تجزیه و تحلیل حساسیت از اثر رفتار مستاجر در مصرف انرژی در خانه های استیجاری

There has been a history of low-energy design failing to translate into low measured energy consumption in domestic buildings. In part this failure can be attributed to occupant behaviour and household variation. It is therefore important to provide a method whereby such variation can be accounted for so that deviations from design values can be identified as natural variation rather than design failure. This paper addresses the likely range of occupant behaviour and the resultant impact on heating energy consumption for domestic Passivhaus buildings. Realistic, quasi-empirical, profiles for different occupancies, lighting, and appliance-use were applied to a set of 100 terraced Passivhaus units, and modelled in a dynamic building simulation programme. Strong correlations between the results and measured data from a large set of similar properties are shown. Multiple regression techniques were used to identify the relationship between space heating load and behavioural variables. This led to the development of a regression equation which can be used to estimate the likely space-heating requirements of a household given particular behavioural variables, and to the test the impact of certain behaviours on annual heating energy demand. It is found that in general passive houses are less sensitive to behaviour than anticipated.

It is known that building characteristics have a significant effect on energy consumption; and governments worldwide have introduced regulations and policies in a bid to improve the energy performance of fabric and systems within the built environment [1] and [2]. The development of non-mandatory energy efficiency schemes has been improving on these mandatory performance requirements by considerable amounts; in terms of reducing space heat demand, none have been so successful as Passivhaus Certification [3]. To meet the low space heating and cooling limits of no more than 15 kWh/m2a that such a philosophy demands, the building fabric has to be constructed to maximise insulation, and to minimise thermal bridging and uncontrolled air exchange. This can be challenging within the design phase, and even more so during construction. One issue with estimating energy consumption based on a specific design philosophy is if the final in-use performance deviates from the expected performance the design can be seen as failed. In reality this may just be an example of a particular user's consumption being above average. In addition, when energy systems are sized for an individual dwelling, or for clusters of dwellings powered over a district-wide scheme, there is a probability that the system will not be able to meet demand if the cluster has above average demand. This is likely to cause the greatest issues when renewable energy systems have been sized to match the expected demand and the number of dwellings is small (the smaller the number of dwellings, the greater the likelihood of a significant deviation from the mean). Although on-going improvements to system efficiency, materials, and construction methods have significantly reduced the amount of energy used for space heating [4] and [5], studies have indicated that as buildings become more energy efficient, the behaviour of occupants play an increasingly important role in consumption [6], [7], [8] and [9]. Homes are a particularly indeterminate energy sink, as a number of factors make accounting for consumption difficult; firstly there is the wide-ranging variety of structure, sizes, and materials found in the UK housing stock; additionally privacy issues limit the collection of data, and detailed sub-metering of houses has a prohibitive cost on a large scale; also, it is known that much of the discrepancy in energy consumption amongst buildings with similar constructions can be attributed to differences in occupancy patterns and occupant behaviour [10], [11] and [12]. In research by Gill et al., in which a post-occupancy evaluation was carried out on a UK EcoHomes site, the contribution of energy behaviours accounted for 51% of the variance in heating energy use. Other studies have indicated higher levels of behavioural effect of energy consumption, as much as 100% for a given dwelling [13]. Karlsson et al. stresses the importance of building occupant behaviour when designing energy simulations, and also highlights the difficulty of doing so [14]. Wood and Newborough propose that behavioural change is a major untapped area for energy savings, however they argue that the diversity of understanding, attitude, and abilities are a salient barrier to change [15]. Several studies have since investigated the effect that the type of thermostatic control has on energy use [16], [17] and [18]. Shipworth found that in households with thermostats the mean temperature was generally lower than in dwellings without a thermostat in a study of over 400 dwellings [18]. Furthermore, they found that households with a programmable thermostat kept the heating system on for longer than households with non-programmable thermostats, though the difference was not statistically significant at the p = 0.05 level. The paper is organised as follows: Firstly the varieties of bottom-up models are introduced, the methodology then describes our modelling approach, beginning with a look at the stochastic behavioural representation methods and is followed by a justification of the thermal modelling tools used. To conclude the methodology, a discussion of the data interrogation techniques used, namely linear multiple regression models. The results are then presented and validated against expected results gathered from various sources, and a regression model is employed to analyse the results from the modelling. Finally, we discuss the salient points of our study and suggest future work in the area.

It has been shown that it is possible to generate varied and representative models of typical UK occupancy patterns and appliance-use behaviour in homes, and to use these within a modelling environment. The simulation of households has been shown to reproduce measured data surprisingly well. Due to the need to create such as large number of profiles the method is clearly not suited to general use unless it is embedded within the interface of industry standard software. However, there is no reason that this could not be done in some form and the resulting technique would go some way to reducing the credibility gap that models and modellers suffer from. In order that others need not complete multiple runs with a spectrum of behaviours for any and every project, and to allow the generation of ‘rules of thumb’ about the sensitivity of Passivhaus designs to variation in use, a regression model was assembled. In the regression analysis set-point temperature, appliance-use, and airflow behaviour were shown to be the major estimators of total heating energy. Occupancy patterns were shown to be less significant factors. The regression model is limited to a single architectural design and hence it would be useful to examine how much the values of its coefficients change for other Passivhaus dwellings. However it is clear that many of the concerns that some have voiced about the Passivhaus approach being overly sensitive to occupant behaviour and therefore not applicable to many sections of society are unfounded.