Building Services Engineering Research & Technology

Buildings typically have a long life span, which can easily reach 50 or 100 years. In the light of expected climate change it is therefore important to look towards the future, and to analyse how buildings will cope with the changes in climate that are predicted by the climatologists. In such long-term predictions of building performance, uncertainties play a large role.

This paper describes a preliminary study that aims to better map out the consequences of dealing with a whole range of uncertainties in the specific case of predicting the effect of climate change on the energy use and thermal comfort (overheating) in the large stock of terraced houses in the UK. Uncertainties in climate change prediction are compared with other variable factors like building occupancy patterns, actual thickness of construction materials and HVAC control settings. Uncertainties have been propagated in a transient model of these terraced houses using the transient simulation program EnergyPlus. In order to explore the large range of input variants use has been made of a genetic algorithm. The current search and solution spaces are discussed, putting the impact of climate change in perspective with regards to other changes and developments that might increase or decrease overheating. Overall, the findings indicate that uncertainties in long term thermal performance predictions run high, with standard deviations of over 100%. However, the robustness of the contemporary dwelling design is fortunately high.

Residential Ground Coupled Heat Pump systems are usually characterised by an ON/OFF behaviour of the heat pump with typical cycling frequencies of 1—4 cycles per hour. The ground loop fluid pump has the same ON/OFF behaviour and the borehole heat exchanger operates either in full flow or no flow conditions. Typical hourly simulations of GCHP systems use steady-state models for the heat pump and the borehole fluid (transient models being used for buildings and heat transfer in the ground). This paper reviews the models used in typical hourly simulations as well as transient models that are available and compares the results obtained using the two classes of models within the TRNSYS simulation environment. Both the long-term energy performance and the optimum system design are compared. It is shown that using steady-state models leads to an overestimation of the energy use that ranges from a few percents with oversized borehole heat exchangers to 75% for undersized exchangers. A simple Life Cycle Cost analysis shows that using steady-state models can lead to selecting a very different design than the one that would have been selected using dynamic models.

The paper describes a modelling study of heat transfer and buoyancy-driven airflow in double skin facades consisting of a glass outer layer, a control device (venetian blind) and a double-glazed inner skin. The modelling study was based on two approaches — a component-based, lumped parameter simulation which used a public domain, open source differential/algebraic equation solver and a detailed, CFD calculation which included air flow, conduction, convection and radiation. The primary objective of the work was to compare the performance of the simplified model with the output of a rigorous CFD calculation.

Validation of simulation models appears from a long time as a key issue in order to promote a more intensive and more efficient use of simulation models in the field of building and HVAC simulation. IEA Annex 34/43 originally targeted a number of specific applications where a more advanced validation was required: ground coupling problems, multizone building, shading, day lighting and cooling load interaction, HVAC components and ventilated facades. These validation exercises were built on the large methodological experience obtained in previous projects and address sometimes very fundamental problems of heat transfer in buildings. Consulting engineers and practitioners might see these exercises as a bit too far from their objectives and it is the reason why an additional activity was proposed with the specific aim of producing, based upon the results of the validation of models, a set of reference simulations. These applications cover a range of building types (residential, commercial) and systems (production, distribution emission) and run in a variety of climates. The paper will describe how models dedicated to these applications were developed, starting from validation results, going through the selection and consolidation of simulation hypothesises and ending with a number that might be considered as reference for the concerned applications. The paper will concentrate on models required by a residential building application (multizone building equipped with a heat pump or a condensing boiler, heat emitted by radiators or floor heating systems). Simulations make use of both EES and TRNSYS software and both software are applied in parallel as far as possible in the different applications in order to get a better judgment of their potential advantages and drawbacks. The use of reference simulations in view of qualifying normative methods currently in development in the frame of the European Energy Performance in Buildings Directive is also addressed and demonstrated in the paper.

Building level energy models are important to provide accurate energy consumption prediction for building performance diagnosis at building level as well as energy efficiency assessment of retrofitting alternatives for building performance upgrading. However, it is a great challenge to establish building energy models of existing buildings for long-term energy performance prediction only using short-term monitoring operation data. In this study, a mixed-mode building energy model is proposed to describe building system for thermal performance prediction at the building level. The model includes two parts. One part is the simplified energy models (3R2C) of building envelopes. The parameters of these simplified models are determined by comparing the frequency characteristics of those simplified models and their theoretical frequency characteristics based on the easily available detailed physical properties of exterior walls and roof. The other part is the simplified 2R2C model for building internal mass, whose detailed physical properties are very difficult to obtain. The parameters of the building internal mass model are optimised using short-term monitored operation data. A genetic algorithm estimator is developed to optimise these parameters. The parameter optimisation of the simplified building internal mass model (2R2C) and the mixed-mode building energy model are validated in a high rising commercial office building under various weather conditions. An application of this model for performance evaluation of alternative control strategies is illustrated briefly.

This paper intends to show how benchmarks can help in the audit of a HVAC system and how these benchmarks can be generated. The work presented is part of the European `AUDITAC' project. How can an auditor declare that a given HVAC is `consuming too much'? He can't make any judgment, if not having some reference, i.e. some `benchmarks' available. Focus is given here to cooling regime, but even then, heating cannot be forgotten (for example, the remaining heating demand can be satisfied thanks some recovery on the condensers of the chillers). The dramatic question is: what should be the consumption(s) for such a building, in such a climate, with such occupancy, such internal loads and such actual indoor environment? Better than to look for a (very hypothetical) global weather index, similar to heating degree-days, it seems more rational to run a simulation model on a few thousands of hours, corresponding to one (or to several) cooling season(s). Current performances of simulation tools make this approach very expedient. The climate can then be considered as it is, without any simplification.

The main simplification is still welcome on the system (building + HVAC) side, in order to get calculation robustness, easy understanding and easy parameter identification. In this perspective, the building has to be subdivided in a very limited number of zones and only a few components of the HVAC system have to be included in the simulation model, with as simple as possible control strategies.