fire safety requirements and external wind on the performance of naturally ventilated multistorey buildings in which the external envelope is dynamically insulated. The theoretical foundation is outlined for a spreadsheet model used to simulate prototype 3-, 4-, 5- and 10-storey buildings all sharing the same rectangular floor plan, with fresh air drawn into the building through the envelope by depressurisation using a fan-driven, ducted extract system. From the analysis, the effects of wall porosity, depressurisation level, extract system deployment, occupant density and distribution, and building orientation have been quantified, confirming the practical feasibility of such a system.

Using computer aided case based reasoning to support clinical reasoning in community occupational therapy

Community occupational therapists have long been involved in the provision of environmental control systems. Diverse electronic technologies with the potential to improve the health and quality of life of selected clients have developed rapidly in recent years. Occupational therapists employ clinical reasoning in order to determine the most appropriate technology to meet the needs of individual clients. This paper describes a number of the drivers that may increase the adoption of information and communication technologies in the occupational therapy profession. It outlines case based reasoning as understood in the domains of expert systems and knowledge management and presents the preliminary results of an ongoing investigation into the potential of a prototype computer aided case based reasoning tool to support the clinical reasoning of community occupational therapists in the process of assisting clients to choose home electronic assistive or smart house technology.

CBR for SmartHouse Technology

SMART HOUSE technology offers devices that help the elderly and people with disabilities to live independently in their homes. This paper presents our experiences from a pilot project applying case-based reasoning techniques to match the needs of the elderly and those with disabilities to SMARTHOUSE technology. The SMART HOUSE problem is decomposed into sub-tasks, and generalised concepts added for each sub-task. This decomposition and generalisation enables multiple case reuse employing a standard decision tree index based iterative retrieval strategy. Documented real situations were used to create a small case base. A prototype implemented using RE CALL 1 with TCL script is evaluated empirically using leave-one-out testing, and separately with the domain expert on newly created test cases. Results show the generated solutions to be comparable to those of a domain expert. Importantly, the iterative retrieval strategy employing multiple indices generated solutions that were significantly better compared to a best match retrieval without indices.